LEAD-ZINC 2000

LEAD-ZINC 2000 Proceedings of the Lead-Zinc 2000 Symposium which was part of the TMS Fall Extraction & Process Metallurgy Meeting, Pittsburgh, U.S.A., October 22-25,2000

Edited by

J.E. Dutrizac, B.A.Sc, M.A.Sc, Ph.D., F.C.I.C, F.C.I.M. CANMET Ottawa, Ontario, Canada

J.A. Gonzalez, B.Sc, M.Sc, Ph.D. Cominco Research Trail, British Columbia, Canada

D.M. Henke, B.Sc. Doe Run Company Herculaneum, Missouri, U.S.A.

S.E. James, B.Sc, M.Sc. Big River Zinc Corporation Sauget, Illinois, U.S.A.

A.H.-J. Siegmund, Dipl.-Ing., Dr.-Ing. RSR Technologies Inc. Dallas, Texas, U.S.A.

A publication of

A Publication of The Minerals, Metals & Materials Society 184 Thorn Hill Road

Warrendale, Pennsylvania 15086-7528 (724) 776-9000

Visit the TMS web site at http://www.tms.org

The Minerals, Metals & Materials Society is not responsible for statements or opinions and is absolved of liability due to misuse of information contained in

this publication.

Printed in the United States of America Library of Congress Catalog Number 00-107368

ISBN Number 0-87339-486-0

Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by The Minerals, Metals & Materials Society for us-ers registered with the Copyright Clearance Center (CCQ Transac-tional ReportingService, provided that the base fee of $3.00 per copy is paid directly to Copyright Clear-ance Center, 27 Congress Street, Salem, Massachusetts 01970. For those organizations that have been granted a photocopy license by Copyright Clearance Center, a separate system of payment has been arranged.

If you are interested in purchasing a copy of this book, or if you would like to receive the latest TMS publica-tions catalog, please telephone 1-800-759-4867 (U.S. only) or 724-776-9000, EXT. 270.

TUS

PREFACE

The Lead-Zinc 2000 Symposium is the fourth in the series of decennial conferences on the processing of lead and zinc; it is organized by the Minerals, Metals and MaterialsSociety (TMS) and is co-sponsored by the Australasian Institute of Mining and Metallurgy (AusIMM), the Canadian Institute of Mining, Metallurgy and Petroleum (CIM), the Gesellschaft fur Bergbau, Metallurgie, Rohstsoff-und Umwelttechnik(GDMB), the Institution of Mining and Metallurgy (IMM), the Mining and Materials Processing Institute of Japan (MMIJ), the International Lead Zinc Research Organization (ILZRO) and the International Zinc Association (IZA). The pioneering first decennial conference was held in St. Louis in 1970 and emphasized global operations in both the mining and processing sectors. The second conference, held in Las Vegas in 1980, reviewed both the practical and fundamental aspects of the production of lead, zinc, and also tin. The third meeting in Anaheim in 1990 discussed the theory and practice of the production of lead and zinc. As a reflection of the changing global situation of both metals, a significant part of the 1990 conference focused on secondary feeds and environmental considerations. The present Lead-Zinc 2000 Symposium builds on the foundations of those previous meetings, as well as on the achievements of similar events organized by the Canadian Institute of Mining, Metallurgy and Petroleum in Calgary, Canada in 1998, by the Mining and Materials Processing Institute of Japan in Sendai, Japan in 1995 and by the Australasian Institute of Mining and Metallurgy in Risdon, Australia in 1993. Like those previous international events, the Lead-Zinc 2000 Symposium brings togethertheworld'sprocessing, engineering and research communities to discuss the latest developments in the hydrometallurgical and pyrometallurgical processing of lead and zinc.

The Lead-Zinc 2000 Symposium will also honor Dr. T.R.A. Davey for his many contributions to metallurgical science and the industry throughout his distinguished career. The recognition of Dr. Davey at the Lead-Zinc 2000 Symposium is especially appropriate as he played a key role in the organization of the first decennial conference in 1970 and contributed to both the 1980 and 1990 meetings.

The proceedings volume of the Lead-Zinc 2000 Symposium is the culmination of over two years of work that included the preparation of the papers by the authors, as well as the refereeing, proof-reading and indexing by the editors. The proceedings volume contains sixty-four papers which cover all aspects of lead and zinc processing, including the global business trends of the metals, plant operations, new processing installations, emerging technologies and environmental considerations. The development of new and

improved smelting technologies continues because such processes can treat both primary and secondary feeds with the stabilization of many of the associated impurities in slag form. Innovative hydrometallurgical processes are being developed, and there is an on-going effort to integrate the best of both hydrometallurgy and pyrometallurgy into effective flowsheets which yield lead and zinc at low cost and in an environmentally friendly manner. Accordingly, it is the editors' sincere hope that this proceedings volume will remain a valuable record of the Lead-Zinc 2000 Symposium and that it will become a standard reference for the processing of lead and zinc.

The production of this proceedings volume was a major undertaking, and many individuals were involved over the course of several months. Accordingly, the editors would like to extend their sincere appreciation to Dave Hardy, Marilyn Harris, LomaPaquette and Hariranja Rakotoarimanga at CANMET, and to Dan Ashman, Lida Gambin and Juris Harlamovs at Cominco Research. The editors genuinely appreciate their assistance with the refereeing and/or editing of the various papers, and we thank them for their assistance in the production of the proceedings volume of the Lead-Zinc 2000 Symposium.

Pittsburgh J.E. Dutrizac October 2000 J.A. Gonzalez

D.M. Henke S.E. James

A.H.-J. Siegmund

EDITORS' BIOGRAPHIES

John E. Dutrizac attended the University of Toronto where he received his B.A.Sc. degree in metallurgical engineering in 1963 and his Ph.D. degree in 1967. Upon graduation, he worked for a short time at the Noranda Technology Centre in Pointe Claire, Quebec. In 1968, he joined CANMET as a Research Scientist, and began to study a variety of hydrometallurgical problems. His current research activities are centred on zinc processing with its associated iron precipitation problems, the leaching of sulphide minerals, and the characterization of the anodes and anode slimes encountered in copper electrorefining. These efforts commonly involve the integration of chemical metallurgy with the techniques of applied mineralogy to improve the understanding of the process. Dr. Dutrizac has over two hundred publications in these and in related fields. John Dutrizac is a former Chairman of the Hydrometallurgy Section, and a Past President of the Metallurgical Society of CIM. He has received many international awards, and is a Fellow of the Chemical Institute of Canada and the Canadian Institute of Mining, Metallurgy and Petroleum.

Jose Alberto Gonzalez was born in Mexico City in 1960. He received B.Sc. and M.Sc. Degrees in Metallurgy from the University of Mexico (UNAM) in 1984 and 1985, respectively. He then pursued Ph.D. studies at the University of British Columbia (UBC) in the area of lead electrorefining, under the supervision of Prof. Ernest Peters. After the completion of his Ph.D. degree he joined Cominco in 1991, where he is currently Principal Research Scientist and heads the Electrometallurgy Section in Cominco Research. Since 1997, he has been on the editorial board of the Hydrometallurgy Journal. In 1994, he became an Adjunct Professor at UBC where he has been involved in Electrometallurgy and Hydrometallurgy Chair research activities. His research focuses on developing energy efficient ways to electrowin zinc and copper and on optimizing the winning and refining of lead from fluosilicate electrolytes.

Daniel M. Henke grew up in Wyoming, U.S.A. and attended college at the South Dakota School of Mines and Technology in Rapid City, South Dakota. Daniel earned a B. Sc. degree in Metallurgical Engineering in 1979. Upon graduation, he joined St. Joe Minerals at the primary lead facility in Herculaneum, Missouri. In 1986, St. Joe Minerals became the Doe Run Company. The many positions he has held over the past 21 years with the Doe Run Company include: Assistant Superintendent Blast Furnaces, SuperintendentBlast Furnaces, SuperintendentRefinery, Production Manager, Facility Manager. His current position is Operations Manager of the products and technical service departments. The managerial responsibilities of the products area include the refining,

vii

alloying, and casting of the lead products, shipping of all the finished goods, and production of lead strip and sheet products. The responsibilities of the technical services group include supporting customer service, directing all quality programs, and supporting other operational processes.

Steven E. James has worked for over 20 years in the United States zinc industry at three different operating plants. A graduate of the Colorado School of Mines in Metallurgical Engineering, he began his career as a Research Engineer for the St. Joe Minerals Corporation in Monaca, Pennsylvania. Among other activities, Steve served as the key process engineer for a lead hydrometallurgical pilot plant during a successful 2-year operation. After organizing and leading the hydrometallurgy research group for St. Joe Minerals, he assumed a position as a process engineer at the National Zinc Division of St. Joe Resources Co. in Bartlesville, Oklahoma. He served in a variety of operating positions there, eventually becoming production manager for the electrolytic zinc plant under its new owners, The Zinc Corporation of America. Zinc production and worker safety improved every year during his tenure as production manager. In 1991, Steve joined the Big River Zinc Corporation in Sauget, Illinois as the Director of Technology, a role he still fills. Steve is a member of The Minerals, Metals and Materials Society, The Mining History Association, The Canadian Institute of Mining, Metallurgy and Petroleum, The North American Zinc Processors, and The American Society for Quality. He is both a past chairman and incoming chairman of the Lead-Zinc-Tin Committee of The Minerals, Metals and Materials Society.

Andreas H.-J. Siegmund was born and raised in Hanau, Germany. After graduation from high school in Hanau, he enrolled in metallurgical engineering at the Technical University of Berlin, Germany where he received the degree of Dipl.-Ing. in 1985. He then worked in the laboratory of Preussag AG in Goslar, Germany in a joint research project between Preussag AG and the Technical University of Berlin. In 1989, he received the degree of Dr.-Ing. from the Technical University of Berlin for work in the area of the electrorefining of ultra-pure cadmium. After completion of his Doctoral degree, he joined Lurgi Metallurgie GmbH were he was involved in the commissioning of all QSL plants for lead smelting, and progressed through various responsibilities to Head of the Non-ferrous Process Department. Since 1998, he is Manager Research and Development at RSR Technologies, Inc. in Dallas, Texas focussing on projects for lead recovery from secondary and primary sources and working with anodes for Cu- and Zn-electrowinning.He is an active member of the Lead Committee of the GDMB and is currently Chairman of the Pb-Zn Committee of TMS.

viii

LEAD-ZINC 2000 ORGANIZING COMMITTEE

B. Blanpain Katholieke Universiteit Leuven Metallurgy and Materials Engineering W. de Croylaan 2 B-3001 Leuven Belgium

J.E. Dutrizac CANMET 555 Booth Street Ottawa, Ontario Canada K1AOG1

J.A. Gonzalez Cominco Research P.O. Box 2000 Trail, British Columbia Canada V1R4S4

D.M. Henke Doe Run Company Herculaneum Smelting Division 881 Main Street Herculaneum, Missouri U.S.A. 63048

S.E. James Big River Zinc Corporation 2401 Mississippi Avenue Sauget, Illinois U.S.A. 62201

J.F. Pusateri Horsehead Resource Development Technical Centre 300 Frankfort Road Monaca, Pennsylvania U.S.A. 15061

V. Ramachandran Asarco Inc. Technical Service Centre 3422 South 700 West Salt Lake City, Utah U.S.A. 84119

ix

M.A. Reuter Delft University of Technology Applied Earth Sciences Mijnbouwstraat 120 2628 RX Delft The Netherlands

A.H.-J. Siegmund RSR Technologies Inc. 2777 Stemmons Freeway Suite 1800 Dallas, Texas U.S.A. 75207

E. Tamargo Asturiana de Zinc S.A. Cardenal Marcelo Spinola 42, 8° 28016 Madrid Spain

Y. Umetsu Institute for Advanced Processing Tohoku University 1-1 Katahira, 2-chome Aoba-ku, Sendai Japan 980-8577

P. Wollants Katholieke Universiteit Leuven Metallurgy and Materials Engineering W. de Croylaan 2 B-3001 Leuven Belgium

SESSION CHAIRS

Session 1 - Global Factors Affecting Lead and Zinc

C.Y. Choi Korea Zinc Company Ltd. 142 Nonhyon-Dong Gangnam-ku, Seoul Korea

D. Magoon Cominco Limited Trail, British Columbia Canada V1R4L8

Session 2 - Modern Lead Smelting Technologies I

P.J. Moor Britannia Refined Metals Botany Road Northfleet, Kent DAI 1 9BG England

V. Ramachandran Asarco Service Center 3422 South 700 West Salt Lake City, Utah U.S.A. 84119

Session 3 - Zinc Operations I

KH.O. Meyer Zinc Corporation of South Africa Plover Street, Struisbult P.O. Box 218, Springs 1560 South Africa

R. Menge Metaleurop Weser Zinc Johannastrasse 2 26954 Nordenham Germany

Session 4 - Modern Lead Smelting Technologies II

N.L. Piret Piret & Stolberg Partners Im Licht 12 D-47279 Duisburg Germany

D.M. Henke Doe Run Company 881 Main Street Herculaneum, Missouri U.S.A. 63048

Session 5 - Zinc Operations II

C. Canoo UM Zinc Zinkstraat 1 B-2490 Balen Belgium

M. Agnew Noranda Inc., CEZinc 860 Cadieux Boulevard Valleyfield, Quebec Canada J6S4W2

Session 6 - Imperial Smelting Technologies

J.F. Pusateri Zinc Corporation of America 300 Frankford Road Monaca, Pennsylvania U.S.A. 15061

Y. Umetsu Tohoku University 1-1 Katahira, 2-chome Aoba-ku, Sendai Japan 980-8577

Session 7 - Zinc Electrowinning

P. Wollants Katholieke Universiteit Leuven W. de Croylaan 2 B-3001 Leuven Belgium

T.J. O'Keefe University of Missouri-Rolla 1870 Miner Circle Rolla, Missouri U.S.A. 65409

M.A. Reuter Delft University of Technology Mijnbouwstraat 120 2628 RX Delft The Netherlands

J.A. Gonzalez Cominco Research P.O. Box 2000 Trail, British Columbia Canada V1R4S4

G.L. Bolton Dynatec Corporation 8301-113 Street Fort Saskatchewan, Alberta Canada T8L4K7

S.E. James Big River Zinc Corporation 2401 Mississippi Avenue Sauget, Illinois U.S.A. 62201

Session 8 - New Developments in Lead and Zinc

A.H.-J. Siegmund D. Ashman RSR Technologies Inc. Cominco Research 2777 Stemmons Freeway P.O. Box 2000 Dallas, Texas Trail, British Columbia U.S.A. 75207 Canada V1R4S4

Session 9 - New Zinc Processing Technologies

A. Malsnes B. Blanpain Norzink Katholieke Universiteit Eitrheim Leuven N-5751 Odda W. de Croylaan 2 Norway B-3001 Leuven

Belgium

Session 10 - New Electrowinning Technologies for Lead and Zinc

Session 11 - Environmental Aspects of Lead and Zinc Production

Xl l

TABLE OF CONTENTS

Preface v

Editors' Biographies vii

Organizing Committee ix

Session Chairs xi

CHAPTER 1 - GLOBAL FACTORS AFFECTING LEAD AND ZINC

Market Fundamentals and the Evolution of Lead and Zinc Supplies C. Hassall and H. Roberts 3

Lead Product Development in the Next Millennium R.D. Prengaman 17

Zinc Applications: A World of Performance E. Gervais 23

Lead and Zinc: A Study of Toxicological Contrasts and Shared Regulatory Concerns C.J.Boreiko 39

CHAPTER 2 - MODERN LEAD SMELTING TECHNOLOGIES I

Primary Lead Production - A Survey of Existing Smelters and Refineries A. H.-J. Siegmund 55

Operations at the Doe Run Company's Herculaneum Primary Lead Smelter N. D. Schupp 117

Modern Lead Smelting at the QSL-Plant Berzelius Metall in Stolberg, Germany R. Püllenberg and A. Rohkohl 127

A Review of Ausmelt Technology for Lead Smelting E. N. Mounsey and N. L. Piret 149

Cominco's New Lead Smelter at Trail Operations D. W. Ashman, D. W. Goosen, D. G. Reynolds andD. J. Webb 171

xiii

Commissioning and Optimisation of the New Lead and Silver Refinery at the Pasminco Port Pirie Smelter P. Kapoulitsas, M. Giunti, R. Hampson, A. Cranley, S. Gray, B. Kretschmer, R. Knight and J. Clark 187

CHAPTER 3 - ZINC OPERATIONS I

Zinc Production - A Survey of Existing Smelters and Refineries S. E. James, J. L. Watson and J. Peter 205

Use of Data Reconciliation: A Zinc Plant Case Study T.J. Auping, M.A. Reuter, S.C. Grund and K. Born 227

Recent Operations at the Hikoshima Smelter T. Iwamoto, H. AkiyamaandK. Eto 241

Improvements in the Leaching Circuit of IMMSA's Zinc Plant in San Luis Potosi, Mexico P.Alfaro, C. Moctezuma and S. Castro 251

Expansion Plans at CPM's Electrolytic Zinc Refinery T. Takayama, W. Magalhäes, J. Welsh, T. Newton and S.J Thiele 261

The Boleslaw Electrolytic Zinc Plant D. Krupka, B. Ochab andJ. Miernik 277

CHAPTER 4 - MODERN LEAD SMELTING TECHNOLOGIES II

Health and Hygiene in the Modern Lead and Zinc Industry D.N. Wilson 289

Cominco's Trail Operations: An Integrated Zinc-Lead Operation ET. de GrootandD.L. Verhelst 307

The Lead Bath Smelting Process in Nordenham, Germany M. Sibony, N. Basin, J. Lecadet, R. Menge andS. Schmidt 319

QSL Lead Slag Fuming Process Using an Ausmelt Furnace M.B. Kim, W.S. LeeandY.H. Lee 331

Recent Developments in the Lead Refining Operations at Britannia Refined Metals Ltd. P.J.Moor 345

xiv

Analysis of Dust Formation in the Oxygen Flash Smelting of Lead-Bearing Feeds in the Kivcet Smelter M.A. Lyamina, V.A. Shumsky, I.P. Polyakov, N.M. Ospanov andN.N. Ushakov 361

CHAPTER 5 - ZINC OPERATIONS II

Recent Operations at the Iijima Zinc Refinery T. Yamada, S. Kuramochi andR. Togashi 373

Improvements at the Electrolytic Zinc Plant of Big River Zinc Corporation, Sauget, Illinois, U.S.A. S.E. James, L.L. Ludwig and B.L. Motycka 385

Recent Process Improvements in the Kokkola Zinc Roaster J. Nyberg, M.-L. Metsärinta and A. Roine 399

Oxygen Enrichment of Fluo-Solids Roasting at Zincor C. MacLagan, M. Cloete, E.H.O. Meyer and A. Newall 417

Method for Processing Siliceous Zinc Ores S.Ikenobu 427

Leaching and Purification at Cominco's Trail Zinc Operations D.J. McKay, G. Sterzik, T.L. Salway and W.A. Jankola 437

CHAPTER 6 - IMPERIAL SMELTING TECHNOLOGIES

The Continuing Evolution of the Imperial Smelting Process R.W.Lee 455

Impurity Distribution in the ISP Process at the Harima Works of Sumitomo Metal Mining Co., Ltd. O. Kitamura and H. Kubota 467

Optimization of the New Jersey Refining Process G. Hanko, A. Lebleu, M. SibonyandJ. Lecadet 481

A Study of the Changes in the Permeability of the Sintering Bed in the Imperial Smelting Process K. Kawanaka and Y. Mori 497

Changes in the Physical and Mechanical Properties of SiC Trays Caused by Ageing in Zinc Refinery Operations A. Piant, M. Fritz, M. Boussuge andM.-D. Dupuits 511

XV

Development of Mechanization Facilities in the Non-Ferrous Metallurgical Field Y. Sako, K. Muraguchi, K. ShibataandE. Nomura 521

CHAPTER 7 - ZINC ELECTROWINNING

Review of Engineering and Construction Factors in Building a Zinc Cellhouse J. G. Cooper, P. Mercille andM. F. Nasmyth 537

Asturiana de Zinc Expansion at the San Juan de Nieva Plant for a Zinc Production of 440,000 Tonnes per Year F. San Martin, F. Tamargo and Y. Lefivre 555

Zinc Autostripping at Falconbridge Limited Kidd Metallurgical Division J. Lenz and D. Ducharme 563

Cellhouse Ventilation J.A. Davis and J. de Visser 579

New Wrought Pb-Ag-Ca Anodes for Zinc Electrowinning to Produce a Protective Oxide Coating Rapidly R. D. Prengaman and A. Siegmund 589

Mechanical Properties and Electrolytic Behavior of Pb-Ag-Ca Ternary Electrodes for Zinc Electrowinning Y. Takasaki, K. Koike andN. Masuko 599

CHAPTER 8 - NEW DEVELOPMENTS IN LEAD AND ZINC

Equilibrium versus Kinetics in Lead Refining T.R.A.Davey 617

Direct Zinc Smelting in an Iron Oxysulfide Bath R.-Q. Li, J.G. Peacey and P. Hancock 637

The Influence of New Technology at Sulphide Ore Mine Sites on Metals Production and Recoveries, with its Commercial Significance H. Fletcher and P. Gray 659

Recovery of Zinc and Cadmium from Lead Smelter Furnace Dusts at Met-Mex Pefioles by a Solvent Extraction Process I. S. Fernandez del Rio 677

xvi

The Kivcet Treatment of Polymetallic Feeds L. V.Slobodkin, Yu.A. Sannikov, Yu. A. Grinin, M.A. Lyamina, V.A. Shumsky andN.N. Ushakov 687

A Mathematical Model for the Electric Furnace in the Kivcet Process G.S. Hanumanth and G.A. Irons 693

CHAPTER 9 - NEW ZINC PROCESSING TECHNOLOGIES

Atmospheric Leaching of Zinc Sulphide Concentrates Using Regenerated Ferric Sulphate Solutions C.J.Ferron 711

Comparison of Direct Pressure Leaching with Atmospheric Leaching of Zinc Concentrates K.R. Buban, M.J. Collins, I.M. Masters andL.C. Trytten 727

Treatment Of Secondary Zinc Oxides For Use In An Electrolytic Zinc Plant S.S. Chabot andS.E. James 739

Upcoming Zinc Mine Projects: The Key for Success is Zincex Solvent Extraction M.A. Garcia, A. Mejias, D. Martin and G. Diaz 751

The Galvanic Stripping Treatment of Zinc Residues for Marketable Iron Product Recovery J. A. Barrera-Godinez, J. Sun, T.J. O'Keefe and S. E. James 763

Separation of Iron from a Zinc Sulphate Electrolyte by Combined Liquid-Liquid Extraction and Electro-Reductive Stripping K. Verbeken, M. Verhaege and E. Wettinck 779

CHAPTER 10 - NEW ELECTROWINNING TECHNOLOGIES FOR LEAD AND ZINC

New Clean Technologies to Improve Lead-Acid Battery Recycling C. Frias, M. Garcia and G Diaz 791

Electrowinning of Lead Battery Paste with the Production of Lead and Elemental Sulphur Using Bioprocess Technologies M. Olper, M. Maccagni, C.J.N. Buisman andC.E. Schultz 803

Performance of a Conventional Cell Design for Zinc Chloride Electro winning C.Allen 815

Spouted Bed Electrowinning of Zinc from Chloride Electrolytes J. W. Evans, A. Roy and C. Allen 831

The Effect of Microstructure on the Electrochemical Behavior of Lead-Silver Alloy Anodes during Zinc Electrowinning S. Jin, E. Ghali, G. St-Amant, V. Cloutier and G. Houlachi 845

Effect of Polymer Additives on Zinc Electrowinning T. Ohgai, H. Fukushima, N. BabaandT. Akiyama 855

CHAPTER 11 - ENVIRONMENTAL ASPECTS OF LEAD AND ZINC PRODUCTION

Effluent Treatment at the Pasminco Clarksville Zinc Plant S. Subhawong 867

Development, Testing and Full-Scale Operation of a New Treatment Method for Selenium Removal from Acidic Effluents G.A. Monteith, G. Houlachi, M. Pineau and M. Laliberte 879

Recovery of Sulfides from Sulfate-Containing Bleed Streams using a Biological Process C.F.M. Copini, G.H.R. Janssen, C.J.N. Buisman andS. Vellinga 891

Goethite: From Residue to Secondary Building Material - Union Miniere's Graveliet® Process J. Winters, L. VosandC. Canoo 903

A Mineralogical Study of Jarofix Products for the Stabilization of Jarosite Residues for Disposal T.T. Chen and J.E. Dutrizac 917

Disposal of a Lead Sludge

R.J. Wesely 935

Author Index 945

Keyword (Subject) Index 949

xviii

Chapter 1

Global Factors Affecting Lead and Zinc

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 3

MARKET FUNDAMENTALS AND THE EVOLUTION OF LEAD AND ZINC SUPPLIES

C. Hassall and H. Roberts CHR Metals Limited

Hamble House, Meadrow Godalming, United Kingdom GU7 3HJ

ABSTRACT

The opening years of the new millennium will be characterised by the closure of many lead and zinc mines because of reserve depletion. However, weak commodity prices through the second half of the 1990s have slowed project evaluation and have made it difficult to raise finances for the development of new mines required to replace lost output and to meet increasing demand for metal in the future. Growth of lead and zinc demand will be determined by the pace of global economic activity, especially in the emerging economies. How rapidly demand grows will have a key bearing on the possible shortfall in mine output and the response of prices to the implied tightening of metal supplies early in the new millennium. These issues are the focus of a discussion of the critical lead and zinc supply developments which will unfold post 2000, taking into account project lead times and the role that recycled metal plays in meeting market demand.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

4 LEAD-ZINC 2000

LEAD AND ZINC SUPPLIES

In this paper we will examine the evolution of lead and zinc supplies, the market factors influencing these supplies and will then seek to draw some conclusions from this analysis about future prospects for both industries.

In looking at the evolution of supply, we take into account the natural links between the lead and zinc industries, which are determined by the fact that their respective minerals are often found and exploited together. We shall also examine those developments that have made this natural linkage far less important today than was the case 30 or 40 years ago.

First of all, we need to clarify some definitions. With the collapse of communism in Eastern Europe and the Soviet Union and the opening up of China, there is a clear need now to look at developments in the lead and zinc industries in global terms; i.e., including every country in the world. However, in looking at historic development trends, one encounters the problem of consistency and accuracy of reporting in the former communist countries, particularly the Soviet Union, as well as continuing problems in the collection and collation of data in some of these countries and in China. For this reason, we continue to make the distinction between the "transitional" economies, which we define as China, North Korea, Vietnam, Cuba, the Former Soviet Union (FSU) and the Warsaw Pact countries of Eastern Europe, and the Western World which comprises all other countries. Much of the subsequent discussion refers solely to the Western World but, where possible, we also quote global numbers and/or numbers for the transitional economies.

As shown in Figure 1, in 1960 Western World production of lead and zinc was 2.5 Mt and 2.4 Mt, respectively. Based on average prices in that year, the output of the lead and zinc industries together was worth a little over $5 billion in today's terms. Unfortunately, we do not have very reliable figures for lead and zinc production in the transitional economies of 40 years ago. The best estimates are those produced by the International Lead and Zinc Study group that show that "socialist" countries produced 0.72 Mt each of lead and zinc. (Following the break-up of the Soviet Union it became apparent that lead and zinc production may well have been over-stated in the USSR, but it is still difficult to determine what output might have been as long ago as 1960.) If we accept these numbers, then this implies that the output of the global lead and zinc industries in 1960 was worth around $6.8 billion.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 5

1960 1999 1960 1999

Figure 1 - Western World Refined Lead and Zinc Output: Volume and Value (Source: ILZSG, CHR Metals)

Returning again to a discussion limited to the Western World, it is worth noting that, in 1960, the value of lead and zinc output taken together accounted for 18% of total non-ferrous production (Al, Cu Pb, Zn, Sn and Ni), with zinc's share standing at 10% and lead's at 8%. How have matters changed? In the last year of the 20th Century, Western World lead production was 4.9 Mt and zinc 5.8 Mt. Now that we have much greater faith in the data flowing from the transitional economies, we estimate that global production of lead and zinc in 1999 was 6.2 Mt and 8 Mt, respectively. Compared with 40 years ago these totals are 93% higher for lead and 153% higher for zinc. Turning again, for a moment, to consider just the Western World, the value of zinc output in 1999 was 120% higher in real terms compared with 1960. For lead the increase was only 5%. Overall their combined share of Western World non-ferrous output was a little smaller than in 1960 at just over 16%. However, zinc increased its share of the total to nearly 12% while lead saw its share fall to only 4.5%.

Clearly much has changed, therefore, in the relative positions of these two metals over 40 years, both in terms of the balance between lead and zinc output and the position of these metals in the global non-ferrous industry. What have been the fundamental issues affecting the demand for these metals which have had such a significant impact on their relative performance, both in terms of output growth and the actual value of production?

For all metals, the most important determinant of demand growth is the pace of economic activity. Usually a measure of industrial production is taken as the most appropriate indicator of this when looking at metal demand. The second factor to examine is the intensity of metal use. In other words how many tonnes of metal are consumed for each unit of industrial output? This key element in determining metal demand reflects, in broad terms, the uses to which a metal is put. The intensity of use value varies from country to country depending on the structure of industrial output. It also varies over time as individual economies pass through different stages of industrialisation and development and increasing per capita

6 LEAD-ZINC 2000

incomes, leading finally to a stage where the service sector becomes the main area of increases in economic activity. In addition, intensity of use is affected by substitution of one material for another, changes in design of products and, increasingly important at the end of the 20th

Century, environmental concerns.

Looking first at lead demand we can see that there have been huge changes in the end-use patterns of consumption over the past 40 years. In 1960 lead-acid batteries accounted for a little under 30% of all uses. Rolled products (lead sheet and pipe and ammunition) had a 21% share of the market and cable sheathing was another major end use with an 18% share. Lead also found significant applications in chemicals (pigments, oxides and tetra-ethyl lead) and alloys. At the end of the century batteries had come to dominate all other end uses, accounting for almost three-quarters of all consumption. Use in pigments and compounds had a 10% share while rolled products represented only 8% of the market. All other uses were relatively insignificant.

This dramatic change in the market for lead had one over-riding cause. This was the appreciation of lead's toxicity in the general environment and, in particular, the effect of exposure to lead on children's mental development. This meant that all uses of lead came under very close scrutiny with many countries taking decisions to restrict severely the use of lead in applications such as paints, pipes, alloys and chemicals, including tetra-ethyl lead used as an additive in gasoline. Thus, while one end use, the requirement for lead in batteries, has seen a significant increase in absolute levels of consumption, reflecting the ever-increasing number of vehicles on the roads and the growing market for industrial batteries, the demand for lead in virtually all other applications has fallen. In fact, the apparently poor performance of lead consumption growth, especially in more recent years, masks very impressive growth in demand for lead in batteries. Over the past 15 years annual consumption growth in batteries has averaged a very respectable 3.9%. This contrasts with only a 2% per annum value for total consumption. In absolute terms there was a 350 kt decline in Western World offtake for all non-battery application between 1984 and 1999. Over the same period, lead use in batteries grew by more than 1.75 Mt. Further significant demand growth in lead use in batteries can be expected for the foreseeable future. While the pace of growth in vehicle fleets has slowed in the mature economies, a large number of new cars, trucks and buses are being added each year to the total of vehicles on the road in the emerging and transitional economies. Furthermore, the infrastructure required to support the rapid spread of mobile telecommunications and the need for back-up power supply systems for electronic equipment everywhere have created important growth markets for lead-acid industrial batteries. The likely increase in lead offtake will have to be met principally from primary sources, in other words, from mine production.

For zinc there have also been some important shifts in the patterns of consumption, but these have not been as significant as for lead. However, it is the case that most of the growth in demand has occurred in one end use. In common with lead, environmental factors have also played a major role in determining the use of zinc in some applications. However, the issue was not toxicity, but rather weight. In the 1960s and early 1970s many components in cars were made from zinc die-castings. These included some structural elements as well as parts such as carburettors and exhaust manifolds. The first oil crisis of 1973-1974 forced automakers to redesign their vehicles to improve fuel efficiency. Reducing the weight of cars was an important step in this process. As a result there was great incentive to re-engineer parts and to substitute heavier metals by lighter metals (aluminium for zinc) and plastics. The die-casting industry responded to this threat by introducing new technology to produce lighter components and developed thin-wall die-casting. While this meant that zinc continued to be used for some parts in cars, the per-unit use of zinc fell. Also this same technology found applications outside of the auto industry, causing a further erosion of zinc offtake. Overall, as a result of direct

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 7

substitution and the use of thin-wall die-casting techniques, zinc consumption in this single application has still to regain the levels achieved in 1973.

Another end-use, which has suffered because of weight considerations in the auto industry and substitution by plastics, has been the use of zinc in brass. While the impact of these developments was not as severe as in the case of die-casting, it was not until 1994 that the consumption of brass surpassed the levels first reached in 1973.

Zinc's major end use, galvanising, has seen its Western World market share increase from just under a third of all consumption in 1960 to around 50% today. Somewhat ironically a significant boost to the use of galvanised sheet came from the auto industry. Pressure from consumers and competition to produce a better product forced the auto companies to adopt the widespread use of galvanised sheet in the construction of cars in order to offer warranties against corrosion. This was especially evident in the growth of zinc demand in the 1980s in the USA and Japan and in the late 1980s/early 1990s in Europe. In many of the emerging economies carmakers have yet to switch to the extensive use of galvanised sheet. This will happen in due course and will provide further gains for zinc demand. Although much attention has focussed on the use of galvanised sheet in the auto industry, it is construction that actually consumes more zinc, not just in galvanised products, but also in the form of brass and zinc sheet. This latter product is particularly important in mainland Europe as a roofing and cladding material. In addition, zinc-coated sheet finds many applications in general construction where it is used as sacrificial shuttering for concrete floors, cable trays, air conditioning ducting, etc. It is these uses that have been responsible for the strong consumption growth rates for zinc that were evident in Asia for much of the 1990s.

Western World growth in zinc consumption in galvanising over the past fifteen years has amounted to 1.7 Mt representing an annual average rate of increase of almost 4%. All other uses together have increased by 0.7 Mt at an annual average growth rate of 1.4%. Thus, in common with lead, growth in zinc demand has largely reflected gains in a single end use. However, in the case of zinc, this single major application accounts for about 50% of total zinc demand and all other applications are still exhibiting some growth, albeit relatively slow in some instances. For the foreseeable future galvanising will remain the principal area of demand growth for zinc, with the best performance almost certainly to be in the emerging/maturing economies, especially in Asia. Infrastructure and other construction activity will underpin these growing markets for zinc.

Before we can consider how the market fundamentals have affected the evolution of lead and zinc supplies, there are a number of other aspects of the market equation that need to be considered. One factor that has played a critical role in determining the characteristics of lead supplies has been the increasing availability and recycling of lead scrap and the treatment of other secondary materials to produce refined lead. Lead is unusual amongst the non-ferrous metals in that the life cycle of the product in which it finds its principal end use is rather short (typically 2 to 6 years for SLI batteries and 6 to 8 years for industrial batteries). Furthermore, batteries contain the metal well and can be collected relatively easily at the end of their useful lives. Finally, there is almost no dissipation of the lead used in batteries during normal working conditions. These factors taken together mean that lead has the highest rate of recycling of any metal. Losses reflect the rather small proportion of batteries not recycled and the inevitable loss of some lead during processing.

Given that the share of the market for lead accounted for by batteries has been increasing steadily, it is no surprise that an increasing proportion of lead supply has been met from secondary sources. Fifteen years ago, secondary sources accounted for only 48% of

8 LEAD-ZINC 2000

Western World lead production. By 1999 this share had increased to over 61%. Perhaps more strikingly, Western World secondary output has grown at an average annual rate of 2.8% since 1984. Primary output (from lead and bulk concentrates) has barely risen at all. We should note that, in our calculation of secondary production, we are including the treatment of secondary materials in "primary" smelters as well as production at dedicated secondary operations. Although it has always been the case that some residues from other metallurgical plants and industrial wastes have been processed in primary smelters, there has been a sharp increase in the volume of non-concentrate feed processed by the primary sector in recent years. The initial spur to this was an acute shortfall in lead concentrates in the years 1993/1994. However, tighter environmental legislation introduced during the 1990s also meant that more lead-containing industrial wastes were offered to lead smelters as a cheaper option to expensive land fill, and the opening up of the Eastern European economies and the former Soviet Union exposed a legacy of waste materials which could be treated economically by lead smelting operations in the West.

The fact that the lead market is now so dependent upon secondary materials to meet demand, with the bulk of scrap derived from batteries, means that there ought to be a self-regulating mechanism to balance lead supplies against demand. If demand is growing strongly, this suggests that sales of replacement batteries are buoyant. And for each replacement battery sold a battery is scrapped. So long as the recycling chain is efficient (which it now is in most mature economies although it is still susceptible, to a degree, to price pressures), growth in replacement battery sales should also be accompanied by an increase in secondary production. The reverse situation should also be true with a sluggish battery market leading to a tightening in scrap supplies. While this theory is fine and could, perhaps, have been seen operating during the second half of the 1980s, developments in the 1990s have upset all earlier assumptions. Figure 2 above shows that there was an absolute decline in secondary refining in the early 1990s and again towards the end of the decade. It is probable that low lead prices played a major role in this. When prices recovered in the mid-1990s, so too did the availability of lead scrap, though weak prices in recent years have again caused the volume of scrap collected to fall. However, a more critical factor has been at work in causing the breakdown in the theoretical natural balance between supply and demand that was seen emerging in the West during the 1980s, as batteries became the dominant end use of lead. This was the sudden appearance of new supplies of primary lead in the early 1990s with exports to the West from the FSU and Eastern Europe and, in more recent years, from China.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 9

4500 1

■ Lead consumption in batteries 4000

Q Refined lead from secondary 350° " materials

3000 "1 - j ~i ~-

2500 - , - i

s Ί 1 1 1 2000 -I -I

1500

1000

500

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Figure 2 - Lead Consumption in Batteries and Volume of Lead Recycled (Source: ILZSG, CHR Metals)

Before discussing further the impact of the break-up of the Soviet Union and the emergence of China as a force in the lead and zinc industries, we need to consider also the role of secondary production in zinc output. Unlike lead, most of the products using zinc have a relatively long Hfe cycle and, in some cases, the zinc used tends to dissipate over the product's useful life. Even where the zinc remains in situ, its economic recovery may be difficult. There has been much attention given to recovering the zinc used to galvanise steel. Where zinc coated steel scrap is smelted in an electric arc furnace, the dust captured is increasingly being treated and the zinc recovered. EAF dust is formally classified as a hazardous waste in much of the world today, but whether it is recycled or not depends on a range of factors, including the relative costs of recycling versus the alternatives (if permitted) of hazardous waste disposal at an approved site and/or stabilisation in a manner that complies with national standards. Some EAF dust is taken directly into the zinc smelting industry (with, for instance, some Imperial Smelting furnaces using such material directly) but the majority is pre-treated, typically though not exclusively in Waelz kilns, before being sold as a crude zinc oxide to both ISFs and to electrolytic smelters. It is expected that EAF dust will become a much more significant source of zinc-containing feed for the zinc smelting industry in the future, a direct result of a higher proportion of zinc in the steel scrap as more intensively galvanised autos are themselves scrapped. However, zinc produced from secondary sources will remain a relatively small part of overall supply, even though there will continue to be an increase in the rate of recovery.

So far we have described some key characteristics of the lead and zinc markets. Traditionally the two metals have been mined together. Zinc demand has grown at an overall faster rate than lead in recent years. Lead consumption is dominated by a single end use which, in turn, currently depends largely on a single end-use sector, vehicles; other end uses for lead are contracting. Zinc also has a major end-use, galvanising, but it is by no means the only end use and demand for its other applications is also growing albeit quite slowly; galvanised steel is used in a wide variety of final end uses. Lead and bulk concentrates in the West now account

LEAD-ZINC 2000

for less than 40% of total lead feed. Secondary output of zinc, although it is growing, remains a relatively small part of overall supply, with zinc and bulk concentrates still accounting for just over 90% of total zinc feed units.

Although we have focussed on developments in the Western World in making our comparisons of the factors that characterise the lead and zinc industries, the situation today is that we have a global market for all commodities, as is suggested by Figure 3. Much has been written about the impact on metal markets of the break-up of the Soviet Union and the integration of China into the world's trading system. In essence, for lead and zinc, the principal shock to the "system" was a sudden and unexpected increase in supply coming into the market, apparently at the bottom end of the cost curve. Certainly for operations in the former Soviet Union (mainly Kazakhstan) the physical assets (mines and smelters) were transferred to new owners at almost no cost and these operations were then subjected to a form of asset stripping for a number of years. In the case of a number of mines that have now closed, previously developed ore was extracted but the very low grades at these operations meant that no one was prepared to make any further investments. For the smelters little was spent on maintenance for a number of years but, more recently, this situation has changed and we are now seeing a recovery in output of both metals in the FSU. This has not, so far, been matched any meaningful rise in domestic consumption.

800

600

400

S 200

0

-200

-400

1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999

Figure 3 - Net East/West Trade in Refined Lead and Zinc (Source: ILZSG, CHR Metals)

China's entry into the world of lead and zinc could not have been more different from the circumstances in the other countries in the former Eastern Bloc. Prior to the mid-1980s China's refined lead production was no more than 225 kt/a and zinc output was around 300 kt/a. From the mid-1980s, however, there was, a sharp increase in mine production of both metals that, initially, was much greater than the rise in refined output. The development of new mines was the direct result of official policy to promote the domestic lead and zinc industries to

■Zinc

E|_ead

West net exoorter to the East

p B *3" West net importer from the "East"

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 11

forestall a significant dependence on imports of these metals from the West. So successful was the policy that China soon found itself facing a glut of lead and zinc concentrates. This prompted the next phase of development that saw a very rapid expansion of smelting and refining capacity, often in very small units. Thus, from being a relatively modest producer in the 1980s, China emerged to become the world's largest primary producer of both lead and zinc in the early 1990s. This obviously had a profound impact on the markets, especially as China's consumption growth failed by a considerable margin to match the increases in production.

The collapse of lead and zinc demand in Eastern Europe and in the former Soviet Union, and the rapid build-up of significant production capacity in China, meant that the transitional economies became net exporters to Western markets whereas previously they had been net importers. With little reliable information available about the situation in China, especially, and at best only partial knowledge about likely developments in the former Soviet Union and Eastern Europe, Western producers were at something of a loss to know how to respond to the changed circumstances. With hindsight it can be seen that a rapid rationalisation of capacity in the West was perhaps required. However, in the first half of the 1990s, it was far from clear how long the transitional economies would continue to be net exporters. There was hope initially that there would be a fairly rapid recovery in demand in Eastern Europe and Russia and it seemed doubtful that China would, or could, continue to expand its lead and zinc production capacity at a break-neck pace. Thus, although there was a temporary cut in mine output in the West starting in 1993, Western smelter capacity largely remained in place. Stocks of both lead and zinc concentrates, accumulated in the early 1990s, were drawn down but the metal produced added to market over-supply and set in train a drawn-out period of depressed prices which persisted over the balance of the decade. (There was a spike in the zinc price in 1997 but this reflected a squeeze on large short positions taken by one Chinese producer.) However, it is worth noting that, while smelters remained in place, there was little expansion of refined lead or zinc capacity in the Western World during most of the 1990s.

Low prices for much of the 1990s, the changing structure of lead and zinc production, both in geographical terms and the role of secondary materials, and variations in consumption growth rates have established the context for the future evolution of lead and zinc supplies. At this stage perhaps the most profound effect on the industry has been the impact that low prices have had on slowing the pace of investment in new mine and smelter capacity in the West. Instead, capacity has expanded rapidly in China in recent years and has played a significant role in meeting the need for greater supply to accommodate growing global market demand. However, this has created a situation where there is a distinct risk of a shortfall in overall mine and/or smelter supply in the opening years of the 21st Century. This arises because low prices and a concern about further developments in China have engendered a degree of wariness on behalf of Western companies and investors to fund the development of new projects. This is especially so in the case of lead mine production.

We noted earlier that lead and zinc are often mined together. However, over the past 40 years there has been a significant change in the typical ratio of lead to zinc mine production, as shown in Figure 4. As recently as 1960 the ratio of lead to zinc in Western World mine output was 0.7:1. It has fallen steadily to the current level of just under 0.4:1. Furthermore, many of the most promising mine projects have prospective lead:zinc ratios of substantially less than this. Indeed, several zinc mine projects will produce very little or no lead at all. These are the coppenzinc deposits or zinc oxide projects. This shift in the kind of ores now being exploited and evaluated for future development is a response both to market fundamentals and to new and improved technology.

12 LEAD-ZINC 2000

0.8 i 1 7

Lead/Zinc mine production ratio

Figure 4 - Ratio of Mined Lead Output to Zinc (Western World Only) (Source: ILZSG)

The market fundamentals are reflected most clearly in the change in the relative prices of lead and zinc, with the zinc price being substantially higher than the lead price since 1980, in contrast to earlier experience. Given that no chronic shortage of lead has appeared over the past 20 years, does this mean that the cost of producing lead has fallen significantly relative to zinc? If this is not the case, what is the explanation for companies remaining in the business of lead production? In attempting to answer these fundamental questions it should be stated that, at the mining level, it is very difficult to disaggregate the cost of mining lead from zinc. In reality, it is only once the concentrates leave the mine site that it is possible to allocate accurately specific costs (freight and treatment charges/smelting and refining costs) to each metal. In practice, where the metals are mined from the same orebody, the costs incurred can be apportioned to the two metals (and other by-products) in a number of ways. There are those who prefer to address each metal separately, setting all costs of production against one metal, offsetting these against any co/by-product revenues that might be earned. Alternatively, costs can be divided on a pro-rata basis, perhaps in proportion to the revenue of each metal stream. Either way, the relative prices of all metals produced will have a significant bearing on the calculated cost of production for each individual metal. With lead prices clearly on a falling trend relative to zinc since the early 1980s, zinc mine output has tended to carry a greater share of the costs compared with lead. In other words, the cost of mining lead has fallen. In fact, mining costs overall have fallen in real terms over the past quarter century because of improvements in productivity, giving further support to the thesis that the cost of producing lead has fallen in the context of a weaker price relationship with zinc, as indicated in Figure 5.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 13

0.2 - -

0 I : . , 1 , , : , ,

1960 1965 1970 1975 1980 1985 1990 1995

Figure 5 - Ratio of Lead to Zinc Prices 1960-1999

Another factor often mentioned when discussing lead production costs is the increase in the proportion of secondary output relative to primary production. Environmental legislation has played a significant role in determining the evolution of lead supply. Legislation in many states in the USA and in some European countries prohibiting the dumping of scrap batteries has undoubtedly caused recycling rates to increase. Measures have been adopted to establish closed loops, including battery manufacturer, distribution chain, final consumer and recycling facility, to ensure that there is both an economic and legislative incentive for scrap batteries to be collected and returned to an environmentally sound plant for recycling. It was realised that, without a legislative framework and left to purely commercial considerations, lead recycling would take place only if it were economically viable. This was evident in the mid-1980s when recycling rates slumped during a period of low prices (just below $650/t or 30c/lb in $1999 terms). In those countries where there are well-organised scrap battery recycling schemes, and some of the cost of collecting this lead scrap is effectively passed onto the consumer or battery manufacturer, it may well be that this has had an impact on lowering costs and maintaining the viability of lead recycling. Set against this, however, are the much higher costs that must be incurred to meet ever more stringent environmental laws. Indeed, the recycling industry has had to invest heavily in upgrading its facilities or face closure. Overall, therefore, we feel that the impact of rising secondary production on average costs has probably been neutral. The inescapable conclusion is that lead prices have been depressed relative to zinc because of a general perception of market oversupply, even at times when lead inventories have been falling or held stable. One might even go so far as to say that zinc production has "subsidised" lead output, resulting in very low margins for the lead/zinc industry overall for much of the past decade. This is a partial explanation for why lead concentrate producers have remained in business. Most are actually zinc companies forced also to produce a low-priced by-product. Thus, while the economics of producing lead may have been poor, so long as the zinc price was reasonable, mines remained in business. However, there have been casualties amongst those mines where lead revenues are important and precious metal values are low. The outstanding example here is Faro. For the most part though, miners have struggled on generating low

14 LEAD-ZINC 2000

margins and certainly not earning enough to finance significant exploration activity. Project development has been at a very slow pace.

If the increase in secondary lead production, together with the flow of lead from the transitional economies, was such as to force prices in 1993 to their lowest level in real terms in 100 years (and hold them close to this level since then), why has there not been a greater response from primary producers, especially custom smelters? For these producers the reasons for remaining in the lead business are complex. Firstly, the lead price is not a key determinant of their profitability; it is principally the concentrate treatment charge and recovery of precious metals and other by-products that generate the revenues. Lead TCs have been relatively high in recent years. Also, there is the question of the liability for environmental clean-up if a lead smelting site is decommissioned. It may well be the case that it is less costly to persevere in this marginal business than trigger an expensive and difficult clean up.

With concentrate supplies tending to fall, primary producers also have sought to maintain output levels by increasing the proportion of secondary material treated. Here again legislation has helped in increasing the flow of lead-containing residues and industrial wastes to smelters. Furthermore, in some countries, the owners of secondary facilities have decided that it is more cost effective to divert sulphur-containing battery pastes to "primary" smelters. This has been especially evident in Europe in France, Germany and Italy. The construction of new, primary smelters employing technology (Ausmelt, QSL and Kivcet) specifically designed to treat a range of raw materials has been a factor in this development. Again, environmental legislation has been the driving force in hastening the adoption of newer technology where greater control of all emissions is possible.

In the case of lead, we can see that the market fundamentals that have determined the evolution and character of supply have reflected principally the impact of environmental legislation. This has forced changes in the patterns of demand and in the character of supply. For zinc we see much less evidence of environmental factors determining both the nature and scale of supply, although there are some important issues in this area that are likely to have a higher profile in the future. (The question of the disposal of EAF dust is just one factor. General environmental concerns are likely to maintain the pressure to increase the amount of zinc being recycled.) Where the mine supply of zinc has exhibited some important changes in recent years, however, has been in relation to its association with lead production. Economic considerations have meant that attention has been given to projects with low lead values, such as Antamina in Peru which will produce principally copper and zinc, Century in Australia and the new Irish mines, Galmoy and Lisheen, none of which will produce much lead in relation to their zinc output, and zinc leach projects such as Skorpion in Namibia and Shaimerden in Kazakhstan. If developed, the last two projects will produce significant quantities of zinc (100 kt+ each) and no lead at all.

It is interesting to note that, while zinc prices have not been very favourable over much of the past decade and, as a consequence, financial returns to the zinc industry have been poor, zinc smelters have not suffered any significant or sustained shortfall in concentrate supplies. This is in sharp contrast to lead smelters that experienced acute pressure on their raw material supply between 1993 and 1995. Indeed, at the end of the century, zinc smelters were able to take comfort that, notwithstanding the fact that several mines are due to close as a result of reserve depletion over the course of the next few years, there was the prospect of a sizeable increase in concentrate availability in the near term with a number of large mines in the process of commissioning and several other mines pursuing expansions. As indicated in Figure 6, it is this latter development that has been especially important for the zinc industry. There had been concerns that, based on the likely schedule of new mine developments and mine closures, zinc

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 15

smelters might have faced a serious shortfall in mine supplies going into the 21st Century. However, as matters have evolved, zinc mine output has increased slightly ahead of smelter capacity. One of the keys to this has been unlocking the potential of brownfield mine expansions, with a much lower capital cost compared with new mine site projects, a feature not hitherto associated with zinc mining. Notable here has been the marked increase in capacity at Cominco's Red Dog mine in Alaska, with a further expansion to follow in the next 12 months.

Figure 6 - Evolution of Lead and Zinc Mine Supply Including Probable and Possible Projects

Although zinc smelters may take some comfort from recent developments affecting the evolution of zinc mine production, a serious problem looms for the lead industry. Many of the mines slated for closure as their reserves deplete over the period 2000 to 2005 represent operations with relatively high lead to zinc ratios such as Hellyer, Sullivan, West Fork and Black Mountain. Based on information available at the beginning of 2000, mine reserves at existing operations and those projects which appeared definite to go ahead suggested that the ratio of lead to zinc mine output was likely to fall from just under 0.4:1 in 1999 to only 0.3:1 by 2005. If this actually happens, Western World lead smelters will face an insurmountable shortfall in concentrate supplies that could well result in smelter closures because of the lack of feed. This would be an unprecedented situation but an inevitable result of a drawn out process affecting the evolution of lead supplies over the past 25 years. In Figure 6, it is easy to see how the various market factors which have affected both lead and zinc over recent years are being translated into a much greater focus on the search for new zinc resources. Although there is a reasonable prospect that zinc mine production will expand significantly over the first decade of the new century, based on an assessment of output from projects now being explored and evaluated, the same cannot be said for lead. Production from new projects is expected barely to offset losses from mine closures through reserve depletion, leaving hardly any scope for an increase, if indeed that is what the lead concentrate market requires.

16 LEAD-ZINC 2000

Ultimately, the need for lead concentrates will be driven by the final demand for metal and developments affecting secondary production. If a real shortage of lead units emerges, then rational economic analysis suggests that the lead price will have to rise to encourage development of lead mine capacity. However, in recent years lead production growth in China has outpaced market growth, even at the low prices prevailing.

Although the lead industry is facing up to the consequences of past and current developments, there are also important implications for zinc producers to consider. It is likely that the viability of new zinc mine production will increasingly be driven by the economics of zinc alone or by a co/by-product other than lead. In other words, new mine capacity to meet the growth in demand for zinc globally may have to be justified on zinc revenues alone in lead/zinc operations. With lead no longer bringing significant value to zinc mines, this means that current plans for the development of new zinc mines, where lead is the main associated by-product, may require a higher zinc price than that seen typically in recent years. This will add to the pressure to bring down the cost of zinc mine production. It is also perhaps prompting the development of much larger operations than have characterised the industry in the past and the expansion of existing operations where possible. New zinc-producing mines may also increasingly be associated with different co-products, most importantly copper. More metallurgically complex projects have been coming to the forefront in recent years, with one such major project, Antamina, now not too far from commissioning. Another, Pirquitas, which could produce zinc, tin and silver, has been promoted actively over the last few years. Finally, new technology, which offers the prospect of very low operating costs for the production through to refined metal, is also being considered for the exploitation of zinc oxide deposits. These and other projects are posing challenges, not just for the zinc mining industry, but also for smelting and refining.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 17

LEAD PRODUCT DEVELOPMENT IN THE NEXT MILLENNIUM

R.D. Prengaman RSR Technologies, Inc.

2777 Stemmons Freeway, Suite 1800 Dallas, Texas, U.S.A. 75207

ABSTRACT

Lead usage in paint, anti-knock gasoline additives, pipes, seals, and solder has declined markedly over the past 30 years. Products utilizing the unique properties of lead and its ability to be recycled have survived and grown. These products include shielding, glass (particularly TV and computer screens), sheet for roofing, stabilizers for PVC, insoluble anodes for metal production, electronic solders and lead acid batteries. With the loss of markets many companies have decreased lead product development efforts. Much of the patent lead materials development is devoted to lead acid batteries. In the past 25 years, the automobile battery has been almost doubled in power and reduced in weight, leading to a nearly three-fold increase in performance. Corrosion-resistant lead alloys and improved battery designs have made battery lead usage more efficient. The same corrosion-resistant battery alloys are utilized to increase the life of electrowinning anodes. The challenge for lead product development is not only to maintain and improve the existing lead products, but also to develop the growing lead acid battery market for telecommunications, uninterruptive power sources, remote area power, hybrid and electric vehicles.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Sicgmund TMS (The Minerals, Metals & Materials Society), 2000

18 LEAD-ZINC 2000

INTRODUCTION

Over the past 30 years, many lead-based products have disappeared from use or the use has been greatly diminished. These products include lead-based paints, lead-based anti-knock compounds for automobiles, lead glazes for pottery, lead sealants for wine battles, solder for sealing the seams of steel cans, automobile body solder, solders for potable water systems, solder for sealing copper radiators, chemical tank linings, lead pipes, lead-based metal bearings, lead sealants such as for drain pipes, power and electrical cable sheathing, and ammunition.

In many cases these products were replaced or removed because of environmental pressures associated with the dispersive and toxic nature of lead. Some products such as anti-knock compounds were banned because lead poisoned the catalytic converters on automobiles which were used to produce improved air quality. The removal of lead reduced the airborne exposure of lead. Solders for steel cans and potable water systems were replaced because of the potential leaching of lead from these products. Copper radiators have been replaced by aluminum for reduced vehicle weight. Lead bearings have been substituted by roller bearings. Lead tank linings have been replaced by stainless steels or plastics. Tile drain lines have been substituted by plastic pipe. Lead cable sheaths have been replaced with plastic/aluminum composites. Lead shot for waterfowl hunting has been displaced by steel shot and leaded steels are being replaced with bismuth containing alloys.

PRODUCT DEVELOPMENT

In most cases the replacement of lead products has been by products which offer lower exposure to toxic metals, limit the release of lead to the environment, are lighter, stronger or cheaper, or have new production methods. In the new millennium, lead-based products will consist only of lead products with a definite purpose. These product requirements will be:

• needed throughout the economy • cost effective • readily recycled • packaged to limit exposure to lead • not readily substituted • have longer life and use lower amounts of lead • offer significant advantages over competitive materials.

Product development in the new millennium will see the development of only a few product areas. Some governments around the world wish to ban all lead products or at least restrict lead to the currently used products. In this universe there will be no new lead products. How does an industry perform good product research and development when there are no new products?

Lead products in the future will fall into several categories where the unique properties of lead enable it to compete with other materials. The special characteristics are protection and electrochemical properties.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 19

Protection

Because of its unique properties, lead finds significant use in protecting the population from radiation. Lead is used to collimate and absorb x-rays and other damaging radiation. Lead is used, and will continue to be used, in containers for radioactive isotopes, shielding in nuclear power and medical treatment facilities. Lead in the form of lead oxide glasses finds use in cathode ray tubes, TV picture tubes, computer screens, and transparent shielding to protect users of these products from ionizing radiation. Other protective products are lead stabilizers to protect PVC products from degradation caused by ultraviolet radiation or other natural environmental processes. Lead sheet is used as waterproofing where long lifetimes are required.

In each of these applications, continued product development will utilize the optimum amount of lead for each shielding material. Improved stabilizers will permit the use of lower amounts of lead while still imparting the same beneficial properties. And finally, means must be developed to recycle the lead shielding materials, particularly those in TV and computer screen glasses. In many lead products, process development to recycle these products will be as important as product research to improve the product.

Electrochemical Properties

Because of the unique properties of lead to exist in three valence states (metal, ion with +2 charge, and ion with a +4 charge), lead is used not only for electrochemical anodes to electroplate other metals from sulfuric acid solution but also to serve as anode, cathode, and active material in lead acid storage batteries. Storage batteries represent over 60% of lead usage worldwide and over 80% of lead usage in the United States. Electrowinning anodes and batteries are two of the unique successes of lead product development efforts.

Lead product development in these two areas has led to the application of novel processes and products which use these newly improved old products.

Anodes

Insoluble lead anodes have been used for electroplating zinc, copper, chromium, nickel, cobalt, and other metals from sulfuric acid for many years. Over the past 20 years, product development efforts have improved anode performance dramatically as seen in Table I.

Table I - Performance of Anodes in Copper and Zinc Electrowinning Systems

EW Date Anode Thickness Anode Life Current Density Process (mm) (year) (A/m2)

Copper

Zinc

1980 2000

1980 2000

10- 15 mm 6 mm

12- 18 mm 9 mm

.5-2 8-10

1 -2 6-8

180 300

380 550

The performance of the innovative lead anode products has enabled new copper SX-EW processes to grow from a very small percentage of the copper market in 1980 to a significant

20 LEAD-ZINC 2000

market share in 2000, with significantly lower production costs than conventional copper production methods and similar cathode purity. Product research has been responsible for the development of new rolled lead-calcium-tin alloy materials with unique grain structures which are extremely corrosion resistant for anodes. This has enabled the anode thickness to be reduced by a factor of two while the average life has been increased by a factor of at least five, even as the current density was almost doubled.

New lead-calcium-silver alloy zinc electrowinning anodes have given longer life with thinner anodes at higher current density. The lead anodes offer significant cost, life and performance advantages over conventional anodes. The lead anodes offer scrap value to recyclers and are much more cost effective than non-lead competitive electrowinning anodes. Much of the success of electrowinning anodes has been the result of new product development aimed at improvements in lead acid battery technology. For lead to be competitive in a toxic/regulatory environment it must offer dramatic improvement over the old product. For lead to continue to be considered a viable product in the new millennium, the product must use less material, have a longer life, and offer greatly improved performance than the current product. Lead electrowinning anodes are an excellent example of the "new" old products.

Lead Acid Batteries

Lead acid batteries are the bright spot in lead product development. Since 1970, the percentage of the lead market occupied by lead acid batteries has risen from 28% in 1960 to over 73% in 1999. The lead market has risen primarily as a result of the dramatic increase in production of lead acid batteries even as other lead products declined or disappeared. The lead consumption has risen from 3,260,000 tons in 1960 to over 6,000,000 tons in 1997. Battery usage increased from just under 1,000,000 tons to 4,400,000 tons per year over the same period. The lead acid battery is readily packaged for recycling and the amount of recycled lead production increased dramatically from under 1,000,000 tons in 1960 to over 3,000,000 tons in 1999.

Many of the improvements in lead acid batteries have been the result of lead product development aimed at improving the performance of the lead acid battery. Novel battery grid alloy materials have led to process changes which resulted in significant decreases in cost to produce the batteries. The process changes not only decreased production costs but also increased productivity and at the same time decreased worker exposure to lead.

In 1975 at the U.S. Super Bowl, the Sears die-hard battery started several cars simultaneously on a frozen lake in Wisconsin. Table II shows how typical lead-acid starter battery performance has improved between 1975 and 2000.

Table II - Lead Acid Battery Performance

Year Weight (kg) Cold Cranking Amps

1975 21 450 2000 15 750

The performance of the battery has almost doubled while the weight of the battery has been reduced by over 30%. The resultant battery product has been improved by a factor of 2.5 times. The improved performance and reduced material consumption have resulted in reduced

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 21

lead consumption per battery. The improved performance combined with improvements in productivity resulting from new lead battery alloys and lead materials have resulted in a decrease in cost to the consumer for automobile starter batteries in 2000 compared to 1975.

The major improvement has been the replacement of lead-antimony alloys with lead-calcium-tin alloys. The new alloys are more conductive and significantly more corrosion resistant than the alloys which they replaced. In addition, the new alloys permitted the construction of maintenance-free batteries which do not require the addition of water during the expected life of the battery. The new alloys permitted dramatic changes in the battery grid and plate manufacturing process from handling individual grids and plates to continuous processes where battery parts are produced at high speeds. The continuous processes reduced worker exposure to lead during battery manufacturing. The maintenance-free batteries, because of the sealed construction, reduced consumer exposure to the acid electrolyte in the battery, further enhancing the environmental appeal of the modern automobile battery.

Automobile batteries have also been designed for recycling. The secure case permits return of the failed battery to the recycler without unique environmental problems. Lead recyclers have developed processes to recover virtually all the lead values in the battery for return to new battery production. In addition, the battery cases and acid have been also recycled to usable products. Lead acid batteries represent the highest recycling rate of any commercial product. Rates of 95-97% have been achieved in the late 1990's.

The greatest growth area for lead products is in sealed valve-regulated lead acid batteries (VRLA). In these batteries, the electrolyte is contained in a gel or the separator between the plates, and gases generated upon charging are recombined within the battery. The totally sealed batteries can be utilized in any orientation and virtually any location and have found use in telecommunications and uninterruptive power sources for hospitals, computer systems, and emergency systems. VRLA batteries are also used for power conditioning, remote area power storage such as from solar or wind power generation, and are being specially developed for electric vehicle and/or hybrid electric vehicles. Much of the research and development for these new battery systems particularly for electric vehicle and hybrid electric vehicle service has been funded by the Advanced Lead Acid Battery Consortium (ALABC).

Lead product development efforts have been devoted to the development of new lead alloys to reduce the rate of corrosion, improve the conductivity between the grids and active material, and constrain the active material to increase life. Lead product development efforts for batteries are also aimed at improved oxides for the active materials leading to improved material utilization. Process development to improve the purity of lead from recycled batteries is also a major factor in improving the life of VRLA batteries.

Because of the development of extremely corrosion-resistant, extremely strong battery grid materials, new manufacturing methods have been devised to produce much thinner grid materials, which not only decrease weight but also improve the performance of the active material. The new lead alloy grid materials are resistant to corrosion and degradation at elevated temperatures. This makes them ideal candidates for extending the life of new batteries operating in hot, severe climates where battery life is extremely short today.

These new alloys have permitted significantly different battery designs. New battery designs such as thin metal foil batteries, using grids as thin as 0.05 mm, offer significant improvements in power compared to conventionally designed batteries using grids of 1-2 mm thick. A 12-V battery of the new design weighs about 3.5 kg and delivers 1000 amperes while the conventional design weighs 15 kg and delivers only 750 amperes. New battery designs

22 LEAD-ZINC 2000

using the newly developed grid materials are being developed to compete with exotic battery chemistries for use in hybrid or electric vehicles.

Product development in the new millennium means not only developing your own material, but also assisting the customer who produces the ultimately used lead-containing materials. The assistance requires even more knowledge of the customer's process and leads to a partnership between the lead product producer and lead consumer in batteries. The lead product development engineer in the new millennium must be as (or more) knowledgeable about the user's process than the customer.

CONCLUSIONS

The number of lead products has decreased significantly over the past 30 years. The remaining lead-containing products offer special advantages over competitive products with regards to performance, cost, life and recyclability. The major areas where lead products remain are those of security and electrochemical applications. These few remaining lead-based products have dramatically expanded lead usage by serving customer needs while at the same time reducing environmental exposure of a toxic metal by extremely high recycling rates. Product development in the new millennium is not only the development of improved products, but also a partnership with customers to improve their productivity as well as to produce lead-based products which are readily recycled.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 23

ZINC APPLICATIONS: A WORLD OF PERFORMANCE

E. Gervais International Zinc Association,

Avenue de Tervueren, 168/Box 4 B-l 150 Brussels, Belgium

ABSTRACT

Modern life in inconceivable without zinc. Zinc provides the most cost-effective and environmentally efficient method of protecting steel from corrosion. Zinc is also used to power electric vehicles and computers, to make brass, automotive equipment, household appliances, fittings, tools and toys, and to support the building and construction industries. It is also used in pharmaceuticals and cosmetics, in rubber, in fertilisers and in animal feeds. The links between properties, applications and performance, as well as zinc demand and market dynamics, will be reviewed along with market development opportunities and successes. The presentation will also include a progress report on the five-year plan of the International Zinc Association to increase the global market for zinc, over and above the natural growth, by 500,000 tonnes/year by 2002.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

24 LEAD-ZINC 2000

INTRODUCTION

Zinc is a versatile and quite remarkable metal which extends steel product life through corrosion protection, which is cast into intricate high precision components, which adds design flexibility in architectural applications, which is a vital micronutrient and so forth. For most of its applications, zinc plays its role in association with other materials, for example:

• Coatings on steel where the zinc thickness may range from a few micron to more than 100 microns

• Alloying with copper to make brass where the zinc content may approach to 40 %, but it is typically 30%

• Rubber which contains on average 2 % zinc oxide • Zinc carbon and alkaline batteries which contain 4 to 25 grams of zinc.

Because of this fact, zinc's impact to our modem world is much larger than a simple reference to the tonnage produced and recycled would suggest.

The zinc industry is engaged in a process to characterize zinc's contribution to sustainable development and to develop the future business plan for the International Zinc Association (IZA) based on the concept of sustainable development. As part of this process, we are initiating a program to characterize the benefits/contributions of zinc to the world. The results will be delivered in terms of material, economic, environmental and social performance. This work will explore zinc's contributions to safe and affordable housing, safer and more energy efficient vehicles, better infrastructure, agriculture, health, etc.

This paper, the first in a series, focuses on the link between properties, applications and performance, explores zinc demand and market dynamics and examines the performance of the co-operative market development efforts of the industry.

KEY PROPERTIES AND APPLICATIONS

Centuries before zinc was discovered in the metallic form, its ores were used for making brass and zinc compounds for healing wounds and sore eyes. Metallic zinc came into use in the Middle Ages becoming the eighth metal known after the seven metals of antiquity: gold, silver, copper, iron, tin, mercury and lead. Zinc's corrosion resistance and the possibility to roll zinc at 100 -150 °C to produce a flexible sheet paved the way, around 1810, for zinc roofing. Hot-dip galvanizing, the oldest steel anticorrosion process, was introduced in the 1830's. The die-casting process was developed in 1849 but it was not until zinc of 99.99 % purity became available in the 1920's that the modem zinc die-casting industry was bom. Sendzimir pioneered continuous steel sheet galvanizing in 1931.

The key properties and uses of zinc are given in Table I. The broad range of end-uses results from zinc's corrosion resistance, galvanic protection, alloying characteristics with copper, aluminum and magnesium, low melting point, formability, essentiality as a nutrient and its healing properties. The economic significance of each first-use and the broad range of end-use markets served by them is illustrated in Figure 1.

Tab

le I

- Pr

oper

ties

and

Use

s of

Zin

c

Prop

ertie

s R

eact

ivity

with

iron

C

orro

sion

res

ista

nce

Ele

ctro

chem

ical

L

ow m

eltin

g po

int

Flui

dity

C

apac

ity fo

r su

rfac

e tr

eatm

ent

Stre

ngth

A

lloyi

ng c

hara

cter

istic

s

Form

abili

ty

Res

ista

nce

to c

orro

sion

E

lect

roch

emic

al

Phys

ico-

chem

ical

Ess

entia

l nu

trie

nt

Hea

ling

Firs

t-U

se/P

rodu

ct

Cor

rosi

on p

rote

ctio

n (g

alva

nizi

ng,

zinc

ther

mal

spr

ayin

g, e

lect

ropl

atin

g,

zinc

-ric

h pa

ints

) D

ie c

astin

g an

d gr

avity

cas

ting

Bra

ss (

copp

er-z

inc

allo

y), a

lum

inum

al

loys

, mag

nesi

um a

lloys

R

olle

d zi

nc

Bat

teri

es

Zin

c ox

ide

Zin

c co

mpo

unds

Z

inc

com

poun

ds

End

-Use

B

uild

ing/

cons

truc

tion,

ene

rgy/

pow

er,

stre

et f

urni

ture

, ag

ricu

lture

, au

tom

otiv

e/tr

ansp

ort

Aut

omot

ive

equi

pmen

t, ho

useh

old

appl

ianc

es, f

ittin

gs, t

oys,

tool

s, e

tc.

Bui

ldin

g/co

nstr

uctio

n, f

ittin

gs,

auto

mot

ive

and

elec

tric

al c

ompo

nent

s, e

tc.

Bui

ldin

g/co

nstr

uctio

n

Con

sum

er p

rodu

cts,

aut

omot

ive/

tran

spor

t, co

mpu

ters

M

edic

al e

quip

men

t T

yres

, all

rubb

er g

oods

Pa

int p

igm

ents

C

eram

ic g

laze

s E

lect

rost

atic

cop

ying

pap

er

Food

indu

stry

, ani

mal

fee

d, f

ertil

izer

s Ph

arm

aceu

tical

ind

ustr

y, c

omes

tic i

ndus

try

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 2

26 LEAD-ZINC 2000

World total 1999: 8,18 million tonnes

Figure 1 - Zinc Consumption: First-use and End-use. Source: Brook Hunt and Outokumpu OY

Zinc's most remarkable quality is its ability to protect steel against corrosion. Zinc has an excellent corrosion resistance in the atmosphere, in hard fresh water, in salt water and in contact with many natural and synthetic substances. The life and durability of steel is greatly improved when coated with zinc. No other material can provide such efficient and cost effective protection for steel. Zinc provides a threefold protection to steel:

• Tough and adherent coating, sealing the underlying metal • Corrodes much more slowly (about 40 times) than steel • Electrochemically protects exposed steel surfaces.

Seven different practical and cost effective methods are commercially available to apply zinc to steel surfaces. Today, continuously galvanized steel with a very fine surface finish is used to produce the parts of car bodies that are vulnerable to corrosion. The surface finish of the coated steel is such that there is no visible difference in appearance, after painting, between panels with and without zinc protection. The building and construction industries use at least two-thirds of all the coated steel strip produced, mainly for roofing and cladding of commercial and industrial buildings. Much of the material used for building has a mill-applied organic coating on top of the zinc. Color-coated steel buildings are a familiar sight, particularly in shopping centers and industrial parks. Color-coated steel can provide lifetime protection for such buildings.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 27

Brasses are the best known copper alloys. They are primarily binary copper-zinc alloys and they contain up to 40 % zinc. The best known brass (yellow or architectural brass) is the 70 % Cu -30 % Zn alloy. All brasses can be cast readily; they are ductile and they can be worked into complex shapes. Brass seems so commonplace in everyday life that it is almost taken for granted. For instance in our houses, it is encountered as doorknobs, taps, valves and lighting fixtures; it is even more frequently encountered in commercial buildings.

Zinc in alloyed form can be cast with fine detail into complicated shapes and can be formed with extra thin walls (0.5 mm). Specifiers can choose from a family of zinc alloys and a range of casting processes that are the most suited for their products. Zinc alloy castings are unique, particularly when produced by the pressure die-casting process. They can be made to extremely close tolerance, with excellent surface finish, have a range of useful physical and mechanical properties and can receive a wide range of applied finishes. Typical applications are precision parts for vehicles, aircraft and communications equipment, plumbing hardware, construction fixtures, household appliance, zippers and toys.

Zinc oxide is commercially the most important compound of zinc and it is used to a considerable extent in the rubber, ceramic and paint industries. It is utilized to activate the organic accelerator of the vulcanization process for natural rubber and for most synthetic rubbers. Also zinc oxide serves as the accelerator for some types of elastomers. Further, it provides reinforcement to the rubber, it improves its heat conductivity, it limits the degradation by UV radiation and it improves adherence. Typically, rubber contains 2 % zinc oxide. A very broad range of other compounds are of extensive interest in chemistry, biology and physics and they make great practical contributions to the chemical, ceramic, fertilizer, paint, plastics, textile and electronics industries.

In many respects, zinc has nearly ideal properties as a material of choice for battery cell anodes. Its high reducing potential results in a reasonably high cell voltage. Indeed, zinc has the highest practical potential for aqueous systems. Various zinc battery system are used commercially; they find very extensive applications as a source of electricity. Primary cells, which mainly have neutral and alkaline electrolytes, are particularly widely used. The most common zinc batteries are the Leclanche dry cell (zinc/carbon), the so-called alkaline batteries and the zinc-air battery. Typically Leclanche and alkaline batteries contain from 4 to 6 grams of zinc for AA cells to 20-25 grams for D cells which are commonly used in flashlights. The usefulness of these batteries is illustrated by the fact that 22 billion Leclanche cells and 11 billion alkaline batteries are produced every year.

The zinc-air battery presents an opportunity to innovate and to create new market niches. These batteries have generated research interest, stimulated by the requirements of the developing electric vehicle market. The zinc-air system offers the highest specific energy and, after the lead-acid battery, the lowest cost.

DEMAND AND MARKET DYNAMICS

Last year, world consumption grew by 5.5 % reaching 8.18 million tonnes of refined metal (>97.5 % zinc); refined metal consumption includes some recycling. Consumption, including all forms of recycling, reached 11 million tonnes in 1999. Consumption has been multiplied by 22 times this century; this is equivalent to an annual compound growth rate of

28 LEAD-ZINC 2000

more than 3 %. In tonnage terms, zinc is in fourth place among all metals, after iron, aluminum and copper.

Broadly speaking, world zinc demand growth has followed the fluctuating fortunes of the world economy and has shown strong periods of growth following global recession periods. The oil crises of the 70s, the recession of the 80s and the major geopolitical shifts of the early 90s, all impacted significantly on refined zinc consumption. Each of these events led to a consumption decline; this is clearly seen in Figure 2. The political changes in Eastern and Central Europe caused the major drop in consumption in the 90s; notably in the former CIS, where consumption collapsed from about 800,000 tonnes/year to about 250,000 tonnes/year.

Figure 2 - World Refined Metal Consumption (1960-1999) Source: International Lead Zinc Study Group (ILZSG) and IZA

Since 1960, refined zinc consumption has been growing at a trend growth rate of 2 % a year. Over the last seven years, world consumption has been growing at the faster rate of 3.2 % a year. It is generally believed consumption will continue to grow at rates exceeding 3 % per year.

Zinc consumption is linked to economic development. There is a close relationship between industrial production and zinc consumption. Over the 39 years from 1960 to 1999, OECD industrial production (IP) grew at 3.8% a year, whereas consumption in the OECD countries grew at 2.4 % a year. In recent years, zinc consumption growth came much closer to that of the OECD industrial production. Since 1994, IP growth has averaged 3.6 % a year, while OECD zinc consumption expanded at an average rate of 3.2 % a year.

Over the last 20 years, consumption in Asia, driven by internal demand, by major investments in public infrastructures and by growing exports progressed at a rate which was about three times that of the growth rate of Western Europe and North America. Altogether, consumption growth in Southeast Asian countries has been spectacular. Similarly, China, now the largest zinc consumer, has been growing for the last ten years at an average rate of 7.5 % a year. The net effect of regional growth rate on consumption is illustrated in Figure 3.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 29

Figure 3 - Net Changes in Refined Zinc Consumption (1979-1999) Across the World Asia/Pacific Includes 11 Countries Excluding China and India. Sources: ILZSG and IZA

Zinc consumption is market driven. Traditionally, consumption is analyzed in terms of six first-uses categories (Table II). Total reported consumption by first-uses provides a rough indication of the combined effects of developments taking place within each first-use market. At 50 % of consumption, galvanizing is the main user of zinc. It is followed by brass (19 %), by zinc base alloys (18 %), compounds (8 %) and semi-finished products (6 %). Since 1960, the galvanizing market share grew from 38 to 50 %; over this period galvanizing also experienced the fastest trend growth rate (3 %).

Table II - Total Reported Annual Consumption (1998) by First-use (*)

First-Use Market

Galvanizing Brass Zinc base alloys Compounds Zinc semi-manufactures Zinc/dust and powder Miscellaneous TOTAL Source : ILZSG & IZA

Consumption ktonnes 3,321 1,327 1,072

576 480

63 236

7,075

Trend Growth Rate (;

1978-1998 3.7 2.4 1.5 3.2 1.9

(1.2) 3.7 2.8

**) % a year 1993-1998

3.4 3.1 2.7 2.3 1.6 1.0 6.7 3.1

(*) Data based on about 65-70 % of world consumption inclusive of recycling (**) Compounded growth rate derived from statistical analysis

30 LEAD-ZINC 2000

Galvanizing protects approximately 100 million tonnes of steel and is driven by both construction and automotive applications. Galvanizing grew the fastest (3.7 % a year) during the 1978-1998 period; galvanizing is the engine of growth for zinc consumption. This first-use breaks down into two major product families: continuous galvanizing and general galvanizing. They represent about 30 % and 20 % of total zinc consumption, respectively. The needs of the automobile industry for corrosion protection and the increased promotional efforts of the steel industry were the major catalysts for developments in continuous galvanizing. During the 90s, world production of coated steel sheet has been increasing at a rate of 5.9 % a year, reaching 71.5 million tonnes of coated steel sheet in 1999. Investments in continuous galvanizing capacity are continuing at a fast pace; coated steel sheet production is expected to reach 86.0 million tonnes by 2004. Concurrently, as a result of product and process refinements, the average zinc coating thickness has been reduced. On average, 35 kg of zinc are used per tonne of hot-dip galvanized steel sheet (76 % of installed zinc coating capacity).

In Western Europe, the largest market for general galvanizing (395,500 tonnes per year) has been increasing at a trend growth rate of 4.5 % a year since 1984. Growth in other regions has been less significant. Altogether, worldwide, the galvanizing market, which is supported particularly by developments in construction and public infrastructure, is expected to grow at a rate of about 3% a year.

Brass, the second largest market, maintained its relative position. However, it grew at a slower rate (2.4 % a year) than the average; the trend growth rate increased to 3 % a year during the last 5 years.

Triggered by the oil crises, consumption for zinc base alloys dropped from 1,140,000 tonnes in the early 70s to 700,000 tonnes during the 70s. Consumption dropped because of car weight reductions and downsizing, because of increased attention to value engineering as well as the impact of increased materials competition. Competition from plastics threatened the market for zinc castings, but the development of process control enabled zinc castings to hold their own in many areas, particularly where strength, accuracy, surface finish and applied finishes are required. A further benefit of the process improvements was that castings could be made much thinner so that much less metal was used for a given product. This in turn meant weight saving - a valuable point, especially in automotive applications - while at the same time, quality and consistency improved and production costs were reduced. So, although far more castings are produced today than 20 years ago, this is not reflected in the tonnage used. Thus, over the 20-year period, although reported consumption by this first-use market appeared stagnant, extremely positive changes were taking place. The trend growth rate (2.7 %) for the last 5 years (1993-1998) suggests that recovery, in terms of tonnage, is underway after the long period of adjustment that followed the oil crises.

To conclude this section, it is important to stress that consumption growth is not simply the result of growing demand for products because of growing economic activity; it is the result of breakthroughs particularly in the making of continuously galvanized steel. Today superior quality galvanized steel produced at higher speed is meeting the exacting requirements of the consuming industries particularly in the automotive industry. Similarly, developments in the die-casting industry, such as improved product design and manufacturing control, led to lighter and better products. Overall economic growth, zinc's remarkable characteristics, outstanding developments in zinc-containing product production technology, and the increased promotional efforts of the steel industry are behind zinc's superior consumption growth rate. Another factor, active promotion by the zinc industry, is covered in the next section.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 31

COLLECTIVE MARKET DEVELOPMENT

Markets are our most important assets; they must be defended and expanded. Principal challenges are:

• Designers/specifiers have a choice • Zinc producers are removed from end-use decisions • Competition from engineering materials is intense • Government regulations impact on material selection.

Opportunities to grow zinc consumption through market development are numerous and significant. Most zinc-based markets are not mature; they are far from saturated. For example, general galvanizing is a growth industry; the market penetration of the galvanizeable steel market is low, less than 20 % according to some measurements. General galvanizing is price and performance competitive with premium paint systems. The market responds to promotion; a sustained growth rate of 5 % or more is experienced when investments are made in promotion. Market development should enable general galvanizing to gain market share against paint. Continuous galvanized sheet steel has been growing at a rate of about 4% during the last 20 years and this trend continues. The market share of zinc coated sheet is currently about 50 % of cold rolled steel sheet. Promotion is required to maintain and expand this market. The die-casting market has been recovering, in spite of limited promotion, from the trend to lighter and smaller parts and from the severe competition from competitive materials. A market study confirmed that zinc die-casting offers many positive mechanical, physical and process attributes that make it the material of choice for many applications. However, many part designers are unaware of zinc's die-casting attributes.

The potential for growth in emerging markets is evident when one compares consumption per head or intensity of use (Table III) in various regions of the world.

Table III - Consumption per Head and Intensity of Use Across the World (1997)

Region or Country

North America South America Asia/Pacific** China India Western Europe Central & Eastern World

Europe

Consumption per kg/person

4.0 1.1 3.0 0.7 0.2 4.8 1.5 1.3

Head Intensity of Use Tonnes per Billion US$ of

GDP, at PPP* 169 217 269 204 150 262 322 265

(*) GDP at purchasing power parity (ppp: a statistical adjustment for cost of living differences) (**) 11 countries

IZA's vision is to increase the use of zinc by 500,000 tonnes/year, over and above the natural growth in the market, in five years (1997-2002) through the support of initiatives which

32 LEAD-ZINC 2000

will expand the applications of zinc and through the transfer of zinc application technology to emerging markets.

IZA has taken a comprehensive approach to market development. A long-range plan setting philosophies, priorities, policies and processes has been developed. Programs are the heart of the market development plan and funding is directed to support selected programs. The approach put in place requires that:

• Programs are focussed on end-use markets • Actions are directed towards specifiers • Programs are developed and implemented in partnership with first-use zinc consumers.

To develop and oversee market development programs, the committee (Figure 4) is organized into World Committees for Continuous Galvanizing, General Galvanizing and Die-Casting. In each case, marketing and technical leaders from the zinc industry are co-ordinating actions and implementing programs which are expected to increase "the choice for zinc".

Chairman

WORLD COMMITTEES

DIE CASTING -

CONT. GALV. -

- GENERAL GALV.

REGIONAL / NATIONAL COMMITTEES

EUROPE h

ASIA/PACIFIC j -

LATIN AMERICA H

USA

INDIA

Figure 4 - Market Development Committee Structure

In addition, regional leaders have undertaken the task of ensuring that market development and technology transfer and taking place within the U.S.A. and Canada, Latin America, Asia-Pacific, Europe and India. The market development committee seeks to use the resources of the industry's regional development agencies, the skills of a world class research organization such as ILZRO and the technical and professional skills of the members in order to obtain the maximum return from the financial resources at its disposal. Programs, which are developed within each of the world or regional committees, must meet tonnage targets and are subjected to a rigorous selection and review procedure before being submitted for funding.

The current 21 programs are directed to three large market segments representing 63 % of refined zinc consumption; namely, continuous galvanizing, general galvanizing and die-casting. The 2000 market development budget is USS 1,150,000. At present, each dollar of IZA funds leverages six dollars from other sources. Typically work proceeds from market research studies and surveys to marketing tool development and to promotion. Further, market

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 33

development programs are backed by IZA's product research programs (US$ 1,200,000), which are led and managed by the International Lead Zinc Research Organization (ILZRO).

General Galvanizing

The joint program with the European General Galvanizers Association (EGGA) is probably the best way to illustrate the impact of our co-operative activities. Three market surveys have been conducted and, as dictated by the surveys, essential information on general galvanizing has been summarized into a set of facts files, comprising three to five pages each. The facts files are available in 11 languages, with two additional translations under consideration. Altogether 46,500 fact files have been distributed worldwide and more will be required. Ultimately, the performance of the market development programs is measured in tonnes. We are pleased to report that the general galvanizing component in the Western Europe program reached, on time (1993-1998), its original target of adding 1,000,000 tonnes of galvanized steel to the market, that is an extra 75,000 tonnes of annual zinc consumption. The full impact of the work conducted has yet to be felt.

IZA and the regional network of five zinc industry associations are formally linked to 21 galvanizers associations and recently they helped the creation of six new galvanizers' associations.

The U.S. market for general galvanizing has been growing, for the period of the 1990's, at 6.1 %/year because of the dynamic promotional work of the American Galvanizers Association. Their work seeks to educate decision-makers in government departments and specifiers.

In the Asia/Pacific region, market surveys of the Philippines, Malaysia and Taiwan have been completed and a survey of Thailand is underway. Promotional seminars on general galvanizing are being organized and national/regional programs are being developed.

A case study library is being developed, and at least 50 European case studies will be available on the Internet by the end of the year. An educational CD-ROM, developed by the French, Dutch and Belgian galvanizers associations is to be translated into English and Spanish. In the U.S.A, galvanized steel utility poles are being promoted.

Another important initiative is the Galvauto group. It is an industrial partnership working to develop hot-dip galvanizing for corrosion protection of automotive underbody components. It is targeting to add 2 kg of zinc on every car in Europe and the U.S.A. The partners are automotive manufacturers, equipment suppliers or chassis engineering companies, galvanizers who are capable of meeting the specific requirements of the automotive industry, and the international zinc industry. The group conducted a series of assessments and field studies on the performance of galvanized underbody components and it is promoting the use of galvanized steel by automobile manufacturers.

In 1997, a consensus was reached, by the Galvauto group, that it was time to develop a new galvanizing single-dip thin coating process. This program is managed by ILZRO. The University of Rome was given the task of developing a process to apply a 30μιτι coating to high strength steels. Researchers made an important discovery; they found the benefits of using a copper pretreatment before galvanizing with a Zn-5%A1 alloy. The new coating, for which patent protection is sought, is applied at a thickness of 30 μπι, much lower than typical

34 LEAD-ZINC 2000

galvanized coatings; the new coating should give 15 years of protection to steel components facing underbody environments.

Continuous Galvanizing

IZA's market development program for continuous galvanizing is designed to support regional steel industry marketing programs. This includes:

• Understanding growth opportunities in construction and automotive markets • Addressing coating issues that affect sheet market development • Creating regional and global market partnership with the steel industry • Educating end-users and specifiers • Facilitating information transfer across regions • Supporting new market growth.

A 1998 worldwide steel industry survey showed that galvanized steel is a value-added product which steel companies regard as strategically important. Substantial growth in consumption is expected in the construction industry and, outside the nearly saturated US market, in the automotive industry. Zinc's critical properties of corrosion protection, recyclability and cost mean that zinc is expected to remain the preferred material for corrosion protection of steel for the foreseeable future. Finally, the zinc industry needs to support coated steel markets by managing zinc-related environmental issues and assisting the development of markets for zinc-coated steel, particularly new markets.

The zinc industry sponsors and is a member of the decision committees of the US Metal Roofing Alliance (2000 Budget = US$ 3,500,000) and of the North American Steel Framing Alliance (2000 Budget = US$ 465,000). They are both targeting a 25 % market share of the huge residential housing market; a new growth market for steel. For the U.S.A. only, these two programs may add 40,000 tonnes of new zinc consumption per year by 2003. Considerations are given to take advantage of these programs in other regions.

IZA's support is helping the Light Gauge Steel Engineers Association (LGSEA) to expand its activities and membership. LGSEA exists to provide technical information about the use of light gauge steel in construction. Throughout the world, there are thousands of engineers who do not have experience with light gauge steel framing and are ill-equipped to support a builder's decision to use steel. The newness of steel as a residential construction material means that there is much to learn about designing and building steel homes. Over the past year, LGSEA membership has increased to 550, with members all over the world. The association has distributed over 30,000 newsletters, 16 technical publications; its seminars have reached over 1,000 engineers and building officials in the U.S.A.

IZA has published a brochure entitled "Zinc Coatings - Protecting Steel" which includes chapters on the environment, the mechanism of corrosion protection, continuous galvanizing, general galvanizing, other zinc coatings, and typical applications of zinc-based corrosion protection systems. The steel industry co-operated in the preparation of the brochure and it is distributing it as a part of the steel industry promotional activities particularly in support of light-gauge coated steel framing for residential construction.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 35

Two publications on zinc recycling "Zinc Recycling - The General Picture" and "Zinc Recycling - Zinc Coating Steel" have been published by IZA-Europe, in co-operation with EGGA and EUROFER.

An U.S. based information centre on the properties, performance and applications of galvanized steel sheet has been established; the Galvlnfo Centre is managed by ILZRO and is co-sponsored by IZA and the steel industry. Further, ILZRO acts as a leader and catalyst in the field of galvanized autobody sheet; it brings together 33 steel producers to sponsor and to direct R&D progams addressing technical challenges arising from the use of coated steel sheet by the automotive industry.

Die-Casting

Two marketing tools, ZincCast and DeZign, have been developed. ZincCast, a training tool for die-casters on CD-ROM, is now available in English and French and two thousand copies are currently being distributed. Spanish, German, Italian and Chinese versions are being developed. DeZign, a design guide for specifiers and designers, is available both on CD-ROM and on the Internet.

A network of market development technical support centers is being established; it involves machine manufacturers, high schools, universities, die-casters' associations and alloyers. Five centers have been established in Europe. Both ZincCast and DeZign are used to initiate the centers. Their mission is to provide technical assistance and training for die-casters.

In North-America, Interzinc focuses on trade shows, seminars, building awareness, information dissemination and college education. Because of regular complaints on the quality of the die-castings produced in the Asia/Pacific region, die-castings were purchased and analyzed against specifications; 18 % were found to be off-spec. An educational program towards die-casters and end-users will be launched to try to avoid these problems in the future and to increase awareness of the need for high quality products. IZA-Latin America organized a satellite conference last November that attracted 73 participants at seven sites in four countries: Mexico, Colombia, Peru and Venezuela. An internet-based "chat-box" enabled the participants to ask questions and make comments during the conference. A similar conference on general galvanizing was organized; it attracted over 200 participants.

Emerging Markets

Altogether, the activities sponsored by IZA are supporting programs and program development activities in 16 emerging markets. Major centers of activity are Mexico, India and the CIS. Here also, programs are directed to general galvanizing, continuous galvanizing and to the die-casting market segments and they benefit from the work done in established markets.

Measurements

Performance measurements are constantly being established for measuring the results of individual programs and of the market development plan. Investments in market research and market intelligence represent 5 to 10 % of the budget. Program teams are requested to set targets based on market research and intelligence and to propose performance measurements including a tonnage target.

36 LEAD-ZINC 2000

Figure 2 is one way to gauge the performance of the overall plan. The current growth rate of 3.2 %/year, compared to the 40 year trend growth rate of 2 %/year, could be interpreted as meeting the target of making a difference of 500,000 tonnes/year by 2002, over and above the natural growth in the market [8,000,000 tonnes x (3.2 %/year - 2 %/year) = 96,000 tonnes/year or 480,000 tonnes in five years]. However, this growth rate is influenced by the current period of economic prosperity, and hence, it is not considered adequate.

We prefer to follow the "intensity of use"; it is a measure that is less dependent on economic cycles and economic activities. Intensity of use calculates the amount of zinc in tonnes per billion dollars of economic activity. The curve since 1968 is shown in Figure 5. It is based on world refined zinc consumption; i.e., it does not include the growing proportion of recycled zinc. GDP rather than industrial production (IP) is used because world IP data are not available over the 1968-1999 period. The graph shows that the intensity of use dropped by about 50 %. Other metals have undergone a similar drop. The intensity of use of steel, for example, was also halved over the past 30 years. This is due to a multitude of factors including major advances in the efficiency of use of zinc and steel. The intensity of zinc use reached a low of 248 tonnes per billion US$ of GDP in 1993, but it has been growing ever since to reach 265 tonnes of zinc/billion US$ in 1999. Hence the intensity of use of zinc is currently growing at a rate of 0.8 per cent/year; i.e., world consumption is increasing by 62,000 tonnes/year or 310,000 tonnes in 5 years. This is short of our target, but progress is being made. Consumption per capita is another indicator that can be used. World annual per capita consumption has also been growing steadily at (1.8 per cent/year) since 1993 reaching 1.34 kg zinc per person in 1999.

tZn/US$ billion

600

500 — \

400

300

200

100

0

1968 1973

265

1978 1983 1988 1993 1998

Figure 5 - World Intensity of Zinc Use 1968-1999

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 37

CONCLUSIONS

The materials world is in constant evolution. Expanding knowledge, innovation, product performance, environmental considerations, cost and marketing efforts all contribute to important shifts among metals, polymers, ceramics and glasses. Each metal has a number of properties which combine to differentiate it sharply from other metals, and often times, a group of properties makes a metal the material of choice for a given application.

Zinc has a growing dominant application: galvanizing (50 %). Zinc consumption growth is primary linked to market development. Recognizing opportunities in market development, the zinc industry has organized itself worldwide to promote major end-uses, in co-operation with its customers. Zinc industry's efforts relative to market and environmental challenges are being pursued by the International Zinc Association, by a world network of national and regional associations and by the International Lead Zinc Research Organization. Specific approaches for the general galvanizing, continuous galvanizing and the die-casting markets are being implemented. Programs are delivering tangible results. Consumption is growing at the targeted rate of adding 500,000 tonnes, over and above natural growth (3.2 % vs. 2.0 % per year), in five years. However, growth according to a world intensity of use measurement would deliver, in 2002, 310,000 tonnes of additional consumption at the current growth rate. As the zinc industry advances in its program to characterize zinc's contribution to sustainable development, detailed information of the role and contribution of each application will be developed and published.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 39

LEAD AND ZINC: A STUDY OF TOXICOLOGICAL CONTRASTS AND SHARED REGULATORY CONCERNS

C.J. Boreiko International Lead Zinc Research Organization, Inc.

P.O. Box 12036 Research Triangle Park, North Carolina, U.S.A. 27709

ABSTRACT

Traditional patterns of lead and zinc utilization are being evaluated in accordance with new paradigms that seek to ensure the long-term compatibility of global industrial activity with precepts of sustainable development and sustainable consumption. As a consequence, both industry sectors are experiencing intense scrutiny from the international community, and numerous regulatory options are being considered that may impact upon international industry practice. This paper reviews the growing expectations being placed upon industry, and highlights some of the common regulatory concerns being expressed regarding both metals despite their differing toxicological profiles.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

40 LEAD-ZINC 2000

INTRODUCTION

Traditional patterns of natural resource utilization are being evaluated in accordance with the principals for sustainable development and consumption. In accordance with concepts formalized at the "Rio Conference" convened by the United Nations Conference on Environment and Development (UNCED) in 1992, and as articulated in Agenda 21, the acceptable production and use of metals by society is no longer based upon compliance with regional regulatory frameworks. Rather, the metal producing industry must now assume accountability for down-stream uses of their products on a global basis. International agencies are increasingly seeking assurances that the production, use and end-of-life disposition of metal-containing products do not pose undue risks to human health or the environment. Moreover, industry is being asked to bear the burden of proof that current practice preserves environmental integrity over timeframes that now extend beyond decades. Patterns of material consumption are expected to be compatible with sustainable development for future generations in a fashion that ensures the availability of resources and preserves the safety and well being of workers, the general population and the environment.

Various United Nations organizations and Conventions are presently seeking to identify and control materials that pose transboundary risks through trade or through emissions to air, soil and/or water. Regulation of trade in "hazardous wastes" is being addressed through the Basel Convention. The United Nations Economic Commission for Europe has worked to develop a Convention on Long-Range Transboundary Air Pollution, the initial targets of which were multiple non-ferrous metals but which ultimately developed international agreements pertaining to lead in gasoline and mercury. The Intergovernmental Forum on Chemical Safety (IFCS) was created, and it identified the need to control persistent organic pollutants (POP's). An international protocol restricting the production and use specific organic substances, including many pesticides, is being formulated.

The successful negotiation of a POP's protocol has suggested the need for a companion "PIP's" protocol for inorganic substances, targeting yet to be specified non-ferrous metals. The United Nations Convention on the Law of the Sea is beginning to examine the impacts of effluents from mining activity that might impact upon estuaries and coastal waters. Accompanying this is a global program of action for the Protection of the Marine Environment From Land Based Sources of Pollution that will begin to address inputs of multiple substances, including lead and zinc, from point and non-point sources. The inclusion of metal-related issues in these, and a host of other, initiatives will evolve over the course of the next several years. Lead and zinc will, to varying extents, be subject to intensive scrutiny.

A focus upon metals is inevitable, largely because of commitments made by the international community to the basic precepts of Agenda 21 and to principles of risk assessment. In 1994, the Intergovernmental Forum for Chemical Safety pledged to the conduct of risk assessment for high production volume substances. The international community made commitments for assessments to be conducted on 500 substances by the year 2000. Although this target is unlikely to be met, broad-based initiatives have been proposed in an effort to satisfy this obligation. For example, the Organization for Economic Cooperation and Development (OECD) has sought international harmonization of classification criteria, risk assessments and risk reduction strategies amongst its member states. Mechanisms are being established for the harmonization of risk assessment methodologies around the world and for the adoption of assessments performed by "regional authorities". For example, the deliberations of the United States Environmental Protection Agency resulted in OECD undertaking a Pilot Risk Reduction Program which evaluated the need for concerted

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 41

international action to address the risks posed by lead. Issues of lead risk reduction were thus the focus of six years of debate amongst the OECD member states, with an initial effort being made to apply broad prescriptive measures to the industry that would harmonize international regulations governing industry practice and product applications. The OECD ultimately adopted a recommendation which took into account differing national priorities, policies, programs and achievements. A Ministerial Declaration on Lead Risk Reduction was issued that focused the attention of the international community upon multi-media strategies to identify, and reduce as necessary, harmful human exposures to lead.

More recently, the Existing Substances Program of the European Union has undertaken an evaluation of high production volume substances used in its member states. Risk assessments performed under the auspices of the Existing Substances Program are now automatically forwarded to OECD for review and action. Within this context, an evaluation is presently being conducted on zinc and five zinc compounds. Just as risk assessments conducted in the United States created concerted OECD action pertaining to lead, the outcome of the EU Existing Substances Program could place zinc upon the agenda of the OECD for international risk reduction activity.

Whereas the lead industry has grown accustomed to international scrutiny, and adjusted its industrial practices and product applications accordingly, the zinc industry has maintained a sense of complacency regarding pending risk assessment activity. As an established industry with a diverse array of product applications of proven utility in modern commerce, it has been tempting for the zinc industry to presume that its products are, and shall continue to be, valued for their cost-effective contributions to modern technology and to the maintenance of the public infrastructure. Zinc and lead have distinctly different characteristics and the fate of lead should have little bearing upon that of zinc. The toxicological profile of lead would appear to justify stringent regulation. More discomfiting have been the actions of end-use sectors choosing to distance themselves from the use of toxic materials such as lead by implementing voluntary substitution programs in the absence of actual risk so as to avoid the use of materials that are "toxic".

Until recently, the actions of regulatory agencies have not had significant impact upon the zinc industry beyond common sense measures that ensure occupational safety and environmental protection. The industry is only now recognizing that the new precepts of sustainable development and sustainable consumption are creating a focus upon the zinc industry. The extent to which zinc-related issues are included in future international deliberations, and the subsequent outcome of these deliberations, will be dependent upon the perceptions developed by the international community concerning zinc and the industries that produce and use zinc. This will in turn be influenced by the process of risk assessment. With an array of important and "practically essential" applications, the zinc industry has been slow to recognize that being "practically essential" is not the same as being indispensable. Indeed the broad array of product applications enjoyed by zinc are the source of growing concern since most are, to varying extents, capable of producing environmental dispersion. Use of zinc by diverse industrial sectors further means that large numbers of workers have potential exposure and that opportunities exist for consumer exposure. The zinc industry must be alert to the evaluation process that is presently being undertaken and work to ensure that a balanced assessment occurs.

42 LEAD-ZINC 2000

THE TOXICOLOGY OF LEAD AND ZINC

Although primary zinc and lead are cogeneration products and, from a strictly metallurgical perspective, are "heavy metals", their toxicological properties are highly divergent. Table I provides a summary overview of the contrasting human health profiles of the two metals. Zinc is an essential nutrient with a known required daily intake and concentrations in blood which must be maintained in order to ensure good health. Occupational exposure limits exist for zinc, but are near those required for nuisance dusts. Finally, there are drinking water limits for zinc, but these are indexed to aesthetic properties such as bitter taste and not to any health related endpoint (1).

Table I - Summary of Human Health Profiles Zinc Lead

Classification Essential nutrient Nonessential, toxic Concentration in blood 13 μπιοΙ/L required 0.5 μπιοΙ/L child limit Diet intake 15 mg/day required 37.5 μg/day limit Water limits 5 mg/L (taste) 15-50 μg/L (health) Occupational air 5-10 mg/m3 50-200 μg/m3

In contrast, lead is a non-essential element that can be toxic. Limits have been recommended for maximum exposure via the diet. Similarly, limits for blood lead levels in the general population have been suggested and health based standards exist for lead in drinking water and for exposure in the occupational setting (2).

The contrasting toxicological profile of the two metals deserves further examination. The statement that "zinc is essential" is derived from nutritional science and connotes a specific set of properties that is shared by few other metals (1). As an essential trace mineral nutrient, zinc is required for the growth and reproduction of virtually all organisms. Biological functions served by zinc are unique and cannot be replaced by any other substance. Atoms of zinc are intimately involved in metabolism and play a role in processes as divergent as the immune system, reproduction and the functioning of the nervous system. More than 300 enzymes have been identified that require zinc for activity, making zinc one of the most important metals utilized within living organisms. Finally, tissue levels of zinc are under homeostatic control -organisms carefully regulate the uptake and excretion of zinc, using these processes to adapt to periods of deficiency and/or excess exposure and to maintain constant internal concentrations.

Although homeostatic controls exist, prolonged dietary deficiency for zinc can have a variety of adverse physical consequences as summarized in Table II. Symptoms of severe zinc deficiency include growth retardation and increased susceptibility to infection (1). More recent studies have demonstrated that zinc deficiency associated with infant diarrhea may be a significant contributor to infant mortality in developing countries and that supplementation with zinc for may enhance survival rates (3). Recent studies have further shown that zinc deficiency is associated with impaired mental development in children and that supplementation with appropriate levels of zinc can reverse these adverse effects (4). Finally, zinc deficiency is now believed to extend beyond developing countries and to the Western world. A significant proportion of very young and the elderly in countries such as the United States is suspected to be at risk of zinc deficiency (1).

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 43

Table II - Symptoms of Zinc Deficiency • Retarded growth • Delayed sexual maturation • Mental lethargy • Anorexia • Pica • Rough, dry skin • Slow healing • Increased infection

Excess zinc exposure, if intense enough or of sufficient duration, can have adverse impacts upon health (1). For example, metal fume fever is association with inhalation exposure to grossly excessive amounts of molten zinc vapor and/or ultra-fine zinc oxide particles. Although metal fume fever can be induced by virtually any metal, the relatively low boiling point of zinc has resulted in greater occurrence of metal fume fever with zinc than for most other metals. Fortunately, metal fume fever is a short-term flu-like illness, characterized by fever and flu-like symptoms which last approximately 24-72 hours. No long-term health effects have been associated with metal fume fever.

Short-term oral intake of large amounts of zinc can also result in nausea and gastrointestinal upset. Such symptoms typically require ingestion of several hundred milligrams of zinc and, while producing temporary discomfort, are similarly not associated with any long-term health consequences. Prolonged exposure to elevated levels of zinc, usually via oral intake, can be associated with subtle metabolic alterations. Although sometimes referred to as "zinc toxicity", such changes are in fact alterations in trace mineral metabolism best viewed from a nutritional perspective. When any single nutrient in the human diet is taken in gross excess to others, imbalances will result. Excess daily intakes of zinc on the order of 100 mg/day or more over a period of months will begin to produce clinical symptoms of copper deficiency. Prolonged dietary imbalance will yield perturbations in copper and iron metabolism that can result in anemia. Finally, case reports exist in which anemia has been clinically significant. Such symptoms, and indeed the impacts of any "toxic effects" of zinc, are readily reversed when exposure is reduced and/or balance nutrition is resumed.

The toxicological profile for lead is in sharp contrast to that of zinc. Exposure to lead is known to be associated with a variety of significant and potentially irreversible health impacts (2). The effects of high level exposure to lead are summarized in Table III. These effects include lethality, massive central nervous system breakdown (encephalopathy) and kidney effects. Although this table includes the exposure ranges that are associated with specific health effects, it is important to recognize that individual variability with respect to responses to lead is very significant. The values depicted in the table are generally the lower values at which impacts are observed. Higher exposures can occur in individuals who are, to all outward appearances, asymptomatic.

44 LEAD-ZINC 2000

Table III - Lead Health Effects. High Exposure All Effects Vary Greatly with the Individual

Lethality: Children: PbB > 125 μg/dL Adults: PbB > 300-400 μg/dL

Encephalopathy: Children: PbB > 80 μg/dL

Adults: PbB > 100 μΒ/dL

Colic: PbB > 60-100 μg/dL

Kidney Effects: Prolonged exposure

> 60-70 μg/dL

Anemia: PbB > 50 μg/dL

Male Reproduction: Impaired at high exposure, subtle changes > 40 μg/dL

Female Reproduction: Miscarriage at high exposure, subtle effects (birth weight, fetal mental development) possible below 30 μg/dL

A variety of other health concerns are sometimes raised with respect to high-level lead exposure. Lead is classified as a possible human carcinogen based upon the fact that exposure of animals to high levels of lead will produce cancer, usually of the kidney. However, epidemiology studies of occupationally exposed workers have failed to observe lead-related cancer and this endpoint is thus seldom the focus of regulatory activity. For other endpoints, such as impacts upon the immune system or altered liver function, there is little evidence for significant effects. Recent studies have begun to focus upon neurological and neuropsychological effects of lead in occupationally exposed adults. Subtle and presumably reversible impacts of lead can be observed at or near the occupational exposure limits in place in some countries. Health professionals are currently investigating whether or not significance should be attributed to these observations. Finally, although much publicity has been directed to the issue of endocrine disrupting substances, there is no evidence that lead can function as an endocrine disruptor. Although perturbations can be observed in hormone systems at high levels of lead exposure, these perturbations are side effects of systemic toxicity and not specific endocrine disrupting effects.

Low-level exposures to lead are associated with a variety of other health endpoints of concern (Table IV). A number of studies have demonstrated that exposure to lead is statistically associated with a one to three IQ point decrement in intelligence for a 10 μg/L increase in blood lead above a baseline of 10-15 μg/dL. It is important to recognize that effects produced by this level of exposure are extremely subtle and not detectable at the level of the individual. Simply put, a deficit of 1 or 2 IQ points does not translate into functional changes that can be measured by current psychometric techniques. However, such changes are postulated to have significance if extrapolated to a large number of individuals or across an entire society. Actual demonstrations of such "societal impact" have not been reported and remain the subject of speculation and scientific debate.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 45

The so-called "level of concern" for blood lead levels in children primarily refers to post-natal blood lead levels, or more specifically to the child aged 1-6 years (5). Pre-natal exposures to lead (exposure in the womb), although initially of concern, no longer appear to be as critical as those that occur after birth. The "level of concern" in place in countries such as the United States is often misinterpreted as an absolute exposure limit. Given that the effects of low-level lead exposure have minimal significance for the individual, policy in many jurisdictions has focused upon maintaining the majority (e.g., 95%) of individuals below this level.

Current health research is attempting to address the issue of whether or not there is a threshold for the impact of lead upon child development. Although it is often claimed that "there is no threshold for the effect of lead upon child intelligence", this somewhat hyperbolic statement should more appropriately be viewed as reflective of limitations inherent in current neurobehavioral research. The effects of lead at low levels of exposure are extremely subtle and current analytical techniques, both in the measurement and statistical analysis of psychometric impacts, are unable to discriminate the presence or absence of effect thresholds (5). For some, the "level of concern" thus represents a practical threshold for the setting of public health policy. If effects exist below this level, they are likely quite small in magnitude.

Potential impacts of lead upon blood pressure have also been of concern in the consideration of low-level exposures. Recent meta-analyses have suggested that a 1 mm Hg increase in blood pressure is associated with a doubling of blood lead (6). As with the neurobehavioral issues, the significance of such a small change for the individual is minimal, but has potential significance if extrapolated to large populations. Hypothetically, modest increases in blood pressure could increase the prevalence of hypertension and subsequent cardiovascular disease in a population. However, it is important to recognize that this hypothetical concern has never been validated by any observation of increased cardiovascular disease or hypertension in the general population.

Table IV - Low Level Lead Exposure Issues Child Intelligence: · 1 -3 IQ point decrement with a 10 μg/dL

increase in blood lead above a 10-15 \igldh baseline

• Effects are subtle - not detectable in an individual

• Could have "societal significance" • Effect threshold not known and difficult to

detect

Blood Pressure: · 1 mm Hg blood pressure increase associated with a doubling of blood lead

• Significance for the individual is minimal • Causality uncertain • Societal significance also uncertain • Thresholds not easy to detect with current

methods

46 LEAD-ZINC 2000

The effects of low-level lead exposure will continue to be an issue upon which reasonable people disagree. Several factors contribute to the continuing scientific debate in the US and Europe. The effects of lead at low levels of exposure are subtle and extremely difficult to detect. Often times, they represent less than 1% of the variance present in a population and are expressed in the presence of strong confounders. In essence, there are many lifestyle factors and environmental circumstances that are known to impact upon lead exposure, neurobehavioral development and blood pressure (Table V). Adequate correction for these confounders, particularly when they exert effects up to an order of magnitude greater than that being attributed to lead, is critical but extremely difficult to conduct. Thus, the presence or absence of a "lead effect" in a given study can be a function of the fashion in which confounder correction is attempted. For example, correlations of childhood development must consider important confounders such as socio-economic status, parental intelligence, parental education, and overall health status. Similarly, studies of lead and blood pressure must consider hereditary factors, dietary habits, alcohol consumption, and physiological parameters such as hematocrit and body mass index. Some scientists suggest that our understanding of confounders is incomplete and that additional confounders exist that should be measured in future scientific studies. The debate thus often times focuses upon what is a lead effect and what is the possible impact of residual confounding. For lead impacts upon blood pressure, in particular, many believe that the statistical associations reported are not causal in nature and are the products of residual confounding.

Table V - Lead and Confounding Endpoint Known or Suspected Confounders I.Q.:

Blood Pressure: «

> Parental IQ > Diet > Health History > Parental Education > Family Integrity > Maternal Depression » Home Environment ► Socioeconomic Status ► Birth Order

» Age » Body Mass Index » Alcohol Intake » Smoking > Stress > Calcium Status • Exercise » Hormonal Status » Blood Hemoglobin

Although the toxicological perspective of lead differs markedly from the nutritional perspectives which characterize zinc, risk assessment has yet to adequately establish methodologies for addressing essential nutrients (7). Indeed, the "rules" which govern nutritional recommendations and the conduct of toxicological risk assessments are sufficiently divergent so as to potentially create problems, even for a substance such as zinc. For example,

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 47

the recommended daily intakes (RDI) of zinc for adults have been set by the nutritional community at 12-15 mg/day. This recommendation has been put forward to avert the significant adverse deficiency effects noted earlier. The actual nutritional requirement of most adults will be satisfied by as little as 6-8 mg/zinc per day (1). However, a modest safety factor is incorporated into dietary recommendations so as to accommodate inter-individual variability.

Risk assessments are conducted under a far different set of guidelines. One of the first efforts to establish an upper limit for zinc intake was proposed by the Environmental Protection Agency (EPA) in the United States during their development of a reference dose for zinc. The reference dose, or RfD, is the level of substance intake that can be sustained for a lifetime without ill-effect. In setting an RfD the toxicologist tends to take a cautious approach. Copper and iron deficiency are known to be induced by high levels of zinc administration; for example, a daily intake of 50 mg/day of supplemental zinc has been reported to produce modest inhibition of a red blood cell enzyme called Superoxide dismutase. Whether or not this is a "toxic effect" remains a matter of conjecture to this day (8, 9). However, toxicologists, in the interest of public health, have tended to interpret this as a potentially adverse effect and 50 mg per day of supplemental zinc intake has become a point of reference for "toxicity".

In addition to exercising caution with respect to what is a toxic effect, toxicologists further apply large safety factors to their observations. These large safety factors are designed to protect all individuals in the general population and are generally on the order of a factor of 10 or more. Thus, the first proposals for an RfD for zinc from EPA suggested that oral intakes of zinc should not exceed 5 mg/day. Had it been implemented, this guideline would have resulted in the banning of vitamin pills and, for individuals who managed to live in compliance with these guidelines, would have induced zinc deficiency in most of the US population. A smaller safety factor was subsequently used and an RfD of 21 mg/day adopted. This RfD is not too distant from the recommended daily intake of 12-15 mg/day and (on a body weight basis) is actually less than the RDI for pregnant and lactating women.

The relatively small difference between the RDI and RfD exists in part because of the conflicting philosophies with which nutritional assessments and risk assessments are conducted (7). However, the RfD can also be interpreted as an indication that little additional exposure should be permitted above and beyond that which is acquired from the diet. This relatively narrow margin of additional exposure can thus lead to the conclusion that occupational exposures to zinc must be sharply limited. During the course of the ongoing EU Risk Assessment for Zinc, concern is being expressed that occupational exposures exceed the toxicologically derived "maximum tolerable daily intake" value which is close to the recommended daily intake level. Whether or not such toxicologically derived exposure limits are appropriate for setting occupational exposure standards will be discussed over the course of the next year or two. Resolution of such issues is clearly in the interests of the zinc industry and serves to demonstrate that even a beneficial substance such as zinc can become the source of regulatory concern when the conservative guidelines of risk assessment are applied to an essential nutrient.

ENVIRONMENTAL ZINC AND LEAD

Matters of human health tended to dominate public health discussions until the past decade. There is now increasing awareness that ecological and environmental issues must be taken into consideration when evaluating the compatibility of lead and zinc with sustainable development. As with human health, there are distinct differences between lead and zinc. All

48 LEAD-ZINC 2000

organisms require zinc as an essential trace element, but no essentiality requirement has been validated for lead. The two substances also differ significantly with respect to environmental inputs. In essence, the human toxicological profile of lead has created an end-use profile more conceptually compatible with sustainable development. In contrast, the favorable image accorded to zinc has permitted multiple end-use applications that may create environmental dispersion. Significant challenges are thus posed for the zinc industry and threaten to impact upon traditional end-use applications.

The significant publicity accorded to lead's toxic profile for humans has resulted in sharp restrictions upon the emissions of lead from point sources and has further restricted diffuse emissions. Lead applications that are environmentally dispersive (e.g., lead in gasoline) have been eliminated, or are being phased out in most countries. Lead also has an extremely high recycling rate (nearly 90%) which further aids in the reduction of emissions to the environment. As a result, existing environmental issues for lead mainly stem from historical contamination problems. Zinc on the other hand is facing scrutiny because of the dispersive nature of multiple uses of zinc and their inputs into the environment. Applications of zinc such as galvanizing, agricultural fertilizers, and as a vulcanizing agent in vehicle tires can result in varying levels of input to the environment. The fate and effects of environmental zinc, if any, are not fully understood. Although zinc is essential for all living organisms, the uptake levels required for optimal environmental health are not rigorously characterized. Further complications arise when consideration is given to issues of bioavailability. Very little zinc in the environment is present in ionic or soluble form. Basic thermodynamics favor the complexation and immobilization of zinc in soils and in aquatic sediments. Thus, although the concentrations of zinc in natural waters are measured in terms of "parts per billion", the concentrations of zinc in sediments underlying these waters will be a thousand-fold or more higher. Finally, the geochemical factors that dictate bioavailability exhibit extreme site specificity. This means that no single numerical value can be identified for total environmental zinc levels that either satisfies essentiality or guards against possible toxicity. Regulatory limits for zinc are thus set using traditional risk assessment parameters to guard against toxicity, with minimal understanding of essentiality. The regulation of lead is solely based upon toxicity. As a consequence, both metals end up being regulated in a very similar fashion.

For both metals, aquatic toxicity is dictated by the amount of metal present in a soluble or ionic form. This is the form of the metal that is bioavailable and capable of interacting with biological organisms. Water quality parameters that modify the bioavailability of zinc and lead in water, listed in order of importance, include pH, dissolved organic carbon and hardness. Current water quality criteria are often based upon the dissolved concentrations of the metal, after mathematical calculation of the effects of only a limited set of these parameters. In the United States, for example, water quality criteria are based upon dissolved concentrations using only hardness to normalize for bioavailability (10). Water quality criteria for lead and zinc are shown in Table VI.

Table VI - Acute Water Quality Criteria for Lead and Zinc Metal USEPA Water Quality Criteria Zinc 120 μg/L Lead 82 μg/L

*Both values calculated using a water hardness of 50 mg/L

In the real world, levels higher than these would likely be required to produce toxicity since factors such as dissolved organic carbon in natural waters serve to attenuate possible

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 49

ecological impacts. Moreover, models exist that permit the derivation of standards which incorporate such factors into site-specific emission limits (11). However, the regulatory community has no incentive to derive such limits. Rather, it is the industry that must take it upon itself to generate the information required to support more sophisticated and mechanistically-based standards. More accurate risk assessment is indeed possible, but only if the industry makes a commitment to generate the science required to support the risk assessment process.

Issues of bioavailability similarly dictate the toxicity of both lead and zinc in sediments (12). In anoxic sediments the most significant factor controlling bioavailability is the acid volatile sulfide content (e.g., iron sulfide, manganese sulfide). Research has shown that if the total amount of acid volatile sulfide (AVS) is greater than the total amount of simultaneously extracted metals (SEM), then no toxicity will occur. Another physical chemical factor that will modify the toxicity of lead and zinc in sediments is organic carbon, both in sediments and pore water. Incorporation of such parameters into the derivation of sediment quality guidelines is feasible and in the industry's best interest. Again, however, there must be commitment on the part of industry to generate the information required for sound site-specific risk assessment.

Similar to the toxicity of zinc in the aquatic environment (water and sediments), the toxicity of zinc and lead in soils is dependent upon the soil parameters controlling bioavailability. The key soil physico-chemical properties controlling bioavailability include pH and organic carbon. Cation exchange capacity has also been shown to be an important factor in some cases. Efforts to model the terrestrial impacts of zinc are in an earlier phase of development than for other environmental compartments. However, considerations of toxicity in terrestrial ecosystems are emerging to be of significant concern to regulatory agencies.

ZINC, LEAD AND SUSTAINABLE DEVELOPMENT

This paper has summarized the toxicological properties of lead and zinc for human beings and for the environment. Such considerations will play an important role in future debates on sustainable development and will pose issues which must be addressed if the viability of both industry sectors is to be ensured. In an era where "precautionary" philosophies are applied in the face of scientific uncertainty, industry is placed at a distinct disadvantage unless efforts are made to supply the detailed information required for sound risk assessment. The conduct of research is particularly acute for the zinc industry since many zinc applications are environmentally dispersive and may result in elevations of zinc levels in soil, water and/or sediments. Indeed, significant environmental damage produced by historical point source emissions has provided industry with a legacy which demonstrates that metals can have environmental impact. Although present industry practices and product applications do not result in environmental loadings remotely comparable to those recorded to produce impacts around point sources, it is difficult to identify with precision those levels of environmental emission which are acceptable. Thus it is that Denmark can propose a ban on most existing applications of lead, predicated on the concern that, over the course of almost 1000 years, adverse increases in soil lead may result. Debates presently being initiated under the Existing Substances Program of the European Union are focusing upon the multiple diffuse sources of zinc and concern is being expressed that environmental impact is being produced, or will result in the future. Although the European theater is the predominant focal point for diffuse source concerns, the issue is not restricted to Europe. Similar questions and concerns are beginning to be voiced by regulatory agencies in North America, Asia and Australia.

50 LEAD-ZINC 2000

The basic policy instruments being developed by the international community are assessing the compatibility of multiple industry sectors with new precepts of sustainable development and consumption. The challenges this poses for the metals industry are both complex and daunting. There are natural fluxes of metals in the environment, and metals such as zinc are indeed essential to life. However, industry is being asked to prove that anthropogenic inputs to ecosystems are not producing significant perturbations in natural processes. The risk assessment burden confronting the international regulatory community is such that simple solutions will be sought to complex problems. In the absence of a commitment from industry to provide the means and information required to conduct complex scientific risk assessments, policy will be formulated in a fashion that reflects uncertainty and assumes the worst.

The zinc industry should take a lesson from lead in this regard. In the face of uncertainty, the regulatory community will adopt a policy of "materials cycle optimization" in which the only safe policy that can be pursued is one in which product applications that cannot be completely recovered for recycling must be discontinued (13). Thus, it is the concern over toxicity and dispersive uses that has resulted in the lead industry being almost totally dependent upon the lead-acid battery for its economic viability. Although the demand for portable energy storage remains strong, lead remains restricted with respect to the applications to which it can be applied. Numerous promising applications of lead, such as the active ingredient in marine anti-fouling paints or as an anti-oxidant to prevent the hardening and cracking of bitumen for roadways, were abandoned because of environmental concerns. If similar criteria were applied to the applications permitted for zinc, the impact upon existing markets would be significant.

The industry must also be aware that the metrics of toxicity are only one component of the sustainable development equation. The international community seeks to identify the means by which the prosperity and well-being of the world's population is enhanced. This entails cost-effective utilization of existing resources and the promotion of product applications that work towards this end. Opportunities exist for both lead and zinc when this broader and more appropriate perspective is adopted. For example, the lead industry has committed to the development of advanced lead-acid batteries for the anticipated growth of electric vehicle markets. Through the Advanced Lead-Acid Battery Consortium (ALABC) the industry has been able to achieve significant advancements in the development of batteries that have longer life, lighter weight and rapid rechargeability, all in a relatively low cost, maintenance-free package. The ALABC is currently expanding its market's interests to include batteries for hybrid electric vehicles, remote area power systems (RAPS) and even the automotive starting, lighting and ignition battery. Virtually all of these new potential markets, in one way or another, represent a repositioning of lead as being a commodity capable of making a positive contribution to sustainable development. Electric vehicles, or hybrid electric vehicles, offer reductions in hydrocarbon-based emissions and more efficient energy utilization. The development of remote area power systems has the potential to help alleviate poverty in areas without electricity. Moreover, through the development of systems which are reliant upon renewable energy, such as solar and wind power to supply RAPS energy, these systems are environmentally friendly and concordant with sustainable development.

Similar opportunities exist for the zinc industry. For example, one of the fastest growing markets for zinc has been the use of zinc-coated steel in automobile bodies. However, there have been recent challenges to this market because of the demand for lighter weight and more fuel-efficient vehicles. Automobile manufacturers have been evaluating lighter-weight potential replacements for steel in autobody panels. The steel industry has been responding to these expectations through efforts such as the Ultra Light Auto Project which is using high strength steels and ultra-high strength steels to improve the structural performance and reduce

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 51

the mass of sheet steels used in automobile construction. The zinc industry has an opportunity to contribute to this effort through the development of zinc coating systems that are compatible with the goals of the Ultra Light Auto Project. Thus, through the ILZRO, 12 zinc companies have joined in partnership with 29 steel and automotive companies in the Galvanized Autobody Partnership (GAP), a comprehensive program of research and development seeking to ensure that the steel industry can produce high quality zinc-coated steel sheet for the automotive market in a cost-effective fashion.

The primary focus of the metals industry has, for many decades, been on increasing the efficiency with which ore bodies can be extracted and the technologies that can be employed for cost-effective metal production. The technological sophistication and metallurgical expertise of the zinc and lead industry has reached new heights. Unfortunately, the criteria that the industry uses to judge itself are not the criteria by which the international community will judge the industry. The paradigms of sustainable development and sustainable consumption will determine the future viability of the non-ferrous metals industry. Recognition of the changing matrix by which industry performance will be evaluated provides the industry with both the opportunity to reposition its markets and to shape the policies that will govern future natural resource utilization and reuse. The lead industry is moving to meet this challenge, but the actions of the zinc industry have been less cohesive.

There are multiple avenues that the zinc industry can pursue to demonstrate that it is capable of making a positive contribution to the laudable goals of sustainable development and sustainable consumption. Regrettably, relatively few zinc producers have shown interest in making substantive contributions to such efforts. Whether it be the support of research to facilitate accurate risk assessment or the design of more effective products, it is encumbent upon the industry to recognize the new challenges being posed by the international community and to respond to these challenges in a proactive fashion. Innovation, strategically directed towards coping with sustainable development goals, will be the key to the survival and health of the lead and zinc industries in the decades to come. This entails a cohesive and collaborative approach to new market development, technical research to retain key markets, and toxicological research to ensure that the guidelines governing markets are appropriate.

REFERENCES

1. C. Walsh, H.H. Sandstead, A.S. Prasad, P.M. Newberne and P.J. Fraker, "Zinc: Health Effects and Research Priorities for the 1990s," Environmental Health Perspectives Vol. 102, 1994, 5-46.

2. International Programme on Chemical Safety (IPCS), Environmental Health Criteria 165 - Inorganic Lead, published under the joint sponsorship of the United Nations Environment Programme, the international Labour Organisation and the World Health Organization, Geneva, Switzerland, 1995.

3. Z. Isani, S. Nizami and Z. Bhutta, "Zinc Supplementation in Malnourished Children with Persistent Diarrhea in Pakistan," Pediatrics, Vol. 103,1999,1-9.

4. X.C. Chen, J.S. Li, H.H. Sanstead and F. Zhao, "Effects of Repletion with Zinc and Other Micronutrients on Neuropsychologic Performance and Growth of Chinese Children," American Journal of Clinical Nutrition. Vol. 68,1998,470S-475S.

52 LEAD-ZINC 2000

5. S. Pocock, M. Smith and P. Baghurst, "Environmental Lead and Children's Intelligence: A Systematic Review of the Epidemiological Evidence," British Medical Journal. Vol. 309,1994,1189-1197.

6. J. Staessen, C.J. Bulpitt, R. Fagard, R.R. Lauwerys, H. Roels, L. Thijs and A. Amery, "Hypertension Caused by Low-Level Lead Exposure: Myth or Fact?," Journal of Cardiovascular Risk. Vol. 1,1994, 87-97.

7. G. Nordberg, B. Sandström, G. Becking and R.A. Goyer, "Essentiality and Toxicity of Trace Elements: Principles and Methods for Assessment of Risk from Human Exposure to Essential Trace Elements." Journal of Trace Elements in Experimental Medicine. Vol. 13,2000,141-153.

8. C. Davis, D.B. Milne and F.H. Nielsen, "Changes in Dietary Zinc and Copper Affect Zinc Status Indicators of Postmenopausal Women, Notably Extracellular Superoxide Dismutase and Amyloid Precursor Proteins," American Journal Clinical Nutrition. 2000, in press.

9. D.B. Milne, CD. Davis, L.M. Klevay and F.H. Nielsen, "Low Dietary Zinc Exacerbates Signs of Copper Deprivation More Markedly than High Dietary Zinc in Postmenopausal Women," 2000, submitted.

10. United States Environmental Protection Agency, "Water Quality Criteria Documents for the Protection of Aquatic Life in Water." 1996.

11. P. Chapman, I. Thornton, G. Persoone, C. Janssen, K. Godtfredsen and M.N.Z. Graggen, "International Harmonization Related to Persistence and Bioavailability," Human and Ecological Risk Assessment. Vol. 2,1996, 393-404.

12. G. Ankeley, D.M. DiToro, DJ. Hansen and W.J. Berry, "Technical Basis and Proposal for Deriving Sediment Quality Criteria for Metals," Environmental Toxicology and Chemistry. Vol. 15,1996,2056-2066.

13. R.U. Ayres, "Toxic Heavy Metals: Materials Cycle Optimization," Proceedings of the National Academy of Sciences. Vol. 89,1992, 815-820.

Chapter 2

Modern Lead Smelting Technologies I

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 55

PRIMARY LEAD PRODUCTION - A SURVEY OF EXISTING SMELTERS AND REFINERIES

A. H.-J. Siegmund RSR Technologies, Inc.

2777 Stemmons Freeway, Suite 1800 Dallas, Texas, U.S.A. 75207

ABSTRACT

Since the last survey of lead smelters in 1987, the lead industry has experienced many changes. Technological innovations, in combination with a dramatic shift in the market structure as well as more stringent government regulations, caused major upheavals in many different areas. Therefore, a new survey was performed to review the progress in the technology of lead recovery over the least decade. The newly conducted survey reflects all current phases of the world's primary lead production, from charge preparation to refined metal shipment. It was prepared by sending questionnaires to individual companies, and the detailed statistical data of this survey are presented. In considering the data developed from the survey, it should be recognized that the primary lead industry is at a crossroad where novel technologies will gradually substitute the sinter machine - blast furnace operation in the new millennium.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

56 LEAD-ZINC 2000

INTRODUCTION

The last detailed TMS survey of primary lead smelters was carried out in 1987 (1). Since then, the decade of the nineties was characterized by dramatic changes in the lead industry caused by economic difficulties, the implementation of more stringent environmental regulations by the authorities, a significant shift in the market structure of the consumption of lead and the implementation of modern smelting technologies on an industrial basis. Therefore, the organizing committee of the Lead-Zinc 2000 Symposium considered performing a second survey to provide general up-dated operating data from lead smelting operations worldwide. As originally conceived by the committee, the survey paper was to present a detailed statistical overview of the primary and secondary lead industries describing the current status of existing lead smelter operations and the progress of the technology of lead recovery over the least decade.

In order to carry out such a large task, a significant amount of operating data was required, and a prepared questionnaire was sent to 41 identified primary smelters and 49 secondary smelters around the world. Eighteen completed questionnaires from primary plants and 14 from secondary operations were received. In addition, two companies reported on their refinery operations. Unfortunately, the committee had to recognize that the secondary lead industry especially, but also some primary lead smelting companies, are very protective about their sometimes obsolete and technically well-known technologies. Except for the secondary smelters belonging to the Quexco-group (12 operations from RSR corporation and Eco-Bat Technologies PLC) as well as the companies Campine NV in Belgium and Kovohute Pribram A.S. (Czech Republic), no further secondary operations responded to the questionnaire. This is somewhat surprising, because in times of dramatic change, one would expect that these companies would be looking for a close and open dialogue with others already applying modern technologies in order to avoid unnecessary expenses for research and development. Several cases in the past demonstrated that communication and co-operation result in cost savings and that this approach is more effective than strictly competitive concerns. Because of the competitive concerns of these companies, the committee therefore decided not to publish the operating data from the available questionnaires in the secondary lead industry and to focus the survey exclusively on primary smelters.

Companies operating new technologies were, except for two smelters, very open and were prepared to share their achievements. It seems that the Kivcet, the QSL, and the Kaldo processes are well established and mature technologies that are introducing a new era of lead smelting. The Ausmelt or the similar Isamelt technology, which is well established in the copper industry, could not be integrated into the survey because of a lack of comparative information. From the three smelters using the latter technologies for the treatment of lead bearing materials, one is permanently closed and the two in operation were not prepared to participate in the survey. A general description of the Ausmelt plant at Metaleurop Weser Blei in Nordenham, Germany is, however, presented elsewhere in this volume.

The eighteen primary lead smelters taking part in the survey, including their geographical distribution and capacity, are detailed in Table 1. The relevant questionnaires are summarized in Appendices 1-4.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 57

Table I - Surveyed Lead Smelters

Lead Refinery Country Type Capacity

Mount Isa Mines Limited Pasminco Port Pirie Smelter

Hachinohe Smelting Company Sumitomo Metal Mining Co., Harima Works Hosokura Smelting & Refining Co. Kamioka Mining & Smelting Co. Toho Zinc Co., Chigirishima Smelter

Korea Zinc Co. Ltd., Onsan Smelter

Hindustan Zinc Co., Chanderiya Smelter

Asarco Inc., East Helena Plant

Cominco Ltd., Trail Operations Noranda Inc., Brunswick Smelter

Australia Australia

Japan

Japan Japan Japan

Japan

Korea

India

U.S.A.

Canada Canada

S & B S & B / P y

ISF ISF / El

S & B / El S & B / El

S & B / El

QSL / El

ISF/Py

S & B

Kivcet / El S & B / P y

160 250

45 40 30 34

100

120

35

75

120 110

Met-Mex Penoles S.A. de C.V., Torreon Smelter Mexico S & B / Py 180

Berzelius Stolberg GmbH MHD-M.I.M. Huettenwerke Duisburg GmbH Norddeutsche Affinerie AG, Hamburg

Boliden Mineral AB, Roennskaer Smelter

Britannia Refined Metals Ltd., Northfleet

Portovesme srl Portovesme srl

Germany Germany Germany

Sweden

U.K.

Italy Italy

QSL / Py ISF Py

Kaldo/Py

Py

ISF/Py Kivcet/Py

100 45 15

75

230

35 100

S & B = Sinter Machine / Blast Furnace; ISF = Imperial Smelting Py = Pyrometallurgical Refining; El = Betts Electrowinning

The surveyed lead smelters comprise a total production of more than 1.5 million tonnes of lead in 1999. Operating data from the Herculaneum smelter of the Doe Run Company, U.S.A. were also integrated into the survey. These data were extracted from the preceding survey as well as from the publication about the smelter elsewhere in this volume. This should be taken into account when analyzing the information. The technical discussion and drawn conclusions are based primarily on the results of the 2000 questionnaire. Altogether, this survey embraces approximately 1.75 million tonnes of produced lead, representing 86% of the total lead production of the primary smelters in the Western World in 1999.

58 LEAD-ZINC 2000

DISCUSSION

Economics of Lead

Notwithstanding the significantly changing patterns in the individual segments of the market, the consumption of lead has grown steadily in most countries, regions and overall during the past three decades. Lead usage as a gasoline additive as well as in paint, seals and solder has declined and virtually disappeared, whereas products like radiation shielding, sheet for roofing, compounds in the glass and plastics industries and insoluble anodes for metal electrowinning have survived and grown. The principal consumption for lead, however, with the strongest and still fastest growing share, is for lead-acid batteries. They are used as automotive batteries in vehicles, but also as industrial batteries in emergency systems, in computers, in fork-lift trucks and to a growing extent for telecommunication systems, uninterruptive power sources, remote access power systems (RAPS), and hybrid and electric vehicles (2). The end use pattern for lead over the last five years is illustrated in Figure 1.

Figure 1 - End Uses for Lead over the Last Five Years (3)

Close to 85% of all products using lead metal are recyclable. One direct consequence is the continuously rising recycling rate of lead-bearing scrap. In 1999, refined lead recovered from secondary materials totaled 2.9 million tonnes, equivalent to 47.2% of total production worldwide or 58.9% in the Western World (4). Total refined lead metal supply rose in 1999 to 6.143 million tonnes. Of this, 4.93 million tonnes were produced in the Western World. Considering that the bulk of the lead from secondary materials is still recovered in the Western World, approximately 2.2 million tonnes of lead from primary sources was produced in the Western World. As shown in Table II, significant increases in production were noted at Pasminco's Port Pirie operation, in Kazakhstan as a result of increased output of the Ust-Kamenogorsk smelter and from China.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Table II - Lead Annual Data; Source ILZG in Thousands of Tonnes (4)

59

Production/Consumption

Refined Production

Europe Canada Mexico Peru United States China Japan Kazakhstan Korea Australia Other Countries

Western World World Total

1996

1,796 311 232

94 1,372

706 287 69

141 228 582

4,771 5,818

1997

1,891 271 259

98 1,417

708 297

82 182 229 571

4,944 6,005

1998

1,846 265 263 104

1,420 757 302 92

180 200 564

4,893 5,993

1999

1,846 265 210 112

1,410 808 296 155 188 268 585

4,925 6,143

Change

0 0

-53 8

-10 51 -6 63

8 68 21

32 150

1998-99

0.0 0.0

-20.2 7.7

-0.7 6.7

-2.0 68.5 4.4

34.0 3.7

0.7 2.5

Refined Consumption

Europe United States China Japan Korea Other Countries

Western World World Total

1,942 1,648

470 330 290

1,307

5,211 5,987

1,968 1,650

485 330 292

1,291

5,234 6,016

1,914 1,726

505 308 236

1,274

5,218 5,963

1,970 1,735

525 293 262

1,355

5,371 6,140

56 9

20 -15 26 81

153 177

2.9 0.5 4.0

-4.9 11.0 6.4

2.9 3.0

Lead Smelting in China

During the last ten years, China has significantly increased its lead output from 300,000 tonnes in 1990 to 808,000 tonnes in 1999; it has become one of the major producers and suppliers of lead. Comparatively to most developing countries, lead is recovered predominantly from primary sources, which account for more than 85% of the total Chinese lead production. The main operating smelters in China are the state-owned Zhuzhou Smelter, Shenyang Smelter, Shaoguan Smelter and the Shuikoushan Mine Bureau as well as the local enterprises Huize Lead and Zinc Mine Bureau and Jiyuan Smelter (5). The technology adopted by most of these smelters is the conventional sinter machine - blast furnace operation. This includes the ISF operation in Shaoguan with a capacity of approximately 70,000 tonnes annually. In addition, there are many small-sized plants existing, which are mostly township enterprises and are spread all over the country. These plants, however, were mainly responsible for the production increase of 50,000 tonnes in 1999 (4). Many of these small-size plants employ electric furnaces for treating the lead-bearing materials from primary sources or apply reverberatory furnaces for battery and lead based scrap. Examples of medium-sized plants treating exclusively secondary materials are Xingxing Smelter in Guangzhou, Xuzhou Smelter and Changchun Smelter (6).

60 LEAD-ZINC 2000

Similar to the rest of the world, the traditional primary smelters also tend to treat an increasing amount of lead scrap in their primary lead smelting operations.

Despite considerable progress in terms of efficient lead production and modernization of the existing plants, the smelters in general, but particularly the medium- and small-sized ones, suffer lower productivity and higher energy consumption compared to plants in the Western World. Moreover, they are facing environmental difficulties because of the lack of adequate facilities (5). It seems inevitable that many of these smelters will have to be modernized or shutdown (small-sized ones) in the near future. On the other hand, the secondary lead industry needs to be considerably developed by building modern and pollution-free plants.

Lead Smelting Technologies

For more than a decade, the lead industry has encountered more and more environmental pressure in almost every country, because of the introduction of stringent environmental legislation as well as health and safety regulations. This, but also rising energy and manpower costs, resulted in the requirement for additional capital, higher energy and waste disposal costs. The consequence was the closure of some older facilities and their partial or complete replacement by novel technologies such as QSL, Kivcet, Kaldo or Ausmelt/Isamelt. Most of these novel processes have in common that the metallurgical reactions are carried out in one furnace, or in two closely-coupled sealed units, with a very high level of automation and significantly reduced number of operators directly exposed to the operation. They all apply oxygen or highly oxygen enriched air to minimize the process gas volume and energy requirements. The new processes are designed to reduce emissions of harmful exhaust gases in order to protect the environment and to meet the environmental regulations in force, and to reduce the amount of hazardous dust and gases released to the ambient work place atmosphere.

The distribution of the different technologies of the 41 contacted primary smelters, and from the operations that responded to or are considered in the survey, is as follow:

Table III - Technology Distribution of Primary Smelters

Technology

Smelting Sinter Machine/Blast Furnace Sinter Machine/Imperial Smelting Kicvet QSL Ausmelt/Isamelt Kaldo Refining Pyrometallurgical Refinery Betts Refinery

Contacted

22 12 3 2 2 1

32 9

Responded

9 5 2 2 0 1

12 6

Responded/ Contacted

41% 42% 67%

100% 0%

100%

38% 67%

The majority of the primary smelters responding tend to treat secondary materials. In conjunction with environmental concerns and the decreasing net lead mine output during the last ten years, which resulted in difficulties in sourcing adequate supplies of concentrates and led to temporarily high treatment charges, the use of secondary materials was intensively

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 61

pursued. This is much more evident in industrialized countries where sufficient scrap material is available and where efficient collecting systems for lead scrap are in place. In general, almost all new technologies permit the treatment of a wide range of raw materials, including secondary feeds.

Sinter Machine / Blast Furnace

The roast-reduction process using Dwight-Lloyd sinter machine and blast furnace operation is still the basis for the majority of the lead produced in primary smelters. Since the last survey (1), however, the total number of smelters based on this conventional technology has been remarkably reduced.

Over the years, the ratio of primary to secondary raw material has changed towards the usage of more secondary materials. The amount of treated secondaries varies and mainly depends on the ownership of the smelter and its location. Two of the responding primary operations are still exclusively concentrate smelters, whereas two Japanese smelters basically converted into secondary plants by treating only secondary materials. The remainder treats secondary materials in the range between 10% and 25% of the total amount of raw material processed.

The sinter composition does not differ significantly in all the plants, averaging 41% to 48% lead, with the exception of one smelter which operates with sinter containing 28% to 36% lead. The sulfur content of the generated sinter is normally less than 2%, and the iron varies between 10% and 18% depending on the nature of the treated raw materials and/or the amount of slag, which may have been recycled. The reported composition indicates the generation of a fayalitic slag with varying basicity (Si02/CaO ratio) and lead contents of less than 3%.

Of the primary operations which responded to the questionnaire, five are employing oxygen enrichment of the blast air. Nevertheless, none of the smelters operate with high oxygen enrichment and the degree of enrichment varies between 1.7% and 6.5%.

With the exception of one operation, all carry out some form of sulfur capture. The majority (six) produces sulfuric acid. One smelter did not specifically define the desulfurization plant installed, and another is scrubbing the off-gas. The average conversion efficiency of the sulfuric acid plants is claimed to be between 94% and 98.5%. Overall recovery figures of sulfur were not provided.

Sinter Machine / Imperial Smelting Furnace

The Imperial Smelting Process (ISP) is a modified variant of conventional blast furnace technology, which is particularly appropriate for treating lead-zinc bulk concentrates. Today, thirteen furnaces at twelve smelters are in operation with a total annual capacity of approximately 470,000 tonnes of lead. The operating Imperial Smelting Furnaces are presented in Table IV. Because the ISP adopts the same principal equipment as the conventional sinter machine/blast furnace process, it faces similar issues of environmental pressure and the necessity to employ cost intensive metallurgical coke. On the other hand, the ISP is predominantly a zinc producer and, therefore, the revenue from the higher priced zinc might offset to a greater extent the production cost.

62 LEAD-ZINC 2000

Table IV - Imperial Smelting Furnaces in Operation (7)

Location Country Shaft Area Year Started Capacity Capacity (m2) Zinc Lead

Avonmouth Chanderiya Cockle Creek Copsa Mica Duisburg Hachinohe Harima Miasteczko Noyelles Godault Portovesme ShaoguanNo 1 Shaoguan No 2 Titov Veles

UK India Australia Romania Germany Japan Japan Poland France Italy China China Macedonia

27.2 21.5 24.2 17.2 19.3 22.1 19.4 19.0 24.6 19.0 17.2 17.2 17.2

1967 1991 1961 1966 1966 1969 1966 1979 1962 1972 1975 1996 1973

109,300 70,000 97,300 29,200 97,400

107,500 82,500 70,000

108,400 76,700 70,600 70,000 65,000

52,000 35,000 40,700 15,260 45,100 49,600 32,700 30,000 42,500 33,800 34,600 30,000 30,500

The sinter composition exiting the sinter machine is similar at all ISP-operations responding to the questionnaire, but differs from the sinter composition of sinter machines preceding the blast furnaces. Depending on the type of raw material used, the reported lead and iron contents vary in the order of 16.1% to 21% and 8.1% to 11%, respectively. The sulfur content in the sinter is generally less than 1%. Four smelters are treating secondary materials. The total amount of charged secondary material varies between 25.5% and 40%. One of the responding smelters uses concentrates as the only raw material source.

None of the furnaces operated with oxygen enriched blast air. Because of the stronger adjusted reduction potential, the lead concentration in the final slag, which is equal to or less than 1.5%, is lower compared to that of blast furnace slags.

All of the responding ISP operations capture the sulfur in the form of sulfuric acid. The process gases coming from the sinter machine have a relatively low concentration of SO2, normally in the range between 5% to 7%. The conversion rate of SO2 to sulfuric acid is reported to be between 94% and 99.7%.

Kivcet

The technology of the Kivcet process is based on the flash smelting principle. Concentrate burners serve to mix the feed with oxygen in order to initiate the roasting and smelting reactions as the charge travels down the reaction shaft. The desulfurized and molten material collects at the bottom of the shaft and passes through a floating coke checker where the bulk of the reduction takes place. The recovered lead bullion and the generated slag flow co-currently underneath a water-cooled partition wall into an electric furnace compartment. The electric furnace acts as a settler and performs additional reduction. Currently three smelters are operating based on Kivcet technology, from which two participated in the survey.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 63

The Kivcet process is able to treat a wide variety of raw materials. The reported ratio of primary to secondary materials was 73 : 27 in one plant (Portovesme) and reversed in the second (Cominco). At the Cominco plant, the bulk of the secondary material consists of zinc leach residues. The flash smelting process, however, requires drying the feed material, prior to its being charged into the furnace, to a moisture content of less than 1%.

Any zinc contained in the raw materials reports predominantly to the slag and enters the electric furnace compartment where it is partly volatilized under the reducing atmosphere and is recovered as fume. This lowers the cost of the slag fuming operation, which is performed subsequently in a separate furnace. The produced slag from the responding smelters contains less than 5% lead. In addition, matte is generated in both furnaces and is discharged.

Compared to the conventional two-step process, the communicated off-gas volumes of 12,000 Nm3/h and 21,000 Nm3/h are much smaller and contain a much higher SO2 concentration of 12% to 25%. This can be attributed to the sealed design of the furnace and the application of nearly pure oxygen. By cooling the off-gases in a waste heat boiler system, energy is recovered. In the case of the Portovesme plant, the thermal energy is converted in a turbine into electricity, which is utilized in the process minimizing third energy requirements. The collected dust is directly recirculated back to the process. The high SO2 concentration promotes sulfuric acid production at reduced cost. Furthermore, the small volume of exhaust gas minimizes the emission of dust and fumes.

QSL

The QSL technology is a continuous bath smelting process. Smelting occurs in a single unit consisting of a kiln-like converter, which is divided by a partition into a smelting zone and a slag reduction zone. Raw materials are charged into the smelting zone of the converter in a moist and agglomerated form, while oxygen is injected through submerged tuyeres at the bottom of the converter. The roast-reaction smelting takes place in the liquid bath, thereby converting some of the lead compounds directly into lead bullion and forming a lead oxide slag. The slag passes into the slag reduction zone where the lead oxide is gradually reduced to metallic lead, using pulverized coal as the reducing agent, as the slag flows to the opposite end of the converter. The coal is also injected into the melt through bottom blowing tuyeres. The lead and the slag flow countercurrently.

Both of the operating QSL plants participated in the survey. The current ratio of primary to secondary material at the smelter in Onsan is 64.5 : 35.5 and in Stolberg it is 80 : 20. Both operations, however, reported a high flexibility of the process in terms of feed composition. The QSL process can deal with a wide variety of feeds, and much higher rates of secondary materials have been charged in the past. Lead scrap materials can be directly charged into the converter.

The QSL process uses nearly pure oxygen, and this contributes, as for the Kivcet technology, to a relatively small volume of highly SCvrich off-gas. The reported off-gas volume for the plant in Stolberg is 15,000 Nm3/h with up to 20% SO2 depending on the type of concentrate. For the Korea Zinc plant, the volume is 18,000 Nm3/h with a SO2 concentration of 24%. The converter at Korea Zinc has a completely closed partition wall separating the off-gases from both zones. Depending on the adjusted gas atmosphere in the slag reduction zone, the resulting process gas contains zinc fumes, which exit the converter through an installed second uptake and are collected for subsequent zinc recovery. This lowers the cost of the slag fuming operation which is carried out in a separate furnace. The produced slags contain 5% lead in the Korea Zinc plant and only 2% to 3% lead at Stolberg. Heat is recovered from the

64 LEAD-ZINC 2000

off-gases in both plants and is utilized to minimize third energy requirements. After passing through several gas-cleaning stages, the SCVrich off-gas enters a sulfuric acid plant to produce sulfuric acid at conversion rates of 99.5%. The collected dust is immediately recirculated back to the process.

Kaldo

Boliden Metall AB developed the Kaldo process, and the operation in Rönnskär, Sweden is the sole application in the lead industry. The process operates on a discontinuous basis and consists of a top blowing rotary converter. It has special features compared to the other processes. It seems to be unsuitable for large-scale operation, but is a very flexible unit and can treat a wide range of secondary materials including battery scrap, residues and recycled dust along with lead concentrates. Currently it uses only concentrates as the raw material. The bone dry lead concentrate is blown through a lance into the furnace, where it reacts in the flame with the oxygen-enriched air that is introduced. The generated slag contains high concentrations of lead oxide, which is reduced with coke to an average lead content of approximately 4%. The converter is completely encapsulated in a vented enclosure, generating off-gas volumes of 25,000 Nm3/h with a SCvconcentration between 4% and 5%. The off-gases are combined with process gas from the adjacent copper smelter to produce sulfuric acid.

Refining

The foundation of all industrially applied primary lead production processes is a pyrometallurgical treatment. Lead bullion is invariably produced in equilibrium with a silicate slag and contains a range of impurities derived from the ore, fluxes, reagents and refractories. Many of these impurities like silver, gold, copper, antimony or bismuth have value and warrant recovery. Also, they need to be eliminated from the lead. To remove these impurities either pyrometallurgical or electrolytic refining techniques are employed. Worldwide, about 90% of the installed lead refining capacity employs pyrometallurgical processing routes. This is not quite reflected in the survey where only eleven of the participating operations run pyrometallurgical refineries and six employ the Betts electrorefining process.

Regardless of the choice of the procedure for refining the lead bullion, the first stage is drossing. This operation is normally performed as the last step of the smelting process and is integrated in the smelter operation and not in the refinery. Here the bullion is simply cooled in order to precipitate mainly copper and nickel.

Pyrometallurgical Refining

In pyrometallurgical refining, the impurity elements are removed one or more at a time in a sequence of steps. In general, all respondents carried out the refining batch-wise in kettles, with the exception of two operations which have continuous furnaces installed for copper removal. The sequence of operations from all the smelters which responded is very similar, as is suggested by the data in Appendices 5-7. After the removal of the residual copper with sulfur or pyrite/sulfur, antimony is removed, predominantly by employing oxygen. Alternatively, oxygen/air or caustic soda treatment is used in some plants. Only one smelter applies a separate treatment with oxygen for removing tin. Hindustan Zinc is the only smelter which performs the softening step after zinc removal, whereas arsenic is removed with caustic prior to desilverizing. The softening is followed by the Parkes process, carried out in one or more stages for adequate silver removal, and a subsequent vacuum dezincing operation. The obtained silver-rich crust passes through a liquation step or is treated further to Dore-silver; i.e., it retorted in Faber- du Faur or vacuum furnaces, and is oxidized in cupellation furnaces, such as a TBRC or

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 65

BBOC. Seven of the reporting operations carry out a bismuth removal step with part or the entire amount of lead bullion. Before casting the lead into ingots or blocks, all traces of any remaining impurity elements are removed by a final refining with caustic and niter.

Betts Electrowinning

The electrolytic method is based on the Betts electrorefining process that eliminates most of the impurities in one operation; it thereby minimizes lead and gaseous emissions. Electrolytic refining is normally confined to special cases or specific areas. Besides all the major Japanese smelters, Cominco and Korea Zinc, which responded to the questionnaire, only La Oroya and some smelters in China are currently using electrolytic refining. The questionnaires from the responding operations are given in Appendices 8 and 9. The reason for choosing either process depends on several factors, including the degree of bismuth removal required, the efficiency and cost of by-product removal, relative power cost, etc. All participating refineries apply the same principle. The lead bullion is cast into anodes, which are submerged in electrolytic cells containing hydrofluosilicic acid and thin lead starter sheets as cathodes. The obtained qualities of the refined lead are at least 99.9%. Most plants, however, produce 99.999% Pb, or almost this purity. Most of the impurities report to the slime on the spent anode. Precious metals, copper, antimony, bismuth, etc. are recovered in separate refining steps. The cost of their recovery may be reasonably inexpensive because of the comparatively small amount of material being processed.

CONCLUSIONS

In his review " Lead Smelting and Refining - Its Current Status and Future" at the third decennial Lead-Zinc symposium in 1990, Moriya (8) concluded that the decade of the nineties would be a very exciting period during which a revolution would be made in lead smelting and the traditional smelting processes would yield to new pyrometallurgical technologies. The lead industry is indeed at a crossroad. Environmental pressures have influenced many of the most significant and recent developments in the lead industry, either directly or indirectly. It seems inevitable that environmental concerns will remain at the top of the agenda. The traditional sinter machine-blast furnace operation is being challenged by direct smelting processes, which after overcoming teething problems at the beginning, are now more than major competitors. They are economically and environmentally viable technologies and offer several advantages. The ongoing introduction of more stringent pollution control regulations in many countries will most likely be the prime motivator for additional modernization efforts in the future.

ACKNOWLEDGEMENTS

The author thanks foremost all those lead smelter companies who were prepared to participate in the survey and the personnel who completed the questionnaire.

66 LEAD-ZINC 2000

REFERENCES

1. J. C. Taylor and H.R. Traulsen, Eds., World Survey of Nonferrous Smelters. The Metallurgical Society of AIME, Warrendale, PA, U.S.A., 1987,169-186.

2. R.D. Prengaman, "Lead Product Development in the Next Millennium", Lead-Zinc 2000. J.E. Dutrizac, AJ. Gonzalez, D.M. Henke, S.E. James and A. Siegmund, Eds., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 2000.

3. ILZG, "Economics of Lead and Zinc", International Lead Zinc Study Group - Statistics. http://www.ilzsg.org/statisticsasp?pg=eco. 2000.

4. ILZG, "Zinc and Lead Review of Trends and Forecasts", Metall. Vol. 5, 2000, 251-255.

5. Z. Liu, "Development and Structure of Chinese Lead and Zinc Industry", The Future of Lead and Zinc - Asia and the World. The Nonferrous Metals Society of China, Beijing, People's Republic of China, 1996, 8-20.

6. J. Wang, "Lead Smelting in China", The Future of Lead and Zinc - Asia and the World. The Nonferrous Metals Society of China, Beijing, People's Republic of China, 1996, 56-61.

7. R. W. Lee, "Production of Lead in the Imperial Smelting Process", Lead into the Future. The Institution of Mining and Metallurgy, Buxton, United Kingdom, 1996, 75-89.

8. K. Moriya, "Lead Smelting and Refining - Its Current Status and Future", Lead-Zinc '90. T.S. Mackey and R.D. Prengaman, Eds., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1990, 23-37.

9. N.D. Schupp, "Operations at the Doe Run Company's Herculaneum Primary Lead Smelter, Lead-Zinc 2000. J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A. Siegmund, Eds., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 2000.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 67

Appendix 1 -Primary Lead Smelter Survey - Smelters

οο

Co

mp

an

y N

ame

An

nu

al P

rod

uct

ion

-

Lea

d

Cu-

Dro

ss

Sb-

Sla

g B

i-cru

st

Dor

e O

ther

Typ

e B

ism

uth

met

al

Silv

er

Cop

per

Sul

phat

e S

odiu

m A

ntim

onat

e C

oppe

r A

rsen

ate

Gol

d

Ap

plie

d T

ech

no

log

ies

Ben

efic

atio

n /

Sep

arat

ion

Sm

eltin

g

Ref

iner

y

Raw

Mat

eria

ls

Con

cent

rate

s %

Pb

%Z

n %

Fe

%S

%

sio

2

%H

20

Sec

onda

ries

Type

of S

econ

darie

s

%P

b %

Zn

%F

e %

S

% S

i02

%H

20

MTP

Y

'ΜΤΡ

Ϋ M

TPY

M

TPY

M

TPY

M

TPY

M

TPY

Μ

ΤΡΫ

M

TPY

M

TPY

M

TPY

O

zPY

MTP

Y

%

%

%

%

%

%

%

%

Mou

nt Is

a M

ines

Lim

ited

156,

000

7,50

0

- - -S

ilver

in B

ullio

n 37

4 m

tpy

Lead

/Zin

c co

ncen

trato

r on

site

. C

rush

ing

follo

wed

by

flota

tion.

S

inte

r M

achi

ne /

Bla

st F

urna

ce &

D

ress

ing

(Cu,

As

and

Sb)

Ref

inin

g pe

rform

ed a

t B

ritan

nia

Ref

ined

Met

als

(BR

M)

Eng

land

.

315,

000

100

51 0

7

7 11

6

23

1 2

1 14

% o

ff dr

um fi

lter

0 - - - - - -

Pas

min

co P

ort P

irie

Sm

elte

r

215,

000

11,0

00

L 18

0

Upd

raug

ht S

inte

r M

achi

ne /

Bla

st F

urna

ce

Pyr

omet

allu

rgic

al R

efin

ing

335,

000

90

65

5 6 18

3 6 10

Lead

Bul

lion

97 . - - - -

Hac

hino

he S

mel

ting

Com

pany

44,2

00

5,30

0

Upd

raug

ht S

inte

r M

achi

ne /

Impe

rial S

mel

ting

Furn

ace

294,

000

66

Ϊ8Ϊ5

38

3

7.8

27.7

2.

6

34

11.1

38

.6

10.4

3.6

Sum

itom

o M

etal

Min

ing

Co.

H

arim

a W

orks

22,0

54

2,14

2

Upd

raug

ht S

inte

r M

achi

ne /

Impe

rial S

mel

ting

Furn

ace

Bet

ts E

lect

rore

finin

g P

roce

ss

179,

539

14.3

41

.8

8.2

27.0

2.

7

7.0

Wae

lz K

iln O

xide

8.2

60.3

4.

Ϊ 6.

3 1.

7 8.

5

Hos

okur

a S

mel

ting

& R

efin

ing

Co.

20,6

76

Yes

Y

es

Yes

Y

es

Upd

raug

ht S

inte

r M

achi

ne /

Bla

st F

urna

ce

Bet

ts R

efin

ery

27,6

48

0 100

Spe

nt L

ead-

Aci

d B

atte

ries

55

-60

5-6

LEAD-ZINC 2000

Co

mp

any

Nam

e

Fur

nace

Flu

x S

i02

Lim

e / L

imes

tone

Iro

n C

oal

Cok

e O

ther

Typ

e

Fee

d P

rep

arat

ion

Ble

ndin

g S

yste

m

Dry

er, I

f ap

plic

able

Ty

pe

Num

ber

Dim

ensi

on

Feed

Rat

e In

let-

H20

Ö

ütje

t-H

2Ö

Fuel

- Ty

pe

Cal

orifi

c V

alue

G

as T

empe

ratu

re

Dry

er In

let

Dry

er O

utle

t G

as V

olum

e D

isch

arge

Tem

pera

ture

Sin

ter

Mac

hine

N

umbe

r D

imen

sion

s, H

earth

Are

a N

omin

al C

apac

ity

Sul

fur

Bur

ning

Rat

e

Fuel

- Ty

pe

Cal

. Val

ue

Feed

Moi

stur

e

Sin

ter C

ompo

sitio

n %

Pb

%F

e %

S

kg /

MT

- bu

llion

kg /

ΝΪΤ

- bu

llion

kg /

MT

- bu

llion

kg /

MT

- bu

llion

kg/

MT

- bu

llion

kg/M

T -

bulli

on

m

MTP

H

%

%

kJ/k

g

"" "

r c *c

N

m3 /h

r ÖC

m5

MTP

H

MT

Sui

fur/m

5 * d

kJ/k

g %

H20

%

%

%

Mou

nt Is

a M

ines

Lim

ited

155

495 - - 181 -

Pad

dle

Mix

er A

nd P

elle

tiser

Not

App

licab

le

- - - - - - . . - - - - - 1 93

Tota

l Fee

d 24

0 2.

59

Die

sel o

il is

use

d to

igni

te th

e su

lfur

in th

e ig

nitio

n la

yer.

18,4

00

4-6

44.1

10

.2

1.7

Pas

min

co P

ort P

irie

Sm

elte

r

100

65

135 - 50 -

Mix

ing

Pla

nt -

Con

ditio

ning

Dru

m

Not

App

licab

le

- - - _ - - - - - - - - - Ϊ 83

.6

85

3.0

Nat

ural

Gas

38,0

00

11.5

48

11.5

1.

7

Hac

hino

he S

mel

ting

Com

pany

Not

App

licab

le

- - - - - - - - - - - - - 1 90

34

Ϊ.78

Hea

vy F

uel O

il

42,0

00

6

19.6

9.

0 0.

6

Sum

itom

o M

etal

Min

ing

Co.

H

arim

a W

orks

200

Mix

ing

Dru

m &

Pel

letiz

er D

isc

Not

App

licab

le

- - - - - - - - - - - - - Ϊ 70

25

Ϊ.63

Hea

vy F

uel O

il

90

6

18.6

8.

1 0.

6

Hos

okur

a S

mel

ting

& R

efin

ing

Co.

- . 107 -

236 -

Not

App

licab

le

- - - - - - - - - - - - - 1 29

27

Hea

vy F

uel O

il

5.5

41.7

3.3

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

ο

Co

mp

any

Nam

e

Exh

aust

Gas

· V

olum

e %

S0

2

Sm

elti

ng

Num

ber o

f Fur

nace

s Ty

pe

Nom

inal

Cap

acity

D

imen

sion

s

Sm

eltin

g

Aux

iliary

Fue

l - T

ype

Am

ount

C

alor

ific

Val

ue

Air

Vol

ume

Oxy

gen

Enr

ichm

ent

Oxy

gen

Vol

ume

Ope

ratin

g Te

mpe

ratu

re

Flux

as

% o

f Fee

d R

efra

ctor

y Li

ning

- Ty

pe

Cam

paig

n of

Life

Furn

ace

Offg

as

Vol

ume

Tem

pera

ture

ex

Furn

ace

%S

02

Gas

Coo

ling

by

Num

ber o

f uni

ts

Out

let T

empe

ratu

re

Dus

t Col

lect

ion

by

Dis

posi

tion

of D

ust

Am

ount

Nm

3 /hr

%

MTP

H

Kg (N

ms j/h

r kJ

/kg

Nrr

Shr

%

Nm

3 /hr

°C

%

year

s

Nm

3 /hr

°C

%

°C

TPH

Mou

nt Is

a M

ines

Lim

ited

80,0

00

2.8 1

Lead

bla

st fu

rnac

e w

ith a

sin

gle

row

tuye

res

(40

tuye

res

in to

tal)

27 m

tph

lead

7.

02m

X 1

98m

Cok

e 4,

600

26,0

00

25,2

00

1.7

450

1,22

0-1,

250

23.5

C

hom

e/M

agne

sium

bric

ks a

re

used

in th

e he

arth

, the

sid

es o

f th

e fu

rnac

e ar

e w

ater

coo

led

jack

ets.

7

92,5

20

130

0.48

In

gres

s ai

r an

d sp

ray

cham

ber

1 sp

ray

cham

ber t

hat r

ecei

ves

sint

er p

lant

and

furn

ace

off

gase

s.

120

Bag

hous

e

Slu

rry d

ust a

nd re

cycl

e to

the

sint

er p

lant

. 2.

6 (S

inte

r M

achi

ne)

+ 0.

4 <B

F)

Pas

min

co P

ort P

irie

Sm

elte

r

108,

000

0.4 1

Ope

n To

p B

last

Lea

d Fu

rnac

e

27

18.8

m2 H

earth

Are

a

Met

allu

rgic

al C

oke

5,70

0 28

,000

26,0

00

6.5

1,80

0

1,15

0-1,

250

2

134.

000

325

0.2

Dire

kt A

ir C

oolin

g

Bag

hous

e

Slu

rry -

recy

cled

0.5

Hac

hino

he S

mel

ting

Com

pany

62,4

00

7 1 Im

peria

l Sm

eltin

g Fu

rnac

e

14.8

27

.3 m

2 Sha

ft A

rea

Met

allu

rgic

al C

oke

50,0

00

no -

68,0

00

450

Gas

Coo

ling

Tow

er

1 50

Thei

sen

Dis

inte

grat

or

2.0

Sum

itom

o M

etal

Min

ing

Co.

H

arim

a W

orks

ii.b'

u'u

6.3 1

Impe

rial S

mel

ting

Furn

ace

4 19

.4 m

2 Sha

ft A

rea

Met

allu

rgic

al C

oke

33,0

00

no -

950

Chr

ome

Mag

nesi

a B

rick

2.5-

3.Ö

45,0

00

1,00

0

470

2.5

Hos

okur

a S

mel

ting

& R

efin

ing

C

o.

8,04

0 2.

5 1 B

last

Fur

nace

23Ö

Ö M

TPM

Cok

e

4,50

0 no

-

Bag

hous

e

0.31

LEAD-ZINC 2000

Co

mp

any

Nam

e

Red

ucin

g A

gent

- Ty

pe

Am

ount

C

alor

ific

Val

ue

Air

Vol

ume

Oxy

gen

Enr

ichm

ent

Oxy

gen

Vol

ume

Ope

ratin

g Te

mpe

ratu

re

Ref

ract

ory

Lini

ng -

Type

C

ampa

ign

of L

ife

Furn

ace

Offg

as

Vol

ume

Tem

pera

ture

ex

Furn

ace

%S

0 2

Gas

Coo

ling

by

Num

ber o

f uni

ts

Out

let T

empe

ratu

re

Dus

t Col

lect

ion

by

Dis

posi

tion

of D

ust

Am

ount

Furn

ace

Pro

duct

s (a

vera

ge)

Lead

Bul

lion

%P

b %

S

%C

u %

Sb

%A

g Te

mpe

ratu

re

Mat

te

%P

b %

As

%C

u %

Fe

%S

Te

mpe

ratu

re

Sla

g %P

b %

FeO

%

Si0

2

% C

aO

%Z

n Te

mpe

ratu

re

—

Kg (

Nm

ä J/hr

kJ

/kg

NiÄ

'r

%

Nm

3 /hr

6C

year

s

Nm

3 /hr

°C

%

°C

TPH

MTP

D

%

%

%

%

%

5 "c

MTP

D

%

%

%

%

%

' ä"c

MTP

D

%

%

%

%

%

... . „

_

Mou

nt Is

a M

ines

Lim

ited

_!a

,§M

£,M

i;n

ifi„

.,.„

M

etal

lurg

ical

Cok

e

510

98.6

5 <

0.05

0.

002

0.11

0.

24

1,07

0

55.5

0.

9 26

.7

1 10

.7

520

2.6

22 8

21

24

5

139

1,22

5

Pas

min

co P

ort P

irie

Sm

elte

r

iij.ffl

asl&

mas

i: M

etal

lurg

ical

Cok

e

650

95

0.4

2.2

0.5

0.02

13Λ5

00Μ

ΤΡΥ

38

2.

5 40

0.

5 15

210,

000

MTP

Y

2.6

32

22

14

16

1,20

0

Hac

hino

he S

mel

ting

Com

pany

Met

allu

rgic

al C

oke

170

93.9

3.6

1,00

0

275

1.0

36.0

12

.3

17.5

7.

0 1,

300

Sum

itom

o M

etal

Min

ing

Co.

H

arim

a W

orks

Met

allu

rgic

al C

oke

75

98.5

0.

1 0.

6 0.

2 0.

2 1,

100

150

1.5

38

6 19

.1

14.2

8.

0 1,

250

Hos

okur

a S

mel

ting

& R

efin

ing

C

o.

JoJM

Eum

ise

Met

allu

rgic

al C

oke

59.8

97

.1

0.13

1.

13

0.12

38.6

0.

8 28

7

24

21 2

2

6

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Co

mp

an

y N

am

e

Sla

g C

lea

nin

g /

Fu

min

g

Nu

mb

er

of

Fu

rna

ces

Typ

?.

Sla

g T

rea

ted

Dim

en

sio

ns

Re

du

cta

nt

Typ

e

Fin

al

Sla

g %

Pb

%C

u %

2n

Fu

me

Pro

du

ctio

n %

Pb

%Z

n

Su

lfu

r F

ixa

tio

n

So

urc

e o

f G

as

Pla

nt-

Ty

pe

Ra

ted

Ca

pa

city

G

as

Vo

lum

e

Ave

rag

e S

02

in i

nle

t g

as

Ave

rag

e C

on

vers

ion

Eff

icie

n

Pro

du

ct G

rad

e

En

erg

y R

eq

uir

em

en

t A

uxi

liary

Fu

el

Ta

ikia

s S

cru

bb

ing

MT

PH

h M

TP

H

%"

%

%

MT

PH

%

%

MT

PD

Nm

3/h

r

%

%

% H

2S

04

KW

h/M

T "

Sul

fur

Mo

un

t Is

a M

ine

s L

imit

ed

No

t A

pp

lica

ble

- - - - . . - - - - - -

No

t A

pp

lica

ble

- - . - - - - . - -

Pa

sm

inc

o P

ort

Pir

ie S

me

lte

r

2 B

atc

h

25

0,0

00

MT

PY

Ste

am

ing

jCo

al

20

5,0

00

MT

PY

0

.05

0.3

2.5

JäO

JJO

O M

TP

Y

12

66

Sin

ter

Pla

nt

Su

lfuri

c A

cid

Lu

rgi S

ing

le C

on

tact

23

0 4

8,2

40

5.0

98

98

.5

Na

tura

l G

as

No

ne

""'

Ha

ch

ino

he

Sm

elt

ing

Co

mp

an

y

1 A

usm

elt

11

.5

2.4

m d

ia x

6 m

hig

h O

il In

ject

ion

10.7

0

.2

0.6

3.5

0.8

9.0

5

7.7

Sin

ter

Pla

nt

Su

lfuri

c A

cid

Lu

rgi

Do

ub

le C

on

tact

50

0 6

2,4

00

7.0

99

.7

77

Ä9

8

No

ne

Su

mit

om

o M

eta

l M

inin

g C

o.

Ha

rim

a W

ork

s

No

t A

pp

lica

ble

- - - - . . - - - - - -

Up

dra

ft S

inte

rin

g G

as

Su

lfu

ric

Aci

d S

ing

le C

on

tact

Typ

e

33

0 5

1,0

00

6.3

97

.5

98

33

0

Am

mo

niu

m A

bso

rpti

on

Me

tho

d

Ho

so

ku

ra S

me

ltin

g &

R

efi

nin

g C

o.

No

t A

pp

lica

ble

- - - . . . - - . - -

Sin

ter

Ma

chin

e,

Bla

st F

urn

ace

De

sulfu

riza

tion

Pla

nt

(no

H2S

Cy

Pro

d.)

65

0

LEAD-ZINC 2000

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 73

Appendix 2 -Primary Lead Smelter Survey - Smelters

Co

mp

any

Nam

e

An

nu

al P

rod

uct

ion

-

Lea

d

Cu-

Dro

ss

Sb-

Sla

g B

i-cru

st

Dor

e O

ther

Typ

e B

ism

uth

met

al

Silv

er

Cop

per

Sul

phat

e S

odiu

m A

ntim

onat

e C

oppe

r Ars

enat

e G

old

Ap

plie

d T

ech

no

log

ies

Ben

efic

atio

n / S

epar

atio

n

Sm

eltin

g

Ref

iner

y

Raw

Mat

eria

ls

Con

cent

rate

s %

Pb

%

Zn

j

%F

e %

S

%S

iÖ2

%H

2Ö

Sec

onda

ries

Type

of S

econ

darie

s

%P

b %

Zn

%F

e %

S

% S

i02

%H

2Ö

MTP

Y

MTP

Y

MTP

Y

MTP

Y

MTP

Y

MTP

Y

MTP

Y

MTP

Y

MTP

Y

MTP

Y

MTP

Y

OzP

Y

MTP

Y

%

%

%

%

%

%

%

%

Kam

ioka

Min

ing

& S

mel

ting

C

o.

34,0

00

I 3,

000 -

Yes

Y

es

. - -Y

es

Bre

akin

g an

d sp

ecifi

c gr

avity

se

para

tion

Upd

raug

ht S

inte

r M

achi

ne /

Bla

st F

urna

ce

Bet

ts R

efin

ery

55,0

00

0 100

Spe

nt L

ead-

Aci

d B

atte

ries

60

5-6

Toho

Zin

c C

o.

Chi

giris

him

a / J

apan

90,0

00

1500

0

Ant

imon

y M

etal

Y

es

165

Yes

Upd

raug

ht S

inte

r M

achi

ne /

Bla

st F

urna

ce

Bet

ts R

efin

ery

141,

000

62

55

5.5

18.0

1.

5 5.

0

Dus

t and

Res

idue

s

53.1

2.

6 0.

9 11

.8

0 5

Hin

dust

an Z

inc

Ltd.

C

hand

eriy

a / I

ndia

35,0

00

375

(Cad

miu

m)

10

74

Upd

raug

ht S

inte

r M

achi

ne /

impe

rial S

mel

ting

Furn

ace

Pyr

omet

allu

rgic

al R

efin

ing

260,

887

100

19

35

7.5

24.0

3.

3 5.

5 0 - - - - - -

Kor

ea Z

inc

Co.

, Lt

d.

Ons

an /

Sou

th K

orea

121,

040

(QS

L) +

18.

445

(SR

F)

Yes

Yes

QS

L-P

roce

ss &

Sho

rt R

otar

y Fu

rnac

e

Bet

ts R

efin

ery

252,

244

64.5

67

4.

3 2.

8 15

.0

2 6.5

35.5

Le

ad-S

ilver

Res

idue

s,

Lead

Res

idue

s, P

aste

33

6 3.6 13

56 25

Asa

rco

Inc.

E

ast H

elen

a P

lan

t/U

SA

65,0

00 -

75,0

00

Sin

ter

Mac

hine

/ B

last

Fur

nace

; D

ross

Rev

erbe

rato

ry

Ket

tle D

ecop

periz

atio

n

235,

000

50

-65

0-7

5 0

-12

0-3

0 1

3-3

0 2

-15

5-1

5

10

-25

Pre

c. M

etal

Sla

g, G

old

Bea

ring

Car

bons

, Zin

c P

lant

Res

idue

s,

0-9

0 0

-15

0-4

0 0

-30

0-9

5 0

-65

LEAD-ZINC 2000

Co

mp

any

Nam

e

Fur

nace

Flu

x S

i02

Lim

e /

Lim

esto

ne

Iron

Coa

l C

oke

Oth

er T

ype

Fee

d P

rep

arat

ion

Ble

ndin

g S

yste

m

Dry

er, i

f ap

plic

able

ly

pe

Num

ber

Dim

ensi

on

Feed

Rat

e In

let-

H20

O

utle

t-H

20

Fuel

- Ty

pe

Cal

orifi

c V

alue

G

as T

empe

ratu

re

Dry

er In

let

Dry

er O

utle

t G

as V

olum

e D

isch

arge

Tem

pera

ture

Sin

ter

Mac

hine

N

umbe

r D

imen

sion

s, H

earth

Are

a N

omin

al C

apac

ity

Sul

fur

Bur

ning

Rat

e

Fuel

- Ty

pe

Cal

. Val

ue

Feed

Moi

stur

e

Sin

ter

Com

posi

tion

%P

b %

Fe

%S

kg /

MT

- bu

llion

kg /

MT

- bu

llion

kg /

MT

- bu

llion

kg /

MT

- bu

llion

kg /

MT

- bu

llion

kg /

MT

- bu

llion

m

MTP

H

%

%

kJ/k

g

°C

°C

Nm

ä /hr

Tc n7

M

TPH

M

TS

uifu

r/m

F'd

U/k

g %

H20

%

%

%

Kam

ioka

Min

ing

& S

mel

ting

Co.

88

163

116 -

222 -

Non

e

Not

App

licab

le

- - - - - - - - - - - - -N

ot A

pplic

able

- - - - - - - - - - -

Toh

o Z

inc

Co.

C

higl

rishi

ma

/ Jap

an

55

161

15 - 18

6 -

2 P

elle

tizer

Not

App

licab

le

- - . - - - . . . - - - - 1 33

24

1.92

Hea

vy F

uel O

il

39,8

00

4.5

41

11

1.8

Hin

dust

an Z

inc

Ltd.

C

hand

eriy

a / I

ndia

340

Mix

ing

and

Con

ditio

ning

Dru

m

Not

App

licab

le

- - - - - - - - - - - - - 1 120

192

Sin

ter

Pro

duct

1.

6

Oil

4.5

- 5.

5

20

-22

10

-12

<1

Kor

ea Z

inc

Co.

, Ltd

. O

nsan

/ S

outh

Kor

ea

69

118

90

147 - -

Wei

ghfe

eder

and

Mix

er

Not

App

iicab

ie

- - . - - - - - - - - - -N

ot A

pplic

able

. - . - - -j

. - - -

Asa

rco

Inc.

E

ast H

elen

a P

lant

/US

A

255-

295

245

- 28

0 2

0-2

5

-35

0 -

375

- ·

12 F

eed

Hop

per,

Impa

ct

Ham

mer

M

ill an

d M

ixin

g D

rum

Not

App

licab

le

- - - - - - - . - - - - - 1 83

3 (U

pdra

ft A

rea)

2

5-3

0 1.

01

7.5

-9.0

28

-36

10

-18

1.3

-1.9

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Co

mp

any

Nam

e

Exh

aust

Gas

- V

olum

e %

so2

Sm

elti

ng

Num

ber o

f Fur

nace

s Ty

pe

Nom

inal

Cap

acity

D

imen

sion

s

Sm

eWna

A

uxilia

ry^F

uel -

Typ

e A

mou

nt

Cal

orifi

c V

alue

Air

Vol

ume

Oxy

gen

Enr

ichm

ent

Oxy

gen

Vol

ume

Op

en

ing

Tem

pera

ture

Fl

ux a

s %

of F

eed

Ref

ract

ory

Lini

ng -

Type

Cam

paig

n of

Life

Furn

ace

Offg

as

Vol

ume

Tem

pera

ture

ex

Furn

ace

%S

02

Gas

Coo

ling

by

Num

ber o

f uni

ts

Out

let T

empe

ratu

re

Dus

t Col

lect

ion

by

Dis

posi

tion

of D

ust

Am

ount

'rim

'/rir

%

MTP

H

Kg_

(Nm

3 )/rir

kJ/k

g

NnA

hr

%

Nm

3 /hr

~'r C

%

year

s

Nm

3 7hr

°C

%

°C

TPH

Kam

ioka

Min

ing

& S

mel

ting

C

o. - - 1

Bla

st F

urna

ce

S

5.2

m2 a

t tuy

ere

leve

l

4,80

0 no

-1,

500

26.4

1

15,0

00

500

(at t

op o

f fur

nace

) 0.

1 B

allo

n du

ct (3

) &

Dus

t cha

mbe

r (8)

3

&8

80

Bag

filte

r

Rec

lrcul

atjo

n

0.4

Toho

Zin

c C

o.

Cht

giris

him

a / J

apan

21,0

00

7 1 B

last

Fur

nace

9.5

5.5m

x 1

.7m

x 4

.2m

Hea

vy F

uel O

il ^

70

39,8

00

8,40

0 no

-1,

250

Chr

ome-

Mag

nesi

te B

rick

20

19,2

00

250

0.8

Scr

ubbe

r

1 115

Bag

filte

r

0.8

Hin

dust

an Z

inc

Ltd.

C

hand

erty

a / I

ndia

105,

000-

120,

000

3.5

- 5.

0

1 Im

peria

l Sm

eltin

g Fu

rnac

e

21.5

m2

Oil

45,0

00 -

50,0

00

no -

950-

1,00

0

2

950-

1,00

0 0.

1 50

ES

P (L

urgi

) fo

r S

inte

r Mac

hine

Th

eise

n D

isin

tegr

ator

for

ISF

R

edrc

ulat

ion

to S

inte

r Mac

hine

7

Kor

ea Z

inc

Co.

, Lt

d.

Ons

an /

Sou

th K

orea

- - 1 Q

SL-

Con

verte

r (C

ylin

dric

al S

hape

)

20 ΪΜ

ΤΡΗ

Bul

lion)

41

.13m

long

x 4

/4.5

m d

ia

Ant

hrac

ite C

oal

1,03

1 28

,000

- 100

7,00

0

1,15

0 9.

9 C

rom

e-M

agne

site

Bric

k

1

i'ä'.ö

'öö

1,25

0 24

W

aste

Hea

t Boi

ler

1

<40

0 E

lect

rost

atic

Pre

cipi

tato

r (E

SP

); Lu

rgi

Rec

ircul

atjo

n to

QS

L-C

onve

rter

7

Asa

rco

Inc.

E

ast H

elen

a P

lan

t/U

SA

50,9

70 -

59,4

15

4-6

2 B

last

Fur

nace

32

-36

7.93

m X

3.0

5m x

6.1

m

h M

etal

lurg

ical

Cok

e

16,5

00

4-6

67

9 -

849

1,20

0-1,

350

15

-17

Wat

er C

oole

d Ja

cket

s

0.25

35,0

00 -

37,0

00

310

0.01

- 0

.03

Dire

ct A

ir C

oolin

g

54

Bag

hous

e

Rec

ycle

d to

Fur

nace

0.7

- 0.

9

LEAD-ZINC 2000

Co

mp

any

Nam

e

Red

ucin

g A

gent

- Ty

pe

Am

ount

C

alor

ific

Val

ue

Air

Vol

ume

Oxy

gen

Enr

ichm

ent

Oxy

gen

Vol

ume

Ope

ratin

g Te

mpe

ratu

re

Ref

ract

ory

Lini

ng-

Type

C

ampa

ign

of L

ife

Furn

ace

Offg

as

Vol

ume

Tem

pera

ture

ex

Furn

ace

%S

0 2

Gas

Coo

ling

by

Num

ber o

f uni

ts

Out

let T

empe

ratu

re

Dus

t Col

lect

ion

by

Dis

posi

tion

of D

ust

Am

ount

Fur

nace

Pro

duct

s (a

vera

ge)

Lead

Bul

lion

%P

b %

S

%C

u %

Sb

...__%

. A&

Te

mpe

ratu

re

Mat

te

%P

b %

As

%C

u %

Fe

%S

Te

mpe

ratu

re

Sla

g %P

b %

FeO

%

S(0

2 %

CaO

"*

%Z

n Te

mpe

ratu

re

——

——

K

g (N

rt?j/h

r kj

/kg

Nm

J /hr

%

Nm

'/hr

"c

year

s

Nm

3 /hr

*"

"»c %

"C

TPH

MTP

D

%

%

%

%

%

~'c

M

TPD

%

%

%

%

%

ec

MTP

D

%

%

%

%

%

~ T

~

Kam

ioka

Min

ing

& S

mel

ting

Co.

„..„„„

„„„Jn

ßlaU

F.kim

s.ss,

,..

.„„

Met

allu

rgic

al C

oke

43,0

00 M

TPY

9

4-9

8

. 0.1

0.9

0.1

800

Non

e - - - - . -20

,400

MTP

Y

3 30

23

25 -

1.20

0

Toh

o Z

inc

Co.

C

higi

rishi

ma

/ Jap

an

JrLB

laäi

.fMffl

sg,,.

,,,,.,

..,,,,

M

etal

lurg

ical

Cok

e

228

97.8

0.03

1.

2 0.

2 1,

000

Non

e . - . - . -28

0

28

21

20

15

1,20

0

Hin

dust

an Z

inc

Ltd.

C

hand

eriy

a /

Indi

a

„Jn

JmR

SM

Sm

fiiP

" Fu

rnac

e M

etal

lurg

ical

Cok

e

100-

120

93.5

3 0.

2 -

0.4

950

250

1-2

35

20

1

4-1

6 8

-10

1,00

0

Kor

ea Z

inc

Co.,

Ltd.

O

nsan

/ S

outh

Kor

ea

Pul

veriz

ed A

nthr

acite

Coa

l 1,

700

28^0

00

2,30

0 55

.6

1,28

0

1,25

0 C

hrom

e-M

agne

site

Bric

ks

2

6,00

0 1,

250

<0.0

5

Asa

rco

Inc.

E

ast H

elen

a P

lant

/US

A

(Hm „

Hnu

J[R

nBJa

jitJju

jjia1c

e 1

Met

allu

rgic

al C

oke

Was

te H

eat B

oile

r &

Gas

Coo

ler

2 200

Bag

hous

e Zi

nc O

pera

tion

2.3

400

98

0.05

1.

5 0.

2 0.

23

1,00

0

13

25

0.3

47

0.5

20

1,00

0

260 5

26.6

Ϊ9

.3

13.2

15

.0

1,21

6

65^0

00 -

75.,0

00 M

TPY

95

.4 -

96.

9

. 0.

02

1.6

-2.1

1

.2-1

.5

9,00

0-11

,000

100,

000-

120,

000

MTP

Y

1.8

25.8

22

.3

19.6

14

.2

1,20

0-1,

300

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Co

mp

any

Nam

e

Sla

g C

lean

ing

/ Fu

min

g

Num

ber

of F

urna

ces

lYPJ

B

Sla

g Tr

eate

d D

imen

sion

s R

educ

tant

Typ

e

Fina

l Sla

g %

Pb

%C

u %

Zn

Fum

e P

rodu

ctio

n %

Pb

%Z

n

Su

lfu

r F

ixat

ion

Sou

rce

of G

as

Pla

nt -

Type

Rat

ed C

apac

ity

Gas

Vol

ume

Ave

rage

S0

2 in

inle

t gas

A

vera

ge C

onve

rsio

n E

fftci

en

Pro

duct

Gra

de

Ene

rgy

Req

uire

men

t A

uxilia

ry F

uel

faiig

as S

crub

bing

MTP

H

MTP

H

%

%

%

MTP

H

%

%

MTP

D

Nm

3 /hr

%

%

% H

2S0 4

KW

h/M

T S

ulfu

r

Kam

ioka

Min

ing

& S

mel

ting

C

o.

Not

App

licab

le

- - - - - . - - - - _

Bla

st F

urna

ce

Scr

ubbe

r

15,0

00

0.1

98

Toho

Zin

c C

o.

Chi

giris

him

a / J

apan

Not

App

licab

le

. . . - . . . - . - . .

Sin

ter

Mac

hine

Sul

furic

Aci

d M

onsa

nto

Sin

gle

Con

tact

13

0 21

,000

7 94

98.5

32

0

Hin

dust

an Z

inc

Ltd.

C

hand

eriy

a / I

ndia

Not

App

licab

le

. . . - - . . - . . - -

Sin

ter

Mac

hine

Sul

furic

Aci

d Lu

rgi S

ingl

e C

onta

ct

300

75,0

00

4-5

94

83

-94

Ligh

t Fue

l Oil

Non

e

Kor

ea Z

inc

Co.

, Lt

d.

Ons

an /

Sou

th K

orea

1 A

usm

elt F

urna

ce

13

4.6m

dia

. X 8

.8 m

hig

h Lu

mp

Coa

l

10.5

1.

5 - 8 2.2

22

49

Sm

eltin

g Zo

ne Q

SL

Con

verte

r

Sul

furic

Aci

d Lu

rgi D

oubl

e C

onta

ct

800

72.0

00

10.4

99

.5

98/9

6 25

6

yes

Asa

rco

Inc.

E

ast H

elen

a P

lan

t/U

SA

Not

App

licab

le

- . - - . - . - . . - .

, S

inte

r M

achi

ne

Sul

furic

Aci

d Lu

rgi S

ingl

e C

onta

ct

325

102,

000

2.5

-3.5

96

93.3

Non

e

LEAD-ZINC 2000

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 79

Appendix 3 -Primary Lead Smelter Survey - Smelters

Co

mp

any

Nam

e

An

nu

al P

rod

uct

ion

-

Le

ad

- ''"

MTP

Y

Cu-

Dro

ss,

MTP

Y

Sb-

Slä

fl.

Μ'Τ

ΡΫ

B

l-cru

st

MTP

Y

Do

re""

M

TPY

O

ther

Typ

e,

MTP

Y

Bis

mut

h m

etal

M

TPY

S

ilver

, "'

M

TPY

"

Cop

per

Sul

phat

e,

MTP

Y

Spd

ium

Ant

jmon

ate

MTP

Y

Cop

per

Ars

enat

e M

TPY

".

1

~1

. G

old

i "O

ZP

Y

Ap

plie

d T

ech

no

log

ies

_

Ben

efic

abon

/ S

epar

atio

n

Sm

eltin

g

Ref

iner

y

Raw

Mat

eria

ls

Con

cent

rate

s %

Pb

%Z

n %

Fe

%S

%

SiO

:

%H

20

Sec

onda

ries

Ty

pe o

f Sec

onda

ries

%P

b %

Zn

%F

e %

S

%S

i02

%H

20

1-

MT

PY

%

%

%

%

%

%

%

%

Co

mln

coL

td.

Doe

Run

Com

pany

(1,

2)

Met

-Mex

Pen

oles

, S.A

. N

ora

nd

aln

c.

| B

erze

llue

Sto

lber

a G

mbH

T

rail

Ope

ratio

ns /

Can

ada

: H

ercu

lane

um /

US

A

Tor

reon

/ M

exic

o B

runs

wic

k S

mel

ter

120,

000'

200

"37

3""

7,

600

1,20

0 90

0 50

,000

KIV

CE

T

Bet

ts R

efin

ery

300,

000

27

58.5

8.

2 7 18

1.4 7 73

6% B

atte

ry P

aste

; 94

% R

esid

ues

77.8

(B

atte

ry P

aste

)

5.8J

Bat

tery

Pas

te)

225~

0M

" '■

-

1g

o0o

o *

" 10

8,00

0 *

ΙΟΟ

,Μθ'

3,

500

_

,_

7,50

0

6,00

0 4,

800

" "

4,00

0 ~

40

0 15

,700

*

4,50

0 *

6,00

0 2,

400

] "

350-

400

\ 20

0 \.

. "

. -2

,500

*

Upd

raug

ht S

inte

r M

achi

ne/

Bla

st F

urna

ce

Pyr

omet

allu

rgic

al R

efin

ing

320,

000

100

72 . 3.

3 16

.1

0.7 8 Ö

Upd

raug

ht S

inte

r Mac

hine

/ B

last

Fur

nace

Pyr

omet

aiiu

rgic

al R

efin

ing

474,

620

77

43.5

6.

5 10

.5

19.8

8.

2 8

-10

23

Zin

c P

lant

Res

idue

s, D

ross

es,

Sm

elte

r R

esid

ues

36.5

4.

5 8.

5 9.

3 4.

4 1

-2

Upd

raug

ht S

inte

r M

achi

ne/

Bla

st F

urna

ce

Pyr

omet

allu

rgic

al R

efin

ing

275,

000

75

11.0

-73.

0 2.

0-12

.0

2.0-

25.0

10

.0-3

5.0

2.0-

12.0

3.

0-8.

0

25

Pb

Sul

phat

es,

Pb-

bear

ing

Res

idue

s,

0-63

.0

O-1

2.0

2.

(ΕΪ5

.δ

2.0-

12.0

4.

0-65

.0

4.0-

30.0

QS

L-P

roce

ss

Pyr

omet

allu

rgic

al R

efin

ing

148,

300

80

70

4-5

1

-5

15

-20

1-3

6 20

B

atte

ry P

aste

, Res

idue

s

60

-70

1 -5

0

-3

5-1

5 0

-3

10

LEAD-ZINC 2000

Co

mp

any

Nam

e

Furn

ace

Flux

S

i02

Lim

e /L

imes

tone

Iro

n C

oal

Cok

e O

ther

Typ

e

Fee

d P

rep

arat

ion

Ble

ndin

g S

yete

m

Dry

er, i

f app

licab

le

Type

N

umbe

r D

imen

sion

Feed

Rat

e In

let-

H20

O

utle

t -H

20

Fuel

- Ty

pe

Cal

orifi

c V

alue

G

as T

empe

ratu

re

Dry

er In

let

Dry

er O

utle

t G

as V

olum

e D

isch

arge

Tem

pera

ture

Sin

ter

Mac

hine

N

umbe

r D

imen

sion

s, H

earth

Are

a N

omin

al C

apac

ity

Sul

fur

Bur

ning

Rat

e

Fuel

- Ty

pe

Cat

Val

ue

Feed

Moi

stur

e

Sin

ter

Com

posi

tion

%P

b %

Fe

%S

kg /

MT

- bu

llion

kg/U

T-b

ulli

on

kg /

MT

- bu

llion

kj/M

T-b

ulli

on

Htg

/ M

T -

bullio

n ""

kg'/M

T-

bullio

n

m

MTP

H

%

%

kJ/k

g

- s c"

""~

"C

Nm

3 /hr

- '"

»c

'"

L—

m

5

MTP

H

MT

Sul

fur/m

2 · d

kJ/k

g %

H20

%

%

%

Com

inco

Ltd

. Tr

ail O

pera

tions

/ C

anad

a

400

250

450

350

100 -

Pro

porti

onin

g B

ins

Rot

ary

one

24 m

long

x 3

m d

iam

eter

64

14

<1.0

N

atur

al G

as

'

650

140

62,0

00

SO

Not

App

licab

le

- - - - - - - - . . -

Doe

Run

Com

pany

(1,2

) H

ercu

lane

um /

USA

Bel

t Mix

ing

Not

App

licab

le

- - - - - - - - - - - -

!_

1 75

Nat

ural

gas

37,6

50

48

15

<2

Met

-Mex

Pen

oies

, S.A

. To

rreo

n /

Mex

ico

- 306

39 . 23

2 -

Load

er /

Con

veyo

r B

elt

Not

App

licab

le

- - - - - - - - . - - - - 2 160

65

1.15

Nat

ural

Gas

66,4

27

4.5

-6.0

41.5

11

.3

Nor

anda

Inc.

B

runs

wic

k S

mel

ter

270

200

usua

lly c

onta

ined

in fe

ed

-10

-12%

cok

e ra

te

-

Bed

ding

, 7 b

in p

ropo

rtion

ing

plan

t

Not

App

licab

le

1

- - - - - - - -""

- - - Ϊ 120

35

42

tpii

outp

ut s

inte

r Ι.

ϊϊ.β

Pro

pane

4

41.s

5s.5

iE

?9

_,

ί.ϊϊ

.β

Ber

zeliu

s S

tolb

erg

Gm

bH

70

200

135

25

Pro

porti

onin

g B

ins

&

Mix

ing

Dru

m

Not

App

licab

le

- - - - - - - - - - - - -N

ot A

pplic

able

- - - - - - - - ■

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Co

mp

any

Nam

e

Exh

aust

Gas

- V

olum

e %

S02

Sm

elti

ng

Num

ber o

f Fur

nace

s Ty

pe

Nom

inal

Cap

acity

D

imen

sion

s

Sm

eltin

g

Aux

iliary

Fue

l - T

ype

Am

ount

C

alor

ific

Val

ue

Air

Vol

ume

Oxy

gen

Enr

ichm

ent

Oxy

gen

Vol

ume

Ope

ratin

g Te

mpe

ratu

re

Flux

as

% o

f Fee

d R

efra

ctor

y Li

ning

- Ty

pe

Cam

paig

n of

Life

Furn

ace

OfF

gas

Vol

ume

Tem

pera

ture

ex

Furn

ace

%S

02

Gas

Coo

ling

by

Num

ber o

f uni

ts

Out

let T

empe

ratu

re

Dus

t Col

lect

ion

by

Dis

posi

tion

of D

ust

Am

ount

Nm

^h'r

%

MTP

H

Kg

(NnT

yhV

kJ/k

jL

Nm

'mr

%

Nm

3 /hr

" " "

r c %

year

s

Nm

3 /hr

°c

%

°C

TPH

Com

inco

Ltd

. Tr

ail O

pera

tions

/ C

anad

a

- - 1 K

IVC

ET

56

24 m

long

x 5

m w

ide

Coa

l 4,

500

kg/h

r 32

,400

- 100

8750

1,40

0 27

W

ater

-coo

led

Cop

per

Ele

men

ts

plus

Chr

ome-

Mag

nesi

te B

ricks

4

21,0

00

1,37

5 1

2to

15

Was

te H

eat B

oile

r

1 350

_j

Ele

ctro

stat

ic P

reci

pita

tor

(Lur

gl)

Rec

ycle

d to

Zin

c P

lant

8

Doe

Run

Com

pany

(1,2

) H

ercu

lane

um /

USA

230,

000

4.0

3 (2

in o

pera

tion)

B

last

Fur

nace

30"

1.76

m

x6.5

m

Cok

e

2550

0 ea

ch F

urna

ce

Yes

1500

7 2

170,

000

0.02

5 D

ilutio

n

300

Bag

Hou

se

Rec

ycle

d to

Sin

ter

Mac

hine

2.5

Met

-Mex

Pen

oles

, S.A

. To

rreo

n /

Mex

ico

91,6

36

5

6 (3

in o

pera

tion)

B

last

Fur

nace

62.5

1.

52m

x6

.4m

Nat

ural

Gas

30

6 66

,427

16.0

00-2

2,50

0 N

o -

Silic

a - A

lum

ina

Bric

ks

45 -

60 d

ays

36,0

00 -

60,0

00

Wat

er N

ozzl

es

120

Noz

zles

iöö-

iiö

Bag

Hou

se

Leac

hing

/ S

olve

nt E

xtra

ctio

n fo

r Zn

& C

d R

ecov

ery

3

Nor

anda

Inc.

B

runs

wic

k S

mel

ter

120,

000

4.0-

5.0

1 IS

F c

onve

rted

to P

b on

ly

37-4

3 tp

h si

nter

thro

ughp

ut

1.68

m x

6.4

m

Cok

e 4.

5

20,5

00

26

800

1,20

0 20

st

eel w

ater

jack

ets

1-2j

fear

s

400-

600

0.02

Bag

Hou

se

Rec

ycle

d to

Sin

ter

Mac

hine

1.25

Ber

zeliu

s S

tolb

erg

Gm

bH

- - 1 Q

SL-

Con

verte

r

33m

long

x 3

/3.5

m d

ia

Pet

role

um C

oke

40Ö

30

,000

600 0

3,50

0

1,15

0 15

C

hrom

e-M

agne

site

Bric

ks

1

15,0

00

*~

1,20

0 20

W

aste

Hea

t Boi

ler

1

420

ES

P (L

urgi

)

Rec

ircul

ated

to

QS

L-C

onve

rter

4-5

LEAD-ZINC 2000

Co

mp

any

Nam

e

Red

ucin

g A

gent

- Ty

pe

Am

ount

C

alor

ific

Val

ue

Air

Vol

ume

Oxy

gen

Enr

ichm

ent

Oxy

gen

Vol

ume

Ope

ratin

g Te

mpe

ratu

re

Ref

ract

ory

Lini

ngj-J

Type

C

ampa

ign

of L

ife

Furn

ace

Offg

as

Vol

ume

Tem

pera

ture

ex

Furn

ace

%S

02

Gas

Coo

ling

by

Num

ber o

f uni

ts

A

Out

let T

empe

ratu

re

Dus

t Col

lect

ion

by

Dis

posi

tion

of D

ust

Am

ount

Furn

ace

Pro

duct

s {a

vera

ge)

Lead

Bul

lion

%P

b %

S

%C

u %

Sb

, %

Ag

Tem

pera

ture

Mat

te

%P

b %

As

%C

u %

Fe

%S

Te

mpe

ratu

re

Sla

g %P

b %

FeO

%

Si0

2

%C

aO

%Z

n Te

mpe

ratu

re

Kg

(Nm

3 )/hr

kJ/k

g

NrA

%

Nm

3 /hr

*"

""c

year

s

Nm

3 flir

°c

%

'c

"

TPH

MTP

D

%

%

%

%

%

r c "

MTP

D

%

%

%

%

%

" "»

c"

" "

MTP

D

%

%

%

%

%

' ""

""c"

~

Com

inco

Ltd

. Tr

ail O

pera

tions

1 C

anad

a

Cok

e 20

0

1,30

0-1,

350

Wat

er c

oole

d co

pper

jack

ets

4

5,00

0 1,

100

Was

te H

eat B

oile

r 1 380

AB

B B

agho

use

Zinc

Ope

ratio

ns, O

xide

Lea

ch

0.84

320

94

1 1.

96

1.42

14

9oz/

mt

850-

950

9 31

.4

0.81

45

.8

0.33

15

.8

1,30

0

650 5 28

20.9

12

.7

17.8

1,

300-

1,35

0

Doe

Run

Com

pany

(1,2

) H

ercu

lane

um /

USA

Met

allu

rgic

al C

oke

800

97

<2

1,20

0

700 2 33

22

12

10

1,20

0

Met

-Mex

Pen

ofes

, S.A

. To

rreo

n /

Mex

ico

Met

allu

rgic

al C

oke

500

96.9

0.06

0.

92

15-1

6ΚΟ

Λ

40

12.8

7.

2 48

.2

6.4

6.7

1,20

0-1,

250

600

1.3

31.4

19

.9

20.5

12

-12.

5 ΪΪ0

0-Ϊ.1

50

Nor

anda

Inc.

B

runs

wic

k S

mel

ter

Met

allu

rgic

al C

oke

360

96.5

1.

3 2.

5 0.

8 0.

45

15

11

1.5 36

31

430

2.0-

5.0

34.0

-38.

0 20

-21

15-1

6 9.

0-11

.0

1,20

0

Ber

zeliu

s S

tolb

erg

Gm

bH

Pul

veriz

ed C

oal

1,40

0 32

,000

1,00

0 0 70Ö

1,25

0 C

hrom

e-M

agne

site

Bric

ks

1.5

Com

bine

d w

ith S

mel

ting

35

0-4

00

94

0.1

-0.3

0.

1 0.

02

0.2

850-

950

2-3

2

5-2

8 21

-2

3 21

-2

3 7

-8

1,25

0

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Co

mp

any

Nam

e

Slap

, Cle

anin

g /

Fum

inq

Num

ber o

f Fur

nace

s Ty

pe

Sla

g Tr

eate

d D

imen

sion

s R

educ

tant

Typ

e

Fina

l Sla

g %

Pb

%C

u %

Zn

Fum

e P

rodu

ctio

n %

Pb

%Z

n

Su

lfu

r F

ixat

ion

Sou

rce

of G

as

Pla

nt -

Type

Rat

ed C

apac

ity

Gas

Vol

ume

Ave

rage

S0

2 in

inle

t gas

A

vera

ge C

onve

rsio

n E

ffici

en

Pro

duct

Gra

de

Ene

rgy

Req

uire

men

t A

uxilia

ry F

uel

Tailg

as S

crub

bing

_____

MTP

H

%

%

t_

*

MTP

H

%

%

MTP

D

Nrr

P/hr

%

%

% H

2S0 4

KW

h/M

T S

ulfu

r

Com

inco

Ltd

. Tr

ail O

pera

tions

/ C

anad

a

1 W

ater

coo

led

stee

l jac

ket

27

6.44

x 2

.4 m

P

ulve

rized

Coa

l

20.5

0.

1 0.

55

2

' 8.

1 22

51

KIV

CE

T O

xida

tion

Sul

furic

Aci

d M

onsa

nto

(3)

1_30

0 -1

^400

26

,000

8.

8 91

93

none

am

mon

ia

Doe

Run

Com

pany

(1,2

) H

ercu

lane

um /

USA

Not

App

licab

le

- - . - . ^

™

-f_

_ . .

Sin

ter

Mac

hine

Sul

furic

Aci

d M

onsa

nto

Sin

gle

Con

tact

34

0 23

0,00

0 4

99.9

93

650

Nat

ural

Gas

N

one

Met

-Mex

Pen

oles

, S.A

. To

rreo

n /

Mex

ico

Not

App

licab

le

- - . - . - . - - . . .

Sin

ter

Mac

hine

Sul

furic

Aci

d P

arso

n S

ingl

e C

onta

ct

570

93,0

00

5-6

98

.5

98.5

62

0 ^

Nat

ural

Gas

B

rink

Nor

anda

Inc.

B

runs

wic

k S

mel

ter

Not

App

licab

le

- - - - - - - - - - -

Sin

ter M

achi

ne

Sul

furic

Aci

d S

ingl

e C

onta

ct

500

120,

000

4.Ö

- 5

.Ö"

98.5

93.5

No

Ber

zellu

s S

tolb

erg

Gm

bH

Not

_Apj

>lica

ble

- - - - - - - . - - - -

QS

L-C

onve

rter

Sul

furic

Aci

d Lu

rgi D

oubl

e C

onta

ct

250

23,0

00

10

-11

99.5

96

90

-11

0

LEAD-ZINC 2000

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 85

Appendix 4 -Primary Lead Smelter Survey - Smelters

CO

Co

mp

any

Nam

e

An

nu

al P

rod

uct

ion

-

Lea

d

Cu-

Dro

ss

Sb-

Sla

g B

i-cru

st

Dor

a O

ther

Typ

e B

ism

uth

met

al

Silv

er

Cop

per

Sul

phat

e S

odiu

m A

ntim

onat

e C

oppe

r Ars

enat

e G

old

Ap

plie

d T

ech

no

log

ies

Ben

efic

atio

n / S

epar

atio

n

Sm

eltin

g

Ref

iner

y

Raw

Mat

eria

ls

Con

cent

rate

s %

Pb

%

Zn

%

Fe

%S

%

SiÖ

2

% H

2Ö

Sec

onda

ries

Type

of S

econ

darie

s

%P

b %

Zn

%F

e %

S

% S

i02

% H

2Ö

MTP

Y

MTP

Y

MTP

Y

MTP

Y

MTP

Y

MTP

Y

MTP

Y

MTP

Y

MTP

Y

MTP

Y

MTP

Y

OzP

Y

MTP

Y

%

%

%

%

%

%

%

%

Bol

iden

Min

eral

AB

R

onns

kar

35,0

00

Cu-

allo

y 6,

000

Ag-

allo

y 40

0

Kal

do F

lash

Sm

elte

r

Pyr

omet

allu

rgic

al R

efin

ing

43,0

00

100

78

1 13

5 5 0 - - - - - -

MH

D-M

.I.M

. Hue

ttenw

erke

D

uisb

urg

Gm

bH

Por

tove

sme

sri /

Ital

y

Impe

rial S

mel

ting

Pla

nt

Kiv

cet P

lant

31

,000

12

5,00

0 9,

000

1 Λ00

0 4,

400

200

Upd

raug

ht S

inte

r M

achi

ne /

Impe

rial S

mel

ting

Furn

ace

Bul

lion

to B

ritan

nia

Ref

ined

M

etal

s / U

K

60

55.7

/ Ϊ2

.5

12.7

/45.

9 3.

11/2

.1

17.6

/25.

8 4

.7/5

.6

2.6/

12.8

R

ed D

qg/M

cArt

hur

40

Mai

nly

Zinc

dom

inat

ed

Sec

onda

ries

Upd

raug

ht S

inte

r Mac

hine

/ Im

peria

l Sm

eltin

g Fu

rnac

e K

ivce

t Fur

nace

Pyr

omet

allu

rgic

al R

efin

ing

205,

000

74.5

16

38

.4

8.5

28.6

3 8

25.5

150,

000

73

64

5.5 6 18.5

2 6 27

B

atte

ry P

aste

, EA

F-bu

st, W

aelz

Oxi

de,

Zinc

Scr

aps,

Zin

c O

xide

s et

c.

5.7

31.5

1.

5 67

.6

2.5

LEAD-ZINC 2000

Co

mp

any

Nam

e

Furn

ace

Flux

S

i02

Lim

e /

Lim

esto

ne

Iran

Coa

l C

oke

Oth

er T

ype

Fee

d P

rep

arat

ion

Ble

ndin

g S

yste

m

Dry

er, I

f app

licab

le

Type

N

umbe

r D

imen

sion

Feed

Rat

e In

let -

H20

O

utle

t-H

20

Fuel

- Ty

pe

Cal

orifi

c V

alue

G

as T

empe

ratu

re

Dry

er In

let

Dry

er O

utle

t G

as V

olum

e D

isch

arge

Tem

pera

ture

Sin

ter

Mac

hine

N

umbe

r D

imen

sion

s, H

earth

Are

a N

omin

al C

apac

ity

Sul

fur

Bur

ning

Rat

e

Fuel

- Ty

pe

Cal

. Val

ue

Feed

Moi

stur

e

Sin

ter

Com

posi

tion

%P

b %

Fe

%S

kg /

MT

- bu

llion

kg /

MT

- bu

llion

kg /

MT

- bu

llion

kg /

MT

- bu

llion

kg /

MT

- bu

llion

kg/M

T-b

uNio

n

m

MTP

H

%

%

kj/k

g

öc

'"

4C

N

mä m

r °C

~ π

?

MTP

H

MT

Su

lfür/

mf *d

kJ/k

g %

H20

%

%

%

Bol

iden

Min

eral

AB

R

onns

kar

- 203

217 - 48 -

Rot

ary

Furn

ace

1 30

5 <

0.1

Oil

800

- 900

9

5-1

05

15,0

00-2

0,00

0 90

Not

App

licab

le

- - . - - - - . - - -

MH

D-M

.I.M

. Hue

ttenw

erke

D

uisb

urg

Gm

bH

'2,1

90

Not

app

licab

le

- - - - - - - - - - - - - 1 72

550

Oil

5j83

3 6 16.1

7.

9 06

2

Por

tove

sme

srl /

Ital

y

26.8

20

5.8

109.

5

Not

App

licab

le

- - . - - - . - - - - - - 1 70

30

2.2

Oil

40,0

00

5

20.5

8.

5 0.

6

... 92

.8

154.

4 23

.2

50

Dou

ble

Cou

nter

-Rot

atin

g P

addl

e S

haft

Rot

atin

g D

rum

1

15m

(len

gth)

x 3

m (d

ia.)

35

8-1

0 0.

1 LP

G G

as

44Λ0

00

35

0-4

20

85

24,0

00

65

Not

App

licab

le

- - . - - - - - - -

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Co

mp

any

Nam

e

Exh

aust

Gas

- V

olum

e %

S02

Sm

elti

ng

Num

ber o

f Fur

nace

s Ty

pe

Nom

inal

Cap

acity

D

imen

sion

s

Sm

eltin

g

Aux

iliary

Fue

l - T

ype

Am

ount

C

alor

ific

Val

ue

Air

Vol

ume

Oxy

gen

Enr

ichm

ent

Oxy

gen

Vol

ume

Ope

ratin

g Te

mpe

ratu

re

Flux

as

% o

f Fee

d R

efra

ctor

y Li

ning

- Ty

pe

Cam

paig

n of

Life

Furn

ace

Offg

as

Vol

ume

Tem

pera

ture

ex

Furn

ace

%S

Ö2

Gas

Coo

ling

by

Num

ber o

f uni

ts

Out

let T

empe

ratu

re

Dus

t Col

lect

ion

by

Dis

posi

tion

of D

ust

Am

ount

s %

MfP

H

Kg

(Nn?

)/hr

kJ/k

g

Nm

ä /hr

%

Nrn

'/hr

°C

%

year

s

Nm

ä /hr

~"

5 c"

%

5c

TPH

Bol

lden

Min

eral

AB

R

onns

kar

- - 1 K

aldo

-Fum

ace

75,0

00

Oil

1,20

0

Yes

Chr

ome-

Mag

nesi

te B

ricks

0.2

25,0

00

4-5

S

crub

ber

1 70

Set

tler

Rec

icul

ated

to F

urna

ce

3

MH

D-M

.I.M

. Hue

ttenw

erke

D

uisb

urg

Gm

bH

41,1

00

62 1

IS-F

umac

e

100,

000

Zn

/ 35,

000

Pb

19.4

Met

allu

rgic

al C

oke

42,0

00

No

4.7

51.0

00

Coo

ling

Tow

er

1 35

Thei

ssen

Des

inte

grat

or

Slu

dge

to S

inte

r M

achi

ne

1.3

Por

tove

sme

srl /

Ital

y

... 60

,000

6

.0-6

.5

2 IS

-Fum

ace

35,0

00

Cok

e 32

5 29

,000

36,0

00

No

1,30

0 6

KR

70C

- D

urita

l

3

60,0

00

450

Scr

ubbe

r

1 50

Thic

kene

r

Rec

ircul

atio

n ba

ck to

Sin

ter

Mac

hine

1.

5

...

- - 1 K

ivce

t-Fur

nace

100,

000

Cok

e 23

7 29

^00

100

6,50

0

1,35

0 17

C

hrom

e-M

agne

site

Bric

ks w

ith

Wat

er C

oole

d C

oppe

r Jac

kets

3-4

12,0

00

1,15

0 2

2-2

5 W

aste

Hea

t Boi

ler

1 &

1

65

ES

P

Rec

ircul

atio

n ba

ck to

Fu

rnac

e 2

5

LEAD-ZINC 2000

Co

mp

any

Nam

e

ΜΜ

ΒΒ

^Μ

ίιΜ

Β^

Ι!Η

ίΗΒ

ΜΐΗ

·"η

Ιιιι

ιιιι

ιιιι

ιι

Red

ucin

g A

gent

- Ty

pe

Am

ount

C

alor

ific

Val

ue

Air

Vol

ume

Öxy

gen

Enr

ichm

enf

Oxy

gen

Vol

ume

Ope

ratin

g Te

mpe

ratu

re

Ref

ract

ory

Lini

ng -

Type

C

ampa

ign

of L

ife

Furn

ace

Offg

as

Vol

ume

Tem

pera

ture

ex

Furn

ace

%S

0 2

Gas

Coo

ling

by_

Num

ber o

f uni

ts

Out

let T

empe

ratu

re

Dus

t Col

lect

ion

by

Dis

posi

tion

of D

ust

Am

ount

Furn

ace

Pro

duct

s (a

vera

ge)

Lead

Bul

lion

%P

b %

S

%C

u %

Sb

%A

g Te

mpe

ratu

re

Mat

te

%P

b %

As

%C

u %

Fe

%S

Te

mpe

ratu

re

Sla

g %P

b %

FeO

%

Si0

2

%C

aO

%2

n Te

mpe

ratu

re

—

- ·—

kg

(N

m'/h

r kj

'/kg

Sl

Ä

%

Nm

ä /rir

8 "c""

' "

year

s

Nm

ä /hr

b "c""

" ~

%

"" s "c

TPH

MTP

D

%

%

%

%

%

r c"

MTP

D

%

%

%

%

%

"""«

c

MTP

D"

%

*"

% %

%

%

" "c

Bol

iden

Min

eral

AB

R

onns

kar

Cok

e 48

Kg/

t-bul

lion

1,05

0-1,

250

Chr

ome-

Mag

nesi

te B

ricks

0.

2

25,0

00

800-

1,20

0 4

Scr

ubbe

r 1 70

S

ettle

r R

ecic

ulat

ed to

Fur

nace

3

1

35,0

00 M

T/Y

ear

(dis

cont

inuo

us)

99.3

0.

2 0.

2 0.

1 0.

05

950

4.2

40

21

28

4.6

1,15

0

MH

D-M

.I.M

. Hue

ttenw

erke

D

uisb

urg

Gm

bH

„Jni

ffijM

äLSm

sllin

af.M

mss

s,,,,

M

etal

lurg

ical

Cok

e

93

95.8

3.1

0.27

0.

15

1,30

0

211

0.97

40

.5

13.8

12

.5

7.4

1,30

0

Por

tove

sme

srl /

Ital

y

„inJm

fiSiia

l.Sm

allin

ftfM

mas

s.

Met

allu

rgic

al C

oke

120

94

Ο.Ϊ

2.7

0.Ϊ5

0.

275

i',30

0

150 1 40

20

ϊ'β

7 i'.ä

öö

—

· 300

98.2

0.

4 0.

6 0.

15

0.1

650

10

40

2 25

5 12

1,10

0

120 4 26

22

20

9 1,

300

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Co

mp

any

Nam

e

Sla

g C

lean

ing

/ Fu

min

g

Num

ber o

f Fur

nace

s Tf

fi".

„_

Sla

g Tr

eate

d D

imen

sion

s R

educ

tant

Typ

e

Fina

l Sla

g %

Pb

%C

u %

Zn

Fum

e P

rodu

ctio

n %

Pb

%Z

n

Su

lfu

r F

ixat

ion

Sou

rce

of G

as

Pla

nt -

Type

Rat

ed C

apac

ity

Gas

Vol

ume

Ave

rage

S0

2 in

inle

t gas

A

vera

ge C

onve

rsio

n E

ffici

en

Pro

duct

Gra

de

Ene

rgy

Req

uire

men

t A

uxilia

ry F

uel

Tailg

as S

crub

bing

MTP

H

MTP

H

%

%

%

MTP

H

%

%

MTP

D

Nm

3 /hr

%

%

% H

2S0 4

KW

h/M

T S

ulfu

r

Bol

iden

Min

eral

AB

R

onns

kar

Cur

rent

ly n

ot U

sed

for

Lead

Sla

gs

1 R

ever

bera

tory

50

Coa

l

250,

000

MT/

Yea

r 0.

03

0.5

1.4

42,0

00 M

Y/Y

ear

5.5

72

Kal

do &

Cu-

Sm

elte

r

Sul

furic

Aci

d D

oubl

e C

onta

ct

1 λ00

0 14

0,00

0 5.

5 Θ

9.8

94.5

No

MH

D-M

.I.M

. Hue

ttenw

erke

D

uisb

urg

Gm

bH

Not

App

licab

le

. - - - . - - - - . - .

Raw

Gas

from

Sin

ter

Mac

hine

Sul

furic

Aci

d D

oubl

e C

onta

ct

280

57,0

00

6.2 96

186

No

Por

tove

sme

sri /

Ital

y

...

Not

App

licab

le

. - . . . . - . - - - -

Sin

ter

Mac

hine

Sul

furic

Aci

d D

oubl

e C

onta

ct

, 45

0 60

,000

6.

5 98

98.8

11

0 Li

ght O

il N

o

...

Not

App

licab

le

- - - - - . - - . . - -

Kiv

cet F

urna

ce

Mix

ed to

geth

er w

ith G

as fr

om

Zn-r

oast

er

Sul

furic

Aci

d D

oubl

e C

onta

ct

180

12,0

00

22

-25

98

98.5

75

Li

ght O

il N

o

LEAD-ZINC 2000

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 91

Appendix 5 -Primary Lead Smelter Survey - Pyrometallurgical Refineries

Co

mp

any

Nam

e

Ref

iner

y -

Typ

e

An

nu

al P

rod

uct

ion

- L

ead

S

oft L

ead'

H

ard

Lead

P

b-C

a-A

lloys

Pyr

om

etaM

urg

ical

Ref

iner

y

Dec

oppe

rfzi

ng

Met

hod

Ket

tle S

ize

Num

ber o

f Ket

tles

Inle

t Cu-

cont

ent

Tem

pera

ture

Am

ount

Am

ount

of D

ross

Tr

eatm

ent o

f Cu-

dros

s

Fina

l By-

Pro

duct

A

mou

nt

%P

b %

Cu

%S

Sof

teni

ng

Met

hod

MTP

Y

%

%

%

T

MTP

Y

%

°c

Kg/

T B

ullio

n K

g/T

Bul

lion

MTP

Y

%

HR

/Ket

tle

ΜΤΡ

Ϋ

MTP

Y

%

%

%

Pas

min

co P

ort P

ine

Sm

elte

r A

ustr

alia

Pyr

omet

allu

rgic

al R

efin

ing

250,

000

89

3 8

Con

tinuo

us D

ross

ing

Furn

ace

i Fu

rnac

e 27

0,00

0 2.

Ö -

2.5

45

0

260,

000

0.01

Cop

per

Mat

te

Ί 3Λ50

0 38

40

15

Oxy

gen

Trea

tmen

t

Toh

o Zi

nc C

o.

Chi

giri

shim

a P

lant

Con

tinuo

us D

ecop

pehz

ing

see

Bet

ts R

efin

ery

Con

tinuo

us D

ross

ing

iöö

3 79

,ÖÖ

Ö

2 34Ö

-

78,0

00

Ö.0

3

i'ö.'ä

15,0

50

Ele

ctric

Fur

nace

2,70

0

ϊέ'

36

2Ϊ.3

'

Not

App

licab

le

Hin

dust

an Z

inc

Ltd

., C

hand

eriy

a / I

ndia

Pyr

omet

allu

rgic

al R

efin

ing

35,0

00

100 - -

Pre

cipi

tatio

n &

Sul

phur

Add

ition

100 4

54,4

39

5 9

-6.3

90

0 -

450

Saw

Dus

t & S

ulph

ur

3 (S

aw D

ust}

2.5

[Sul

phur

) 44

,601

0.

035

20

-22

10,4

90

Pro

duct

ion

of C

uSO

< fo

r E

W

9,91

7 50

34

.4

trace

s

Oxi

datio

n af

ter V

acuu

m

Dez

inci

ng

Cau

stic

Tre

atm

ent

Asa

rco

Inc.

E

ast H

elen

a P

lant

Pyr

omet

allu

rgic

al R

efin

ing

65,0

00 -

75,0

00

Pre

cipi

tatio

n &

Sul

phur

Add

ition

125 3

65,0

00 -

75,0

00

0.2

-0.5

37

0 C

oke

Bre

eze

& S

ulph

ur

0.02

6

27,0

00

Rev

erbe

rato

ry F

urna

ce

Mat

te /

Spe

iss

9.4

/11

.1

46.7

/ 3

6.4

17

.5/0

.1

Not

App

licab

le

-

Met

-Mex

Pen

oles

, S.A

. To

rreo

n /

Mex

ico

Pyr

omet

allu

rgic

al R

efin

ing

180,

000

72

28

Pre

cipi

tatio

n &

Sul

phur

Add

ition

250 6

180,

000

0.06

Sul

fur

176,

260

0.00

6 3

3,50

0 R

ecyc

le to

Sm

elte

r

3,50

0 78

.8

5.4

Rev

erbe

rato

ry F

urna

ce

LEAD-ZINC 2000

Co

mp

any

Nam

e

Ke|fe

,SiS

6,

Num

ber

of K

ettle

s T

reat

ed B

ullio

n In

let S

b-co

nten

t Te

mpe

ratu

re S

tart

Tem

pera

ture

End

R

eact

ion

Age

nt(s

) A

mou

nt

Fina

l Bul

lion

Fina

l Sb-

cont

ent

Dur

atio

n of

Tre

atm

ent

Am

ount

of

Dro

ss/S

lag

Trea

tmen

t of S

b-D

ross

/Sla

g

Fina

l By-

Pro

duct

A

mou

nt

%P

b %

Sb

%N

a

Det

inni

ngi

Met

hod

Ket

tle S

ize

Num

ber o

f Ket

tles

Tre

ated

Bul

lion

Inle

t Sn-

cont

ent

Tem

pera

ture

Sta

rt T

empe

ratu

re E

nd

Rea

ctio

n A

gent

(s)

Am

ount

Fina

l Bul

lion

Fina

l Sn-

cont

ent

Dur

atio

n of

Tre

atm

ent

Am

ount

of

Dro

ss/S

lag

Trea

tmen

t of

Sn-

Dro

ss/S

tag

Fina

l By-

Pro

duct

A

mou

nt

%P

b %

Sn

%N

a

I M

TPY

%

°C

°C

KgT

T B

ullio

n K

g/T

Bul

lion

MTP

Y

%

HR

/Ket

tJe

MTP

Y

ΜΤΡ

Ϋ

%

%

%

T

MTP

Y

%

°c

°c

Kg/

T B

ullio

n K

g/T

Bul

lion

MTP

Y

%

HR

/Ket

iie

MTP

Y

MTP

Y

%

%

%

Pas

min

co P

ort P

lrie

Sm

elte

r A

ustr

alia

4,o,p,

2

260,

000

0.55

45

5 47

0 O

xyge

n 1.

5

250,

000

<0.0

1 20

10

A000

S

lag

Red

uctio

n

Ant

imon

y A

lloy

6,50

0 94

5.

5

-N

ot A

pplic

able

- - - - - - - - - - - - - - - - - . - -

Toho

Zin

c C

o.

Chi

glri

shim

a P

lant

. - - - - - - - - - - - - - _ - -

Not

App

licab

le

- - . - - - - - . - -1

- - - - - - - -

Hin

dust

an Z

inc

Ltd

., C

hand

eriy

a / I

ndia

,i,§a

„,„„

1 40

,125

0.

49

490

460

Cau

stic

and

Nite

r 5.

2 1.

6 39

,581

00

005

15

975

Pro

duct

ion

of A

ntjm

onal

Lea

d

975

55.3

3 18

De-

arse

natin

o.

150 1

44,7

23

0,27

.(As)

40

0 46

0 C

aust

ic S

oda

4.9

44,5

97

0.01

(A

s)

12

613

Sm

elte

r Tre

atm

.for

Pb-

reco

very

613

17.3

7 17

.69

(As}

Asa

rco

Inc.

E

ast

Hel

ena

Pla

nt

. - - - - - - - - - - - - - - - - -

Not

App

licab

le

- - - . - - - - - - - - - - - - - -

Met

-Mex

Pen

oles

, S.A

. To

rreo

n /

Mex

ico

„ „3

P,

2 17

6,62

0 0.

89

Air

+ O

xyge

n

171,

350

0.04

C

ontin

uous

4,

800

Rec

ycle

Bla

st F

urna

ce

4,80

0 5

28

29.7

-N

ot A

pplic

able

- - - - - - - - - - - . - - - - - -

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Co

mp

an

y N

am

e

De

sil

ve

rizi

ng

M

etho

d

Ket

tle S

ize

Num

ber o

f K

ettle

s A

mou

nt o

f S

teps

T

reat

ed B

ullio

n In

let A

g-co

nten

t Te

mpe

ratu

re S

tart

Tem

pera

ture

End

R

eact

ion

Age

nt(s

) A

mou

nt

Fin

al B

ullio

n Fi

nal A

g-co

nten

t D

urat

ion

of T

reat

men

t A

mou

nt o

f Dro

ss/C

rust

Tr

eatm

ent o

f Ag-

Dro

ss/C

rust

Fina

l By-

Pro

duct

A

mou

nt

%P

b ...

...%

Ag.

%

Cu

%Z

n

Dez

inci

ng, i

f app

licab

le

Met

hod

Ket

tle S

ize

Num

ber o

f Ket

tles

Tre

ated

Bul

lion

Inle

t Zn-

cont

ent

Tem

pera

ture

R

eact

ion

Age

ntfs

) A

mou

nt

App

lied

Vac

uum

Fi

nal B

ullio

n Fi

nal Z

n-co

nten

t D

urat

ion

of T

reat

men

t A

mou

nt o

f D

ross

/Met

al

%P

b %

Zn

T

MTP

Y

%

°C

°C

Kg/

T B

ullio

n K

g/T

Bul

lion

MTP

Y

%

HR

/Ket

tle

MTP

Y

MTP

Y

%

%

%

%

T

MTP

Y

%

°C

Kg/

T B

ullio

n m

bar

MTP

Y

%

HR

/KeS

e M

TPY

%

%

Pas

min

co P

ort P

irle

Sm

elte

r A

ustr

alia

Tw

o S

tage

Par

kes

Pro

cess

'4Ö

Ö

3 2 2

50

,00

0 0

.25

460

324

Zin

c

12

24

5,0

00

0.0

00

5 20

[se

con

d S

tag

e)

10

(1st

Sta

ge

9,6

00

Liq

ua

tio

n

Liq

ua

tio

n H

igh

Gra

de

Allo

y 1,

720

8 25

1 66

Va

ccu

um

D

ezi

nci

ng

400 2

24

5,0

00

0.4

580 - -

0.0

1

245,

005

0.0

4

18

1,20

0

95

4

To

ho

Zin

c C

o.

Ch

igir

ish

tma

Pla

nt

No

t A

pp

lica

ble

- - - - - - - - - - - - - - - - - - - -N

ot

Ap

plic

ab

le

- - - - - - - - - - - - - - -

Hin

du

sta

n Z

inc

Ltd

.,

Ch

an

de

riy

a /

Ind

ia

Tw

o S

tag

e P

ark

es

Pro

cess

Ϊ8Ϊ

2 2 4

4,5

98

0.2

46Ö

32

Ö

Zin

c (9

9.9

9%

) 1

2.8

44

,70

0

ö.oö

i Ϊ4

(Is

täg

e)

&2Ö

'p.S

täge

) 1,

672

Liq

ua

tio

n

Silv

er

Ric

h C

rust

1,

672

73

6 Ö.6

1

9.7

6

Va

cuu

m

De

zin

cin

g

ISO

Ϊ

44,7

00

Ö.5

6

59

0-6

10

Ο.Ο

Ϊ3

44

,33

0

Ö.Ö

5 13

Ϊ6

Ό

Ϊ 9

7.5

As

arc

o I

nc

. E

as

t H

ele

na

Pla

nt

No

t A

pp

lica

ble

. - - - - - - - - - - - - - - - - - - - - -N

ot

Ap

plic

ab

le

. - - - - - - - - - - - - - -

Me

t-M

ex

Pe

no

les

, S

.A.

To

rre

on

/ M

ex

ico

Th

ree

Sta

ge

Pa

rke

s P

roce

ss

30Ö

6 3

18

1,4

50

1.4

Zin

c i"2

- iä

17

0,2

00

Ö.ÖÖ

ÖS

12

Ϊ5,9

βδ

Liq

ua

tio

n

Pa

rke

s A

lloy

15

,96

0 65

Ϊ2

Ϊ5

Ö.Ö

31

2Ö.i

Va

cuu

m

De

zin

cin

g

280 2

17

0,2

00

0.5

4

50

mm

Hg

16

9,3

00

0.05

Ϊ 8 924

Ϊ00

LEAD-ZINC 2000

Co

mp

any

Nam

e

MÄn

irtlM

stoK

i M

eth

od

Ke

ttle

Siz

e N

um

be

r o

f K

ett

les

Am

ou

nt

of

Ste

ps

Tre

ate

d B

ulli

on

Inle

t B

i-co

nte

nt

Te

mp

era

ture

Sta

rt

Te

mp

era

ture

En

d

Re

act

ion

Ag

en

t's)

A

mo

un

t

Fin

al

Bu

llio

n F

ina

l B

i-co

nte

nt

Du

rati

on

of

Tre

atm

en

t A

mo

un

t o

f D

ross

/Cru

st

Tre

atm

en

t o

f B

i-D

ross

/Cru

st

Fin

al

By-

Pro

du

ct

Am

ou

nt

%P

b %

Bi

%C

a %

Mg

Fin

al

Re

fin

ing

M

eth

od

Ke

ttle

Siz

e N

um

be

r o

f K

ett

les

Am

ou

nt

of

Ste

ps

Tre

ate

d B

ulli

on

Te

mp

era

ture

Sta

rt

Te

mp

era

ture

En

d

Re

act

ion

Ag

en

t(s)

A

mo

un

t

Fin

al

Bu

llio

n

Du

rati

on

of

Tre

atm

en

t A

mo

un

t o

f D

ross

T

rea

tme

nt

of

Dro

ss

T

MT

PY

%

°c

°c

Kg

/T B

ulli

on

Kg

/T B

ulli

on

ΜΤΓ

ΡΫ

%

HR

/Ke

fie

MT

PY

MT

PY

%

%

%

%

T

MT

PY

°C

°C

Kg

/T B

ulli

on

Kg

/T B

ulli

on

MT

PY

H

R/K

ett

le

MT

PY

Pa

sm

inc

o P

ort

Pir

ie S

me

lte

r

Au

str

ali

a

Kro

ll B

ett

ert

on

200 2 3

21

0,0

00

0.5

6 41

5 40

0 C

a/M

g/S

b C

a: 0

.32

(cru

de

} 0

.35

(ful

l) M

g:

1.1

(cru

de

) 1.

8 (f

ull)

21

0,0

00

02

9m

ax

(cru

de

) 0

05

ma

x (f

ull)

1 20

0 E

xte

rna

l B

y- P

rod

cut

Bi

Cru

st

200

40

40

10

10

Ca

ust

ic R

efi

nin

g

200 5 1

24

5,0

00

430

430

Ca

ust

ic S

od

a 1 1

24

2,0

00

1 8

,00

0 S

inte

r P

lan

t

To

ho

Zin

cC

o.

Ch

igir

ish

ima

Pla

nt

.Nffl

t.ifti

PB

liSiS

.ftfS

,

- - - - - - - - - - - - - - - - - - - - -

No

t A

pp

lica

ble

- - - - - - - - - . - - -

Hin

dust

an Z

inc

Ltd.

, C

hand

eriy

a / I

ndia

, M

,oJ,A

DR

||caj8

|g

„„„

- - - . - - - - - - . - - - - - - . -

Oxi

datio

n

150 1 1

39,4

24

450

430

Cau

stic

Sod

a &

Nite

r 0.

5 0.

1 39

.244

10

-12

120

Asa

rco

Inc.

E

ast

Hel

ena

Pla

nt

, iü,

°i;.A p,R

lisäte

!s

- - - . . - - - . - . . . . - . - . - . -

No

t A

pp

lica

ble

. - . - - - - . - - - . -

Me

t-M

ex

Pe

no

les

, S

.A.

To

rre

on

/ M

ex

ico

Kro

ll-B

ett

ert

on

250 4 1

16

9,3

00

0.7

04

Ma

gn

esi

um

/ C

alc

ium

Allo

y 0

.49

Kg

/Kg

Bi

16

6,5

00

0.0

1 26

1

5,7

00

Bis

mu

th

Pro

cess

15

,70

0 8

7.7

7.

8 1.

15

2.5

Ca

ust

ic S

od

a

250 4 1

16

6,5

00

Ca

ust

ic S

od

a 0.

8

16

3,5

00

2 4

,20

0 R

ecy

cle

to S

me

lte

r

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Co

mp

any

Nam

e

Cas

ting

Mac

hine

- T

ype

Num

ber o

f Mou

lds

Rat

e C

astin

g Te

mpe

ratu

re

Cas

t Pro

duct

In

got S

ize

Ingo

t Wei

ght

Rej

ects

as

% o

f New

Mat

eria

l

MTP

H

°C

mm

K

g %

Pas

min

co P

ort P

irie

Sm

elte

r A

ustr

alia

Stra

ight

Lin

e M

ould

ing

/ B

lock

Cas

ting

36 (S

LMM

) SO

(bl

ock

whe

el)

420

refin

ed le

ad/le

ad a

lloy

600x

80x7

5 25

5

Toho

Zin

c C

o.

Chi

glrl

shlm

a P

lant

H

indu

stan

Zin

c Lt

d.,

Asa

rco

lnc.

M

et-M

ex P

enol

es, S

A

Cha

nder

iya

/ Ind

ia

Eas

t Hel

ena

Pla

nt

Torr

eon

/ M

exic

o

Stra

ight

Lin

e M

ould

ing

Stra

ight

Lin

e M

ould

-Cha

in

260

20

40

440

Pur

e Le

ad

540

x 95

x 7

5 25

0.

5-1.

0 [

44

0.3

LEAD-ZINC 2000

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 97

Appendix 6 -Primary Lead Smelter Survey - Pyrometallurgical Refineries

Co

mp

any

Nam

e

Ref

iner

y ■

Typ

e

An

nu

al P

rod

uct

ion

- L

ead

S

oft L

ead

Har

d Le

ad

Pb-

Ca-

Allo

ys

Pyr

om

etal

lurg

ical

Ref

iner

y

Dec

oppe

nzin

g

Met

hod

Ket

tle S

ize

Num

ber o

f Ket

tles

Tre

ated

Bul

lion

Inle

t Cu-

cont

ent

Tem

pera

ture

R

eact

ion

Age

nt's

) A

mou

nt

Fina

l Bul

lion

Fina

l Cu-

cont

ent

Dur

atio

n of

Tre

atm

ent

Am

ount

of

Dro

ss

Trea

tmen

t of C

u-dr

oss

Fina

l By-

Pro

duct

A

mou

nt

%P

b %

Cu

%S

Sof

teni

ng

Met

hod

MTP

Y

%

%

%

T

MTP

Y

%

°C

Kg/

T B

ullio

n K

g/T

Bul

lion

MTP

Y

%

HR

/Ket

tle

MTP

Y

MTP

Y

%

%

%

Nor

anda

Inc.

B

runs

wic

k S

met

ter

Pyr

omet

allu

rgic

al R

efin

ing

108,

000

85 - 15

Cop

per

Dro

sstn

g-R

ever

bera

tory

+

CS

D

230 3

130,

000

1.5

-3.0

90

0 so

da a

sh

12.3

123,

500

0.07

co

ntin

uous

17,0

00

Rev

erbe

rato

ry F

urna

ce

Mat

te /

Spe

iss

10

-21

%

38

Ber

zeiiu

s S

tolb

erg

Gm

bH

Sto

lber

g /

Ger

man

y

Pyr

omet

allu

rgic

al R

efin

ing

100,

000

iö'ö - -

Coi

cord

(P

yrite

& S

ulph

ur)

18

0/1

50

3-2

i'iö

'.ö'öö

0.

1 -5

.0

44Ö

-47Ö

P

yrite

/ S

ulph

ur

Pyr

ite 4

.5

Sul

phur

1.2

ide

.ö'uö

<0

.005

1

8-4

8 [in

ci. m

eltin

g)

5,00

0 R

otar

y Fu

rnac

e M

eltin

g to

Mat

te

Cu-

Mat

te

80

0-1

00

0 3

8-4

2 4

0-4

5 1

5-1

7

Air/

Oxy

gen

Trea

tmen

t O

xyge

n Tr

eatm

ent

Bol

iden

Min

eral

AB

R

onns

kar /

Sw

eden

B

rita

nnia

Ref

ined

Met

als

Ltd.

N

ort

Mle

et/U

K

Brit

anni

a R

efin

ed M

etal

s Lt

d.

No

rth

flee

t/U

K

: j

Pyr

omet

allu

rgic

al R

efin

ing

35,0

00

100 - -

100 1

35,0

00

0.2

340

- 42

0 S

ulph

ur

0.4

35,0

00

<0.

03

900

Rec

ycle

to S

mel

ting

Furn

ace

Not

App

licab

le

-

Pyr

omet

allu

rgic

al R

efin

ing

Bul

lion

from

M.I.

M. A

ustr

alia

15

1.00

0 88

12

Cop

per

Rem

oval

take

s pl

ace

in

Aus

tralia

- - . - . - - - - - - - - - - - - - -P

rimar

y D

e-S

b ta

kes

plac

e in

A

ustra

lia

-

Pyr

omet

allu

rgic

al R

efin

ing

Bul

lion

from

ISF-

Pla

nts

74,0

00

100

Pre

cipi

tatio

n

250 4

86,0

00

3.5

490

- 33

0 P

ulve

rized

Flu

e A

sh

3.5

76,0

00

0.05

2 36

10,0

00

Scr

eene

d, C

ondi

tione

d an

d S

old

to U

nion

Min

ere

9,80

0 54

.5

25.8

«

1

Take

s P

lace

afte

r D

ezin

cing

Cau

stic

Sod

a

LEAD-ZINC 2000

Co

mp

any

Nam

e

ßS

ßl«

Ä8.

.™„.

, ,

„„„„

„„„„

N

umbe

r of K

ettle

s T

reat

ed B

ullio

n In

let S

b-co

nten

t Te

mpe

ratu

re S

tart

Tem

pera

ture

End

R

eact

ion

Age

nt(s

) A

mou

nt

Fina

l Bul

lion

Fina

l Sb-

cont

ent

Dur

atio

n of

Tre

atm

ent

Am

ount

of

Dro

ss/S

lag

Trea

tmen

t of

Sb-

Dro

ss/S

lag

Fina

l By-

Pro

duct

A

mou

nt

%P

b %

Sb

%N

a

Det

inni

ng.

Met

hod

Ket

tle S

ize

Num

ber o

f Ket

tles

Tre

ated

Bul

lion

Inle

t Sn-

cont

ent

Tem

pera

türe

Sta

rt Te

mpe

ratu

re E

nd

Rea

ctio

n A

gent

fs)

Am

ount

Fina

l Bul

lion

Fina

l Sn-

cont

ent

Dur

atio

n of

Tre

atm

ent

Am

ount

of D

ross

/Sla

g^

Trea

tmen

t of S

n-D

ross

/Sla

g

Fina

l By-

Pro

duct

A

mou

nt

%P

b %

Sn

%N

a

I M

TPY

%

°C

°C

Kg/

T B

ullio

n K

g/T

Bul

lion

MTP

Y

%

HR

/Ket

tle

MTP

Y

ΜΤΡ

Ϋ

%

%

%

T

MTP

Y

%

°c

°c

Kg/

T B

ullio

n K

g/T

Bul

lion

MTP

Y

%

HR

/Ket

tle

MTP

Y

MTP

Y

%

%

%

Nor

anda

Inc.

B

runs

wic

k S

mel

ter

„„„„

„ffiff

, 2

123,

500

500

620

Air

+ O

xyge

n

117,

500

0.00

7 12

6,

000

SR

F s

mel

ting

Pb^

Sba

iioy

3,50

0 85

- 9

2%

7-1

3%

-N

ot A

pplic

able

- - - - - - - - . - - . - - - - . - . -

Ber

zellu

s S

tolb

erg

Gm

bH

Sto

lber

g /

Ger

man

y

„1,5,8

, 1

16,0

00

0.4

- 0.

8 60

0 65

0 O

xyge

n 2.

5 (N

m3 )

15,7

00

<0.1

12

40

0 S

ale

Pb

/stS

lag

4ÖÖ

5

8-6

2 2

5-2

9

Not

App

licab

le

- - - - - - - - - - - -" 1

- - . . - -

Bol

iden

Min

eral

AB

R

on

nsk

ar/

Sw

eden

. - - - - - - - - . - - - - - . - -N

ot A

pplic

able

- - - - - - - - - - - . - - - - . - - -

Bri

tann

ia R

efin

ed M

etal

s Lt

d.

No

rth

flee

t/U

K

. - - . - - - - - - . - - - - - . . -N

ot A

pplic

able

- - - - - - - - - - _ - _ - - - - . - -

Bri

tann

ia R

efin

ed M

etal

s L

td.

No

rth

flee

t/U

K

„„

2.P.0,

1

77,5

50

0.35

51

0 48

0 C

aust

ic S

oda

/ N

iter

Cau

stic

Sod

a 4.

2 N

iter

1.1

76,6

40

0.00

01

12

2,10

0 R

otar

y Fu

rnac

e

Not

App

licab

le

- - - - - - - - - - - - - - - - - - - -

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Co

mp

any

Nam

e

Des

ilver

izin

g M

etho

d

Ket

tle S

ize

Num

ber o

f Ket

tles

Am

ount

of S

teps

Tr

eate

d B

ullio

n In

ietÄ

g-co

nten

t Te

mpe

ratu

re S

tart

tem

pera

ture

End

R

eact

ion

Age

nt(s

) A

mou

nt

Fina

l Bul

lion

Fina

l Äg-

cont

ent

Dur

atio

n of

Tre

atm

ent

Am

ount

of

Dro

ss/C

rust

Tr

eatm

ent o

f Ag-

Dro

ss/C

rust

Fina

l By-

Pro

duct

A

mou

nt

j %

Pb

......

% A

g....

. %

Cu

%2

n

Dez

inci

ng,,.

if a

pplic

able

M

etho

d

Ket

tle S

ize

, N

umbe

r of

Ket

tles

Trea

ted

Bul

lion

4

Inle

t Zn-

cont

ent

Tem

pera

ture

R

eact

ion

Age

nt(s

) A

mou

nt

App

lied

Vac

uum

Fi

nal B

ullio

n Fi

nal Z

n-co

nten

t D

urat

ion

of T

reat

men

t A

mou

nt o

f D

ross

/Met

al

%"p

'b

1

%Z

n

T

MTP

Y

%

■c

°C

Kg/

T B

ullio

n K

g/f

Bul

lion

MTP

Y

%

HR

/Ket

Ue

MTP

Y

MTP

Y

%

%

%

%

T

MTP

Y

%

°c

Kg/

T B

ullio

n m

bar

MTP

Y

%

HR

/Ket

tie

MTP

Y

%

%

Nor

anda

Inc

. B

runs

wic

k S

mel

ter

Tw

o S

tage

Par

kes

Pro

cess

260 4 2

117,

500

0.03

7 50

0 32

0 S

HG

zin

c 6

4

116,

500

0.00

08

16

4,40

0 liq

uatio

n

Dor

e S

ilver

35

0 0.

06

> 99

.7

0.1

Vac

uum

Dez

inci

ng

200 1

116,

500

0.5

600

0.01

5 11

5,90

0 0.

07

6 500 2 98

Ber

zeliu

s S

tolb

erg

Gm

bH

Sto

lber

g / G

erm

any

Par

kes-

Pro

cess

150 4 1

106,

000

0.1

-0.5

48

0 <

320

Zinc

3

3

105,

500

< 0.

001

26

7 U00

0 Li

quat

ion

/ Cup

eiat

ion

Dor

e S

ilver

20

0 <

0.1

>99

.7

<0.

2

Vac

uum

Dez

inci

ng

150 2

105,

500

0.6

530

<0.

01

ϊϊδ,Ο

'Ο'Ο

<

0.05

1

2-1

4 50

0 10

90

Bol

iden

Min

eral

AB

R

onns

kar

/ Sw

eden

Par

kes-

Pro

cess

120 1 Ϊ

35,0

00

0.05

50

0 32

5 Zi

nc

4.5

35,0

00

< 0.

001

Bri

tann

ia R

efin

ed M

etal

s L

td.

No

rth

fle

etf

UK

Mod

ified

Par

kes-

Pro

cess

340 3 2

150,

000

0.23

5 47

0 31

8 G

OB

Zin

c ΪΪ

.8

154,

000

0.00

03

10 h

(1st

sta

ge);

18

h (2

nd s

tage

] 2,

005

Tea

Pot

Zn-P

b-A

g-A

lloy

170 5 10

3 62

Vac

uum

Dez

inci

ng

120 1

35,0

00

0.55

56

0

<0.

1 35

,000

<

0.05

200

10

90

4,10

0 Li

quat

ion

w

"Low

grad

e" L

ead

2,60

0 96

0.

25

0 3.6

Vac

uum

Dez

inci

ng

340 2

154,

000

0.54

59

0

153,

000

0.01

5 6 650

95

26

Brit

anni

a R

efin

ed M

etal

s L

td.

No

rth

fle

et/

UK

Mod

ified

Par

kes-

Pro

cess

250 3 2

76,0

00

0.17

up

to >

0 4

5 47

0 32

0 G

OB

Zin

c 14

.8

78,0

00

0.00

05

10 h

(1st

sta

ge);

18

h (2

nd s

tage

4,

200

Liqu

atio

n

Low

grad

e" L

ead

2,60

0 95

.7

0.25

- 4

Vac

uum

Dez

inci

ng

210 1

78,0

00

0.54

59

0

77,5

50

0.01

5 6 35

0 95

2.

6

LEAD-ZINC 2000

Co

mp

any

Nam

e

(Mto

nutK

ilsin

s,

Met

hod

Ket

tle S

ize

Num

ber o

f Ket

tles

Am

ount

of S

teps

T

reat

ed B

ullio

n In

let

Bi-c

onte

nt

Tem

pera

ture

Sta

rt Te

mpe

ratu

re E

nd

Rea

ctio

n A

gent

(s)

Am

ount

Fina

l Bul

lion

Fina

l Bi-c

onte

nt

Dur

atio

n of

Tre

atm

ent

Am

ount

of D

ross

/Cru

st

Trea

tmen

t of

Bi-D

ross

/Cru

st

Fina

l By-

Pro

duct

A

mou

nt

%P

b %

Bi

%C

a %

Mg

Fina

l Ref

inin

g

Met

hod

Ket

tle S

ize

Num

ber

of K

ettle

s A

mou

nt o

f Ste

ps

Tre

ated

Bul

lion

Tem

pera

ture

Sta

rt Te

mpe

ratu

re E

nd

Rea

ctio

n A

gent

js)

Am

ount

Fina

l Bul

lion

Dur

atio

n of

Tre

atm

ent

Am

ount

of D

ross

Tr

eatm

ent o

f Dro

ss

T

MTP

Y

%

°C

°C

Kg/

T B

ullio

n K

g/T

Bul

lion

MTP

Y

%

HR

/Ket

tJe

MTP

Y

MTP

Y

%

%

%

%

T

MTP

Y

°C

°C

Kg/

T B

ullio

n K

g/T

Bul

lion

MTP

Y

HR

/Ket

tle

MTP

Y

Nor

anda

Inc.

B

runs

wic

k S

mel

ter

Kro

ll-B

ette

rton

260 2 1

116,

600

0.25

48

0 32

0 30

%C

a-70

%M

g al

loy

3.4

111,

500

0.00

8-0.

015

16

5,40

0 O

xida

tion

4,00

0 93

7

Cau

stic

Sod

a / N

iter

230 2 1

111,

500

450

550

0.8

1.3

108,

000

4 4,

000

Rec

ycle

to S

inte

r P

lant

Bar

zeliu

s S

tolb

erg

Gm

bH

Sto

lber

g /

Ger

man

y

Kro

ll-B

ette

rton

Pro

cess

150 2 1

70,0

00

ό'.ο'έ

'-ο.ϊ'έ

50

0 34

0 M

agne

sium

/ C

alci

um A

lloy

2.4

69,0

00

<0.

01

16

6,00

0 E

nric

hmen

t by

Kro

ll-B

ette

rton

Pro

cess

Pb/

Bi-A

llloy

200

86

-89

10

-13

- -

Cau

stic

Sod

a / N

iter

150 3 1

104,

000

45Ö

52

0 -

580

Cau

stic

Sod

a / N

iter

Cau

stic

Sod

a 4.

0 N

iter

0.1

100,

000

12

5,00

0 R

otar

y Fu

rnac

e

Bol

iden

Min

eral

AB

R

onns

kar/

Sw

eden

MflH

.ftfiR

liSSH

Sm»,

- - - - - - - - - - - - - - - - - - - -

Cau

stic

Sod

a / N

iter

100 1 1

35.0

00

550

500

Cau

stic

Sod

a / N

iter

Cau

stic

Sod

a 40

N

iter

10

35,0

00

1,30

0 R

ecyc

le to

Sm

eltin

g Fu

rnac

e

Bri

tann

ia R

efin

ed M

etal

s Lt

d.

No

rth

flee

t/U

K

MslÄ

Hißl

ifiäte

lS.

,

- - - - - - - - - - . - - - . - - - -

Cau

stic

Sod

a /

Nite

r

300 3 1

153,

000

510

480

Cau

stic

Sod

a /

Nite

r C

aust

ic S

oda

Ϊ.9

Nite

r 0

4 15

1,00

0 6

1,98

0

Bri

tann

ia R

efin

ed M

etal

s L

td.

No

rth

flee

t/U

K

Kro

ll-B

ette

rton

200 1 1

73,0

00

0.21

47

0 33

0 M

agC

ai e

x Ti

mm

inco

2.

3

71,5

00

0.01

4 12

2,

100

Rem

eltin

g an

d O

xida

tion

of

Cal

cium

Mag

nesi

um C

onte

nt

2,10

0 93

.5

6.5

Cau

stic

Sod

a /

Nite

r

180

μ 2 1

71,5

00

330

530

Cau

stic

Sod

a / N

iter

k C

aust

ic S

oda

0.6

Nite

r 0.

7 70

.000

9

2,00

0 R

otar

y Fu

rnac

e

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Co

mp

any

Nam

e

ftast

inu

„„„

,. „„

„„ C

astin

g M

achi

ne -

Type

Num

ber o

f Mou

lds

Rat

e C

astin

g Te

mpe

ratu

re

Cas

t Pro

duct

In

got S

ize

Ingo

t Wei

ght

Rej

ects

as

% o

f New

Mat

eria

l

MTP

H

°C

mm

K

g

%

Nor

anda

Inc.

B

runs

wic

k S

mel

ter

20

-30

500

ingo

t and

jum

bos

50 lb

-1

tonn

e 4

Ber

zellu

s S

tolb

erg

Gm

bH

Sto

lber

g /

Ger

man

y

Mou

ld-C

hain

(S

hepp

ard)

400

50

45

0-5

00

Ingo

ts

50

<2

Bol

iden

Min

eral

AB

:

Bri

tann

ia R

efin

ed M

etal

s Lt

d.

Ro

nn

skar

/Sw

eden

N

ort

hfl

eet/

UK

I M

ETP

RO

j C

ontin

uous

Bel

t, C

astin

g W

heel

240

i 36

4 &

269

(Bel

t); 2

5 (W

heel

) 40

50

, 40

, 40

44

Ö-4

6Ö

43

0&

5ΪΟ

In

gots

In

got &

Blo

ck

75

x12

2x6

40

42

1t&

40

kg

2 0.

2

Bri

tann

ia R

efin

ed M

etal

s Lt

d.

No

rth

flee

t/U

K

Con

tinuo

us B

elt,

Cas

ting

Whe

el

269

(Bel

t); 2

5 (W

heel

) 40

4S

Ö

Ingo

t & B

lock

1t &

40

kg

0.2

LEAD-ZINC 2000

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 103

Appendix 7 -Primary Lead Smelter Survey - Pyrometallurgical Refineries

Co

mp

any

Nam

e

• R

efin

ery

- T

ype

An

nu

al P

rod

uct

ion

- L

ead

S

oft L

ead

Har

d Le

ad

Pb-

Ca-

Allo

ys

Pyr

om

etal

lurg

ical

Ref

iner

y

Dec

oppe

nzin

g

Met

hod

Ket

tle S

ize

Num

ber o

f Ket

tles

Tre

ated

Bul

lion

Inle

t Cu-

cont

ent

Tem

pera

ture

R

eact

ion

Age

nt's

) A

mou

nt

Fina

l Bul

lion

Fina

l Cu-

cont

ent

Dur

atio

n of

Tre

atm

ent

Am

ount

of D

ross

Tr

eatm

ent o

f Cu-

dros

s

Fina

l By-

Pro

duct

A

mou

nt

%P

b %

Cu

%S

Sof

teni

ng

Met

hod

MTP

Y

%

%

%

T

MTP

Y

%

■c

■

Kg/

T B

ullio

n K

g/T

Bul

lion

MTP

Y

%

HR

/Ket

tle

MTP

Y

MTP

Y

%

%

%

Nor

ddeu

tsch

e A

ffine

rle A

G

Ham

burg

/ G

erm

any

Por

tove

sme

srl /

Ita

ly

i py

rom

eta"

urg'

cal

Ref

i"in

g

15,0

00

50

50

Pre

cipi

tatio

n &

Sul

fur

Trea

tmen

t

200 1

20,0

00

0.08

-0.1

5 45

0 -

320

Sul

fur

1.5

19,0

00

<0.

01

8

ϊ,'0'Ö

Ö To

Sec

onda

ry C

oppe

r S

mel

ter

NA

1,00

0 80

5

-10

5

Oxy

gen

Trea

tmen

t

Pyr

oTet

aHur

gica

1 Refi ni

pg

125

000

98

06

1 4

Pb-

Βι a

lloy

Pre

cipi

tatio

n &

Sul

fur

Trea

tmen

t

180

4 13

5,00

0 1.

2 45

0 -

320

Sul

fur

3

126,

400

0.01

8

9,00

0

9,00

0 65

1

65

4.5

Oxy

gen/

Air

Trea

tmen

t

LEAD-ZINC 2000

Co

mp

any

Nam

e

KsK

s.S

ks,

„ ,„„

„ N

umbe

r of

Ket

tles

Trea

ted

Bul

lion

Inle

t Sb-

cont

ent

Tem

pera

ture

Sta

rt Te

mpe

ratu

re E

nd

Rea

ctio

n A

gent

(s)

Am

ount

Fina

l Bul

lion

Fina

l Sb-

cont

ent

Dur

atio

n of

Tre

atm

ent

Am

ount

of

Dro

ss/S

lag

Trea

tmen

t of S

b-D

ross

/Sla

g

Fina

l By-

Pro

duct

A

mou

nt

%P

b %

Sb

%N

a

Det

inni

ng

Met

hod

Ket

tle S

ize

Num

ber o

f Ket

tles

Tre

ated

Bul

lion

inie

t Sn-

cont

ent

Tem

pera

ture

Sta

rt Te

mpe

ratu

re E

nd

Rea

ctio

n A

gent

(s)

Am

ount

Fina

l Bul

lion

Fina

l Sn-

cont

ent

&ΐίΡ

.Ρ_ο

ίΣ(®

?*Π

ϋ®.!?

.Ι A

mou

nt o

f D

ross

/Sla

g Tr

eatm

ent o

f S

n-D

ross

/Sla

g

Fina

l By-

Pro

duct

A

mou

nt

%P

b %

Sn

%N

a

„,,„„

..„I,.

.„

MTP

Y

%

°c

°c

Kg/

T B

ullio

n K

g/T

Bul

lion

MTP

Y

%

HR

/Ket

tle

MTP

Y

MTP

Y

%

%

%

T

MTP

Y

%

"c

°c

Kg/

T B

ullio

n K

g/T

Bul

lion

MTP

Y

%

HR

/Ket

tle

MTP

Y

MTP

Y

%

%

%

Nor

ddeu

tsch

e A

ffln

erie

AG

H

ambu

rg /

Ger

man

y

jsa_

i,

1 [

18,0

00

1.0

-2.5

63

0 65

0 O

xyge

n 1

Por

tove

sme

sri /

Ital

y

358 2

126,

400

0.15

60

0 70

0 A

ir +

Oxy

gen

17,0

00

< 0.

05

16

-24

1,00

0 S

ale

125,

650

0.06

3

-4

1,00

0 R

otar

y Fu

rnac

e

Har

d Le

ad

1000

5

0-6

0 3

0-3

5

700

85

15

Not

App

licab

le

Oxy

gen

Trea

tmen

t i

200 1

I 18

,500

0

.1-0

.5

580

600

Oxy

gen

1 -

i 18

,000

<

0.03

16

50

0 -

Sal

e

PbO

-SnO

Dro

ss

500

70

6-8

i

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Co

mp

any

Nam

e

Des

ilver

izin

g

Met

hod

Ket

tle S

ize

Num

ber o

f Ket

tles

Am

ount

of S

teps

Tr

eate

d B

unio

n In

let A

g-co

nten

t Te

mpe

ratu

re S

tart

Tem

pera

ture

End

R

eact

ion

Age

nt(s

) A

mou

nt

Fina

l Bul

lion

Fina

l Ag-

cont

ent

Dur

atio

n of

Tre

atm

ent

Am

ount

of

Dro

ss/C

rust

Tr

eatm

ent o

f Ag-

Dro

ss/C

rust

Fina

l By-

Pro

duct

A

mou

nt

%P

b %

Ag.

,

%C

u %

Zn

Dez

lnci

ng, i

f ap

plic

able

M

etho

d

Ket

tle S

ize

Num

ber o

f Ket

tles

Trea

ted

Bul

lion

Inle

t Zn-

cont

ent

_, Te

mpe

ratu

re

Rea

ctio

n A

gent

's)

^ A

mou

nt

App

lied

Vac

uum

Fi

nal B

ullio

n Fi

nal Z

n-co

nten

t D

urat

ion

of T

reat

men

t A

mou

nt o

f Dro

ss/M

etai

%

Pb

%Z

n

T

ΜΤΡ

Ϋ

%

°C

°C

Kg/

T B

ullio

n K

g/T

Bul

lion

MTP

Y

%

HR

/Ket

tle

MTP

Y

MTP

Y

%

%

%

%

T

MTP

Y

%

°c

Kg/

T B

ullio

n m

bar

MTP

Y

%

HR

/Ket

tie

MTP

Y

%

%

Nor

ddeu

tsch

e A

ffin

erle

AG

P

orto

vesm

e S

ri /

Italy

H

ambu

rg /

Ger

man

y

»'■··»·

■»

f »

»

Par

kes-

Pro

cess

[

Tw

o S

tage

Par

kes

Pro

cess

200

220

1 2

Ϊ -

3 2

17,0

00

134,

250

1.0

-1.6

0.

15

450

440

320

32Ö

Z

inc

Zinc

1

8-2

2 13

16,0

00

125,

850

< 0.

0015

0.

0006

1

6-2

4 16

4,

000

10,0

00

Liqu

atio

n &

Cup

elat

ion

Liqu

atio

n

Ag-

rich

Bul

lion

\ Li

quat

ed A

g-A

lloy

300

850

20

15

70

22

<5

2 <

5 ί

60

Vac

uum

Dez

inci

ng

Vac

uum

Dez

indn

g

200

220

1 2

16,0

00

125,

850

0.3

- 0.

5 0.

52

500

- 53

0 60

0

0.04

bar

0.

001

16,0

00

125,

200

< Ö.

Ö5

Ö.Ö

2 7

8-1

0 70

65

0 <

10

5 >

90

95

LEAD-ZINC 2000

Co

mp

any

Nam

e

Met

hod

Ket

tle S

ize

Num

ber o

f Ket

tles

Am

ount

of S

teps

Tr

eate

d B

ullio

n In

let B

i-con

tent

Te

mpe

ratu

re S

tart

Tem

pera

ture

End

R

eact

ion

Age

ntjs

) A

mou

nt

Fina

l Bul

lion

Fina

l Bi-c

onte

nt

Dur

atio

n of

Tre

atm

ent

Am

ount

of

Dro

ss/C

rust

Tr

eatm

ent o

f B

i-Dro

ss/C

rust

Fina

l By-

Pro

duct

A

mou

nt

%P

b %

Bi

%C

a %

'lCig

Fina

l Ref

inin

g

Met

hod

Ket

tle S

ize

Num

ber

of K

ettle

s A

mou

nt o

f Ste

ps

Tre

ated

Bul

lion

Tem

pera

ture

Sta

rt Te

mpe

ratu

re E

nd

Rea

ctio

n A

gent

(s)

Am

ount

Fina

l Bul

lion

Dur

atio

n of

Tre

atm

ent

Am

ount

of D

ross

Tr

eatm

ent o

f Dro

ss

T

MTP

Y

%

Is.

°c

Kg/

T B

ullio

n K

g/T

Bul

lion

MTP

Y

%

HR

/Ket

tle

MTP

Y

MTP

Y

%

%

%

%

T

MTP

Y

°c

°c

Kg/

T B

ullio

n K

g/T

Bul

lion

MTP

Y

HR

/Ket

tle

MTP

Y

Nor

ddeu

tsch

e A

ffin

erie

AG

H

ambu

rg /

Ger

man

y

Kro

ll-B

ette

rton

200 1 1

16,0

00

0.3

-1.0

52

0 32

0 C

a-M

g-A

lloy

3-5

15,5

00

0.01

0-0.

015

16

500

Ref

inin

g an

d C

astin

g P

b/B

i-Allo

y

Pb/

Bi-A

lloy

1,60

0 94

6

Cau

stic

Sod

a /

Nite

r

200 1 1

15,5

00

400

450

Cau

stic

Sod

a / N

iter

Cau

stic

Sod

a 2.

0 N

iter

1.0

15,0

00

8 500

To S

econ

dary

Cop

per-

Sm

elte

r N

A

Por

tove

sme

srt /

Ital

y

Kro

ll-B

ette

rton

220 2 1

125,

200

0.12

48

0 32

5 C

alci

um &

Mag

nesi

um

Cal

cium

0.8

6 M

agne

sium

2 0

1 12

3,90

0 Ö

.Ö7

- Ö

.Ö25

10

4,

400

Ref

inin

g an

d C

astin

g P

b/B

i-Allo

y

Pb/

Bi-A

lloy

1,70

0 93

.5

6.5

Cau

stic

Sod

a /

Nite

r

220 1 1

123,

900

400

600

Cau

stic

Sod

a / N

iter

Cau

stic

Sod

a 0.

7 N

iter

0.3

122,

500

6 2,

200

Rec

ycle

d to

kiv

cet P

lant

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Co

mp

any

Nam

e [

Cas

ting

Mac

hine

- Ty

pe

Num

ber o

fMoü

jds

^ _R

ate

Cas

ting

Tem

pera

ture

C

ast P

rodu

ct

Ingo

t Siz

e _

4

Ingo

t Wei

ght

Rej

ects

as

% o

f New

Mat

eria

l

MTP

H

" °C

mm

Jfa

Nor

ddeu

tsch

e A

fftne

rie A

G

Ham

burg

/ G

erm

any

Stra

ight

Lin

e M

ould

-Cha

in

236

" 20

" "

450

Ingo

ts

70

0x1

15

x80

" 50

1

-2

Por

tove

sme

art /

Ital

y

Stra

ight

Lin

e M

ould

-Cha

in

400

50

380

Ingo

ts

57

0x1

05

x90

42"

o

00 LEAD-ZINC 2000

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 109

Appendix 8 -Primary Lead Smelter Survey - Betts Electrorefineries

Co

mp

an

y N

am

e

Re

fin

ery

- T

yp

e

An

nu

al

Pro

du

cti

on

- L

ea

d

Sof

t Le

ad

Har

d Le

ad

Pb-

Ca-

Allo

ys

Hy

dro

me

tall

urg

ica

l R

efi

ne

ry

Ave

rage

Pb

cont

ent

Num

ber

of T

anks

-

Dim

ensi

ons

- C

ontr

actio

n M

ater

ial

Num

ber

of A

node

s

Ele

ctro

lyte

Le

ad c

onte

nt

Tot

al H

2SiF

e

Fre

e H

2SiF

s

Oth

ers

Add

itive

s C

onsu

mpt

ion

Alo

es

Lign

in S

ulph

onat

e G

lue

Oth

ers

Tem

pera

ture

C

ircul

atio

n A

ppar

atus

ΜΤ

ΡΫ

%

%

%

%

m

g/i

g/i

g'i

mg/

l

9't

g't

9't

9't °c

Rec

ircul

atio

n R

ate

Aci

s Lo

ss

l/min

K

g/t

Cu

rren

t C

atho

de

Cel

l Vol

tage

C

urre

nt p

er G

ener

ator

Am

p/m

2

V

KW

Su

mit

om

o M

etal

Min

ing

Co

. H

arim

a W

ork

s

Bet

ts E

lect

rore

finin

g P

roce

ss

22,0

54

gg.g

gg

1.1

5x0

.92

x1.4

5 as

phal

t lin

ed r

einf

orce

d co

ncre

te t

ank

90

-10

0

110-

115

1,00

0

40

-43

1 8

m3 /m

in x

20

m H

ead

30

4 185

0.55

5,

000

amp

@ 5

0V

Ho

soku

ra S

mel

tin

g &

Ref

inin

g

Co

.

Bet

ts E

lect

rore

finin

g P

roce

ss

21,5

76

99 9

98

180

2.6

x0

.72

x1

.15

Rei

nfor

ced

Con

cret

e

23/c

ell

45

98

67

1,08

8

41

32

153

0.45

Kam

ioka

Min

ing

& S

mel

tin

g

Co

. T

oh

o Z

inc

Co

. C

hlg

iris

hlm

a P

lan

t

Bet

ts E

lect

rore

finin

g P

roce

ss

Bet

ts E

lect

rore

finin

g P

roce

ss

30,0

00

100

90,0

00

87

13

99.9

99

78

5.0

x1

.37

x1

.6

Con

cret

e lin

ed w

ith P

VC

Res

in

43/c

ell

120

160

70 -

600

1,00

0 - 40

50

1.5

127

0.6

1,20

0

99.9

99

364

3.0

x1

.0x

1.5

P

VC

with

Ste

el F

ram

e

28/c

ell

70

150

90 -

100

1 Λ03

0

- 40

40

4 142

0.46

2,

000

LEAD-ZINC 2000

Co

mp

an

y N

am

e

Cur

rent

Effi

cien

cy

Ene

rgy

Con

sum

ptio

n E

lect

roly

sis

Mec

hani

cal

An

od

es

Cas

ting

Met

hod

Com

posi

tion

%S

b %

As

%B

i %

Cu

%A

g %

Sn

Tot

al D

imen

sion

D

imen

sion

im

mer

sed

Sur

face

Mod

e of

Sus

pens

ion

Ano

de L

ife

Scr

ap

Ano

de S

paci

ng

Wei

ght

Cat

ho

des

S

tart

ing

She

et

Pro

duct

ion

Met

hod

Thi

ckne

ss

Wei

ght

Life

An

od

e S

lime

Com

posi

tion

%S

b %

As

%P

b %

Cu

%B

i %

'Äg

%

KW

h/T

Lea

d K

Wh/

T L

ead

%

%

%

%

%

%

mm

m

m

Day

s

%

mm

K

g

mm

K

g D

ays

%

%

%

%

%

%

Su

mit

om

o M

etal

Min

ing

Co

. H

arim

a W

ork

s

93

17

5-1

80

Cas

ting

Whe

el w

ith W

ater

Coo

ling

on T

op a

nd B

otto

m o

f M

ould

, 15

Mou

lds

0.5

0.05

0.

13

0.1

0.05

970

x 74

0 x

35

cast

lugs

8 26

.5

110

280 1 10

4

Ho

soku

ra S

mel

tin

g &

Ref

inin

g

Co

.

93

142

Wal

ker

850

x 57

0 x

30

8 36

.6

102

190

0.8 4 4

13.2

3.4

Kam

ioka

Min

ing

& S

mel

tin

g

Co

.

94.6

170

' H

oriz

onta

l Cas

ting

Mac

hine

with

H

oriz

onta

l M

ould

1 0.1

0.3

0.1

0.1

0.1

1,1

40

x9

85

x2

5

Sho

ulde

r T

ype

5-6

50

11

0 30

0

Ref

ined

Lea

d C

ast

0.7

12

-13

5-6

40

3 20

2 10

5

To

ho

Zin

c C

o.

Ch

igir

ish

ima

Pla

nt

97

140

18

Ver

tical

Aut

omat

ic C

astin

g M

achi

ne

1.2 0

0.25

0.

03

0.2 -

1,3

90

x80

0 1

,19

0x8

00

Sho

ulde

r T

ype

7 35

100

300

Ref

ined

Lea

d C

ast

0.9

14

7 50 - 12

1 10

7

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Co

mp

an

y N

am

e

Rem

oved

afte

r P

erce

ntag

e o

f Ano

des

Scr

ubbi

ng M

etho

d

Met

hod

of S

lime

Tre

atm

ent

Pro

duct

s

Cas

tin

g,

Mel

ting

Fur

nace

C

apac

ity

Add

itive

s A

mou

nt

Am

ount

of O

ross

T

reat

men

t of

Dro

ss

Cas

ting

Mac

hine

- T

ype

Num

ber

of

Mou

lds

Rat

e C

astin

g T

empe

ratu

re

Cas

t P

rodu

ct

Das

_ %

T

Kg/

T B

ullio

n M

TP

Y

MT

PH

°C

Ingo

t S

ize

^ m

m

Ingo

t W

eigh

t R

ejec

ts a

s %

of

New

Mat

eria

l

mm

m

m

%

Su

mit

om

o M

etal

Min

ing

Co

. H

arim

a W

ork

s !

Ho

soku

ra S

mel

tin

g &

Ref

inin

g j

Co

. K

amio

ka M

inin

g &

Sm

elti

ng

j

Co

. T

oh

o Z

inc

Co

. C

hig

iris

him

a P

lan

t

1.3

Γ_

__

__

Σ_

__

_Ι

5 I

5 7 2.8

Rot

atin

g br

ush

-

•

Ket

tle

80

Cau

stic

Sod

a 5

3 48

0

.

78

460

Ingo

ts

,

50

'_

Filt

ratio

n

•

Ket

tle

100

Cau

stic

Sod

a 0

3 25

8 R

ecyc

le t

o B

last

Fur

nace

Whe

el

80

16

430

Ingo

ts

„

50

[

Pyr

omet

allu

rgic

al R

efin

ing

Pb/

Sb-

Allo

y, S

ilver

, B

ism

uth

Ket

tle

140

Cau

stic

Sod

a 0

5 1,

700

Rec

ycle

to

Bla

st F

urna

ce

11

500

Ingo

t 8

30

x1

70

x1

20

50

21

LEAD-ZINC 2000

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 113

Appendix 9 -Primary Lead Smelter Survey - Betts Electrorefineries

Co

mp

an

y N

am

e

Ref

iner

y -

Ty

pe

An

nu

al

Pro

du

cti

on

·

'

Le

ad

S

oft L

ead_

H

ard

Lead

g

Pb-

Ca-

Allo

ys_

Hy

dro

me

tall

urg

ica

l R

efi

ne

ry

Ave

rage

Pb

cont

ent

I N

umbe

r of

Tan

ks

- D

imen

sion

s -

Con

duct

ion

Mat

eria

l

Num

ber

of A

node

s

Ele

ctro

lyte

Le

ad c

onte

nt

Tot

al H

2SiF

e

Fre

e H

2SiF

e

Oth

ers

Add

itive

s C

onsu

mpt

ion

Alo

es

Lign

in S

ulph

onat

e G

lue

Oth

ers

Tem

pera

ture

C

ircul

atio

n A

ppar

atus

- . - '

MT

PY

%

%

%

%

m

g"

9/1

mg/

l

3«

3«

g/t

g/t

°c

Rec

ircul

atio

n R

ate

Ad

s Lo

ss

l/min

Kflr

t

Cu

rren

t C

atho

de

Cel

l Vol

tage

C

urre

nt p

er G

ener

ator

Am

p/m

2

V

KW

Ko

rea

Zin

c C

o.,

On

san

Bet

ts E

lect

rore

finin

g

142,

343

84

4 12

99 9

Ltd.

Pro

cess

542

2 98

x 0

98

x 2

98

PV

C w

ith S

teel

Fra

me

28/c

ell

90-1

10

65-7

5

500

1,00

0

35-3

8

40-4

5 3.

8

14

5-1

70

0.5

-0.7

Co

mln

co L

td.

Tra

il O

per

atio

ns

Bet

ts E

lect

rore

finin

g P

roce

ss

120.

000

99 9

95

768

,2 7

5 x

0 81

x 1

20

FR

P

24/c

ell

17

0-1

80

85

-95

210

275

38

-42

Pum

p ta

nk,

Wilf

ley

pum

ps &

se

ctio

n he

ader

s.

Cas

cade

flo

w

20

-25

3

17

8-1

92

0.45

- 0

.5

5,00

0

LEAD-ZINC 2000

Co

mp

an

y N

am

e

Cur

rent

Effi

cien

cy

Ene

rgy

Con

sum

ptio

n E

lect

roly

sis

Mec

hani

cal

An

od

es

Cas

ting

Met

hod

Com

posi

tion

%S

b %

As

%B

i %

Cu

%A

fl

%S

n

Tot

al D

imen

sion

D

imen

sion

Im

mer

sed

Sur

face

Mod

e of

Sus

pens

ion

Ano

de L

ife

Scr

ap

Ano

de S

paci

ng

Wei

ght

Cat

ho

des

S

tart

ing

She

et

Pro

duct

ion

Met

hod

Thi

ckne

ss

Wei

ght

Life

An

od

e S

lime

Com

posi

tion

%S

b %

As

%P

b %

Cu

%B

i %

Ag

%

KW

h/T

Le

ad

KW

h/T

Lea

d

%

%

%

%

%

%

mm

m

m

Day

s

%

mm

.«a

.

mm

K

g D

ays

%

%

%

%

%

%

Ko

rea

Zin

c C

o., L

td.

On

san

94

175

20

Ver

tical

Cas

ting

with

Wat

er

Coo

ling

0.8

0.02

0.

25

< Ö

.Ö5

0.2 -

1,2

50

x8

00

x2

5 Ϊ 3

70x

800

Cas

t iu

gs

6 38

100

270

Ref

ined

Lea

d (2

Θ)

Rot

ary

drum

cas

ting

mac

hine

0.8 13

6 35

0.2

T

8 3 20

ΪΟ-Ϊ

2

Co

mln

co L

td.

Tra

il O

per

atio

ns

92

Ϊ20

Hor

izon

tal w

/wat

er c

oolin

g on

bo

ttom

1.3

0.26

0.

4 0.

004

0.3 -

880

x 67

0 x

30

876

x 67

0 x

30

Hun

g fr

om c

ast

in e

ars

6 40

100

200

Ref

ined

lea

d R

oll c

ast

embo

ssed

0.5

8-9

6 40

7.

8 7.

8 0.

16

14

15.5

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Co

mp

an

y N

am

e

Rem

oved

afte

r D

ays

Per

cent

age

of A

node

s %

S

crub

bing

, Met

hod

Met

hod

of S

lime

Tre

atm

ent

Pro

duct

s

Cas

tin

g

Mel

ting

Fur

nace

C

apac

ity

, T

A

dditi

ves

Am

ount

,K

g/T

Bul

lion

Am

ount

of

Dro

ss

_ M

TP

Y

Tre

atm

ent

of D

ross

Cas

ting.

Mac

hine

- T

ype

Num

ber

of M

ould

s ~

Rat

e ]

MT

PH

C

astin

g T

empe

ratu

re

°c

Cas

t P

rodu

ct

Ingo

t S

ize

A

mm

m

m

mm

In

got

Wei

ght

kg

Rei

ects

as

% o

f N

ew M

ater

ial

%

Ko

rea

Zin

c C

o.,

Ltd

. O

nsa

n

6 1.2

Rot

atin

g br

ush

with

wat

er s

pray

Co

min

co L

td.

Tra

il O

per

atio

ns

6 40

Scr

ap w

ash/

brus

h

Pyr

omet

ailu

rgic

ai &

'

Filt

ratio

n E

lect

rore

finin

g

Ag-

, Au

-, B

i-mgo

ts, S

b/P

b-al

loy

_ C

ake

4 M

eltin

g F

urna

ces

Ket

tle

120

' 20

0 C

aust

ic S

oda

Cau

stic

sod

a 0

3 ~

1 4,

775

* S

hort

Rot

ary

Fur

nace

,

Rec

ycle

to S

mel

ter

Hon

zont

al

Whe

el

252

* 15

0 (p

igs)

/16

(jum

bos)

16

'

16

-30

430-

450

" 4

10

-46

0 In

got

Pig

s &

jum

bos

84 x

555

x 6

0 [

90 x

460

x 8

5 (2

5kg)

, 1

10

x5

55

x8

5 (4

5 kg

) *

590

x 59

5 x

300

(910

kg)

25

*

25 &

45

& 9

10

0 45

ve

ry s

mal

l

LEAD-ZINC 2000

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 117

OPERATIONS AT THE DOE RUN COMPANY'S HERCULANEUM PRIMARY LEAD SMELTER

N. D. Schupp The Doe Run Company

881 Main Street Herculaneum, Missouri, U.S.A. 63048

ABSTRACT

The Doe Run Company's Herculaneum, Missouri lead smelter has existed in continuous operation since 1892. Throughout the history of the plant extensive developments have been made to improve throughput, increase productivity, decrease operational costs and improve technology, while remaining an environmentally responsible producer of high quality primary lead in the international market. This paper provides an overview of the Herculaneum smelter, focusing on the most recent configuration of the operations. The paper includes discussion on all the aspects of primary lead production at the Herculaneum smelter.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

118 LEAD-ZINC 2000

INTRODUCTION

Lead mining and smelting has been practiced in Missouri since the early 1700's. Initially, the mining and smelting occurred at the shallow deposits of the old lead belt in southeastern Missouri. In 1892 the Herculaneum smelter was built at the present site. Since 1892, the smelter has been continuously improving its processes while working with customers to build strong marketplace-driven products. Technology at the site has changed, from calcine furnaces, to Salzburg converters, to downdraft sinter machines, and finally to a modern updraft sintering machine. Furnace technology has followed, from cupolas, to 36 by 100 inch blast furnaces, to the latest 66 by 336 inch blast furnaces. Matte converters were replaced with reverberatory furnaces, which have been idled with the advent of dry dross production. Refining has moved from simple metal cleaning, to modern methods producing 99.99%+ Pb, to the current practice of producing specialized alloys meeting a large variety of customer specifications. Likewise, environmental controls at the smelter have evolved from simple baghouses installed as early as 1910, to wet scrubbers and modern baghouses by the 1950's. Our present system incorporates an electrostatic precipitator, modern baghouses with membrane bags, and an acid plant to treat the sulfur gases produced in the process.

The smelter was expanded in the 1960's to a capacity of approximately 250,000 short tons per year. Since the 1960's the smelter has seen continuous improvements in the efficiencies of the operations, safety and hygiene, environment, and quality.

RAW MATERIALS HANDLING

The Herculaneum smelter is one of the largest primary lead smelters in the world. Historically, the smelter was supplied with lead concentrates from lead mines in the lead mining districts of southern Missouri. Lately the smelter has also been processing lead concentrates from around the world.

Doe Run receives lead concentrates by railroad or truck. Once the concentrate arrives at the smelter it is managed with very specific quality procedures. The concentrate is either sent directly into holding bins or is maintained on a controlled stockpile. The smelter utilizes a rota-side railroad car tippler that can handle up to 150 tons. Once dumped, the concentrate is transferred by belt, to bins in the sintering plant mix room.

Fluxing materials are handled in a fashion similar to concentrates. The limestone rock, sand, and other fluxing agents are generally delivered to the plant by truck. Metallurgical coke is generally delivered to the plant in bottom-dump railcars. The coke is unloaded into storage bins via a railroad trestle or it can be stockpiled. Secondary materials are also processed at the smelter. These are generally materials that contain lead or have value as a flux.

SINTERING PLANT

The lead concentrates and fluxes are transported to the mixing room bins, as previously discussed. Under each mixing room bin, an apron feeder with a pre-set gate continually withdraws the material and discharges it onto a conveyor belt. Scale systems are used to

LEAD-ZINC 2000

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 119

maintain the proper mix of new feed material (lead concentrate and fluxes), recycled materials (in-house fume, water treatment filter cake, etc.), and blast furnace slag. Along the way sinter returns are added to the feed at the desired ratio. This combined mixed feed is tumbled in a mixing drum and discharged to shuttle belts, which fill the ignition and main layer feed hoppers of the sintering machine.

The up-draft designed sintering machine is 10 feet by 100 feet. The ignition layer of feed (approximately 10% of the total feed) is charged onto the machine pallets and is leveled by a plate to the desired depth. This layer of feed passes through a downdraft ignition area where natural gas burners ignite the layer. Next, the main feed layer is added to give the total height of the sinter bed of approximately 12 to 14 inches. The material then enters the updraft portion of the machine. After burn-through, the sinter is cooled before discharging off the machine.

The sinter is discharged as a large mass and falls into a claw breaker. From the breaker, the sinter is crushed and sized through rolls. A portion of the sinter is crushed to about one-half inch and stored for future use as return sinter in the mixed feed. The finish sinter is transported to storage bins for smelting in the blast furnaces.

Typically, except for planned maintenance, the sintering and associated processes operate 24 hours per day, 7 days per week, and 11 Vi months per year. Planned maintenance is generally one eight-hour period each week. A typical analysis for the sinter is given in Table I.

Table I - Typical Sinter Analysis Element Content (wt%) Lead 48 Sulfur <2 Copper 1 Silica 10 Iron (oxide) 15 Lime 6 Magnesium (oxide) 2 Zinc 5

BLAST FURNACE

The Herculaneum smelter has three blast furnaces, of which two are operating at any one time. The furnace crucible is built of castable refractory. The furnace sides are built of water-cooled steel jackets. Each furnace has 21 tuyeres on each side, for 42 tuyeres total. Air is supplied to a furnace by one of three blowers capable of supplying 15,000 ft3/min of air. The blast air is directed to the furnace by a series of bustle lines. The air blast is enriched with oxygen that is controlled at each furnace.

Finish sinter and coke are stored in the blast furnace trestle bins. These bins are loaded with bottom dump rail cars. From the trestle bins, the sinter and coke are transported by conveyor belt to the blast furnace charge bins. The coke (typically 8 to 10% of total charge) and sinter are discharged with vibratory feeders from the charge bins, and are weighed on scale belts. From the scale belts, the sinter and coke are transported by a series of conveyor belts to a

120 LEAD-ZINC 2000

traveling tripper which discharges to the shuttle belts supplying each side of each furnace. An operator controls the feeding operation from the feed floor.

The blast furnaces continually discharge lead bullion and slag through a slag tapper. The lead bullion and slag falls into a gas-fired, refractory-lined settler. The settler is constructed to allow a separation of slag and lead bullion. The slag overflows the settler and travels down a launder to the granulation system. The lead bullion, being denser than slag, underflows into a lead well, then overruns another weir into refractory-lined lead pots. The lead pots sit on a carriage, which travels on a Y-shaped track arrangement. When a pot is full, the carriage is pulled back onto an "arm" of the track and the empty pot is then pushed forward immediately to receive lead. After cooling, the full pot is transported by overhead crane to the dross plant. The slag is granulated with a closed-loop water system. The granulated slag is slurried to one of two dewatering bins where the slag settles from the slurry. The water is cooled and recycled to the granulation water tank. The majority of the slag is returned to the sintering process and the balance of the slag is stockpiled.

A blast furnace can typically treat up to 1000 tons of sinter per day, producing approximately 450 tons of lead bullion, and 350 tons of slag. The blast furnaces operate 24 hours per day, 7 days per week, 11 'Λ months per year and maintenance is accomplished on an as needed basis. A typical analysis of the slag is contained in Table II.

Table Π - Typical Slag Analysis Element Content (wt%) Lead 2 Sulfur 2 Copper <1 Silica 22 Iron (oxide) 33 Lime 12 Magnesia 4 Zinc 10

DROSS PLANT

The drossing operations at the Herculaneum smelter are conducted entirely in kettles. The dross plant uses four 250-ton capacity kettles. The lead bullion is received from the furnaces, and is allowed to cool. This cooling causes the copper, sulfur, and trace amounts of other impurities to precipitate and rise to the surface of the bullion, producing a copper-rich dross. This dross is skimmed off and put through a wet screw classifier and sold as a by-product to copper smelters. The drossed bullion is transferred to another kettle where it is further cooled and fluxed with sulfur for final decopperizing. This dross is skimmed and recycled to the blast furnace. The decopperized lead bullion is then pumped to the refinery. Maintenance in the dross plant is minimal and is managed on an as needed basis.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 121

REFINERY

The Herculaneum refinery consists of eleven 250-ton capacity kettles. Decopperized lead is received from the dross plant into one of the first two kettles. The refinery uses a modified Parkes process to clean the lead bullion of contained silver and the remaining copper. The third through sixth kettles are used to perform the desilverizing process. Zinc metal is added to the lead bullion in preset amounts to obtain the high purity level of Doe Run quality lead. Throughout the entire batch process, lead samples are taken to analyze the impurity levels and to maintain the stringent quality control procedures. During the desilverizing process, silver-rich drosses are produced and sold as a by-product to silver refineries.

Following desilverizing, the lead is bailed to kettle seven or eight to remove the remaining amount of zinc. The lead is heated over 1000°F, then vacuum dezinced using a flattop dezincer. The zinc recovered in the vacuum dezincing is recycled in the desilverizing process. After dezincing, the lead is bailed to the final three kettles (nine, ten or eleven). At this point, the lead has very minor levels of impurities. The final stage in refining is accomplished by stirring in sodium hydroxide, sodium nitrate and charcoal.

The refined lead is sampled for quality assurance. The lead can be cast as Doe Run quality corroding grade or one of a wide spectrum of specialized alloys. Following alloying, the lead is sampled again to assure that it is within customer specifications. If adjustments are needed, they are made at this point. The lead is then cast into one-ton ingots, half-ton ingots, sixty-pound pigs, or one hundred-pound pigs. The sixty-pound and one hundred-pound pigs are cast on straight-line casting machines. The half- and one-ton ingots are cast on a 26-foot diameter bench style-casting wheel.

The refinery can process and cast three to five 250-ton kettles of lead per day. A typical analysis of the refined, but unalloyed (corroding grade), lead is given in Table III. The alloys produced in the refinery are typical of the alloys required in modern battery production and a wide variety of specialty lead products.

Table III - Typical Refined Lead Analysis Element Content (ppm) Lead >99.99 wt% Silver <10 Bismuth <30 Copper <5 Zinc <5 Calcium <5 Tin <5 Aluminum <5 Antimony <5 Arsenic <5 Cadmium <2 Nickel <2

122 LEAD-ZINC 2000

SHIPPING DOCK

Once the lead is cast, it is transported to the shipping dock. The pigs are stacked in approximately 1-ton stacks on the casting machine, and then the stacks are weighed, banded with four steel bands, and arranged into truck or rail lots for shipment. The ingots are also weighed and arranged into truck or rail lots for shipment. Before shipment, the dock personnel are responsible for the final inspection and marking according to customer specifications.

STRIP MILL

The Herculaneum smelter also produces precision wrought lead strip. Lead is melted and alloyed in one of two 75-ton kettles. From the kettle, the lead is gravity fed to a feed box and then flows through a slotted opening in the bottom of the box. The slotted area of the box is cooled with water sprays and the lead freezes as it is continuously pulled through the slot by a series of drive rolls. The lead, now in a slab-form, is pushed onto a 110-foot roller conveyor. When the slab reaches the desired length, it is cut and advanced to a two-high reversible rolling mill.

The rolling mill is arranged between two payoff reels that coil or uncoil the lead as it is rolled. A typical pass schedule for the mill requires up to seven passes to produce the desired thickness. The reduction in thickness for any pass is generally held between 40 and 60 %, excluding the final pass, which is generally 10 to 30 %. The mill rolls are cooled and lubricated with an emulsion of mineral oil and water. Final thickness of the finished lead strip is measured by using an electronic micrometer, which passes across the width of the strip. During the rolling operation, the strip is continually observed for defects from rolling or casting.

After rolling, the lead is slit to the final width desired by the customer. The slitter can produce material from approximately 2 inches wide through 26 inches wide. As described previously, the slabs are cast at varying widths; the width is selected to minimize the edge waste from the slitter. The coils of strip are transferred to a turnstile, then to an upender where they are palletized. The palletized coils are then packaged, weighed, and grouped into truck lots for shipment. The strip product is generally used in battery applications, as well as fabricated specialty applications (radiation shielding, construction flashing, chemical resistant linings, etc.). Maintenance is accomplished on a scheduled basis and as required,

ACID PLANT

The Herculaneum smelter operates a 340-ton per day Monsanto type single contact acid plant on the site. This plant treats the sulfur-rich gases produced in the sintering process. The 93% black sulfuric acid produced is stored in tanks on site. Sulfuric acid is shipped to customers from the smelter by truck, rail or river barge. The acid plant operates and performs maintenance on a schedule similar to the sinter plant.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 123

WATER TREATMENT PLANT

The water treatment plant at the Herculaneum smelter treats all water to be discharged from the plant. The water from the plant is collected in two large storage tanks. The water used within the plant is recycled from one of these tanks. The excess water is pumped to the treatment plant. In the treatment plant, the water is initially pumped to a thickener where the solids settle out. Next, the overflow water is treated with lime to control the pH, and a polymer flocculant is added. The water is then pumped to a reactor/clarifier where the heavy metals precipitate, then settle out of suspension. The underflow sludge is pumped back to the thickener. The overflow is treated to remove any trace heavy metals. The water is then filtered before discharge. The discharge water is tested regularly to ensure that it meets all standards for water quality. The sludge collected from the thickener is filtered, the solids are recycled through the sintering process, and the liquid returns to the water treatment plant.

BAGHOUSES

The Herculaneum smelter has two primary baghouses. The #3 baghouse cleans the updraft air from the cooling portion of the sintering plant process. It has 12 cells containing 384 bags each. The #3 baghouse is served by two 650 HP blowers, and cleans approximately 300,000 ft3/min of air. The #5 baghouse treats the gases from the blast furnace and dross plant operations. It has 17 cells containing 384 bags each. The #5 baghouse is served by two 1500 HP blowers, and cleans approximately 500,000 ft3/min of air. There are also a number of smaller baghouses to treat the emissions from various processes throughout the plant. All told, these smaller baghouses treat approximately 200,000 ft3/min of air. From the baghouses, the cleaned air is transferred to a 550-foot stack. All the dust collected at these baghouses is recycled through the sintering plant.

PRODUCTION LABORATORY

The production laboratory at the Herculaneum smelter performs quality analytical work for all the processes. The laboratory equipment includes a spark emission spectrophotometer, inductively coupled plasma atomic emission spectrometer, wavelength dispersive X-ray fluorescence spectrometer, sulfur analysis equipment, and wet chemistry equipment. The laboratory conducts the analysis for all process materials, including sinter, blast furnace slag, lead bullion, all finished lead products, and environmental samples.

MAINTENANCE, ENGINEERING AND SERVICES

The Herculaneum smelter has a maintenance department that handles electrical and mechanical support activities. However, contractors are called upon for specialized electrical, mechanical and heavy equipment maintenance. An engineering department manages all the equipment design requirements of the smelter. The service department manages all materials handling outside of the process areas. Materials are moved through the plant by rail and truck.

124 LEAD-ZINC 2000

SMELTER WORKFORCE AND SUPPORT STAFF

The smelter workforce has a team-oriented, higher performance organization culture. Performance-based incentives are in place for the employees. These incentives are based on performance areas that the workforce can control, such as environmental performance, safety and hygiene, cost control, production optimization and process improvement. The smelter philosophy is to "make tomorrow better than today".

Total smelter employment is approximately 280 hourly personnel among the various departments, including maintenance. The smelter's supervision and support staff includes accounting, payroll, purchasing, environmental technicians, safety and hygiene, medical, human resources, information systems, technical services, engineering, and management. In total, there are approximately 90 support staff personnel.

SAFETY AND HYGIENE

The Doe Run Company strives to be one of the safest employers in the world. As such, comprehensive programs focusing on safety and hygiene are in place for the benefit of the employees. The Doe Run safety philosophy is continuously taught and practiced at the Herculaneum smelter. As a result, the frequency and severity of the accidents in the plant have continually decreased in recent years.

Hygiene at the Herculaneum smelter is also of primary importance to the Doe Run Company's operating philosophy. The employees are fitted and trained in the use of various respirator styles according to their exposure levels, personal exposure history, hygiene habits, and personal preference. The employees are provided a modern change house and cafeteria, and are encouraged to be proactive with their hygiene habits. The results of the hygiene program at Herculaneum have shown that our workforce's hygiene performance is one of the best in the world, as shown in Figure 1.

T>

rn 3

m <1> _1 TJ o 0

m (1) σ> ra (1)

4

70

60 SO

40

3U

20

10

0

1975

♦ ♦ ♦ ♦ * * ♦

♦ ♦ * ♦ ♦ ♦ * ♦ ♦ * *

1980 1985 1990 1995 2000

Figure 1 - Average Blood Lead versus Year

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 125

ENVIRONMENT

The smelter has undergone extensive changes to minimize its environmental impact. Most notably, the smelter has taken steps to reduce fugitive emissions while maintaining worker hygiene standards. Some results of this program are the elimination of wet scrubbers from the plant, enclosure of the sintering plant building, and increased baghouse capacity through the use of Teflon membrane bags. Plans for future reductions of fugitive emissions are being developed and implemented continuously. The results of the smelter environmental projects have shown a reduction in the average community exposure levels, and are continuing to improve.

CONCLUSIONS

The Doe Run Company's Herculaneum smelter has existed for over a century as a leading producer of high quality primary lead products in the international market. Our 1997 quality certification as an ISO 9002 registered producer has further secured our position in this market for many years to come. The smelter personnel have always viewed change as an opportunity to increase safety, minimize environmental impact, and improve product quality and customer service, while optimizing operations and lowering costs. Only through continuous change will the Herculaneum smelter continue to be a dominant player in the world's primary lead market.

ACKNOWLEDGEMENTS

The author would like to thank The Doe Run Company for the opportunity to publish and present this paper and the Herculaneum staff who provided input.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 127

MODERN LEAD SMELTING AT THE QSL-PLANT BERZELIUS METALL IN STOLBERG, GERMANY

R. Piillenberg and A. Rohkohl Berzelius Metall GmbH

Binsseldhammer 14 D-52224 Stolberg, Germany

ABSTRACT

The QSL plant at "BERZELIUS" Stolberg GmbH in Stolberg, Germany was commissioned in 1990, introducing a new era of lead smelting in the more than 150-year-old history of the lead smelter. The lead smelter in Stolberg has changed its appearance several times during its existence. Therefore, at many locations on the site, old and obsolete buildings and equipment parts can still be seen. They are often located side by side with new modern constructions like the QSL-plant and the mercury removal plant, and reflect the specific contrast between tradition and modern technology. After a brief description of the development of the smelter and its decision to implement the current novel plant, the most recent operating results are reported. Following some modifications to the process and equipment, which had become necessary along with the implementation of the new technology, the plant is in operation today at approximately 30% above its design capacity. These achieved production levels are not restricted by the QSL-reactor but by precedent and subsequent available equipment. The QSL-plant entirely complies with the demands of modern, energy-saving, ecologically compatible lead production technology. Stringent environmental standards are maintained on a continuing basis. The demonstrated viability and flexibility of the overall process permits the operation to adapt quickly to the prevailing economical situation worldwide.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

128 LEAD-ZINC 2000

INTRODUCTION

The lead smelter Binsfeldhammer of "BERZELIUS" Stolberg GmbH was founded by interested French parties in 1848. From the beginning, blast furnace technology was applied and the lead bullion was refined in a pyrometallurgical refinery. The annual capacity of the plant was nominally between 3,000 and 6,000 t of lead until 1892. At this point in time the plant had approximately 180 employees (1). In 1892 the existing blast furnace was redesigned and the plant capacity increased to 15,000 t/y of lead. With the installation of the first electrostatic precipitator in Europe for cleaning blast furnace off-gases in the year 1905, the Stolberg smelter dedicated itself, almost from the beginning, to the use of the latest state-of-the-art technology for environmental protection (1). At this time the smelter produced about 18,000 t of lead annually. With the closing of mining activities in the Stolberg region, the smelter was converted to a "tolling" smelter in 1920. During the following years, however, the capacity of the smelter was increased to more than 30,000 t/y lead by expanding the refinery and the silver removal facility, the erection and commissioning of reverberatory furnaces and a sulfuric acid plant and other technical modifications.

Table I - Technical Development at the Lead Smelter "BERZELIUS" Stolberg GmbH 1854 Construction of a 100m high stack 1892 Modernisation of Blast Furnace 1903 Construction of New Sinter Plant 1905 Implementation of first Electrostatic Precipitator 1920 Commissioning of Rotary Furnaces 1956 Construction of Up-draught Sinter Machine 1960 Construction of Sulphuric Acid Plant 1964 Commissioning of Battery Scrap Breaking Facility 1967 Start of Construction of Bagfilter Facilities 1979 Installation of Two New Rotary Furnaces 1990 Commissioning of the QSL-Plant

In 1956 the first up-draft sintering machine in the Northern Hemisphere for roasting lead concentrates went into operation in Stolberg. This increased the annual capacity of lead to about 54,000 tonnes. Moreover, an almost complete replacement of the blast furnace took place in 1960 in conjunction with the installation of a new sulfuric acid plant. A plant applying a new patented process for separating lead storage batteries was erected in 1964. This plant, however, was shut down in the year 1984, after a new battery breaking and separating system based on the latest technology was installed and started up at the affiliated company in Braubach, Germany. Between 1955 and 1980, the pyrometallurgical refinery was also gradually replaced, expanded and modernized in order to accommodate the increased production of lead from the smelter. The annual output of refined lead at the beginning of the eighties was between 80,000 and 90,000 tonnes.

The most fundamental modification to the lead smelter, however, was the construction and start up of the QSL-converter in 1990, a novel technology for the recovery of lead. Integrated into this project was the erection of a new double absorption sulfuric acid plant, a facility for leaching flue dust for cadmium removal and a turbine for the generation of power utilizing the energy from the process off-gas (2). With more than 150 million DM capital expenditures, this project was the most expensive but also the most progressive investment. The technology was introduced in order to meet the demands of modern energy-saving and

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 129

ecologically compatible lead production. Figures 1 and 2 show an overview of the present layout of the plant.

Figure 1 - Overview of the Lead Smelter "BERZELIUS" Stolberg, Germany

Figure 2 - The QSL Plant

130 LEAD-ZINC 2000

In the course of an extensive restructuring program Metallgesellschaft AG sold their three German lead smelters ("BERZELIUS" Stolberg GmbH, BSB Recycling GmbH and Muldenhütten Recycling und Umwelttechnik) in 1995 to Eco-Bat Technologies PLC, a subsidiary of Quexco. The new owner continued the modernization program to eliminate the still existing problems and to convert it into an economically operating plant. A new mercury removal plant was implemented in 1997 to significantly improve the quality of the generated sulfuric acid (2). A new effluent treatment plant was built in 1998. Until 1999 the logistic of raw material transportation was gradually optimized and an up-to-date crane was installed in the raw material storage area. Moreover, a larger casting machine was installed in the refinery, increasing the production output of the plant. This year, the construction of a new wet gas electrostatic precipitator will follow.

All these technical innovations resulted in a decisive improvement of plant capacity and efficiency and are the foundations of an economically ensured operation. The current capacity of the plant is more than 100,000 t of lead and lead alloys annually. The Berzelius Stolberg GmbH thus possesses the most modern technology for producing primary lead worldwide and will continue its tradition of leadership in concentrate processing even in the future.

MOTIVES FOR THE IMPLEMENTATION OF THE QSL-TECHNOLOGY

For more than one hundred years, lead has been extracted from primary sources applying a two-step process (sinter machine and blast furnace) which, although having been constantly improved, has essentially remained unchanged. The smelter in Stolberg, like more than 95% of all primary smelters in the beginning of the eighties, used that standard technology. However, its side effects have been, as in most of the other cases, emissions of heavy metals and sulfur dioxide containing gases to the environment (3). Despite all considerable improvements to the standard process, the absolute demands of compliance with the most stringent pollution control regulations, while achieving high plant efficiencies at reasonable capital and operating costs, made it necessary to replace the classical process by a modern technology. After having developed the QSL-technology to a semi-industrial scale, it was concluded that the investment cost for a "greenfield" lead smelter based on the QSL-process with a capacity of 100,000 t/y lead are only 60 - 70 % of those of a conventional smelter. It was also shown that the operating costs are approximately 40% less compared to the conventional process. In addition, the QSL-technology has the following advantages compared to the conventional sinter machine/blast furnace operation:

• Production of lead bullion low in sulfur as well as a disposable low-lead slag in one continuous operating converter at a lead recovery rate higher than 98%.

• High specific smelting rates by means of oxygen application resulting in a more compact design of the process.

• Capability to timely and reliably monitor and control the operating parameter of the entire smelter complex for optimized plant operation.

• Efficient power utilization by recovering energy from the waste gas.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 131

• The application of oxygen technology significantly reduces the process gas volume resulting in SC>2-contents up to 15% in the off-gas depending on the treated types of concentrates. The gas, however, is well suited to be treated in a subsequent sulfuric acid plant for effective sulfur removal.

• The handling of raw materials and additives to be charged is much simplified. Drying of raw materials is not required, only moist and finely grained materials, and no hot dusty materials, are processed. Breaking and screening of material is not required.

• Instead of expensive metallurgical coke or coal, a broad range of lower-cost bituminous coals and, optionally, additional minor amounts of low-cost petroleum coke is used. High sulfur containing coals can be used because of the subsequent SO2 removal in the sulfuric acid plant.

• Emissions of SO2 and lead are reduced by 95% and 75%, respectively, and the imposed "TA-Luft" legislation (technical instructions on air quality control) is met on a permanent basis.

• Since handling of dry and dusty materials is minimized and carried out only in encapsulated and vented conveyors, the hygienic conditions in the plant have been significantly improved and regulations regarding the maximum permitted concentration of materials in the air at the working place are attained.

When the German government imposed the new "TA-Luft" legislation in 1986, Metallgesellschaft AG decided to replace the existing conventional technology with the QSL-process in order to meet the revised environmental legislation (4). The QSL-plant at "BERZELIUS" Stolberg GmbH was commissioned in 1990, and was followed by some modifications to the process and equipment which had become necessary along with the implementation of this new technology (5). Today the plant is operating above its nominal design capacity of 75,000 t/y and it was demonstrated that the anticipated quantum leap was successfully achieved with the QSL-technology for lead recovery.

DESIGN OF THE QSL-PLANT IN STOLBERG

The QSL-technology is a continuous operating process. Lead-bearing materials are treated with oxygen in a single smelting unit consisting of a kiln-like converter. The final prod-ucts of the overall process are lead bullion and a disposable slag. Raw materials such as metal sulfide concentrates and secondary materials, fluxes, recirculated flue dust and, if required, solid fuel are thoroughly blended and agglomerated. The agglomerated feed mixture is then charged homogeneously with minimized dust evolution through feed ports into the converter. It falls into a dispersed molten mixture of slag and lead. The handling facilities are shown in Fig-ures 3 and 4.

132 LEAD-ZINC 2000

Figure 3 - Raw Material Storage Figure 4 - Charging System to the QSL Reactor

As shown in Figure 5, the QSL converter consists of a slightly sloped, 33m long, up to 3.5m in diameter refractory-lined cylinder, which can be tilted by 90° about its longitudinal axis when the process is interrupted. The entire converting process is incorporated into two compartments, a smelting zone and a slag reduction zone in the single bath-smelting unit, and separated only by a weir. In the first section of the converter, the charged material is smelted by blowing tonnage oxygen into the bath through bottom blowing, gas-cooled injectors. The initiated autogenous roast-reaction smelting leads to the formation of low-sulfur primary lead bullion, a slag with a lead content of 30 - 40% and a sulfur dioxide rich off-gas.

The lead oxide-rich slag is passed through an opening at the bottom of the weir to the slag reduction zone. In this section, pulverized coal is introduced with a carrier medium in parallel with supplementary oxygen through a series of bottom blowing injectors in order to reduce the lead oxide to metallic lead. The injectors for the carbothermic reduction are arranged to provide a continuous mixer settler regime over the entire length of the zone. This configuration allows a progressive decrease of the bath oxygen activity towards the slag discharge end and achieves the desired compositional and temperature gradients by the possible individual adjustment of heat and mass transfer in a series of regulated bubble plumes (6). The metallic lead settles to the bottom of the zone and flows countercurrently to the slag back into the smelting zone. It combines with the primarily produced lead bullion and is discharged through a siphon system. The low-lead final slag is tapped at the end of the slag reduction zone and is granulated, as indicated in Figures 6 and 7.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 133

Figure 5 - Schematic Drawing of the QSL Converter at "BERZELIUS" Stolberg

Required postcombustion, mainly of the gases in the slag reduction zone is accom-plished with oxygen-enriched air, which is injected into the gas atmosphere via lances.

Figure 6 - Lead Syphon with Decopperizing Figure 7 - Slag Tap with Granulation

The SCVrich off-gases leave the smelting zone of the converter at a temperature of approximately 1100° - 1150° C. The gases are first cooled in a vertical pass of a waste heat boiler system to a temperature below 700° C before passing a convective heat recovery section. Subsequently, the gases are conveyed to a hot gas electrostatic precipitator for dedusting. The collected flue dust is recirculated back to the converter with the exception of a bleed to extract cadmium. The gas cleaning operations are shown in Figures 8 and 9.

134 LEAD-ZINC 2000

Figure 8 - Hot Gas Electrostatic Precipitator Figure 9 - Evaporation Cooler

Exiting the electrostatic precipitator the off-gas still has a temperature of 320° - 350°C and a dust content of less than 100 mg/Nm3. In an evaporation cooler the temperature of the gas is lowered to 70°C. It then passes a two-step wet gas electrostatic precipitator unit where the gas is cooled to 30° - 35°C as well as dedusted to levels below 1 mg/Nm3.

A new mercury removal unit based on the Boliden-Norzink process has been in operation since 1997. Mercury vapors are removed from the off-gases in a continuous hydrometallurgical procedure, where the gases are treated in a countercurrent absorption tower with an acidified solution of HgCb-complex dissolved in water, as shown in Figure 10.

The dedusted and demisted SCVcontaining gases are conveyed and passed through a double catalysis sulfuric acid plant, illustrated in Figure 11. After drying the gases, the SO2 is converted into SO3 in a 4-tray converter. Between the first and the second trays the converter gases are cooled in a steam superheater. After the third tray, the gases with a defined preliminary SO2 conversion are cooled in a gas/gas intermediate heat exchanger and are led to the intermediate absorber. Subsequent to absorption of SO3, the residual SO2 gases are preheated again to the conversion temperature in the above mentioned heat exchanger, prior to entering the fourth tray. Between the third and fourth layer, the gas is cooled in a steam generator. The hot gas coming from the fourth tray is passed to the heat exchanger heating the fresh gas before entering the final absorption stage. The gases exit the sulfuric acid plant with a remaining SCvcontent of less than 600 mg/Nm3. The generated sulfuric acid has a concentration between 96 - 99 % and a residual mercury content of less than 0.5 ppm. The overall conversion efficiency of the acid plant is higher than 99.7%. Various views of the equipment associated with the acid plant are shown in Figures 12-14.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 135

Figure 10 - Hg Removal Unit Figure 11 - Sulfuric Acid Plant

Figure 12 - Automatic Acid Loading

136 LEAD-ZINC 2000

Figure 13 - Boiler Energy Recovery Figure 14 - Steam Generator Unit

The general flowsheet of the smelter and acid plant is shown in Figure 15, and recent operational data are summarized in Table II.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Raw Material Storage

137

Flue dust, Coal w

02, N2, Air, Coal

►

4 - 5 t/h Flue Dust

v

Charge Preparation Bin

T

Mixer

1 r 50 t/h

QSL-Converter

Ϊ

1200 °C

r

Waste Heat Boiler

1 380 "C

Hot Gas-ESP

1 320 °C

Washing Tower

i , 75 °C, 28.000 N

Wet Gas-ESP

' r 30 °C

Hg-Removal

' 30 °C

Sulfuric Acid Plant

Dryer

Converter 1st-3rd Tray

Interm. Absorber

Converter 4th Tray

Final Absorber

1

Clean Gas < 200 ppm S 0 2

^ Leaching Residue

Final Slag 60.000 t/a *

-1 Cu-Dross ^

Lead Bullion

>100.000 t/a " 200 t/a Ag

Steam 15 t/h, 45 bar, "

256 "C

Flue Dust 1 t/h r

Ti3/h

Kalomel ^ I

H2S0496,5%

70.000 t/a

Disposal

Rotary Furnace

Refinery

Power Generation 3,5 MW

Cd-Leaching

Mischkarbonat

Disposal

Disposal

2000 m3-Tank < 0,5 ppm Hg

Figure 15 - Flow Diagram of the QSL-Plant at "BERZELIUS" Stolberg GmbH

Atmosphere

138 LEAD-ZINC 2000

Table II - Technical Data of the QSL-Plant at "BERZELIUS" Stolberg in 1999 (7)

Converter

Total Throughput Throughput Raw Material Lead Bullion H2S04 96,5% Disposable Slag Length Diameter Smelting Zone Diameter Slag Reductions Zone Decopperizing Kettles Temperatur Smelting Zone Temperatur Slag Reductions Zone

Off-gas

Temperatur Volume (moist) Composition SO2

0 2

C02 H20 N2

Waste Heat Boiler

210.0001 165.000 t 110.0001

> 70.0001 60.0001

33 m 3,5 m 3,0 m

2 x 3001 1100°C 1150 °C

1200 °C 28.000 Nm3/h

12 -14 % 12 -14 % 13 -15 % 13 - 15 %

Remainder

Gas Temperatur after WHB 3 80 °C Steam Production WHB 15 t/h Pressure WHB 45 bar Water Temperatur 256 °C

Hot Gas-ESP

Gas Temperatur beyond Hot Gas-ESP 320 °C Dust Content beyond Hot Gas-ESP < 100 mg/Nm3

Auxiliary Plant Units

Gas Temperatur after Washing Tower 70 °C Gas Temperatur after Wet Gas-ESP 30 - 35 °C Hg in Off-gas after Hg Removal Unit < 0,05 mg/Nm3

Temperatur Off-gas prior H2S04-Plant 30 - 35 °C H2S04-Production 250 t/d

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 139

PRODUCTION RATES OF THE QSL-PLANT DURING THE PREVIOUS YEARS

An important criterion for the economic operation of the QSL-plant was the availability of the plant, which could be permanently increased by implementing several technical improvements. The availability is shown in Figure 16.

Figure 16 -Availability of the QSL-Plant (7)

Corresponding to the improved plant availability, the specific throughput rates of materials and related lead production could also be increased, as illustrated in Figure 17.

Figure 17 - Raw Material Throughput and Lead Bullion Production (7)

140 LEAD-ZINC 2000

It is well know from the literature (5, 7) that the relatively short service life of the injectors and its adjacent refractory lining in the smelting zone and the slag reduction zone represented one of the major problems of the QSL-process. The plant downtime for replacing the injectors affected the availability to a great extent. The algorithm of several technical modifications to the injector design and operating mode as well as the refractory design successfully overcame these shortcomings. Today, a lifetime of 800h on average for the oxygen injectors in the smelting zone and 1300h on average for the coal injectors in the slag reduction zone is achieved without a problem. The trends are shown in Figures 18 and 19. The injectors are replaced in scheduled preventive maintenance plant shut downs after the above service life has elapsed in order to avoid unscheduled downtime.

Figure 18 - Service Life of Oxygen and Coal Injectors

Figure 19 - Wear Rate of Oxygen and Coal Injectors

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 141

Figure 20 shows the relation between the power production in the plant turbine utilizing steam from the waste heat boiler, the power consumption of the QSL-plant, including the preceding and subsequent plant facilities but without the refinery, as well as the required power from the public network. It is obvious that the power consumption of the QSL-plant including waste heat boiler, electrostatic precipitator, washing tower, mercury removal, acid plant and cooling towers amounts to only 75-80 % of the in-house generated power. More than 75% of the total power requirement of the entire smelter is generated in-house.

Figure 20 - Power Consumption and Generation at "BERZELIUS" Stolberg

SPECIFIC CONSUMPTION FIGURES

Because of the increased production rates, higher plant availability and the optimization of the overall process parameter, the specific consumption figures for energy carriers and addi-tives could be partially decreased to a greater extent and a reduction in the related cost realized. Table HI indicates the development and current status of these figures:

142 LEAD-ZINC 2000

Table III - Changes in the Plant Operating Parameters

1997/1998 1999 1999 It Feed Material It Feed Material It Bullion

Feed Rate Lead Production Additives Si02

CaO FeO Nitrogen Energy Carrier Oxygen Power Consumption Power Production Petroleum Coke Powdered Coal Natural Gas Compressed Air

35-38 13

23 70 52 46

168 86 105 21 31 6 61

t/h t/h

kg/t kg/t kg/t Nm3/t

Nm3/t kwh/t kwh/t kg/h kg/h Nm3/t Nm3/t

42 -48 16

20 63 57 38

142 79 95 9

35 4

80-90

t/h t/h

kg/t kg/t kg/t Nm3/t

Nm3/t kwh/t kwh/t kg/h kg/h Nm3/t Nm3/t

47 150 135 93

365 185 218 20 80 10

200 - 210

kg/t kg/t kg/t Nm3/t

Nm3/t kwh/t kwh/t kg/h kg/h Nm3/t Nm3/t

HEALTH, SAFETY AND ENVIRONMENTAL (HSE) ASPECTS

Because of the location of the smelter in the immediate vicinity of the city of Stolberg, decreasing the amount of emission was an absolute must and one of the driving forces to im-plement the QSL-technology. Figures 21 and 22 indicate the positive development in reducing the emission of SO2, Pb and dust during the last two decades.

Figure 21 - Development of SO2 Emissions (8)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 143

Figure 22 - Development of Lead and Dust Emissions (8)

Focussing on the implementation of new strategies for improving the situation in the areas of safety, the environment and quality control was one of the major issues to which time was devoted. New, improved respirator systems were introduced and the change, wash and break rooms were redesigned according to the most recent experiences. In addition, several industrial safety programs were initiated resulting in lowering the incidents of accidents (first aid, medical and lost time), the number of reported illnesses and the blood lead levels. The trend is illustrated in Figure 23. Since 1997 "BERZELIUS" Stolberg GmbH is DIN EN ISO 9001 certified. In 1998 the plant received the QS 9000 and VDA 6.1 certification and this year the ISO 14001 environmental certificate.

144 LEAD-ZINC 2000

Figure 23 - Development of Occurred Incidents at "BERZELIUS" Stolberg

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 145

TECHNICAL MODIFICATIONS, DEVELOPMENTS AND INVESTMENTS OVER THE LAST 10 YEARS

1990 Commissioning of the QSL-Plant 1991 Shut down of the Electro Furnace and Replacement by a Forehearth 1991 - 1993 Modification to Refractory Lining:

• Removal of two Partition Walls in Slag Reduction Zone • Modification of Injector- und Post Combustion Lance Locations • Relocation of Feed Ports in-between Oxygen Injectors

1993 Dismantling of Forehearth and Transition to Slag Granulation System at QSL-Slag Tap

1994 Modification of Waste Heat Boiler: • Modification at Transition Radiation Channel to Convective Pass • Installation of Additional 100 m2 Heat Exchanging Panels in Lower

Part of Radiation Channel 1994 Improvement to Rapping Device at Hot Gas-ESP 1996 Installation of Additional 80 m3 Heat Exchanging Panels in Upper Part of

Radiation Channel 1996 - 1998 Introduction of Program to Minimize the Amount of Waste Water 1997 Commissioning of Hg-Removal Plant based on Boliden-Norzink Process 1997 Certification DIN EN ISO 9001 1998 Expansion of Lower Heat Exchanging Panels from 100 m3 to 140 m3

1998 Replacing of Knocking Devices at Waste Heat Boiler to Pneumatically Driven Knocking Devices

1998 Construction of Effluent Treatment Plant 1998 Installation of Pneumatic Mailing System for Sample Transportation 1998 Certification QS 9000 und VDA 6.1 1998 - 1999 Introduction of Program to Optimize the Specific Energy Consumption 1998 - 1999 Reconstruction of Social Facilities and Break Rooms as well as Change to

Improved Respirator System 1999 Installation of New Crane in Raw Material Storage 1999 Commissioning of New Casting Machine and Expansion of Refinery Ca-

pacity to 130,000 t/y 1999 Installation of New Overhead Crane in Refinery 1999 Expansion of H2S04-Capacity by Application of Cs-Catalyst 1999 Expansion of Plant Capacity by Process Optimization, Enlargement of De-

copperizing Kettles to 2 x 300 t-Kettle 2000 Construction of New 2-step Wet Gas-ESP 2000 Construction of Building for Lead Storage 2000 Certification DIN EN ISO 140001 2000 Expansion of Sulfuric Acid Plant to more than 300 t/day

FURTHER DEVELOPMENT

The previously described improvement in plant availability and the achieved increase in material throughput are also attributed to extended furnace campaigns. The refractory lining of the converter was qualitatively and geometrically reformed over the last years as shown in Figures 5 and 24. The insertion of a so-called "step"-type injector resulted in a significantly

146 LEAD-ZINC 2000

longer lifetime and, consequently, higher plant availability. The reinforcement of the brick thickness at the transition slag line / gas atmosphere, which are areas of intensified attack and wear, will extend the period of a furnace campaign to 18-24 months in the future.

Figure 24 - Cross-section of the Refractory Lining in Slag Reduction Zone

Areas of higher refractory wear are also at the opposite side of oxygen and coal injector locations. The installation of water-cooled elements in these critical areas led to a higher lifetime of the brick lining. Currently, further investigations are being carried out to make the cooling elements even more efficient and to achieve a more stable performance of the critical areas so that furnace campaigns of more than 2 years can be realized.

The major section in the plant currently restricting higher production rates is the sulfuric acid plant. Therefore, an expansion of the sulfuric acid plant to rates higher than 300 t/day of monohydrate is in the planning stage.

The major task of all these measures is to elevate the throughput of raw material to 180,000 t/y to obtain an annual lead and lead alloy production to 120,000 - 130,000 t. Some forecasts even predict a possible raw material capacity of more than 200,000 t/y as realistic.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 147

For the thorough mixing and blending of the feed to satisfy steady-state charge preparation requirements and homogenizing the furnace operation, the entire material weighing and mixing system was changed to a continuous operating system. In this connection it is scheduled to implement a process control system for further optimization, in the near future. In order to become totally independent from weather influences, it is planned to cover the raw material storage.

Based on infrastructural conditions in the past, concentrates have been delivered to the plant completely by truck. After the completion of negotiations with the Federal Railway Organization, the partial shipment of concentrates by railroad directly from the seaport of Antwerp, Belgium to Stolberg began in 1998. Today, the amount of shipped concentrates by railroad totals 50%. From the year 2001 on, all purchased concentrates will be transported by rail.

The rising performance of the QSL-plant inevitably led to the requirement of expanding the existing capacity of the pyrometallurgical refinery. Aside from the installation of additional kettle capacity and the erection of a new casting machine, a covered and larger area for the final lead products has to be set up in order to complete all necessary preparations for a plant capacity of 130,000 t/y.

CONCLUSIONS

The lead smelter at "BERZELIUS" Stolberg GmbH in Stolberg / Germany has applied QSL-technology since 1990. Following some modifications to the process and the equipment, the plant demonstrated the viability and flexibility of the QSL continuous oxygen converter process. Optimum energy exploitation can be achieved by synchronization of the auxiliary plant sections with the QSL-converter and, thus, the operating costs are lower than those of the conventional sinter machine and blast furnace process and most other technologies. The design and construction of the plant guarantees minimum heavy metal emissions and, therefore, stringent environmental legislation can be attained and permanently maintained. In 1999, the following production figures have been obtained:

Lead Bullion > 100.0001 Dore Metal ca. 2001 Sulfuric Acid ca. 70.0001 Final Slag ca. 60.0001

At the same time, the emissions of SO2 and lead could be significantly reduced

S02-Emission Standard Process 2.1001 S02/a QSL <100tSO2/a

Lead-Emission Standard Process 13 t Pb/a QSL < 3 t Pb/a

148 LEAD-ZINC 2000

With the transformation from the conventional process to the state-of-the-art QSL-technology, "BERZELIUS" Stolberg GmbH now operates a process which simultaneously satisfies the three demands for modern metal production of improved pollution compliance, lower energy consumption and efficient production. "BERZELIUS" Stolberg GmbH possesses the most modern technology for producing primary lead worldwide and will continue its tradition of leadership in concentrate processing in the future.

REFERENCES

1. 150 Jahre "BERZELIUS" Stolberg, Publication for the 150th Anniversary of the Lead Smelter Berzelius Stolberg GmbH. Stolberg, Germany, 1998.

2. Siegmund, R. Piillenberg and A. Rohkohl, "QSL - Stolberg / Germany - Lead Smelting in the New Millennium", Zinc and Lead Processing. J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton and P. Hancock, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 825 - 842.

3. Victor Tafel, Lehrbuch der Metallhüttenkunde. S. Hirzel Verlagsbuchhandlung, Leipzig, 1953.

4. H. Mazcek, "Bleihütte Binsfeldhammer, Stolberg", Presentation during the 150th

Anniversary Celebration. Stolberg, 1998.

5. L. Deininger, K.C. Choi and A. Siegmund, "Operating Experience with the QSL-Plants in Germany and Korea", EPD Congress 1994. G. W. Warren, Ed., The Metallurgical Society of AIME, Warrendale, PA, U.S.A., 1994,477 - 502.

6. P. E. Queneau and A. Siegmund, "Industrial-Scale Leadmaking with the QSL Continuous Oxygen Converter", Journal of Metals. Vol.48, 1996, 38 - 44.

7. Rohkohl, "Process Data Survey", Berzelius Stolberg GmbH. Stolberg, 2000.

8. P. Arthur, A. Siegmund and M. Schmidt, "Operating Experience with QSL Submerged Bath Smelting for Production of Lead Bullion", Savard-Lee Symposium. Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1993.

9. G. Offermanns, "HSE-Data Survey", Berzelius Stolberg GmbH. Stolberg, 2000.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 149

A REVIEW OF AUSMELT TECHNOLOGY FOR LEAD SMELTING

E. N. Mounsey Ausmelt Limited A.C.N. 005 884 355

12 Kitchen Road, Dandenong Victoria, 3175, Australia

N. L. Piret Piret & Stolberg Partners

Im Licht 12, D-47279 Duisburg, Germany

ABSTRACT

Ausmelt technology is successfully used to process primary and secondary materials to produce lead in large-scale commercial operations and to meet stringent environmental standards. The development and commercial data covering these applications and the results with respect to metallurgical, cost and environmental performance are reviewed in commercial operations. A new smelter is scheduled for commissioning in 2000 at Korea Zinc's Onsan site to process lead fume and other secondary materials to produce 50,000 tonnes per annum of lead bullion. The unique features of Ausmelt technology also allow for the efficient processing of polymetallic concentrates and secondary materials, particularly copper and lead, through the effective separation of these elements into metal, slag and fume species. This capability has been successfully demonstrated at a commercial scale and the results and implications for the industry are discussed.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

150 LEAD-ZINC 2000

INTRODUCTION

Ausmelt technology for lead smelting was successfully commercialised at the Metaleurop lead smelter in Nordenham, Germany (1996) and at the Gold Fields of Namibia Tsumeb lead smelter in Namibia (1997). Since these key developments, Korea Zinc Co Ltd has contracted with Ausmelt for a further lead smelter. This will be commissioned in the second half of 2000 and will process lead oxide fume, sulphate residue and battery paste to produce lead bullion. These three operations are reviewed in this paper. They exemplify the operational flexibility provided by Ausmelt technology, covering oxide, sulphide and mixed feed regimes, and all meet strict environmental requirements with respect to in-plant emissions and dioxin levels in the offgas.

These commercial developments have occurred after many years of development work at Ausmelt. In excess of sixty pilot plant tests processing high, medium, and low grade sulphide concentrates, battery pastes, dusts and high grade slags have been carried out at the Ausmelt pilot facility in Dandenong, Australia.

This body of knowledge provides a significant process and operational data bank for the design, commissioning and operation of Ausmelt technology smelters and the application of top submerged lancing (TSL) technology to all aspects of lead smelting.

Furnace Description

The general arrangement of the Ausmelt commercial smelters broadly follows that shown in Figure 1. The key features of the design are:

• Top submerged patented Ausmelt shrouded lance • Cylindrical, refractory lined, shower cooled furnace • Metal and slag discharge via weirs or tapholes. • Afterburning in the gas space above the bath • All of the process fuel and air (may be oxygen enriched) injected into the bath via the

lance • Pugmilled or pelletised feed material directly fed to the bath from the feed port.

The superior process flexibility of the Ausmelt system emanates from the ability to accept varying feed types, to operate at varying but precisely controlled oxygen system potentials and to recover heat from the post combustion reactions to the bath via cascading splashed slag, thereby reducing the system fuel demand. The system effectively separates volatile species to fume, valuable non volatile species to metal and low value non volatiles to slag.

The superior environmental performance derives essentially from the totally enclosed furnace operating under negative pressure and from the sealing of all ports via specially designed devices. When processing secondary lead materials, the post combustion of process gases at temperatures in excess of 1300°C results in levels of dioxins and furans that are well below the statutory levels. These unique features are particularly well suited to the requirements of the lead smelting industry.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 151

Figure 1 - Schematic View of Ausmelt Technology Furnace

PROCESS DEVELOPMENT

Lead Smelting

The redox potential verses temperature chart for various Ausmelt technology processes is shown in Figure 2. The chart shows high, medium and low grade smelting as well as slag reduction regimes.

Sulphide Systems

With the commissioning of the third Ausmelt technology smelter at Korea Zinc in Onsan, Ausmelt technology lead smelters will process high, medium and low grade feeds for a wide range of sulphide material blends.

152 LEAD-ZINC 2000

Figure 2 - Redox Potential vs Temperature for Ausmelt Systems

The deportment of lead in an Ausmelt sulphide system between metal, fume and slag varies with the system oxygen to concentrate ratio as generally shown in Figure 3.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

The maximum direct lead yield to metal increases with: 100 ' Increasing Pb in Concentrate

% of "Stoichiometric" Oxygen / Concentrate

Figure 3 - Lead Distribution to the Different Phases for 50% Lead Concentrate

This relationship can provide an indication of the system distribution under the varying conditions encountered in operating Ausmelt technology plants. This is summarised in Table I for high, medium and low grade materials.

Table I - Lead Recovery during Smelting as a Function of Feed Grade Feed Grade

High

Medium

Low

Pb Content %

60-80

45-60

<45

Temp Smelting °C

1000-1100

1100

1150-1250

Direct Recovery To bullion

75-85

35-60

20-50

Smelting Slag Grade %

40-60

25-40

20-30

Distribution %

Slag

5-15

25-45

30-50

Fume

5-15

15-20

20-30

Overall Pb Recovery with fume recycl. (%)

90

55-75

50-70

Oxide/Sulphate Systems

The processing of oxide materials is carried out under mildly reducing conditions using coal as the reductant. Blends of oxide, sulphate and sulphide materials require reducing or oxidising conditions within the limits set by all-sulphide or all-oxide systems. The processing of sulphide and oxide blends can be carried out successfully, but this results in greater fume production and a higher sulphur content in the bullion than that which occurs when these materials are processed separately.

T3

ÖJ0 CO

2 o c o S

Q •a υ

"■5

154 LEAD-ZINC 2000

Slag Reduction

Lead Recovery

The oxidation smelting process step is followed by a slag reduction process to produce bullion and a discard slag as in conventional lead processes.

Precise control of the system oxidation state is a key feature of Ausmelt lance technology and this, together with the addition of a reductant, such as coal, allows the reduction step to be precisely tailored to meet specific requirements. The system is as efficient for slag reduction as it is for smelting.

The liquidus temperature of lead oxide-containing slags increases rapidly from 1000°C -1100°C to 1150°C - 1250°C as the lead in slag drops from about 40% Pb to about 15% Pb. Fluxing with limestone may also be needed to replace lead oxide by calcium oxide and to prevent the precipitation of zinc ferrite or magnetite from zinc- and iron-rich slags having lead levels in the region of 15% - 40% Pb in the temperature region of 1200°C. In effect, lead oxide acts as a flux to allow operation in the lower temperature region, and once the level drops, operations must move to a higher temperature region.

Slag reduction is a less economic exercise than smelting of high-lead content material directly to bullion, particularly with solid slag processed on a campaign basis. This step will, however, usually need to be carried out to meet TCLP requirements for long term storage of the slag.

Lead oxide fume is produced in increasing quantities with increasing temperature; it is collected in a baghouse or ESP and is recycled to the smelting furnace, with or without external fume treatment for impurity removal.

Zinc Recovery

If the zinc level in the slag after lead reduction is sufficiently high to make its recovery economically attractive, or if the slag is cleaned to a low metal content to facilitate its use in a specific market requiring such low metal contents, a zinc fuming stage is carried out.

The fuming of zinc from slag is achieved by feeding lump coal into the furnace whilst raising the temperature to the range of 1250°C-1300°C. This reduces the zinc oxide in the slag to zinc metal vapour. The vapour is removed in gases above the bath, where it is re-oxidised, together with hydrocarbons and carbon monoxide from the coal reactions, by injection of post combustion air into the gas space above the bath using the Ausmelt shrouded lance. This allows recovery of some of the energy released by the post combustion reactions back to the bath via the cascading splashed slag. The zinc fume is collected in a baghouse or ESP and is processed for zinc production.

Depending on the local circumstances of the plant, the zinc fuming step can be operated as a separate batch fuming stage in the slag reduction furnace, or it can be operated continuously in a third furnace. For economic reasons, none of the Ausmelt technology plants currently operating or planned will recover zinc to fume as part of the lead slag reduction step. The circumstances under which zinc fuming from slag is economically attractive arise where maximum value can be extracted from the process, in particular where:

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 155

Molten slag is fed to the Ausmelt fumer Zinc levels in the slag are in excess of 10% Zinc fume is relatively free of deleterious impurities, such as arsenic, antimony, cadmium, and halides Zinc fume can be consumed onsite or nearby.

PROCESS FLOWSHEETS

General Approach

Ausmelt's preferred approach to lead smelting is to use two furnaces where slag flows in a continuous manner from the smelting vessel (oxidising) to the reduction vessel (reducing) via weirs, siphons and launders, as shown in Figure 4.

Figure 4 - Continuous Process Two Furnace Lead Smelter

The second furnace is generally smaller than the primary vessel, particularly when smelting high grade feed, and the operating costs are low because of the low fuel and process air requirements resulting from the treatment of molten material. However, even though the second furnace is small, it requires a separate gas handling system typically incorporating an

156 LEAD-ZINC 2000

evaporative gas cooler and baghouse, and this increases the plant capital cost above that of the single furnace, two-stage operation.

In the development of each plant process design, Ausmelt uses its extensive in-house database drawn from eleven operating plants and over thirty smelters built to date incorporating Ausmelt technology.

The final slag product from the second furnace, in this situation, will typically contain low levels of lead (<3% Pb) but may well contain appreciable levels of zinc (8-12% Zn). In this circumstance, the slag can be granulated and held for long term storage or a second slag reduction can be carried out to recover the zinc as a high grade (>50% Zn) fume, as previously described, and to produce a discard slag containing low levels of lead (<1% Pb) and zinc (<3% Zn).

The selection of either single, batch, two furnace continuous, two furnace continuous and batch, or three furnace continuous process option(s) as generally indicated in Figures 5, 6 and 7 is discussed below.

Air Fuel (pi)

Concentrates Battery Paste Secondary Materials Fluxes I Reductant ^ Coal ^

Lead Fume to Recycle

Gas Cooling

—►

1 t A.

Gas Cleaning

(Slag Reduction)

-1* Sulphur Recovery

(stack)

High Level Slag

Discard Slag

""► Continuous Process '*" Batch Process

Figure 5 - Single Furnace Batch Process with Discard Slag and no Zinc Recovery or Single Furnace Continuous Process with no Discard Slag and no Zinc Recovery

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 157

Air Fuel

Lead Fume to Recycle

Concentrates Battery Paste Secondary Materials Fluxes Reductant Coal

Reductan Coal

Lead

Gas Cooling

Gas Cleaning Sulphur

Recovery

Gas Cooling

Lead Fume to Recycle

1 Gas

Cleaning

Zinc Fume

Discard Slag <l%Pb, ~3%Zn

Lead Discard Slag <3%Pb,~10%Zn

Stack

~* Continuous Process

-► Batch Process

Figure 6 - Two Furnace Continuous Process with no Zinc Recovery or Two Furnace Continuous/Batch with Zinc Recovery

Air Fuel (02)

Concentrates Battery Paste Secondary Materials Fluxes Reductant Coal

Lead Fume to Recycle

Gas Cooling

Gas Cleaning Sulphur

Recovery

Reductar i Coal

Lead

«ductant Ooai I

Lead

Gas Cooling

Gas Cleaning

> Lead Fume to Recycle

Gas Cooling

Gas Cleaning

Zinc Fume

Stack

Discard Slag < 1 % Pb, - 3 %

Figure 7 - Three Furnace Continuous Process with Zinc Recovery

158 LEAD-ZINC 2000

Table II provides a summary on the mode of operation of the single, two and three furnace Ausmelt systems.

Table II - Overview of Ausmelt Lead Smelting Operating Systems. Design of operation

Single Furnace

Two Furnaces

Three Furnaces

Furnace and mode of operation 1 Batch

1 Continuous

1 Continuous 2 Continuous 1 Continuous 2 Batch

1 Continuous 2 Continuous 3 Continuous

Number of stages

2

1

2

3

3

Final slag quality

Discardable

High lead

Discardable

Discardable very low

lead Discardable

very low lead

Smelting slag handling

Liquid

Granulation/ solid to external

reduction Liquid

Liquid

Liquid

Zinc Recovery

No

No

No

Yes

Yes

Commercial application

Tsumeb

Metaleurop

Korea Zinc

-

-

Note: No lead bullion is produced in stage 3, since the third stage is operated as a fumer.

Single or Multiple Furnace(s)

Many factors can influence the furnace selection for each specific operation.

Capital Cost

This is the most obvious criterion and is often the key driver in the adoption of a single furnace design.

Unit Operating Cost

The general trend with plant unit operating costs is for this to increase with reduced throughput and with more than one process operation in an Ausmelt furnace. However, in some circumstances unit operating costs may actually be reduced when switching from a two furnace to a single furnace operation because of a reduced labour requirement, when this can be sustained.

Overall Cost Profile

The capital and operating costs can be expressed via the concept of plant capital and operating cost per tonne of product based on a ten year plant operation; e.g., overall cost = (investment / 10 + annual operating cost) /annual production. This is shown in Figure 8 for single and two furnace approaches to lead smelting for different throughputs based on Australian conditions.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 159

x> Cu OJ C e o H < to O

U

I

"Single Furnace No Sulphur Capture

"Single Furnace Sulphur Capture

"Two Furnace Sulphur Capture

"Two Furnace No Sulphur Capture

0 20 40 60 80 100 120 1-

Lead Production (In Thousand Tonnes per Year)

Figure 8 - Single and Two Furnace Lead Smelting Systems

The single furnace options incorporate flue gas desulphurisation (FGD) scrubbing whereas the two furnace options incorporate double contact sulphuric acid plants.

The general shape of the total cost profile curve is similar for single and two furnace systems in the region 50,000 - 70,000 tonnes of lead production. Above 70,000 tonnes of lead per year, a two furnace system is the preferred option because of the operability difficulties in carrying out smelting and slag reduction steps at large scale in a single vessel. A cost advantage occurs in the region of 50,000 - 70,000 tonnes per year lead production with a single furnace without sulphur capture; e.g., Tsumeb.

There may be other reasons why two furnaces would be preferred to a single furnace operation for a particular project. For example, environmental constraints and interest on high lead inventories may prevent the long term storage of a high-lead slag prior to campaigning through the furnace to recover lead and other metal values. Alternatively, a single furnace may not be suitable on the basis of, for example, a need to provide continuous consistent strength SO2 to an acid plant.

ENVIRONMENTAL PERFORMANCE

Feed Systems

Feed systems must be specifically designed to minimize material spills via the use of belt scrapers, cleaners and seals. All cross-over points must be hooded and vented to a hygiene circuit incorporating a baghouse. These features are common requirements in all lead smelters,

160 LEAD-ZINC 2000

though the feeding of concentrates and secondary materials to an Ausmelt unit is simple to implement in engineering terms.

Fugitive Emissions

The Ausmelt reactor is totally enclosed with up to four small openings in the furnace roof to accommodate the furnace feed, lance, standby burner and sample tool entry. The furnace is operated under negative pressure, typically -20 to -40 Pa, to minimise fugitive emissions caused by pressure fluctuations. These must be further reduced to very low levels by the use of additional sealing devices on the feed and lance ports. Sealing systems used in current operations include a rotatory feed valve and spinning disc feeder on the feed ports, and water and compressible bellow seals on the lance ports.

Slag Quality

Slag produced from an Ausmelt technology smelter is a homogenous material containing typically low levels of lead and higher levels of zinc. The flushing action of the gas together with the enhanced metallurgical performance possible in the system, via the precise control of process temperature and system oxidation state, provide a slag, which when rapidly quenched, is suitable for long term storage on site. At lead in slag levels below 1%, the slag will meet TCLP criteria for lead and other deleterious elements such as cadmium, antimony and arsenic.

SOx

High levels of SOx are generated in the furnace offgas to optimise sulphuric acid production. The use of oxygen enrichment to 50% in the process gas results in SOx concentrations of 12% -16% at the furnace offtake. The effect of air ingress on the gas strength at the acid plant is to reduce this figure to typically 10% -14%.

Dioxins

Pilot plant data on lead battery paste, computer boards and scrap portable battery material smelting, taken on a number of occasions by NATA or USEP A/ISO registered analysts, has indicated dioxin and furan levels very much lower than the statutory limit of 0.1 ng/Nm3.

The Ausmelt system is extremely well suited for the processing of lead secondary and other materials with the minimum production of dioxins and furans, because of:

• Submerged slag bath combustion via the FeO/FeOi .5 couple • Intense smelting at temperatures from 1100°C to 1200°C • High post combustion temperatures in the region of 1300°C - 1500°C because of the

oxidation of hydrocarbons and lead and zinc fume by the Ausmelt lance shroud air • Complete combustion of all carbon bearing material at high temperature • Rapid cooling from temperatures in the region of 1300°C to the region of 350°C, in

either an evaporative cooler or a suitable waste heat boiler.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 161

Lance Cleaning

An automatic lance cleaning device is in use at one operational Ausmelt smelter. All slag is safely removed while the lance is being withdrawn, and is consigned back to the furnace.

Liquid Effluents

There are no liquid effluents from the smelting or slag reduction steps. Effluent treatment is, however, very important with respect to fume washing or the leaching of deleterious impurity elements, and in the gas cleaning and sulphuric acid systems.

Smelter Dusts

The Ausmelt system has a low rate of dust generation, typically less than 1% of the feed input. This process feature tends to concentrate volatile impurities, such as arsenic, cadmium, antimony and bismuth in the lead fume, thus providing a relatively small quantity of fume for further treatment.

ENGINEERING DESIGN

The furnace and gas handling system equipment selection and design are critical in obtaining a functional and cost effective plant. Two key areas are furnace containment and gas handling.

Containment

The conventional Ausmelt design of a conductive refractory lining and shower cooling is suitable for low to high grade lead smelting operations as confirmed at Tsumeb and Metaleurop. The afterburning reactions of lead to lead oxide are adequately contained by shower or channel cooling of the furnace to the offtake transition and the inclined offtake sections.

The afterburning reactions of zinc are more exothermic and a waste heat boiler type of offtake and roof system is recommended. This type of system is in successful use at Korea Zinc on zinc fuming plants.

Gas Handling

The quantity of lead fume to be handled varies significantly with the type and grade of material and the operating conditions, as generally shown in Figure 3 for sulphide systems.

The fume generation in the Tsumeb operation is significantly higher than at Metaleurop because of the different operating conditions with low (Tsumeb) and high (Metaleurop) feed grade regimes. Cooling and fume recovery was adequately handled at Tsumeb by the conventional evaporative gas cooler (EGC), spray cooler and baghouse system, following modifications including improved access for fume and accretion removal and spray nozzle performance. The Metaleurop design incorporated the use of a circulating fluidised bed boiler, and two ESPs with the gas then routed to the acid plant. The system is operating well after

162 LEAD-ZINC 2000

modifications in December 1998 addressed waste heat boiler operating issues which had occurred in the first eighteen months of plant commissioning and operations.

Future Operations

Either of the general systems used at Tsumeb or Metaleurop will be appropriate for use depending on local conditions.

For long term operational stability and cost effectiveness, particularly where high temperature operations associated with zinc fuming occur, a containment system incorporating a waste heat boiler type offtake, upper furnace and roof, with enhanced copper cooling elements in the furnace bath and slag splash areas as developed by Korea Zinc and Oschatz, and as schematically shown in Figure 9, may be an optimum configuration.

Boiler Tube Cooling

Copper Cooling

Figure 9 - Ausmelt Furnace with the Enhanced Containment System

Edited by J.E. Dutrizac, JA. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 163

AUSMELT TECHNOLOGY COMMERCIAL DEVELOPMENTS

Key reasons for the selection of Ausmelt technology for lead smelting by Metaleurop, Tsumeb and Korea Zinc are the excellent environmental performance, capital and operating cost profiles and process flexibility provided in a compact high intensity smelter system.

Metaleurop

The main driving force necessitating a change of smelting technology at Nordenham was environmental, with significant pressure emanating from the local farming community with respect to lead and other emissions to the surrounding area.

The Ausmelt technology plant, designed to process 120,000 tonnes per year of mixed concentrates and secondary materials to produce 90,000 tonnes of bullion, commenced operations in March 1996. After initial teething problems were overcome, the smelter is now achieving the results envisaged at the project commencement, as reported by Metaleurop (1).

The data shown in Table III indicate a significant reduction of emissions into the atmosphere compared to the sinter plant blast furnace technology operated earlier.

Table III - Atmospheric Emission Reduction using Ausmelt Technology Lead 83% Thallium 76% Cadmium 97% Mercury 65% Antimony 35 % Sulphur dioxide 90 %

The environmental performance of the Metaleurop furnace in regard to fugitive emissions has been optimised via the use of a sealed feed port and the use of a lance to lance port water seal.

The smelter capital cost of FF290 million in 1995 and a 20% reduction in operating cost when compared to the earlier technology plant have been reported by Metaleurop for the current operation (1).

The furnace is of conventional Ausmelt design approximately 4 metres internal diameter and 10 metres high, with external water cooling on the shell, roof and offtake. The furnace uses natural gas as fuel and the Ausmelt lances are oxygen enriched to approximately 40% to optimise process efficiency.

The lead bullion produced is refined to pure lead and a range of alloys mainly for the battery market. The high grade slag (-50% Pb) from the smelting process is routinely tapped and granulated, and is subject to further processing for lead and other metals recovery. The original design called for the campaigning of this slag through the Ausmelt furnace. Current practice is that the furnace is used for concentrate and battery paste smelting to bullion throughout the year, and the slag sold to third parties for metal recovery. This change to the originally envisaged method of operation is economically driven in that it is advantageous to maximise the time the smelter processes concentrates and battery pastes.

164 LEAD-ZINC 2000

The offgas is cooled in a waste heat boiler and de-dusted in two electrostatic precipitators before being fed to the existing acid plant. Total sulphur recovery is in excess of 99%.

Tsumeb

The Tsumeb project to process 130,000 tonnes per annum of low grade lead concentrates and secondary materials to produce approximately 30,000 tonnes of bullion shared many similarities with the Metaleurop Nordenham project. Essentially, changing feedstocks, increasing maintenance and operating costs and decreasing environmental performance of the existing plant necessitated a technology change. Ausmelt technology was selected after a comprehensive technology review by Gold Fields of Namibia as the most suitable for processing the predicted low grade feed to lead bullion with optimum environmental performance.

The plant design incorporated a single-furnace batch operation smelting low grade concentrates and secondary materials directly to bullion, with a separate smelting slag reduction step in the same furnace to produce additional bullion and a discard slag in a typical eight hour cycle. Extensive use was made of existing equipment in the project. Existing feed bins, coal dryer conversion to pelletiser, air blower, gas handling equipment including the baghouse, spray coolers and induced draught fans were modified for the new duty to keep the capital cost to a minimum level. The capital cost for the smelter installation is estimated at approximately AS20 million in 1996 (2).

The initial process design incorporated lead feeds in the region of 40 to 50% Pb, but the grade was progressively reduced through the development phase of the project. A major source of high grade lead concentrates was lost from the predicted feed mix when the Rosh Pinah mine changed ownership in 1996 and the lead grade of the available feed fell from approximately 30% to below 20% Pb.

The economics of smelting very low grade lead concentrates to bullion proved marginal at best, and in early 1998 the plant was used successfully to smelt copper concentrates to produce matte and a discard slag as a means of increasing revenue and profitability. Unfortunately the plant, as part of the Tsumeb complex, remains on cold standby following closure in March 1998 when Gold Fields of Namibia ceased operations.

Typical results from lead smelting are shown in Table IV.

Table IV - Feed and Product Assays

Feed

Bullion*

Discard Slag

(Assay) (Dist) (Assay) (Dist)

% Pb

27.8

63.4 98.1 2.8 1.8

% Cu

6.6

16.5 97.9 1.7 1.7

% As

3.4

8.4 98.4 0.3 1.6

% Sb

0.3

0.6 77.1 0.4 24.8

ppm

133

344 99.0

8 1.0

ppm Au

1

1.88 96.9 0.1 3.1

% Si02

5.43

0.28 2.6 29.7 97.3

% CaO

2.8

0.58 9.2 14.2 90.8

* Includes deportment to bullion and stockpiled fume.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 165

In the slag reduction step, molten slag produced in the smelting of the low grade 30% Pb feed, assaying typically 20% lead, is reduced to 4% lead on a cycle basis within 2 hours using coal as the reductant. This slag is granulated and held in long term storage.

Recovery of zinc was not required by Gold Fields of Namibia and the slag contains approximately 11% zinc. Levels below 4% lead are not required in this design in order to minimise fuming of zinc, though in practice, levels of 2% lead were routinely obtained.

The operating temperature of the smelting stage is 1180°C, and the slag reduction step is carried out at 1250°C. The higher operating temperature is required because of the low lead level in the discard slag.

The furnace does not incorporate any external feed or lance to lance port sealing systems, but ensures low level emissions by operating under conditions of adequate suction in the furnace.

Korea Zinc

The flexibility inherent in the Ausmelt technology will be further demonstrated by the operation of a single furnace Ausmelt system to process approximately 100,000 tonnes of secondary sulphate feeds, including fume, lead residue and battery pastes, to produce 50,000 tonnes of bullion and a high grade slag (-50% Pb) at Korea Zinc's Onsan plant. This slag will be processed for lead recovery in the existing QSL smelter (3).

The main driving force for this development is the overall process improvement to be gained by removing many internal secondary recycle streams, thereby freeing the QSL smelter for primary lead smelting activities. This smelter is the third Ausmelt technology smelter installed at Korea Zinc's Onsan complex and will commence operations in the second half of 2000.

The fume from the process will contain significant levels of cadmium, arsenic and antimony, and will require an acid leaching process step to reduce the impurity levels to allow recycle of the fume to the Ausmelt furnace.

The smelting process temperature is designed to be in the region 1000°C - 1050°C and a mildly reducing smelting regime will be utilised. The design mass balance for the process is shown in Table V.

166 LEAD-ZINC 2000

Table V - Design Mass Balance INPUT Weight % % % % % % % %

tpa Pb Zn Fe SiQ2 CaO S A12Q3 MgO Lead Feed Materials Recycle Fume

100,000 55.6 1.7 0.2 2.03 0.1 11.3 0.82 0.1

Coal (Fuel) Coal (Reductant)

11,494 8,293

0.6 0.5

10.2 9.2

0.2 0.2

0.2 0.4

4.1 4.0

0.2 0.2

OUTPUT Lead Bullion Slag

51,081 12,228 50.0 11.3 1.9 16.6 0.5

0.9 1.0 6.7 0.6

PLANT COMMISSIONING AND PRODUCTION ESTABLISHMENT

Project Delivery and Plant Startup

Both the Tsumeb and the Metaleurop plants had teething troubles during start-up. Ausmelt direct involvement in the commissioning activities was limited to the provision of a small number of commissioning shift support personnel. In both projects, the Ausmelt package was a component deliverable within an overall project controlled and managed by the client.

Recent start-ups have improved in delivering the design or close to design production levels within six to twelve months of start-up. The specific changes in the delivery of Ausmelt project services involve a significant up stream and down stream design overview role by Ausmelt, such that the impact of equipment or process step choices, outside the core Ausmelt furnace package, can be evaluated and changes made before any problems occur.

A feature of the early problems of both Tsumeb and Metaleurop was the high downtime of the gas handling equipment. This contributed significantly to plant downtime, in the case of Tsumeb, centering on the operation of an evaporative gas cooler and in the case of Metaleurop, a Fluxflow waste heat boiler. In both cases, the process equipment selection occurred directly between the client and supplier without a formal review step by Ausmelt. This review has now been incorporated into Ausmelt's procedures for future projects, with a view to provide a benefit to both the client and technology supplier.

The project management procedures in use at Ausmelt have been significantly improved and modified over the past three years. In addition to the project design overview, Ausmelt has an extensive in-house data bank of commissioning and operations knowledge gained from the start up of the eleven currently operating smelters, as well as from the over thirty Ausmelt technology pilot plant and development smelters designed, constructed and commissioned from the early 1970s. This knowledge was called upon directly during the commissioning of the Tsumeb plant when a problem of high refractory wear in the offtake transition duct was overcome by the installation of a cooling jacket around the area. The specific information for this development was provided by another Ausmelt technology plant which had experienced and solved the problem separately. Sharing of knowledge from operating plants can provide substantial benefits in new Ausmelt plant start-ups.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 167

OPERATIONAL FLEXIBILITY - THE PROCESSING OF POLYMETALLIC CONCENTRATES

Tsumeb Operations

The successful processing of polymetallic feed materials through a single furnace system to achieve marketable products has long been of significant interest to the metals industry, particularly with regard to the processing of secondary materials which tend to be of lower quality and mixed grade when compared to primary materials. The Ausmelt system has already achieved commercial application in this mode when the Tsumeb plant operated successfully as a polymetallic smelter for a period of approximately three months in early 1998.

With no changes to the plant required, the smelter operation was changed from lead smelting to copper smelting using high-lead copper concentrates operating on a cycle basis. The process was quickly established to smelt 100 tonnes of concentrate over four hours to produce, typically, a 50% copper matte, a 50% lead fume, and a discard slag of 0.5% copper at a temperature of 1150-1200°C. When the bath level of approximately one metre was attained, the melt was held for matte and slag separation. The matte and slag were then tapped from the metal and slag tapholes respectively, consecutively. This process takes typically ninety minutes with a further ninety minutes contingency on each cycle to cover any other activities.

Typical assays produced in normal operations during this campaign using polymetallic concentrates are shown in Table VI.

Table VI - Feed and Product Assays %Cu %Pb %As %_S

Feed 20 12 5 20 Fluxes Silica and limestone as required Matte 54 7 Slag 0.5 0.2 Fume 50 14

The copper matte was transferred in ladles to the copper smelter and the slag was granulated and discarded. Of particular note is the copper in slag content of 0.5% at matte grades of 50-55 % copper. The matte produced from the smelting step was converted to blister copper in existing Peirce-Smith converters in the adjacent copper smelter.

The high lead and arsenic fume produced was collected from the balloon flue, spray cooler and baghouse and was roasted in the Tsumeb arsenic plant to reduce the arsenic levels in the fume and to produce arsenic trioxide as a separate product. In due course this fume was to be smelted in a separate lead campaign, together with other lead bearing materials, to produce bullion in a similar manner to that previously established.

The operation over the full period was essentially trouble free. The operating mode was provisionally established as successive campaigns of one month polymetallic concentrate smelting to matte followed by two months of lead production from high grade fume and lower grade secondary materials.

168 LEAD-ZINC 2000

Ausmelt Technology Converting

Where no copper smelter is available, a logical development would include the production of blister copper from matte in the same Ausmelt furnace by using the Ausmelt converting process, successfully commercialised at Bindura Nickel Corporation, Bindura, Zimbabwe (1995) and the Zhong Tiao Shan copper smelter in Houma City, Shanxi Province, People's Republic of China (1999). There is a significant environmental benefit in the processing of polymetallic materials to blister copper via Ausmelt Technology compared to the traditional Peirce-Smith converter route (4,5). In this regard, the predicted results for the converting of the Tsumeb matte via Ausmelt Technology are shown in Table VII.

Table VII - Predicted Products from Ausmelt Technology Converting at Tsumeb

Matte Blister Copper Converter Slag Fume

%Cu 54

98.5 7.2 10

%Pb 7

0.6 3 60

%As 2

0.2 0.9 20

%S 24 0.1 0.2 6

This process requires two or three distinct steps depending on whether a final slag cleaning step for complete base metal recovery is required. The steps for single and two furnace operation are:

• Smelting to produce matte, discard slag and a high lead fume, also containing any other volatile elements, such as arsenic. The matte can then be converted in the same or a separate furnace to produce blister copper, converter slag and fume of the general assay shown in Table VII.

• In a single furnace after tapping the blister copper, the converter slag can then be either tapped and granulated for recycling to the smelting stage or reduced using coal as the reductant to produce additional copper and a discard slag.

• With two furnaces, the smelting furnace would operate continuously. A separate, smaller converting unit would operate either continuously to produce blister copper and a converter slag for granulation and recycle to the smelting furnace, generally as indicated for lead smelting in Figure 4, or the converter could operate in a batch mode to produce blister copper and a discard slag generally as indicated in Figure 5.

• The fume products are recycled if possible, but will be removed for separate treatment if the impurity content is too high.

The final process choice will be governed by economics, environmental or other issues which may be peculiar to each operation.

Economic Benefits

The benefits of this technology, via the inherent flexibility of the process to handle very different materials in the same furnace, provides, in effect, an economic multiplier for a plant owner. This feature is likely to be of significant interest to operators considering suitable technologies for lead and polymetallic smelter projects.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 169

REFERENCES

1. S, Karpel, "Greater Flexibility and Lower Costs at Nordenham", Metal Bulletin Monthly. December 1998,40-45.

2. G.P. Swayn and E.N. Mounsey, "Ausmelt Technology Versatility: The Design and Operation of the Ausmelt Lead Smelter and the Subsequent Development of the Unit to Successfully Produce Copper Matte at Tsumeb Smelter, Namibia" GDMB Conference, 58th Meeting of The Copper Committee. Alpbach, Austria, September 24-25, 1998.

3. K.R. Robilliard P.J. King and W.E. Short, "The Application of Ausmelt's Top Submerged Lancing Technology to the Lead Industry", Eleventh International Lead Conference. Venice, Italy, 25-27 May, 1993.

4. J.M. Floyd and H. Li, "Modernisation of Chinese Lead Industry - Application of Ausmelt Lead Technology," Ausmelt Presentation to The State Nonferrous Metals Industry Administration of the People's Republic of China. Beijing, China, February 2000.

5. T.C. Hughes, "Ausmelt Technology - The Answer to Toxic Organohahde Production during High Temperature Treatment of Wastes", 31 May, 1999.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 171

COMINCO'S NEW LEAD SMELTER AT TRAIL OPERATIONS

D. W. Ashman Cominco Research

P.O. Box 2000 Trail, British Columbia, Canada VIR 4S4

D.W. Goosen, D. G. Reynolds and D. J. Webb Cominco Limited P.O. Box 1000

Trail, British Columbia, Canada V1R 4L8

ABSTRACT

Cominco Limited operates an integrated zinc-lead complex at Trail, British Columbia, Canada. Central to the successful operation of this complex is the treatment of zinc leach residues through the lead smelter. In 1997, Cominco started up a new lead smelter using KTVCET flash smelting, a state-of-the-art conventional slag fuming furnace, and a new drossing plant. The smelter has fully met the requirements for environmental performance, reduced energy consumption, and lowered labour costs. This paper provides a description of the new smelter and discusses the problems encountered during the start-up, as well as the solutions to those problems. Recent operating results are presented.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

172 LEAD-ZINC 2000

INTRODUCTION

Cominco has operated a zinc-lead smelter at Trail since the early 1900's. A close integration between the lead and zinc operations has developed over the years to maximize metal and co-product recoveries. Modernization of the operations began in the late 1970's, and this resulted in the capacity of the zinc operations increasing faster than the ability of lead operations to treat the corresponding iron residues. The increasing size of the iron residue stockpile, along with increased labor and fuel costs and increasing environmental pressures, justified replacement of the existing sinter machine/blast furnace technology.

The selection of smelting technology was very arduous for Cominco because of the unique metallurgical requirements for smelting zinc leach residues. The QSL process and a new dressing plant were installed and commissioned in 1989. Unfortunately, the start-up was unsuccessful, and after further technical study, the decision was made to abandon the QSL process and to install a KIVCET furnace. Snamprogetti was retained as the license and engineering services supplier. The KTVCET furnace, along with a new slag-fuming furnace, was commissioned in the spring of 1997 (1,2,3).

The new lead smelter is now operating at design rates. The majority of the charge is made up of zinc plant residues, Sullivan mine lead concentrate, and precious metal concentrates. The smelter has enabled Cominco to meet its objectives of reduced energy and labour costs and reduced environmental impact.

THE NEW LEAD SMELTER

Feed Preparation

A simplified flowsheet of the new smelter is given in Figure 1. The feed preparation plant has a 30,000-tonne storage capacity. Railcars and trucks deliver feed materials into twenty-one main storage bins. The material is distributed into twelve proportioning bins according to a predetermined master charge program. The master charge program determines the feed rate from each proportioning bin to the elemental set points and calculates the theoretical amount of oxygen and carbon required to fuel the charge. This provides operating personnel with guidelines for furnace operation.

The smelter treats all of the zinc plant residues, which constitute approximately 45-50% of the charge. The percentage of zinc plant residues is a function of the smelter elemental limits and maintains a specific heat balance in the KTVCET furnace. Residues from the oxide leach plant and sulphide leach plant are blended and pumped to the feed plant at a specific gravity of 1.75. The residues are filtered through two Ingersoll-Rand Lasta plate filter presses. The total residue filtering capacity is 825 dry t/d, with the filter cake containing 20-22% moisture.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 173

υ > o -4-»

13

a

■a

I

174 LEAD-ZINC 2000

The feed from the proportioning bins, including the residue filter cake, has an average moisture content of 13-14% H20. The feed material is fed by a series of belt conveyors to a rotary dryer reducing the moisture to less than 1%. The dryer is a modified version of an existing 3.04-m diameter by 24.4-m long rotary dryer formerly used in the sinter plant operation. It contains a 20,000 Mcal/h natural gas burner. Dilution air is used to control the dryer inlet temperature at 625-675°C to reduce sulphur oxidation, which causes corrosion problems in the dryer exhaust baghouse.

The coarse feed from the dryer is conveyed into a Fuller ball mill with a capacity of 70 t/h. The normal ball charge consumption is 0.75 kg per tonne of feed. This achieves a product grind of 100 percent minus 1 mm. The balls are added to maintain a power draw from the mill of 900 kW. The fine feed entrained in the gas stream from the dryer is collected by two parallel cyclones and a Wheelabrator baghouse. The cyclone and baghouse dust catches are added to the ball mill product. The feed is conveyed to a 53-m tall charge bucket elevator with a capacity of 113 t/h. This is delivered to two 80-t day bins or a 1750-t storage bin, depending on the KTVCET operation.

Typically, a small percentage of the feed to KTVCET is from the storage bin and the remaining feed is from either day bin. This practice ensures that the storage bin is regularly turned over. The feed is weighed and delivered to the furnace in two separate feed chains at a rate up to 28 t/h each. Coke and recycle dust are added to each feed chain and are mixed in mixer-levelers. The coke size ranges from 5-15 mm and constitutes 1.5-2.5% of the feed. The recycle dust is treated at a rate of approximately 8 t/h. The feed from the mixer-levelers drops into splitters where the feed is equally divided and delivered to four charge burners from the two feed chains.

KIVCET Furnace

The KTVCET furnace is divided into three sections: the reaction shaft, the electric furnace, and the uptake shaft, as illustrated by Figure 2. The reaction shaft is made entirely of water-cooled copper elements. There are four charge burners in the reaction shaft roof. In addition, there is a 12,000 Mcal/h natural gas auxiliary burner in the middle of the reaction shaft used to heat the shaft after prolonged shutdowns. All four burners normally operate. However, the furnace can operate with one feed chain at reduced rates. The feed is mixed with 99.9% oxygen at a rate of 12-14 t/h. The burners are typically cleaned once during each 12-h shift. The feed-to-oxygen ratio in the burner is critical in maintaining a stable reaction shaft operation. Excess oxygen results in an oxidizing atmosphere that increases the lead content in the slag and possibly promotes magnetite formation. A deficiency in oxygen will result in unburnt feed that forms a matte phase and sulphide deposits in the uptake shaft.

The reaction shaft is monitored by taking hourly flame temperatures. The flame temperature is controlled between 1380-1420°C. A low flame temperature indicates that the feed is under fueled, requiring an increase in carbon and/or oxygen. A high flame temperature indicates the feed is over fueled, requiring a decrease in carbon and/or oxygen. Normal coal adjustments made for fueling requirements are 2.5-3.5% of the total coal on charge. The feed is analyzed for carbon every three hours to monitor feed consistency. In addition, smelt samples are taken from the reaction shaft under each burner when required. The smelt samples indicate the efficiency of each burner in terms of unburat feed. The smelt samples are assayed for their sulphur content. The typical amount of sulphur remaining in the smelt is 0.5-1.2%.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 175

Coke added to the charge floats on the slag layer in the reaction shaft to form a "coke checker". The purpose of the coke checker is to reduce the lead oxide in the smelt. The coke is added at a rate of approximately 1-1.5 t/h to maintain a coke layer thickness of 100-150 mm. The coke checker thickness is measured with a steel bar on an hourly basis. The temperature of the coke checker temperature is maintained between 1100-1200°C. Temperatures above 1200°C indicate a thin coke checker, and temperatures below 1100°C indicate a thick coke checker.

Figure 2 - The KIVCET furnace

The smelt from the reaction shaft flows through the coke checker and under a submerged water-cooled copper partition wall. The partition wall separates the reaction shaft from the 55-m2 electric furnace as illustrated in Figure 2. The electric furnace acts as a settling furnace to provide the added retention time for the metallic lead to settle through the slag. Three 900-mm diameter electrodes, in line with each other, deliver 5-7 MW of power to the slag. The slag temperature is maintained between 1320-1360°C and the lead temperature is maintained between 850-950°C. Varying the immersion of the electrodes controls the slag and lead temperatures. A high bullion temperature is required to prevent matte formation in the electric furnace. Matte is very difficult to tap from the furnace and it creates several downstream problems. The total bath level in the furnace is maintained between 1570-1820 mm from the base of the hearth. The slag depth varies with slag tapping between 600-1000 mm. Approximately 80-95 tonne of slag is tapped from KIVCET every 3-3.5 hours to the slag-fuming furnace. The lead bullion level is maintained below 900 mm and is tapped through three specially designed 50-mm diameter tapping inserts located at an elevation of 730 mm. A fourth tap-hole at a lower elevation is used for draining the furnace. The tap-holes are opened and closed with a mechanical drill and mud gun.

176 LEAD-ZINC 2000

The hearth of the KTVCET furnace is built in the form of an inverted arch utilizing three types of brick. Magnesite-chrome bricks are directly in contact with the lead bullion. The last four outer perimeter bricks are made of alumina-chrome to minimize hydration problems, which cause hearth expansion. The magnesite-chrome bricks rest on a l-mm thick sheet of stainless steel. Below the stainless steel sheet is a layer of 150-mm graphite bricks that uniformly distribute the heat from the hearth. The hearth is air-cooled through fourteen channels below the graphite bricks with three variable-speed fans with a maximum capacity of 60,000 Nm3/h. The operation of the cooling fans is controlled to maintain the hearth temperature between 300-350°C. The roof of the electric furnace is an arch design built entirely of magnesite-chrome brick. Thermal expansions of the hearth and electric furnace roof are accommodated for through the use of sixty-eight springs; fifty-two for the hearth and sixteen for the electric furnace roof. The springs are measured frequently to monitor movement.

The sidewalls of the KTVCET furnace are constructed with water-cooled copper jackets. A total of two hundred and seven independent cooling elements are supplied through a central distribution system. The cooling water is maintained at 40°C and the outlet temperature is measured frequently to monitor jacket integrity. Typical outlet temperatures are 43-50°C depending on jacket location. Tap-hole insert temperatures are monitored during tapping. Typically each tap-hole insert will have a service life of two hundred taps.

Drossing Plant

Lead bullion is tapped to the continuous drossing furnace (CDF). The purpose of the CDF is to precipitate copper into matte by cooling the bullion. Recirculating the lead through a cooling machine cools the bullion. The freeboard temperature is maintained at 1280-1320°C by three natural gas burners. A matte layer is maintained in the furnace to insulate the slag from the cold lead. Sulphur is added to the return pot to prevent the formation of speiss. The temperature of the lead is reduced to 375°C.

The lead is transferred from the CDF to a series of induction-heated steel pots, a softener, and then to a continuous sulphur drossing circuit (CSD). The softener controls the antimony and arsenic at 1.2% and 0.7%, respectively. The remaining copper from the CDF is reduced to less than 0.005% in the CSD. The lead bullion is cast into 3-tonne buttons for shipment to the lead refinery. The copper matte and softener slag are treated in downstream plants.

KIVCET Gas Handling

Process gas from the reaction shaft enters the uptake shaft at 21,000 Nm3/h, 1375°C, and 14-18% SO2. The uptake shaft is constructed entirely with water-cooled copper elements. It is joined to the reaction shaft by four copper jackets called the "bullnose" as shown in Figure 2. The KTVCET waste heat boiler is divided into three sections: radiant, downcomer, and convection. The radiant section consists of 717-m2 of membrane wall and reduces the process gas temperature at the top of the uptake shaft to 800-830°C. The process gas flows through the down comer section with a surface area of 486 m2, and is cooled to a temperature of 600-630°C followed by the convection boiler with a surface area of 1012 m2. The outlet gas temperature from the waste boiler is 325-350°C. Typically, the waste heat boiler produces 23-25 t/h of steam at charge rates of 56 t/h. The on-line cleaning of the waste heat boiler consists of spring hammers for the down comer and radiant sections and pneumatic rappers for the convection section.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 177

The process gas flows through a four-field electrostatic precipitator (ESP) of Lurgi design. The dust caught by the convection boiler and the ESP is recycled back to the charge burners. The gas flows from the ESP at 300-325°C into an adiabatic spray tower with a dust loading of less than 5 mg/m3. The adiabatic spray lowers the gas temperature to approximately 65°C with a recirculating water flow of 6,600 L/min. The gas passes through a packed cooling tower, which reduces the gas to a temperature of 18°C. The gas enters a SO2 fan and exits into a 900-mm diameter by 500-m long flue to the zinc operation acid plants. A condensed water purge from the gas scrubbing system is passed through a SO2 stripping tower before being pumped to the effluent treatment plant.

Slag Fuming Furnace

The normal slag charge is 80-95 tonne with a typical backpressure of 75-85 kPa at the tuyeres. The time required to charge the slag fumer is dependent on the slag level in KTVCET and typically takes between 60-90 minutes. The KTVCET slag, at a temperature of 1320-1360°C, contains 16-18% Zn. During charging, the furnace is operated with coal rates up to 80 kg/min to increase fuming of all the elements. The slag temperature after charging is 1200°C and is further reduced to 1140-1160°C when the furnace is tapped, because of the endothermic reactions. The furnace is tapped when the zinc in slag is between 1-2.5%. The normal tapping time is 25-30 minutes from two 100-mm tap holes. The barren slag tapped from the fuming furnace is granulated at a rate of 3-5 t/min with 2200 m3/h of water into a granulation system. A complete fuming cycle is approximately 205 minutes.

The slag furnace is a rectangular water-cooled box with the inner dimensions of 2.44 m by 6.4 m. The slag enters the furnace 5 m above the hearth that consists of seven water-cooled sole jackets. A protective layer of frozen slag is formed to reduce slag erosion during charging. The water enters each sole jacket at 24°C at a flow rate of 270 L/min and exits at a temperature between 28-32°C. The lower sidewalls contain fourteen mild steel jackets that each contain three 38.1-mm tuyeres. The tuyeres deliver coal and air to the slag furnace. The primary air required to carry the coal to the furnace is 4,000 Nm3/h; the secondary air required for combustion is supplied by a Roots blower at 18,000-20,000 Nm3/h.

The process gas leaves the slag-fuming furnace at a rate of 26,000 Nm3/h. Additional tertiary air, approximately 30,000 Nm3/h, is added to burn the 12-16% carbon monoxide in the off-gas. The slag fuming furnace waste heat boiler consists of the radiant, down-comer, and convection sections. The gases are cooled from 1275-1325°C to 350-450°C in the waste heat boiler. The boiler is cleaned with eighty spring hammers and fifteen saturated steam soot blowers. The process gas is further cooled by ambient tempering air to 170°C. The gas flows through the new Asea Brown Boveri Inc. (ABB) baghouse where the fume is collected. Fume production is approximately 180-200 t/d and is leached in a sodium carbonate dehalogenation process before being pumped as a slurry to zinc operations.

COMMISSIONING AND START-UP

The first feed went into the KTVCET furnace on March 31, 1997. The start-up of the new slag-fuming furnace was three months later in June 1997. The initial period of operation was hampered by mechanical problems that prevented sustained periods of operation until a major shutdown in November 1997 when several mechanical deficiencies were corrected. This

178 LEAD-ZINC 2000

initial "mechanical commissioning" period was followed by a "process commissioning" period until July 1998. The second major shutdown occurred, and the remaining mechanical deficiencies were repaired. Following this second shutdown, the KTVCET furnace was ramped up to full design operating rate by December 1998. Approximately 20 months expired from the first feed to operation at the design rate.

Mechanical Commissioning

During the initial period of KTVCET operation, the most significant mechanical problems were external to the KTVCET furnace itself. These were related to the feed system, slag and bullion launders, and the slag granulation system.

Feed System

During design, a synthetic dry feed was tested to determine the material's handling characteristics. Unfortunately, the tests did not give sufficient indication of the tendency of the actual dry feed to fluidize. This, together with the damage caused by coarse material in the feed originating from battery scrap, resulted in major fluctuations in the feed rate and the feed-to-oxygen ratio. In July 1998, a 1.5 by 3-m Rotex vibrating screen with a 6-mesh screen deck was installed downstream of the ball mill. This screen separated out the coarse battery metallics and entrained plastic. Elimination of this coarse material prevented jamming and wear of the rotary valves on the feed system. Fluidization of the feed has been virtually eliminated by a number of measures including; air permeation nozzles on the bins to control aeration, carefully designed rotary valves which act as air locks, and changes to the control strategy.

Slag Launders

Based on past experience at the Trail smelter, the original slag launders were constructed of water-cooled cast iron launders. During the initial start-up period, the poor furnace operation resulted in high levels of lead in slag. With the high lead in slag and the high temperature of the KTVCET slag, corrosion attack of the cast iron resulted. In December 1997, water-cooled cast copper launders were installed which virtually eliminated the launders as a source of downtime.

Slag Granulation

Many of the problems with the slag granulation system were due to excessive production of fine slag particles which accumulated in the circuit. Modifications to the granulating water sprays have helped to produce coarser slag particles. Installation of several submersible pumps in the dewatering tanks has prevented solids accumulation. The cooling tower was found to be inadequate. It was replaced with a shell and tube heat exchanger. A Buhler drag conveyor was installed as a rock-catcher across the granulation tank. This prevented occasional large slag chunks from blocking the slag slurry pumps. Additional measures were taken to reduce the amount of abrasive wear on the pumps and pipes.

Reaction Shaft Cooling Jackets

The reaction shaft cooling jackets are constructed of copper tubes cast in copper with swallowtails on the hot side to anchor a layer of protective refractory. During the furnace heat up, much of the protective refractory spalled off. The spalling and excessive heat fluxes led to

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 179

premature failure of some of the jackets. Inspection of some of the failed jackets revealed that a small gap developed between the copper tubes and the castings. New jackets were designed which incorporated the following features:

• Adequate allowance for thermal expansion of the refractory • Dual cooling water circuits.

Heat loads have also been reduced by avoiding high flame temperatures and by using modified burners. This reduces the amount of flame impingement on the sidewalls. No further problems have been experienced since the new jackets were installed in July 1998.

KTVCET Boiler Corrosion

Because of the experience at the KTVCET plant at Portovesme, Italy, chloride attack at the bottom part of the radiation boiler was anticipated before start-up. This is the reason for using replaceable boiler panels at the bottom of the vertical boiler. A protective flame spray coating was also applied. The coating failed after a few months and it became necessary to apply an Inconel 625 weld overlay. This overlay now covers 160 m2 of the boiler surface and appears to be working well.

Slag Fuming Furnace Boiler

Once the KTVCET furnace was operating consistently enough to provide a steady flow of hot slag to the fuming furnace, it became apparent that fouling of the slag furnace boiler was preventing the off-gas from being cooled below the design of 350°C at the boiler outlet. Since quench air is used after the boiler to cool the gases to 170°C for baghouse filtration, the coal rate had to be kept below the design rate to prevent draft restrictions. A tenacious, thick, fluffy layer of fume was fouling the boiler tubes. This did not come off easily by rapping. In April 1999, an additional tube bundle and fifteen soot blowers were installed. The fuming furnace can now operate above its design rate.

Oxygen Plant

Oxygen supply interruptions to the KTVCET furnace have caused a number of extended furnace outages. The supplier of the oxygen has made major efforts to reduce the interruptions with further improvements planned in 2000. Cominco brought a 1,500-t oxygen storage tank into service in January 2000 to provide a backup supply.

Process Commissioning

During mechanical commissioning, intermittent and unsteady-state operation resulted in poor process results. The mechanical and process issues became entwined and it was difficult to distinguish them. The process issues were as follows:

• Poor combustion of the charge causing piles of unburnt feed at the base of the reaction shaft

• Poor operation of the coke checker resulting in high lead in slag • Deep bath accretions resulting in excessive level fluctuations between taps, formation of

a matte layer, and difficult tapping conditions • Poor boiler performance because of accretions and fouling

180 LEAD-ZINC 2000

• Furnace outages to remove accretions in the boiler and electrostatic precipitator • Blockages of the "bullnose" (passage from the reaction shaft to the vertical uptake to the

boiler) • Excessive electrode breakage and consumption • Restricted slag furnace operation to avoid corrosion of the steel jackets by high lead in

slag.

Cominco had gained a strong knowledge base through its collaboration with the process licensors during the pilot plant and industrial scale testing of the KTVCET process (4). However, when faced with the challenge of getting the KTVCET furnace to design rate, a technical team was created in October 1997 to develop the process understanding required to achieve design operating rates. The team consisted of members of Lead Operations and Cominco Research. A number of parallel investigations were initiated including:

• Characterization of furnace operation through extensive furnace sampling campaigns • Burner development including cold modeling with a full-scale burner, CFD modeling,

development of a prototype burner, and studies aimed at developing a single charge burner to replace the four in use

• Thermodynamic modeling of the reaction shaft, coke checker, waste heat boiler and electric furnace using F*A*C*T, and an extended slag database for the silica-lime-iron oxide-zinc oxide-lead oxide system developed by the University of Queensland. (5)

• Fundamental studies of coke checker reduction mechanisms including mathematical modeling and laboratory studies at the University of Queensland

• Evaluation of boiler performance through water modeling and studies of accretion formation mechanisms and management strategies

• Accretion management by CFD modeling of heat dissipation (McMaster University) and electric furnace control strategies (Mintek)

• Supplemental reduction through plant testing of pulverized coal injection into the KTVCET electric furnace.

As additional knowledge of KTVCET operation was developed, a plan to resolve the process issues was implemented. Were it not for the close co-operation between Cominco Research and Lead Operations, and the willingness to try new things, the ramp-up to design rate would have taken much longer.

New Charge Burners

During mechanical and process commissioning, a number of issues prevented the KTVCET furnace from operating at the design rate. The lower oxygen flow through the burners prevented operating as designed and resulted in the piling of unsmelted feed on the coke checker. The horizontal oxygen tuyeres just above the coke checker, originally installed to heat the coke checker, helped to burn down the piles. Coal additions had to be increased so that higher oxygen flows could be maintained ensuring good dispersion of the charge. A new burner design was developed based on experience in copper and nickel flash smelting. After testing, the new burner design was installed in all four burner positions in June 1998. The new burners eliminated the formation of un-smelted piles and made the oxygen tuyeres redundant. This allowed the coal content of the feed to be reduced to design levels. This also reduced the off-gas volume and reduced the load on the waste heat boiler. The new burner design is recognized to be one of the major factors in the rapid ramp-up to design rate in the remainder of 1998.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 181

Because the new burners were so successful, the original burners have not been re-introduced so that their operation could be checked at the designated conditions.

Feed Variability

In addition to the mechanical problems preventing constant flow to the burners, the composition of the feed was also extremely variable. This was traced to three things:

• Variation in fixed carbon content of the fuel coal • Variable residue composition because of poor blending of the different zinc plant

residues • Variable flow of recycle dust

To steady out these fluctuations a better coal supply was sought. An inexpensive residue blending facility was installed upstream of the smelter residue filters and recycle dust bins with loss-in-weight discharge were installed. Also, to provide a faster measure of the carbon content in the feed, a Leco carbon analyzer was installed in the plant.

Lead Reduction

When the KTVCET furnace bath is accreted, it is very difficult for heat to be transferred under the water-cooled partition wall from the electric furnace to the coke checker. Thus, lead in slag is difficult to control. The reaction shaft can be operated in a more reducing environment by raising the coal in the feed and increasing the feed-to-oxygen ratio. In this manner, it is possible to operate with lead in slag at 2-5%. This practice has many detrimental side effects. Accretion growths block the bullnose and limit the draft on the reaction shaft.

Experience has shown that, if furnace accretions are controlled, the operation of the coke checker improves dramatically. It is possible to average less than 5% Pb in slag consistently. The key factors to control accretion growth are:

• Steady operation of the charge burners at high rate • Close control of flame stoichiometry • Control of slag fluxing • Close monitoring of the coke checker.

With furnace accretions low, the heat losses increase, and the electric furnace transformer becomes limiting. A three-in-line electrode system is unstable and requires extra attention to control. A Minstral electric furnace controller was purchased from Mintek and has stabilized the operation of the electric furnace. One can operate closer to the transformer current limit when required. The reduction in furnace accretions has also improved the lead tapping conditions. Less drilling and oxygen lancing are required. Electrode usage has also decreased as the lead in slag has dropped.

Boiler Performance

During normal operation, the KTVCET operating rate is limited by the capacity of the radiation section of the waste heat boiler. A key variable to prevent the buildup of accretions in the downcomer section of the boiler is the temperature at the top of the radiant shaft. A

182 LEAD-ZINC 2000

temperature of 830°C must not be exceeded, as this is the melting point of a basic lead sulphate eutectic.

It is also important to control the chemistry in the bottom part of the radiant boiler. One must avoid condensation of insulating deposits on the boiler walls. For example, while operating with high reduction in the reaction shaft, lead sulphide can condense and deposit on the boiler walls.

PERFORMANCE HISTORY: KIVCET AND SLAG FUMING

Throughput and On-Line Time

Improved mechanical availability and better understanding of the process chemistry has contributed to improved KIVCET on-line time. Design availability of 89% and an average feed rate of 56 t/h were achieved 20 months after start-up. On-line time since the plant started up in April 1997 is shown in Figure 3.

c a 1) 0.

loo -I 90 " 80 " 70 -60 " 50 " 40 " 30 -20 " 10 -

I 1—i 1 — i — i — i — i — i 1—i—i 1 — i — i — i — i — i — i — i — i — i — i — i — i — I

97 Feb Apr Jun Aug Oct Dec Feb Apr Jun Aug Oct Dec

1998 1999

Figure 3 - KIVCET On-Line Time (Design = 89%)

Major shutdowns occurred October 1997, July 1998 and April 1999. There were also oxygen supply interruptions in December 1997, May 1998 and October 1999.

Throughput in the KTVCET furnace since start-up is shown in Figure 4. The furnace is now operating at sustained rates above the design capacity of 56 t/h. The main bottlenecks facing the plant in 2000 are elevated radiant boiler outlet temperatures and accretions in the boiler/precipitator areas.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 183

Figure 4 - K.IVCET Furnace Throughput as a Percentage of Design (1208 t/d)

Both on-line time and throughput have benefited from improved accretion control in the furnace. Reduction in accretions in the reaction shaft area has led to better operation of the coke checker. Accretion management has been affected mainly by the throughput in the furnace, proper control of the oxygen potential in the reaction shaft, and proper fluxing targets.

KTVCET was designed to consume all zinc plant residues with the zinc plant operating at 290,000 tonne per year zinc. Residue treatment for 1999 was in balance with production from zinc operations with over 288,000 tonne of zinc produced.

Carbon on charge has decreased from 10-12% to 7-8%. Improved coke checker performance has reduced the necessity of flame reduction. The lower carbon levels have reduced the heat and dust loading which has led to improved boiler and ESP performance.

Reduction of lead has been a main factor in measuring furnace performance. Operation of the coke checker is greatly affected by coke thickness, coke temperature, and the total active non-accreted area. The improved reduction has led to better fume quality and improved slag furnace throughput. This has contributed to the slag furnace operating at approximately 120% of design. Other contributing factors are improved slag chemistry and increased boiler capacity. Figure 5 outlines the fuming performance.

Figure 5 - Fume Produced as a Percentage of Design (160 t/d)

184 LEAD-ZINC 2000

Environmental

The smelter has met its internal environmental objectives. Ambient lead levels in the smelter have decreased from an average of 0.45 mg/m3 to 0.27 mg/m3. The reduced ambient lead levels, combined with a comprehensive Health Protection Plan (HPP), has reduced biological leads in smelter workers from an average of 42 μg/dL in 1990 to 29 μg/dL in 1999. The HPP includes process, engineering, administrative, and housekeeping controls. The HPP was jointly developed with input from management, the union and the Workers' Compensation Board.

Figure 6 outlines the reduction in emissions from Trail operations since the new smelter was started. Figure 7 further outlines lead emissions to air. The reduced lead emissions, as well as increased public awareness, have reduced the geometric mean blood lead levels in the community's children (aged 0.5 to 5 years) from 11.5 μ§/άΧ to 7.7 μg/dL.

Figure 6 - Percent Reduction in Emissions (1996 to 1999)

Figure 7 - Lead Emissions to Air (1990 to 1999)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 185

CONCLUSIONS

After a challenging start-up, Cominco's new lead smelter is a success. It has fully met the requirements for environmental performance, reduced energy consumption, and lowered labour costs. The new smelter has positioned Cominco Trail Operations well for the future and it is expected that further gains in throughput can be achieved through de-bottlenecking.

ACKNOWLEDGEMENTS

The authors would like to thank Cominco Ltd. and Snamprogetti for permission to publish this paper. The assistance of Snamprogetti, Enirsorse, Vniitsvetmet and KAZZINK J. St. Co. during commissioning and start-up of the KTVCET furnace is also gratefully acknowledged. Last but not least, the efforts of the many Cominco employees who worked hard to make the new lead smelter a success must be recognized.

REFERENCES

1. M.J. Walker, "KTVCET Smelter On-stream at Trail", Mining Magazine. Vol. 178, No.4, 1998, 256-263.

2. A.R. Babcock, R.A. Franco, A.C. Mikrovas, S. Bharmal and M.I. Cecchini, "Preparation and Start-up of Cominco's New Lead Smelter", Zinc and Lead Processing. J.E. Dutrizac, G.L. Bolton and P. Hancock, Eds., The Metallurgical Society of CTM, Montreal, Quebec, 1998, 795-809.

3. D.W. Goosen and M.T. Martin, "Application of the KTVCET Smelting Technology at Cominco", paper to be presented at the Mining Millennium 2000 Convention. Toronto, Canada, 8 March 2000.

4. D.W. Ashman, "Pilot Plant and Commercial Scale Test Work on the KTVCET Process for Trail's New Lead Smelter", Zinc and Lead Processing. J.E. Dutrizac, G.L. Bolton and P. Hancock, Eds., The Metallurgical Society of CTM, Montreal, Quebec, 1998, 783-794.

5. E. Jak, B. Zhao, P.C. Hayes, S. Degterov and A.D. Pelton, "Coupled Experimental and Thermodynamic Modeling Studies of the System PbO-ZnO-FeO-Fe203-CaO-Si02-AI2O3 for Lead and Zinc Smelting". Zinc and Lead Processing. J.E. Dutrizac, G.L. Bolton and P. Hancock, Eds., The Metallurgical Society of CJM, Montreal, Quebec, 1998,313-333.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 187

COMMISSIONING AND OPTIMISATION OF THE NEW LEAD AND SILVER REFINERY AT THE PASMINCO PORT PIRIE SMELTER

P. Kapoulitsas, M. Giunti, R. Hampson, A. Cranley, S. Gray and B. Kretschmer

Pasminco Pty. Ltd. Level 7, 380 St Kilda Road Melbourne, 3004 Australia

R. Knight and J. Clark Britannia Refined Metals

Northfleet England

ABSTRACT

In 1997 Pasminco's Port Pirie Lead Smelter began a modernisation of its lead and precious metals refinery operations. Driven by the anticipated closure of its Broken Hill mine, the new refining process replaced the locally developed continuous process, which had been in operation since the 1950's. Commissioning of the new plant was staged, beginning in late 1997, and finishing early in 1999. The new plant provides for batch refining of lead, and the use of vacuum induction retorting and bottom blown oxygen cupel technology in the silver refining area. The project has resulted in an increase of lead production capacity of 11% and allows the treatment of a wider range of concentrates than was formerly possible. The most significant change has been to double the smelter's silver capacity to over 400 tonnes per annum. This paper describes the successful implementation and optimisation of the new refinery, which is now consistently meeting design capacity.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

188 LEAD-ZINC 2000

INTRODUCTION

Pasminco Port Pirie Smelter is part of the Australian company Pasminco which was formed in 1988. It has a smelting and refining history based on the discovery of the rich lead, zinc and silver deposits at Broken Hill, Australia, over 100 years ago.

In 1996 the decision was taken to modernise the lead refining plant to increase production and to allow a wider range of feed materials to be treated. Some historical notes, the reasons leading up to the refinery modernisation and a brief description of the process to be installed have been previously documented (1).

CONSTRUCTION SCHEDULE

The Pasminco Board gave project approval in November 1996. Kinhill Engineers (now owned by Brown and Root) of Adelaide, South Australia were engaged as project engineering, procurement and construction managers with engineering consultancy also coming from the MIM Technology group at Britannia Refined Metals (BRM) in England. As shown in Table I, preliminary engineering had commenced in early September 1996 and a construction timetable was scheduled to ensure commissioning of all major plant and equipment by mid 1998, to coincide with the expected delivery of new high silver concentrates and extra crude lead bullion to Port Pirie. The schedule and co-ordination of the construction phase was developed and monitored to ensure key dates were maintained and that all operations, both construction and lead refining, continued in a safe manner. The 22-month project implementation period tested the resolve of the engineering, construction and operating teams as they became familiar with unseen equipment and overcame the challenges of operating the existing plant and building the new one in and around those operations. It was a significant achievement in that, although there was some movement in the construction of some items of plant, all key units were available for commissioning to begin on the dates originally scheduled. Production losses were limited to those originally anticipated and the works were completed with no major safety issues arising.

Table I - Pasminco Port Pirie Smelter - Refinery Modernisation Key Construction and Commissioning dates.

Date Key Activity

September 1996 Start design and construction timetable May 1997 Construction started September 1997 Liquation furnaces commissioned November 1997 First VIR & BBOC commissioned May 1998 Second VIR and all PMR commissioned June 1998 Single stream batch refining commissioned October 1998 KB A commissioned November 1998 Dual stream batch refining commissioned February 1999 New crane commissioned

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 189

The first step was to demolish the existing continuous desilverising kettle (CDK) maintenance shed at the northern end of the refinery to begin construction of four liquation kettles. This allowed enrichment of the existing CDK high grade alloy, thus increasing throughput to the precious metals refinery (PMR) when three of the existing five Faber de Faur retorts were demolished and replaced by one of two vacuum induction retorts (VIR's) and the bottom blown oxygen cupel (BBOC). Construction of the second VIR furnace and the parting plant upgrade commenced once the remaining two retorts and the Sirosmelt cupellation furnace were removed.

Essentially one stream of the old continuous refinery plant was demolished to make way for the installation of five 420 tonne pans, allowing batch refining to commence at reduced capacity. It was a significant achievement that single stream batch refining was successfully commissioned while operators lacked access to walk fully around three of the new pans.

Process Description

Lead Refinery

A simplified process description of the new batch lead and silver refinery is shown in Figure 1. De-copperised lead bullion is transferred to the refinery by rail in an 18 tonne ladle. The bullion is accumulated in a 420 tonne pan at 460°C prior to transferring to one of two softening pans. In softening, the bullion is heated to 570°C over a 3 hour period at which point an 85% oxygen/air gas mix is injected into the bath through two lances and ignition commences. The bullion temperature is allowed to increase in a controlled manner over the course of the blowing cycle to reach a maximum of 680°C. During the blowing cycle, the slag produced is periodically removed. Blowing is terminated when the antimony assay of the bullion is determined to be less than 0.07%. At this point, the lances are withdrawn and the bullion is force cooled to 460°C over a 6 hour period. Stirring also assists cooling.

Desilverising is modelled on the Britannia Refined Metals (BRM) desilverising process carried out at Northfleet, United Kingdom. Softened bullion, containing 0.2-0.3% Ag, is transferred via a submerged pipe entry to the first stage desilverising pan, and is mixed with low-grade crust from the previous batch. The resultant Parkes crust, containing 6.5% Ag, 17% Zn and entrained lead, is skimmed into one tonne moulds and directed to the liquation kettles. Fresh zinc is then added, mixed in to the bath, and the bullion transferred to the second stage desilverising pans. Zinc additions vary according to the Ag content of the softened bullion and are adjusted in order to maintain the Ag/Zn ratio of the skimmed crust at 0.37-0.40. In second stage desilverising, the crust from the previous batch is allowed to soak at 460°C for 4 hours in the new batch of bullion, after which it is thoroughly mixed into the bath. The new crust formed is skimmed and returned to the empty first stage pan for reaction with the next batch. This counter-current recycling of low Ag crusts enables better zinc utilisation and has resulted in a net reduction of new zinc consumption of approximately 1000 tonnes for twice the silver production. After crust removal, the second stage bullion is allowed to cool to 320 °C by slow stirring and forced cooling. The bullion, which now contains 0.54-0.56% Zn and less than 6 g/t Ag, is transferred to the vacuum de-zincing pans, while the low grade crust remains in the pan and becomes part of the next batch.

190 LEAD-ZINC 2000

Vacuum de-zincing takes about 15 hours and is carried out using BRM style vacuum de-zincing domes. The bullion is heated to 500°C, dressed, and the dome placed in position. Heating is continued and vacuum is applied once the temperature reaches 590°C. De-zincing continues for 5-6 hours until the zinc level in the bullion falls to 0.04 %. The de-zinced bullion is then subjected to caustic refining after which it is moulded as market metal or further refined in the debismuthising plant.

The debismuthising plant has been fully described elsewhere (2) and the process metallurgy was not changed during the modernisation project. Opportunity was, however, taken to increase the size of the alloying and debismuthising kettles, to improve plant layout and to fully automate the burner and cooling systems for improved temperature control. This has resulted in the capacity to debismuthise all the production if required.

Silver Refinery

Silver refining to dore specification is modelled on BRM practice, and is followed by a conventional parting plant operation employing Balbach Thum cells.

The Parkes crust is liquated in four cast iron kettles at 650°C, utilising the miscibility gap in the Ag-Zn-Pb system. The upgraded alloy, which contains approximately 25% Ag, 10% Pb and 65% Zn, is tapped into 350 kg moulds and is delivered to the precious metals refinery. Low-grade alloy from the liquation process is returned to the first stage desilverising pan.

Zinc contained in the high-grade alloy is extracted by distillation in one of two Junker Gmbh vacuum induction retorts (VIR's), using the Union Miniere (UM) licensed process (3). Distilled zinc is tapped from the liquid zinc condensers into 250 kg moulds and is returned to the de-silverising process. The VIR bullion, which now contains approximately 65% Ag, 25% Pb and less than 4% Zn, is upgraded to silver grade dore in a bottom blown oxygen cupel (BBOC) furnace. Zinc, lead and copper impurities are preferentially oxidised in that order by submerged injection of oxygen into the bath at 700°C. The BBOC slag, which consists essentially of ZnO and PbO, is returned to the blast furnace. The silver dore is poured into 10 kg anode plates on an adapted Outokumpu casting machine and is transferred to the electrorefining plant.

Dore is electrorefined in conventional Balbach Thum cells to produce 99.99% silver cathode which is then cast or granulated according to market requirements. Anode slimes are periodically recovered from the cells and are leached to produce a 98% gold product.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 191

Lead Bullion 0.7 % Sb 0.4 % As 0.2 % Ag I

Q2/Air Softening

Softener slag w

Zn

l l l k

Desilverising

VDZ Zn

Parkes alloy

Liquation slag

Low grade alloy -<

\

Liquation

High grade Alloy

Dezincing

J\.

NaOH, NaN03

VIRZn

Caustic Refining

Caustic dross ►

Moulding (99.97 % Pb)

O,

Vacuum Induction Retorts

VIR dross

Bottom Blown

Oxygen Cupel

BBOC Slag

t Dore

♦ Parting Plant

Au Ag

Figure 1 - Flowsheet of Pasminco's Port Pirie Smelter New Batch Lead and Silver Refinery

192 LEAD-ZINC 2000

COMMISSIONING

Commissioning of the plant was staged to begin as soon as the various key items of equipment had been installed, culminating in the sequential operation of all batch refining steps in the overall refining process. This proved to be a significant challenge for all operating staff, previously used to the operation of a continuous process. Issues associated with scheduling and sequencing of each step of the operation, development of skills needed to break up and skim the desilverising crusts and the orderly transfer of lead between the refining pans using pumps and pipes needed to be overcome. A commissioning team comprising technical and operating personnel enabled significant progress to be made during this phase of the project. Representatives of BRM, UM, and Brunswick Mining and Smelting (BM&S) also provided on-site assistance at various key times.

Prior to commissioning, key technical, maintenance and operations personnel travelled overseas to BRM Northfleet and UM Hoboken to be trained in the new operations. This enabled a better understanding of the operating and maintenance requirements of the new processes and allowed key job safety procedures to be developed for training of additional operators before commissioning began.

ead Refinery

Softening

Based in part on the batch softening process used by BM&S (4), the softening process incorporates several unique design features which make it simpler, cleaner and more intense than the process on which it was based. In particular, the lance design has been simplified and the availability of tonnage oxygen has allowed the oxygen level in the reaction gas to be increased to higher levels than used by BM&S. The lance configuration was developed during the design stage at Port Pirie and further modified during the commissioning period, during which time the level of oxygen in the reaction gas was increased from 50% to over 85%. This allowed the number of lances to be lowered from four to two, which has provided benefits in reduced pan wear. Lance life is currently of the order of 2-3 weeks. The move to higher oxygen levels also resulted in reduced agitation in the reaction pan, leading to reduced fume evolution and less metal entrainment in the slag.

The blowing time required is predicted by an empirical equation relating gas rate to bullion assays. Developed during the commissioning phase after trials with various gas flowrates and 02/Air ratios, this equation is now used by the operators to control the blowing time.

The original process concept provided for bullion at the end of one blowing cycle to be cooled, with a final clean up using a small addition of caustic soda before pumping it to the desilverising pans. The solid dross was to be skimmed with a BRM style dedrosser. On commissioning, it was evident that this operation lengthened the overall cycle time such that the plant throughput could not be achieved. As commissioning progressed, it was found that the target softening levels could be reached during the blowing stage, without need of caustic soda addition. It was also found that the solid slag crust formed during the later stages of blowing could be absorbed by the richer slag formed during the first two hours of blowing of the next batch. This allowed the de-drossing step to be eliminated. Slag is displaced from the reaction

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 193

pan by adding extra bullion in a controlled manner during the blowing cycle. The slag is tapped onto a water-cooled oscillating tray and is collected in a bunker for further treatment.

Desilverising

The batch desilverising stage represents the core operation in Pasminco's new batch refining process. The process is essentially an isothermal 2-stage batch operation and is illustrated schematically in Figure 2. The metallurgy of silver removal from lead using zinc has received lengthy attention, and excellent reviews on the subject can be found elsewhere.

Figure 2 - Pasminco Port Pirie Smelter Batch Desilverising Flowsheet

Parkes crust is removed using an overhead crane and a custom made scoop. Early problems with high lead entrainment, detrimental to liquation performance, were eliminated as operators developed improved techniques of crust removal, and by increasing the settling time after mixing. Another problem encountered was an excessive build up of the cooling crust in the second stage desilverising. This resulted in unacceptable times being required for soaking and mixing. This issue was settled as a better understanding of the zinc balance around the process was developed and consistent operation with respect to crust removal and temperature control was achieved.

During the early stages of commissioning, transfer problems between the first and second stage desilverising were encountered, resulting in substantial batch cycle delays. This was found to be caused by slower than expected pumping rates which allowed Ag-Zn crystals to form at the entry of the pump and inside the transfer pipe. Batch pump over times were designed to be completed in 30 minutes, and required a transfer rate of 640 tonnes per hour. In practice the transfer rates achieved were only 400 to 450 tonnes per hour, contributing to the blockage problems as well as adding to the overall cycle times. These problems lead to an investigation whereby the mixing step immediately following the addition of zinc was avoided. These trials proved successful and resulted in the elimination of pump and pipe blockages, and reduced the first stage batch cycle time. The new operating method, however, changed the overall zinc balance, and resulted in greater levels of zinc being directed to liquation, and a

194 LEAD-ZINC 2000

reduction in the amount of zinc being returned from second stage desilverising. As a result, after about 60 batches, the residual silver content in the cooled bullion rose from 4 g/t to over 150 g/t while the zinc to silver ratio at the beginning of the cooling stage had also been found to be reduced.

To offset these effects, computer modeling suggested that extra zinc additions should be made in the second stage desilverising operation, while the amount of zinc added in the first stage is reduced accordingly. Residual silver levels quickly returned to the expected values and the net zinc addition was kept constant, such that the new operating procedure was adopted as standard.

During mid-1999 a design engineering study was initiated to consider means for increasing the pumping rate to 650 tonnes per hour. This would further reduce cycle times as well as allowing the opportunity to revert to the BRM style zinc addition procedure. A shortfall of the current procedure and the controls that are in place is that it is still somewhat difficult to predict when the process is nearing a zinc depletion phase. As a result of this phenomenon, batches of product are occasionally produced which are out of specification. At the time of writing this paper, trials using an alternative pump design had just begun.

Another issue which culminated in several problems becoming apparent on start up resulted from the need to design and retrofit the new plant around the existing continuous process which was housed in two adjacent sheds, one running North-South and the other East-West. Firstly, this meant that the most critical pumping procedure involving the transfer of cold lead (320°C) from the second stage desilverising pan to the two vacuum dezincing (VDZ) pans occurred over the change of shed direction. In this area, overhead crane access was limited and a semi-permanent pipe which incorporated gas trace heating had to be set up which could be easily moved from one VDZ pan to the other. A tight bend after the pump pipe cleared the desilverising pan also attracted a lot of attention by the operators who would install a large burner in this area to avoid blockages. However, this caused frequent damage to the pump motor and power lead, and also created a hazard for operators coupling the heated transfer pipe to the pump. However, after a short time the operators became very confident in this procedure such that trace heating is now only applied to the straight section of the pipe and at a much lower intensity.

A study of crane operations during the process design phase indicated that two cranes would be sufficient to meet the needs of the operations. As the continuous process had utilized two cranes, one servicing each shed, it was reasoned that the process would be adequately covered. However as design engineering progressed, and especially during the early weeks of commissioning, it became evident that the target cycle times, and thus design throughput, could not be met because of a higher than anticipated number of crane movements in the North-South shed. Approval to install a second overhead crane in the North-South shed was obtained, and the design and installation were fast tracked to enable the new crane to be commissioned by the end of February 1999.

The throughput rate of the new plant is measured by the total time taken for first stage desilverising, as all metal must go through this one pan. Table Π shows the cycle times achieved during the various stages of plant commissioning. It can be seen that the target cycle time of less than 10 hours was achieved soon after the second crane installation, once operators became accustomed to sequencing the operations of the two cranes in the North-South shed.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 195

Table II - Pasminco Port Pirie Smelter - Refinery Modernisation First Stage Desilverising Cycle Time

Commissioning Period Hours

Immediately after start of single stream 20.0 Optimisation of single stream 16.0 Immediately after start of dual stream 14.0 Optimisation of dual stream 12.0 After second crane installed <10

To date, the batch size achieved has been only 300 tonnes compared to a design of 320 tonnes. It is thought that the lack of pump head capability, which is also affecting the pump capacity, is the major contributing factor. It is expected that a further plant capacity increase will be possible once transfer rates of over 650 tonnes per hour are achieved.

Other issues experienced during the optimisation stage were poor stirrer shaft life and a high rate of consumption of thermocouples. The change in zincing procedure improved the stirrer life, and thermocouple life was improved by an increased focus on operator procedures; however, room for improvement in both areas remains.

Dezincing

The vacuum dezincing stage is based on BRM's VDZ process and features a water-cooled flat surface condenser with an integral stirrer and rubber seal. The rim of the pan is water cooled and further protected by a solidified layer of condensed zinc. Commissioning of the vacuum dezincing unit was a relatively easy task, and to date, no major issues affecting production have become apparent. Some initial problems with thermocouple failures were addressed by using alternate materials of construction. The VDZ dome life has also been variable and target BRM life has not consistently been achieved because of the excessive build up of dross in the stirrer shaft and shroud. Various options are being considered in this area, and regular consultation with BRM personnel has been maintained in an effort to develop an appropriate solution to the problem.

Debismuthising Plant Upgrade

Pasminco's Kroll-Betterton antimony debismuthising plant has been in operation for almost 20 years and the process has been well documented. The upgrade was carried out simultaneously with the batch refinery stages. Modifications to kettle size, burner control systems, and equipment used to dissolve the calcium and magnesium reagents and to remove the bismuth crusts have all enhanced the ability of the plant to run at the design rate of up to 40 t/h, and remove bismuth from input bullion at 0.05-0.06% down to <0.005% Bi.

Additional refining pan capacity has improved plant flexibility and enabled a range of soft lead and alloy products to be manufactured more effectively. Mechanical dross removal units fitted to the refining pans have minimised the need for manual dross removal, while considerably improving the hygiene conditions within the plant area. Savings in calcium and magnesium reagent consumption have been noted in the order of 8-9% compared to the previous operation.

196 LEAD-ZINC 2000

Silver Refining

Liquation

There were four main areas of difficulty encountered in commissioning and optimising the operation of the liquation kettles:

• excessive dross make and silver losses • scaling of the surface of the kettles • poor burner control • alloy accretion growth in the kettle.

The kettles were initially commissioned to treat 'thick' CDK alloy from the continuous desilverising kettle of the old refinery. The alloy was well oxidised by the time it was treated in the liquation circuit and thus, required additional fluxing of the dross by addition of a calcium/sodium chloride. This practise was abandoned early in the commissioning stage because of the prohibitive amount of flux required and the lack of capacity in the kettle to accommodate this quantity. During this period, wet dross containing a significant quantity of silver was produced which was recycled through the blast furnace. When batch-refining operations were commissioned, silver crust quality improved and the practice of fluxing the kettles was successfully re-instituted.

When first put into service, the surfaces of the cast iron kettles (which are made of grey cast iron of a composition chosen for its resistance to zinc attack), slowly oxidised to form a mill scale coating on the upper section of the kettle. This resulted in poor kettle life, reduced treatment rate of the Parkes crust, increased gas consumption, increased heat losses to the operating environment and poor life of other associated equipment around the liquation kettles. The scaling of the kettle surface was determined to be a result of the operation of the burners with an oxidising flame, and chloride flux attack caused by overflow when charging. The problem has been mostly overcome by tuning the burners to operate with sub-stoichiometric air / gas ratios to produce a slightly reducing flame, and by better procedures for fluxing.

Control over the firing rate of the top burners on the liquation kettle was also improved via trials using direct measurement of the temperature of the high grade alloy layer though an alloy resistant thermocouple sheath. This method of control produced a marked increase in crust treatment rates, provided gas savings and gave better control over the high grade alloy temperature; it has been adopted as standard practice.

Formation of alloy shelves and pads resulted in restrictions to the discharge of the low-grade alloy from the kettles and resulted in a tendency to overflow the kettles as they were charged. In the worst instances, there was a loss of connection to the siphon pipe which prevented low-grade alloy from being discharged. In both cases, the result was for low grade alloy to be tapped from the high grade alloy tapping spout and to report to the vacuum induction furnaces, where it reduced crucible life and depressed the silver grade of the VIR bullion product. Shelving of the kettles was noticed to be particularly predominant when the Parkes crust butts were excessively large, often weighing up to 1400 kg. When the mass of the butts was standardised and reduced to around 1000 kg, the operation of the kettles was stabilised and the formation of shelves ceased.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 197

The formation of alloy pads is thought to be a result of slow accretion build up, and is possibly related to thermal cycling which occurs at the base of the kettle while it is in operation. As the low-grade alloy cools, it rejects silver and zinc in a solid phase some of which forms on the wall and floor of the kettle. This build up over time restricts the flow of low-grade alloy from the kettle resulting in the discharge of the low-grade alloy from the high grade tapping spout. The occurrence of alloy pads was reduced by firing the bottom burner only when required. Accretion of the kettles is still a problem but is managed by using the 'heat treatment' procedure which involves firing the kettle to a uniform temperature of about 600°C. Generally this is enough to remove any restrictions.

Silver recovery at liquation is a major consideration of the kettle operation. As discussed earlier, large losses of silver (as high as 30%) to the wet dross occurred when the kettles were first operated without flux. Reintroduction of fluxing procedures, when the kettles began to treat the Parkes crust from the batch refinery, resulted in a significant increase in silver recovery, to greater than 85%.

Vacuum Induction Retorts

Long distillation times and frequent crucible failures made the VIR furnace the bottleneck for silver production in the early days of operation. The main issues dealt with through the commissioning and optimisation of the VIR furnaces were:

• Excessive zinc distillation times caused by high zinc levels in the feed, and high dross carryover in the feed

• Crucible failure because of the high lead content in the alloy from the liquation furnaces • Poor detection of the distillation end point • Poor vacuum.

The first of the VIR furnaces was commissioned well before the batch refinery, using feed that was a mixture of the CDK 'liquid' alloy and unfluxed liquated CDK 'thick' alloy, to produce a charge typically of composition 16% Ag, 75% Zn and 7% Pb. The high level of dross carry-over in high grade alloy from the liquation kettles and the higher fraction of zinc than would be typical of the Parkes crust for which the process was designed, resulted in distillation times of the order of 14 to 15 hours.

Carry over of low-grade liquation alloy (predominately lead) with the upgraded Parkes alloy was found to have a detrimental effect on the VIR performance. On melting this material in the VIR furnace, the two alloy layers would once again separate with the denser low-grade alloy settling to the bottom of the crucible. Over the course of the distillation, the two recombine to form silver-lead bullion with a much higher lead content. The result is reduced crucible life, because of increased lead penetration through the porous clay graphite crucible, and a reduction in the potential silver production capacity of the plant.

Assessing the end point of the distillation process was initially difficult with much variability in the level of zinc in the product bullion. Ultimately a technique from Union Miniere, where the .final temperature of the distilled zinc, furnace vacuum and the zinc assay from the previous charge are used to predict the endpoint of the next charge, was adopted as standard procedure once the new batch refinery was commissioned. This was found to reliably predict the endpoint of the distillation phase.

198 LEAD-ZINC 2000

The most critical aspect of operating the VIR furnaces successfully is the level of vacuum that is applied to the furnace. There are a number of potential leakage points where air ingress can reduce the level of vacuum applied to the furnace, leading to increased and variable distillation times and unpredictable levels of zinc in the product bullion. Higher metal temperatures are also required for distillation of the zinc. Both factors cause accelerated crucible wear. Poor vacuum also results in loss of zinc as a zinc oxide dross, which subsequently causes blockage problems as it accumulates in the condenser and on the porous brick filters. This further reduces the vacuum and extends the distillation time. Initial vacuum problems were found to result from poor mechanical seals that were generally attributable to the use of under sized or incorrect o-rings, poor technique when closing flanges, and poor installation of the porous brick filter. Tests using blanking plates to progressively prove the vacuum over the system after rebuilding a furnace were devised, assisted by the use of a directional ultrasonic leak detector. It is now common practice to perform vacuum tests when restarting the furnaces. Poor vacuum is not tolerated and the furnaces are shutdown to identify and to rectify leaks as they are detected. Extended periods of operation with vacuum levels of 80 mbar or less are now typical.

After two years operation of the VIR's, the major issue still to be overcome is poor crucible life. Expected target life of over 40 charges has not yet been achieved despite using a variety of crucible types.

Bottom Blown Oxygen Cupellation (BBOC)

BBOC commissioning issues arose from poor VIR bullion quality, with both zinc and lead levels being higher than their respective target maxima of 2% and 20%. Mechanical problems were largely associated with jamming of the oxygen lance and moulding of the dore product.

Early commissioning trials showed that the level of zinc tolerable in the VIR bullion was dependent on the amount of lead in the bullion. Zinc oxidises preferentially to form a solid oxide dross on the surface of the bath, which is then progressively fluxed by the litharge formed as the lead oxidises. The viscosity of the resultant slag affects the degree of silver metal entrainment and was found to be dependent on the initial zinc/lead ratio in the charge. Silver recovery to dore at the BBOC furnace was poor until liquation and VIR operations were brought under control.

Higher than planned zinc input to the BBOC also caused increased refractory wear and excessive lance consumption because of its more strongly exothermic reaction compared to lead. Furnace cycle times through the commissioning period were longer than predicted as a result of the higher zinc content (more frequent replacement of the lance) and the increased lead content of the feed (increased oxidation duty on the furnace) which was a result of the liquation performance. Once the batch process was bedded down, allowing the VIR to produce a lower zinc and lead product, the above issues were resolved and cycle times were reduced to BRM design figures.

Good furnace operation is reliant on proper operation of the mechanical lance feeding system. This can be quite prone to jamming, an occurrence which increases cycle times and causes excessive refractory wear as the lance burns back into the furnace base. Anecdotal evidence suggests that this is more common with higher zinc charges or if the molten bath is close to freezing when the lance is changed.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 199

The BBOC unit has proven to be quite capable of reducing dore impurity levels to as low as 0.01% Pb and less than 0.1% Cu at a temperature of approximately 1050°C. However, the need to cast BBOC product into 10 kg plates for subsequent electrorefining generally requires it to be heated to 1100-1150°C. Under these conditions, dore typically contains around 0.2% Pb and 0.6% Cu. This practice has resulted from the adaptation of an existing dore casting conveyor into which metal was manually spooned from a reverberatory furnace. Because the dore is so thermally conductive it tends to freeze before reaching the moulds if not given enough super heat. The higher levels of impurities in the dore have increased the required purification duty in the electrorefining plant. Design of an improved casting machine has been initiated, and this is expected to result in improved safety and production performance, and to enable the handling of the dore at lower temperatures.

Parting Plant

Prior to the refinery modernisation, the parting plant consisted of 80 Balbach Thum cells, and produced up to 220 tonnes of silver per year. Plant scale trials had shown earlier that a 450 tonne per year capacity was achievable through the same number of cells. Wiring, cell switching and electrical distribution systems were upgraded as appropriate to allow operation at the higher current and at a higher silver concentration in the electrolyte, and the plant is now approaching design capacity. There are, however, issues to be overcome that are associated with the increased number of dore plates to be handled, and the increased requirement to rake cathode silver from the cells. A project has been initiated to address these areas of concern.

CONCLUSIONS

Over a period of just two and a half years the Pasminco Port Pirie Smelter lead and silver refining operations have been replaced by new and upgraded processes capable of producing 250,000 tonnes per annum of refined lead and 450 tonnes per year of refined silver. The successful design, construction and commissioning phases of the plant were completed on time and within original budget. Original targets and results achieved are shown in Tables ΙΠ andlV.

Table III - Pasminco Port Pirie Smelter - Refinery Modernisation Target Comparison - Lead Refinery

Unit Operation Target Achieved

First Stage Desilverising 10.3 hours 8.2 hours Second Stage Desilverising 17.0 hours 17.0 hours Vacuum Dezincing 17.0 hours 14.5 hours Batch size 320 tons 310 tons Zinc consumption 540 tpa 490 tpa

200 LEAD-ZINC 2000

480 480 480 450

580 600 650 415

Table IV - Pasminco Port Pirie Smelter - Refinery Modernisation Target Comparison - Silver Refinery

Unit Operation Target (tpa) Achieved (tpa)

Liquation VIR BBOC Parting Plant

In addition to the criteria listed above, significant improvements in hygiene and occupational safety conditions have also been achieved in the new lead and precious metals refining plants, with minimal fume emissions and the eradication of nearly all of the heavy manual workload characteristic of the old plant. A significant reduction in zinc consumption compared to the previous process has been observed. In addition, the recycle rate of dross material has been significantly reduced, and the overall predicted increase in plant capacity was achieved. These achievements along with the production capacity increases were accomplished with virtually the same labour compliment as with the old plant.

ACKNOWLEDGMENTS

The authors wish to thank the management of Pasminco Port Pirie Smelter for the permission to publish this paper. The assistance of R. Knight and J. Clark of BRM and K. Vertongen and Y. Mampey of UM through the initial commissioning phase is also duly acknowledged.

REFERENCES

1. B. Kretschmer, G. Burgess and I. Sanderson, "Recent changes to Pasminco's Lead Smelter at Port Pirie", Zinc and Lead Processing. J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton and P. Hancock, Eds., The Metallurgical Society of CIM, Montreal, Canada, 1998,455-469.

2. T. Banytis, C.A. Waters and J.V. Happ, "Debismuthising of Lead at the BHAS Lead Refinery, Port Pirie, SA", The Aus.I.M.M. North Queensland Branch, Smelting and Refining Operators Symposium. May 1985, 77-85.

3. P.R.M. Van Negen and P. Weertes, "Vacuum Dezincing of Parkes Triple Alloy in an Induction Furnace ", Electric Furnace Techniques in Metallurgy. Bad Salzdetfurth, 19-20 March, 1987,169-176.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 201

4. H. Perez, R. Sinclair, P. Hancock and J. G. Lenz, "Recent Process Intensification Efforts at the Brunswick Lead smelter", Challenges in Process Intensification, C.A Pickles, PJ. Hancock and J.R. Wynnyckyj, Eds., The Metallurgical Society of CIM, Montreal, Canada, 1996, 247-265.

Chapter 3

Zinc Operations I

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 205

ZINC PRODUCTION - A SURVEY OF EXISTING SMELTERS AND REFINERIES

S.E. James Big River Zinc Corporation

2401 Mississippi Avenue Sauget, Illinois, U.S.A. 62201

J.L. Watson and J. Peter Department of Metallurgical Engineering

University of Missouri-Rolla Rolla, Missouri, U.S.A. 65409

ABSTRACT

A survey of operating zinc smelters and refineries throughout the world has been done in conjunction with the Lead-Zinc 2000 Symposium. This survey follows the examples set by the Japanese Mining Industry Association in 1985 and 1995. Questionnaires were sent to operating zinc plants that covered feed materials, process details, energy consumption and labor requirements. The voluntary responses to the survey were tabulated. Key operating parameters including feed grade, production capacity, quality of the intermediate products and quality of the final products have been compared for both hydrometallurgical and pyrometallurgical processes.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

206 LEAD-ZINC 2000

INTRODUCTION

In 1985 and 1995, the Japanese Mining Industry Association presented papers entitled "A Survey of Operating Situation of Smelters and Refineries in the World" (1,2). These surveys have proved extremely useful in documenting trends and changes within the zinc industry. In this symposium, the tradition of examining the state of the zinc industry is continued. A list of zinc smelters and refineries was compiled from data from the International Lead Zinc Study Group (ILZSG) (3) and from the book "Zinc and its Markets" (4). A questionnaire was prepared covering the following topics:

• Metal production capacity • Types of metal products made • Types of feed materials used • Zinc concentrate roasting • Sulfuric acid production • Calcine sintering • Smelting and thermal refining • Leaching • Purification • Electrolysis • Casting • Distribution of labor.

Unfortunately the 1985, 1995 and 2000 surveys have different populations of plants. Hence, making direct comparisons among the survey results is difficult. However, it should be possible to identify operational changes and trends in many areas.

SURVEY RESULTS

Responses

Fifty surveys were sent to individual smelters and refineries. Twenty-two plants responded to the survey, or 40 % of the plants contacted. This compares to 41 responses to the 1985 survey and 32 responses to the 1995 survey. The names of the plants that responded appear in Table I. Table II lists all the operating smelters and refineries and their production capacities as compiled by Brook Hunt (5). The respondents to the 2000 voluntary survey represent a total production of 2,825,000 t/y or 42.6 % of the total world zinc production.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 207

Table I - Smelters Participating in the Voluntary Survey

Smelter/Refinery Sumitomo Metal Mining Korea Zinc, Kidd Metallurgical Division Pasminco Shaogun Smelter Industrial Minera Doe Run Outokompu Zinc Zinc Corporation Portovesme, JSC Chelyabinsk KCM Hachinohe Binani Zinc Toho Zinc Cominco Pasminco-Clarksville Big River Zinc Hudson Bay Mining & Smelting Canadian Electrolytic Zinc

Country Japan Korea Canada Australia China Mexico Peru Finland South Africa Italy Russia Bulgaria Japan India Japan Canada USA USA Canada Canada

Table II - World Zinc Smelters and Refineries

Smelter Springs Zhuzhou Huludao (VR) Shaoguan Huludao (EL) Baiyin Changsha Shenyang Onsan Sokpo Tak Chanderiya Debari Kayseri Vizag Binanipuram Risdon Cockle Creek Chelyabinsk

Total Slab Zinc Production for 1999 (kt)

107 249 210 131 68 95 20 17

288 102 88 54 49 36 35 33

200 78 119

Country South Africa

China China China China China China China Korea Korea

Thailand India India

Turkey India India

Australia Australia Russia

Region Africa Asia Asia Asia Asia Asia Asia Asia Asia Asia Asia Asia Asia Asia Asia Asia

Australia Australia

Eastern Europe

Process Type Electrolytic Electrolytic

Vertical Retort ISF

Electrolytic Electrolytic

Vertical Retort Electrolytic Electrolytic Electrolytic Electrolytic

ISF Electrolytic Electrolytic Electrolytic Electrolytic Electrolytic

ISF Electrolytic

208 LEAD-ZINC 2000

Total Slab Zinc Production for

Smelter Vladikavkaz Miasteczko Slaskie Boleslaw Plovdiv Szopienice San Juan De Nieva Auby Budel Balen Kokkola Odda Nordenham Avonmouth Crotone Datteln Duisburg Noyelles-Godault Porto Vesme (EL) Porto Vesme (ISF) Cartagena Iijima Annaka Hachinohe Harima Hikoshima Kamioka Torreón Cajamarquilla Tres Marias San Luis Potosí Juiz De Fora La Oroya Itaguai Trail Valleyfîeld Kidd Metallurgical Division Monaca Clarksville Flin Flon Sauget Total

1999 (kt) 75 77 68 58 31 308 232 217 209 200 137 133 79 91 103 76 92 75 68 44 179 122 94 74 70 61 129 113 108 101 69 72 4

274 226 145

147 99 99 95

6635

Country Russia Poland Poland

Bulgaria Poland Spain France

Netherlands Belgium Finland Norway Germany

United Kingdom Italy

Germany Germany France Italy Italy Spain Japan Japan Japan Japan Japan Japan

Mexico Peru

Brazil Mexico Brazil Peru

Brazil Canada Canada Canada

USA USA

Canada USA

Region Eastern Europe Eastern Europe Eastern Europe Eastern Europe Eastern Europe

Europe Europe Europe Europe Europe Europe Europe Europe Europe Europe Europe Europe Europe Europe Europe Japan Japan Japan Japan Japan Japan

Latin America Latin America Latin America Latin America Latin America Latin America Latin America North America North America North America

North America North America North America North America

Process Type Electrolytic

ISF Electrolytic Electrolytic Electrolytic Electrolytic Electrolytic Electrolytic Electrolytic Electrolytic Electrolytic Electrolytic

ISF Electrolytic Electrolytic

ISF ISF

Electrolytic ISF

Electrolytic Electrolytic Electrolytic

ISF ISF

Electrolytic Electrolytic Electrolytic Electrolytic Electrolytic Electrolytic Electrolytic Electrolytic Electrolytic Electrolytic Electrolytic Electrolytic

Electro thermic Electrolytic Electrolytic Electrolytic

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 209

Plant Capacity and Production

Table III presents the 1999 capacities and productions of the surveyed zinc smelters and refineries. The total capacity for the 20 plants was 3.2 million t/y of zinc, and the production rate was 2.8 million t/y. The average reported plant production was 141,250 t/y. The major by-products are also given in Table III. Over 90% of the operations produced cadmium and sulfuric acid. Other important by-products include copper and gypsum.

Table III - Zinc and By-product Production

Zinc Capacity (t/year) Zinc Production (t/year) Cadmium (t/year) Copper (t/year) Gypsum(t/year) Sulfuric Acid (t)

No. of Plants

20 20 19 12 5 18

Minimum

30,000 29,500

46 301 150

42,000

Maximum

422,000 330,500

1,350 3,830 68,000

440,000

Average

161,000 141,250

430 1,660

24,500 224,000

Total

3,219,000 2,825,000

8,160 19,900 122,650

4,030 ,000

The distribution of the overall zinc recovery from the feed materials appears in Figure 1.

-■Ill 85- 87.5- 90- 92.5- 95- 97.5-87.5 90 92.5 95 97.5 99

% Zinc Recovered

c iS 0. o o .O

E 3 Z

Figure 1 - Distribution of Overall Zinc Recovery from the Feed Materials

210 LEAD-ZINC 2000

Feed Materials

Table IV summarizes the major constituents in the sulfide and oxide concentrates, and in the secondary materials processed at the zinc smelters and refineries. Sulfide concentrates represented 90% of all the feed materials of the 20 plants surveyed. The sulfide concentrates had higher iron contents, whereas the secondary zinc material had significantly higher cadmium and lead contents than the other feeds. Figure 2 illustrates the relationship between iron and zinc in the sulfide concentrates.

Table IV - Plant Feed Data

ZnS Concentrates (t/year) %Zn %S %Fe %Pb %Cd %Cu

g/tAg ZnO Concentrates

(t/year) %Zn %Fe %Pb %Cd

g/tAg Secondary ZnO

(t/year) %Zn %Fe %Pb %Cd

g/tAg

No. of Plants

18

19 18 18 19 19 19 13 3

3 3 3 3 2 12

12 8 8 9 5

Maximum

51,400

41.9 27.4 1.4 0.5 0.13 0.11 3.5

5,400

55.0 0.02 0.03 0.001

12 1500

21.0 0.1 0.2

0.01 3

Minimum

550,00

62.1 33.5 11.0 14.0 0.58 1.53 380

70,000

77.8 0.88 20.0 0.55 31

275,700

84.0 4.4 16.0 8.5 180

Average

271,000

52.3 31.5 7.5 2.5 0.22 0.57 120

27,350

64.3 0.47 7.7

0.21 21.6

36,200

55.4 1.4 6.2 1.2

79.8

Total/y

4,879,000

82,000

434,600

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 211

4 6 8 10 % Fe in Concentrates

12 14

Figure 2 - Relationship Between Zinc and Iron in Sulfide Concentrates

Roasting

Table V presents the operating and equipment data for 37 roasting units operated by 16 plants.

Table V - Roasting Operation Data

Hearth Area (m2) Air Vol (Nm3/h) %Oxygen Feed Rate (t/h) Operating Temp °C Calcine: S as Sulfide Calcine: Total S

No. of Units 37 37 19 38 38 33 33

Minimum 6

725 1

0.8 900 0.2 0.9

Maximum 130

86,400 35

26.2 1030 0.6 2.8

Average 48.6

27,300 16.9 13.1 948 0.3 2.0

Sintering

The survey data for 5 sinter units operated by 5 plants is given in Table VI, and of the 5 units four are updraft and one is downdraft.

212 LEAD-ZINC 2000

Table VI - Sintering Operation Data

Tons/month Sinter Zn% Sinter S% Sinter Fe% Sinter Pb%

Acid Plant

(capacity) No. of Units

5 5 5 4 5

Minimum 17,000

5.4 0.5 8.3 18.1

Maximum 27,000 46.8 2.0 11.2 45.3

Average 20,000

30.9 0.9 9.8

29.2

The use of an acid plant to treat roaster and sinter process off-gases was practised at IS of the surveyed plants, and Figure 3 illustrates the type and number of acid units operated.

Single Adsorption Double Adsorption

Tail Gas Scrubber

Figure 3 - Types of Sulfur Removal Equipment Used in the Acid Plants

Smelting

Of the 20 plants surveyed, 8 operated smelters with 5 having a sinter feed and 3 a briquette feed. Table VII gives the operational data for the smelters.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 213

Table VII - Smelter Operational Data

Sinter Charge (t/month) Briquette Charge (t/month) Coke (t/month) Zn Metal (t/month) Pb Metal (t/month) Slag (t/month)

No. of Units

5

3

5 4 5 5

Minimum

15,450

1,100

6,900 7,150 2,050 4,150

Maximum

36,300

2,700

20,000 13,400 10,800 13,000

Average

24,050

1,900

10,700 9,500 5,200 7,850

Leaching

Seventeen plants reported leaching operations, and Figure 4 illustrates the number of units of the various types of leaching processes operated by the plants. Three-stage leaching was the most common process, but one-, two-, four-, and five-stage leaches were reported. Table VIII presents the leaching plant data including the operating conditions and the major additives employed. Leach solution analyses appear in Table IX, and the leach residue analyses are summarized in Table X. The methods of iron disposal after leaching are illustrated in Figure 5; jarosite precipitation remains the major process.

\hn Neutral Weak Hot Acid Super Pressure Other

Acid Hot

Figure 4 - Leaching Steps Used

214 LEAD-ZINC 2000

Table VIII - Leaching Process Data

No. of Units Minimum Maximum Average

pH Stage 1 Stage 2

H2S04 g/l Stage 1 Stage 2

Zn Recovery % Stage 1 Stage 2 Stage 3

Oxygen kg/day Air (m3/day) Mn02 (kg/d) KMn04 (kg/d)

15 9

5 9

12 8 8 6 12 8 4

1.2 0

1 10

77.5 7.5

68.3 1,250

65 560

1

5.4 4.0

147 70

98.0 99.4 98.5

112,000 4,000 4,500 330

4.4 2.6

37 34

88.5 75.4 88.4

35,000 1,100 1,450 128

Table IX - Leaching Solution Analyses

Zn g/l Stage 1 Stage 2 Stage 3

Fe g/l Stage 1 Stage 2 Stage 3

Mn g/1 Stage 1 Stage 2 Stage 3

Mg g/l Stage 1 Stage 2 Stage 3

Cd g/l Stage 1 Stage 2 Stage 3

Cu g/l Stage 1 Stage 2 Stage 3

No. of Units

15 7 6 14 10 7 12 5 3 11 3 3 14 4 3 15 5 6

Minimum

133 55 65 0.5 0.6 5

3.0 3.9 2.0 1.8 8.5 2.0 0.5 270 300 1.3 0.9 100

Maximum

160 160 50

5,000 30,000 30,000

10.5 10.0 10.0 13 15 10

1040 650 600 1880 1500 3000

Average

147 108 104 889

7,900 12,000

6.2 6.5 6.5 7.9 11.2 6.5 478 394 467 613 850 1108

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 215

Table X - Leach Residue Analyses

Zn% Stage 1 Stage 2 Stage 3

Fe% Stage 1 Stage 2 Stage 3

Pb% Stage 1 Stage 2 Stage 3

Cd% Stage 1 Stage 2 Stage 3

Ag g/t Stage 1 Stage 2 Stage 3

No. of Units

13 9 7 13 9 7 13 8 5 9 7 4 6 4 3

Minimum

4.0 2.0 2.5 12.0 16.6 7.4 0.9 1.2 0.5

0.02 0.01 0.02 140 48 130

Maximum

32.5 21.0 20.0 34.5 35.0 35.0 9.9 10.9 15.4 0.54 0.40 0.50 1393 460 780

Average

19.8 11.9 7.0

20.7 24.1 23.3 4.2 4.4 6.3

0.22 0.14 0.15 411 221 437

1%

Figure 5 - Iron Disposal Processes Used

216 LEAD-ZINC 2000

Purification

Solution purification was practiced by 17 plants, with up to 4 stages of purification being reported. Table XI gives the data relating to the common additives used in solution purification, and Table XII reports the purified solution analyses. The solid product from the purification process is a cake, and the cake analyses, by stage, are summarized in Table XIII.

Table XI - Additives for Solution Purification

Zn Dust kg/day Stage 1 Stage 2 Stage 3

Copper Sulfate kg/day Stage 1 Stage 2

Sb Tartrate kg/day Stage 1 Stage 2

As203 kg/day Stage 1 Stage 2

No. of Units

14 14 4

1 8

1 6

3 2

Minimum

750 1,750 2100

130 75

1 7

360 300

Maximum

13,600 33700 21,100

130 1,750

1 36

634 715

Average

6,200 9,650 8,850

130 650

1 17

521 508

Table XII - Purified Solution Analyses

Zng/1 Stage 1 Stage 2 Stage 3

Femg/1 Stage 1 Stage 2 Stage 3

Cumg/1 Stage 1 Stage 2 Stage 3

Cdmg/1 Stage 1 Stage 2 Stage 3

No. of Units

11 12 5

13 12 5

10 11 4

13 12 5

Minimum

137 139 133

2.0 2.3 0.1

0.07 0.02 0.04

0.43 0.05 0.1

Maximum

167 167 169

1,000 11.6 21

350 1,240 0.10

1200 600 15

Average

152 152 151

84.3 5.6 8.2

112 113 0.09

271 80 3.5

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 217

Table XIII - Purification Cake Analyses

No. of Units Minimum Maximum Average

Zn% Stage 1 Stage 2 Stage 3 :% Stage 1 Stage 2 Stage 3 >%1 Stage 1 Stage 2 Stage 3 1 %

Stage 1 Stage 2 Stage 3 1% Stage 1 Stage 2 Stage 3 )% Stage 1 Stage 2 Stage 3

16 13 6

8 5 2

9 6 4

15 11 4

14 11 6

7 8 3

1.5 0.4 14

0.04 0.002

0.4

0.3 0.09 0.5

0.7 0.3 0.2

0.05 0.03 0.2

0.004 0.0004 0.003

95 98 28

15.2 15.2 25.1

11.6 13.8 2.5

80 50 38

42 67 85

6.4 3.5 0.7

22 31 19

2.2 3.2 12.7

4.5 5.3 1.4

35.0 13.6 10.6

7.8 16.5 30.3

1.3 1.5 0.4

Electrolysis

Seventeen of the reporting plants operated electrolytic cells, and data for the electrolysis process are given in Tables XIV and XV. Figure 6 illustrates the relationship between cell power consumption and current density. The current efficiency for electrolysis is compared with the average current density in Figure 7. No correlation between the two variables is apparent. The electrolyte concentrations of the plants are reported in Table XVI, and the electrolyte additives are shown in Table XVII.

218 LEAD-ZINC 2000

Table XIV - Electrolytic Cell Data

Cell Voltage Room 1 Room 2 Room 3

Deposit Time (h) Room 1 Room 2 Room 3

Current density (A/m2) Room 1 Room 2 Room 3

Current Eff. (%) Room 1 Room 2 Room 3

kwh/Zn(t) Stage 1 Stage 2 Stage 3

No. of Units

16 4 3

16 4 3

16 4 3

16 4 3

15 4 3

Minimum

3.10 3.35 3.49

16 24 24

281 430 500

87.3 89.0 89.0

2970 3070 3150

Maximum

3.75 3.7 3.7

72 48 48

640 617 608

92.0 92.0 92.0

3400 3477 3300

Average

3.38 3.51 3.56

34 33 40

505 509 536

90.3 90.1 90.3

3174 3276 3240

Table XV - Cell Configurations

Cathode Area (m2) Room 1 Room 2 Room 3

Anodes/Cell Rooml Room 2 Room 3

Anode Life (month) Room 1 Room 2 Room 3

Cathode Life (month) Room 1 Room 2 Room 3

No. of Units

16 4 3

15 4 3

15 4 3

15 4 3

Minimum

1.0 1.2 1.6

19 28 31

18 18 30

6.5 8.0 15.0

Maximum

4.5 3.2 3.2

85 101 113

84 60 60

60 21 20

Average

2.0 2.0 2.7

41 52 82

40 42 48

19.1 17 18

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 219

c o "H. E 3 V) c o Ü

I o 0.

3400

3350

e 3300

3250

3200

3150

3100

3050

3000

2950

200

0

0

0

o

0 0

0

0

0

0 0

o 0

300 400 500 600

Current Density (A/m )

700

Figure 6 - Electrolysis Power Consumption versus Current Density

93

92

>> o c <D () £ LU

*-» c 0} u. 3

o

91

9U

89

88

87

o

CD

O O

O

O O

o o

o o o

o

o

o o o

CD

200 300 400 500 600 700

Current Density (A/m )

220 LEAD-ZINC 2000

Figure 7 - Current Efficiency versus Current Density Table XVI - Electrolyte Concentrations

Zn g/1 Room 1 Room 2 Room 3

H2SO4 g/i Room 1 Room 2 Room 3

Mg g/1 Room 1 Room 2 Room 3

Mn g/1 Room 1 Room 2 Room 3

Cl mg/1 Room 1 Room 2 Room 3

F mg/1 Room 1 Room 2 Room 3

No. of Units

16 3 2

15 3 2 14 3 2 14 3 2 14 3 2 13 3 1

Minimum

47 47 50

0 165 175 2.0 6.5 11.0 1.3 5.0 10.0 0.2 130 240 1.5 10 50

Maximum

150 70 51

200 193 193 15.0 15.0 15.0 11.5 10.0 11.0 550 500 500 115 50 50

Average

63 56 51

159 176 184 8.1 11.8 13.0 5.6 6.8 10.5 233 310 370 20.6 27.5 50

Table XVII - Electrolyte Additives

No. of Units Low High Average

Glue kg/day Stage 1 Stage 2 Stage 3

SrC03 kg/day Stage 1 Stage 2 Stage 3

Licorice kg/day Stage 1 Stage 2

Na Silicate kg/day Stage 1

10 1 1

14 3 3

5 1

5

5 20 25

1.5 121.7 121.7

5.7 20

27

120 20 25

905 500 200

40 20

280

26.9 20 25

330 257 164

21.8 20

127

Casting and Products

The relative distribution of the types of zinc metal product produced and sold is shown in Figure 8. Just under half is Special High Grade Zinc. Production of alloys for galvanizing, die casting, and other casting totals 23%. Figure 9 shows the distribution of cast zinc shapes and sizes. A little more than half of the production is in the form of 25-kg slabs.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 221

Galvanizing Alloys 18%

Other Alloys 5%

Die Casting Alloy 3%

Dust& Powder

2%

SHG 48%

Figure 8 - Types of Zinc Produced

Slab 20· 25kg 56%

Ingot > 1200 kg 9%

Ingot < 20kg 9%

Ingot 500-1200kg

26%

Figure 9 - Distribution of the Shapes and Sizes of Zinc Metal Products Cast

222 LEAD-ZINC 2000

Labor Use

Data for the labor force was reported by 16 plants and this information is presented in Table XVIII. Figure 10 shows the relationship between zinc productivity and capacity.

Table XVIII - Zinc Plant Work Force

No. of Units Minimum Maximum Average

Total Maintenance Contract

16 12 2

130 10 131

1879 284 199

594 108 110

c CD

E o c

N

.5 3 ■σ o

1600

1400

1200

1000

800

600

400

200

0

o o 0°

o 8

o οθ°ο

0

o o

0

0 100,000 200,000 300,000 400,000 500,000

Capacity (t/y)

Figure 10 - Zinc Productivity versus Capacity

COMPARISON WITH PREVIOUS SURVEYS

Although the number of reporting plants for this survey fell below that of the previous two surveys, some general comparisons can still be made. The average plant capacity has increased by 65% since the 1985 survey. Table XIX shows the production data from the plants responding to the present survey. Figure 11 illustrates the distribution of zinc capacity for the plants that responded to the survey.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 223

Table XIX - Reported Plant Capacity

Year 1985 1995

2000

CO

c (0 Q-

O

ω A

E Z

Total Reported Plant Average Unit Capacity Capacity Slab Zn per Plant

(kt/y) Number of Plants (kt/y) 3968 41 97 3433 32 107

3219 20 161

12

10

a 6

4

2

0 ■ 1 I

π L m ■ ■ . ■

: ■ . ■ I I 1

r-B ^H I 1 1 1

Q1985!

^1995

■20Q0

<50,000 >50,000 >100,000 >150,000 <100,000 <150,000

z inc :Capa cit /(t/y)

Figure 11 - Distribution of Capacity of Plants Responding to Survey

The average plant feed changed as shown in Table XX. The zinc grade improved slightly and the iron content decreased slightly. The lead assay, however, increased over the values for both the 1985 and 1995 surveys.

Table XX - Average Feed Quality of Reporting Plants

Year Average % Zn Average % Fe Average % Pb 1985 1995 2000

52.7 51.2 51.7

7.4 7.9 7.5

1.8 2.4 3.3

Zinc recovery increased by almost half a percent since the 1995 survey. As shown in Figure 12, improvements in processing show a steady increase in zinc recovery since the 1985 survey.

224 LEAD-ZINC 2000

95.5

93.0

1980 1985 1990 1995 2000 2005 Year

Figure 12 - Average Recovery of Zinc from Feed Materials

Conditions for the electrolysis of zinc are compared to those from the previous two surveys in Table XXI. If the conditions used in the plants in the survey are representative of the remainder of the electrolytic operations, then the average current density has increased over that reported in 1995, probably because plants have worked to maximize their zinc production. The power consumption has risen with the current density.

Table XXI - Comparison of the Electrolysis Condition with the Earlier Surveys

Year 1985 1995 2000

Average Cathode Average Current

Aream 1.75 2.20 2.09

Density A/m 527 484 510

Average Current

Efficiency %

89.2 90.3

Average DC Power Consumption kWh/t

Cathode 3181 3191 3202

CONCLUSIONS

A survey of operating zinc smelters and refineries throughout the world has been done in conjunction with the Lead-Zinc 2000 Symposium. Although the number of reporting plants for this survey fell below that of the previous two surveys, some general comparisons can still be made. The average plant capacity has increased by 65% since the 1985 survey. The average plant feed has also changed. The zinc grade improved slightly and the iron content decreased slightly. The lead assay, however, increased over the values for both the 1985 and 1995 surveys. Zinc recovery increased by almost half a percent since the 1995 survey, and improvements in processing show a steady increase in zinc recovery since the 1985 survey. If

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 225

the conditions used in the plants in the present survey are representative of the remainder of the electrolytic operations, then the average current density has increased over that reported in 1995, probably because plants have worked to maximize their zinc production. The power consumption has risen with the current density.

REFERENCES

1. F. Yamada, T. Terayama and S. Enomoto, "A Survey of Operating Situation of Zinc Smelters and Refineries in the World", Zinc '85. K. Tozawa, Ed., The Mining and Metallurgical Institute of Japan, Tokyo, Japan, 1985,61-92.

2. T. Kajiwara, T. Iida and T. Hino, "A Survey of Operating Situation of Zinc Smelters and Refineries in the World", Zinc and Lead '95. T. Azakami, N. Masuko, J.E. Dutrizac and E. Ozberk, Eds., The Mining and Materials Processing Institute of Japan, Tokyo, Japan, 1995,27-53.

3. World Directory: Primary and Secondary Zinc Plants. International Lead Zinc Study Group, London, England, 1999.

4. A. Moreno and R. Hinchcliffe, Zinc and its Markets. Metal Bulletin Books, Ltd., Surrey, England, 1999.

5. Zinc Smelters and Projects: Processes. Costs, and Profitability. Brook Hunt Mining and Metallurgical Consultants, Surrey, England, 1999.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 227

USE OF DATA RECONCILIATION: A ZINC PLANT CASE STUDY

T.J. Auping and M.A. Reuter TU Delft, Raw Materials Processing

120 Mijnbouwstraat Delft 2628RX, The Netherlands

S.C. Grand Consultant

Alter Postweg 12 Dorsten 46282, Germany

K. Born Ruhr-Zink GmbH Wittener Strasse 1

Datteln 45711, Germany

ABSTRACT

Process control, environmental monitoring, budgeting and accounting are based on process data. Without appropriate tools for data analysis, the process data could remain unused and wasted. This paper demonstrates how data reconciliation can be used as a tool to develop closed mass balances. Mass balances are the initial step in plant optimisation (technical, economical or environmental). From mass balances, fundamental process models can be developed. These balances link process parameters (e.g., feed materials) to process performance (e.g., recovery of valuable elements). Some of these models could find their way into simulators useful to plant operators. Classical regression methods regularly fail when they are applied to complex non-linear problems involving a large number of variables. Alternatively, multi-parametrical regression analysis can be successfully applied with the aid of neural networks based on reconciled data. Both methods have been applied to plant data originating from the hydrometallurgical zinc plant at Ruhr-Zink, which in this paper serves as a case study. Very promising and useful results were obtained from this study.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

228 LEAD-ZINC 2000

INTRODUCTION

Ruhr-Zink is the name of a German zinc plant located in the Ruhr area. This plant has an annual production of 105,000 tonnes of zinc. About 30 different concentrates are treated in this plant via the RLE process (roasting, leaching and electrowinning) to produce 6-t zinc ingots.

The leaching process at Ruhr-Zink has gone through a number of technological developments since its start-up in 1968. Initially, iron was separated as jarosite, later as hematite. Pressure leaching was applied for a period to increase production. Today, Ruhr-Zinc is operating using more classical processes, which in the near term might turn out to be one of the ways to develop more modern and future-oriented technologies. Although most zinc plants still precipitate iron as jarosite (which is dumped in ponds or caves), Ruhr-Zink halts the leaching step just before the dissolution of the iron in the calcine. Thus, a zinc-iron concentrate (ZIC) is produced, which is shipped to an Imperial Smelting facility for recovery of Zn, Pb and Ag. Iron in the Ruhr-Zinc process reports mostly to the slag obtained from the Imperial Smelting furnace.

In addition to treating concentrates, Ruhr-Zink also treats Wälz oxides. With this diversity of feed materials, Ruhr-Zink has entered the world of recycling and thus has created a new promising field of operation.

The flowsheet of Ruhr-Zink is complex with a large number of internal recycle flows. The effects on the process performance caused by minor elements present in the concentrates or in the Wälz oxides are difficult to detect and to assess. Limits on sampling and analysis, as well as economical boundaries, forbid a continuous and complete measurement of all relevant data. Nevertheless, optimisation requires closed mass and energy balances as an initial part of the investigation. Even if a complete database were available, process-modelling techniques formerly lacked the ability of handling complex problems that might incorporate an extremely large number of parameters.

The objective of this investigation was, therefore, (i) to assimilate the appropriate plant data from various sources, (ii) to determine and estimate their accuracy, (iii) to calculate closed mass balances for defined intervals of time, and (iv) to develop quantitative model to predict the production of ZIC and its composition based on the feed materials and their analysis. In the course of the investigation, the mass balance model was developed first (Model 1) with the aid of data reconciliation. Then, the prediction model (Model 2) was calculated based on the results of Model 1 incorporating a multi-parametrical regression analysis. In this paper, the capabilities of the combination of these two modern computer-aided tools will be demonstrated as applied to the Ruhr-Zink flowsheet. With these methods, a plant simulator can be updated in a relatively short time that incorporates the latest changes in the flowsheet.

PLANT MODELLING AND OPTIMIZATION TECHNIQUES

Mathematical modelling and simulation have been part of extractive metallurgy and minerals processing for a long time. However, developments have often been slow for a number of reasons:

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 229

• Multi-phase nature of the systems • Poor parameterization of feeds and products • Multi-dimensionality of the systems • Poor (mineralogical) characterization of the paniculate systems.

Plant Modelling

Because of ever-stricter demands on product quality, demanding environmental legislation, tighter profit margins, and a flood of process data from process computers, it becomes ever more important to create models that can assist in meeting the process expectations. Computers are becoming faster, making it possible to treat more data and perform simulations. The main question is whether it is necessary to develop complex models of systems for which the parameters are often only poorly defined.

There are various modelling techniques available (1):

• Fundamental models (white/gray box models); i.e., models based on the physical and chemical fundamentals of the processes (e.g. E=mc2).

• Semi-empirical models (white/gray box models); i.e., fundamental models, whose parameters are estimated by calibration using least squares fitting of existing data (the validity of these models is limited to the range represented by existing data and may not be generalised).

• Empirical models (black box models); i.e., models based on empirical equations whose parameters often do not have physical meaning (mostly application specific).

• Regression analysis (black box models); i.e., statistical methods that permit the approximation of functions and the classification of data using non-parametric methods (application specific).

• Knowledge based modelling (black box models); i.e., models based in symbolical representation of data and information (application and domain specific).

• Combined or hybrid models (black/gray box models); i.e., a combination of two or more of the above mentioned models.

For modelling mass flows and the link between concentrate input and neutral leach residue output, a regression analysis using non-parametric methods (in this case, neural networks) was applied. Intensive analysis and preparation of the available plant data (data reconciliation) had to be performed to produce the database used for modelling.

Data Management at Ruhr-Zink

In a metallurgical plant many parameters are controlled, producing an overwhelming quantity of data. The data are often spread over different places, have different accuracy and standard deviations Furthermore, often, important data are hard to obtain because of non-existent sampling points, etc. The Ruhr-Zink database can be divided into several categories. Data can be classified based on their sampling frequency and how they were made available. Also data can be real-time via a distributed control system (DCS), sampling data from the laboratory, or empirical data originating from the operator. This large body of information often does not find its way into a common database. Most of the assumptions for the mass balance models discussed below are based on such information.

230 LEAD-ZINC 2000

Data Reconciliation

It is well known that measured process data are always noisy and must therefore be described by an average value and a standard deviation. Often, crucial data are recorded because flow rates were not measured and samples were not taken. It would be clear that under such conditions it is difficult to produce closed mass balances for a plant. However, data reconciliation is a technique that permits balancing of a system as a function of the measured data taking into consideration the average value and standard deviation of that data. Also, it permits inspection of process dynamics. According to Wills (2), data reconciliation is often done with minimisation of the sum of squares of the closure residuals of the mass balance equations. This method calculates the best-fit values for closing mass balances which are a function of measured process data with a given mean and standard deviation. For this project, software Usim Pac 2.1© (BRGM, France) is used to do the mass balances. This software uses data reconciliation techniques.

The first step in data reconciliation is to measure the flow rates (if possible) and sample the streams to analyse variations in the concentrations of elements, compounds, etc. In all cases, average values are obtained and their respective standard deviations are estimated. These data form the input to the data reconciliation model with the aim of finding a solution within the boundaries of the data. An analysis of adjustments calculated in data reconciliation for each stream forms an integral part of the analysis. This is a very helpful tool (i) to find deficiencies in the model, (ii) to establish inadequate assaying and sampling practices, (iii) to determine model variance, and (iv) to determine coincidental or systematic errors.

Neural Networks

Aldrich et al. (3) and Reuter et al. (4) discussed the use of neural networks in metallurgy and mineral processing. Neural networks are applied in this paper as a non-parametric regression-modelling tool to develop the prediction model. The program Basic ModelGen™ (Crusader Systems (Pty) Ltd, South Africa) is used to develop the neural nets used in this research.

MASS FLOW ANALYSIS - MODEL 1

Ruhr-Zink can roughly be divided into seven process sections (Figure 1). These are:

• Roasting • Leaching • Wälz oxide leaching • Neutral leach product filtration • Purification • Electrolyte storage and mixing • Electrowinning.

For each of these sections, models were developed to provide a closed and coherent mass balance. The partial flowsheets were linked to produce a complete model for Ruhr-Zink that permits closure of all measured and unmeasured flow rates and analyses. During zinc production not only zinc is recovered as a product, but also, concentrates suitable for sale or for further processing are generated. These by-products include concentrated sulphuric acid, zinc-

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 231

iron concentrate, lead sulphate, copper arsenate, cadmium, thallium cake, gypsum, neutral leach and cleaned process water.

In order to perform the mass balances for Model 1 various assumptions were made, some of which are summarised below:

• Laboratory and instrumentation provide reliable data, with an assumed relative error of 10% for poorly defined streams diminishing to 1% for well-sampled streams. These errors have to be estimated and are used in the data reconciliation procedure. The calculated errors after reconciliation are very valuable for process identification.

• The process is assumed to be at steady state, otherwise the time lag between inputs and outputs should be taken into account.

• All filtration steps are only solid-liquid separations implying that no chemical reactions

With these assumptions and using five elements (zinc, iron, sulphur, cadmium and copper), mass balances were calculated for each of the sections on a monthly basis. Model 1 is capable of estimating the flow rates and analyses of all the measured and unmeasured streams included in the model.

Roaster+Cflbas Cleaning

WO Leaching

PbSOl Concentrate

Spent BecMyte

IL H Leading

BectrowfTinq

Si I dlii u and Casting

"Cflges-H-ESOl

~ Π Neutia1 Leach Rssjote Filtration I "*■ Neutral Leach Residue

~ * O K A S Cement

Figure 1 - Schematic Overview of the Present Flowsheet of Ruhr-Zink

Roasting

During roasting, metal sulphides are mostly converted to oxides, ferrites and silicates. The chemical analysis and mineralogical composition of the calcine have a large impact on the performance of the leaching step. During the period of the project, the roaster feed consisted of about 15 different concentrate sources. The mix also contained about 5% of zinc-containing recycled plant residues and some zinc oxides containing relatively high levels of halogens. Airflow through the fluidized roaster bed is oxygen-enriched by about 1.5%.

232 LEAD-ZINC 2000

The roaster sub-model of Model 1 consists of three major process steps: the roaster itself including the dry off gas cleaning facility, the wet off gas cleaning system and the sulphuric acid plant. The following major assumptions are made to establish this model: (i) 100% of the roasted solids are recovered in the calcine, (ii) water recovered in the washing acid originates mainly from moisture content of the concentrate, (iii) the produced H2SO4 is measured exactly, (iv) the nitrogen concentration in the gas flows is specified exactly.

Leaching

The leaching model (Figure 2) consists of a number of leaching stages and "Thickener 4", where the ZIC is filtered and recovered. The whole leaching and neutralization circuit is summarised in one reactor for the model. Calcine is leached with spent electrolyte. From this sub-model of Model 1, an accurate estimate could be made for the amount of spent electrolyte required to leach one tonne of calcine. This value will later be used to develop a model for predicting the amount of produced ZIC. In addition, unmeasured streams were estimated.

1. Calcine

2. Spent Ekctrolyte

3. Neutral Leach Residue Filtrate

A. Waelz Oxide Filtrate

Sump Steam Cu F i l t ra tF icmCd Plant Outvkumpu Sump Cleaned Cell Slimes 0 2

1. Upmake Tank and Leach Tanks Qa-11

L 9. Solids to Neutral Leach Residue Filtration

Figure 2 - Flowsheet Leaching

Wälz Oxide Leaching

The Wälz oxide is delivered as a washed raw material that still contains excess chloride and fluoride; it is processed in the Wälz oxide treatment plant. The Wälz oxides are washed in an alkaline environment, and subsequently leached with spent electrolyte. After leaching, the filtrate is added to the calcine leaching stage.

In order to develop this sub-model of Model 1, it is assumed that the composition of the Wälz oxide feed remains unchanged during the whole month. Also it is assumed that during washing, the weight loss during the dissolution of the solids (except chloride and fluoride) is negligible. From this model, the required amount of spent electrolyte per tonne of Wälz oxide could be calculated. In addition poorly defined streams could be characterised.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 233

Zinc Iron Concentrate (ZIC) Filtration

De-moisturising of the ZIC is defined in three steps (Figure 3). First, the solids are filtered in two parallel drum filters, and then the solids pass through the centrifuges. The ZIC from the centrifuges contains 20% moisture. (= "dry" neutral leach residue) The filtrate is filtered in presses to recover the remaining solids, which are not recovered in the centrifuges. This produces a ZIC with 30% moisture (= "wet" neutral leach residue).

Purification

Direct or hot-cold purification is done in three stages. In the first stage, copper arsenate is cemented; a fraction of which is filtered and two-thirds of this residue is recycled. The remaining part is washed and dried to produce a Cu-As residue. In the second stage, cadmium is cemented in an Outokumpu process, producing a cadmium and zinc residue. The third stage is a backup stage used only in case of failure and is integrated into the first stage in this sub-model of Model 1.

Electrolyte Storage and Mixing and Electrowinning

In the electrolyte storage and mixing station, the neutral leach solution is mixed with spent electrolyte to adjust the composition of the electrolyte fed to the cellhouse. In the cellhouse, zinc is plated and then it is melted and cast into six-ton ingots.

Plant Flowsheet of Study: Ruhr-Zink Neutral Leach Residue Filtration 1. Underflow Thickener 4

1 1. Drumfllter 1 &3

a Filtrate to Leach 8 -

2. Centrifuges 1-7 VC & Thickener (2)

3

3. Filters 1,3.4 VC & Thickener 3

4 ® *dn 4. Filtrate to Leach

10

I 1 ΰ Wet Neutral Leach Residue

Figure 3 - Flowsheet of the ZIC Filtration Operation

Total Plant

A mass balance model of the whole plant was constructed after the seven sections were completed and reconciled with the sub-models. The sum of the seven sub-models produced a reconciled mass balance. Although no specific numerical data/results are provided here for confidentiality reasons, it can be said that it was possible to balance the whole plant and to generate complete and reconciled data sets for the period in which the analysis was done. It was also possible to generate useful data, which was subsequently used to develop Model 2.

234 LEAD-ZINC 2000

ZINC IRON CONCENTRATE PREDICTION - MODEL 2

A known phenomenon at Ruhr-Zink is that, when ZIC production increases, the concentration of zinc remains constant, which results in increasing zinc losses. The mechanism is explained by an increase in the concentration of silica in the calcine as shown in Figure 4.

▲

U SI

SiCh calcine

Figure 4 - Assumed Relation Between ZIC Production and S1O2 in the Calcine

It is important to identify the relationship between concentrates and other input materials, and the resulting ZIC. This allows the optimization of purchasing practices of concentrates and oxide materials while minimising ZIC production. It is then possible to predict the amount of ZIC and its costs in addition to the environmental burden of the plant represented by this product.

Mineralogy

According to Chen et al. (5), the main phases in calcine are ZnO and ZnO»Fe2C>3 while minor phases such as ZnS, S1O2, PbS04, Fe203 and Z^SiCh are also present. Silicon appears in the concentrate as silica, Al-silicate and K-silicate. During roasting, zinc rich Zn-Fe-Pb-silicates are formed as described by Equation 1.

xZn2Si04 + yFe2Si04 +zPb2SiO„ ->(ZnxFeyPb2)2Si04 (1)

The main phase in the ZIC is zinc-ferrite, while other phases include ZnO, Zn2Si04, gypsum, Pb-K jarosite, PbS04, ZnS and Fe-hydroxide. Silica (gel) can be formed as described in Equations 2 to 4. The zinc ferrite is free of silica. The mineralogical assays, however, do not provide enough information to correlate the formation of ZIC with the composition of the calcine.

Zn2Si04 + 2H2S04 -> 2ZnS04 + H4Si04 (2)

H4Si04-»(Si02)(Gel) + 2H20 (3)

Fe£, + 6H20 -> Fe(OH)3(Gel) + 3H30(+

a (4)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 235

Model 2

The mass balance of the whole plant (Model 1) discussed in section 3 was used as a basis for establishing the relationship between the calcine and ZIC production. This model supplied information on various flows enabling construction of a simplified flowsheet (Figure 5). This flowsheet represents the whole production of ZIC starting with calcine, applying the two depicted units. The first represents ZIC with 20% moisture ("dry" ZIC) and the second ZIC containing 30% moisture ("wet" ZIC). It is evident that this flowsheet combines leaching and ZIC sub-flowsheets into one flowsheet.

2. Spent Electrolyte \

3 WO Leach ^ \ \

- 5 — ^ J

1. Leaching & Centrifuges & Thickener 1 (T)

3. Filter 1,3,4 VC & Thickener 3

g 8 Neutral Leach Residue Dry

f® 9 Neutral Leach Residue Wet 9 ε»

h 7 Crude Leach

Figure 5-Adapted flow sheet of the ZIC Prediction Model

To develop this model various assumptions were made, among which are the following:

The silica concentration in the calcine is the main cause of the increase in ZIC production but other minor elements also have an influence on the products. The time lag is approximately three days. The spent electrolyte volumes used for calcine leaching and Wälz oxide leaching are applied as derived in section 3.2 and 3.3. The composition of the Wälz oxide leach is constant. The moisture in both products is crude leach, but the crude leach content is not used for the calculation of the compositions of the products.

Data

The basis of the results produced by Model 2, are 34-weekly reconciled data series based on the total mass flows and the analyses of zinc, iron, silica, lead and alumina in the indicated streams. The reconciled data based on the relevant statistical information of the data were subsequently used for regression modelling using neural networks.

236 LEAD-ZINC 2000

30

3

< -20 ! ,

-30 - :

O e « 5 < 0 0 > C \ I C \ i m c 0 T - ' * r ^ . O ■ * ■ * ■ * · * i n * - τ - « - CM

Week (#)

Figure 6 - Adjustments on the S1O2 Concentration of the ZIC that is Produced in the Filter Presses

Figure 6 shows the adjustments, which were calculated with the aid of data reconciliation, for the Si02-content in the ZIC, which is produced in the filter presses. These adjustments indicate that this analysis is mainly attributable to systematic errors. The diagram shows clearly that in week 8 an additional error occurred, which in this case was an analytical error. It is also clear that the large calculated errors resulting from the data reconciliation indicate rather large sampling errors on the data. Similar diagrams for other adjustments, which were necessary to calculate a closed mass balance, are valuable indicators of malfunctions in daily plant practice.

Prediction Model

The adjusted weekly values can now be used for multi-parametric regression using neural networks. The objective is to define the non-linear relation for the amount of ZIC produced in the presses and centrifuges. It is also aimed to model the contents of Zn, Fe, Pb, S1O2 and AI2O3 in the ZIC as a function of the chemical composition (Zn, Fe, Pb, S1O2 and AI2O3) of the feed material.

The following relationships were calculated:

Centrifuges f / C a l c i n e \ 1?^· R 1 0 / fS^ mZIC20% — 1Neural NetVm(Fe,Pb,AI203,Si02)^ Λ ■ O l / O (J)

mZIC30% ^ N e u r a l Nel(m(SiO„Fe,Zn)) ^ : 9 4 % ( 6 )

Similar corrrelations were calculated for the moisture and the Zn-, Fe-, Pb-, S1O2- and AI2O3-contents in the ZICs. The accuracy of the predictions regarding the filter presses ranges from 88 to 98%), which is quite precise for a regression analysis based on plant data. The prediction accuracy for the ZIC produced in the centrifuges ranges from 70 to 84 %, which indicates that in this case, important parameters were not taken into account (e.g., size distribution of the concentrate, rotation speed of the centrifuges).

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 237

The resulting correlations, calculated with the neural network functions for Equations 5 and 6, are depicted in Figures 7 and 8, respectively. They show the strong non-linearity of the system. It lies in the nature of a 3D figure that the graph is restricted to only two variables each, which are Pb and Fe in the case of the ZICCenmfuges and S1O2 and Fe in the case of the ZICpresses.

The application of these equations in August 1999 allowed a prediction of the ZIC production with a deviation of only 3% from the measured production. Similar predictions have now found their way into daily plant practice.

Figure 7 - Prediction Model for the ZIC with 20% Moisture (Centrifuges), R2: 81%, [t/week]

Figure 8 - Prediction Model for the ZIC with 30% Moisture (Filter Presses), R2: 94%, [t/week]

238 LEAD-ZINC 2000

A sensitivity analysis has been made for the neural net functions defining Equations 5 and 6. Each of the input variables mFeRG, (mass of iron in the calcine or in German Röstgut), ms» (mass of silica in the calcine), mpbRG (mass of lead in the calcine), and mAi!OjRG (mass of alumina in the calcine) was perturbed independently by 5%, and the resultant effect on the output of each function was recorded (Figures 9 and 10). From Figure 9 it is evident that production of ZIC with 20% moisture (centrifuges) is principally dependent on the iron and lead concentration in the calcine. On the other hand, production of ZIC with 30% moisture (presses) depends mainly on the silica and iron concentration in the calcine.

The prediction model was tested against weekly, unseen data (Figure 11). It is clear that the model performs rather well. It is now possible to investigate changes in the ZIC under a number conditions. The model has successfully predicted the ZIC production on a monthly basis. Concentrates have to be blended and purchased in such a way that the silica concentration in the calcine does not exceed 2% in order to maintain a minimum production rate of ZIC.

Figure 9 - Sensitivity Analysis of the Effect of the Components in the Calcine on the Composition of the ZIC with 20% Moisture

Figure 10-Sensitivity Analysis of the Effect of the Components in the Calcine on the Composition of the ZIC with 30% Moisture

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

50

140

o 5 30 ■o o ΰ | 20

10 10 20 30 40 50

Measured value (t/w)

Figure 11-Comparison Between Measured and Predicted Values for the Total Amount of S1O2 in the ZICpreSsesm Tonnes per Week

It is qualitatively known (6) that increased silica contents in the feed materials could increase the production of neutral leach residue. However, now it possible for Ruhr-Zink to quantitatively examine not only the influence of silica on the production of ZIC but also the synergistic influence of alumina, iron, lead, silica, and zinc on the production of ZIC.

CONCLUSIONS

The main characteristics of plant data are their noise because of systematic and coincidental errors and their incompleteness for practical and cost reasons. Computer-supported methodologies allow calculation of closed mass balances even for complex plants, and thus, development of computer models, which relate process parameters such as the composition of the feed material to the operational results; e.g., recovery of valuable elements.

Application of modern mass balancing techniques (data reconciliation) and multi parametrical regression modelling (neural nets) to plant data from Ruhr-Zink demonstrate the capability of the combined methodologies. Based on incomplete and noisy plant data, closed mass balances were calculated for the complex flowsheet of the hydrometallurgical plant. The reconciled data gave not only valuable information about the accuracy of the sampling and analysis but also permitted updating of existing simulation models and modelling of a quantitative relation between the composition of the concentrate feed and the amount and composition of the resulting neutral leach residue. This model is now used to optimise the industrial process.

At present the information technology structure in the plant is being upgraded to integrate modelling techniques into metallurgical control and accounting. For example, this will permit monitoring and optimisation of zinc dust additions in the purification section and improve environmental control. In the future it will also be possible to monitor individual (minor) elements and investigate their effects on the metallurgical and economic performance of the plant on a sound statistical basis.

240 LEAD-ZINC 2000

REFERENCES

1. M.A. Reuter, "Some Aspects of the Control Structure of Ill-defined Metallurgical Furnaces", Erzmetall. Vol. 51(3), 1998,181-194.

2. B.A. Wills, "Mineral Processing Technology : An Introduction to the Practical Aspects of Ore Treatment and Mineral Recovery", 6th edition, Butterworth-Heinemann, 1997, ISBN: 0750628383.

3. C. Aldrich, J.S.J. Van Deventer and M.A. Reuter, "The Application of Neural Nets in the Metallurgical Industry", Minerals Engineering. Vol. 7(6), 1994, 793-809.

4. M.A.Reuter, "Hybrid Neural Net Modelling in Metallurgy", Second International Symposium on Metallurgical Processes for Early Twenty-First Century. H.Y. Sohn, Ed., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1994,907-927.

5. T.T. Chen, J.E. Dutrizac and C. Canoo, "Mineralogical Characterization of Calcine, Neutral Leach Residue and Weak Acid-Leach Residue from the Vieille-Montagne Zinc Plant, Balen, Belgium", Trans. Instn. Min. Metall.. Vol. 102,1993, C19-C31.

6. R.M. Grant, "Zinc Concentrate and Processing Trends", World Zinc '93.1.G. Matthew, Ed., The Australasian Institute of Mining and Metallurgy, Parkville, Australia, 1993, 391-397.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 241

RECENT OPERATIONS AT THE HIROSHIMA SMELTER

T. Iwamoto, H. Akiyama and K. Eto Hikoshima Smelting Co. Ltd.

1-1-1 Nishiyama-cho, Hikoshima Shimonoseki, Yamaguchi, Japan

ABSTRACT

The Hikoshima smelter carried out an expansion project in 1997 and increased its production from 59,200 t/y to 71,000 t/y. The leaching process and the casting process had production capacities of 84,000 t/y, but because of the electric power situation in Japan, the production capacity in the electrolysis process was limited to 59,200 t/y to justify the production cost. Electrolysis was the bottle neck of the plant. In order to eliminate a part of this bottle neck, and considering the electric power situation in Japan, a way to increase the production during the time when the electric power rate is low was sought. In practice, we increased the number of cells by 36, and further improved the operation procedures in the cellhouse. For example, we improved the working efficiency of the cells, improved the productivity during the night time, decreased the unit power consumption and improved labor productivity. Accordingly, this paper describes the recent operations at the Hikoshima plant.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

242 LEAD-ZINC 2000

INTRODUCTION

The Hiokoshima plant started zinc operations in 1916 with a horizontal retort process, and in 1971, switched its production to an electrolytic process. In 1986, the plant was separated and became an independent company, Hikoshima Smelting Co., Ltd., a wholly owned subsidiary of Mitsui Mining & Smelting Co., Ltd. The production capacity was approximately 55,000 t/y until 1994. In 1994, it was increased to 59,200 t/y as the result of an energy saving and innovation project. A further production expansion was carried out in 1997 and the production was increased to 71,000 t/y.

The leaching and casting processes at the Hikoshima plant had a capacity to meet 84,000 t/y production, but because of the electric power supply situation in Japan, the cellhouse had only 59,200 t/y capacity to justify the production cost. This was the bottle neck of the plant. Thus, in 1997, an expansion of the electrolysis process was carried out to obtain 71,000 t/y of zinc production. The major works for this expansion were increasing the number of cells, improving the working efficiency of the cells, and increasing the current density. The chronological changes of electrolytic zinc production at the Hikoshima plant are shown in Figure 1.

Figure 1 - Production of Electrolytic Zinc (t/y)

ELECTRIC POWER SUPPLY SITUATION IN JAPAN

First, a short explanation of the electric power supply and demand situation in Japan is given. Japan is said to have a higher electric power rate than other countries. Further, at times when there is a high demand, the power rate becomes very high. For this reason, the power rate is different depending on the season, the day of the week and even during different hours of the day. Therefore, speaking of the production cost, the most reasonable production method is to increase the production during those hours when the power rate is low and to minimize the power consumption as much as possible during the hours when the power rate is high.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 243

Figures 2 to 5 show some examples of the electric power rate and the variation of the current density in a day.

7.0

6.0 ■a .£ 5.0

•R 4.0

I 3.0

| 2.0 o *- i.o

0.0

-

- i r L 0 3 6 9 12 15 18 21 24

Time

ε <

53 c u -a

i

700

600

500

400

300

200

100

0

1

0 3 6 9 12 15 18 21 24

Time

Figure 2 - Example of the Power Rate (Weekday)

Figure 3 - Example of the Current Density (Weekday)

0 3 6 9 12 15 18 21 24

Time

0 3 6 9 12 15 18 21 24

Time

Figure 4 - Example of the Power Rate (Night Time)

Figure 5 - Example of the Current Density (Night Time)

The percentage of the electric power cost in the total production cost is rather high, and most of the electric power is consumed in the electrolysis process. For the expansion of production, it is ideal to increase the production without increasing the electric power cost. Therefore, the most suitable method is to increase the rate of production during the hours when the power rate is low.

244 LEAD-ZINC 2000

OUTLINE OF THE FACILITIES BEFORE THE EXPANSION

Electrolytic Cells and Rectifier

The cellhouse is composed of 3 lines with a total 324 cells and 3 rectifier circuits. Figure 6 shows the outline of the facilities. The No. 1 line was 18 cells/row x 8 rows for a total of 144 cells; a silicon rectifier with 32.5 kA 550V was used. The No. 2 line had the same facilities as the No. 1 line. The No. 3 line was composed of 18 cells/row x 2 rows for a total of 36 cells, and a thyristor rectifier of 140V 36 kA was used. Of these cells, 2 rows, 36 cells, were normally short circuited for cell maintenance.

No.3 Rectifier

Stripping machine No.l No.2

Rectifier No.l No.2

c iftto -Θ H H h -H- J □ □ □ C

Circurating hoist

3 r~T—i □ □ □ □ □ □ □ □ □ □ 1=1 Automatic transfar crane

EBBB BBBB EBBB EBBB

BBBBB BBBBB BBBBB BBBBB BBBBB BBBBB BBBBB B BB BB □ □ □ CZD Γ Τ Ί □ □ □ □ □ □

^ _ B B BBBB EBBB EBBB

BB BB BB BB BB BB BB BB BBBB BBBB BBBB BB BB

BBBBB BBBBB BBBBB BBBBB BBBBB BBBBB BBBBB BBBBB □ □ □ □ □

Figure 6 - Outline of the Cellhouse

Electrodes and Equipotential Bars

Cathodes have an effective depositional surface area of 1.67 m2 and there are 33 cathodes in a cell. A knife switch type is used for the cathode contacts. This type allows the cathode header bar to come in contact with both the clip and the equipotential bar, resulting in a wide contact area. With a self-weight contact type, the contact resistance is increased because the weight of the cathode is light at the initial stage of electrolysis. To minimize this disadvantage, the knife switch type has been consistently in use to the present time.

The anode is made of 0.75% Ag-Pb alloy. It is cast with holes, and 34 anodes are charged in a cell. The anode is a self-weight contact type. A copper strip of 140 mm width is used as the equipotential bar. The center to center distance between the anodes is 75 mm.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 245

Cell Cleaning Anode Maintenance

Two rows of 36 cells are always shutdown for cell cleaning and anode maintenance. Manganese crusts sticking to the anodes were removed by a reductive dissolution reaction using ferrous sulfate, from 1970 to 1982. This method required time to prepare the ferrous sulfate as well as for reaction, and therefore, in 1982, an anode preparation machine which removes the manganese crust by mechanical impact was developed. However, this machine was a fixed type requiring manpower to move the anodes by forklift trucks to the machine. Further, the surface of the hammer was made of metal and this caused partial damage to the surface of the anodes. Because of these problems, the machine was modified in 1989 to a self-traveling type which allowed direct charging of the anodes by an overhead travelling crane. The material of the impact hammer was also changed from metal to plastic to minimize damage to the anode surface.

Stripping Machine

Stripping machines are of the hammering-wedge descending type of our own development, and there are two units, each unit operated by one person. How to introduce the wedge is the most important point in stripping the deposited zinc, and in this system, a part of the zinc deposit is loosened from the cathode plate by an impact given by the hammer to allow an easier introduction of the wedge. At the discharge outlet of the stripping machine, a brushing/polishing device was installed to polish the surface of those parts of the cathodes which came in contact with the electrolyte.

Switching of Current Density and Concentrated Stripping

In order to increase the rate of utilization of inexpensive electric power, the current density during day time is dropped to a minimum of 70 A/m2 but is increased to 580 A/m2 at night time. However, for the initial stage of deposition of zinc on the aluminum cathode, the current density has to be 150 A/m2 or higher, otherwise, sticking of the deposit on the cathode will occur during stripping. Thus, the stripping operation is not done during the hours of low current density but is concentrated during the night time. The electrode exchange and harvesting cycles are dependent on the electric power situation and differ depending on the season and the day of the week, varying from 24 to 48 hours.

Conveying Facilities

Up to 1992, cathodes were pulled from the cell by an overhead travelling crane and were moved to a transfer car. On this car, the cathodes were manually transferred to a circulating hoist. Plated cathodes coming from the stripping machine were also manually transferred from the circulating hoist to the overhead travelling crane on the transfer car and were then returned to the cells. In 1992, the circulating hoist and overhead travelling crane were modified to have an automatic connection, thus eliminating the manual transfer operation on the transfer car. An automatic transfer to the stripping machine was established. With this improvement, the electrode exchange operation now can be done with only one operator.

246 LEAD-ZINC 2000

IMPROVEMENTS FOR THE PRODUCTION EXPANSION

Increasing the Number of Cells

Initially, the number of cells was 18 rows x 18 cells. It was subsequently increased by 2 cells per row, giving 18 rows x 20 cells = 360 cells. This increased the production by 11%.

In order to carry out the operation at 580 A/m2 as before, the total voltage exceeded 550V which was the capacity of the rectifier. However, by dropping the cell voltage, both No. 1 and No.2 lines remained less than 550V which is within the capacity of the rectifier. A drop in the cell voltage can be expected by increasing the number of electrodes in a cell, but this procedure was not implemented in the expansion project. If the electrode exchange and harvesting cycles remain the same, the increase in the number of cells means more cathodes have to be stripped. To cope with this, improvements were made to minimize the loss time during stripping, and also, the settings of the sequencers were changed to shorten the stripping time. Further, for easier stripping, the polishing of the aluminum cathode blanks was modified to polish their entire surface.

Improving the Working Efficiency of the Cells

At the Hikoshima plant, two rows of cells are short circuited for cell cleaning and anode maintenance. Before the production expansion, 2 rows of 36 cells were always shut down for maintenance. However, the power rate is comparatively inexpensive on Saturdays, Sundays and holidays, and if all the cells are operated, production with cheap power can be increased. Thus, it was decided to have all the cells operating on weekends and holidays. On the other hand, with this change, cell cleaning and anode maintenance cannot be done on weekends and holidays, resulting in a longer anode maintenance cycle. A longer anode maintenance cycle means an increase of operation disturbances such as a rise of the cell voltage and more short circuits of the electrodes. In order to solve these problems, cell cleaning and anode maintenance efficiencies were improved to allow, as much as possible, the extension of the maintenance cycles. That is, first, an automatic cell maintenance machine was developed for cell cleaning, and the suction of the sludge, which was previously done manually, was automated.

Further, the anode maintenance process was made more efficient by the following procedures. First, the number of anode preparation machines of the mechanical impact type was increased to two units. Then, noticing that the anode transfer time was long, the transfer system was made more efficient. The things that were done, are as follows. First, the anodes and cathodes from two cells are set on a rack to empty two cells, and then the automatic cell maintenance machine is set into the cell by the overhead travelling crane. This automatic cell maintenance machine cleans the cells at a speed of 14 minutes per cell. During this time, the anodes from the adjacent cell are transferred by the overhead travelling crane to the two anode preparation machines for maintenance. The anodes after maintenance are returned to the cleaned cell. These procedures are repeated one after the other resulting in efficient cell cleaning and anode maintenance. Figure 7 shows this work procedure.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 247

Cleaned cell

Empty

T Empty

Uncleaned cell

Automatic cell maintenance machine

Anode maintenance machine

Figure 7 - Outline of Cell Cleaning and Anode Maintenance Procedures

Anode maintenance used to take 3 days for 36 cells, but it is now possible to clean 40 cells in 2 days. This practice helped to minimize any significant extension of the overall maintenance cycle. The history of the changes in the anode maintenance procedure is listed below.

By these changes, the cell working efficiency was improved and production was increased by 8%. In the year 2000, a plan to establish procedures for maintenance during energizing will be implemented to further improve the working efficiency of the cells.

Increasing the Current Density

Before the expansion, the rectifier in the No. 3 line had an allowance in its capacity, but in order to maintain a constant running current, it was running at the same current as the No. 1 and No. 2 lines. In order to utilize the equipment capacity more fully, the current density was increased. By this change, production was increased by 1%.

IMPROVEMENT OF THE OPERATION

Because of the different electric power rates in the day time and at night in Japan, the power consumption is changed from the day time to the night. Comparing the same production tonnage, the unit consumption of electric power is in proportion to the current, and therefore, it is higher than having a constant electrolysis current. Accordingly, the following measures were taken to decrease the unit consumption of electric power.

The unit consumption of electric power is mainly determined by the cell voltage and the current efficiency, and Table 2 shows the control standard of the electrolyte which has a great effect in decreasing the cell voltage. Zinc concentration and sulfuric acid concentration were reviewed and their control standards were changed from 60 g/liter to 55 g/liter and 160 g/liter to 185 g/liter, respectively, to reduce the solution resistance. Further, impurities in the circulated electrolyte were also reviewed, and the control standard for the Mg concentration was changed from 10 g/liter to 5 g/liter and that of the Mn concentration from 4 g/liter to 2 g/liter to reduce the solution resistance.

Tab

le 1

- H

isto

ry o

f Ano

de M

aint

enan

ce P

roce

dure

s

Yea

r M

etho

d M

aint

enan

ce C

ycle

M

anpo

wer

M

aint

enan

ce D

ays

Mai

nten

ance

Cel

ls

Eff

icie

ncy

of M

aint

enan

ce

(day

s)

(man

pow

er/

(day

s/ba

tch)

(c

ells

/bat

ch)

(cel

l /m

anpo

wer

) ba

tch)

19

70-1

982

Che

mic

al

40

60

5 36

0.

6 19

82-1

989

Mec

hani

cal

24

18

3 36

2

1989

-199

7 M

echa

nica

l 24

12

3

36

3 19

97-2

000

Mec

hani

cal

28

6 2

40

6.7

LEAD-ZINC 2000

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Table 2 - Changes in the Concentration of Elements in the Electrolyte

Element Zn H2S04

Mg Mn

April 1994 60 160 10 4

April 1999 55 185 5 2

The current efficiency has been maintained around 90%.

Cooling of the electrolyte is done in a TCA cooler and the control standard for the electrolyte temperature is 40°C. Heat generated in the cells has increased by 12% with the increase in the number of cells. The TCA cooler is influenced by atmospheric temperature, and in summer time under full load, it needs a reinforcement of cooling capacity. Thus, improvements were made to reduce the heat generation by dropping the cell voltage, and at the same time, the gas-liquid ratio of the TCA cooler was changed to control at the point where the cooling heat volume is greatest. By these measures, the heat balance is maintained even during the summer time when the load is maximum, and without any expansion of the cooling facilities.

CONCLUSIONS

In 1997, the Hikoshima smelter carried out an expansion of the cellhouse, which then was the bottle neck of production, and increased its production from 59,200 t/y to 71,000 t/y. The major features of this expansion project were increasing the number of cells, increasing the working efficiency of the cells and raising the current density. We plan to carry out further improvements, such as anode maintenance while energizing the cells, to establish ourselves as a cost competitive smelter in the world.

Edited by J.E. Dutrizac, I.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 251

IMPROVEMENTS IN THE LEACHING CIRCUIT OF IMMSA'S ZINC PLANT IN SAN LUIS POTOSI, MEXICO

P. Alfaro, C. Moctezuma and S. Castro IMMSA, Planta de Zinc C.P. 78180, Apdo. 1305

San Luis Potosi, S.L.P., Mexico

ABSTRACT

IMMSA operates a conventional roasting-leaching-purification-electrolysis plant for the production of zinc from concentrates in San Luis Potosi, Mexico. The leaching circuit was designed to take into consideration a feed mixture to the roaster with a maximum iron content of 7%. Since the start-up of the plant in 1982, the zinc recovery increased to 94.5%. However, an increase in the iron content in 1997 caused an important decrease in the global zinc recovery. Two new steps were added to the leaching circuit and modifications were made to the other steps in order to increase zinc extraction. Two new automated press filters were installed to replace the original vacuum drum filters and to minimize soluble zinc losses. Additionally, the residue treatment steps in purification were modified and a calcium removal circuit is just starting operation. These actions, besides improving the operations, have as the main objective to increase the global zinc recovery to levels above 94.5%.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

252 LEAD-ZINC 2000

INTRODUCTION

The operation of the electrolytic zinc refinery of Industrial Minera Mexico has been described previously by P.Alfaro et al. (1); consequently, the intent of this paper is to present the latest additions and modifications that were done in the wet circuit to increase zinc recovery.

Since the start-up of the zinc plant, the leaching circuit comprised three main steps: neutral leaching, hot acid leaching and jarosite precipitation, as shown in the flowsheet of Figure 1.

Spent - Calcine

^ Impure Neutral Solution

**—Na* I Lead-Silver to storage

m LÜT

Jarosite to storage

Figure 1 - Original Leaching circuit (1982 - 1989)

This circuit operated without major changes until 1990. In that year an intermediate step, or low acid leach was incorporated into the circuit, see Figure 2. The objective of the low acid leach is to extract the zinc oxide content of the neutral leach residue which originates from the use excess of calcine in neutral leaching. In so doing, the remaining zinc ferrites will have a better chance to be destroyed in the high acid leaching step.

Following this modification, the global zinc recovery increased from 92% to 94%, and remained over this figure until 1996 when it decreased along with an increase in the iron content of the concentrate mixture fed to the roaster. This caused a strong impact in the 1997 and 1998 zinc recoveries. Figure 3 shows how an increase in the iron content of the calcine to leaching correlates with a decrease in the global zinc recovery. Even though the global extraction in the leaching circuit re-established itself to the levels of the previous years (see Figure 3), the global recovery diminished because of the additional amount of residue produced (see Figure 4).

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 253

A study of the wet circuit operations, with the technical assistance of Asturiana de Zinc, showed that the majority of the zinc loss was due to the low wash efficiency of the existing vacuum drum filters. In addition, it was observed that there still was an amount of recoverable zinc in both the lead-silver and jarosite residues, and that several other steps of the wet circuit could be improved.

Soent - Calcine

. Impure Neutral Solution

NH„*

L^. Jarosite residue to storage pond

Figure 2 - Leaching Circuit with the Incorporation of Low Acid Leaching (1990)

99 , , 10

s u < o z z o

1994 1995

— Recovery — Iron in Calcine

1996 1997 1998

—£—Global Extraction

Figure 3 - Zinc Recovery and Zinc Extraction in Relation to the Increase in the Calcine Iron Content

LEAD-ZINC 2000

90 I i i 1- 50

1994 1995 1996 1997 1998

—X—Recovery —■—Lead and Silver residue — ♦ — Jarosite residue —A—Total residue

Figure 4 - Zinc Recovery in Relation to the Residues Produced.

IMPROVEMENT PROJECT

In order to realize a substantial increase in the zinc recovery, it was decided to carry out an integral improvement project. This improvement project included the following features: equipment substitution, the addition of new steps and modifications to the existing steps.

Equipment Substitution

Filter Press

To justify the substitution of the installed vacuum rotary filters by more efficient equipment, a pilot filter press was installed close to the lead-silver and jarosite residues underflow lines. Tests performed with the press on-line indicated that both residues were easily filtered with high soluble displacement efficiencies. On this basis, two new automated filter presses with 85 recessed 1500 x 1500 mm plates were bought. These filters substituted six 4.2 x 4.9 m vacuum rotary filters.

Addition of New Steps

Third Acid Attack

The objective for this new step was to increase the global zinc extraction by 0.6% with another zinc ferrite leach, and also to increase the iron extraction by partially destroying the jarosites formed in the acid circuit. This produced a higher quality lead-silver residue better suited for high silver recoveries in the subsequent flotation step. Two installed tanks of 80 m3

capacity were refitted as reactors and one thickener of 500 m3 capacity, previously used in the counter-current jarosite wash, were connected to form this new step.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 255

Jarosite Acid Wash

The objective of this step is to extract at least an additional 0.5% of the total zinc going to leaching. This is achieved in a high temperature and moderately acidic acid wash of the jarosite residue, to leach part of the zinc ferrites that are present because calcine is used as the neutralizing agent in the jarosite precipitation step. To install this step, two existing 80 m3

tanks were refurbished as reactors and a new 500 m3 thickener was built.

Modifications to the Existing Operations

Neutral Leaching

A new tank, of 80 m3 capacity, was installed to collect the cell cleaning manganese sludge which is pumped intermittently from the cellhouse sump. From this collector tank the manganese sludge is added to the neutral leach in a controlled manner thus avoiding the abrupt flow variations which had previously caused control problems and an inefficient use of the manganese dioxide contained in the sludge.

Low Acid Leaching

An existing 80 m3 tank was added to low acid leaching to increase the residence time, thus assuring the extraction of the zinc oxide present because of the use of excess calcine in neutral leaching.

Jarosite Precipitation

To maintain the required residence time, a new reaction tank of 80 m3 capacity was installed to replace a tank that has been used as a fourth purification step. Also, a pumping tank was refurbished as a reaction tank to replace another one used in this project in the fourth purification step. A calcine feeder was also installed to deliver calcine to the fifth tank, as previously, and to establish a better control over the acidity in this stage.

Fourth Purification

A new reaction tank was installed and another was taken from the jarosite cascade, both of 80 m3 capacity. A pumping tank was built to send the pulp towards an existing 500 m3

thickener conditioned for this stage.

The objective of this modification is to increase the residence time to maximize the oxidation of ferrous ions, contained in the jarosite solution, using pure oxygen. In doing so, this stage also helps in the neutral leaching step in removing some impurities from solution such as arsenic and antimony.

The leaching flowsheet in operation since November 1999 in shown in Figure 5.

256 LEAD-ZINC 2000

Spent „ „ „ Mn02 Cell House n ._ ..

, . Purification manganese sludge s o l u ( j o n s • Calcine

^ Impure Neutral solution

Figure 5 - Current flowsheet

Other Modifications

Purification Cement Treatment

Modifications are in progress for the treatment of the cements from both purification steps. The hot purification cement treatment will be modified to include a semi-continuous double acid wash, for the extraction of zinc and cadmium, followed by an alkaline wash, for the extraction and recycle of arsenic. For the cold precipitation cement, a semi-continuous single acid wash will be included. The objective is to increase the global zinc recovery by 0.3%, to decrease the arsenic trioxide consumption by 75%, and therefore, improve the quality of the copper cement. The cold purification cement treatment area will undergo an equipment redistribution because two reactors previously used for leaching the cement were refurbished and assigned to the jarosite acid wash circuit.

Gypsum Removal

The typical analysis of calcium in the purified solution is 550 mg/1 and in the spent electrolyte it is 350 mg/1; this indicates that an important amount of gypsum is collected in the

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 257

cooling towers, pipes, distribution launders and anodes. This makes it necessary to carry out periodic cleaning of this equipment. The consequence is a decrease in the availability of the electrolyte cooling and distribution system and an additional requirement for manpower for this work.

Consequently, a calcium removal system for the purified solution was installed and it includes a 150 m3 mix and pumping tank, three cooling towers, one of which is new, and a 1500 m3 capacity thickener. This modification will reduce gypsum deposition in the electrolysis circuit, and thus, increase the availability of the electrolyte distribution system.

Investment

The total cost for the improvement project is 6.6 million U.S. dollars with the distribution shown in Table I.

Table I - Investment Cost Distribution Step

Neutral Leaching Low Acid Leaching High Acid Leaching Third Acid Attack Jarosite Precipitation Fourth Purification Jarosite Acid Wash Pb-Ag and Jarosite Filtration Purification Cements Treatment Gypsum Removal Ancillaries Eng., Superv., and Administration

Cost (thousands U.S. dollars) 205 91 81

417 236 319 778

2,531 559 872 292 293

RESULTS

The incorporation of the third acid attack in July 1999 and the jarosite acid wash in October 1999 have had a positive impact on the global zinc extraction in the leaching circuit increasing it by over 1%, as shown in Table II.

Table II - Effect of the Incorporation of the Third Acid Attack and the Jarosite Acid Wash on Zinc Extraction

Year

1995 1996 1997 1998 1999

Nov '99 - Feb Ό0

Neutral + Acid Extraction

(%) 98.9 98.3 98.1 98.7 99.2 99.5

Global Extraction

(%) 97.9 97.1 96.6 97.3 98.2 98.4

258 LEAD-ZINC 2000

The third acid attack has had the most impact on the global zinc extraction, as evident in Table III. This occurs because its operation parameters have been better controlled in comparison with the jarosite acid wash. In the latter, temperatures are still below target because it has not been possible to achieve a good balance in the distribution of steam to every unit of the process.

Table III - Zinc Extraction per Step and the Contribution of Each Step to the Global Zinc Extraction for the Period Between November 1999 to February 2000

Individual Zinc Extraction

Contribution per step to the Global Zinc Extraction

Neutral

%

86.0

65.1

Low Acid

%

58.8

6.2

Hot Acid

%

68.3

3.0

Third Attack

%

71.2

1.0

Fourth Purif.

%

70.6

5.3

Jarosite

%

92.0

17.6

Acid Wash

%

19.7

0.3

Global

%

98.4

98.4

With regard to the main objective of this improvement project, which was to increase the global zinc recovery, it can be concluded from the data in Table IV that we are very close to the minimum expected value of 94.5%. It can also be seen that, in comparison with the years previous to 1998, the total amount of residue produced is higher. This means that the soluble zinc recovery has been substantially improved with the operation of the automated press filters.

The identified zinc losses in the residues indicate that, with an adequate adjustment of the operating parameters in the jarosite acid wash, it is possible to increase the extraction by at least an additional 0.5% of the total zinc. As well, it is feasible to obtain a similar increase in the recovery of the soluble zinc that is lost to the residues, see Table V.

Table IV - Zinc Recovery Compared with Residue Production

1995 1996 1997 1998 1999

Nov '99 - Feb '00

Iron in Calcine

(%) -

7.4 7.9 8.7 8.9 8.9 8.9

Lead-Silver residue

kg ton of Calcine

122 121 131 109 96 87

Jarosite residue

kg ton of Calcine

180 186 193 237 243 244

Total residue

kg ton of Calcine

302 307 324 346 339 331

Global Zinc Recovery

(%) 94.4 93.8 92.6 91.9 92.4 94.2

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 259

Table V - Identified Zinc Losses for the Period July 1999 to February2000 Zn insol. Pb-Ag Zn sol. Pb-Ag Zn insol. Jarosite Zn sol. Jarosite

Zn in feed Zn in feed Zn in feed Zn in feed

(%) m (%) (%) 0.11 0.29 1.60 0.75

CONCLUSIONS

The increase in the iron content in the zinc concentrate mixture fed to the roaster caused a noticeable decrease in the global zinc recovery of the plant. This occurred mainly because of the higher amounts of residue generated that were not properly washed in the existing filtration equipment. Additionally, the zinc extraction decreased because of the lack of a sufficiently strong attack of the zinc ferrites and the increased use of calcine in the jarosite step that is caused by the increased amount of jarosite precipitated.

An improvement project was initiated which consisted of important additions and modifications to the hydrometallurgical section of the process. To date, all of the changes in the leaching circuit are finished and in operation. The gypsum removal circuit is to be started soon, and the purification residue treatment changes are in progress.

With the corrective actions taken to date, it has been possible to increase by 1.1% the global zinc extraction. The global zinc recovery has increased by 2.3% and it is predicted that with the better control of the circuit and with an optimal residue wash, the proposed goal of 94.5% zinc recovery will be easily obtained during this year. The success of the improvement project will more than justify its cost of 6.6 million U.S. dollars, with an annual production increase of 2,500 tonnes of refined zinc. As well, a decrease in maintenance cost is expected along with an increase in the availability of the equipment because of the substitution of filtration equipment and the operation of the gypsum removal step.

REFERENCES

1. P. Alfaro and S. Castro, "The Zinc Refinery of IMMSA in San Luis Potosi, Mexico", Zinc and Lead Processing. J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton, P. Hanckock, Eds., Canadian Institute Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 71-83.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 261

EXPANSION PLANS AT CPM'S ELECTROLYTIC ZINC REFINERY

T. Takayama, W. Magalhäes and J. Welsh Companhia Paraibuna de Metais

Br 267, km 119, Igrejinha 36.001-970, Juiz de Fora, Minas Gerais, Brazil

T. Newton and S.J. Thiele Egis Consulting Australia Pty Limited

19* Floor, Central Plaza Two, 66 Eagle Street Brisbane, Queensland, 4001, Australia

ABSTRACT

Companhia Paraibuna de Metais (CPM) is presently pursuing a zinc production expansion project. Besides increasing production, the goal of the project is to address existing problems while providing a platform for a proposed twofold increase in production in 2003/2004. Therefore, the current expansion strategy was formulated to address present and future requirements. Computer simulation modeling and full scale plant testing are being used to generate data to optimize the operating conditions and to perform a sensitivity analysis. The initial phase of the project, which will boost CPM's capacity to 86,000 t/y of marketable zinc, is to be completed in September 2000.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

262 LEAD-ZINC 2000

INTRODUCTION

Zinc production at CPM's plant in Juiz de Fora has increased since commissioning in 1980. Market predictions suggest that increased demand for zinc in Brazil will continue and this, coupled with the recent commissioning of the CPM-owned Sobragi hydroelectric plant, prompted management to embark on yet another expansion. This expansion, scheduled for completion at the end of 2000, will boost CPM's capacity to 86,000 t/y of zinc. Originally, this represented an increase of 16,000 t/y, but plant trials performed under "overload" conditions, indicated that a production increase of 5,000 t/y could be obtained almost without additional investment.

The expansion mechanism adopted makes maximum use of the existing equipment and layout. On the other hand, future expansions will require major investments in each of the main production areas. Current thinking favours a duplication of the existing plant. This can be done in conjunction with construction of a new hydroelectric plant at Picada (about 23 km from the zinc plant). At present, it is envisaged that this project will be realized during the latter part of 2003.

PUTTING CPM'S HOUSE IN ORDER

At the time at which the 86,000 t/y project was conceived, CPM was experiencing environmental problems. In addition, the government's environmental body had to approve the expansion plans before the project could be built and operated. More importantly, a license to operate the existing facility had never been issued. A submission to regularize the current situation was being prepared at the time of the expansion and this was submitted together with the preliminary expansion plans in order to obtain temporary approval for the project. This joint submission addressed the problems of jarosite disposal, quality of liquid effluents, inadequacy of the melting furnace exhaust system, the freon R22 gas refrigeration system used in the liquid SO2 plant, and the quality of the gaseous emissions from the roasting and sulfuric acid plants.

The emission levels from the acid plant received significant attention from the government and the public. The original fluid bed roaster off-gas cleaning equipment and interconnecting pipes were made out of lead. In early 1997, these were in such a precarious state of repair that a decision was made to replace the entire gas cleaning system during the 1998 biennial roaster/acid plant maintenance shutdown. It was also decided to take advantage of this shutdown to do other work:

• Replacement of the serpentine acid coolers with plate and frame heat exchangers • Replacement of the badly corroded No. 4 gas-gas heat exchanger • Renewal of all of the converter catalyst, including incorporation of some cesium-

based material in the fourth catalyst bed. This was done to modify the acid plant to meet the legislated emission limit of 2 kg of SO2 per t of H2SO4 produced. (Emission levels at the time were of the order of 3.0 - 3.5 kg of SO2 per t of H2SO4 produced)'.

• Additions and improvements to the control system, including partial transfer of the back-up panel to a computer-based supervisory system.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 263

During decommissioning of the plant after its overhaul, an accident occurred during calibration of the level gauge in the drying tower of the acid recirculation tank. Acid bounced over into the new No. 4 gas-gas heat exchanger and badly corroded some of the tubes. This corrosion resulted in openings in a number of places allowing a fraction of the cold SC^-rich gas to bypass the acid plant and join the hot SCh-depleted gas regularly exhausted via the acid plant stack. The consequences of this event were extremely elevated SO2 emissions and a visible plume caused by the generation of a fine mist upon contact of the two gas streams. This situation prevailed for some time, until, under extreme public pressure, the plant was shut down to address the problem. The No. 4 gas-gas heat exchanger was opened and tubes where iron sulfate deposits were found (indicating the entry of acid) were blocked off. Then, having passed a provisional pressure test, the heat exchanger was reconnected and the plant restarted. The plume was no longer visible; however, SO2 emissions continued to exceed the legislated limit. Attention turned towards identifying the causes of these high emissions and a number of contributing factors were reviewed:

• A small leak in the No. 4 gas-gas heat exchanger. Uncertainties as to the integrity of this unit led to the decision to replace it.

• Short-circuiting of gas through one or more of the converter beds either by uneven distribution of the catalyst, or by passage of some gas between the shell and the brick lining.

• Insufficient catalyst, or a drop-off in efficiency of the catalyst because of chemical and/or physical degradation.

A plan was formulated to replace the heat exchanger during a routine monthly shutdown and to expedite procurement of additional catalyst. Also, the actual performance of each conversion and adsorption stage was compared with the theoretical values. Disappointedly, replacement of the heat exchanger did not decrease the emission levels. Thus, attention turned towards studying means of increasing the catalyst load and/or sealing the converter vessel with a minimum of downtime.

As it turned out, ideas such as injecting catalyst into the beds via feed tubes were never realized because of a forced roast/acid shutdown in July 1999. Failure of the boiler water supply system resulted in a boiler tube rupture necessitating a total stoppage for three days. During this time, the No. 4 catalyst bed was opened. A sample of catalyst taken from the bed was analyzed and was found to have close to 100% activity. Some 2,000 liters of new catalyst were added and the plant was restarted. Little or no decrease in emissions was found and, consequently, a study to finally solve the problem was initiated. The options considered included:

• Addition of more catalyst - The existing converter beds still had some room for more catalyst. In the event that completion of existing beds failed to provide adequate conversion, then a fifth catalyst bed could be installed in a separate vessel. There were two negative aspects of this option. The first was that the acid plant blower was being operated at its maximum capacity and any additional pressure drop in the acid plant (from the addition of more catalyst), would require a blower upgrade, to maintain production levels or else production would drop. The second negative aspect was on a fall on public confidence if yet another attempt to solve the problem proved unsuccessful. This could happen if the converter top-up turned out to be insufficient.

264 LEAD-ZINC 2000

• Installation of a tail gas scrubber - With this option, gases leaving the acid plant would be treated with a chemical reagent to reduce their SO2 content to levels sufficiently low to allow their safe discharge to the atmosphere. The advantage of this option was that CPM had made a commitment to the environmental authorities to limit SO2 discharges during plant shutdowns and startups: A well designed scrubber could serve the dual role of treating gases during shut-downs as well as during normal operations.

Not surprisingly, the latter option was favoured by management. However, it was not until after a court order had been issued to close the plant that a commitment to have the scrubber up and running by November 7, 2000 was made. It was only as a result of this commitment and intense lobbying by employees, trade unions and pro-CPM community groups that the court order was revoked. Shutdown procedures had already commenced at the time that the court decision was reversed.

SCRUBBER OPTIONS

A survey of SO2 tail gas scrubbing systems in current operation indicated a wide range of reagent options. These included NaOH (or Na2C03), NH3, H202) seawater, CaC03) Ca(OH)2, MgC03, ZnO and zinc oxide calcine. Naturally, CPM's distance from the sea precluded the use of seawater. It was also decided to exclude reagents which could generate a solid residue requiring on-site storage. The reagent specification was left open to the project bidders; the only restriction being that its use should generate a saleable or recyclable product. On the receipt of bids, three options were analyzed that were based on H2O2 (Figure 1), NaOH (Figure 2) and zinc oxide calcine (Figure 3).

Gas Inlet

<7~ Blower

Hydrogen

Peroxide

c

-7K—7TC—7TC-

» / Λ,

Peroxide Tank

Stack

Venruri and Scrubber Tower Weak Acid

Figure 1 - Hydrogen Peroxide Scrubber Flow Diagram

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 265

NaOH Addition

Oxidation Tower

Air

r~\

v y

Sulphite Bleed

Gas Inlet

Absorption Tower

Heat Exchanger Circulation Tank

Blower

Stack

Gas Inlet <?" Blower

Figure 2 - Caustic Scrubber Flow Diagram

Absorption

Tower

Circulation Tank

Stack

Basic Plant

Water

" <

i<3 M

"

ZnO Calcine

"

jrindin

ixing T

Aci

Storage Bin

y-> g

ank

d

To Leach

Decomposition

Figure 3 - Zinc Oxide Calcine Scrubber Flow Diagram

266 LEAD-ZINC 2000

Choosing among these options proved to be a complicated procedure, which considered capital and operating cost, technical risk, flexibility and consistency with future strategies. Visits were made to sites using these reagents to carry out a first hand evaluation. The information obtained was collated and used to prepare Table 1. Then, monetary values were assigned to the indicators shown in Table 1.

Table 1 - Comparison of Scrubber Options H202 NaOH

Complete NaOH Simplified

Calcine Complete

Calcine Simplified

Capex Opex - Initial

Later Technical Risk Flexibility with Respect to Future Environmental Requirements

High High

Low High

Med/High High

Low Med

Med

Med Med

V.High Low

Low/med Med

Med/High Low

Med Med

H2o 2«-»2

Hydrogen peroxide technology has a high initial capital cost because of the need to have two scrubber vessels. The operating costs, initially high because of the cost of the reagent, may drop in the future as a result of either of the following events:

• Installation of an additional new acid plant designed so that its emissions, combined with those of the existing unit, meet legislated levels.

• Use of Caro's acid, H2SO5, generated in situ by electrolysis as the scrubbing reagent. The new desulphurization reaction may be written as:

H2SO5 + S02 + H20 -> 2H2S04 (1)

The fact that the current efficiency for the generation of Caro's acid is likely to be low would be of little consequence if it were produced using cheap hydroelectric power. A strong point of using hydrogen peroxide (or Caro's acid) as the scrubbing reagent is that the reaction product (dilute sulfuric acid) is amenable for recycle to the acid recirculation towers in the main acid plant or can be used in the leaching plant SAL stage. Another highly favourable factor of this process is that the extremely rapid reaction kinetics allows very low SO2 emission levels to be attained. This is important in light of the ever increasing environmental limits. This will also put pressure on our competitors in the sulfuric acid and liquid SO2 markets to improve their operations and limit their SO2 emissions.

NaOH

A sodium hydroxide tailgas cleaning plant offered by a local engineering company did not have any of the advantages shown by hydrogen peroxide. From an operating cost perspective, NaOH is expensive. Although the partial, or complete, substitution, of ammonium jarosite with its sodium analogue would help offset this expense, the future possibility of implementing iron removal by pyrometallurgical means would neutralize this gain.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 267

In addition, the by-product (sodium sulfite) could cause problems if the simplified version of the flowsheet were to be adopted. In addition, the recycle of sodium sulfite to the roaster may contribute to the build-up of accretions in the roaster and in the boiler. Also, operating problems that could be experienced at higher rates could be blamed on this recycle whether or not it was the underlying cause of the operating problems. Direct incorporation of this material into the leaching plant could generate SO2 gas from tanks and launders. This could result from incomplete oxidation of the sulphites to sulphates. The use of a dedicated oxidation tower, or of an acid decomposition stage, represented an extra step and an additional capital expenditure.

Also of consideration when comparing technologies is the possibility that, in the future, environmental authorities may specify that the best available practice must be adopted rather than defining specific numerical limits. In that case, sodium hydroxide scrubbing would lose out to its hydrogen peroxide competitor because Sulfites in the recirculating solution would always allow some SO2 to escape because of the higher SO2 partial pressure in the gas stream.

Calcine

Calcine scrubbing, characterized by very low operating costs because of the use of "free" calcine, imposes a large trade-off between capital expenditure and process control. Although it is assumed that the calcine grinding and/or acid decomposition stages could be eliminated from the flowsheet, the risk of having to retrofit one or the other, or both of these, was unacceptably high. Another concern is the tendency for the calcine pulp to settle out in tanks and pipelines. For example, the operating unit of the Hindustan Zinc Vizag plant requires regular cleaning of the recirculating tank, tower base and distribution manifolds.

Final Decision

After evaluation of the three processes, the final decision was made based on practical considerations. The calcine scrubber, being offered under license by Mesco in Japan, placed the onus of realization of the project within CPM. Since CPM is not set-up to carry out this type of project, turnkey alternatives were preferred. Then, comparing the hydrogen peroxide and sodium hydroxide options, the possibility of the participation of the hydrogen peroxide bidder in other projects was the deciding factor. A partnership of this nature would decrease mobilization and other costs, and would simplify the administration of the project. The hydrogen peroxide scrubber is scheduled to come on-line in early November 2000.

EXPANSION OPTIONS

Since the inception of the concept to expand production, the actual mechanism has been continuously refined, resulting in radical changes in certain areas and "polishing" in others. A summary of the original ideas for the main unit operations together with the alternatives, which evolved during the course of the evaluation, is presented in Table II. The decision path was as follows:

268 LEAD-ZINC 2000

Small new FBR + small new electrolysis 4-Turbocharging of existing FBR + small new electrolysis I Turbocharging of both existing FBR and electrolysis

The major considerations that affected the outcome of the study were:

The lack of availability of capital. Because of the extreme difficulties encountered in trying to obtain a bank loan, the decision was made to finance the entire project from CPM's operating budget. The desire to maximize reutilization of the equipment for the proposed twofold increase in plant capacity. Or, in other words, to avoid buying new equipment or carrying out modifications that could become redundant in 2-3 years. The trend towards using operating equipment in the most economical way. Experience has shown that in many cases it was possible to gain higher throughputs in certain areas with relatively small increases in maintenance cost and downtime. This factor had a large impact on the capital cost needed to double the capacity of the plant.

Table II - Summary of the Original Ideas for Expansion Along with the Alternatives Area Original Idea Alternative(s) Roasting H2S04 / S02 (1)

Leaching

Electrolysis

Casting

Add a small roasting train Replace the existing SO2 (1) plant with a new, larger unit

Add a new jarosite filter Add a new jarosite tank Upgrade pumps/piping lines where necessary Automate Install a small rectifier and additional banks of cells

Install additional spent electrolyte cooling towers

Install a new small melting furnace

"Turbocharge" existing unit Upgrade the capacity of the existing plant Install a new smaller SO2O) plant alongside the existing unit Convert the existing "basics" filter to jarosite and install a new "basics" filter

Install a large rectifier and additional banks of cells Install a small rectifier and "turbocharge" existing cells Convert existing pure solution cooling tower to spent electrolyte and install a new larger pure solution cooling tower Increase the cooling capacity of the existing towers "Turbocharge" the existing furnace with the addition of a further inductor Install a new larger melting furnace

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 269

EXPANSION DETAILS

Roasting

The Roaster

Lurgi designed the existing roaster. It has a grate area of 46.4 m2. During its first years of operation it was fitted with a refractory ring to reduce the effective grate area to 40.4 m2. In 1985 this ring was removed and in 1987 oxygen enrichment commenced. In 1988, a fourth serpentine cooling coil was installed in the bed itself. Since then, roaster throughputs were restricted by concentrate quality, which, in turn, varied significantly.

High iron concentrates, in particular, with their large heat generating characteristics limited the maximum roaster feed rates to approximately 20 t/day of dry concentrate per 1% of iron. In 1996, when CPM became part of the Paranapanema group, the decision was made to purchase only concentrates so that the annual average concentrate iron level did not exceed 6%, and also to arrange shipments in such as way to avoid months in which the average iron level would be significantly higher than 6%.

Also at this time studies were made to identify ways of increasing the roaster throughput. Computer models were set up and it was identified that significant gains could be obtained by either raising the bed temperature (Figure 4) and /or by increasing the addition rate of overbed cooling water (Figure 5).

_ 410 T3

a. ? 400 -' a Q>

LL

S 390

§ 380

940 960 980 1000 1020

Bed Temperature (°C)

Figure 4 - Concentrate Feed Increase with Bed Temperature

The bed temperature operating range, which had been typically set at 920 - 940°C was adjusted to 940 - 960°C with a resultant throughput gain of approximately 10 t/day (2.5%) of concentrate. The possibility of increasing the addition rate of overbed water, however, became a contentious issue and was subjected to further studies. A visit was made to the Pasminco Operations at Risdon in Australia where free use of overbed water is practiced in their two fluid bed roasters. Operating problems which were being experienced there, whether related to the use of overbed water or not, were used to construct worse case scenarios where the possible impact of increasing overbed water could be closely evaluated. This analysis showed that even in extreme circumstances (e.g., under increasing periods of plant shutdowns required for

270 LEAD-ZINC 2000

cleaning or unplanned maintenance) the gains to be had by increasing overbed water usage were significant.

___ 450 ■σ

a. S" 430 <D a> u_ » 4 1 0 E § 390 j

c o O

370 -1,0

Figure 5 - Concentrate Feed Increase with Overbed Water (Bed Temperature 950°C)

The decision was eventually taken to increase overbed water usage in small incremental amounts and to closely monitor any resultant changes in performance. A simple reduction in the addition of water could then be quickly made if the model predictions could not be validated. At the time of writing this paper, overbed water was being added at a rate of between 4 and 5 m3/h. This represents a throughput gain of an estimated 40 t/day (11%) of concentrate with no apparent ill effects.

To further increase the capacity of the roaster to obtain the rates required to produce 86,000 t/y of marketable metal, the addition of more overbed water and the installation of a fifth serpentine cooling coil in the bed were considered.

The first option was simple and easy to implement, but it would increase the cooling load on the gas purification section. The second option would produce more steam, but it could crowd the roaster bed and impose tighter time limits on future routine shutdowns because of the increase in the fixed rate of bed cooling. Calculations indicated that the steam produced in the first option would be sufficient to meet the demands of the leaching plant, and consequently, the use of additional overbed water was adopted as the preferred strategy.

Boiler

It was soon realized that the only effective means of removing the additional heat from the boiler would be to install additional tube bundles. Boiler manufacturers were consulted and they confirmed that there was sufficient room within the existing unit, in the space presently used as a radiation chamber, to install two additional bundles. Foster Wheeler, with their spring-hammer designed cleaning system, was the successful bidder.

Gas Purification

The recently replaced gas purification section was already operating outside its design parameters because of the extra heat load placed on the system as a result of increasing the overbed water usage in the roaster. The plate heat exchangers of the humidifying tower/wash

2,0 3,0 4,0 5,0 6,0

Overbed Water (m3/h)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 271

tower recirculation system and their associated water cooling towers did not have sufficient capacity to cope with the requirement to recondense this additional steam, and to cool the associated gases to the maximum temperature necessary for subsequent treatment in the sulfuric acid plant. In order to compensate for this, fresh cold make-up water was being added directly to the wash tower at rates significantly higher than planned. The subsequent bleed of acid water to the effluent treatment plant was suspected of being a major contributor to the operating problems within that section.

With the expansion, the heat removal burden in the gas purification section will increase. This in particularly so because of the decision to use direct water addition to control the roaster bed temperature will result in an equivalent quantity of hot water vapor entering the gas purification section along with increased amounts of SO2 and O2 gases. The progression of the heat removal duty of the gas purification section derived from the modeling studies is shown in Table III.

Table HI - Heat Removal in the Gas Purification Section Roaster Overbed Heat Removal Water (Gas Purification)

Equipment Requirements

M3/h GJ/h Heat Exchangers Cooling Towers Original Design

Intermediate Operating Stage Present Operations

Future Operations

1.0

3.0

4.0

6.0

30.0

33.4

37.4

40.3

One (50% capacity) One (75% capacity) One (100% capacity) Two

The following steps will be taken to remove this additional heat:

• Completion of the plates in the existing heat exchanger together with the addition of another complete heat exchanger unit

• Installation of an additional cooling tower • Upgrade of the recirculating pumps • Upgrade of the nozzles in the humidification tower.

In the expanded plant it is intended to dramatically reduce the volume of acid water generated in the gas purification section as part of CPM's environmental commitment to its long-term "zero effluent" target.

Oxygen Supply

When oxygen enrichment of the roaster air first came into practice, liquid O2 was purchased from a nearby company. In 1999 this same company installed a swing absorption gaseous oxygen plant (VPSA) adjacent to CPM's roaster. The roaster operation was then modified to use this gas, with liquid oxygen being maintained as a back-up in case of VPSA outages, and as a top-up during peak consumption periods. Modeling showed that the oxygen requirements would roughly double upon expansion to 86,000 t/y. The best means of achieving this oxygen supply capacity increase is highly dependent on the philosophies that would be

272 LEAD-ZINC 2000

adopted for the proposed plant duplication. In rum, these philosophies need to consider the cost of oxygen. For example, the obvious concept of alleviating the existing roaster on duplication of the roaster circuit, so as to no longer use oxygen enrichment, would have implications on the requirement for a much larger new roaster. This might not be the most cost effective option. A study of the situation identified the five scenarios listed below:

• Use locally purchased liquid oxygen for both present and future expansions. • Use locally purchased gaseous oxygen for both present and future expansions. • Use locally purchased liquid oxygen for the present expansion and then modify the

roasters to operate without oxygen enrichment upon duplication. • Install our own VPSA gaseous oxygen plants for both present and future expansions and

operate them using power purchased from the state grid. • Install our own VPSA gaseous oxygen plants for both present and future expansions and

operate them with our own power.

The capital and operating cost estimates for each of these options were used to prepare Table IV.

Table IV - The Influence of Duplication Options on Oxygen Supply Option Investment Operating Cost Net Present Value

(US$ millions) (US$ millions /year) (US$ millions) Purchase 02 (1) 0 3.5 (26.7) Purchase 02 (g) 0 2.7 (20.6) Bigger Roaster 10.0 0 (10.0) Produce 02 with state 3.5 0.7 (8.1) power Produce 0 2 with own 6.5 0.2 (7.9) power

It is clear that the best option is to operate two "turbo-charged" roasters using oxygen from our own VPSA plants powered with our own electricity. This is also consistent with CPM's strategic plan of being energy self-sufficient. However, because of capital restrictions, if the duplication project goes ahead unaccompanied by a new hydroelectric plant and the larger roaster scenario is selected, it is imperative that CPM be free of any long term obligation to purchase, or produce, gaseous oxygen. For this reason, and because of the obvious impracticability of installing a VPSA plant plus a small hydroelectric generating facility at the present time, the use of locally purchased liquid oxygen was chosen for the 86,000 t/y expansion.

Roasting Peripherals

A 24-h test demonstrated that the materials transport equipment and the calcine rotary drum cooler had sufficient capacity to operate at the necessary expanded rates but that the calcine ball mill might have to be upgraded. At the time of writing, the ball mill performance was being closely studied.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 273

H2S04/Liquid S0 2

Acid Plant Blower

Gas flow rates through the roasting, gas purification and acid plant sections are currently limited by the capacity of the existing KKK blower. The requirement to increase these flow rates upon expansion will need to be met. Four possible means of achieving this were indicated:

• Replacement of the existing blower with a new larger unit • Upgrade of the existing blower • Installation of a new blower in parallel • Installation of a new "booster" blower in series.

The cost of the first option was prohibitively high, whereas the second, which was also expensive, left doubts as to whether an old unit was capable of integrating new components in a reliable way. Option 4 was already effectively being carried out as part of the scrubber project. However, to guarantee that the production rates would not be reduced in case of a scrubber blower outage and to leave open the option to increase the amount of catalyst in the beds (to save on scrubbing reagent yet increasing the pressure drop across the acid plant), option 3 was adopted in tandem. Having excess acid plant blower capacity could result in oxygen savings if future operating experience showed that the plant could cope with higher flow rates so that oxygen usage could be partially, or completely, substituted by air.

Liquid S02

One of the key attractions of the 86,000 t/y expansion project was the possibility of increasing our share in the lucrative local liquid S02 market. Since the inception of the original expansion idea, it was always intended that the additional SO2 generated as a result of increased roasting rates would be converted to liquid SO2 rather than to sulfuric acid. As in most other cases, there existed a number of options for increasing the production capacity of liquid S02:

• Replacement of the existing plant with a new larger plant • "Turbo-charging" the existing plant • Installation of a new, smaller plant along side of the existing unit.

A complicating factor in evaluating the options was the fact that the existing plant uses freon R-22 as refrigerant and that its use, although not prohibited, is being phased out. Budget prices for 60 t/day liquid S02 plants were sought and were found to be extremely expensive. To "turbo-charge" the existing plant, it would be necessary to convert it to another type of refrigerant and then carry out the necessary additions/modifications. Alternatively, an effort could be made to convince the local environmental authorities to allow us to expand the capacity using freon R-22 in exchange for a commitment to replace it with another refrigerant at some future date. Neither of these alternatives was attractive, nor was the concept of scrapping a functional plant with many years of potential life.

Thinking was turned towards the third option and tenders were called for a 35 t/day liquid S02 plant. Bids were received for two radically different technologies; namely, a conventional partial condensation process using ammonia and absorption using an aqueous amine solution.

274 LEAD-ZINC 2000

The problems of amine emissions at levels detrimental to health and an objectionable odor, that had plagued the latter process in the past, were said to have been overcome. In addition, the latter process was claimed to be capable of carrying out the dual role of tail gas scrubber as well as liquid SO2 producer if positioned to take acid plant tail gases. However two glaring problems existed with this option. No full scale commercial plant using this particular amine is yet in existence. Even though pilot scale tests had been carried out with some success, and considering that this technology is closely analogous to the Sulfidin process (based on a mixture of xylidine and toluidine), the Solinox process (based on polyethylene glycol, dimethyl esters), the UCAP process (triethanolamine) and a process based on Ν,Ν-dimethylaniline which was developed by the American Smelting and Refining Co., CPM did not want the risk associated with being a proving ground. CPM's remote location was also a concern. If problems were to occur, it would be almost inevitable that delays would result in obtaining technical assistance and/or spare parts. In light of these problems, the decision was made to install a 35 t/day liquid SO2 plant using the partial condensation process based on ammonia as the refrigerant.

Leaching

The existing leaching plant was in the comfortable position of having excess capacity in most stages. Thickeners, each of 600 m3 working volume, were more than adequate for their respective duties. Residence times in each process stage were adequate, with the possible exception of the jarosite precipitation section. The decision to install an additional tank in this section was made upon the advice of an expert in jarosite technology. The limitations of jarosite filtration capacity were well recognized from operating experiences during times when relatively high iron concentrates were treated. The decision to install an additional jarosite filter was made on the basis of a request from operations personnel. The need to amplify/modify areas such as pumping systems, launders, heating systems, filtration duties, etc. was put to the test during a 24 hour trial in which the leaching plant was operated at rates consistent with the production of 86,000 t/y of zinc. This trial identified the need for few, relatively minor, changes in these areas.

In the area of automation and control, automatic on-line analyzers for Co and Cd were provisionally included as first-cut items on the basis that they would be purchased if the must-have items would bring the project under budget. Further automation items such as level controllers, flow controllers and variable speed drives would also be included if funding permitted.

Electrolysis

The two existing silicon diode rectifiers each have a nameplate rating of 26 kA. However, after relatively minor modifications carried out by the supplier (Siemens) and with their close technical support it has been possible to operate both units at 28 kA. This amperage is very close to the 31.5 kA that would be required for the expansion. The idea of turbo-charging the existing cells, rather than installing new ones emerged. Tests in which a number of cathodes was removed from closely monitored cells at various points in the cellhouse to simulate operation at high current density were carried out with success. Calculations indicated that the busbar system was capable of handling the higher loads. Consequently, the turbo-charging option was chosen: 2 small 6 kA rectifiers would be connected in parallel with the existing rectifiers effectively increasing the current density to 650 A/m2 during normal operations.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 275

Also, to minimize additions and modifications, it was decided not to alter the flow rate of recirculating spent electrolyte, but rather to change the cooling system to allow the cells to operate with a higher temperature differential. Studies showed that the cooling capacity of the existing cooling towers could be increased to levels consistent with those required at the expanded production rates by modifying the spray nozzles, and that additional towers would not be needed.

Melting/Casting

The existing single Demag induction furnace has a melting capacity of 9 t/h of cathode zinc, yet it presently treats close to 10 t/h by operating in a rather precarious fashion with bath temperatures typically in the range 440 - 450°C. A feature of this furnace is that it operates with only 3 inductors. Although provision was made on the furnace to house a fourth inductor, the electrical system was designed to serve only 3 inductors. Attempts to commission a fourth inductor using a separate power supply were unsuccessful. After a number of attempts, this path was abandoned. Two alternatives were to install a small melting furnace of about 3 t/h capacity or to install a larger melting furnace of 9 - 10 t/h capacity and operate the casting section only during Monday to Friday.

The second option was adopted because it provided more catch-up capacity, and it was consistent with conceived future expansions. Labor savings from reducing the number of operating shifts more than pay for the extra investment. Casting and zinc blowing facilities are adequate for the expanded production rates.

Other Areas

Areas such as concentrate storage and transport, air compression, cementation residue treatment, CUSO4.5H2O production, silver concentrate flotation, water treatment, maintenance workshops, anode casting and cathode fabrication were all considered to be capable of meeting the additional demands that would be placed on them. The effluent treatment plant was already undergoing an upgrade and so it was easy to ensure that it would be capable of handling the increased duty.

THE NEXT PHASE

Once the de-bottlenecking exercise was completed, the next step was to look at doubling the capacity of the plant. This large expansion depends on five fundamental factors, the first of which is the cost of electrical energy. With the commissioning of its hydroelectric plant at Sobragi, CPM is effectively energy self-sufficient at existing production levels. It is intended that CPM will retain this self-sufficiency at the duplicated production rate and for this reason several alternative hydroelectric schemes are being studied. Secondly, labor costs in Brazil are low and are likely to remain so for many years. This indicates that a cellhouse with manual stripping would be economically favoured over one equipped with automatic stripping. Furthermore, CPM has a social responsibility to provide employment to the local community whenever practical. Environmental restrictions are also a factor. At present, land disposal of jarosite is accepted by the environmental authorities as an appropriate means of eliminating iron. This suggests that the current RLE technology with iron removal as ammonium jarosite should be maintained. A possible changeover to the jarofix process should be considered as a

276 LEAD-ZINC 2000

medium term option (10 - 15 years), with pyrometallurgical options such as Ausmelt, or a Waelz kiln, reserved as long-term possibilities (25 - 30 years). At present, the Brazilian markets for sulfliric acid and liquid SO2 are healthy and growing. Therefore, it is safe to assume that increasing our sulfuric acid production will not result in potential production cuts as a consequence of the inability to sell these products. Finally, the availability of financing is a major concern. This is likely to be a limiting factor in determining the means by which CPM ultimately expands. The fact that the RLE option with manual stripping will be the least expensive option weighs heavily in our favour, if the investment were limited to a bare minimum by financial restrictions.

ACKNOWLEDGEMENT

The authors wish to thank the management of Companhia Paraibuna de Metais for permission to publish this paper.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 277

THE BOLESLAW ELECTROLYTIC ZINC PLANT

D. Krupka Silesian Technical University

2A Akademicka Street 44-100 Gliwice, Poland

B. Ochab and J. Miernik ZGH BOLESLAW 37 Kolejowa Street

32-332 Bukowno, Poland

ABSTRACT

The electrolytic zinc plant, with a capacity 75,000 tonnes per year, is part of Zaklady Gorniczo-Hutnicze BOLESLAW. It is the biggest zinc producer in Poland. The electrolytic zinc plant started up in 1955 with a capacity 15,000 tonnes per year, and multiplied its production by 5 till nowadays. At the beginning, the plant processed only calcine from zinc oxides (60% Zn, 3-4% Pb, 0,03 % Cd) which were produced from Waeltz fumes by calcination in a rotary kiln. Participation of calcine from zinc sulfide concentrates in the feed grew during the BOLESLAW plant expansion. Currently, only roasted sulphide concentrates, calcine from fluid-bed roasters, is processed. This paper presents a general description of the ZGH BOLESLAW plant and discusses the processes of roasting, leaching, purification and electrolysis. Also shown are the chemical analyses of the various process materials.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

278 LEAD-ZINC 2000

INTRODUCTION

The electrolytic zinc plant started up in 1955 as a part of the mining and metallurgical complex Zaklady Gorniczo - Hutnicze "BOLESLAW" which had a mining and metallurgy tradition rooted in the ΧΕΠ century. Currently existing as a government enterprise, ZGH BOLESLAW owns two Zn-Pb mines OLKUSZ and POMORZANY, a concentration and flotation plant OLKUSZ - POMORZANY, a pyrometallurgical waste treatment plant, RECYCLING S.A. a subsidiary company, a roasting plant with a sulfuric acid plant, an electrolytic zinc plant and auxiliary plants. The Zn-Pb ores from local deposits are poor in comparison to world deposits. Their global zinc and lead contents are to 6 - 7%. Sphalerite and galena are the main minerals. The ores are concentrated by selective flotation with prior concentration in heavy liquids. This prior concentration allows the separation of coarse, dolomite, in quantity about 40 % of the total mine output. The whole production of dolomite in quantities exceeding 1,000,000 tonnes per year is sold as a valuable commercial product. Zinc sulphide concentrates from the concentration plant contain about 53 - 55 % Zn and 1.5 % Pb, and the galena concentrate contains about 50 % Pb and 1.5 % Zn.

ZGH BOLESLAW produces:

• Electrolytic zinc • Sulfuric acid • Zinc concentrates • Lead concentrates • Dolomite • Zinc oxides (as fume and as calcine) • Zinc alloys • Magnesium sulfate • Zinc sulfate • Cadmium sponge.

ZGH BOLESLAW is the biggest zinc producer in Poland. It recently undertook many development and modernization works. These activities aimed at a zinc production of over 75,000 tonnes annually, about 5 times more than the starting production of 15,000 tonnes per year.

RAW MATERIALS

Over the period 1955-1961, the electrolytic plant processed only calcined zinc oxide having a composition of 60% Zn, 3-4% Pb and 0.03% Cd .The zinc oxide was produced by calcination of the fumes from the Waelz process at temperatures from 1420 to 1470 K. Zinc carbonate ores, Zn-Fe residues and other zinc-bearing materials were used as raw materials in the Waelz process. After enhancement of the electrolytic zinc plant in 1962, the importance of roasted sulphide concentrates was continuously growing in the feed. Currently, only the calcine from fluid-bed roasters is processed. Table I shows the analysis of the raw materials processed over the period 1998-1999.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 279

Table I - Chemical Analy

Concentrate

Calcine

Zn 54.57

63.56

sis of Zinc Raw Materials at ZGH BOLESLAW (1997 -

Pb 1.5

1.87

Cd 0.27

0.31

Fe 4.4

4.97

MgO 1.0

1.03

CaO 1.63

2.25

Cu 0.028

0.048

Si02

0.7

0.65

Ss 31.5

0.1

1998)%

SSO4

0.24

2.33

ROASTING AND GAS TREATMENT

The roasting and sulfuric acid plant was commissioned in 1969 as a facility equipped with two fluid-bed roasters having a hearth surface 15 m2 each and single contact acid facilities. During the following years the plant was modernized and expanded. Currently, two fluid-bed roasters, each of a hearth area of 27 m2, with waste heat boilers are in use. Gases are processed in converters with four shelves of MONSANTO vanadium pentoxide catalyst. A double conversion system increases the conversion of SO2 to SO3 and reduces SO2 emissions. A ROSEMOUNT System was introduced to control the roasting and flue gas processing in 1994. The system provides precise, remote and central control of the whole process including: the furnace with its waste-heat boiler, the scrubber, the electro-precipitators, conversion apparatus, sorption towers, blowers, acid drains treatment and sulfuric acid storage.

The application of the Rosemount System results in a good quality of roasted concentrates for subsequent processing and ensures the compliance of the stack gasses and waste waters with the new environmental regulations (SO2 content in stack gasses is kept below 120 mg/Nm3). The wastewater from the sulfuric acid plant is processed in the acid waste water plant (commissioned in 1996) in connection with the acid drains from other plants.

WAELZ PLANT

The Waelz Plant was commissioned in 1952 and is equipped with thirteen Waelz kilns and two rotary kilns for calcination of the Waelz fumes. The specifications of both kilns are as follows:

• Length 40 m • Outer diameter 3.0 m • Lining thickness 250 mm • Inclination 2°/4° • Rotational speed 0.67 rpm.

The Waelz plant processed 400,000 tonnes/year of zinc carbonate ore typically containing 10 % Zn and 2 % Pb and Zn-Fe residues with zinc contents of about 16-18 %. After running out of zinc carbonate ore deposits and because of environmental reasons, nine kilns were liquidated in 1989. Currently, six kilns are operated: 4 as Waelz kilns and 2 for calcination of the Waelz fumes.

At the same time, changes were introduced in the feed: residues from the electrolytic zinc plant are processed with some addition of zinc-bearing materials from the waste water treatment plant, and other secondaries like EAF dusts coming from steel mills (electric arc

LEAD-ZINC 2000

furnace dusts). Table II shows the typical analysis of the residues from the electrolytic zinc plant.

Table II - Chemical Analysis of Fe-Zn Residues, %

Zn 14.0

Pb 5.58

Cd 0.25

Fe 17.9

MgO 1.49

CaO 8.68

Cu 0.12

Si02

2.7 Ss

0.31 SSO4

8.08

About 50,000 tonnes per year of different zinc-bearing raw materials are processed with typical contents of 19 % Zn and 4 % Pb . This allows the production of about 13,000 tonnes of calcinated ZnO (55-62 % Zn and 12 % Pb) as well as 2,600 tonnes of Pb-Zn-Cd concentrate per year. Coke in the feed approximately equals 35 %. Waelz fumes contain 47-55 % Zn, 15-20% Pb and 0.5-1 % Cd. Slag is formed in quantities of about 38,000 tonnes per year with a typical analysis of about 2 % Zn and 0.4 % Pb. Addition of silica to the feed results in the formation of a glassy slag allowing its utilization in mining (for excavation filling) and road building. Table III shows the chemical analysis of the Waelz slag.

Table III - Chemical Analysis of the Waelz Slag, %

Zn 2.0

Pb 0.4

Cd 0.06

Fe 21.0

MgO 2.5

CaO 11.0

Cu 0.2

Si02

15.0 Sog 4.9

C 28.0

LEACHING

After the commissioning of the electrolytic zinc plant, the leaching process was adopted to treat only the calcine of the zinc oxides (60 % Zn, 3-4 % Pb and 0.03 % Cd) produced from the Waeltz fumes by calcination in a rotary kiln. This material is very difficult for zinc extraction by hydrometallurgical methods. Leaching was carried out by reverse leaching at low acidity. Because of the high level of colloids in the slurries, solid/liquid separation created many problems. Slurries with high viscosity were practically impossible to clarify and the filtration rate was rather low. The high specific volume of the residues resulted in a great amount of occluded zinc. For these reasons, reverse leaching was recognized as correct, but resulted in low zinc yields. Participation of calcine from zinc sulphide concentrates grew during the BOLESLAW plant expansion. Currently only roasted sulphide concentrates, calcine from the fluid-bed roasters, are processed.

The leaching process has been modernized. Primarily, the new leaching flowsheet was prepared on a basis of our own research and development work connected with the experiences of the world's zinc electrolysis plants which apply iron precipitation in the form of goethite. This allows zinc yields at the level of 95 to 98 %. The high zinc yield is the main advantage of the goethite method. From the other side, the main problems at the erection of the plant and its operation are connected with the severe requirements for the leaching and precipitation equipment, heating to 360-370 K and especially with difficulties in the utilization/storage of goethite residues. It was found that the goethite method would probably be the best, but only for new plants that are processing local concentrates. An additional limitation was caused by fact

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 281

that the chosen method has to be applied in an old, existing plant in rather difficult technical and economical conditions. It was necessary to consider the presence of existing buildings and facilities, local resources, and the possibility of residue treatment in the Waelz plant, which was situated in the neighborhood. Accordingly, it was decided to modernize the process in the way shown in the flowsheet in Figure 1.

The first stage of leaching, the so-called acid-neutral leach, is operated continuously. About 90 % of the zinc is leached from the calcine. The slurry is prepared for further processing by solution purification and zinc electrolysis. Neutral slurry is classified in hydrocyclones for zinc-bearing sands recovery. The main objective of this stage is to recover the surplus of calcine that was used for neutral leaching. This stage has to be prepared and operated in a very precise and careful way. The overflow from the hydrocyclones is transported by gravity to Don-thickeners. The overflow from the thickeners is pumped to a solution purification plant and the underflow is pumped to low acid leaching tanks. After classification, the pulp is washed and filtered at Larox"filters. The Fe-Zn cake from the filters is transported by conveyors to the Waelz plant. Underflow from the hydrocyclones is pumped to next stage of leaching. Zinc-bearing sands from the hydrocyclone underflow are processed using REDOX acid leaching for recovery of all zinc compounds from the sands (including zinc ferrites), as well as for the preparation of manganese sulfate and ferric sulfate solutions having the concentrations required in the first stage of calcine leaching. Sands leaching is carried out in two stages. In the first stage, the sands are leached with spent electrolyte and sulfuric acid (initial acidity: 20-25 % of H2SO4) in the presence of a reducing agent. As reduction agents can be used ZnS concentrates, Waelz fumes or crystalline ferrous sulfate. Same MnC>2 is added to the leaching tank in the second stage of leaching (oxidation stage), for the oxidation of iron compounds according to the following reactions:

Fe2(S04)3 + ZnS = 2FeS04 + ZnS04 + S (1)

2FeS04 + Mn02 + H2S04 = Fe2(S04)3 +MnS04 +2H20 (2)

ZnS + 4Mn02 + 4H2S04 = ZnS04 + 4 MnS04 +4H20 (3)

The leaching process requires temperatures of about 360 K and a reaction time of 4 hours (2 hours in the reduction stage and 2 hours in the oxidation stage). When sands with an analysis of 39.9 % Zn, 8.94 % Fe, 2.32 % Pb and 1.76 % Mn were leached, the best results were obtained with some addition of Waelz fume as a reductant. The zinc yield from the zinc bearing sands was obtained at the level 93 % with a solution analysis: 157 g/dm3 Zn, 24 g/dm3 Fe, 20 g/dm3 Mn, 40 g/dm3 H2S04 for an addition of 10 % fume. Both the research work and industrial practice made it clear that there is no need for sands grinding before leaching to obtain a good yield of zinc and to produce a solution with the ferric contents at the level required in the first stage of calcine leaching. The leaching plant flowsheet is shown in Figure 2.

Calcine is transported from the roasting section to the leaching plant by railway cars for a distance of about 200 meters and next is unloaded either in six silos with a total capacity 3,200 tons of calcine or directly to a mixer. Unloading directly to mixer is the most common practice. Spent electrolyte (free acid content 10-15 % H2S04) is added to the mixer in the proportions from 1:3 to 1:4. The outlet is pumped to the leaching tanks. Spent electrolyte from a launder situated above the leaching tanks continuously feeds the mixer and three in-line leaching tanks, each with a capacity 80 m3. The first leaching tank is fed with slurry from the mixer, solution from sands leaching, manganese dioxide and spent electrolyte to produce a slurry of 3.0 pH. The

282 LEAD-ZINC 2000

outlet from the first leaching tank is pumped to the second tank and subsequently to the third leaching tank. The slurry pH from the second tank is 4.8, and from third one is 5.1.

ZnS concentrate

spent

electrolyte

ROASTING (fluid-bed roaster)

calcine

NEUTRAL LEACHING

'

neutral slur

CLASSIFICATION (hydrocyclones)

' '

THICKENING (Dorr thickeners)

' overflow

PURIFICATION FILTRATION

'

purified , electrolyte

COOLING (cooling towers)

' '

ELECTROLYSIS

'

cathode zin

MELTING CASTING"

}

process gasses SULFURIC ACID PLANT

sulfuric acid

ry

underflow

Γ ■

' » ■

underflow

c

'

1 H2S04.

HIGH ACID LEACHING

acid slurry

LOW ACID LEACHING

' '

CLASSIFICATION (hydrocyclones)

1 overflow

THICKENING (Dorr thickeners)

< underflow

FILTRATION &

WASHING

< Fe-Zn cake

1

W A E L Z

P R O C E S S

' '

<4 1

underflow

overflow

filtrate

Waelz slags

ELECTROLYTIC ZINC Clinker for ISP

Figure 1 - Flowsheet for ZnS Concentrate Processing at ZGH BOLESLAW

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 283

u M u u

Sands leaching

overflow

overflow

Dorr Thickener

Φ9,5 m ©

underflow

Fe-Zn cake for Waelz i © CLZD 0 o ()

() 0 ()

(X D o—o π—n o—n

filtrate

Pressure Filters

Suspension

160 m2 © _ a

Zn dust

Filter Presses

® © 40 m3

2nd stage purification tanks

Purified ZnS04

solution

160 m2 ©

, i ,t |t t

80 m3 ©

Slime leaching

Zn dust

Filter presses

© © ® © © IP V

40 m3

1 stage purification tanks

Cd cake

Figure - 2. Flowsheet of the Leaching and Purification Sections

284 LEAD-ZINC 2000

Neutral slurry is classified in hydrocyclones (350D/120). Underflow from the hydrocyclones is pumped to acid leaching, and the overflow to five Dorr thickeners, each with a diameter of 9.5 meters. Magnafloc 338 is used for this process (0.017 kg per tonne of cathode zinc). Underflow from each Dorr thickener is pumped to low acid leaching, and after classification and washing, to Larox filters. The filter cake with a maximum moisture content of 17 % is fed to the Waelz kilns.

SOLUTION PURIFICATION

Solution clarified in the Dorr thickeners is next purified by cementation. Purification is carried out continuously in two stages, at temperatures of 323-328 K, and with filtration after each stage. First stage equipment consists of five purification tanks, each with a capacity 40 m3, partly in-line and partly in-parallel, equipped with mechanical mixers with a rotary speed 70 rpm. Output from the last purification tank is filtered in four plate filters with a surface area of 160 m2. This filter cake is the raw material for cadmium production. Solution from the first stage, with an analysis varying from 20 to 50 Cd mg/dm3, is purified with Zn dust in the second stage. The purification is carried out in three tanks in-line. Each of them has capacity 40 m3. The Zn dust is added in the form of a water suspension. Overflow from the purification tanks is filtered in three Hoesch filter presses with a surface 160 m2. The cake is returned to the first stage of purification, and the purified solution is pumped to Hamon fan cooling towers via a 50 m3 tank. Zn dust is consumed in a quantity of 25 to 27 kg per tonne of cathode zinc. Table IV presents the chemical analysis of the zinc sulfate solution before and after purification.

Table IV- Chemical Analysis of the ZnS04 Processing Solution, g/dm3

Zn Cd. Cu Fe As Ge Co Cl Mn Mg Before 135 1.2 0.15 0.007 0.0002 0.0004 0.0035 0.15 1.5 14.0 purification After 135 0.0004 0.0001 0.005 0.0001 0.0001 0.0035 0.12 1.6 14.0 purification

The cooled solution is transported by gravity to four 800 m3 in-line tanks from which it is fed to the cellhouse.

ELECTROLYSIS AND MELTING

The cellhouse is equipped with 528 cells having a capacity of about 2.5 m3 each in cascade-lines. Each cell is equipped with 30 cathodes and 31 Pb-Ag anodes and two water coolers. Inlet water at a temperature 286 K is delivered from a local mine. The cells are grouped in two units. Each unit with 22 double cascades of six cells in each cascade-line. Electric current goes to the electrodes through copper bus bars. Each cascade is connected in-series, and the electrodes in the cells are in parallel. The dimensions of the electrodes are the following: for the Al cathodes 1080 x 620 x 6 mm, and for the Pb-l%Ag anodes 1000 x 580 x 6 mm. Neutral electrolyte is fed to each cell, and the overflow from the upper cell flows through the lower cells of the cascades. From the lowest cell, it finally flows to the spent electrolyte tank

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 285

through the collecting launder. The temperature of electrolysis is about 310 K. The zinc deposit is removed manually in 24-hour cycles. The current density is variable, in the range from 300 to 500 A/m2, because of power cost optimization. The current efficiency is 91 %.

Zinc plates are transported to the induction melting furnaces by railway cars. Zinc plates in the form of stacks are fed to the furnaces with a roller-conveyer. The yield of zinc during melting is 97.6 %. There are produced 25 kg ingots and 1000 and 2000 kg blocks on casting machines. The temperature of the zinc during casting is 793-813 K. Zinc ingots are stacked in 9-layer stacks weighing 1,000 kg, strapped with steel tape, and transported to storage by fork lifts.

Zinc alloys are also produced at the zinc foundry. The foundry is equipped with three 20 tonne induction furnaces, one 30 tonne furnace, one 12 tonne furnace (for alloys), and three casting machines (in-line system) for 25 kg ingots (2) and blocks (1).

Metal is separated from the drosses on a drum screen and is recycled to the furnace. Ashes are utilized in the production of Zn compounds. The BOLESLAW Electrolytic Zinc Plant produces zinc with the analysis shown in Table V.

Table V - Chemical Analysis of the Electrolytic Zinc %

Grade

Zl Z2

Zn

99.995 99.99

Pb max.

0.003 0.005

Cd max.

0.003 0.005

Femax.

0.002 0.003

Snmax.

0.001 0.001

Cu max.

0.001 0.002

Total impurities 0.005 0.01

Chapter 4

Modern Lead Smelting Technologies II

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 289

HEALTH AND HYGIENE IN THE MODERN LEAD AND ZINC INDUSTRY

D.N. Wilson Lead Development Association International

42 Weymouth Street London WIN 3LQ, United Kingdom

ABSTRACT

The potential risks of working with lead have been recognised for centuries, but it is only with the development of sensitive measuring techniques in relatively recent years that it has been possible to establish effective approaches for the monitoring and control of lead exposure. This paper examines the principal control measures employed today - lead in air and lead in blood - and reviews the evolution of numerical standards, with particular reference to the latest understanding of health effects. In addition to lead, a large number of other metals are commonly encountered when producing lead and zinc - both primary and secondary - and the health effects of many of these are discussed. Modern hygiene practices and technologies, which have proved successful in reducing worker exposures, are reviewed. The paper concludes with a look at how exposure standards are likely to change in the near to medium future.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

290 LEAD-ZINC 2000

INTRODUCTION

Health and hygiene are important considerations for all industrial operations, but particularly for sectors such as lead and zinc production which involve the handling of potentially hazardous materials. Here, in addition to the more general physical risks associated with noise, lifting of heavy weights, handling of hot materials, etc., workers may be faced with risks resulting from exposure to chemical agents which must be routinely encountered, often in very large quantities. Within the lead and zinc industries, it is lead that is generally recognised as the chemical agent most likely to result in adverse health effects amongst workers, and around which an extensive body of legislation has been developed. However, it should not be forgotten that almost all chemicals have some potential to cause harm if absorbed in sufficient quantities and that attention needs to be paid to the possibility of excessive exposure to zinc and to a number of the other metals which occur naturally with lead and zinc, such as arsenic, cadmium, silver and thallium.

This paper therefore reviews briefly the adverse health effects associated with a number of metals encountered in the lead and zinc industries. It then reviews the hygiene precautions which have been evolved to protect the workforce - principally from the effects of lead exposures - but are generally effective against other metals as well since the exposure routes are similar. Finally it examines the evolution of standards for the protection of workers and the changes in these standards which might be encountered in the short to medium term.

HEALTH EFFECTS OF METALS

In addition to lead and zinc, the ores from which the two metals are extracted contain small quantities of a wide range of other metals. Some of these, for example silver and cadmium, are important co-products in their own right, whilst others have no particular value to the producer and may indeed be regarded as a nuisance, having to be reduced to very low levels to meet quality specifications. Most metals remain present at trace or alloying levels in the lead and zinc produced, so that they will also be encountered again during recycling. In consequence it is not only the potential health effects of the lead and zinc which must be considered in the occupational environment, but also the potential effects of all the other metals which may be present.

Of the range of metals encountered in lead and zinc plants, some are widely recognised to possess potentially harmful effects if absorbed into the body in too large a quantity. Lead and cadmium are probably the best known of these but others can be equally, or indeed more harmful. On the other hand, certain metals are generally regarded as benign; zinc and copper are especially so since both are essential for life. However, even essential elements can have harmful effects when absorbed in excessive quantities. The potential health effects of lead and zinc, and of all the metals most commonly found in conjunction with them, are outlined below.

Antimony

Antimony is absorbed slowly in the body when inhaled or ingested and can concentrate in a number of tissues such as the lung, the thyroid and adrenal glands, the kidneys, the liver and the blood. It inhibits the actions of certain enzymes, which can result in alterations to glucose metabolism and nerve transmission.

Edited by IE . Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 291

In the occupational setting, chronic inhalation of low concentrations can cause rhinitis and irritation of the trachea. Inhalation of high concentrations can result in acute pulmonary oedema and possibly bronchitis, leading to emphysema. Owing to its tendency to concentrate in lung tissue, pneumoconiosis with obstructive lung disease has been recorded. Some evidence of lung tumours has been found in rats, but there is no firm evidence of carcinogenicity in humans.

Biological limit values are not laid down for antimony. The American Conferences of Governmental Industrial Hygienists (ACGIH) sets a threshold limit value (TLV) for antimony of 0.5 μg/m3, measured as a time-weighted avaerage (TWA). ACCGIH also classifies antimony as a suspected human carcinogen.

Arsenic

In the general population the major source of arsenic is drinking water, which, in some regions of the world, contains naturally high levels of arsenic. Absorption of arsenic through the skin is negligible, and inhalation is a relatively minor pathway, except for workers who are occupationally exposed through processes such as smelting and the production of arsenic-containing pesticides.

Arsenic is readily absorbed from the gastrointestinal tract and lungs and concentrates mainly in the liver, kidneys, lungs and skin. Acute ingestion can lead to a variety of symptoms ranging from nausea and diarrhoea to cardiac arrhythmias and other heart effects which may be fatal. Effects on the peripheral nervous system and symptoms such as anaemia are reversible once ingestion has ceased. Chronic exposure can cause jaundice, and may lead to cirrhosis of the liver. The skin is also a major target with symptoms ranging from dermatitis and melanosis to skin cancer.

The International Agency for Research on Cancer (IARC) and the American Conferences of Governmental Industrial Hygienists (ACGIH) both list arsenic as a confirmed human carcinogen. The element has been associated with lung cancers in copper smelter workers and with a range of cancers reported in populations which drink water containing high arsenic concentrations. Blood levels are not a good indicator of exposure because of the short half-life of arsenic compounds. However, urine levels provide a reliable method of determining recent exposure. Owing to its carcinogenic effects, exposure should be kept to an absolute minimum, and the aim should be to maintain levels of arsenic in urine as close as possible to background concentrations of about 50 μg/l. The ACGIH TLV is 0.01 mg/m3 for arsenic and all inorganic arsenic compounds, except for arsine which is 0.16 mg/m3.

Bismuth

Outside the working environment, the main sources of bismuth are from medicinal preparations and, in some areas, drinking water. Bismuth compounds are generally not well absorbed by the body, but once absorbed, bismuth binds to plasma proteins and concentrates in the kidneys, liver and skin. Bismuth can displace bound lead resulting in the release of lead into the circulatory system.

The mechanisms of bismuth toxicity have not been identified. Symptoms of bismuth poisoning include fever, weakness and rheumatism-like pain. Acute toxic effects, including renal failure, can develop following the ingestion of large concentrations of bismuth. Chronic exposure may cause gingivitis and black spots on the gums.

292 LEAD-ZINC 2000

The results of animal experiments have shown that bismuth interferes with copper and zinc metabolism but do not suggest that bismuth is carcinogenic. Industrial air standards exist only for bismuth telluride and bismuth telluride doped with selenium, for which the Occupational Safety and Health Administration (OSHA) and ACGIH TLVs are 5 mg/m3.

Cadmium

The primary routes of cadmium exposure are ingestion and inhalation with absorption of cadmium through the skin being negligible. Cigarette smoke is an important source of cadmium via inhalation. Absorption of cadmium occurs in the gastrointestinal tract (4-7% in adults) and in the lungs (15-30% in adults).

Cadmium accumulates in the liver, bones and in particular the kidneys, with a half-life of up to 30 years. It exhibits several toxic effects including inhibition of enzymes and DNA repair, which is believed to be associated with the induction of tumours. Acute toxic effects can result from the ingestion of large amounts of cadmium in beverages or food. The effects are apparent immediately and include vomiting and abdominal pain. Inhalation of cadmium fumes also produces local irritant effects and may result in pneumonitis and pulmonary oedema which could be fatal. Chronic exposure from any route can cause adverse effects to a range of organs including the heart, lungs (emphysema), bones (osteomalacia), gonads (suppressed testicular function) and especially the kidneys (reduced glomeruli filtration and tubular damage). Renal damage is unlikely to occur until total cadmium concentrations in the renal cortex reach 200 ppm. In order to control occupational exposure, two biological parameters are commonly employed in tandem - cadmium in blood and cadmium in urine. Blood cadmium levels approximately reflect the level of exposure over the previous 3-6 months, whilst the concentration of cadmium in urine provides a means of protecting against kidney damage. Maximum limits tend to be set in the range 10-20 μg/l (blood) and 11.2-22.4 μg/l (urine). The relatively new technique of neutron activation analysis provides a means of measuring cadmium concentrations in vivo. It is non-invasive and readily transportable, but has not yet evolved into a routine monitoring instrument.

Animal studies have shown cadmium to be teratogenic. It is also carcinogenic to rats through the injection of cadmium metal, sulphide, oxides and salts, and through the inhalation of cadmium chloride fumes. The ACGIH lists cadmium as a suspected human carcinogen. The ACGIH TLV is set at 0.01 mg/m3 for elemental cadmium and inorganic compounds (total dust/particulate) and at 0.002 mg/m3 for the respirable fraction.

Copper

The main exposure route for copper is ingestion of food and drinking water. Inhalation may be of importance in industrial settings.

Approximately 50% of copper is absorbed from the stomach, with lesser amounts from the gastrointestinal tract. Copper binds to metallothioneins and is stored in the liver and bone marrow. Although copper is an essential element for many biological processes, too much absorption can result in interference with a range of cell processes, disrupting cell membranes and reducing normal cell viability. Ingestion of large amounts of copper can result in a range of effects from non-specific symptoms such as nausea and vomiting to more severe symptoms such as jaundice, coma and even death. Apart from metal fume fever, however, adverse effects resulting from exposure to copper in industry are rare.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 293

The ACGIH TLV is 0.2 mg/m3 for copper fumes and 1 mg/m3 for copper dusts and mists.

Lead

In the occupational environment, lead may be encountered as both fume and dust. Inhalation and ingestion are both significant pathways for intake, but skin absorption is negligible for inorganic compounds. Because of its wide range of uses and natural occurrence as an impurity or contaminant, workers may also take in some lead outside the working environment.

Once absorbed in the body, lead is transported by the blood. Most is excreted (mainly via urine) but some is stored in body tissues, notably the bone where it has a half life of many years. Lead can cause a number of adverse health effects. At high levels of exposure, it can cause fatigue, aches and pains in muscles and joints, abdominal pain, colic and constipation. At very high levels, encephalopathy and even death can occur. In the modern occupational settings such effects are rarely, if ever, seen and the medical focus falls on the possibility of anaemia, renal insufficiency or neuropsychological effects.

Reproductive toxicity is a potential concern. Effects on sperm morphology have been noted at blood lead levels of 40 μ§ΛΙ1 but these are of questionable significance. For women the risk of miscarriage at high exposures and effects on foetal development below 30 μg/dl give more cause for concern. There is no epidemiolgical evidence of carcinogencicity in man, but high levels of lead have induced cancers in animal experiments. On this basis, lead has been classified as a possible human carcinogen.

Occupational exposure limits vary from country to country. For lead in air, limits are generally set in the range 0.05 to 0.15 mg/m3 whilst lead in blood limits are normally between 50 and 70 μg/dl for men and between 20 and 40 μg/dl for women.

Selenium

Selenium is an essential trace element involved in the detoxification of free radicals. Soluble selenium compounds are readily absorbed through the lungs and from the gut. They are methylated in the liver and rapidly excreted in the urine. Following very high exposure levels, methylated compounds may also be eliminated from the body via the lungs, giving rise to a characteristic garlic smell on the breath.

The effects of selenium vary with both its physical and chemical form. Selenium dioxide dusts are irritants to the eyes (some workers may develop conjunctivitis), nose and upper respiratory tract. Inhalation of high concentrations may result in pulmonary edema. Repeated or prolonged skin contact with selenium may cause dermatitas. Chronic effects include respiratory irritation, gastrointestinal disturbances, hair loss, mood swings and garlic breath. Hydrogen selenide gas is extremely toxic and exposure can cause a range of acute effects similar to those reported for selenium dioxide dusts.

The results of animal experiments suggest that selenium may both protect against some cancers (thought to be associated with its antioxidant properties) whilst causing others. However, there is no evidence that selenium is carcinogenic to humans. The ACGIH TLV for selenium is 0.2 mg/m3.

294 LEAD-ZINC 2000

Silver

Ingestion and inhalation are primary routes of exposure for silver. Silver is not normally found in food or drinking water and therefore toxic effects are generally associated only with workers who are occupationally exposed.

Approximately 10% of ingested silver is absorbed, and it can also be absorbed from the lungs. Once absorbed it tends to precipitate in tissues and forms complexes with serum proteins which are accumulated in the liver. Chronic exposure in silver workers has resulted in industrial argyria (blue-grey patches on the skin), with the amount of discolouration increasing with increasing exposure time.

Monitoring of blood silver levels is the recommended method to determine the level of exposure, and levels below <25 μg/l are not thought to lead to argyria. The ACGIH TLV is 0.01 mg/m3 for soluble silver compounds and 0.1 mg/m3 for silver metal.

Tellurium

The main exposure pathway for tellurium is ingestion of food (in particular fatty foods and garlic), although inhalation may also be a significant pathway in industry.

Tellurium is poorly absorbed by the gastro-intestinal tract but once absorbed it concentrates primarily in the bones and kidneys, and also in the liver and adipose tissues. The toxicity of tellurium varies with the oxidation state, with tellurite, Te(IV), having greater toxicity than either tellurate, Te(VI), or elemental tellurium. Acute toxic effects of inhalation range from sweating, garlic breath and loss of sleep to kidney and liver damage and, in extreme cases, coma and death.

In animal experiments, tellurium has been found to be teratogenic and to cause testicular damage but no such effects have been observed in humans. Tellurium has not been bioassayed for carcinogenicity. It is thought that the concentration of tellurium in urine reflects recent exposure levels. To avoid garlic breath the concentration of tellurium in urine should remain below <1 μg/l and, in view of the reported teratogenic effects, tellurium exposure should be avoided or minimised by women of childbearing age. The ACGIH TLV for tellurium and its compounds is 0.1 mg/m3.

Thallium

Thallium is a highly toxic metal, being readily absorbed by any route including the skin. Once in the body, approximately 70% of thallium binds to the red blood cells. It behaves in a similar manner to potassium, which it displaces competitively.

Acute effects, following high exposure ingestion, initially result in hypotension and bradycardia. However, this is subsequently followed by hypertension and tachycardia and in extreme cases death. Chronic exposure causes a dying back peripheral neuropathy, gastroenteritis, hair loss, abdominal pain and bloody diarrhoea. Necrosis of the liver and renal tubules may also occur.

There are no recommended means for monitoring exposure to thallium. However, it is suggested that thallium in urine is a better indicator of recent exposure than blood thallium. The ACGIH TLV for soluble thallium compounds is 0.1 mg/m3

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 295

Tin

The most important exposure pathways for tin are inhalation (in industrial settings) and ingestion (although large amounts of tin must be ingested before levels are detectable).

Inorganic tin compounds are not readily absorbed from the gastrointestinal tract. Inhaled tin is transferred from the lungs to the liver and kidneys by red blood cells. There is little data available on the toxic mechanisms of tin. Inhaled tin results in mild pneumoconiosis (known as stannosis) and the absorption of tin results in symptoms ranging from nausea and muscle twitching to paralysis.

The results of animal experiments have shown that tin accumulates in the kidneys, liver and bone, but have not found any evidence of carcinogenicity or teratogenicity. The ACGIH TLV for tin metal is 2 mg/m3 and for tin oxide is 0.1 mg/m3.

Zinc

Zinc is an essential element. The primary routes for zinc absorption are the ingestion of food and drinking water, and inhalation in zinc producing and using industries.

Up to 30% of ingested zinc is absorbed from the small intestine, although absorption is controlled by a homeostatic mechanism. Despite its essentiality, too much zinc interferes with iron and copper metabolism which may result in copper deficiency. Excessive absorption can result in symptoms such as nausea, cramps and vomiting. Inhalation of zinc, zinc oxide and zinc chloride fumes by occupationally exposed workers causes pulmonary oedema and metal fume fever, with symptoms such as fever, chills, sweating and weakness occurring within 4-6 hours of exposure. Chronic inhalation of zinc compounds can result in liver damage.

Zinc is not carcinogenic to humans. The ACGIH TLV for zinc oxide fumes is 5 mg/m3.

PROTECTION OF THE WORKFORCE

Since lead and several of the other metals routinely encountered in the workplace are widely acknowledged to be potentially harmful, there is universal acceptance of the need to employ control measures to limit the exposure of individuals involved in the production and use of lead and zinc. However, there is less agreement over the degree of protection which is required, and so, exposure standards vary from country to country. In lead and zinc plants, the greatest concern is invariably related to lead and an extensive body of legislation has been developed to ensure adequate protection of workers. Generally speaking, the techniques employed to limit lead exposure are effective also for other metals and so the discussion in this section of the paper focuses on lead.

In most countries, the primary control employed is a limit on the amount of lead in the air breathed by the worker. This in turn influences the amount of lead to which the worker may be exposed elsewhere in the workplace (on surfaces, etc.) and considerable emphasis is usually placed also on general cleanliness and good working practice to limit exposure from such sources. The ultimate purpose in all cases is, of course, to protect the health of the individual worker and so health monitoring of the workforce is also conducted. This can take the form of

296 LEAD-ZINC 2000

direct exposure measures such as the level of lead in blood. A number of other biological parameters can also be monitored to assess the effects of lead on the body.

Biological Monitoring

As discussed above, lead exposure can result in a number of adverse health effects of varying severity and significance. Thus, whilst the initial focus must always be on limiting exposure, the only practical way of confirming the effectiveness of such measures is to monitor the health of individual workers to ensure that relevant biological parameters do not exceed safe levels.

Biological Indicators

Several biological parameters exist which may readily be measured and which vary in a predictable manner. Parameters which fulfil these requirements may be either direct measures of absorption or indirect measures which reflect the effect of lead on some biological function.

Lead is present in the body in two main compartments; namely, bone (-95%) and blood (-5%). Because lead accumulates in bone with a half-life of many years, bone lead represents the most accurate measure of long-term exposure to the metal. Unfortunately, however, its measurement is relatively complex and methods of reliable non-invasive testing are not yet sufficiently established for routine monitoring purposes. The best and most commonly employed direct measure of lead absorption is thus the level of lead in blood. Lead which is absorbed into the body as a result of inhalation or ingestion is transported by the blood, where it has a relatively short half-life of about 30-50 days, until such time as it is either excreted or deposited more permanently in bone or body organs. Levels of lead in blood, therefore, give a fairly accurate reflection of recent exposure to lead and, since blood is readily sampled, provide a convenient means of monitoring exposure. Blood lead levels are consequently the most widely used measure of exposure, with the great majority of countries setting exposure standards on the basis of this parameter.

Indirect measures of lead absorption include the rate at which lead is excreted from the body and a range of measurements associated with the haematological system. Lead excretion occurs mainly in urine; consequently, levels of lead in urine (PbU) provide a more simple monitoring method. However, various factors can affect PbU so it is not a sufficiently predictable measure for use as a primary control. The small amounts of lead which are excreted in sweat, hair and nails do not provide a practical means of monitoring.

The other indirect measures which can be used to monitor lead exposure are changes in the levels of a range of enzymes and metabolites involved in the synthesis and operation of haem. Thus, increases in the levels of free erythrocyte protoporphyrin (FEP) or of zinc protoporphyrin (ZPP) in blood can be associated with increased levels of lead in blood, as can decreased levels of activity of the enzyme delta-aminolaevulinic acid dehydratase (ALAD). Similarly increases in the levels of urinary coproporphyrin (CP) and urinary aminolaevulinic acid (ALAU) also reflect increased lead exposure. These measures are not, however, always reliable since they can be affected by other factors, for example ZPP may be increased by iron deficiency. Measurements of these parameters tend, therefore, to be used only in conjunction with, and to provide supplementary data to, blood lead measurements.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 297

The Use of Blood Lead Levels for Monitoring

Because of its convenience and ease of measurement, and because it provides the most accurate and predictable measure of lead exposure, blood lead has been adopted almost universally as the primary biological parameter for monitoring the health of lead-exposed workers. It is thus easy to overlook the fact that it has only become available as such a convenient tool within the last 50 years. Before about 1950 analytical techniques were insufficiently precise to provide the necessary degree of accuracy and reproducibility, and as a result, the historical record of occupational blood lead levels is relatively short.

Prior to the 1950s the main technique used in assessing the amount of lead in blood was identification of basophilic stippling. This required a blood slide to be stained and then painstakingly examined by visual microscopy to identify the number of red blood cells which contained blue stained speckles or dots (stipples). Stippling could be detected at a level of exposure which equated to a blood lead level of the order of 120 μ§Λ11 and this was therefore effectively employed as the acceptable level of lead exposure. With the advent of direct blood lead measurement, standards of the order of 100 μg/dl tended to be adopted (voluntarily by industry rather than by legislation) in order to protect workers from the onset of some of the more readily recognisable clinical symptoms of lead exposure. As time progressed there was a gradual lowering of what was regarded as an acceptable blood lead and levels of the order of 70-80 μg/dl begun to appear in legislation, these being seen as levels below which no clinically adverse effects on health could be detected. Many countries now employ lower levels still, 60 and 50 μg/dl becoming not uncommon and designed to avoid most of the reversible health effects and effects of uncertain clinical significance, as well as subclinical effects of lead, such as changes in enzyme function and resulting alterations to metabolite levels.

The permissible blood lead levels laid down in national legislation for a wide range of countries are set out in Table I, in which it can be seen that the most common levels employed today are in the range of 50 to 70 μg/dl. Inevitably this situation is not static and all the indications are that the trend will be towards lower levels with the object of protecting against ever more subtle changes in biological systems and against as yet unsubstantiated health effects (blood pressure, fertility, etc.) which have been suggested purely on the basis of epidemiological studies. Such studies may indicate statistical significance across large populations but are not necessarily clinically significant for the individual. It is also the practice in some of the legislation to require not only that workers be removed from lead work if the specified limit has been exceeded, but also that their blood lead has fallen below a lower level before they may again commence work which exposes them to lead. A typical example of this is the US legislation which mandates removal from lead work at a blood lead over 50 μg/dl and allows return only once it has fallen below 40 μg/dl, and which industry is now extending voluntarily to suspension at 40 μg/dl and return below 35 μ&ά\.

298 LEAD-ZINC 2000

Table I - Blood Lead Limits for Occupational Exposure (Men)

Maximum Lead Level (Mg/100ml)

80

70

60

50

Country

India Namibia South Africa Belgium Denmark EEC France Germany Greece Ireland Italy Luxembourg Netherlands Spain Thailand Israel Japan Morocco Peru United Kingdom Australia Canada Finland Norway Sweden USA

Whilst on the subject of blood leads, it should be noted that some countries impose different limits for men and women. This occurs because lead can be particularly harmful to the developing foetus, especially during the first three months of pregnancy when many women do not even know they are pregnant. Thus to avoid potential problems, lower limits are sometimes set for the permissible blood lead levels for women (sometimes only for women of childbearing capacity), usually around 30 μg/dl (see Table II). This is an issue which raises much debate, particularly in countries where a strong lobby exists for equal treatment of men and women and pressures do exist in some places for blood lead limits for all workers to be reduced to the level applying for women. However, to demonstrate just how divided opinions are on this issue, it should also be noted that some countries adopt a totally different view and actually prohibit the employment of women in jobs with particularly high lead exposure potential. An example is the United Kingdom where women may not be employed in jobs such as smelting, battery pasting or oxide manufacture.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 299

Table II - Blood Lead Limits for Occupational Exposure (Women)

Maximum Lead Level μg/100ml

40

30

20

Country

South Africa Germany Israel Sweden United Kingdom Australia

Medical Supervision/Screening

The detailed requirements of medical supervision vary from country to country but broadly revolve around monitoring of blood lead and other biological criteria to ensure that values of individual workers remain within acceptable limits and if possible to monitor trends and take corrective action before statutory limits are reached. The frequency of monitoring usually increases with increasing blood lead levels.

Medical supervision should be conducted by a competent medical practitioner who, in addition to performing regular blood lead measurements, would also normally conduct a periodic clinical examination of each worker to ensure continued fitness to work with lead. He would therefore look for any type of condition which might be exacerbated by continued exposure to the metal; examples include anaemia, kidney disorders and nervous disorders. It is usually the medical practitioner who has the ultimate responsibility for determining whether or not continued lead exposure is acceptable, although in some countries the rules require that the company management be informed of blood lead levels and take the actual decision on suspension at the recommendation of the doctor.

The preferred method of taking blood samples is by venepuncture, i.e., by taking a sample directly from the vein, rather than by capillary sampling such as pricking a finger. This is to minimise the possibility of contamination by particles of lead on the skin. The skin should be cleaned thoroughly before sampling by washing with soap and water and if necessary swabbing with surgical spirit. The blood should be stored in a lead-free sample tube with a lead-free stopper (lead-stabilised plastics must be avoided) and should be analysed by a laboratory which preferably participates in an inter-laboratory testing circuit to ensure that its analytical procedures are consistently accurate and reliable. A variety of analytical procedures are available including spectrophotometry, atomic absorption, anodic stripping voltammetry and X-ray fluorescence.

In the specific case of pre-employment medical screening, a similar set of considerations apply. The doctor would look for any history of diseases which might be aggravated by exposure to lead or which, conversely, might later be attributed to lead. Thus, workers with a history of anaemia, kidney problems or nervous disorders might well be considered unsuitable for lead work. Similarly, habitual nail biters might well be rejected because of their increased likelihood of ingestion and smokers might be deemed unsuitable because of their impaired ability to remove dust from the lungs. The wearing of beards or moustaches is also undesirable because of their potential retention of dust and interference with close fitting respiratory apparatus. Indeed it is a condition of employment in some companies that workers be clean-shaven. Finally, a blood lead should always be determined before

300 LEAD-ZINC 2000

employment both to provide a base line for future monitoring and to identify individuals whom it might be preferable not to employ.

Control of Exposure

The principal routes of absorption of lead are through inhalation of lead fume or dust in air and ingestion of dust which has settled on the face or hands, and so control measures to protect the workforce are usually designed with these factors in mind. Thus, the levels of airborne lead in the workplace are usually subject to legal limits and plants are designed to ensure minimum exposure to this source. Much attention is also paid to the avoidance of incidental exposure to lead dust on surfaces, to the provision of protective clothing and respiratory equipment to workers, and to good work practice and housekeeping to ensure that opportunities for ingestion/inhalation are kept to a minimum.

Controls on Lead in Air

Although there is not a good correlation between levels of lead in air and the level of lead in the blood of individual workers, there is a general acceptance of the desirability of placing numerical limits on the permissible levels of lead in air in the workplace. Even if the contribution of air lead to blood lead is not strictly quantifiable, the more dust there is present in the air the more will settle on surfaces and workers, with resultant increases in the likelihood of exposure through re-entrainment to air or through manual contact. Common sense and experience both dictate that a clean workplace is preferable to a dirty one and is more effective at limiting exposures.

Lead in air levels is usually measured using samplers which are placed at strategic points in the workplace (static samplers) or which are worn by individual workers with the sampling head as close as possible to the individual's breathing zone (personal samplers). Static samplers are efficient at highlighting failures or efficiency losses in air extraction systems, whilst personal samplers give a more accurate reflection of the actual exposure of individual workers. In many countries, limits are set on the permissible levels of exposure, usually on the basis of personal samplers.

Over the years maximum limits have gradually been reduced. Today the majority of countries with limits employ 0.15 mg/m3, frequently specified as an 8-hour time-weighted average (see Table ΙΠ). A number of countries have adopted tighter standards, for example 0.1 and 0.05 mg/m3. The rationale behind the setting of air lead limits is less clear cut than might be hoped. In fact, 0.15 mg/m3 was first commonly adopted 20 or 30 years ago when it was supposedly correlated with a blood lead level of about 80 μg/dl. However, as already stated, the existence of good correlations is now very much in doubt, and this undermines the fundamental basis of the standard. Nevertheless, as lower blood lead limits have been introduced, there has been a tendency to impose decreases in air lead limits so that several countries with blood lead limits of 50 or 60 μg/dl employ an air lead standard of 0.1 or 0.05 mg/m3.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 301

Table III - Air Lead Limits for Occupational Exposure

Maximum Lead Level (Mg/m3)

0.2

Country

India Morocco Peru Thailand Argentina Australia Belgium Canada EEC

0.15 France Ireland Italy Mexico Namibia South Africa Spain United Kingdom Austria Denmark Finland

0.10 Germany Israel Japan Netherlands Sweden Czech Republic

0.05 Federal Public of Yugoslavia Norway United States

Lead in air samplers typically comprise a sampling head with an inlet orifice of a few millimetres diameter, a lead-free filter capable of retaining particles down to about 0.2 - 0.3 μπι with very high efficiency, and a lightweight portable pump that can be worn by individuals whilst they carry out their normal work and which is capable of running continuously throughout a working shift at a flow rate of the order of 2 litres/minute. The sampling head is best positioned close to the worker's breathing zone, for example on his lapel or shoulder, and preferably with the filter held in a vertical orientation. The pump is run continuously throughout the desired sampling period at a constant rate so that the total air volume passing through it can be calculated. The sampling head should be covered with a protective cap both before and after the sampling period in order to avoid accidental contamination. For similar reasons it is most important that the filter and sampler should not be made of materials capable of building up an electrostatic charge since it has been demonstrated that these can attract dust particles and give false high readings. Analysis of the samples can be conducted by a variety of techniques including atomic absorption spectrometry, X-ray fluorescence spectrometry or colorimetry.

302 LEAD-ZINC 2000

Protection Against Airborne Lead

Lead dust or fume in the workplace atmosphere should, in so far as possible, be controlled to levels within the legal limits by means of effective exhaust ventilation coupled with suitable arrestment facilities. In situations where exposure cannot be adequately controlled by these means, the use of respiratory protective equipment on the part of the workers becomes necessary. Indeed as blood lead limits become lower, the wearing of respirators has become routine in many plants, not just to protect the worker from airborne lead but because it has the added advantage of preventing hand-mouth contact and consequent ingestionof lead.

Exhaust ventilation is necessary at all points where process fumes and dust may be released into the workplace atmosphere; for example, during mixing or charging material to a furnace, at tapping or pouring points, over kettles, casting machines, etc. At these and many other points, there is either a continuous or a discontinuous opportunity for emissions which can be prevented from creating a problem by installation of complete enclosures or of hooding, with air being drawn in from the workplace at a sufficient face velocity to eliminate dust or fume leakage. Process operations which generate particularly large quantities of dust or fume, such as smelting furnaces, should be exhausted directly to an arrestment plant, with the equipment in question being maintained under negative pressure relative to the workplace, in order to prevent leakage. At fixed workstations where enclosure is impractical, effective protection of workers can be provided by situating exhaust extraction points, preferably on the opposite side of the workstation from the worker and as close to the source of dust/fume as possible, such that the lead-containing air is not drawn through the worker's breathing zone on its way to the extractor.

Various technologies are available for cleaning the dust-laden air extracted from process operations and from the workplace. The type needed in a particular situation depends on various factors such as the size of the particles to be trapped, the moisture content of the gas stream and the nature of the impurities present (organics, halides, etc.). Thus where coarse particles and acid mist are present (as in many of the operations in a battery factory), wet collection may be the preferred technology, with the gas stream being passed through or impacted onto an enclosed water bath. For the dry dusty operations encountered in lead and zinc production, however, bag filters are the preferred, indeed the only effective, technology with the gas being drawn through very fine fabric filters to trap the dust. Other technologies such as electrostatic precipitators (dust particles removed by electrostatic attraction) or cyclones (particulate separation by use of centrifugal forces) are also available but rarely find application in the lead and zinc producing industries.

Despite the very best efforts to control levels of lead in air, there are always likely to be specific situations in which it is impossible to achieve the statutory exposure limits. Examples of such situations might include hand charging or drossing operations, clearing of spillages, maintenance or repair of dusty equipment, or cutting or welding of lead contaminated surfaces. In such situations it is necessary for workers to wear suitable respiratory protective equipment (RPE) to ensure that the air they are actually breathing contains less that the maximum permitted level of lead. A wide selection of RPE is available, ranging from simple disposable mouth and nose filters to high efficiency ventilated visors and masks, and indeed full breathing apparatus. When selecting the type of respirator required in a particular situation, it is necessary to take into account not only the concentration of lead in air against which protection is required, but also whether any other toxic risks, such as sulphur dioxide, stibine, arsine or oxygen deficiency, will also be encountered.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 303

One other important point to consider when dealing with the question of controlling exposure to airborne dust is that reducing emissions is not the total answer. Dust will inevitably build up in any workplace if insufficient attention is paid to regular cleaning. This dust can easily be re-entrained into the air and create unacceptable exposure, apart from additional problems it can present through manual contact and contamination of clothing, etc. It is therefore extremely important that an effective programme of cleaning be adopted and indeed that the plant be designed from the outset to minimise dust traps and to aid ease of cleaning. Design feature which serve this purpose include the use of smooth surfaces (floors, walls, etc,) and the avoidance of ledges and inaccessible corners. Good housekeeping practice in regard to cleaning would include rapid clean-up of spillages and regular vacuuming (preferably wet) of all floors and contaminated surfaces.

Personal Hygiene Measures

In addition to the engineering and housekeeping measures which can be used to reduce exposure to airborne lead, there are a number of areas in which employees themselves can make significant contributions towards their own protection, provided they are given the appropriate equipment and facilities for the purpose.

One of the simplest and most familiar measures in this category is the provision and use of protective clothing. Boiler suits made of cotton or other suitable material (sometimes necessary if molten metal or chemical splash is a possibility) should be provided to each worker, together with suitable footwear (e.g., safety boots, Wellington boots), eye protection (spectacles, goggles, visors), ear protection, safety helmet and gloves as appropriate. Ideally, changing facilities should be provided which are situated between designated clean and dirty areas of the plant. On entering the changing facilities from the "clean" side, workers change into their protective clothing, place their own clothing in lockers and enter the "dirty" lead exposure side. Later on leaving the dirty area, they change out of their contaminated work clothes, shower and put back on their clean personal clothing. The dirty work clothes are stored in their own separate lockers (to avoid cross-contamination of the two sets of clothing and are washed frequently by the employer. In the best equipped plants this laundering is conducted on site, with the water subsequently passing to the effluent treatment facility prior to ultimate discharge. If laundry facilities are not available on site, the clothes must be sent in impermeable bags to outside laundry facilities which are equipped to handle contaminated items.

Another very important aspect of personal hygiene concerns the questions of eating, drinking and smoking. If any of these activities is undertaken in an area of potential lead contamination, then there is a very real risk of ingesting or inhaling lead unnecessarily. Thus, in a well designed plant, canteen facilities will be provided on the clean side of the washing/changing rooms and employees will be required to wash and change into their clean clothes before eating meals. If plant layout does not permit this ideal arrangement, then at the very least, a canteen should be provided as far away as possible from the areas of greatest lead exposure and employees should be required to wash and change before entering the room. Food should never be taken into the workplace. Drinking other than at mealtimes is a slightly different matter. Obviously water or other non-alcoholic beverages must be available, particularly in hot working conditions. The best arrangement in this case is to provide a vending machine with disposable containers or a drinking water fountain in an area conveniently close to the source of refreshment to enable workers to wash their hands and face before drinking.

304 LEAD-ZINC 2000

Smoking presents more of a problem. The smoking of cigarettes at the workplace creates a serious risk of increased exposure because any lead dust which contaminates the cigarette, either from atmospheric fallout or from the worker's hands or clothing, is vaporised during the combustion of the tobacco and is inhaled directly into the lungs in the form of fume - the form in which it is most readily absorbed into the bloodstream. Smoking is therefore highly undesirable where there is the possibility of lead exposure, and indeed, many companies either prohibit smoking (with the possible exception of designated low exposure locations) or, so far as is practicable, follow a policy of employing only non-smokers. Both policies have their drawbacks. Prohibition can be a very unpopular measure with the workforce, although once the risks are properly understood good cooperation often ensues. On the other hand, employing only non-smokers is a form of discrimination which may be unacceptable. Nevertheless, the health advantages to be gained from avoidance of smoking are highly desirable.

Another personal hygiene issue with certain similarities to smoking is the wearing of moustaches or beards. Apart from the fact that these provide a point close to the breathing zone in which particles of lead may lodge and be subsequently dislodged and inhaled, facial hair can interfere badly with the tightness of fit of respiratory protective equipment, which in turn can result in a poorer level of protection. Like smoking, beards and moustaches may be prohibited by some companies.

One final and very important consideration in relation to personal hygiene concerns the employee's duty to look after himself/herself and others. The mere provision of protective equipment will serve no beneficial purpose if the employee does not make proper use of, and help to ensure proper maintenance of, that equipment. A respirator which is not worn provides no protection; rules which are not obeyed are useless. In consequence, legislation in some countries includes not only a duty on the employer to provide a safe working environment and protective equipment, but also a duty on the employee to make full and proper use of all measures provided and to advise his/her employer of any defects he or she discovers. Active participation by the employee in ensuring his or her own welfare is every bit as important as the protective measures afforded by the company.

FUTURE EVOLUTION OF STANDARDS FOR LEAD

As has already been described, differing values for occupational exposure standards and biological limit values apply in different countries. The reasons for these differences lie in the varying interpretations and significance attributed to health data by different authorities, and also in the lengths of time between reviews of standards. Hence as new evidence of health effects emerges, or as more sensitive measurement and analytical techniques evolve, new (almost invariably lower) limits may be adopted in certain jurisdictions which may trigger a slow general movement downwards around the world.

In the case of lead, there are a number of independent health effects which can occur at different levels of exposure, any of which have the potential to become a driving factor in standard setting. Many reviews have been conducted of the health effects literature. Most derive similar conclusions although notable differences do arise, most commonly in regard to the importance of changes which, whilst detectable, may be of questionable significance. One of the most up to date and objective reviews of this nature was published in 1995 by the International Programme on Chemical Safety (IPCS) - an agency of the World Health

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 305

Organisation - in its monograph "Environmental Health Criteria 165 - Inorganic Lead." Some of the key data from this review are summarised in Table IV.

Table IV - Health Effects of Lead

Health Effect Blood Lead Level fcg/dl) Encephalopathy >80 Renal function >60 Anaemia >50 Neurobehavioural performance 40-50 Nerve conduction velocity >30 Reproduction (men) >40 Reproduction (women) 15-30

The most serious effects which occur in the region of 70-80 μ§Λ11 (for example encephalopathy and damage to the central nervous system) have already played their part in standard setting since, with very few exceptions, almost all countries have long employed blood lead limits of 70 μg/dl or less. In addition, the risks of damage to the kidney at about 60 μg/dl and anaemia at about 50 μg/dl have also influenced standards in some countries. The debate now, therefore, is focusing on less overt signs of damage such as impaired neurobehavioural performance (40-50 μg/dl) and possible effects on reproductive ability (40 μg/dl).

The subject of neurobehavioural (or psychological) performance has been examined in many studies. The tests performed on workers address such matters as short term memory, dexterity, reaction times and reasoning ability. Whilst a great deal of variability has been found, and some contradictory findings have been noted, there does appear to be evidence of a small negative effect at blood lead levels of the order of 40-50 μg/dl. The clinical significance of such effects is uncertain, and the effects are generally reversible when bloods leads fall again, but the observations are detectable and are having an influence on standard setting.

The question of reproductive ability is, inevitably, a sensitive issue and a particularly difficult area for meaningful research. For women there is historical evidence for increased rates of spontaneous abortion at high blood leads and some evidence of increases in pre-term delivery and reduced foetal growth and maturation at blood lead levels between 15 and 30 μg/dl. These are major determining factors in setting limits for women of reproductive age. For men the detectable effect is an increase in the number of abnormal sperm above 40 μg/dl, but the significance of this for reproductive ability in unknown. Studies are underway, but it is as yet unclear whether they will shed any light on the ability to father children. Nevertheless, the topic is undoubtedly a potential driver in the standard setting process.

Several other health end-points have the potential to influence exposure standards but current understanding does not enable them to do so. Carcinogenicity is a case in point. Lead has been shown to cause cancers in laboratory animals but, as the IPCS concludes, "the evidence for the carcingenicity of lead and inorganic lead compounds in humans is inadequate". Similarly it has been suggested that lead may be an endocrine disrupter but the IPCS says that "there is no strong evidence of an effect of lead on the immune system". Carcinogenitciy and endocrine disruption are not, therefore, influencing occupational standards at present. One other suggested health effect which should not be forgotten is blood pressure. Statistically, a doubling of blood lead levels has been associated with a 1mm Hg increase in systolic blood pressure. However, there are doubts about whether lead actually causes the effect and, in the

306 LEAD-ZINC 2000

words of the IPCS, "there is no clear evidence that lead has an impact of public health importance as regards hypertension or risk of cardiovascular disease". None of these "effects" is therefore influencing the settings of blood lead limits.

The net result of these various considerations is that neurobehaviour and reproductive ability are the major factors in the health debate at present, with 40 μg/dl looking increasingly likely as the blood lead limit towards which legislation will gravitate. In practice this figure has already been anticipated by industry. In North America, the Lead Industries Association and Battery Council International launched a voluntary commitment in 1997 under which the medical removal level will be reduced to 40 μg/dl (with a return to work below 35 μg/dl) by 2001. In Europe, several major lead producers have pursued internal programmes to reduce the blood lead levels of all their employees to less than 40 μ§Α11 by 2000 or thereabouts. These initiatives are not unique, so it is to be hoped that the likely downward trend in legal limits will not cause the industry too many problems.

The question of lead in air limits is more difficult to predict. This is largely a result of the poor correlation which exists between lead in air and lead in blood. Thus whilst some authorities consider it logical to reduce lead in air limits in parallel with lead in blood limits, others take the view that work practice and good hygiene regimes provide a more effective means of protection and see little point in reducing lead in air limits in the absence of evidence to prove the effectiveness of such measures. The net result is likely to be that a range of values (50-150 μg/m3) will continue to be employed in different countries, with a somewhat slower drift to the lower end of the range than might be expected on the basis of movements in bloods lead limits.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James andA.H.-J. Siegmund 307

COMINCO'S TRAIL OPERATIONS: AN INTEGRATED ZINC-LEAD OPERATION

E.T. de Groot and D.L. Verhelst Cominco Limited

Trail, British Columbia, Canada VIR 4L8

ABSTRACT

With the successful start-up of Cominco's new KIVCET smelter, the final key component is now in place for Trail Operations to operate efficiently as a modern integrated zinc-lead facility. Unique metallurgical, operational, planning and management issues apply to this operation. Along with presenting the current flowsheet for Trail Operations, this paper discusses issues such as the impact of recycles, operational re-stabilization following significant process changes, plant operation with just-in-time inventories, real-time process optimization, and management philosophy.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

308 LEAD-ZINC 2000

INTRODUCTION

Incorporated in 1906, Cominco Limited is an integrated natural resource company whose principal activities are mineral exploration, mining, and metal production. Cominco is the world's largest zinc concentrate producer and is also a major producer of refined zinc metal. Trail Operations is vertically integrated with Cominco's zinc-lead mining activities. Trail's zinc and lead production facilities are also closely integrated, and are in effect a single operation.

Geography of the Trail region has significantly impacted the evolution of the metallurgical operation. Located in a narrow river valley, the operation has a long history of scrubbing smelter gases to produce numerous sulphur products. The Columbia River Valley has enabled the production of low cost electrical power. Limited availability of space to stockpile waste slag and residues has encouraged a policy of minimizing the stockpiling of any material, and of establishing a customer base for all materials that are produced. Ongoing high transportation costs associated with Trail's inland location have encouraged efficient plant operation. To meet the challenges of geography, Cominco and Trail Operations have maintained a strong technical focus.

Product History

Operations began in 1896 with the start-up of a small copper-gold smelter treating ore from the nearby Rossland mines. The smelter's base was broadened in 1899 by adding lead furnaces to serve the growing number of lead mines in the area. Lead refining was added in 1902 when the world's first Betts electrolytic lead refinery was commissioned. Cominco acquired the Sullivan mine in 1910 to ensure a reliable supply of lead concentrates, and this mine has proven to be an excellent source of both lead and zinc concentrates over the decades. By 1916, Cominco developed a method of producing zinc by electrolysis.

Over the years, copper metal production stopped for economic reasons, but numerous other products were added for environmental reasons and to maximize economic return from the concentrates treated. In many cases, technical innovation played a significant role in developing the required processes. The list of products sold today includes: zinc, lead, silver, gold, indium, germanium, bismuth, cadmium, calomel, sulphuric acid, liquid sulphur dioxide, ammonium sulphate fertilizers, copper sulphate, copper arsenate, sodium antimonate and ferrous granules.

PROCESS OVERVIEW

The Trail facility is primarily divided into two production streams - zinc and lead operations. A number of additional products are associated with each of these production streams as indicated in Figure 1.

Four process streams closely link the zinc and lead operations. Iron and lead residues produced in the zinc operation are treated in the lead operation, and comprise nearly half of the total feed to the lead operation. Zinc-rich smelter fume makes up about 15% of the total feed to the zinc operation. Lead smelter sulphur dioxide off-gases are fed to the sulphur gas handling plants in the zinc operation, and waste water streams from both lead and zinc operations are treated in the effluent treatment plant.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 309

Zinc Lead Fuels & Concentrates Concentrates Fluxes

Zinc, Cadmium, Indium, Lead, Silver, Germanium & Sulphur Gold, Bismuth

Products & Copper Products

Figure 1 - Trail Operations Overview

Zinc Operations

Cominco's Zinc Operations produces 290,000 t/a of refined zinc, as well as cadmium, indium, germanium, calomel and a variety of sulphur products. The zinc process is based on standard roasting, leaching and electrolytic technology supplemented by a zinc pressure leaching (ZPL) plant. Integration with the lead circuit allows residues produced by the leaching processes to be treated in the lead process. Zinc oxide fume produced by the lead process is treated in the oxide leaching plant (OLP). About 550,000 t/a of zinc concentrate and 70,000 t/a of zinc oxide fume are treated. The process flow sheet, as shown in Figure 2, has been well described previously (1). Key aspects of the flow sheet are described herein.

About three-quarters of the zinc concentrate treated originates from Cominco's Red Dog and Sullivan mines. Custom concentrates are purchased to enhance the production of various by-products. The roasters treat about three-quarters of all zinc concentrates received, with the remainder treated by ZPL.

Three roasters operate in parallel to produce calcine. There are two 84-m2 fluidized bed roasters and a suspension roaster. Off-gas from the roasters is mixed with off-gas from the KIVCET furnace and is then processed in the sulphur gas handling area. By-products include calomel, sulphuric acid, liquid sulphur dioxide, and ammonium sulphate fertilizers.

310 LEAD-ZINC 2000

Figure 2 - Zinc Operations Flowsheet

The ZPL process consists of an autoclave, in which zinc concentrate reacts with pure oxygen under pressure, and a sulphur separation area. This plant produces a zinc sulphate/jarosite residue slurry which is pumped to the sulphide leaching plant for further processing, and the production of elemental sulphur.

Smelter fume is dehalogenated using a soda ash leach, and is then acid leached in the oxide leaching plant (OLP) to produce an impure zinc electrolyte. Indium and germanium are recovered as by-products, and the lead residue is fed to the lead smelter.

The sulphide leaching plant (SLP) receives feed from all of the front-end zinc plants: roasters, ZPL, and OLP. This plant treats calcine, ZPL slurry, and OLP electrolyte using a weak acid and neutral leaching process to produce impure SLP electrolyte and residue. The residue consists mainly of zinc ferrites, paragoethite, jarosites, lead sulphate, as well as coprecipitated impurities. The residue slurry is fed to the lead smelter.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 311

Impure SLP electrolyte is treated using cold-hot-polish zinc dust purification. The cold stage cement is treated in the cadmium plant to produce cadmium metal and a copper cake that is fed to the smelter. Hot stage cement is leached to minimize the loss of zinc and is then directed to the smelter.

The electrolytic plant plates zinc metal from the purified solution. The cellhouse contains four sections of cells, each with its own rectifier. Individual cells contain 50 cathodes and 51 anodes. Cathodes are mechanically stripped every 72 hours. Each section in the cellhouse contains its own double-masted crane for pulling cathodes, which are fed to any one of four stripping machines.

The casting facility melts the refined cathode zinc and adds alloying elements. Zinc is produced in a variety of shapes for customer use. This facility has been certified as meeting all requirements of the ISO 9002 quality standards.

Lead Operations

Lead Operations treats zinc plant residues, lead concentrates, recycled batteries, and lead bullion to produce silver, gold, bismuth, arsenic metal, copper sulphate, copper arsenate, sodium antimonate and about 100,000 t/a of refined lead. The lead smelter has been well described previously (2). Key aspects of the lead process flowsheet, as shown in Figure 3, are described herein.

The KTVCET flash furnace treats a prepared mixture of zinc plant residues, lead concentrates, lead refinery recycles, fluxes and fuels. Preparation of the feed mixture begins with proportioning of all the feed materials to meet the metallurgical requirements of the lead smelter. Proportioned feed is then dried to less than 1% moisture and pulverized in a ball mill.

The prepared feed mixture is smelted in an oxygen fed flame to produce sulphur dioxide gas, lead bullion and slag. The lead content of the slag is reduced as it passes through a coke checker. Lead and slag flow under a partially submerged partition wall into the electric furnace. Impure lead bullion is periodically tapped from the electric furnace to the bullion treatment area. Slag is tapped to the slag fuming furnace where it is fumed using coal to produce a fume containing zinc and lead oxides, and barren slag. The barren slag is granulated and sold as ferrous granules to the cement industry.

Impure bullion from the KTVCET furnace is decopperized and partially softened in the bullion treatment area in order to produce lead bullion meeting the requirements of the electrolytic lead refinery. Decopperizing is effected using a continuous dressing furnace (CDF), and a continuous sulphur drossing (CSD) process. Resulting copper matte is further processed by the copper products plant. Excess antimony and arsenic are removed by oxygen, producing softened bullion and softener slag.

Lead bullion is electrorefined to produce refined lead and slimes. The slimes are treated in the silver refinery where a series of furnaces process the slimes to produce dore metal and bismuth, as well as various slags and baghouse dusts that contain antimony and arsenic. Some of the baghouse dust is directed to the copper products plant and the lead alloys plant to recover antimony and arsenic values, and the remainder of the baghouse dust and the slags are recycled to the smelter. Dore metal is electrorefined to produce pure silver and gold bullion.

312 LEAD-ZINC 2000

Also included in Lead Operations are a lead alloys plant, a copper products plant, and an effluent treatment plant. The lead alloys plant produces arsenic-lead and antimony-lead alloys from the treatment of softener slag and silver refinery baghouse dust. The copper products plant produces copper sulphate, copper arsenate and sodium antimonate from the treatment of copper matte and refinery baghouse dusts. The effluent treatment plant treats effluents from the zinc and lead operations as well as surface runoff from throughout the metallurgical operation.

Zinc Plant Residues, Lead Concentrates, Zinc Oxide Ferrous

Fluxes & Fuels Fume Granules

Figure 3 - Lead Operations Flowsheet

METALLURGICAL AND BUSINESS IMPLICATIONS

Process Efficiencies

The complex system of recycle streams between and within each of the operating units ensures a high overall recovery of the major products. For example, lead, bismuth and precious metals are recovered from both the lead concentrates and the zinc concentrates treated. Recovery of lead is also significantly enhanced by operation of the slag fuming and oxide leaching circuits. Lead not recovered directly to bullion in the KIVCET furnace is recycled

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 313

through the oxide fume circuit, a practice which ensures that the overall recovery of lead is about 99.8%.

Although the recycle streams ensure a high recovery of lead, control of the lead concentration in KTVCET slag remains an important process consideration. Unless the lead concentration in the slag is well controlled, recycle streams will contribute a significant tonnage to the process streams, thereby utilizing capacity that could otherwise be used for the treatment of other feedstocks.

Interestingly, one of the impacts of the recycle streams is process simplification in some of the operating units. Nowhere is this more evident than in the leaching area of the zinc plant, which uses a very simple weak acid leach process circuit for recovering zinc from calcine. Zinc recovery at this stage is typically 85-90% because of the presence of significant quantities of zinc ferrites. However, the overall recovery of zinc is about 98-99% because of zinc recovery by the slag fuming and oxide fume operating units.

It is also important to note that the smelter recovers zinc from zinc plant residues and from zinc contained in lead concentrates. Very few lead smelting operations today recover zinc from lead concentrates.

Impurity Control

Balancing some of the benefits of the recycle streams is the need to ensure that process metallurgy is well understood and well controlled with respect to impurities. Significant impurities include iron, arsenic, antimony, tin, bismuth, cadmium, germanium, indium, silver, gold, chlorine, fluorine, mercury and thallium. Impurities contained in the mixture of zinc concentrates treated in the roasters or in zinc pressure leaching will have an impact on all of the other metallurgical plants. Similarly, impurities contained in the lead concentrates treated will impact many of the metallurgical plants. Thus, the production engineers who plan the feed mixture to any one of these plants must consider the impact on all downstream operations, recognizing the capacity of each operating unit for processing impurities.

Control of impurities in the process is a dynamic issue. This fact can be illustrated by examining the changing process deportments for arsenic during the initial operation of the KIVCET furnace. Process control difficulties during the start-up phase resulted in a significant deportment of arsenic to slag. Arsenic in slag is readily fumed and thus would report to the oxide fume, and return to the lead smelter by way of the zinc plant residue. This resulted in an extremely elevated recycle of arsenic between lead and zinc operations. Compounding this recycle was the effect of a number of internal recycle streams within the oxide leaching plant. To control this impact required strict limits on allowable new arsenic inputs to the process. Initially the situation was eased by process changes within the oxide leaching circuit that decreased the amount of arsenic recycling within this plant. Control of the process metallurgy at the KIVCET furnace has improved considerably so that today the arsenic limit is once again based on limits within the bullion processing circuit, rather than being set by the oxide leaching circuit.

With increased focus on the recovery of metal values in stockpiled slags and residues, impurity control becomes an even more important consideration. Operation of the KTVCET furnace at design capacity and the restart of one of the older slag fuming furnaces have enabled treatment of these stockpiles. As the stockpiles were being accumulated they formed natural bleeds for undesirable impurities such as the halides, and now the challenge will be to control

314 LEAD-ZINC 2000

the overall metallurgy as these materials are returned to the process. Inevitably, process changes will be required to allow for increased concentrations of certain elements.

Business Opportunities

As seen from the process flowsheets there are many possible entry points for raw materials. This provides flexibility, as well as complexity, to the acquisition of raw materials. The list of input materials contains not only base metal concentrates but also semi-finished products such as unrefined lead bullion, and process by-products such as dust or sludge from other companies. Recycling of consumer products is also an important factor in feed material selection because Cominco has a strong commitment to environmental stewardship, as reflected by the quantity of lead battery products processed annually.

A full understanding of the metallurgical behavior of the major and minor elements allows for successful changes in the technologies used in metal production. This understanding also allows changes to the flow streams in continuing efforts to economically improve recoveries, or enable bleed streams of these elements. It is important to know how process changes will affect metallurgical behavior since the economic mix of raw materials will need to be changed as a result. Conversely, if one can recognize the potential options available for changing the process, then it will be possible to dynamically evaluate different mixtures of raw materials that could be treated in the future.

RAW MATERIALS AND IN-PROCESS INVENTORIES

Inventories

Upsets in one part of a large integrated plant will have repercussions in the operation of upstream and downstream units because the affected plant is either the recipient or a supplier of an intermediate product. Having large in-process inventory capacities between the operating units can mitigate such impacts. In the zinc circuits, there is very little in-process capacity with the typical retention time of approximately 4 days from concentrate treated in the roasters or ZPL to zinc in solution to the cellhouse. Similarly, several impurities contained in the zinc concentrate will be part of the KIVCET smelter feed about 4 days later. Thus the mixture of zinc materials treated can quickly have an adverse impact on many of the significant operations at Trail unless the feed mixture is well controlled. This control is maintained with typically only a one-week supply of zinc concentrate supply in the Trail area. A close watch of the analysis of incoming materials is essential to maintain metallurgical control.

In the lead circuits, raw materials inventory for the smelter feed preparation plant is 2 weeks, excluding the large on-site stockpile of residues. The processing time between the feed preparation and smelting circuits is 1 day. Thus feed proportioning is a critical activity. Though there is capacity to store up to one-week's supply of unrefined lead bullion between the smelting and electrorefining circuits, typical inventories are less than 3 days.

There is only one day of capacity between the zinc fume recovery process in the smelter and the fume leach process in the zinc plant. Therefore, if there is an upset in the leach circuit, it is necessary to stop the slag fuming process for zinc recovery and stockpile smelter slag for recovery in the future. The same is true for the flow of residue from the zinc plant circuits to the smelter. Residue slurry is transported from the zinc plant to the smelter, where it is

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 315

processed in pressure filters to produce a manageable residue for input into the smelter feed blend. An outage of the smelter feed preparation process will result in the need to stockpile the filtered residue. However, an outage of the filter presses will result in the need to store slurry within the zinc plant thickeners, a practice which is limited to approximately 2 days, depending on current operating inventories.

Operating Implications

From a business point of view, it is preferable to minimize inventories so as to improve the timing of metals recovery. Control of a complex metallurgical process containing numerous recycles requires strict adherence to established process standards, and a robust process. Many process improvements over the years were driven by the need to control the process and improve metal recoveries. Recovery of the zinc plant residue inventory has also become a focus for the operation as a whole.

To ensure maximum on-line time for all plants in an operation with limited inventory between the plants requires close co-ordination and co-operation of all units. It must also be recognized that the sheer size and complexity of the entire operation does not easily allow for a complete shutdown of all units at one time. Thus creative approaches must be taken to plan maintenance shutdowns.

PROCESS OPTIMIZATION

Issues in Process Optimization

A key issue in process optimization at any site is to ensure that the entire operation and all stakeholders are considered. Optimization of individual process steps can lead to suboptimal performance if the impact on the rest of the operation is not considered. The Trail metallurgical operation, with its many recycle streams impacting numerous process areas, is particularly susceptible to the problem of suboptimization if individual process areas are changed without considering the impact on all affected process and product streams, including the environment.

Optimization starts with determining the zinc concentrate mix to be treated. This determination must be made considering all recoverable metal values and also the downstream metallurgical and operational constraints in the rest of the zinc and lead operations. Zinc concentrates with greater precious metal values typically also contain more iron and other impurities, resulting in a lower zinc metal production and an increased production of iron residues, which reduces available smelter capacity for other feed materials. Thus the optimum zinc concentrate feed mixture is based on all final products produced considering all downstream capacity implications.

Optimization of the smelter feed mix must account for the recoverable metal values in each of the feeds available within the constraints of the KTVCET furnace, bullion treatment and slag fuming. Similarly, optimization of the slag fuming operation must be done within the context of maximizing recoverable metal values and ensuring that customer quality requirements for barren slag are met without restricting the operating rate of the KIVCET furnace.

316 LEAD-ZINC 2000

Metallurgical Model

Process optimization must also include the option to modify the existing operation by changing process parameters or making engineered process modifications. Identification of all significant impacts of such changes is often not obvious to even very experienced technical personnel. To guide decision making, Cominco has developed OPTIMET, a metallurgical and financial model of the Trail operation. This is an empirical model that is based on historical experience. The model uses linear programming to achieve an optimal solution given a set of constraints. Steady state material balances on 26 elements are produced for individual plants, plant areas, and the entire operation. Operating profitability for the operation is also calculated. The model is used to optimize feed mixtures, optimize process targets and to evaluate projects that impact on the Trail metallurgy. This model is calibrated periodically to ensure accuracy.

Real Time Optimization

There is also a real time aspect to process optimization. In real time OPTIMET may be useful as a metallurgical guide, but the process knowledge and experience of the technical and management staff become much more important. As discussed earlier, arsenic impurity levels were strictly controlled during the early phase of KIVCET operation. Recognizing that a problem was being encountered in the oxide leaching circuit, a shutdown of the slag fuming furnace would provide a window of opportunity to treat materials elevated in arsenic. Capital investment decisions were being made on the installation of a circuit to bleed arsenic from the oxide leaching circuit while process changes were being made within the oxide leaching plant. Decisions were also being made regarding the level of arsenic impurity in concentrates that could be purchased. Process optimization in this type of dynamic environment is achieved by a high level of process knowledge and by clear communication between key management and technical staff.

One other aspect of process optimization involves consistent operation to established process standards. The methodologies of Integrated Process Management as proposed by Brian Slater (3) are being used to address this issue. This methodology is entirely consistent with the ISO 9002 standards.

MANAGEMENT PHILOSOPHY

Management Training

Effective management of an industrial operation requires detailed process knowledge of each individual plant to understand its potential, as well as having a clear understanding of the entire operation so that cross impacts are well understood. In addition to process knowledge, it is important to understand the marketing issues for each product, environmental issues, human dynamics, as well as numerous other business factors.

Cominco has developed various programs to develop its technical staff to meet the required standards. Recently graduated metallurgical, chemical and mechanical engineers begin their careers as engineers-in-training during which time they are assigned to work three six-month work terms in different process areas. This rotation is geared to ensure that the engineer-in-training gains an understanding of several process areas, an appreciation of the

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 317

interconnected nature of the entire operation, and a set of important contacts throughout the operation.

Production engineers also move between different plants every two to five years, depending on experience level, career development needs, and operational requirements. During this stage of one's career, short-term assignments in other process areas may be provided to broaden an engineer's experience. As an alternative to a plant assignment an engineer may also be assigned to work with the Trail OPTBViET metallurgical model or in the central technical group. The central technical group has a mandate to work on critical projects that typically impact two or more different process areas. Assignments to the environmental group, the finance department, or the sales division may also be given.

Production engineers typically progress to line management or senior technical roles. Exposure to other areas of the company continues throughout one's career. This exposure ensures that decision making is not done in isolation, but rather recognizes the integrated nature of Trail Operations and of Cominco as a company. Critical decisions are also typically made by developing a consensus rather than by one senior manager. Communication and interpersonal skills thus become even more important for management personnel.

Business Training

Trail Operations also works together with Simon Fräser University to provide business administration training to its professional employees. This training has included graduate level courses in finance, marketing, organization effectiveness, contract law, and economics. Additional courses are currently being developed. Courses are offered on a regular basis to ensure that training can be taken at an appropriate time in a professional's career development. These courses are available to employees working in Trail and Kimberley operations, as well as to employees of other industries in the community. Studying and working with people from various backgrounds allows one to appreciate other perspectives of complex issues, and to increase networking opportunities.

Employee Involvement

It must also be recognized that management alone cannot achieve efficient operation. The active involvement of all employees working throughout the operation is required. This is especially important in an environment where the success of an individual plant can be based on adherence to standards in several other plants in the operation. Employees throughout the operation are eligible to receive training that develops basic business, technical, and interpersonal skills. It has also become apparent that lifelong learning is an important contributor to employee performance and satisfaction, and thus, Cominco has established a learning center in Trail that provides non-work-related training for all employees.

It is also important to recognize that there must be a focus on productivity in an operation that produces numerous different products. To facilitate productivity gains, Trail Operations has developed a gainsharing program. A gainsharing plan was introduced using a simple formula that was focused on residue treatment, zinc production, lead production, and cost control. This plan allowed employees to focus on the key success factors, which were easily understood. The gainsharing formula is currently being revised to reflect how individual efforts impact overall performance, as well as including quality and safety factors. As time progresses it is certain that the gainsharing program will continue to evolve as the relative value of each of the products changes.

318 LEAD-ZINC 2000

THE FUTURE

Trail Operations has successfully produced metals for more than one hundred years. Geography of the region has provided significant challenges that have been overcome by ensuring a strong technical focus, and by the involvement and development of employees throughout the operation.

The Trail operation has started to recover stockpiles of intermediate materials that have accumulated over many years. As this program continues it is certain that additional metallurgical challenges will arise and will be overcome. Inevitably, this will result in producing even more products, and perhaps in the future, the facility may become a destination for intermediate materials that cannot be treated by other operations. In this regard the Trail operation continues to develop a unique model of operational efficiency.

ACKNOWLEDGEMENTS

The authors would like to thank Cominco Limited, for permission to publish this paper.

REFERENCES

1. MJ. Brown, E.T. de Groot, M.G. Heximer, A.J. Karges, G.N. Masuch, and CM. Okumura, "Zinc Capacity Increase at Cominco's Trail Operations", Zinc and Lead Processing. J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton, and, P. Hancock, Eds., The Metallurgical Society of CIM, Montreal, Canada, 1998, 41-54.

2. A.R. Babcock, R.A. Franco, A.C. Mikrovas, S. Bharmal, and M.I. Cecchini, "Preparation and Start-up of Cominco's New Lead Smelter", Zinc and Lead Processing. J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton, and P. Hancock, Eds., The Metallurgical Society of CIM, Montreal, Canada, 1998, 795-809.

3. B. Slater, Integrated Process Management: A Quality Model. McGraw-Hill Inc., New York, NY, U.S.A., 1991.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 319

THE LEAD BATH SMELTING PROCESS IN NORDENHAM, GERMANY

M. Sibony, N. Basin and J. Lecadet Metaleurop Recherche

1, avenue Albert Einstein - BP 120 78193 Trappes Cedex, France

R. Menge and S. Schmidt Metaleurop Weser Blei

Johannastrasse, 2 26954 Nordenham, Germany

ABSTRACT

In 1996, Metaleurop replaced its conventional sintering and blast furnace lead smelting activity in Nordenham, Germany by Sirosmelt/Ausmelt technology. Following some alterations to the process and the equipment, the bath smelting plant reached its design capacity for the treatment of lead concentrates, lead sulphate paste from spent batteries, secondaries and internal by-products. A reliable metallurgical process has been established, and thus higher process stability has been achieved, increasing simultaneously the refractory life. Nevertheless, improvements are needed to control the process with a better automation system and reliable expert-systems, because of the intensity of the smelting process. The introduction of the Sirosmelt process for the pyrometallurgical production of lead instead of the prior conventional process has also demonstrated the following advantages: reduced emissions of heavy metals and gaseous compounds, reduction of the specific energy consumption, minimization of residues and a reduction of staff and maintenance costs. The environmental performance of the plant meets the requirements of the German regulations.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

320 LEAD-ZINC 2000

INTRODUCTION

Traditional smelting methods using sinter machines-blast furnaces are now being replaced by intensive smelting processes that are generally superior from the viewpoints of capital and operating costs, energy usage, and especially, environmental compliance.

Stricter regulations concerning lead emissions and ambient lead in air levels have forced the development of other lead smelting processes to replace the combination of the sinter plant and the blast furnace. The Metaleurop Weser Blei Company closed the former sinter plant-blast furnace plant in 1995, selecting the bath smelting route under Ausmelt license (1) to process a range of high grade feed materials containing 65-80 % Pb. This submerged lancing technique was developed at CSIRO (Commonwealth Scientific and Industrial Research Organisation) during the 1970s, and has been applied to a wide range of applications and uses, including slag treatment, smelting and converting operations (2). The technology is based on the top-entry, submerged lance system, providing natural gas, process air and oxygen below the surface of liquid slag bath. This process has the following advantages:

• Considerable environmental improvements • High rates of reaction and excellent heat and mass transfer efficiency • Flexibility in feeds • Simple control and operations.

The furnace, located in Nordenham, Germany, is now operating at design capacity, using a wide range of lead concentrates and secondary materials.

PROCESS DESCRIPTION

Process Chemistry

The full treatment cycle involves two stages; namely, the smelting step (oxide and sulphide smelting) followed by the slag reduction step with carbonaceous materials.

The overall reactions in the first stage are given by :

PbS+1.5 02->PbO + S02 (1)

ZnS + 1.5 02 -+ ZnO + S02 (2)

FeS+1.5 02->-FeO + S02 (3)

PbS + 2 0 2 - > PbS04 (4)

PbS + 02 -► Pb + S02 . (5)

PbS + 2 PbO ->· 3 Pb + S02 (6)

PbS04 -> PbO + SO2 + 0.5 02 (7)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 321

The second stage involves the reduction of the slag. During the reduction, the following reactions occur:

Fe203 + C -► 2 FeO + CO (8)

PbO + C -► Pb + CO (9)

C02 + C ->■ 2 CO (10)

The slag also contains some zinc. When the lead in the slag is reduced below 6 %, the zinc oxide reduction process becomes significant and zinc metal volatilization takes place. Zinc reports, after oxidation, in the post combustion area above the slag bath as an oxide fume.

An on-line process-control system taking into account the feed materials composition, the thermal balance of the system and the targeted operating conditions has been successfully implemented. The metallurgical model has been developed to assist in establishing the optimum process conditions and to permit process stability.

Oxidation Smelting

Crucible-scale and pilot plant testwork was carried out in Trappes, France, by Metaleurop Recherche (Metaleurop Research Centre), in order to improve the overall process understanding of the technology. The results have allowed further industrial plant process optimisations.

As indicated, the extreme versatility of the process offers a large choice of process conditions for different feed sources. A precise control of the qualities and amounts of the feeds is needed, in order to achieve precise process control. The DTS measurements have indicated a relatively short residence time in the furnace, and expert systems are needed to detect instabilities in the process. The oxidation rates are very high, and the material feed rates are limited only by the capacity of the off-gas handling system.

Slag Composition

Intensive research and development have led to slag composition improvements in order to:

• Prevent refractory wear by promoting a slag coating • Decrease the lance wear by a slag coating • Ensure proper slag fluidity, thus allowing good mass transfer in the furnace and

avoiding blockage problems at slag tapping • Avoid slag foaming phenomena.

The combination of experimental investigations and thermodynamic computer modelling was used to examine suitable slag compositions. The computer system MTData and NPL thermochemical database were employed to calculate phase diagrams and crystallization paths in the industrial slag system including PbO, CaO, Si02, FeO, Fe203 and ZnO (3). A phase diagram example is given in Figure 1.

322 LEAD-ZINC 2000

CaO+Si02 Liquidus Contours - Triangular Section

PbO-CaO+Si02-FeO+Fe203

1173(100)1673 K

0.2 1.01325E5Pa

m PbO FeO+Fe203

Figure 1 - Liquidus Lines in the PbO-Si02-CaO-FeO-Fe203 System with CaO/Si02 = 1 wt and FeO/Fe203 = 0.5 wt

A typical slag composition during the smelting step is given in Table I.

Table I - Typical Slag Composition during the Smelting Stage Element

Pb Zn

Si02 CaO

FeO + Fe203

Range (wt%) 40-60 5-15 10-20 5-10 10-30

Slag Reduction

High-lead slag from the smelting of battery paste and concentrates is stockpiled for separate campaigns of slag smelting and zinc fuming. A single furnace is used for these various campaigns.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 323

PLANT DESCRIPTION

A flow diagram of the plant is presented as Figure 2.

A Raw materials storage B Dust filter C Dosing unit D Dust filter E Screw feeder F Pelletiser G Feed distributor H BSF furnace I Fluxflow heat exchanger J Electrofilter K Venturi L Gas cooling M Electrofilter N Settler O Gas dryer

P Intermediate absorber Q Final absorber R Heat exchanger S Contact vessel T Decopperising vessel U Slag granulation V ZnO exit 1 Concentrates 2 Battery paste 3 Coal 4 Refinery products 5 Limestone 6 Sand 7 Hematite 8 Flue dust

9 Natural gas 10 Process air 11 Oxygen 12 Shroud air 13 Slag 14 Lead bullion 15 Steam 16 Condensate 17 Water 18 Sulphuric acid 19 Exhaust gas 20 Filtrate 21 ZnO dust 22 Leached dust

Figure 2 - Flow Diagram of the Plant

Feed Preparation

The different concentrates, battery paste, recycled products such as dusts, refinery intermediates and copper dross are transported using closed-belt conveyers, the so-called pipe conveyers, thus drastically reducing emissions.

A photograph of a closed-belt conveyor is presented in Figure 3.

324 LEAD-ZINC 2000

Figure 3 - Closed-belt Conveyor Device

The Furnace

A photograph of the plant which houses the furnace is presented as Figure 4.

The metallurgical furnace consists of a cylindrical steel shell, 4.2 m in outer diameter and 9.5 m high. In the furnace the specific conditions during the smelting step result in the deposition of a layer of protective slag on the furnace walls. This coating has proved its efficiency and, with an increased process stability, has increased refractory life to more than one year. The reactor shell is cooled with an external water circulation system to form a protective slag layer.

The lance comprises four tubes (natural gas, air, oxygen, post combustion air), and is tightly sealed to the furnace roof as illustrated in Figure 5. This lance is protected by a layer of slag that builds around it. The slag layer is formed and preserved by the cooling action of the process air. Internal swirlers increase the heat and mass transfer rates, making the slag bath extremely turbulent, thus allowing low residence times.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 325

Figure 4 - Bath Smelting Furnace Plant in Nordenham, Germany

326 LEAD-ZINC 2000

Figure 5 - Upper Lance Device and Lance Port

Lead and Slag Launders

Lead and slag are tapped continuously out of the furnace through two siphons; the slag is then granulated and the primary lead is drossed in two vessels.

Gas Treatment

The gas treatment system is based on FluxFlow technology (4). The operating principle of FluxFlow is shown in Figure 6. The system comprises three main components:

• Mixing chamber • Heat exchanger water/steam • Separation of dusts from the cooled gas (cyclone).

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Process Gas Out T»150-400'C

Cyclone

SoBds Return

Make Up

Dust Out

iL Water/Steam ►

Water

Heat Exchanger

Mixing Chamber T^.ZtXWQO-C

Process Gas In T „ - 600-1,600Ό

Dust, Droplets, Vapor 10-500 g/m5

Figure 6 - Schematic Diagram of the Operating Principle of FluxFlow (4)

The gas volume of approximately 30,000 Nm3/h exits the furnace (temperature between 1000°C and 1500°C, S02 content between 0.5 % and 12 %) through a water-cooled transition piece before entering a circulating fluid bed waste heat boiler. The gases enter the mixing chamber of the FluxFlow waste heat boiler at about 70-80 m/s, and are rapidly quenched to the mixing temperature by the circulated cooled dusts. The mixing temperature is controlled to maintain it below the sticking point of the dusts at about 400°C.

From the mixing chamber the dusts are entrained with the gas flow and are transported as a suspension through the cooler. The heat is recovered as high pressure steam. In the primary cyclone separation stage, the coarser dusts are returned to the mixing chamber. The process gases are then cleaned using two electrofilters. To maintain the amount of dust in circulation at the desired level, the system is equipped with a controllable outlet for dust. The precipitated flue dust removed form the FluxFlow system is fed back to the process after treatment for chlorine and cadmium removal.

Finally, after conversion to sulphuric acid, the gases are rejected to the atmosphere with no particulate emissions; final SO2 conversion efficiencies of more than 99.5 % are achieved.

ENVIRONMENTAL PERFORMANCE

The implementation of the bath-smelting technology instead of the sinter machine-blast furnace option has led to strong improvements in terms of environmental performance.

328 LEAD-ZINC 2000

Air emissions

Table II indicates a drastic decrease in the emission of elements such as Pb, Cd, Sb, As, Tl, Hg and SO2 linked with the new process development, and Figure 7 illustrates these emission reductions on a graphical basis.

Table II - Heavy Metals and SO2 Emissions for the Sinter Machine-Blast Furnace and the Bath Smelting Furnace

1990 Blast furnace

1997 Bath Smelting

Furnace Evolution

(in %)

Pb kg/y

24791

1451

-94.1

Cd kg/y

572

4.05

-99.3

Sb kg/y

460

27.52

-94

As kg/y

219

5.58

-97.5

Tl kg/y

38

1.27

-96.7

Hg kg/y

17.2

0.87

-94.9

S02

t/y

7085

140.4

-98.0

Figure 7 - Heavy Metals and SO2 Emissions Before and After the Technology Improvements

As stressed in the initial discussion of this project, CO2 emissions per tonne of produced lead have been reduced from 1.095 t/t of lead to 0.450 t/t of lead (see Table III).

Table III - CO2 Emissions for the Sinter Machine -Blast Furnace Process and the Bath Smelting Furnace

CQ2 t/t of Lead 1990

Sinter Machine-Blast Furnace 1997/1998

Bath Smelting Furnace

1.095

0.450

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 329

Water Consumption

Because of the new cooling system of the furnace, the water consumption decreased by about 3 million m3/y.

Residues

The flexibility of the bath smelting furnace permits the operation to adapt quickly to the prevailing economic situation. A higher rate of treated battery paste, achieved by lowering the amount of concentrates processed, results in a lower tonnage of primary slag produced. The latter is either processed in a second stage operation or is sold.

Energy Consumption

The implementation of the new bath smelting process in Nordenham has led to a reduction of 35 % of the specific energy consumed per tonne of lead.

CONCLUSIONS

The lead plant in Nordenham used the conventional two-stage sinter machine-blast furnace process until the development of bath smelting technology in 1996.

In the 1980s, Metaleurop studied several new processes, with the following objectives:

• Environmental protection • Decreasing energy consumption • Reducing costs.

These studies, associated with pilot-plant trials, indicated that Sirosmelt technology, because of its flexibility, versatility and simplicity, was the most suitable process to replace the conventional sinter machine-blast furnace. The process has been optimized since the start-up of the furnace based on increased industrial experience and pilot plant trials at the Metaleurop Research Centre. Improved automation and expert systems are necessary to stabilize the process because of the short residence times involved. The bath smelting process operates with a large range of primary and secondary feed materials, and carries out different metallurgical operations in the same reactor.The exhaust gases produced are processed in a sulfuric acid plant, and final SO2 conversion efficiencies of more than 99.5 % are achieved. All the emissions have been drastically reduced; CO2 emissions have decreased by 60 %.

With the bath smelting technology, Metaleurop Weser Blei now owns and operates a modern lead smelter, which gives Metaleurop a competitive advance in terms of environmental protection and recycling capacity. Metaleurop Weser Blei now produces at design capacity.

REFERENCES

1. "Submerged Injection of Gas into a Liquid-Pyrometallurgical Bath", US Patent. No. 4,251,171,1981.

LEAD-ZINC 2000

E.N. Mounsey and K.R. RobiUiard, "Sulfide Smelting using Ausmelt Technology", Journal of Metals. August 1994, 58-60.

MTData Handbook. National Physical Laboratory, Teddington, Middlesex TWl 1 OLW, UK.

K. Westerlund, H. Holopainen and K. Westerlund, "Increasing Profit by Improving Gas Handling Systems", Journal of Metals, September 1992, 46-49.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 331

THE QSL LEAD SLAG FUMING PROCESS USING AN AUSMELT FURNACE

Myung Bae Kim and Woll Seung Lee Korea Zinc Co., Ltd.

142 Nonhyun-Dong, Seoul 135-749, Korea

Yong Hack Lee Chonbuk National University

Chonju 561-756, Korea

ABSTRACT

The Slag Fumer is the first Ausmelt furnace system adopted at the Onsan Refinery of Korea Zinc. This plant continuously treats the QSL lead smelting slag in the molten state. It has developed progressively since its first commissioning in October 1992 and has generated a confidence that it can produce an environmentally acceptable slag for use in cement manufacture and recover the zinc and lead in the slag as a fume oxide. This paper provides the operating experiences accumulated during the last eight years of operation, a summary of problems and solutions, and the current status of the operation.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

332 LEAD-ZINC 2000

INTRODUCTION

In the course of increasing the production capacity of zinc at the Onsan Refinery from 50,000 tons per year in 1978, the first year of operation, to 350,000 tons per year in 1999, the treatment of iron-containing residue, which is currently a goethite residue, has been an important issue because of the limited space at the plant. Korea Zinc investigated the available technologies for a process that could produce an inert material and meet the limits of the pollution restrictions. The requirements for the new process installed at Korea Zinc were as follows:

• Pyrometallurgical process with fuming reactions • Energy source should be cheap coal • Operation should be continuous.

Ausmelt technology was concluded to be one of the best processes that can meet the above requirements. The first Ausmelt furnace system adopted by Korea Zinc was the Slag Fumer for treating lead smelting slag produced from the QSL smelter. The main objectives of this plant were to recover valuable metals from the QSL slag and produce a cleaner slag, and to evaluate the performance and applicability of Ausmelt technology to the commercial process for treating zinc plant iron residues including goethite, jarosite and primary leach residue (the second Ausmelt plant).

The plant established a level of confidence that it can produce an inert clean slag for use in cement industries while recovering valuable metals as a fume oxide. This successful result allowed Korea Zinc to build the second Ausmelt plant, the so called the Zinc Fuming Plant or the Zinc Fumer (1).

Most lead concentrates contain a portion of zinc and vice versa, so that both metals are impurities from the viewpoint of the prime metal contained in the concentrates. Therefore, each smelting activity inevitably produces disposal materials containing the other metal. Korea Zinc has been committed to maximize metal recoveries in lead and zinc production and to reject all non-valuable materials in the concentrate feed in a stable form such as slag. In this respect, Korea Zinc has developed a concept integrating lead and zinc production plants as shown in Figure 1.

In this figure, lead- or zinc-rich intermediates are given or taken between the lead and zinc plants. The QSL smelter treats lead concentrates as well as the lead-silver residue produced from the zinc hydrometallurgical process. The QSL smelter also produces a zinc-rich oxide to be sent to the zinc plant. The Ausmelt technology at Onsan is employed to accomplish this integrated circuit. The Fumer in this figure actually means two Ausmelt systems - one is the Slag Fumer and the other is the Zinc Fumer.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 333

Secondary Material Zinc Concentrate

1Γ U V

Zinc Plant

1

\

'

fron nOuti side

Lead Bearing Intermediates

Zinc Bearing Intermediates

4

IT V

Fumer

1

Lea id Cor lcentrate

" \r

Lead Plant

▼

Zn, Cu, Cd, Alloy, HiSOt Clean Slag Pb, Ag, Au, Bi, Alloy, HiSOt

Coal Burning Power Plant

Maintenance Division

Engineering & Construction

R&D Center

Fine Chemical Production

Figure 1 - Smelting Activity at the Onsan Refinery

PLANT DESCRIPTION

The Slag Fumer was designed and built to treat 100,000 tons per year of the QSL slag. Figure 2 shows the flowsheet of the plant.

OQ

'

c 3 5:

9 ft

CO

Red

uct

ant

Co

al

r- S

oli

d S

lag

^-

H

op

per

— F

uel

Co

al +

Car

rier

Air

—

Oxy

gen

Co

mb

ust

ion

Air

Aft

erb

urn

ing

Air

f P

an C

on

ve

yo

r')^

Lei

d

Fu

"»"»

B

ull

ion

S

lag

Fu

mer

Sla

g

Off

gas

t St

ack

Sp

ray

Wat

er

iv r

Gas

Q

uen

cher

ΓΠ7

Bag

Filt

er se

I

DF

an

> σ

N

5 n 8

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 335

Feed Handling

This plant has two hoppers; one is for storing coal to be used as the reducing agent and the other is a solid slag hopper which supplies the pre-granulated QSL slag. The solid slag is fed to the Ausmelt furnace when the QSL reactor cannot produce liquid slag temporarily because of maintenance. The feed rate can be controlled from 1.2 to 12 t/h. The lump coal used as a reductant is screened to 5-20 mm. Its feed rate is controlled in the range of 100-1,000 kg/h. Both materials are weighed by weigh feeders, and are delivered to the feed port in the furnace roof by a belt conveyor and a pan conveyor.

Lance System

The Ausmelt patented shrouded lance is used in the furnace. This lance is designed to introduce compressed combustion air (2.3 barg), pressurized pure oxygen (2.3 barg) supplied from a neighboring company in gaseous form and afterburning air (0.3 barg) supplied by blowers. Fuel coal is pneumatically supplied from two fuel coal bins with coal carrier air. The lance is operated by an electrical hoist attached to the lance trolley to hold the lance, and to control the depth of the lance in the furnace.

Furnace System

The inner diameter of the furnace is 3.9 m, and the furnace is lined with magnesite-chrome brick to a thickness of 300 mm. The outer shell of the furnace is cooled by water cascading from the top of the furnace to the bottom. Collected water after furnace cooling is sent to a water cooling facility and is recycled to the furnace. The QSL liquid slag is naturally introduced by level differences through launders into a side hole of the furnace, which is positioned above the slag level, and the Slag Fumer slag is discharged by a siphon system.

Gas Handling System

The furnace offgas is drawn from the furnace by an induced draft fan. The temperature of the offgas is kept below 800°C by spraying water at the inlet of the waste heat boiler. The outlet temperature of the offgas is around 350°C. The boiler typically produces 5-6 t/h of steam. A gas quencher uses a water spray system to cool the offgas to a low enough level to be sent to a bag house. The collected fume oxide from the bag house is processed in the existing oxide leaching plant. The lead content in the fume oxide is removed as a lead sulfate residue. This is sent to the QSL smelter for recovery of the lead. The cooled and cleaned offgas is discharged to atmosphere.

Metal Handling

Because of a strong reduction atmosphere in the furnace, some lead metal forms a layer in the bottom of the furnace. The metal is tapped every two or three days through a hole surrounded by a copper block. Molten metal is tapped into kettles through a castable-lined launder. The production rate is 10-12 tons per each tapping. It is sent to the present anode casting plant to be used as an antimony additive. Typical analysis of the metal is shown in Table III and the minor elements in the metal are:

336 LEAD-ZINC 2000

Arsenic : 2.0-3.0% Copper : 1.5-3.0% Silver : 70-200 ppm Gold : trace

Slag Handling

The slag discharged from the furnace is granulated by water jets and is transported into a pit. A grab crane driving along the rails on the top of the pit takes out the slag from time to time and loads trucks to stack on a stockyard or to transport to users.

Control System

The furnace and its associated infrastructure (offgas cooling, bag house, etc.) are controlled by an upgraded Programable Logic Control (PLC) system. Several monitors located in the control room display all process variables associated with the operation. Field instruments which comprise the furnace control system provide 4-20 mA inputs and outputs to indicate gaseous as well as liquid flows, pressures, temperatures and levels. The displays include snapshot indications, controllers, process trends and alarms as well as graphics.

PLANT HISTORY

Campaign 1 (October 1992 - April 1993)

The original commissioning was started on October 20, 1992. During this campaign, there were a number of operational problems to be solved. There were many challenges to introduce liquid slag into the Slag Fumer. Finally, liquid slag feeding was abandoned and the QSL reactor started to produce solid slag which was directly granulated. Accordingly, the Slag Fumer treated solid slag and re-smelted it to maintain a continuous and steady feed rate of raw material from April 1993.

Liquid slag charge can be maintained when the QSL reactor and the Slag Fumer are both operating well at the same time. However, because the lead smelting operation was also started in May 1992 and was not yet operating to design when the Slag Fumer was commissioned, liquid slag feeding to the Slag Fumer was frequently interrupted. Additionally, there was a long length of launder without adequate heating.

Campaign 2 (April 1993 - December 1996)

The operation of the Slag Fumer was much developed during this campaign. Average feed rate of solid slag was 6-8 t/h. However, the treatment of solid slag was strategically stopped on December 9, 1996 because the operation was uneconomical. Its feed rate was less than half of the original capacity for liquid slag and the fuel consumption was much higher than the design because of the re-melting of the solid slag. Instead of the Slag Fumer, the Zinc Fuming Plant treated solid slag by mixing it with goethite residue or conventional two-step leach residue. This operation had the merit that it recovered more valuable metals than the Slag

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 337

Fumer's operation by using two furnaces. Mixing residues with the slag also provided a better handling of the feed mixture, as the moisture content in the feed mixture was decreased.

Campaign 3 (September 1997 - May 1999)

After Campaign 2, there were many attempts to treat liquid slag. Two basic concepts of the study were 1) providing the shortest possible connection configuration between the slag outlet of the QSL reactor and the inlet of the Slag Fumer, and 2) installation of a device on the slag launder to maintain the temperature of the slag. Based on these concepts, the inweir, which was a refractory lined box to introduce liquid slag into the furnace with a siphon system, was removed and an overflow type hole was installed in the side of the furnace shell. This hole was located on the shortest straight line connecting the Slag Fumer and the QSL reactor as shown in Figure 3. Heat-up burners were installed on the cover of the launder to maintain the slag temperature. This change solved the problems that occurred in the earlier campaigns.

In the beginning stage of this Campaign, the quantity ratio between the charges of liquid and solid slags was 70:30. However, the ratios improved to 94:6 in 1998 and 87:13 in 1999.

There was a modification program during June 25 - August 14, 1998 that included capacity enlargements of the bag house and an induced draft fan to widen the operational flexibility. Metal production was started from August 29, 1998. Another modification was carried out in the period May 28 - July 5, 1999 to increase the retention time of the slag. The inner diameter of the furnace was increased from 3.3 m to 3.9 m.

f 1 '■ Steel Launder

- Original - - Present -

Figure 3 - Modification of the Launder for Liquid Slag Feeding

338 LEAD-ZINC 2000

Campaign 4 (July 1999 - Present)

Recoveries of zinc and lead were increased significantly as the retention time of the slag in the vessel became longer. Additionally, a higher fuel coal feed rate was also adopted. However, there was a limitation on the recovery of most of the valuable metals as fume oxide and metal, because of the higher content of their inputs which exceeded the original capacity of the furnace. Therefore, a plan is now projected for the installation of another furnace which will treat the QSL slag in two stages. An updated control system was installed in January 2000.

OPERATION

Operating Conditions

Typical operating conditions are shown in Table I. The QSL slag feed rates varied depending on the QSL operating conditions. However, the feed rate below is based on the normal operation of the QSL smelter at present.

Table I - Typical Operating Conditions of the Slag Fumer

QSL Liquid Slag Input Fuel Coal Rate Reductant Coal Rate Combustion Air Rate Fuel Coal Carrier Air Rate Oxygen Rate Afterburning Air Rate Afterburning Air Rate* Slag Temperature Inlet Slag Temperature Outlet

12.5 t/h 1,000 kg/h

300-400 kg/h 2,200-2,300 Nm3/h

400 Nm3/h 350 Nm3/h

1,500 Nm3/h 2,500 Nm3/h 1200-1250°C

< 1300°C * supplied through separately installed pipe lines into furnace

Performance

Some operating figures are given in Table II. Light oil is consumed for coal crushing and drying, and nitrogen is used for storing fuel coal in bins and delivering it from the coal crushing plant to the Slag Fumer. Liquefied natural gas (LNG) is used for burners heating the launder and outweir.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 339

Table II - Operating Figures for the Slag Fumer lNov'99to31Dec'99

Availability QSL Liquid Slag Treated QSL Solid Slag Treated Nitrogen Used LNG Used Light Oil Used Electricity Used

89.6% 12,011 t 1,967 t 1,413,647 Nm3

40,212 Nm3

16,323 1 570,183 kW

(1,312 hours) (12.36 t/h; 972 hours) (5.79 t/h ; 340 hours) (1,077 Nm3/h) (30.6 Nm3/h) (12.41/h) (435 kWh)

Current materials processed in the Slag Fumer are given in Table III. These compositions are slightly different from the figures previously reported (2). The assays are still changing as the feed rate and bullion production of the QSL smelter are continuously increasing. For example, this plant treated 165,200 dry tons of raw materials and produced 64,500 tons of lead bullion in 1994 and these figures were 339,300 and 121,000, respectively, in 1999 (3).

Table III - Typical Material Compositions (wt%)

QSL Slag

SF Slag

Fume Oxide

Metal

Zn

14-16

8-10

39.4

-

Pb

5-8

2-3

28.9

80.0

Sb

1.0

0.4

1.9

12.1

S

0.1

0.1

0.7

-

Fe

21.0

24.5

0.7

-

Si02

20.3

23.1

0.6

-

A1203

4.0

5.0

0.1

-

CaO

13.2

15.2

0.4

-

Distributions of Impurities

The distribution of impurities shown in Table IV was measured during Campaign 4. The compositions of arsenic and antimony in the QSL slag were both around 1%; however, their distributions were different. Arsenic mostly went into the fume oxide and most of the remainder was retained in the slag. Very little arsenic reported to the metal. The distribution ratios of antimony to the fume oxide and the slag were similar, but the metal also contained a significant amount of antimony. The copper content of the QSL slag is usually quite low; consequently, making an accurate assessment of its distribution is difficult. However, the copper content in the fume oxide was less than 0.1% and copper was mainly recovered to the metal.

340 LEAD-ZINC 2000

Table IV - Distribution Ratios of Arsenic, Antimony and Copper

Arsenic Antimony Copper Fume Oxide 80 44 Metal 3 18 25 Slag Fumer Slag 17 38 75

IMPROVEMENTS

Recovery Improvement by Using Finer Coal

Korea Zinc has a coal crushing plant which mills raw coal and delivers it to the QSL smelter, the Zinc Fumer and the Slag Fumer. The coal crusher produces two kinds of fuel coal classified by their particle sizes. One is called fine coal and the other is powdered coal. Their typical particle size analyses are shown in Table V. Because the QSL smelter consumes only powdered coal, the two Ausmelt plants use a 50:50 mixture of the coals. The Ausmelt lance was believed to have a range of flexibility in fuel coal size. However, a test operation was performed for several days to find the effect of fuel coal size on the operation. As shown in Table VI, using 100% of the powdered coal as fuel in the Slag Fumer operation improved the results significantly. However, this implementation could not be adopted permanently because of the production condition of the coal crushing plant.

One of the merits of Ausmelt technology is the flexibility in fuel types; the lance can be designed for using gas (2, 4), oil (2, 5) or coal (2, 6). In Korea, coal is the cheapest fuel. This test suggests that other plants which use coal for fuel through an Ausmelt lance should check the balance between the additional cost of making finer coal and the greater recovery which might be achieved.

Table V - Particle Size Analyses of Korea Zinc's Crushed Coals

Tyler Mesh + 6

+ 10 + 18 + 25 + 35 + 60

+ 100 + 140 + 200 + 230 + 325 -325

Powdered Coal (wt%) 0 0 0 0 0 0.4

10.9 14.8 13.0 2.2 7.5

51.2

Fine Coal (wt%) 0 0.1 1.4 2.3 6.4

32.6 42.3

9.6 4.2 0.6 0.4 0.1

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 341

Table VI - Comparison of Operational Results Depending on the Fuel Coal Size

Slag Input Total (t/h) Liquid Solid

Slag Input (wt%) Zn Pb

Slag Output (wt%) Zn Pb

Recovery (%) Zn Pb*

Coal Consumption Fuel (kg/h)

Reductant (kg/h)

Mixture Rate for Fuel Coal

50% of Powdered Coal + 50% of Fine Coal

13.3 11.3 2.0

13.9 7.8

8.6 2.9

49 69

923 330

100% of Powdered Coal

13.2 11.2 2.0

15.3 5.4

7.0 1.5

64 78

970 383

* includes recoveries to Fume Oxide and Metal

Metal Production

The metal layer in the bottom of the furnace reduces the volume of slag in the furnace. The metal could escape the furnace only through several holes in the bottom steel shell before the installation of the metal tapping system. Regular metal tapping provided improved profits and was one of the factors that improved the fuming reaction in the vessel.

Reduction of Labor Cost

At the original stage of commissioning in 1992, there were five operators per shift. The number was reduced to three operators per shift in 1994, and since May 1999, a shift has been covered by two persons. One operator controls the whole process from the control room and the other takes care of all machines at the site.

POSSIBLE FUTURE DEVELOPMENTS

Construction of the Second Furnace

As mentioned above, the capacity of the QSL smelter has been increased significantly and the capacity of slag treatment should be enlarged accordingly. In this respect, Korea Zinc started a project for the installation of one more furnace system to be operated in series with the present furnace. This will include building an Ausmelt furnace and its offgas handling system. The expected composition of the final slag is 3.8% Zn and 0.5 % Pb based on 13 t/h of the QSL

342 LEAD-ZINC 2000

liquid slag charge into the two-furnace system. The capital cost will be recovered in one year of operation.

Optimum Furnace Configuration

The Zinc Fuming Plant has devised an optimum furnace configuration developed from Korea Zinc's own idea for providing higher refractory life while increasing the operation temperature. The configuration implies a copper cooling system at the lower part of the furnace and a steam-cooled boiler tube system above the lower part including the furnace wall and offtakes, as shown in Figure 4. This configuration was installed at the Zinc Fuming Plant in 1999, and has stimulated the Slag Fumer to adopt a similar system in its furnace. This new arrangement for the vessel will be considered for the second furnace of the Slag Fumer.

Figure 4 - New Furnace Configuration of the Zinc Fumer

Oxygen Potential Measurement in the Furnace

The control of oxygen potential in the bath is one of the key factors which affects the operational performance. Therefore, measuring the oxygen potential in the bath is very important to respond instantly to varying bath conditions. Korea Zinc is currently measuring oxygen potentials in the bath of the QSL reactor, the Zinc Fumer and the Slag Fumer and some information was given in a previous publication (7). The measurement system currently used is a portable device. The main part of the device is a probe. The disposable probe tip reads the emf and temperature of the bath simultaneously. Therefore, the second stage of this measurement system will be developed towards the design of an automatic and long term device fixed in the furnace to measure the bath oxygen potential continuously.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 343

SUMMARY

Ausmelt furnaces were adopted in the Onsan Refinery to accomplish the integrated circuit between the zinc and lead plants. These furnaces have progressed steadily to fulfill this objective. Disposal materials from the refinery are all environmentally acceptable slags. From the economic viewpoint, the Slag Fumer makes a significant profit even now and it will grow larger as improvements are implemented. The Slag Fumer is still undergoing changes which will develop its metallurgical, mechanical and economic aspects.

REFERENCES

1. C.Y. Choi and Y.H. Lee, "Treatment of Zinc Residues by Ausmelt Technology at Onsan Zinc Refinery," Global Symposium on Recycling. Waste Treatment and Clean Technology (REWAS '99) Vol.11, I. Gaballah, J, Hager and R. Solozabal, Eds., Publication of TMS and INASMET, San Sebastian, Spain, 1999, 1613-1622.

2. J.M. Floyd and G.P. Swayn, "An Update of Ausmelt Technology for Zinc and Lead Processing," Zinc and Lead Processing, J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton and P. Hancock, Eds., Canadian Institute Mining, Metallurgy and Petroleum, Montreal, Canada, 1998,861-874.

3. Lead Smelting Team, "Annual Operation Result," Korea Zinc Internal Report, 2000.

4. J.M. Floyd and G.A. Johnson, "The Design of the Ausmelt Technology Smelting Unit for the Processing of Spent Pot Lining (SPL) to Portland Aluminium," Global Symposium on Recycling. Waste Treatment and Clean Technology (REWAS '99) Vol.11,. I.Gaballah, J, Hager and R. Solozabal, Eds., Publication of TMS and INASMET, San Sebastian, Spain, 1999, 1005-1014.

5. T. Sekiguchi and S. Azuma, "Slag Fuming at the Hachinohe Smelter," Zinc and Lead Processing. J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton and P. Hancock, Eds., Canadian Institute Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 299-311.

6. E.N. Mounsey, H. Li and J.W. Floyd, "The Design of the Ausmelt Technology Smelter at Zhong Tiao Shan's Houma Smelter, People's Republic of China," Smelting Operations and Advances. Copper 99 - Cobre 99. Vol.V. D.B. George, W.J. Chen, P.J. Mackey and A.J. Weddick, Eds., TMS, Warrendale, PA, U.S.A., 1999, 357-370.

7. Y. Lee, N. Moon and C.Y. Choi, "Oxygen Potential Measurements of Lead and Zinc Smelting Slags," Journal of the Mining and Materials Processing Institute of Japan. Vol. 116,2000,147-150.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 345

RECENT DEVELOPMENTS IN THE LEAD REFINING OPERATIONS AT BRITANNIA REFINED METALS LTD.

P.J. Moor Britannia Refined Metals Limited

Botany Road Northfleet, Kent, United Kingdom DA11 9BG

ABSTRACT

Britannia Refined Metals Ltd. was built in 1931 to refine the lead bullion produced at the Mt. Isa mine in Queensland, Australia. Several periods of expansion and development have taken place over the intervening years, leading to major increases in the refining capacity for both lead and silver. A plant for treating recycled lead-acid batteries was added in 1976. This was redeveloped in the early 1990's to utilise Isasmelt technology for the smelting operations. Most recently the No. 2 refinery kettle floor was extended to enable it to process lead bullion produced from two Imperial Smelting Furnaces (ISF), at Britannia Zinc Ltd. in Avonmouth, England and at MIM Hüttenwerke in Duisburg, Germany. This expansion included the provision of a plant for the hygienic handling of drosses and skims. This paper briefly describes the present primary refining operations and gives an update on the most recent developments in the process metallurgy and operating practices that have taken place.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

346 LEAD-ZINC 2000

INTRODUCTION

Britannia Refined Metals Ltd. (BRM) is a wholly owned subsidiary of MIM Holdings Ltd. and operates within the MIM (Europe) group that also includes Britannia Zinc Ltd. (BZL), MIM Hüttenwerke Duisberg (MHD) and Britannia Recycling Ltd. (BRL) at Wakefield, in Yorkshire. The BRM site is located on the south bank of the River Thames, 35 km east of London. There is a deep-water jetty adjacent to the plant where the crude lead from Mount Isa and the Imperial Smelting Furnace bullion from Duisberg are unloaded and stored.

The original plant was built in 1931 with a capacity of 72,000 tonnes per annum (TPA). The capacity was expanded in the 1960's to 140,000 TPA (1). A further refurbishment took place in the mid 1980's, both increasing the capacity to 180,000 TPA and introducing new technologies (2). Other improvements during the 1990's resulted in a capacity for Mt. Isa bullion in excess of 200,000 TPA. An overview of the flowsheet of the operation is given in Figure 1.

A separate plant for treating scrap batteries was constructed in 1975 and 1976, utilising the blast furnace process developed by P. Bergsöe & Sons of Denmark with six 120-tonne kettles. (3). Operations began in 1977. This operation was replaced by conventional battery breaking/short rotary furnace technology in the early 1980's; this technology operated until 1991. In 1991, this operation was replaced with a 'CX' battery breaking plant, supplied by Engitec Impianti of Italy, and an Isasmelt furnace for the smelting operations. This was the first commercial application of Isasmelt technology in secondary lead smelting (4,5). The plant has a capacity of 35,000 TPA lead. Two new 120-tonne refining kettles and a 25-kg ingot-casting machine were provided to refine the metal produced by the Isasmelt process. The short rotary furnace was retained to smelt drosses (6).

The kettle floor was no longer required for the secondary operation and was converted to process decopperised ISF bullion supplied by Commonwealth Smelting Ltd. (CSL) at Avonmouth. Two additional 160-tonne kettle settings were added. This became known as the BRM refinery. Initially the floor processed 6,000 TPA, on a toll basis, supplying metal for topping up the lead-splash condenser system on the CSL furnace. This was further increased to 24,000 TPA of which 18,000 tonnes was purchased and the same 6,000 tonnes continued to be tolled. In 1993 CSL (then operating as part of the Pasminco Europe Smelting Division) was purchased by MIM and the total lead bullion output, 45,000 TPA, then became the feed for this kettle floor. In 1994 MHD became 100% owned by MIM, and the possibility of processing the lead output from both ISF furnaces within the group was under consideration.

Trial work took place during 1996 on treating ISF bullion cast directly from the furnace. This work resulted in an extension to the kettle floor such that it is now capable of treating 55,000 TPA of direct cast ISF bullion from BZL and 36,000 TPA of similar material from MHD. Around 11,000 TPA of copper dross is generated from the revised operation and is sold. It is treated through a purpose-designed plant so that it can be hygienically transported by sea.

The end products are refined lead, lead alloys, refined silver and dore bullion. Refined lead of >99.99% purity is produced from the Mt. Isa refinery and 99.985% and 99.97% purity grades are moulded in the BRM refinery. A wide range of lead alloys is produced, particularly for the battery, lead sheet and cable industries. With the progressive increase in the silver and

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 347

MOUNT ISA LEAD

BULLION

PREMELTING I&2K

»J DROSS HANDLING^. i PLANT

COPPER DROSS "ST0 *-- »"«ii» & Ϊ4Κ " " " ' " ' ™

BZL BULLION

As ROTARY METAL

PRIMARY ROTARY

FURNACE I

SLAG TO BZL

CS

DESILVERING SKIMMING

3&4K

DESILVERING COOLING 3A&4AK

ISALO\M GRADE1"

BRMDORE 450kg BILLETS

99.9% Ag (1000 TROY OZ BARS)

±

REHEATING 32K

BRM LOW DESILVERING " GRADE ~ ~ " SKIMMING

29K

DESILVERING COOLING 27K&28K

J .

SECONDARY ROTARY

FURNACE

HIGH ANTIMONY fBISMUTH METAL

DEZINCING 25K

%

DEZINCING 5&6K

OXYGEN SOFTENING (KETTLE S3)

. HIGH BISMUTH METAL *

ANTIMONY REMOVAL

26K

TO SECONDARY

DEBISMUTHING 24K

ALLOYING 7,8,9 & UK

^ %*.

99.99% REFINED LEAD & ALLOYS

ISA KETTLE FLOOR

'X. * TOSECONDARY

SECONDARY ROTARY

FURNACE

HIGH BISMUTH LEAD

(SOLD)

_ _ LOW BISMUTH CAUSTIC SKIM

TO SECONDARY

FINAL REFINING 21&23K

99.985% AND 99.97% REFINED LEAD AND ALLOYS

PART REFINED LEAD RETURNED TO

BZL/MHD

BRM KETTLE FLOOR

Figure 1 - Primary Operations Flowsheet

348 LEAD-ZINC 2000

gold inputs from the ISF bullions, the percentage of dore produced has grown. It now exceeds the amount of >99.9% silver produced by fire refining the silver contained in the Mt. Isa crude lead. The combined output annually for the site is 260,000 tonnes of lead and lead alloys, 280 tonnes of 99.9 % silver, and 320 tonnes of dore, containing 250.kilograms of gold.

MT. ISA PRIMARY REFINERY

The operations in the Mt. Isa primary refinery have been covered by previous papers (2,6). The Mt. Isa bullion is produced by a conventional sinter plant/lead blast furnace operation. The resulting crude bullion is decopperised to less than 0.003 % Cu and is then given a caustic soda treatment to remove arsenic to less than 0.003%, prior to casting into 4-tonne slabs. This impure lead is shipped by railroad to Townsville and then by sea in 12,000- to 14,000-tonne shipments to BRM for further processing.

The impurities found in Mt. Isa bullion are particularly low and have been very consistent for many years. Bismuth is typically 1 0 - 1 5 ppm. and debismuthising is not necessary to achieve a 99.99% lead product. The refining stages on the Isa kettle floor consist of melting, desilverising, (in two stages), dezincing, and final refining using caustic/nitre. The refined lead is then either moulded or utilised to produce a range of alloys.

The silver/zinc crust produced in the desilverising operation is skimmed into 1000-kg moulds and is transferred to a separate silver plant where it is liquated to remove most of the lead. The resulting triple alloy is then retorted under vacuum to remove the zinc. The penultimate stage is cupellation in the bottom blown oxygen converter (BBOC), followed by a final conditioning in a conventional cupel and moulding into 1000-troy oz bars. The silver/zinc bearing by-products generated in the refining operation are treated through a short rotary furnace to produce a high-lead/zinc slag and metal containing all the precious metal values. The slag is sent to BZL for reprocessing through the ISF furnace. The metal is recycled through the appropriate kettle floor.

BRITANNIA REFINED METALS (BRM) PRIMARY REFINERY

The original process on this floor treats ISF bullion from BZL that has been tapped from the forehearth, cooled and drossed in a kettle and then moulded into 1.5-tonne blocks. The operating philosophy at the smelter regards the lead as a by-product. Should there be any likelihood that zinc production might be compromised by the drossing operation, the lead would be cast into blocks before it had been fully cooled in the drossing process. The quality of the copper dross that was generated was not particularly good and the hygiene on the drossing aisle did not meet current environmental standards, as they were experiencing major employee blood lead problems on their drossing aisle. The total copper content of the bullion blocks was typically around 0.5% Cu. On remelting, a dry skim was produced that contained 8 - 10 % Cu. A sulphur drossing stage was necessary to achieve the desired copper level prior to desilverising. The dross generated was typically 5-7% copper. These two skims had to be treated through a smelting operation to produce a low-grade copper matte that was sold at virtually no advantage.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 349

Pilot plant scale trials were carried out in May 1994, and these indicated that the treatment of direct cast bullion through the BRM kettle floor was feasible. The major issue that the trial raised was the sheer quantity of copper dross (0.4 - 0.5 m thick) that would be generated on a full size kettle together with all the problems of drying up and skimming. Following on from these trials, the concept of taking the lead output from MHD in addition to that from BZL was developed and discussed with MHD in November 1994. Investigation of the practicalities of direct casting at BZL was undertaken in April and June 1995. The Noyelles Godault plant of Metaleurop was visited to view their direct casting of ISF bullion.

The decision to proceed with the overall project was taken in July 1995. Detailed design work then began for the extension to the existing kettle floor. This included the provision of five additional kettle settings and a new overhead crane. A significant element within the overall project was the provision of a plant designed to treat the copper dross and make it suitable for shipment abroad. Once the decision to proceed had been taken, it was necessary to undertake a programme of testwork with some degree of urgency, as the new operation was required to be in place by January 1997.

A trial batch of 540 tonnes of MHD material was processed in December 1995. This was followed by several trial batches during 1996, ending in a trial in September 1996 of material supplied by both smelters. The objective of the trials was to develop an operating procedure to provide metallurgical information on the expected analyses and cycle times and to evaluate the kettle mixing requirements for a full-scale operation. The major concerns the trials highlighted were:

• Potential inclusion of large lumps in the blocks of iron speiss. Although the smelters looked at minimizing the amount of speiss that would report to the metal, it was stated that the material could not be guaranteed to be speiss-free.

• A 132-kW mixer was needed to mix in the initial dross layer. This raised additional issues associated with the size of the power cable, plug, etc.

• Removal of very thick dross using the standard dedrosser was very slow and the dross quality was poor. The machine also proved incapable of handling lumps that frequently broke the drive chain.

In spite of these concerns, however, the overall assessment of the final trial was positive. The concerns relating to the third item above prompted the decision to change to the Worswick dedrosser (see Figure 2). This has proven to be an excellent choice, in retrospect. It was necessary to redesign the three dressing kettles and settings to accommodate an inclined beach to suit the different type of skimmer.

350 LEAD-ZINC 2000

132 kW Mixer

Figure 2 - Side Elevation of a Melting/Drossing Kettle

Deliveries of direct cast ISF bullion from BZL commenced in late December 1996, ready for the planned start up of the new kettles in January 1997. Deliveries of material from MHD were delayed until the end of February 1997. There still needed to be significant modifications to the original eight kettle settings to fit into the new flowsheet.

The commissioning of the new process was accomplished successfully. A problem was encountered with speiss lumps contained in the bullion blocks. The thickness of the dross layer made it impossible to detect this speiss before it came into contact with the paddle on the kettle mixer. Such contacts lead to damage to the equipment and long process delays while efforts were made to remove the lumps from the kettle. The practical difficulties in removing the lumps and the poor hygiene made it an urgent requirement to find a method of identifying blocks containing speiss. Analysis of the speiss revealed that it is iron arsenide, containing -70% iron, -20% arsenic. The speiss was magnetic and a suitable process was developed by May 1997. This enabled "reject" blocks to be identified and segregated. Until May 1997, the overall stocks of bullion on site rose to over 6,800 tonnes. Arrangements were made by MHD to sell 1000 tonnes per month until June 1997 and another 500 tonnes per month until December 1997 to reduce stocks. Once a screening facility was in place, all of this tonnage was tested and separated into "good" and "reject" blocks. The refinery was then able to operate at a normal rate, confident that there would not be lumps in the kettle. However, the screening left a large stock of rejected blocks. These progressively accumulated while a suitable operating procedure for melting and removing the speiss was devised and tested. Procurement and installation of the necessary equipment took until December 1997, by which time over 3,500 tonnes of rejects were on hand. Processing the MHD rejects was then carried out and was completed by the end

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 351

of December 1997. The BZL rejects were treated from January 1998 on, as well as keeping the stocks of MHD rejects to a minimum. By April 1998 the backlog of rejects had been treated. Operational difficulties at BZL limited the total tonnage being delivered thus giving additional processing time to speed the reprocessing of the rejected material. MHD has continued to sell 500 tonnes per month although this is due to cease in December 1998. The rejects are now segregated and treated on a routine basis from week to week.

The overall refining process is currently working well, with over 1900 tonnes of bullion charged in a week, when supplies of material permit. The maximum capacity of the plant is still to be established.

Screening and Speiss Handling

Both of the ISF furnaces are tapped on an intermittent basis so that there tends to be a rush of speiss at the start of a cast followed by a reduced amount throughout the tap. The objective is for the speiss to go with the slag and then to disposal. This leaves the metal relatively clean. The process is very dependent on the forehearth condition and the operator attention to the different levels in the forehearth. The two different bullion sources resulted in obvious differences in operating conditions. The casting system at BZL uses a cascade of 5 to 6 moulds, such that the speiss is concentrated in the first block filled with some carryover into the second and third moulds. The casting system at MHD uses a flat rail-mounted carriage and the speiss shows up in the first mould filled and can spread across several blocks. The BZL speiss tends to occur as quite large lumps, whereas MHD speiss tends to be more plate-like. The surface of the MHD blocks is significantly flatter than the BZL blocks. This led to different screening criteria being used, especially as the scale of the operation was modified.

Both BZL and MHD produce their material in batches of approximately 500 tonnes. This is in line with an agreed sampling procedure for the metal as it is cast. The MHD bullion is delivered by ship, with up to three 500-tonne batches on a vessel. The BZL bullion is delivered by truck in 24-tonne lots. BZL numbers the blocks as they are cast so that their position on the cascade can be identified. This allows blocks from moulds 4 and 5 to be put straight to the "good" stock area, because they do not contain speiss. All of the remaining material is screened.

The screening process is very simple. A 1.2-meter diameter electromagnet is suspended from an overhead crane in the crude lead storage building. A digital crane scale is placed between the magnet and the crane hook. The block to be tested is positioned on top of a block of Mt. Isa crude lead. This assists the forklift driver to locate the test block and removes any influence on the test from the reinforcing steel in the concrete floor. The magnet is lowered as close as possible to the top surface of the block, without touching it. The magnet is turned on. If there is any speiss in the block, there will be a downward force on the magnet, thus giving a reading on the crane scale. The larger speiss lumps result in a higher readings. This is obviously only a qualitative test but it is sufficiently selective to remove "bad" blocks. The cut-off for BZL material was set at 150 kg and for MHD at 500 kg. The cut-offs were established by screening large quantities of blocks and segregating them according to their crane scale reading. The blocks were then treated through the kettles in specific batches and the chemical analysis was reported. Thus, any blocks at this or a higher level would be rejected. The difference between the two levels is related to the fact that it is not possible to position the magnet as close to a BZL block as it is to a MHD block. This is due to the much rougher surface finish on a BZL block and the field strength decreasing according to an inverse square relationship.

352 LEAD-ZINC 2000

The rejected blocks are kept separated (BZL and MHD) and are reprocessed in small batches by carefully adding them to a dross-free bath in a melting kettle. The speiss floats and is removed from the bath using a 1.4-meter diameter electromagnet. The speiss lumps are then carefully dropped down a ventilated chute into a skip. The cooled speiss is weighed and sent to BZL for disposal. This method has proven to be both safe and hygienic. Considerable efforts have taken place at BZL to improve speiss separation in the forehearth, and the speiss levels have decreased in recent months. It is not unusual to find that a batch of 125 blocks contains no rejects.

Kettle Floor Operations

The revised kettle floor consists of thirteen kettle settings, as shown in Figure 3. The five new settings hold 250-tonne kettles whereas the eight older settings ranging from 160 to 230-tonne capacities. They are assigned as follows:

No. 34 Melting/Copper Drossing No. 32 Reheating No. 31 Melting/Copper Drossing No. 30 Melting/Copper Drossing No. 29 Desilverising - Skimming No. 28 Desilverising - Cooling No. 27 Desilverising - Cooling No. 26 Antimony Removal No. 25 Dezincing No. 24 Debismuthising No. 23 Final Refining/Alloying No. 22 Bismuth Dross Treatment No. 21 Final Refining/Alloying

250 tonnes 250 tonnes 250 tonnes 250 tonnes 250 tonnes 230 tonnes 230 tonnes 200 tonnes 210 tonnes 200 tonnes 180 tonnes 160 tonnes 180 tonnes

(with beach)

(with beach) (with beach)

1-Tonne Moulding Wheel

DeZn

40-kg Ingot Casting Machine

Unit Stand

V

N*o Storage

Area

\ 1

i/l ( 2 1 W / 2 3 )

(22Y« (24V*

/

(25V.

h-(26)

1

K ( 2 7 W

V28J

Office

Crude Charge

B ' \

-(29) (31)

(ioVl "VV"

Storage Area

«-(34)

Figure 3 - BRM Kettle Floor Layout

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 353

Melting and Dressing

Blocks of either BZL or MHD bullion, having the composition shown in Table I, are charged two at a time to the desired melting kettle using a crane operated hydraulic grab. Rotary metal, from the smelting of silver bearing by-products, is returned to the process via the melting kettles. There is normally a heel of ~40 tonnes left in the kettle from the previous charge. At present, this has to be allowed to freeze because of the risk of explosions from water that may be absorbed in the dross layer on the top of the block. A storage facility, that includes a new screening station, was constructed to minimise this problem. Once sufficient blocks have been charged, the melting begins. More blocks are charged until the kettle is full. A hydraulically operated hood provides satisfactory hygiene ventilation for the kettle. The bath is stirred at 460°C. Pulverised fuel ash (PFA), from a coal-fired power station, is used to help dry up the copper dross. Once the dross is in a suitable condition, the mixer is slowed to minimize dross movement. The Worswick dedrosser is then started and the dross is skimmed into 1-cubic meter skips. The dross is transferred to the by-products treatment plant for further processing. A keel plate is inserted through the kettle hood to help collect the final dross. This allows the kettle to be skimmed virtually clean. Rejects may then be added, when available. The copper content of the lead bath is typically 0.25% Cu at this stage. Most of the tin content reports to the copper dross, as does virtually all of the iron, nickel, tellurium and indium. In order to achieve the desired level of copper, the charge is cooled to 330°C and is pumped to the reheating kettle. The maximum copper requirement for the dore is 3%; consequently, the final copper content achieved by cooling is acceptable.

Time Cycle 30 hours Copper Dross Make 11 -14 % Fuel Consumption 150kWh/tonne Final Copper Level 0.05-0.06 %

Pump Out Temperature 330°C

Table I - Typical Analyses for Mt. Isa Bullion, the Two ISF Bullions and the Copper Dross.

Element Mt. Isa Bullion BZL ISF Bullion MHD ISF Bullion Copper Dross (wt%) (wt%) (wt%) (wt%)

Antimony Arsenic Copper Bismuth

Zinc Silver Gold

Indium Iron Tin

Nickel Lead

0.08 < 0.003 < 0.003 0.0015

nil 0.23 nil nil nil nil nil

99.5

0.61 0.17 3.46 0.13 0.25 0.40

0.0011 0.016 0.10 0.18 0.04 94.2

0.35 0.10 3.20 0.10 0.25 0.18

0.0002 0.014 0.08 0.34 0.04 95.0

2.0 1.4

26.0 - 30.0 0.12

2.0-3.0 0.25

0.0040 0.10

2.0-3.0 1.0 0.3

50.0-55.0

354 LEAD-ZINC 2000

Reheating

The purpose of the reheating kettle is to raise the temperature of the charge to 470°C before transferring the bullion to the first stage desilverising kettle. A sample is taken to establish the impurity levels contained in the batch. The silver and copper contents are of particular interest for the next stage of the refining operation. In spite of the antimony level contained in the lead going to desilverising (up to 0.4 % Sb), no particularly adverse effects have been observed. Previous experience suggests that arsenic contents in excess of 0.01% As have a major impact on silver recoveries in the liquation process. Fortunately, there is virtually no arsenic left in the lead after copper drossing. Arsenic is typically 0.005% or less. Trials of direct transfer from the drossing kettles have occurred in the past, but problems with sample quality and the transfer of copper dross through the pump resulted in a reversion to the use of this step. Time cycles in the skimming kettle were also extended as a result of filling at the much lower temperature.

Desilverising

Desilverising follows the same two-kettle process originally developed on the Mt. Isa kettle floor in the early 1980's. The lead from the reheating kettle is pumped into the first stage kettle at 460°C. The kettle contains a low-grade silver/zinc crust on top of a large heel. The batch is mixed and the crust reacts with the new lead. This upgrades the silver content in the crust to -7% Ag. This crust is skimmed off into 1000-kg moulds and a steel anchor is inserted into the block. When cooled, the alloy is removed from the moulds and is transferred to the silver plant for reprocessing. The silver/zinc/gold material is treated separately through the same process as the equivalent Mt. Isa material, except that the resulting dore is cast into 450-kg billets and is sold to a precious metals refiner for silver-gold separation. Following skimming, an addition of zinc is made to the kettle to replenish the zinc content of the system. This bath, after stirring, has a regenerated low-grade crust layer. The lead is transferred to the second stage kettle by a centrifugal pump leaving the low-grade crust behind.

The second stage kettle contains a frozen crust of -30 tonnes, left from the previous charge. The kettle is filled such that this crust is covered by hot metal. The kettle is covered and the crust is remelted at 450 to 460°C. The kettle is given a short mix and 5 to 6 tonnes of low-grade skim is removed. This is returned to the first stage kettle, when it is empty. The kettle is cooled to 320°C. A mixer attached to a variable speed drive unit is used during this phase. This keeps a hole in the centre of the crust as it forms and prevents silver/zinc crystals freezing on the kettle walls. At 320°C a preheated transfer pump is used to transfer the lead to the next stage.

Cycle Time - First Stage 8-10 hours Cycle Time - Second Stage 12-14 hours

Total Zinc Addition 15 kg/tonne New Zinc Consumption 2.4 kg/tonne

Fuel Consumption 75 kWh/tonne Final Silver Content 4 g/t

Dezincing

The zinc content of the desilverised lead is -0.55% Zn. This is recovered by vacuum distillation, using the standard BRM-developed technology, illustrated in Figure 4. The kettle

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 355

has a water-cooled rim with a machined top surface. There is an offtake below this rim that connects to a vacuum pump. The rubber seal of the water-cooled dezincing unit sits on the kettle rim providing a vacuum seal. A mixer passes down through the centre of the dezincing unit, providing agitation during the dezincing process. The lead is reheated to 590°C and the surface is skimmed. The dezincing unit is placed in position and a vacuum is created. The mixer is started and dezincing commences. At the end of the process, the vacuum is released and the dezincing unit is removed. The zinc is scraped from the condensing surface and is recycled to the first stage of desilverising. Typically the vacuum during dezincing is less than 0.05 mbar and residual zinc is around 0.01% Zn.

Total Cycle Time Dezincing Time

Dezincing Temperature Vacuum

Final Zinc Zinc Recovery

Fuel Consumption

14 hours 5 hours 590°C

0.05 mbar 0.01% 96%

98 kWh/tonne

water cooled condenser

cooling water feed & return

stirrer drive

condensed zinc

kettle floor control panel

kettle

two stage vacuum pumps

Figure 4 - Vacuum Dezincing System

Antimony Removal

The dezinced lead will contain up to 0.4% antimony. This is removed by stirring at 520°C with caustic soda and nitre. The resulting dross is skimmed using a continuous belt dedrosser into standard 1 -cubic meter skips for reprocessing through a short rotary furnace. The antimonial lead produced is oxygen softened in a kettle (S3) and the soft lead is recycled to the process. The high-antimony slag is smelted to give 25% antimony lead for alloy production.

356 LEAD-ZINC 2000

Cycle Time Reagent Usage

Antimony in Skim Skim Produced

Fuel Consumption

11-13 hours 5.6 kg/tonne

10-12% 4%

50 kWh/tonne

Debismuthising

The bismuth contained in the lead is removed using the conventional Kroll-Betterton process. Proprietary MagCal™ is added to the lead at 490-500°C. The input bismuth content and the desired final analysis determine the level of addition. The bath is cooled and charcoal is added to the dross layer at 420°C. The purpose of the charcoal is to maintain the temperature in the dross layer and to promote separation from the lead. The kettle is cooled by forced air to 340°C, at which temperature the dross is skimmed. As soon as this is completed, the cooled lead is transferred to the final refining stage. The high-bismuth dross is skimmed directly to a kettle where it is stirred to oxidise the reagents and to give a 6% bismuth lead. The resulting dross is reprocessed through a short rotary furnace and the metal is recycled to the dross treatment kettle.

Cycle Time Input Bismuth Content

Final Bismuth Achieved Reagent Usage

Bismuth In Dross Fuel Consumption

12 - 14 hours 0.1 - 0.25 %

0.005 % 2.1 kg/tonne

6 -10 % 25 kWh/tonne

Final Refining

The residual reagents contained in the lead are removed using caustic soda and nitre, leaving either 99.985 or 99.97% lead. The refined lead is then either moulded directly as 40-kg ingots or 1-tonne blocks, or it is alloyed first to make a range of low alloying element alloys. The kettles used for this process have recently been uprated to 180 tonnes capacity, so as to permit extra tonnage per batch. This upgrade is approximately 15 to 20 tonnes per batch or about 7,500 tonnes per annum.

Cycle Time (Includes Moulding) 12 hours Refining Temperature 490°C

Skim Produced 3.2 % Reagent Consumption 1.5 kg/tonne

Fuel Consumption 50 kWh/tonne

Partially Refined Lead

There is a requirement to produce -1000 tonnes per month of partially refined lead for return to the two smelters as top-up for their lead-splash condenser circuits. In essence, the process route is the same as given above except for the omission of the debismuthising and final refining stages.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 357

By-products Treatment Plant

Within the scope of the overall project the provision of a plant to treat the various on-site by-products was a major component. The problem of the hygienic handling of dry, dusty by-products was one that had exercised people at BRM for a number of years. The end result of much trial and test work centred on the use of a dilute solution of molasses to bind and de-dust the material. The plant, as eventually constructed, consists of two main elements; namely, a screening section and a storage/blending section. The plant is divided into two lines, one for high-silver materials and one for low-silver materials. Each line is equipped with two storage silos, one for high bismuth materials and one for low bismuth materials. Problems were experienced in the early operations from the reaction between the free caustic in the caustic skims and the dilution water with the molasses. This caused the molasses to break down and become ineffective. Trials with other binding agents resulted in the selection of calcium lignosulphanate for caustic-bearing by-products. The plant is illustrated in Figure 5, and operates as follows.

Storage Silos

Low Silver Line Molasses Mixer

Figure 5 - By-Products Treatment Plant

The various by-products are stockpiled to accumulate sufficient material to fill a silo. By far the largest amount of material requiring treatment is the BRM floor copper dross. The material is allowed to cool for several days to less than 30°C. The molasses will degrade at

358 LEAD-ZINC 2000

higher temperatures. The individual skips of material are lifted by an overhead crane and are tipped by an automatic tipping system into the screening plant. The first screen separates any lumps greater than 0.05 m. This oversize is returned to the kettle floor for remelting. The second screen is a 0.005 m mesh. The oversize from here, called "middlings", is checked for any tramp materials and then is added back to the stockpile of treated copper dross. The fines discharge onto a drag-link conveyor, which deposits them into the high-silver, high-bismuth storage silo. When full, it is discharged via a loss-in-weight controller through a mixer in which a 20% solution of molasses is blended with the fines. This produces a material that resembles breadcrumbs and is non-dusting. The treated material discharges onto an enclosed conveyor belt that transfers the finished copper dross to the by-product storage shed. There, it is stockpiled until sufficient has accumulated to make up a shipment. The other three silos discharge through similar molasses mixers and then back into skips for transfer to various covered storage locations. When a shipment of copper dross is made, four half-height, 30-ft containers are used. These have been strengthened and modified with an inclined plate at one end, to be similar to a skip. They are loaded onto the back of a standard flat bed trailer, filled with -21 tonnes of dross using a front-end loader, and then taken to a ship. The containers are finally lifted off of the vehicle, lowered into the ship's hold and tipped out. A 1200-tonne shipment can be dispatched within two days with no dust emissions or requirement for extraction.

Environmental

During the early months of operation of the new process, lead-in-air results for the area rose dramatically. The problems were associated with the removal of lumps from the kettles and the amount of dry copper dross left in the kettle between charges. Once screening started, the lead-in-air figures decreased in a rapid manner. Since January 1998, lead-in-air has been consistently below 0.10 μ$Ίη3. At the same time, there was a significant correlation with employee blood leads across the site. Considerable efforts have been made to identify and target actions to reduce operational exposure to lead-in-air. The current site average employee blood lead level (January 2000) is 25.3 μg/dl. We currently have no employee with a blood lead greater than 40 μg/dl. The company has recently achieved accreditation to the ISO 14001 standard.

CONCLUSIONS AND FUTURE DEVELOPMENTS

The current operations are not regarded as definitive. There are a number of areas that can be improved. The initial trials indicated that the grade of copper dross that is present when the dross is first dried up can exceed 35-40% copper. It is intended to carry out further trials to establish the highest grade that it is possible to achieve on a regular basis. There are considerable financial advantages to be gained from maximising the copper content of the dross. There is a major investigation underway to look at ways to minimise the use of caustic soda/nitre in the refining operations and to explore alternative retreatment routes that do not result in soda slag material to be disposed. New conveyors are to be installed in the by-products plant to improve plant availability and minimise cross contamination of materials. With the trend in battery manufacture towards the elimination of antimonial alloys, there is a programme to examine ways of upgrading antimony by-products, and hence, to find alternative outlets for antimony. Similarly, the upgrading of bismuth dross, and hence the resulting bismuth lead, is another area currently being investigated.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 359

ACKNOWLEDGEMENTS

I should like to express my thanks to Britannia Refined Metals Ltd. for permission to present this paper and also my appreciation to those of my colleagues who assisted in its preparation. In particular I should like to mention Tony Piper who is a constant source of ideas and inspiration.

REFERENCES

1. K.R. Barrett and T.G. Morgan, "Expansion of Britannia Lead Co. Ltd. Refinery", Proceedings Ninth Commonwealth Mining Congress, 1969.

2. K.R. Barrett and R.P. Knight, "Lead Bullion Refining and Precious Metal Recovery", Extraction Metallurgy 85, IMM, London, England, 1985, 683-708.

3. J. Taylor and P. Moor, "Secondary Lead Smelting at Britannia Lead Company Ltd.", Lead-Zinc-Tin '80. J.M. Cigan, T.S. Mackey and T.J. O'Keefe, Eds., Metallurgical Society of AEVIE, Warrendale, PA, U.S.A., 1980, 1003-1022.

4. K. Ramus and P. Hawkins, "Lead-Acid Battery Recycling and the New Isasmelt Process", 3rd. European Lead Battery Conference. M.G. Mayer and D.A.J. Rand, Eds., Munich, October, 1992, 299-314.

5. F. Ahmed, "The Battery Recycling Loop: A European Perspective", Journal of Power Sources. Vol. 59, 1996, 107-111.

6. R.P. Knight and R.J. Reader, "The Refining of Lead Bullion and Precious Metals Recovery at Britannia Refined Metals Ltd.", Lead into the Future. IMM, London, England, 1996,53-73.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 361

ANALYSIS OF DUST FORMATION IN THE OXYGEN FLASH SMELTING OF LEAD-BEARING FEEDS IN THE KIVCET SMELTER

M.A.Lyamina, V.A.Shumsky, I.P. Polyakov, N.M.Ospanov and N.N.Ushakov VNIItsvetmet Institute 1, Promyshlennaya St.

Ust-Kamenogorsk 492020, Kazakhstan Republic

ABSTRACT

Studies on the kinetics of the competitive processes of PbS volatilization and PbS oxidation, as applied to the heat and mass transfer conditions in the flame of the Kivcet furnace, revealed that the rate-determining factor is the supply of heat. Therefore, low yields of fume in the autogenous flash smelting of lead sulphide feeds are noted despite the high volatility of PbS at 1,300-1,400°C. Dust formation in the autogenous flash smelting of Kivcet-smelters does not exceed 5-7%, of which fumes amount to ~ 80%. Using extra fuel during autogenous smelting involves the mismatch of the sulphide oxidation rate and the release of heat in the flame, and increases the fume yield. This is a major cause of the high dust formation in pilot plant trials. The analysis of dust formation in the Kivcet smelting of oxidized lead feeds in an industrial-scale smelter shows an overall increase of dust formation (up to 10% of the feed), with fumes constituting ~ 80% of the dust. A pronounced increase of the zinc component of the fumes is also noted.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

362 LEAD-ZINC 2000

INTRODUCTION

For more than 20 years, the process of oxygen flash smelting of lead-bearing feeds (the Kivcet process) has been progressing as a result of various innovations, but its key concept has not changed. However, the issue of dust formation in the flash smelting of lead-bearing feeds is still not clearly understood. That is the reason that there is still no optimum solution for the problems occasionally arising in relation to increased carry-over of dusts in Kivcet units, as a consequence of the wide range of feeds treated.

The development of the Kivcet process and other processes for the flash smelting of lead sulphide feeds progressed during the period from the 1970s until the early 1980s, as based upon numerous experimental data obtained during the operation of small pilot units. Those studies showed a high formation of smelting dust, as much as 15-25% (sometimes up to 55%) of the charge. Because of this, it was believed for a long time that the high carry-over in the flash smelting of finely divided sulphide feeds is determined by the large surface area and the high specific rate of PbS volatilization at temperatures of 1,300-1,400°C, characteristic of this process. In contrast, the industrial-scale flash smelting of lead sulphide feeds in the Kivcet smelters owned by the Ust-Kamenogorsk Lead and Zinc Operations (Ust-Kamenogorsk, Kazakhstan) and "Nuova Samim" (Portovesme, Italy) showed that, in the stable smelting process, the carry-over of dust did not exceed 5-8% of the charge (1, 2). The experience of the operating Kivcet smelters suggests an increased average yield of the flash smelting dust when changing from sulphide feeds to oxidized lead-bearing materials. Thus, for example, in the smelting of oxidized feeds, based on lead-containing residues from hydrometallurgical zinc processing, in the Ust-Kamenogorsk Kivcet smelter, the carry-over averaged 10-12 % of the charge and often amounted to 15-17% of the charge. Similar and even higher carry-over values have been observed during the smelting of oxidized lead-bearing feeds at the Cominco Kivcet smelter. At first sight, however, this runs counter to the experimentally found ratios of the volatilization rates of lead oxide and lead sulphide. These volatilization rates are shown in Figure 1, where the plotted curves for PbS and PbO volatilization are based on experimental kinetic relations of the following type:

kpbs = 3.36106exp(-23960/T), mole/(m2s) kpbo = 0.541CCexp(-19630/T), mole/(m2-s)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 363

101

_ 10° in

1 JUT* βΓ

3 _ S io-2

o

I io"3

ft O

* 1 . «

900 1000 1100 1200 1300 1400 1500 Surface temperature, °C

Figure 1 - Temperature Dependence of the Specific Volatilization Rate of Lead Compounds

We assume, that at the heart of the dust formation problem in the flash smelting of lead feeds lies the general laws governing the carry-over values actually observed that depend on the scale (pilot or industrial) of the Kivcet unit and the quality of the lead-bearing feeds treated.

Accordingly, this paper presents an analysis of dust formation during the flash smelting of lead-bearing feeds in the Kivcet process, using Kivcet experimental smelting data and special-purpose laboratory studies.

THE NATURE OF THE CARRY-OVER PROBLEM

Formation of flash smelting dusts is a combined effect of the fuming of volatile compounds from the feed-oxygen flame and the mechanical carry-over of fine particles with the gas stream.

The role of volatile lead compounds in the formation of flash smelting dusts is easy estimated if a comparison of the ratios of the lead/non-volatile slag-forming compounds (S1O2, CaO, AI2O3, Fe) in the feed and in the smelting dusts is made. As a rule, oxidized lead compounds constitute the majority of the dust. Therefore, the proportions of the lead volatilized and the lead mechanically carried over with the dusts, found by simple calculation, reflect the actual ratio of volatilization and mechanical carry-over of the feed in the overall yield of the smelting dusts. The estimates made on the data obtained from the industrial smelting of lead feeds in the Kivcet smelter at the Ust-Kamenogorsk Lead and Zinc Operations showed that 75-90% of the total amount of lead in the flash smelting dust is accounted for by the fuming of volatile lead compounds. Similar estimates made for the Kivcet smelting of lead feeds at the

364 LEAD-ZINC 2000

pilot unit installed at the Pilot Lead Operations (Ust-Kamenogorsk, Kazakhstan) showed that the percentage of lead fume amounted to 50-90% of the total quantity of lead in the flash smelting dust.

In general, the ratio of the volatilized material and that carried over mechanically in the dusts generated during the flash smelting of lead-bearing feeds depends on a number of factors, and therefore, is not stable. In particular, the total amount and structure of the carry-over is influenced by the composition of the feed treated, the temperature of smelting, and the geometry of the Kivcet unit which determines the heat transfer in the reaction shaft and gas dynamics of the off gas and dust streams. The correlations mentioned are seen clearly from the experimental findings obtained in the Kivcet smelting of lead-bearing feeds, having various compositions, at the pilot-scale unit and in the industrial smelter (see Figure 2). It is seen from the experimental results presented, that no matter how the factors governing the total carry-over values change, the key role of the volatilization of lead compounds is consistent in the overall process of dust formation during flash smelting. Therefore, the macrokinetic features of the volatile lead compounds in the feed-oxygen flame are of principal interest in the analysis of the flash smelting carry-over dust.

Figure 2 - Correlations of the Carry-over Factors with the Quality of the Lead Feeds, Flame Temperature and the Scale of Kivcet Smelting

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 365

Because of this, the flame smelting conditions (autogenous or with the use of extra fuel) are important. In the oxidation of sulphides, the relevant desulphurization reactions (PbS, ZnS, FeS), that are also major contributors of heat, are of the type:

MeS+1.5 02 = MeO + S02 0)

If the thermal effect of an individual reaction of the type shown in Equation 1 is Qi( the heat release rate in a volume V is determined by the following relations:

q = ZQi-v[°2jn i -v· koi -exp R-T: 1 /J

(2)

The feed desulphurization rate of reaction (1) is determined by the following relation:

w = £sr[02]n '-V koi ' e x P _

R-T: (3)

where Sj is the surface area of the i sulphide phase (PbS, ZnS, FeS) being in volume V; [O2] is the current oxygen level; n, is the i reaction order with respect to oxygen; koi and Ej are parameters of the equation for the i oxidation reaction; R is the universal gas content; and Tj. is the temperature of the 2 sulphide phase surface.

The released energy is used to heat the sulphide particles and the gas phase, and it is lost in part through the reaction shaft sidewalls. Equation (2) differs from Equation (3) only in the factor of Qj. Thus, the metal sulphide oxidation processes and heat release from the feed particles during autogenous smelting are unambiguously interrelated.

Oxidation of the lead sulphide feed occurs with coincidental volatilization of the lead compounds PbS and PbO. Therefore, a maximum rate of PbS volatilization from the flame is noted during the intermediate step of feed oxidation. This follows from the relations of the lead sulphide volatilization rate:

W PbS »sp •( l -a)2 / 3 -k,

'sp •exp Jsp

R-T, spy (4)

where s0s is the initial surface of the PbS particles; a is the degree of PbS conversion, k0s

and Esp are the Arrhenius equation parameters for PbS volatilization (see Figure 1); and Tsp is the surface temperature of the PbS particles.

First, the velocity of PbS volatilization increases rapidly with the rising temperature of the particles along the flame height because of the exothermic sulphide oxidation reaction (1). At the same time, because of the concurrent oxidation and volatilization processes, the surface temperature of the PbS particles decreases. As a result, once the PbS volatilization rate has reached a maximum value in the high temperature zone of the flame, it decays (actually as low as zero for complete feed desulphurization) despite the character of the temperature change of the particles along the whole of the flame. Thus, under any condition of flash smelting, the lead sulphide volatilization rate is of extreme importance, with the maximum rate occurring within the intermediate stage of feed oxidation.

366 LEAD-ZINC 2000

In contrast, the volatilization rate of lead oxide in the flame increases montonically and the pace of volatilization is dictated by the heating rate of the particles along the flame height. This follows from the kinetic expression for PbO volatilization:

W PbO=s 0 x -ko o x " e x P

f E

V * ^ ' *■ ox J

(5)

where Sox and Tox are the area and the temperature of the PbO phase surface, and k0 and E, ox

are the Arrhenius equation parameters for PbO volatilization ( which appear on Figure 1). Generally, the value of the PbO surface is variable and its change is different for the PbO phase formed by the oxidation of PbS particles (when treating sulphide feeds) relative to the PbO formed by lead sulphate decomposition (when treating oxidized feeds).

Agglomeration of mixed feed particles may also influence the volatilization rate of the lead oxide in the high-temperature zone of the flame, because PbO reacts with the associated S1O2 particles at temperatures higher than 800°C, forming less volatile lead silicates. Special examination of the flame samples provides evidence that similar interactions occur in the flame.

The mechanism of lead compound volatilization described above explains the relative insensitivity of the total carry-over dependence on the temperature of the lower flame (Figure 2 (b)). At the same time, based only on these laws and having no regard for the phenomena of heat and mass transfer in the gas phase of the flame, it is impossible to explain either the reduced carry-over during the industrial-scale flash smelting of sulphide feeds or the increased carry-over when smelting oxidized feeds (See Figure 2 (a) and (c)).

The above statement suggests, that during the flash smelting of sulphide feeds, the lead fume yield and also the total carry-over depend greatly on the temperature of the particles in the area of the maximum PbS volatilization rate. If the temperature of the PbS particles and gas phase in the flame were equal (that is, if diffusion and heat transfer in the gas phase did not limit the rates of the physical-chemical processes at the particles), the overall stream of volatilized PbS calculated from expression (4) and the parameters k0 and Esp/R (see Figure

1) found experimentally, would actually fit the value of the initial PbS stream entering with the feed. Thus, observance of the set kinetic laws for the volatilization of lead compounds with no limiting factors would result in the dominant volatilization of PbS in the flame with major lead passing into the flash smelting dusts.

The discrepancy with the experimental data makes it necessary to account for the phenomena of heat and mass transfer in the gas phase which may dictate the actual change in the temperature of the surfaces of the PbS particles along the height of the flame in expression (4). For this purpose, the Frank-Kamenetsky equation (3) can be used. It is founded on the proposition of full compensation for the heat consumed for the volatilization of the condensed volatile phase by the external heat flux onto the surface of the particle:

[(P-Po)/(P-pn)](Sh/Nu)(D/a) = 1 + (M-Cp/LHTg-Ts) (6)

Where P is the total pressure of the gas phase in the system; p0 is the partial pressure of the vapours of the volatilized component far from the particle; p„ is the partial pressure of the saturated vapour of the volatilized component; Sh and Nu are the Sherwood and Nusselt numbers; D is the diffusivity of the volatilized component in the gas phase; a is a coefficient of the thermal diffusivity of the gas phase; M is the molecular mass of the volatilized component;

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 367

cp is the molar thermal capacity of the vapour; L is the latent heat of volatilization; and (Tg-Ts) is the temperature gradient between the gas phase and the condensed phase surface. Therefore, to estimate the volatilization of PbS by fitting the conditions for heat and mass transfer in the flame, (Sh/Nu)(D/a)=l can be assumed.

The mechanism determined by Equation (6) for lead sulphide volatilization was found by Sannikov (4). His calculation shows, that at temperatures lower than 900°C, the rate of PbS volatilization to the gas phase is limited by the diffusion of PbS vapours. For temperatures of the particles higher than 1,000°C, heat transfer is the limiting factor. Therefore, the temperature of the surface of the sulphide particles is much lower than the gas temperature (but does not exceed the boiling point of PbS, which is 1,281°C). This causes an obvious decrease in the lead sulphide volatilization rate in the flame. In addition, the calculations suggest that the maximum PbS volatilization rate depends on the initial particles size and that it shifts to the lower temperature zone of the flame as the particle size is decreased. For an initial size of the particles of 10 μπι or less, the PbS volatilization rate drastically drops because of the considerable increase in the velocity of the competing oxidation process. The calculated curves in Figure 3 illustrate the above point, where the specific rate of PbS volatilization given in the figure is the calculated volatilization rate related to the initial surface of the PbS particles.

900 950 1000 1050 1100 1150 1200 Surface temperature, °C

- - 300

U,U*f

1 Jd 0,03

n

CO

fi «tH

o 01

*" 0,02

a o

! O

t ο,οι m u 0) P.

n

1 1 1 1 1

r = 0,005/ 0,011

r - radius of PbS particles in mm / /

/ / J-^'m

/J^*-/Cr■ 0.005 Y

=~~~^r~~~ i - ^ ^ ^ y \ \ \

)\ f

0,03 Γ

0,05 \

>+1 0,0i

V 1 u \ f ° 0,03 Λ / -

■"^ CO \ /W c%va

■*—\ i l l *""' \ i N i

1 / / <U

Ϋ I ^ A h w f\ I\ ^

-4-i

■M o

^ 3

P,

1 \ \

-

—

-

_

1250

500

4)0

o

- 200

H I

td

H

100

1300

Figure 3 - Calculated Dependence of the PbS Specific Volatilization Rate and the Temperature Gradient of the "Gas Phase - PbS Surface" on the Surface Temperature and Particles Radius

(Calculations are Based on (4) and (6) for Autogenous Smelting)

368 LEAD-ZINC 2000

The above findings suggest that the PbS volatilization rate in the flash smelting of sulphide feeds is retarded by heat transfer through the gas phase to the surface of the PbS particles. With the observed autogenous principle establishing an unambiguous relationship between the temperature of the gas phase and the desulphurization rate of lead sulphide feeds (at fixed heat losses), the processes of heat transfer and consumption on the surface of the PbS particles become self-regulating. This allows an essential decrease in the volatilization rate of the lead compounds and eliminates the maximum flame temperature restrictions.

Thus, a speculation on the limiting effect of heat transfer on the PbS volatilization rate allows a theoretically valid explanation of the observed low rates of volatilization under the conditions of flash smelting of lead sulphide feeds.

Additional heat contributors present in the charge (carbonaceous fuel and iron sulphides) disturb the above relationship between lead sulphide oxidation and the gas temperature. Through this, the heat input to the surface of volatile lead compounds is not as limited and their volatilization rate increases. The discussed relationship explains the high production of smelting dusts in pilot-scale Kivcet units, which in the majority of cases use extra carbonaceous fuel to compensate for the great heat losses in the reaction shaft. Besides, the distinct carry-over in the pilot-scale units takes place because of the rather high mechanical carry-over of the feed constituents (see Figure 2 (c)). This is caused by the short gas ducts carrying away the gas and dusts streams from the reaction shaft.

SMELTING OF OXIDIZED FEEDS

In the smelting of oxidized lead-bearing feeds, a carbonaceous fuel, which is often powdered coal, is employed as a major heat contributor. It breaks the unity of the target chemical conversions and the heat release in the flame. Coal combustion is an auxiliary reaction providing thermal conditions for the target endothermic reactions of sulphate decomposition:

PbS04 = PbO + S02 + 0.5 02 (7)

3PbS04 + PbS = 4PbO + 4S0 2 (8)

The rates of the mentioned reactions and the coal combustion reaction, which is:

C + 02 = C02 (9)

are governed by different mechanisms. Therefore, their values change differently with the changes in temperature along the flame height. Besides, reactions (7) and (8) and reaction (9) have different signs with respect to their thermal effects and are spatially separated by the gas phase.

An interface can be imagined in a hypothetical gas suspension which consist of extremely overheated particles of fuel and "cold" particles of oxidized feed. At rather high average values of the gas suspension temperature only an illusion of flash smelting of the feed is created, since the thermal conditions are not provided for the target reactions (7)-(8) to occur. This is not possible during the oxidation of sulphide feeds. Thus, the given artificial example points to the necessity of purposefully co-ordinating the rates of heat release and the

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 369

consumption processes in the flame, during the flash smelting of oxidized feeds as they are not self-regulating in this case.

Since in the flash smelting of oxidized feeds a major heat contributor is the carbonaceous fuel combustion process which is characterized by a different mechanism of heat release than the sulphide oxidation process, the general pattern of the temperature distribution along the flame height can change fundamentally when changing from a sulphide feed to an oxidized feed. According to the mechanism described above, for the flash smelting of sulphide feeds, the temperature distribution along the flame height and a maximum PbS volatilization rate is determined by the ratio of sulphides in the charge (i.e., lead feed quality) and the oxygen consumed for smelting. In the flash smelting of oxidized feeds, the temperature distribution along the flame height is governed primarily by the kinetic characteristics of the combustion of carbonaceous fuel, and the rate of PbO volatilization (5) monotonically increases as the temperature of the particle surface increases. Therefore, it is easy to imagine the possibility of the PbO volatilization rate increasing when changing from sulphide to oxidized feed. However, this does not explain the increased total carry-over observed in the industrial-scale unit (see Figure 2 (a)). This is especially so considering the difference in the specific rates of lead oxide and lead sulphide volatilization, when no possibility for essential development of surfaces for PbO volatilization in the lead sulphate decomposition process is admitted.

At the same time, the observed dependence can be explained by the mechanism of PbS volatilization, taking into account the considerable amount of lead sulphide concentrate (~25-30% of the charge) added to the oxidized feeds treated. The discussed mechanism of lead sulphide volatilization suggests that the heat supply, with no restrictions because of the use of fuel as a major contributor of the heat, causes a considerable increase in the PbS volatilization rate. This explanation agrees with the data obtained in the pilot-scale unit (as appears on the same figure). In that case, the amount of sulphide concentrate in the oxidized feed varied in the range of 0-60% of the initial materials and the heat exchange conditions in the flame did not differ much. This was dictated by the necessity to use extra fuel to compensate for the heat losses in the reaction shaft despite the composition of the raw materials treated. As a result, the lower carry-over corresponded to the reduction of PbS in the feed.

One more peculiarity of dust formation in the flash smelting of lead feeds in the Kivcet unit should be noted. A distinct dependence on the zinc content can be seen: zinc fume in the smelting dusts increases as zinc in the feed increases (see Figure 2 (d)). The observed dependence on the correlating content of lead and zinc in the smelting dusts (Pb 28-54% and Zn 6-25%) can be explained by metallic zinc volatilization since the specific rate of ZnS volatilization is much lower than the PbO volatilization rate which itself is much less than PbS volatilization rate. It is most likely that the majority of the metallic zinc fume is formed by reduction of the high-zinc flame smelt at the layer of carbonaceous reductant (coke checker) floating on the surface of the slag bath in the reaction shaft. However, metallic zinc formation in the flame is probable during the Kivcet smelting of lead-containing residues arising from zinc production. Owing to peculiar properties of these residues to bind the mixed constituents of the feed together into stable agglomerates, a reduction process can proceed in the agglomerate microvolumes even at rather high oxidizing potentials in the surrounding bulk gas phase (5).

In general, there are many factors influencing the relationship between the rates of the exothermic and endothermic processes, that affect the carry-over in the flash smelting of oxidized lead-bearing feeds. They range from the composition and dispersivity of the feed and carbonaceous fuel to the particular methods of feed preparation and the smelting conditions. Therefore, it is rather challenging to match the rates of heat release and heat consumption in the

370 LEAD-ZINC 2000

flame to the smelting of oxidized feeds. Nevertheless, it seems possible to regulate the smelting process and total carry-over to a certain extent using solid fuel with a set fraction composition and suitable kinetic parameters for the combustion reaction (9).

Thus, the basic physical-chemical mechanisms of the volatilization of lead compounds and the factors governing the formation of flash smelting dusts, depending on the quality of the lead feeds treated and the dimensions of the Kivcet unit, are explained on the basis of generalizations and the kinetic analysis of a wide range of experimental data on the production and composition of the Kivcet smelting dust.

REFERENCES

1. A.P. Sychev, A.S. Kulenov, Yu.I. Sannikov, L.V. Slobodkin, V.A. Lysenko and Yu.A. Grinin, "An Important Step of the Kivcet-CS Process Introduction is Made at the Ust-Kamenogorsk Lead and Zinc Smelter", Non-ferrous Metals (USSR), No.l, 1988,14-19.

2. A. Perillo, "KSS Lead Smelter in Portovesme, Italy", Paper presented at the 28th CIM Annual Conference of Metallurgists. Halifax, Canada, Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, August 20-24,1989.

3. Frank-Kamenetsky, Diffusion and Heat Transfer in Chemical Kinetics. Nauka, Moscow, 1967.

4. Yu.I. Sannikov, "Some Theoretical Issues on Lead Feeds Smelting in the Kivcet-CS Unit with the Coke Checker", Non-ferrous Metals (USSR). No.5,1990,19-24.

5. Yu.I. Sannikov, M.A. Lyamina, V.A. Shumsky, Yu.A. Grinin and M.V. Radashin, "A Physical and Chemical Description of the Kivcet Process for Lead Smelting", CIM Bulletin. Vol. 91, No. 1022, 1998, 76-81.

Chapter 5

Zinc Operations II

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 373

RECENT OPERATIONS AT THE IIJIMA ZINC REFINERY

T. Yamada, S. Kuramochi and R. Togashi Iijima Zinc Refinery, Akita Zinc Co., Ltd.

217-9 Shimo-Kawabata Furumichi, Iijima, Akita 011-0911 Japan

ABSTRACT

In 1972, the Akita Zinc Co., Ltd. commenced operations with an annual capacity of 86,000 tonnes of electrolytic zinc, and it has operated with an annual capacity of 156,000 t-Zn since 1974. The business surroundings for non-ferrous metals production have become more severe with the high Yen rate for the dollar and low domestic demand in the last few years. Consequently, the company tried to survive with some improvements and succeeded to reduce operating costs and to improve profitability as follows: 1) expansion of its production capacity to. 196,000 t-Zn annually, 2) saving man-power by the introduction of fully-automated machines for cathode handling in the cellhouse, 3) improvement of the hematite process by means of the recovery of rare-metals. These improvements made Akita Zinc a competitive refinery in the modern world. We have been trying to improve the operations more and more. One of the challenges is to purify the iron oxide made in the hematite process by applying an As removal process, which is coming into operation in the near future.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

374 LEAD-ZINC 2000

INTRODUCTION

Akita Zinc Co., Ltd. was established in 1971 by six domestic companies which were producing non-ferrous metals. In 1972, the Iijima Refinery operated with an annual zinc production capacity of 78,0001, but its capacity expanded to 156,0001 since 1974.

In the last decade, the business environment for non-ferrous metals production in Japan experienced severe conditions. The Yen became stronger against the dollar very rapidly and the tariff for metals became lower step by step. These factors caused a fall in the domestic zinc price (Figure 1). In addition, domestic demand of zinc stayed below 800,000 t/y as shown in Figure 2.

kYen/t

1989 1990 1991 1992 1993 Calendar Year

1994 1995

Figure 1 - Trends in the Zinc Price and Exchange Rate

500

1989 1990 1991 1992 1993 Fiscal Year

Figure 2-Trends in Domestic Zinc Consumption

1994 1995

In recent years, these conditions became worse. In Japan, two zinc refineries have ceased to treat concentrates. Under these conditions, Akita Zinc Co., Ltd. was challenged to reduce its costs by improving each process (1). However, cost reduction by decreasing

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 375

consumable costs had been limited. Accordingly, our next challenges were the reduction of the fixed cost by expanding production capacity and increasing profitability by recovering additional revenue.

Automation is an effective way to reduce costs by means of saving labor. Especially, house keeping in the cell room is most labor intensive because of the transportation of numerous deposited cathodes and blanks with manually operated cranes. We developed a new type of automatic machine that was introduced into the operations since 1994, and the last machine was installed in 1998. We developed a new process for the recovery of rare metals in the hematite process in order to increase revenue. Production for zinc was expanded by 30,000 t/y in 1997 and by 10,000 t/y in 1999.

PROCESS DESCRIPTION

Figure 3 shows the outline of the process employed at the Iijima Zinc Refinery. Two Dorco-type roasters with a set of waste heat boilers and cyclones are operating. The waste gases from the boilers go through a hot fan to a Peabody scrubber to cool and wash the gases. The cooled gases from the two roasters are mixed and de-misted in two stages ofmist precipitators.

Figure 3 - Schematic Flowsheet

376 LEAD-ZINC 2000

Mercury is removed in a Boliden-Norzinc tower. The acid plant is of the double contact type and its production capacity is 900 t-concentrated sulfuric acid daily. A part of the SO2-containing gas is sent to a liquid SO2 plant.

In the neutral leaching stage, calcine is leached with spent electrolyte and low-iron solution containing sulfuric acid, and the terminal pH of the leach solution is about 4 after the reaction. The solution from the neutral leaching stage is sent to the purification plant which consists of three stages. In the first stage, most of copper is removed with adding zinc powder. Then, the remaining copper, cobalt and nickel are removed in the second stage by using zinc powder and arsenic trioxide. In the last stage, cadmium is removed, with zinc powder, as sponge cadmium metal with a Cd content of 90% using fluidized bed reactors supplied by Outokumpu technology. This crude metal is treated by a leaching - electrowinning process to produce high purity cadmium metal.

Fully-automatic production of Cd metal pencils is performed using a new technology with a continuous casting machine. This technology is the first application for Cd metal casting that was designed originally by Akita Zinc (2). This application has been filed in the Japanese Patent Office.

The purified zinc sulfate solution is cooled to separate gypsum by crystallization. The zinc electrowinning plant consists of 680 cells with conventional Pb-Ag anodes and Al cathode sheets with 1.67 m of deposit area. The deposition period is changed from 24 to 48 hours depending on the total electricity passage. In Japan, the power price during the daytime is very expensive in comparison with night time price (Figure 4). Therefore, we apply the lowest current density during the daytime on normal days, and the full current is applied during night hours on normal days and at all times on holidays. Consequently, the current density is ten times, or more, different between daytime and night hours. Therefore, the production of electrolytic zinc on normal days is only half of that during holidays.

Figure 4-Power Price Index

Deposited zinc sheets are stripped and are sent to melting furnaces to cast SHG slabs of 20 kg. Those slabs are bundled in about 1 ton lots, and the production data, such as the date, furnace number, lot number and weight are printed with a fully automated system.

Zinc residue from neutral leaching is treated in the SO2 leaching stage with spent electrolyte,and SO2 gas as the reductant for zinc ferrite. The Zn, Cd and Fe are dissolved, though Au, Ag, Pb and Cu remain in the residue. This residue is delivered to a copper smelter as a Cu and Pb raw material. The solution after SO2 leaching is neutralized in two stages with

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 377

CaCC>3 to remove the excess acid as gypsum, and to breed some elements. Indium and gallium are concentrated in the gypsum from the second neutralization stage. The neutralized solution is fed to autoclaves with oxygen for iron removal as hematite. The solution after iron removal, containing 70 g/1 of free acid, is recycled to leach calcine.

AUTOMATION OF THE CELLHOUSE OPERATION

The material handling operations in the cellhouse, transportation of aluminum cathode sheets with deposited zinc from the cells to the stripping machines and the blanks back to the cells, used to be the most man-power intensive work in hydrometallurgical zinc plants. Therefore, newly operating electrolysis plants have been installed with automatic cranes in the cellhouse that are based on emerging technologies. However, application of these technologies is limited to newly built plants. Then, a large investment is essential for employing these automatic systems. In Japan, the power price is expensive during the daytime as mentioned above. Consequently, we will need almost doubled capacity compared with European refineries when new technology is applied for the same production size. In general, it is difficult to achieve economic feasibility by constructing the above automatic systems in Japan.

In 1993, we started a project for developing a new system for automatic materials handling, which is applicable to a cellhouse of conventional configuration. The details were described previously by Yamada et al. (3). In addition, a monitoring system for machine operation was installed at the same time, to watch machine movements precisely. As a result, we are now able to check the machine and to improve the operating software and to carry out mechanical problem shooting.

The first automated system was installed in 1994. This was effective for reducing the number of operators for the manual cranes. The whole system was completed in 1998 and the system has functioned without operators, as shown in Table 1.

Table 1 - Operators for the Cellhouse (persons/shift)

Hoist crane Stripping ^ machine

manual crane 4 4 8 automatic system 0 4 4

Saving man-power in the cellhouse brought our refinery to one of the highest productivities in the world (Figure 5).

378 LEAD-ZINC 2000

1985 1988 1991 1994

Fiscal Year

1997 2000

Figure 5 - Productivity in the Cellhouse

EXPANSION

We have maintained a zinc production capacity of 156,000 t/y since 1974. On the other hand, the Japanese zinc market needed to import of about 100,000 t of zinc annually because of the short domestic supply. Expansion of the plant capacity had been repeatedly examined since the early 1990's. During that time, the demand for zinc in Japan has been falling slightly, and two Japanese refineries stopped production. In 1996, we decided to start an expansion project.

Concept of Expansion

The target point of the capacity expansion was set at 30,000 t-zinc annually, with minimum investment. At first, we checked bottle necks at each process, when all the equipment was in full action. Then we examined measures to increase the operation of every piece of equipment up to the maximum. Mainly, gas handling in the roasting process, iron removal capacity in the hematite process, rectifiers for zinc electrolysis, cooling capacity for the electrolyte, and a transformer at the main transforming station, were considered.

Gas Handling System in the Roasting Process

When an expanded production is considered, we have to treat more concentrate in the roasters. This leads to the generation of more process gas. From the view point of acid production capacity, the bottle necks were 1) insufficient capacity of the gas handling equipment, such as the air blowers, waste gas fans and gas blowers in the roasting process, 2) imbalance of heat in the roaster caused by increasing feed.

At first, excess 02 for roasting was reduced to prevent the proportional increase of the gas volume corresponding to the increase of concentrate feed to the roaster. The details were

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 379

reported by Kuramochi et al. (4). Simultaneously, the cooling system in the roaster bed was reinforced in order to lower the water injection volume for controlling the temperature of the roasting bed. Because the injected water to the roaster was converted to steam, it caused the process gas volume to become larger.

The reinforcement of the cooling system resulted in not only a reduced gas volume but also a lightening of the load for the gas cooling equipment after the waste heat boiler. Figure 6 shows the trends in the feed rate of concentrate and the amount of the cooling water injected. The feed to the roaster was going up according to the expansion of zinc production; however, the requirement for water injected for cooling the fluidized bed was decreased by use of the additional cooling system.

Figure 6 - Trends in Roaster Feed and Injected Water

As a result of these improvements, we could minimize the investment for the reinforcement of blowers, fans just for gas handling in the roasting process and the sulfuric acid plant. In addition, the operating power cost was kept the same as before.

Iron Removal

The zinc concentrate contains iron as the main impurity. It is necessary to increase the capacity for iron removal according to the expansion of zinc production. We installed an additional autoclave of the vertical type in one of the three existing hematite lines. As a result, the capacity for iron removal increased from 2,303 t/month to 2,650 t/month. The details about the improvements concerning the hematite process were presented at the Zinc and Lead Processing Conference in 1998 (5).

Electrolysis

Production capacity depends on the current passage during operation. According to the Japanese system for power prices, we have to keep the lowest current during daytime and to operate at full capacity during the night hours. Before the expansion, we already reached the maximum utilization of night hours at the full capacity of the rectifiers. Accordingly, we had to reinforce the rectifier capacity of five lines out of seven (Table II).

380 LEAD-ZINC 2000

Table II - Capacity Changes at Each Rectifier

Year

1996 1997 1999

1 45kA 45kA 45kA

2 31kA 45kA 45kA

Circuit Number 3 4 5

31kA 31kA 61kA 31kA 45kA 61kA 48kA 45kA 61kA

6 45kA 58kA 58kA

7 45kA 58kA 58kA

Bold-italic figures: reinforced rectifiers

The production capacity was expanded by 1.25 times. The more the capacity, the higher the current density is without enlargement of the effective area for deposition. Also heat generation during electrolysis increases. Therefore, in order to maintain the operating temperature, two cooling towers of a new type were installed and the existing cooling system was improved. Furthermore, two additional cathode sheets and two anodes were placed in each cell by reducing the pitch of the electrodes. This enabled us to reduce the cathode current density by increasing the deposition area.

Summary of the Expansion

We expanded the zinc production capacity from 156,000 t/y to 196,000 t/y in two steps. In the first step (1997), the capacity expansion gave rise to a production increment of 30,000 t/y. After the second step of an additional 10,000 t/y arising from technical improvements, our production capacity reached almost 200,000 t/y.

The operating cost decreased as a result of the reduction of the fixed cost, and productivity was improved as shown in Figure 7 and Figure 8, respectively.

Total investment for the 40,000 t/y expansion was 30 million U. S. dollars, which is equivalent to 750 $/t. This may be one of the most economical expansion projects in the world.

Figure 7-Trends in Operating Costs

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 381

t/man-year kt/y

1985 1987 1989 1991 1993 1995 1997 1999 Fiscal Year

Figure 8 - Trends in Total Productivity and Zinc Production

HEMATITE PROCESS IMPROVEMENTS

The Iijima Refinery is the only refinery in the world which applies the hematite process for iron removal. This process is excellent for recovering Au, Ag, Cu, Pb and rare metals, and for the precipitation of iron oxide usable in cement production, because of its high iron content. Furthermore, we are trying to develop a new process to recover more rare metals and to purify the iron oxide to broaden its application base.

Rare Metals Recovery

Domestic concentrate produced from Kuroko-type sulfide ore contains much gallium and indium. Until the early 1980's, many domestic mines operated and the Iijima Zinc Refinery treated domestic concentrates as 50% of the total charge, as show in Figure 9.

Figure 9- Ratio of Domestic Concentrate Treated

382 LEAD-ZINC 2000

The content of rare metals was relatively high and these were concentrated in a by-product at a sufficient level for recovery. In the middle 1980's, Japanese mines were closed after a rapid increase of the Yen rate against the U.S. dollar and relatively low LME price of non-ferrous metals. At that time, we had to increase imported concentrates to maintain operations. Therefore, the concentration of rare metals in the gypsum fell rapidly, and recovery costs overtook income and lead to the stoppage of operations. The recovery cost of the previous process was relatively high because of the use of expensive chemicals and complicated procedures.

However, we were developing new procedures in order to produce rare metals at an adequately low cost, such that it would be feasible to recover them at a low content. In 1994, we developed a new hydrometallurgical process for recovering indium without using any solvent and/or chealating reagent. The schematic flowsheet is shown in Figure 10. At the same time, one of stock holders of our company stopped its own refinery, and wanted to bring their domestic concentrates to the Iijima Refinery. This concentrate is rich in indium; therefore, the concentration of indium in the raw material was expected to increase. Under these circumstances, we decided to build a new plant for recovering indium at the Iijima Refinery.

Gypsum

Leaching, Purification & Precipitation

I Leaching,

Purification & Cementation

I Leaching,

Purification

Figure 10-Schematic Flow Diagram for Indium Recovery

The plant was commissioned in November 1995, and production started in January 1996. Indium metal refined in this plant is pure at a level of 5N, and can be utilized for ITO (indium tin oxide) target material. A typical analysis is shown in Table III.

Table III - Typical Analysis of the Final Indium Product

element In* Ni Sn Pb Ge Cd Cu content >99.999 <2 <1 <1 <1 <1 ND

element Fe Tl Al Sb Zn As Bi content ND ND ND ND ND ND ND

ND; not detected, * content is shown as ppm except In (%)

Electrowinning

Melting

T Casting

Indium ingot (5N,lkg)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 383

Recently, we developed a new process for recovering gallium and a commercial plant was commissioned. Production will start in the third quarter of the 2000 fiscal year.

Iron Purification

Arsenic in iron oxide is one of the major limitations for applying this material in some industries. Since the hematite process started, we have examined many procedures to remove arsenic in the iron oxide production process. In 1998, we developed one of the best processes to purify iron oxide and reported the results (5). After that, we examined in more detail ways to reduce the investment and operating costs. Now, the process is in the final stage of commissioning.

CONCLUSIONS

The Iijima Zinc Refinery is being challenged to become more profitable by recovering all by-products, such as Au, Ag, Cu, Pb, rare metals, iron oxide and so on. An increase in zinc production caused an increase in these metals as well. Simultaneously, the fixed cost decreased significantly with an expansion and cellhouse automation. We achieved one of the most competitive costs in the world.

It remains essential to reduce the quantity of disposed materials from smelters and refineries in the near future. In addition, the primary zinc business must recover valuable metals as well as zinc itself. Recovering such metals leads the way to reduce the disposal of the associated elements. Therefore, we will accept the challenge to develop new processes and to utilize automated systems further.

ACKNOWLEDGEMENTS

The authors greatly thank the Akita Zinc Co., Ltd. and Dowa Mining Co., Ltd. for granting permission to publish this paper. The authors would also like to express their appreciation to Dr. Y. Umetsu, of Tohoku University, for his advice in preparing the manuscript.

REFERENCES

1. T. Yamada, T. Nagata, A. Hosoi, M. Kato and R. Togashi, "Recent Improvement of Electrolytic Zinc Production at Iijima Refinery," Zinc & Lead 95, T. Azakami, N. Masuko, J. E. Dutrizac and E. Ozberk, Eds., The Mining and Materials Processing Institute of Japan, Tokyo, Japan, 1995, 589-598.

2. T. Murakami and S. Sudou, "Development of Cadmium Continuous Casting Machine," Journal of the Mining and Materials Processing Institute of Japan. Vol. 115, No. 5, 1999, 354-357.

3. T. Yamada, R. Togashi and T. Aichi, "Development of Automatic Material-Handling and Monitoring Systems in an Existing Electrolytic Zinc Plant," Aqueous Electrotechnologies. D. Dreisinger, Ed., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1997, 89-100.

384 LEAD-ZINC 2000

4. S. Kuramochi, K. Kusakai, Y. Ishiyama and Y. Kai, "Increasing of Roasting Capacity in Iijima Refinery," Journal of Mining and Materials Processing Institute of Japan. Vol. 114, No. 5, 1998,309-312.

5. T. Yamada, S. Kuramochi, S. Sato and Y. Shibachi, "The Recent Operation of the Hematite Process at the Iijima Zinc Refinery," Zinc and Lead Processing. J. E. Dutrizac, J. A. Gonzalez, G L. Bolton and P. Hancock, Eds., The Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 627-638.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 385

IMPROVEMENTS AT THE ELECTROLYTIC ZINC PLANT OF BIG RIVER ZINC CORPORATION, SAUGET, ILLINOIS, U.S.A.

S.E. James, L.L. Ludwig and B.L. Motycka Big River Zinc Corp

2401 Mississippi Avenue Sauget, Illinois, U.S.A. 62201

ABSTRACT

Significant modifications to the electrolytic zinc plant operated by Big River Zinc Corporation were made at the end of the 1990's resulting in increased capacity and improved efficiencies. A new cell room, with the innovative use of acid-resistant concrete cells and a modern ventilation system, was installed to raise the capacity from 80,000 tonne/y to 106,000 tonne/y of finished zinc. Improvements to the leaching section include the conversion from a single to a two-step process, with associated equipment changes, to improve the quality of the residues produced. Another key change is the construction of a facility to allow the use of zinc oxides from secondary sources as a feed to the plant. This paper describes these improvements and the current operations.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

386 LEAD-ZINC 2000

INTRODUCTION

Big River Zinc Corporation (BRZ) operates a zinc refinery located in Sauget, Illinois, U.S.A.. Its central location gives it ready access to a significant portion of the market for zinc in the United States. Because of the proximity to markets and to domestic mines, zinc has been produced at the plant site since 1929 when the Evans-Wallower Zinc Company built and operated a demonstration electrolytic plant. Because of technical problems and deteriorating business conditions, Evans-Wallower closed the plant in 1931. The American Zinc Company purchased and reopened the plant with the support of the US government in 1941. American Zinc expanded the plant capacity several times during its tenure as owner and eventually achieved an annual production rate of 73,000 tonnes (1). In 1971 American Zinc closed the plant. A year later, AMAX acquired the facility with a plan for rehabilitating and modernizing the facilities. This decision allowed AMAX to close its old horizontal retort plant in Oklahoma and to process concentrates from its mines in southeast Missouri. During the early part of the 1970's, AMAX invested over $35 million to upgrade the Sauget plant. In 1988, AMAX exited the zinc business and sold the plant to a private investment group known as the Big River Minerals Corporation. Big River focused on improvements to increase the capacity of the plant and finally achieved a production rate of 80,000 tonnes per year of finished zinc.

In April 1996, Korea Zinc acquired the zinc assets of Big River Minerals but retained the name of its new US subsidiary as the Big River Zinc Corporation. Korea Zinc saw in the Sauget refinery an opportunity for growth for the following reasons:

• The United States has a large and growing demand for zinc, but depends heavily on foreign imports.

• For a period of 15 years, the plant had received very little capital investment by its owners and offered the potential for significant improvements in productivity with modest capital investment.

• Even with the limited investments, the zinc plant had shown its ability to remain profitable during times of low zinc prices when some of its competitors lost money.

Members of the BRZ and Korea Zinc staffs soon developed an improvement program to once again expand and update the Sauget zinc refinery. This program consisted of two phases: operating the existing equipment more efficiently to lower costs, improve product quality, and minimize environmental exposure; then expanding production to a rate of 106,000 tonnes per year.

PLANT IMPROVMENTS

Phase I: Operating Efficiency and Quality Improvements

Much of BRZ's zinc is sold for the continuous galvanizing of steel sheet. During the early 1990's, the steel industry in the Untied States began emphasizing the need for higher and more consistent quality in its operations and from its suppliers. BRZ found that the increased demands for establishing formal quality and improvements programs also offered the opportunity to examine its operations for ways to reduce waste and lower costs. BRZ hired a full-time quality manager and made the commitment to achieve ISO-9002 certification for its finished zinc products. Initial quality programs focused on the melting and casting areas of the

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 387

plant. However, after seeing the positive effects of these programs, they were extended to sulfuric acid production and into the leaching and purification sections of the plant. BRZ obtained certification for its zinc metal and alloys in 1998 and then for its sulfuric acid in 1999.

BRZ established a continuous improvement program whereby goals are set at the beginning of each year for each major operating area. These goals may include cost savings, productivity increases, improved safety and housekeeping, or research and development. The goals are reported and carefully reviewed by management every calendar quarter. Each employee receives a copy of a report showing the status of all the goals. In 1999, BRZ began a formal employee suggestion program, whereby employees, either singly or in groups, can submit ideas that focus on cutting costs or improving the safety of the workplace. The focus of this program is the identification of changes in practice or equipment that require little investment but yield immediate dividends. A joint committee of management and hourly employees carefully considers each suggestion. The implementation of ideas approved by the committee receives high priority.

Critical reviews of the staffing have resulted in the consolidation of job responsibilities. Changes have been made to minimize the number of salaried staff.

Phase 2: Expanding the Production Capacity

When the improvement team began studying alternatives for expanding the capacity of the Sauget refinery, it focused on maximizing the benefits of the existing plant while raising production to the most cost effective level. An initial evaluation concluded that the leaching and purification sections of the plant needed only minor changes as long as the annual capacity remained below 110,000 tonnes. The timing of the expansion came shortly after the Zinc Corporation of America decided to permanently close its zinc refinery in Bartlesville, Oklahoma. This allowed BRZ access to numerous pieces of useful equipment from the Bartlesville site, such as rectifiers and transformers. Extensive application of this used equipment kept the overall costs low.

Roasting and Sulfuric Acid Production

To support the increased zinc production, BRZ had to increase the utilization of its roasters. The roasting and acid plants were built in 1966 and no major changes had been made to the facilities since that time. BRZ engaged Dr. Gordon Cameron of Cecebe-Noram to analyze its acid plant and to develop options for increasing production. The roaster feed rate had increased steadily over the years but faced limits every summer when higher cooling water temperatures hindered heat removal in the acid plant. In the original design, a Peabody scrubber combined both humidifying and cooling of the gas in a single vessel to minimize construction costs. To increase cooling, BRZ installed a gas-quenching vessel ahead of the existing Peabody scrubber. Figure 1 shows this installation in relation to the rest of the sulfuric acid plant equipment.

388 LEAD-ZINC 2000

Existing MistPrecipitatofS D ( l u t i o n

Cross-Flow Stripper

Heat Exchangers

Figure 1 - The Revised Sulfuric Acid Plant for Big River Zinc

Hot gases from the electrostatic precipitators in the roasting plant enter the quencher at 300-310°C. Weak acid solution sprayed into the Venturi section of the quencher cools the gases to 60°C. The existing Peabody scrubber now functions as a two-stage gas cooler that lowers the gas temperature to less than 35°C. To help cool this solution, BRZ added a new plate and frame heat exchanger fitted with Hastelloy C276 plates and upgraded the capacity of the pumps on the water cooling towers. These changes increased heat removal by 10 million BTU per hour. Additional heat removal was required to cool the hot gases exiting the third pass of the converter. A long section of uninsulated steel duct, called the "long pass," was installed to provide more radiant cooling of the hot SO3 gas.

To boost the conversion of SO2 to SO3 in the acid plant, additional catalyst was put into three of the four catalyst beds of the converter. In essence, all of the space in the catalyst beds has now been filled, and the acid plant has reached its limit for converting SO2 into sulfuric acid. Further improvements will require optimizing the operation of the converter, and eventually, rebuilding the acid plant.

These modifications had the desired effect of raising production, especially during the summer months. Figure 2 shows how acid production has steadily increased since the modifications to the acid plant were made in June 1997. The three dips in the curve reflect the annual maintenance outages that BRZ conducts in its roasters and acid plant. Even taking these periods into account, the average acid production has increased by about 10%. Because of the additional gas cooling available, production no longer drops sharply during the hot summer months. Since the gas quencher went into operation, the plate heat exchanger was replaced with a unit manufactured by Alfa-Laval. The design of the original plates allowed solids to become trapped in the channels, an occurrence which accelerated corrosion. Two plate heat exchangers from· Alfa-Laval had been installed for cooling weak acid from the Peabody scrubber about 15 years ago and continue to perform well.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 389

Figure 2 - Trends in Sulfuric Acid Production at Big River Zinc

Figure 3 shows the trend in the total feed rate to the zinc roasters over the same period of time. Once again, the dips reflect the scheduled maintenance outages. A similar 10% increase occurred since the acid plant improvements were made in 1997. Even more encouraging is the increased feed rate and acid production that occurred after the maintenance outage in September 1999. These increases reflect the employees' ability to operate the equipment at the maximum rates.

Figure 3 - Trends in the Roaster Feed Rate at Big River Zinc

390 LEAD-ZINC 2000

Use of Zinc Secondaries

To further increase the capacity of the roasters, BRZ evaluated the use of secondary zinc oxides in the roaster feed. The use of crude zinc oxide as a roaster feed offers a number of advantages compared to rebuilding the acid plant. The oxides bring in additional zinc units without more sulfur units; consequently, then do not require more converter capacity. In contrast to zinc sulfide concentrates, the zinc oxide absorbs heat in the roaster. This helps to maintain the desired roaster bed temperatures without injecting more water. Water injected to the roaster beds increases the heat removal burden of the acid plant where it must be condensed. BRZ became actively involved with several companies operating plants that treated electric arc furnace (EAF) dust from steel mills. The first company generated a crude zinc oxide that it washed with water to remove the majority of the chlorides and fluorides. When this company ceased operations, BRZ negotiated agreements to buy crude zinc oxide from three other companies. Because none of these companies operate a washing plant, BRZ built a washing plant at Sauget with a capacity of 18,000 tonnes per year of crude zinc oxide. This facility began operating in May 1999. It is described in detail in another paper presented at this meeting.

Leaching Modifications

The Sauget plant originally employed a batch leaching process. During American Zinc's tenure, the leaching circuit was converted to a single-stage continuous process. The benefits of this simple approach were offset by zinc losses into the final residue. The improvement team concluded that the addition of a second leaching stage would permit as much as a 1.5% higher zinc recovery for the same type of calcine used with the existing single leaching stage. Other advantages would include better precipitation of impurities such as arsenic and germanium in the neutral leaching step. Operation of the single neutral leach always involved making compromises between achieving a high zinc extraction and controlling the precipitation of these impurities.

BRZ installed two new 25 m3 tanks that operate as a continuous weak acid leaching step on the thickener underflow from neutral leaching. Their inclusion in the leaching flowsheet appears in Figure 4. The entire calcine continues to go to the first neutral leaching tank because the team decided not to move the present calcine handling system. A portion of the neutral leaching thickener underflow is directed into the second weak acid leaching tank to control its final pH. The weak acid leach slurry goes to a rebuilt 15-m diameter thickener. This thickener has been built from concrete with a cast-in-place plastic liner. The weak acid thickener overflow is pumped to the copper recovery circuit. Because of the high concentration of copper in this solution, the improvement team decided to move the copper precipitation step from the impure neutral solution to this location. The copper recovery stage was modified from three reactor tanks to two tanks and a settling vessel. Zinc dust added to the first tank precipitates metallic copper cement. It is recovered from the settling vessel underflow in recessed plate filter presses.

The underflow from the weak acid leach thickener is pumped into the feed launder of an existing washing thickener. Filtrate and used residue washing water mix with the underflow to dilute its zinc content before the residue slurry flows into "aging" tanks. The residue slurry remains in these tanks for as long as 24 hours before filtration on Eimco membrane filter presses. During the "aging" step, some of the cadmium in the residue becomes soluble and is easier to wash. The dilute zinc stream in the thickener overflow is pumped back to the first of the weak acid leaching tanks. When the new leaching circuit was first built, the thickener

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 391

overflow went directly into the weak acid leach thickener. Redirecting its flow to the weak acid leach has allowed better control of the pH and impurities throughout the circuit.

Operation of the new circuit involved a considerable amount of experimentation to find the best combination of operating conditions. The present scheme controls the pH in the first tank of neutral leaching by monitoring the pH in the third tank and by measuring the concentrations of copper and germanium in the solution leaving neutral leaching. Copper concentrations are set based on the level of cobalt in the solution leaving neutral leaching. Because the zinc concentrates from the southeast Missouri region of the United States have very high contents of cobalt, the neutral solution can contain up to 80 mg/1 of cobalt.

The pH in the weak acid leach circuit is controlled based on the slurry leaving the second tank. This level is varied depending upon the load of impurities such as arsenic and germanium. Raising the pH forces these elements into the lead/silver concentrate and prevents them from building to unacceptably high levels in the circuit. Lowering the pH allows a larger proportion of these elements to recycle to neutral leaching, but increases the extraction of zinc, cadmium and copper.

Gypsum Removal

BRZ's calcine contains a relatively high level of calcium. Therefore, the improvement team saw a significant potential for labor savings if the gypsum that formed throughout the leaching, purification and cell room circuits could be collected in just one location. With the help of engineers from Korea Zinc's Onsan plant and from Cominco's plant in Trail, the team designed a gypsum removal circuit that made maximum use of the existing equipment. Purified solution flows into two 411-m3 outdoor storage tanks, then to a mixing tank. From the mixing tank, it is pumped across two Hamon-Sobelco cooling towers that had originally been installed at the now closed Bartlesville zinc plant. The cooled purified solution falls by gravity into a plastic-lined concrete thickener. Following the practice first established by Cominco, the thickener underflow is pumped to the mixing tank where it serves as seed for precipitating gypsum. Periodically, the employees divert a portion of the thickener underflow to a tank ahead of an automated filter press. A small amount of cell acid is added to this tank to prevent the precipitation of zinc salts. The resulting filtered gypsum is white and contains very few impurities.

BRZ started the gypsum removal process in November 1999. Almost immediately, employees in the cell room saw a dramatic decrease in the amount of scale formation on the launders, pipes and cooling towers. Before installing the gypsum removal process, we had to stop the electrolyte flow into the cell room every 4 to 6 weeks to manually remove scale from the distribution launders. Now, the launders have been cleaned only once after approximately 6 months of continuous service, and we have confidence that we can continue on a similar schedule in the future. The rate at which the accretions grew in the electrolyte cooling towers also fell dramatically. Now, the time to clean a tower is only about one-third of that required before the gypsum removal operation began.

392 LEAD-ZINC 2000

_ P E n j P i ; iCir.

Γ · -U—1

| ==: J

-'-,1

COPPER CAKE

STORAGE

L F - C H I I I G CIRCUIT

C C P F E F F E O . C . E F ,

ClFC'- ' lT

-¥"-

_ T V |

PI) AS ,1 , :|ii '!,: CONCENTRATE I 1 i l S l l ^ j FILTER PRES2E?

* T t

LEAD SILVER CONCENTRATE CAR

Figure 4 - The Modified Leaching Flowsheet at Big River Zinc

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 393

Installation of a New Cell Room

The Sauget refinery had two existing cell room units, each arranged with 192 cells in electrical series. The cells were grouped into "banks" of 12 cells, with each bank fed by a single distribution launder. A diverse collection of transformers and rectifiers provided power to each unit with a limit of approximately 26,000 DC amperes per unit. Work over the years had shown that the cathode size could be increased to a maximum of 1.5 m2 of plating area, a size which could still be handled by employees engaged in manual stripping. Increasing the production of the cell room would require adding more cells and rectifiers. Because of the arrangement of the existing cell rooms, no more cells could be added to the existing units. Because the current density in the existing cell rooms had already been raised to 625 A/m2, the improvement team concluded that adding more rectifiers to the units would not be feasible.

The team eventually designed a third cell room unit with 120 more cells. To ensure compatibility with the existing cathodes and anodes, we retained the same size of electrodes and continued to use manual stripping of the zinc deposits. However, we increased the number of cathodes in each cell from 27 to 30. To reduce the overall construction cost, the team designed large unicell sections cast from polymer concrete. Because of the excellent corrosion resistance, high strength, high thermal shock resistance, and excellent electrical resistance of the polymer concrete, installation of the cells was greatly simplified. These unicells require no internal PVC liners, and this results in reduced maintenance costs during normal operation. To assure ourselves that polymer cells would function reliably, we installed four prototype unicells in one of the old cell rooms in March 1997. These cells have been in continuous service ever since without presenting any problems whatsoever.

The new unicells have a monolithic construction and are 7.9 m x 1.1 m x 1.1 m. They were cast at Ancor-CTI in Green Bay, Wisconsin and were shipped to the Sauget refinery for installation. Each unicell has two internal dividers so it acts like three separate cells. Two unicells are arranged lengthwise and are connected with a small plastic overflow trough. This provides a layout of six cells. Four unicells are grouped together with a common electrolyte feed launder to form a "bank" of 12 cells. Like the old cell room units, a fiberglass reinforced plastic distribution launder feeds each compartment in the cells. The resulting arrangement is a modified cascade where each of the successive 6 cells receives all of the flow from the preceding cells plus some fresh electrolyte. Copper bus bars for the electrodes are anchored to one side of the unicells and have polymer concrete insulating supports. Figure 5 shows a view of two adjacent unicells during construction. The internal dividers that split the unicell into three electrical cells are clearly visible. The support ledge for the copper bus bars on another unicell is visible on the left side of the photograph.

394 LEAD-ZINC 2000

Figure 5 - Polymer Concrete Unicells Used at Big River Zinc

Figure 6 shows a view of the completed cell room. A key feature of the new cell room is that it was designed with a proprietary ventilation system developed by Desom Environmental Systems Ltd. The system uses a combined push-pull action to draw acid mist from the cells out of the cell room building. A specially designed plenum located on the roof of the building supplies forced fresh air that sweeps across the cathode stripping area. During the cold winter months, natural gas heaters temper the air before it enters the cell room. A supplemental fan blows air through specially designed distributors and out across the top of the electrodes in the cells. The air flows across the length of the cells then exits through a set of seven wall fans. This design traps the cell acid mist near the top of the electrodes and sweeps it out of the building. Acid mist from the cells stays well below the breathing zone of the employees working in the cell room. Tests conducted in the cell room show that the system keeps the concentration of acid mist for the workers well below the present regulations. Since the cell room began operating in December 1997 the system has worked flawlessly. The only problems have involved cracking of the cell top air distributors, when employees step on them, if they are not properly supported.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 395

Figure 6 - View of the New Cell Room Unit at Big River Zinc

BRZ and Desom worked together to develop equipment to capture the acid mist leaving the building (2). All of the wall exhaust fans have been specially designed with mist collection bags. These are polypropylene fabric bags attached to the fan exhaust outside the building. Impingement of the mist onto the fabric prevents the acid mist from escaping into the environment. The operation of these mist collectors has proved exceptionally reliable. Only a few minor problems related to the stitching of the fabric have been observed. By changing the sewing pattern and reinforcing the fabric these problems have been eliminated. These units require periodic washing with water to keep the fabric clean. An employee can quickly wash the mist eliminators by spraying water from a hose on the outside of the units. An automatic internal spray system has been successfully tested and will eventually be added to all of the mist eliminators.

Power is supplied by two ASEA tap-changing transformer-rectifiers obtained from the closed Bartlesville zinc plant. When the third cell room unit went into operation, we lowered the current in the other two cell room units for the first year until the other expansion projects had been completed. The cell room current is now adjusted to match the cathode production targets for each month and to take full advantage of different electric power costs between day and night time operation. At full production rates, the current density averages 560 - 570 A/m2

over each 24-hour period. The production trend since starting and operating the new cell room is shown in Figure 7.

396 LEAD-ZINC 2000

320 I

Λ 300 .

I 280 ^ I 260 jr....:....-."

-^7z***?^ss^. -^^^0fr.......

| 240 /

o 220 /

200 '

180 J C O O O C O O O C D C D C D C D < 0 C D Qj C D QJ C D CD C D OJ C D QJ C D QJ CD O> CD CD C D QJ

Figure 7 - Trend for Big River Zinc Cathode Production

Casting Department Upgrade

To handle the increased cathode production, BRZ evaluated its casting department and decided to initiate three projects. The first involved boosting the cathode melting capacity. BRZ operated a single melting furnace with four Ajax Magnothermics 300 kVA inductors. The improvement team located suitable used inductors at the closed Bartlesville plant and installed four 450 kVA inductors on the existing furnace. A new electrical substation was installed to handle the higher power demand of these new inductors. Examination of the cathode conveying and charging equipment showed them capable of handling the increased production. In December 1997, all cathode melting ceased for 10 days to permit installation of the new inductors. During this outage, two 72 kVA inductors on the small furnaces that receive the molten zinc were also changed. This upgrade went smoothly, and the melting rates the following month were high enough to handle the cathode production from a fully expanded plant.

The second casting project involved installation of a new alloying furnace and molds for casting ingots. Sales of continuous galvanizing grade (CGG) alloys for coating steel sheet have become a major portion of BRZ's business. Prior to 1998, BRZ made CGG alloys in 2.5 ton batches. An employee manually added aluminum and stirred the molten metal in a small ladle before pouring the metal into a mold. The use of small batches contributed to an excessive variation in the aluminum content of the alloy and resulted in inefficient use of labor. To address these deficiencies, BRZ installed a 25-ton furnace for making alloys. The furnace is fed with molten metal from the main casting holding furnace through a new launder system. An electric mixer can be lowered into the furnace to create a circulation that incorporates aluminum into the molten metal and ensures a uniform composition. The alloying procedure consists of filling the furnace, adding a weighed quantity of aluminum to the molten zinc, mixing for a prescribed time, and taking a sample for analysis. Samples are checked using a LECO glow discharge spectrometer installed in the casting department. When a sample meets the specification for the particular alloy being made, the metal is pumped into stationary water-cooled molds. A set of 4800 lb (2180 kg) molds extend in a single row to one side of the alloy

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 397

furnace with a crane running overhead for pulling the cooled ingots out of the molds. Water is continuously circulated throughout the molds and over a small atmospheric cooling tower.

The alloy furnace went into service in June 1997 and the consistency of the alloys improved immediately. During its initial installation the furnace had 6 stationary molds for 2400 lb (1090 kg) strip jumbo ingots in addition to the 4800 lb ingots. Sales of this size of CGG alloy began to increase, and a second line of stationary molds was laid out parallel to the first and containing 20 strip jumbo molds. Construction began in late 1997 and the new set of strip jumbo molds became operational in April 1998. With these changes, BRZ can efficiently cast a mixture of as much as 1500 tons per month of strip jumbos and 2000 tons per month of the larger ingots.

The third project involved replacing the equipment in the old casting dross plant. A new air swept ball mill designed by Korea Zinc was installed in place of the vertical ring mill. Dross from the main melting furnace is taken to the dross treatment plant and is put through the ball mill to separate particles of zinc metal from zinc oxide. The metal is remelted to make Prime Western or other low-grade alloys, whereas the zinc oxide is fed to the roasters.

CONCLUSIONS

The improvements made over the period 1997 to 1999 have allowed BRZ to significantly increase its zinc production capacity. Additional personnel were required only for manually stripping the cathodes in the new cell room and for ingot casting. To increase its productivity, BRZ has begun actively evaluating mobile stripping machines that could fit into the existing cell rooms without adding an elaborate and expensive cathode conveying system. During the first quarter of 2000, BRZ demonstrated that it could produce finished zinc at a rate of 111,000 tons (101,000 tonnes) per year, which is equivalent to 96% of our final goal. As more zinc oxide from secondary sources becomes available, BRZ will reach its goal of full production.

ACKNOWLEDGEMENTS

The authors would like to acknowledge the hard work of all of the BRZ employees who contributed to the success of the plant expansion projects. Special appreciation is extended to to Mr. K. W. Jeon of Korea Zinc who served as the on-site liaison between the two companies for over two years.

REFERENCES

1. O.H. Bahne, R.K. Carpenter and C.R. Paden, "Electrolytic Zinc Plant of American Zinc Company," AIME World Symposium on Mining & Metallurgy of Lead & Zinc - Vol. II Extractive Metallurgy of Lead & Zinc. C.H. Cotterill and J.M. Cigan, Eds., The Metallurgical Society of AIME, New York, NY, U.S.A., 1970, 308-328.

398 LEAD-ZINC 2000

2. W. Spitzmiller, and J. de Visser, "Reduction of Sulfuric Acid Emissions from Electrowinning Tankhouses," Paper presented at the Zinc Processors Group Conference, Trail, B.C., Canada, October 4-5,1999.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 399

RECENT PROCESS IMPROVEMENTS IN THE KOKKOLA ZINC ROASTER

J. Nyberg Outokumpu Zinc Oy

P.O. Box 26 FIN-67101 Kokkola, Finland

M.-L. Metsärinta and A. Roine Outokumpu Research Oy P.O. Box 60, Kuparitie 5 FIN-28101 Pori, Finland

ABSTRACT

The Kokkola roasting plant uses more than 20 different zinc concentrates every year. During recent years there has been an increase in the variation in concentrate quality; in particular, the amount of fine-grained and low-grade concentrates has increased. Therefore, a great deal of development work has been carried out to improve the operational flexibility, on-line availability, production capacity and calcine quality of the roasting plant. This work consisted of a heat and material balance model of the roasting based on HSC Chemistry software, a new water admixture system for the concentrate feed, a new oxygen feed system for the process air line, improved cooling of the furnace and boiler and improved operational practices, dust removal and concentrate mixing. These new process control tools and other improvements, as well as some test results, will be described in this paper. The oxygen feed may be used to improve production capacity and to decrease the sulfide content of the calcine. The water feed may be used with fine concentrates to decrease both dust carry-over and the temperature in the top part of the furnace. Improved cooling of the furnace and boiler helps to increase production capacity.

Lead-Zinc 2000 Edited by J.E. Dutrizac, JA. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

400 LEAD-ZINC 2000

INTRODUCTION

Outokumpu Zinc Oy produces zinc at the Kokkola plant using the conventional roasting -leaching - solution purification - electrowinning process. The plant came on the stream in 1969. Since start-up, the process has been further developed by Outokumpu. The design capacity of the plant was 70,000 tonnes of zinc per year, but by 1974, the throughput had been increased to 90,000 tonnes per year by improving the design criteria. In 1974 an expansion to 150,000 tonnes of zinc per year was completed and this has been followed and characterized by continuous improvement. The focus has been on increasing throughput and productivity by removing bottlenecks and improving both the process itself and process control.

The capacity was 170,000 tonnes of zinc per year after the 1988 modernisation of the plant. Today the current roasting capacity has been increased to about 190,000 tonnes, see Figure 1. After the 1998 expansion, based on the direct leaching of zinc concentrates, the new total capacity of the Kokkola plant is about 225,000 tonnes of SHG zinc (99.995 %) per year. The latest on-going expansion will further increase both the leaching and roasting capacity to 260,000 tonnes of zinc per year.

."2 u Λ

•g a 3 CO

U

1* kH

s o U

Ü ω CO

u o

a u 3 p

400000

350000

300000

250000

200000

150000

100000

50000

0

1965 1970 1975 1980 1985 1990 1995

Years

100

90

80

70

60

50 -g Ji

40

30

20

.1 ■ * - *

00

Ut O

0) _> 8 e w

S u _N '55 tu

"u

1 B.

10

2000

0

2005

Figure 1 - The Effective Working Time, Amount of Roasted Concentrate, Sulfuric Acid Production, Concentrate Grain Size and Amount of Leached Concentrate in the Kokkola Plant.

(Note: the annual shutdown is included in the effective working time.)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 401

CONTINUOUS DEVELOPMENT

Concentrates treated at the Kokkola zinc plant arrive mostly by sea almost directly to the plant area, because there is a deep-water harbour next to the plant. Only one mine, Pyhäsalmi, still produces zinc concentrate in Finland and this concentrate arrives by train. Concentrates are unloaded and transferred by belt conveyor into two storage buildings with a total capacity of 100,000 tonnes of concentrate.

Roasting Lines

The feed-mix is prepared using 10-20 different concentrates or other feed materials, which have become finer and finer (average particle size D50 = 20-30 μιη) during the last few years, see Figure 1. In 1999, 25 different concentrates and four secondary feeds were used, but the total number of different concentrate grades is about 40-50.

Concentrates are added with a weighing front payloader which charges the concentrate onto a conveyor belt system. This belt conveyor takes it up to the four day-bins, each of which has a capacity of 250 to 300 t. The transport and mixing system installed in 1997, including screening and crushing, is sealed to control fugitive emissions. Now it is possible to mix controlled ashes from the zinc casting operation and other fine secondary feed materials with the concentrates. Other fine feed materials may also be used because the belt was recently provided with a water spray system to decrease dusting on the belt by virtue of the micropelletizing effect. The temperature in the upper part of the furnace may be controlled with a separate water spray system on the belt, just before the rotating table feeder.

Leak Air ,

Leak Air

Oxide Dusts

Water Feed

ROUSTING FURNACE

Zh-concentrate

Process Air

Gas Dust BOILER

^y

Water Circ.

coatMt

Gas Dust ._ ESP

Gas.

Calcine

Steam

CYLINDER (stcan ta*>

Water Feed

)

Steam

7 Flue Dust

Calcine Leaching

Figure 2 - The Process Flowsheet Outline of the Roasting Line

As shown in Figure 2, the roaster plant consists of two lines with two 72 m2 Lurgi roasters (diameter 9.6 m). Under normal operating conditions, the bed temperature is about 950°C, the fluidising air feed is 42,000-44,000 Nm7h and the wind box pressure is about 2550

402 LEAD-ZINC 2000

mm H2O. The bed is cooled using cooling coils. Cooling coils receive about 30 % of the total heat, which is recovered in the furnace and waste heat boiler. The area of the cooling coils in the furnace was increased by 5 % in the summer of 1999 after the oxygen feed installation for the second roaster and first oxygen enrichment tests in the spring of 1999. At the beginning of this year it was possible to use oxygen enrichment in both of the roasting lines. The capacity of the oxygen feed is 1500 Nm3/h per line maximum.

The feed mix is thrown into the furnaces using special slinger belt feeders developed by Outokumpu. The pneumatic remote-controlled moving slinger belt feeders are closed, and thus, leakage air and fugitives are minimized. This improves workplace and environmental hygiene as well as making the furnace operation easier and steadier. The bed height and wind box pressure are controlled manually by changing the number of "doorstep slices" in the furnace overflow. An auxiliary small grate has also been developed to control the overflow. The calcine is cooled to 150-300 °C using a fluid bed cooler developed by Outokumpu.

The process gases enter a forced-circulation-type waste heat boiler at 800-1000 °C. Several changes have been made to the mechanical rapping equipment and cooling pipe bundles of the boiler to improve the performance and maintenance access. The gases exit the waste heat boilers at a temperature of 300-350 °C and pass through the hot cyclone separators and electrostatic precipitators into the mercury removal towers. Two electrostatic precipitators are installed simultaneously in both lines to ensure continuous processing.

The furnace and boiler calcines are combined. Then the mixture is conveyed to a ball mill for grinding to a particle size of 50 % below 40 μπι. The ground calcine, cyclone and electrostatic precipitator flue dust are combined and fed to calcine leaching. A typical calcine contains 0.2-0.3 percent sulfide and 1.5 percent sulfate, see Figure 2.

The mercury removal process was developed by Outokumpu after start-up in 1969, but its operations have been improved in recent years with theoretical equilibrium calculations, laboratory and full scale tests. Currently as much as 400-800 ppm mercury content feed can be used and high quality acid is still produced containing only 0.04-0.08 ppm mercury (17).

The electricity supply is quite good in Kokkola because of a stand-by "power plant". However, a separate power backup system has been installed in the roasting plant in order to secure a safe shutdown of the process in case of possible interruptions in the electricity supply. A sprinkler fire extinguisher system has also been provided for safety reasons.

Process Control

The whole roasting process is controlled with an Alcont - Honeywell process automation system. This system has been the subject of continuous development because the control system must cope with large variations in the feed mixture grades. The optimum roasting temperature depends on the feed mixture grade and the latest fuzzy logic-based control algorithm is able to maintain the furnace temperature at the set point with very small fluctuations (+ 3°C). Accurate temperature control stabilizes the furnace bed and makes the operation and production of the furnace steadier. The fuzzy logic-based control algorithm uses fifteen different rules to regulate the furnace temperature with the concentrate feed and oxygen content of the gas. The changes are carried out at thirty second intervals. The mercury-recovering system from the gas is also controlled by a knowledge-based fuzzy controller. The latest version of the control system has proved to be effective and versatile (1,2).

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke. S.E. James and A.H.-J. Siegmund 403

Calculation Models

A calculation model, RAHA2 ("MONEY2"), has been used for optimizing the proportions of the concentrates in the feed mixture. The feed mixture is prepared on the basis of the calculation results. The condition in the furnace is also taken into account in the calculations. The same calculation model is used to make long-term plans for plant production. The HSC software-based material and heat balance models have been used to forecast the effects of different feed mixtures, water spraying and oxygen enrichment on roasting. These results were used for scaling the equipment and planning the test campaigns (3).

Another new development project is the construction of a three dimensional water/air flow model of the furnace and boiler in order to optimize the construction, dimensions and shape of the process gas line as well as to minimize the formation of build-ups. This work is being carried out in the Flow Laboratory of Outokumpu Research Oy.

A lot of effort has been invested in training the operators and employees. One of the newest training tools is the process control system-based simulation model, which may be used to practise critical start-up and shutdown operations without safety risks or expensive damage to the equipment.

Distribution fallout calculations for sulfur dioxide emissions at breakdown situations have also been carried out by a computer model specially modified for these calculations (12).

Personnel

Human resources are usually the most valuable tools of industrial production. The continuous development of processes is more effective if all these resources are utilized to improve the production methods and modes of working. An innovative and positive working atmosphere may be created by sharing responsibility, encouraging personnel to discuss problems and taking care of safety, tidiness and environmental aspects.

Teamwork is widely utilized in the Kokkola plant to improve production efficiency. The main idea is that the operators change their areas of responsibility periodically. Working tasks have been divided into five groups, and every operator is able to take care of all these different tasks and duties. This system activates the continuous development of working routines. A lot of effort has also been invested in the continuous training of personnel to make them understand the background and consequences of their working routines. Five shifts, five operators per shift and five duties have been used. Leaders of the shift teams are also able to supervise the by-product plant, water plant and foundry operations.

Special attention has also been paid to the tidiness of the workplace and environmental hygiene, which both have a positive effect on industrial safety. For example, the roaster plant is provided with a central vacuum cleaning system. The roaster plant is also divided into five cleaning zones, with each shift responsible for its own zone. The cleaning zones are rotated every two months after checking the quality of the cleaning.

The results of the continuous process development and investments in personnel training and welfare may be seen in the operation time, production capacity and efficiency figures shown in Figure 1. Roasting capacity and sulfuric acid production have increased, in spite of the negative changes in concentrate grain size and impurity levels. New production records have been established in most of the last few years.

404 LEAD-ZINC 2000

TEST CAMPAIGNS

The variable concentrate grade choice and other secondary raw materials pose a major challenge to process development. In particular, the small grain size of the concentrates and many impurity elements such as lead, copper, iron, silicates, calcium, sodium or potassium may decrease the quality of the calcine and may also cause operational difficulties. Long-term experience has shown that the behavior of zinc concentrates in a fluidized bed furnace is influenced by the mineralogy, chemical analysis, moisture content, particle size and particle size distribution (8,9). However, it was decided to carry out special test campaigns to generate more detailed information on these variables (4). Research ideas were preliminary tested with the HSC Chemistry-based heat and material balance models, which were described in a previous paper (3). Using these calculations, oxygen enrichment and water feed were found to have positive effects on the process (3). On the basis of the HSC calculations and reference data, it was planned to carry out systematic test campaigns on an industrial scale.

The first industrial scale test campaign was carried out in February - June 1999 (4). These tests were continued in January-February 2000 to obtain more data on the very fine concentrates which were not available in the first campaign (16).

Quite long test periods must be used to achieve steady process conditions and to see the effects of the selected process variables on the calcine grain size and agglomeration behavior. In the first test campaign in 1999, three different step changes were studied in one week. In the second campaign in 2000 only two step changes were tested in one week in order to ensure steady-state conditions. On the other hand the test periods should be shorter, because it is difficult to keep the concentrate mixture composition constant because of the many different types of concentrates used as a raw material. It was not possible to keep the composition of the concentrate mixture exactly the same in each of the test periods. However, within one period the change was quite small. The average feed-mix analysis in the 1999 test campaign was: Zn 53.9+0.7 %, Cu 0.46+0.08 %, Pb 1.8±0.08 %, Cd 0.20±0.02 %, Fe 6.6+0.5 %, S 28.1+0.5 %, As 0.15+0.03 %, Ca 0.58+0.10 %, Mg 0.24±0.04 %, K 0.1610.03 %, Na 0.07 ±0.03 %, Si 0.66±0.14 and Cl 0.0710.02 %.

The effects of the following parameters were tested:

• Bed temperature • Process air flow-rate • Amount of oxide material in the feed mix • Oxygen enrichment • Feed mixture moisture • Water added on the feed belt • Height of the bed • Cooling of the bed. The amount of cooling coils capacity was increased by 5 % before

the last test periods in 1999.

The particle size distribution of the calcine is quite large; small particles from some ten microns up to 30 mm pellets may be found. Because of these large particles, the theoretical size of a representative sample is very large; however, smaller size samples had to be used for practical and technical reasons. This should be remembered when the test results are analyzed. The test procedures and some of the test results are described in this paper.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 405

Effect of the Fluidized Bed Temperature

The first tests were carried out at different temperatures (930, 945 and 960 °C) using a constant process air feed of 42,000 Nm3/h, 250 1/h of water spray on the belt and a constant height of the bed. The bed temperature increase was found to have the following effects:

• Sulfide sulfur, zinc and copper contents increase in the calcine overflow • Lead, cadmium and arsenic contents increase in the boiler calcine • Calcine recovery from the furnace overflow increases • Mixture feed increases and material retention time in the bed decreases • Temperature before the boiler increases • Grain size of the calcine increases.

The calcine retention time decreases with increasing temperature because the feed rate increases simultaneously. The calcine grain size was found to increase with temperature; the larger particles were also found to contain more sulfide, Figure 3. This observation is in agreement with higher sulfide contents at higher bed temperatures, Figure 4. These effects could not be explained by the small changes in feed mixture impurity contents.

0.5 1 1.5 2 2.5 Particle Size in Calcine mm

3.5

Figure 3 - Sulfide Content of the Furnace Calcine as a Function of Particle Size

406 LEAD-ZINC 2000

925 930 935 940 945 950 955 960 965 Bed Temperature °C

Figure 4 - The Content of the Furnace Calcine as a Function of Temperature

Effect of the Process Air Flow Rate

Three tests were carried out using different process air flow-rates (40,000, 42,000 and 44,000 Nm3/h). The bed temperature (950 °C), water feed (100 1/h) on the belt and the height of the bed were kept constant. The results are shown in Figures 5 and 6. The process air was found to have the following effects on the process:

• The sulfide content of the furnace and boiler calcine increases • The concentrate feed increases and the retention time of the calcine in the bed decreases • Steam production increases when the process air flow increases.

0.6

0.5

i Ϊ 0.4

0.3

0.1

0.0

Sulfide in Boiler Calcine

Sulfide in Furnace Calcine

39000 40000 41000 42000 43000 44000 Process Air Flow Rate Nm3/h

45000

Figure 5 - Sulfide Content of the Calcine

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 407

30

30

29

29

B 28

s 28

27

27

26

39000

Steam Production t/h

40000 41000 42000 43000 44000 Process Air Flow Rate Nm3/h

45000

Figure 6 - Concentrate Feed and Steam Production as a Function of the Process Air Flow Rate

Effect of Oxygen Enrichment

Several tests were carried out using different oxygen additions (300, 600, 900 and 1500 Nm3/h). The air feed (44,000 Nm3/h), temperature (950 °C) and water feed (100 1/h) were kept constant. Oxygen enrichment was found to have the following effects on the process:

• The feed increases and the sulfide content of the calcine decreases, see Figures 7 and 8 • Heat generation increases because of the increased production capacity, Figure 9.

Oxygen enrichment increases the oxygen content of the process gas after the boiler if the bed temperature is kept constant by the furnace automation system by increasing the concentrate feed. However, if the cooling of the furnace is simultaneously improved, then the oxygen enrichment does not increase the oxygen content of the process gas, see Figure 10.

408 LEAD-ZINC 2000

31

30

- 2 9 <u <a u. <u 3 28 fc C D U

J 2 7

26

25

Mixture D & Cooling

Mixture A

20.5 21 21.5 22 22.5 23 23.5 24 Oxygen Enrichment vol-%

Figure 7 - Effect of Oxygen Enrichment on the Concentrate Mixture Feed Rate

0.6

0.5

:0.4

§0.3 υ <υ χι 3 0.2 t/5

0.1

0

Furnace Calcine Mixture C & Cooling

Furnace Calcine, Mixture A

Boiler Calcine Mixture C & Cooling

Boiler Calcine, Mixture A

20.5 21 21.5 22 22.5 23 23.5 24 Oxygen Enrichment vol-%

Figure 8 - Effect of Oxygen Enrichment on the Sulfide Content of the Calcine

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 409

36

:34

30

28

Mixture C & Cooling

Mixture D & Cooling

Mixture A

20.5 21 21.5 22 22.5 23 23.5 24 Oxygen Enrichment vol-%

Figure 9 - Effect of Oxygen Enrichment and Improved Cooling on Steam Production

/

O ^ 6 >-_u 'S CD

<5 « 5 c u * J

e 0 υ es 0 4 ' S3 ü

1

Mixture A , '

φ' * * Φ.' . - ' *

. ' ' ' " . - ' ' . · " . ■'' ■ MixtureB

, ' . ' . -" ~ * - * ■ * . . ·

t * , *

.'-'*" · ■ Mixture D & Cooling A A

4 ■

, ■ "

Mixture C & Cooling *

20.5 21 21.5 22 22.5 23 Oxygen Enrichment vol-%

23.5 24

Figure 10 - Effect of Oxygen Enrichment on the Oxygen Content of the Process Gas After the Boiler

410 LEAD-ZINC 2000

Oxygen enrichment increases the bed pressure, see Figure 11. According to fluid dynamic evaluations, the bed pressure increase seems to be caused by the increase of the density and grain size of the bed. Many calcine particles were found to have a dense oxidized surface layer on the sulfide core, Figure 12, in the microstructural studies.

3000

2900

2800

fi «) 5 w) Cfl u

OH

T) 9)

a

2700

2600

2500

2400

2300

Mixture B

Mixture A

Mixture D & Cooling

20.5 21 21.5 22 22.5 23 23.5 24 Oxygen Enrichment vol-%

Figure 11 - Effect of Oxygen Enrichment on Bed Pressure

Figure 12 - SEM Photograph of a Roasted Calcine Grain Showing the Dense Oxidised Surface Layer on the Sulfide Core. Oxygen Enrichment = 21.4 vol%

The particle size distribution of the bed has a strong effect on the stability of the furnace operation. If the amount of the fine fraction (<56μηι) is high, then the bed pressure decreases

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 411

and the temperature gradients in the bed increase; i.e., the bed is unstable and difficult to control. A small oxygen enrichment increased the particle size of the fine fraction and the bed became stable again, see Figures 13 and 14. Such effects have also been observed in other roasting plants (14,15).

900

800

700

" 6 0 0

2 •O500 u <u u. c400 u bO

I 3 0 0

200

100

Oxygen (Left Axis)

Bed Pressure

v^r"; ^■v/

Λ 320

300

280

260

240

220

200

180

160

D,

υ CQ

17/12/99 21/12/99 25/12/99 29/12/99 02/01/00 06/01/00 10/01/00 Date

Figure 13 - Effect of Oxygen Enrichment on Bed Pressure

Figure 14 - Effect of Oxygen Enrichment on the Grain Size Distribution of the Calcine. Oxygen Increases the Bed Pressure and Grain Size

412 LEAD-ZINC 2000

The microstructures of the calcine and build-ups in the boiler and furnace were studied during the yearly shutdown. On the basis of the microscopic and SEM analyses, it seems that the build-ups contain mainly zinc sulfates which are formed from the primary oxides of the calcine. Some fine metal oxides such as those of iron and copper seem to catalyze the formation of the SO3 gas, which reacts with the oxides to form sulfates (10,11,13,16). Oxygen enrichment may be used to increase the particle size of the calcine, and this was also found to decrease dust and sulfate formation in the process gas lines.

Effect of Water Feed

Tests were carried out using very fine feed mixtures (D50 = 7-13 μπι) and different water additions in the "mixing" of the feed materials, 0.15 - 0.51 % of the feed mixture. The other variables were kept constant: process air 42,000 Nm3/h, bed temperature 930 °C, water feed on the belt 400 1/h, oxygen enrichment 22 %, as well as the height of the bed. Moistening of the feed mixture is known to decrease flue dust carry-over to the boiler because of the micro-pelletizing of the feed mixture (8,9).

Increased moistening of the feed mixture was found to increase the particle size of the overflow calcine as long as the moisture content of the feed was lower than 11 % H2O. The larger particles drop to the bed and the oxidation reactions increase both the bed temperature and furnace heat production. Simultaneously, the furnace control system decreases the feed rate in order to keep the bed temperature constant, Figure 15. The feed rate may be increased with more effective cooling of the furnace if the sulfide level is not too high. The process gas temperature in the top of the furnace decreases because fewer fine particles burn above the bed, while simultaneously dust carry-over decreases, Figures 16.

9 10 11 Feed Moisture wt-%

12

Figure 15 - Effect of Moisture on the Feed Mixture Rate

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 413

υ V

o CD o

<s <t>

m S F=* «) 0. b H

905

900

89S

890

885

880

87b

870

865

860

855

8 9 10 11 12 Feed Moisture wt-%

Figure 16- Effect of Feed Moisture on the Temperature of the Process Gas Before the Boiler

The sulfide content of the furnace calcine was found to increase with an increase of the furnace calcine grain size; however, the sulfide content of the boiler calcine was found to decrease at the same time. The total sulfate content of the calcine mixture was found to decrease because the amount of sulfate-containg boiler calcine decreases. On the other hand, the micropellets in the feed mixture may create very fine particles because of explosions caused by gas formation inside the pellet. In the microscopic studies, several exploded micropellets were observed, see Figure 17. The pellet may explode if a dense melted surface is first formed on the pellet and then the oxidation reactions and the moisture generate gas inside the pellet (16,10).

Figure 17 - SEM Photograph of a Calcine Agglomerate Produced with 9.5 % Moisture and 22.4% Oxygen Enrichment

LEAD-ZINC 2000 414

The effect of water sprays on the belt conveyer just before the table feeder was also tested in 1999 by using 0, 0.7 and 1.4 % water in the feed mixture. The other variables were constant: process air 44,000 Nm3/h, temperature 950 °C, water in mixing 0.5 %, oxygen enrichment 22.6 %, as well as the bed height. Water spraying was also found to decrease the temperature of the process gas entering the boiler in these tests.

CONCLUSIONS

The production capacity and efficiency of the Kokkola roasting plant has gradually been improved by the continuous development of the process, the control automation system and personnel skills. The latest development work consists of a new oxygen enrichment system and the new water feed line. Oxygen enrichment may be used to increase the capacity of the roaster plant and simultaneously to decrease the sulfide content of the calcine, to increase the particle size by using fine concentrates and to stabilize the fluidised bed. More cooling elements are needed in the furnace and boiler to compensate for the effect of the higher feed rates. Water feed may be used to decrease the flue dust carry-over to the gas line with very fine concentrates and also to decrease the temperature before the boiler. Personnel training and effective co-operation help to remove the bottlenecks in the working routines and process equipment. The HSC Chemistry software-based roasting model is a useful tool to forecast the effects of different process variables on the heat and material balances of the process. The current results show that continuous development work will be needed in the future in order to cope with the challenges of raw material changes, the demand for higher production capacity, better efficiency and lower environmental emissions.

REFERENCES

1. T. Rauma, J. Nyberg and J. Herronen, "Fuzzy Control of Furnace Temperature in a Zinc Roaster", IFAC Workshop on Future Trends in Automation in Mineral and Metal Processing, 22-24 August 2000 Finland.

2. T. Rauma, "Fuzzy Modelling for Industrial Systems", Technical Research Centre of Finland, Espoo 1999.

3. A. Roine and J. Nyberg, "Heat and Material Balance Model of the Kokkola Zinc Roaster Based on the HSC Chemistry 4.0 Software", General Non-Ferrous Pvrometallurgy, P.R. Taylor et al., Eds., TMS 129th Annual Meeting, Nashville, 2000.

4. P. Lepistö , Master of Science Thesis, Kokkola 19, 8,1999.

5. J.W. Graydon and D.W. Kirk, "A Microscopic Study of the Transformation of Sphalerite Particles during the Roasting of Zinc Concentrate", Met. Trans. B. 19B, 1988,141-146.

6. M.-L. Metsärinta, "Report of Calcine Microstructure in Zinc Roaster", Internal Report 99100-ORC-T, Outokumpu Research Ov, Pori, 1999.

7. A. Roine, "Heat and Material Balance Model of Zn-Roaster", Internal Report 97095-ORC-T, Outokumpu Research Ov. Pori, 1997.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 415

8. R.J. Longton, "Productivity at Savage Zinc", Zinc and Lead Processing. J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton and P. Hancock, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Quebec, 1998, 471-483.

9. J.W.M Filho, T.T. Filho, and J. Welsh, "Debottlenecking Experiences at Companhia Paraibuna de Metais (CPM's) Smelter in Juiz de Fora", Zinc and Lead Processing. J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton and P. Hancock, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Quebec, 1998, 485-499.

10. A. Berg and F. Pape, "Neuere Erfahrungen bei der Wirbelschichtröstung von Zinkkonzentraten". Erzmetall Bd.31 (1978).

11. M.-L. Metsärinta, "Build-ups of the Boiler at Zinc Plant ", Internal Report 99182-ORC-T. Outokumpu Research Oy. Pori, 1999.

12. H. Pietarila, "Sulfur Dioxide and Particle Content Caused by the Interruption Emissions of Kokkola Zinc Plant", Finnish Meteorological Institute. 10, 5, 1999.

13. R. Dimitrov, The Improvement of Zinc Concentrates Roasting in a Fluidized Bed, Rudarsko-Metallurski Zbornik, Mining and Metallurgy Quarterly. Ljubljana 1983.

14. M.-L. Metsärinta, J. Nyberg and T. Hakala, "Effects of the Iron and Other Impurities on Roasting of the Zinc Concentrates", Internal Report 00012-ORC-T. Outokumpu Research Oy. Pori, 2000.

15. M.-L. Metsärinta and J. Nyberg, "Effects of the Particle Size Distribution on the Stability of the Bed in Kokkola Roaster", Internal Report 00011-ORC-T. Outokumpu Research Ov. Pori, 2000.

16. M.-L. Metsärinta, J. Nyberg, A. Rytioja and H. Natunen, "Effects of the Water and Oxygen Use on the Roasting of Fine Zinc Concentrates", Internal Report 00010-ORC-T, Outokumpu Research Ov. Pori, 2000.

17. H. Peltola, P. Taskinen, H. Takala and J. Nyberg, "Scrubbing of Mercury from the Roasting Gases in Kokkola Roaster, Final Report", Internal Report 00005-ORC-T. Outokumpu Research Ov. Pori, 2000.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 417

OXYGEN ENRICHMENT OF FLUO-SOLIDS ROASTING AT ZINCOR

C. MacLagan, M. Cloete and E. H. O. Meyer Zinc Corporation of South Africa Limited

P. O. Box 218 Springs, Republic of South Africa, 1560

A. Newall Process Consultants

P. O. Box 97955, Petervale, Republic of South Africa, 2151

ABSTRACT

This paper initially gives a brief overview of Zinc Corporation of South Africa Limited (Zincor). The original design parameters of the Zincor roasters are compared to the current operation, and the requirements for efficient roasting are listed. Roaster operational problems experienced during the early to mid 1990's are outlined. A description of the investigation and the outcome to increase overall roaster availability are given. The history and the development of the use of oxygen at the Zincor roaster operation, with an account of the associated operational experiences, are explained.

Lead-Zinc 2000 Edited by J.E. Dutrizac, JA. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

418 LEAD-ZINC 2000

OVERVIEW OF ZINCOR

The Zincor electrolytic zinc refinery was established in 1968 near Springs in the Republic of South Africa. The site was the old gold and uranium mine operated by Vogelstruisbult Metal Holdings. The reasons for choosing the site are the following. Much of the existing equipment could be utilised for zinc production thereby reducing capital costs. Some 90% of the zinc and acid customer base are within a 100 km radius, and the road, rail and power infrastructure was already in existence. Finally, suitable labour was readily available locally.

The Zincor process is based on the roast-leach-electrowinning process. The zinc sulphide concentrate is roasted in four Lurgi fluo-solids roasters to produce zinc oxide and sulphur dioxide gas. The SO2 gas is fed to two Monsanto double absorption acid plants.

The zinc oxide, or calcine, undergoes a three-stage leach with spent electrolyte from the electrolysis section. The zinc sulphate solution is then purified in a two- or three-stage purification process depending on the impurity levels. The main impurities precipitated are copper, cobalt, cadmium and lesser amounts of nickel, germanium and arsenic. The purified zinc sulphate solution is subjected to an electrolysis step to plate the zinc onto aluminium cathodes. The anodes are made from lead alloy. The cathode zinc is stripped every 24 hours, melted and cast into one or two tonne jumbos and into 25 kg slabs which are strapped into one tonne pallets ready for marketing. Zincor produces various grades of zinc, ranging from SHG to grade 4 zinc, depending on the customers' requirements.

TECHNICAL DISCUSSION

Fluidized bed roasters are sized according to the concentrate throughput requirements considering the mineralogical, chemical and physical properties of the available feed concentrates. The physical construction and dimensions of the roaster must at least allow for the following.

Sufficient feed and air must enter the fluidized bed to ensure that the chemical reactions are satisfactorily completed and thermally autogenous conditions are maintained. This continuous process must proceed whilst maintaining good particle fluidization in the bed. Disengagement of certain and sufficient particle size groups in the roaster freeboard that return to the bed, contributing to a satisfactory fluidized bed inventory, must be ensured. Sufficient residence time must be provided for all particles, including elutriated particles, ensuring reaction completion in the roasting process. Sufficient temperature is needed for optimum roasting.

Figure 1 shows a schematic diagram of a fluo-solids roaster and its associated equipment.

en

m

O2A

NA

LY

ZE

R

Yn

\ P

RO

DU

CT

GA

S &

CA

RR

YO

VE

R

SOL

IDS

FR

EE

BO

AR

D

SLIN

GE

R

\

O2

FE

ED

BE

LT

INJE

CT

ION

^

P

OIN

T-^

930-

960°

C '

\

llll

llll

fl

},,

WA

STE

HE

AT

~L

_T

BO

ILE

R

! C

YC

LO

NE

WIN

DB

OX

CA

LC

INE

CO

OL

ING

M

EL

EC

TR

OST

AT

IC

PR

EC

IPIT

AT

OR

G

ASE

S T

O W

ET

SCR

UB

BIN

G

SEC

TIO

N A

ND

AC

ID P

LA

NT

PN

EU

MA

TIC

CO

NV

EY

AN

CE

OF

SO

LID

S T

O

ZIN

C P

LA

NT

a o s σ ε

Figu

re 1

-Sch

emat

ic D

iagr

am o

f a

Fluo

-Sol

ids

Roa

ster

and

Its

Ass

ocia

ted

Equ

ipm

ent

420 LEAD-ZINC 2000

Over the years sources of feed concentrates have changed and the concentrate particle size has become finer. Figure 2 illustrates the size data of the concentrates received.

80

75

1 70 f 65 S. 60

55

50

Figure 2 - Characteristics of Concentrate Sizing with Time

Classification of Zincor Roasters

Zinc producers often refer to roaster capacity either in terms of the specific fluidizing air rate Nm3/m2 bed area/h or the specific concentrate feed rate in dry tonnes/m2 bed area/day, and typically such numbers are 500 to 600 and 7 to 9, respectively. Table I below shows the comparison between the design and actual roaster operating parameters.

Table I - Zincor Roaster Design and Operational Parameters Parameter Bed area Design air rate 1996 air rate Design feed rate 1996 feed rate

Unit (m')

(Nm3/m2/h) (Nm3/m2/h) (t/m2/day) (t/m2/day)

Roasters 1 & 2 18

520 367 6.5 5.3

Roasters 3 & 4 35

463 442 5.7 6.5

Impact of Feed Concentrates

Studies have shown that certain impurities within the concentrates can lead to the formation of low melting point substances. This leads to particle agglomeration within the fluidized bed. Such particle enlargement in the bed, together with large quantities of very fine material leaving the freeboard, can lead to difficult roaster operation and ultimately to bed de-fluidization.

Traditionally Zincor treats concentrates from only four mines, each of which has diverse characteristics. Blending options at the site are limited because of the original plant design and stockyard limitations

January 1989 to December 1999

—*—Actual —«—Design

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 421

Size Analysis and Particle Elutriation

Roaster freeboard diameter, for a given gas volume flow, determines particle elutriation and this increases proportionally to the product gas volume. At Zincor, particle elutriation rates often reached +90% of the feed. The design figure is 70% elutriation.

Figure 2 illustrates the dramatic change in the size distribution of the feed concentrates since project implementation. Together with the changes in the mineralogical and chemical properties of traditional concentrates and incidences of roaster losses, the size distribution was thought to play a significant role in limiting capacity at Zincor.

Effect of Altitude and Climate

Zincor is situated 1500 m above sea level and the absolute air pressure is 850 mB. Thus, under the roaster operating conditions, the actual volume of air increases over the normal volume by an additional 17.6%. This factor further increases particle elutriation from the fluidized bed. Also, there are significant variations in temperature and humidity during the year that also affect roaster operations, as shown in Table II.

Table II - Variations in Humidity and Temperature between Summer and Winter Parameter Summer Winter

Maximum relative humidity 100 40 Minimum relative humidity 50 10 Maximum temperature (°C) 40 25 Minimum temperature (°C) 10 -7

Development of the Model

It became necessary to thoroughly understand the limits of the roaster operation and to develop a fully descriptive model. The model was commissioned in 1996 as a management tool, to be continuously updated and referred to in the day to day running of the plant.

Fundamental to the basis for the calculation of the roaster operational data is a full mineralogical analysis and a complete definition of the chemical and physical properties of each concentrate. Data available from such analyses form the basis for an accurate mass and energy balance. The model then calculates all process input data required for any prescribed output product requirements. Consistent calcine quality can be achieved only if the concentrates are properly assessed and predictions and consequent adjustments are made to the roaster input parameters.

ADVENT OF OXYGEN ENRICHMENT

Before 1996, roaster operation was spasmodically interrupted by occurrences of low bed pressure, highly coarse bed calcine, bed de-fluidisation preceded by bed temperature breaks, high levels of particle elutriation and gas train blockages.

422 LEAD-ZINC 2000

From the computer model, it was shown that oxygen enrichment of the fluidizing combustion air was a feasible option to increase roaster throughput without an increase in roaster air volume.

Zincor adopted a maximum particle elutriation target of 85% and increased oxygen in air until the design feed rates were attained. Off-gas composition, a control output from the model, was monitored using reliable oxygen analysis. This control parameter together with good temperature control allowed consistently increased production. The fluidizing air rate to the roasters was in a single stage reduced by 6%, and the wet roaster feed rate was increased by a minimum of 9%. The practice of oxygen enrichment has been sustained to the present. Oxygen in the fluidizing air is maintained at 23.4 to 25.1%. Table III gives the current roaster operational parameters

Table III - Current Roaster Operational Parameters at Zincor Parameter Unit of Measure Roasters 1 & 2 Roasters 3 & 4 Air rate N m V / h 4Ϊ7 471

Feed rate t/m2/day 6J5 Τ_Λ

The oxygen injection site is on the suction side of the roaster blower fan, upstream of the air volume flow-measuring cell. The commissioning of oxygen enriched air to the four roasters was done in stages from March 1997 to July 1997 by manual control. The automation of oxygen addition to the roasters was completed in February 1998. This is achieved by a control loop that compares the actual oxygen concentration at the cyclone outlet to a set value and adjusts the oxygen inlet valve accordingly.

Oxygen Measurement and Safety Precautions

The decision to use oxygen enrichment meant that oxygen control, measurement and safety precautions had to be implemented. These are briefly discussed below.

Oxygen Control and Measurement

The control philosophy of oxygen injection and oxygen measurement resulted in much debate. The basic requirement is the accurate measurement of the oxygen concentration that controls the oxygen injection rate. Oxygen injection must be proportional to both the roaster feed and air rates.

The ideal positions for oxygen measurement are as close to the roaster outlet as possible, or at the windbox air inlet duct. At the roaster outlet, the gases are poorly mixed and high gas temperature and dust loading make conditions impractical for reliable oxygen measurement. Measurement of the oxygen concentration at the windbox inlet was considered, but at that time, depended on the outcome of the continuous use of oxygen enrichment. Improvements to the concentrate blending system also have to be addressed. The site eventually chosen for the measurement of the oxygen concentration was at the cyclone outlet where the temperature and dust conditions are less aggressive than at the roaster outlet and the conditions for representative oxygen measurement are met. At this site, the oxygen probe still has to be cleaned three times per week.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 423

The probe used is an in-situ type stabilized zirconia oxygen analyzer. The zirconium oxide cell is able to measure oxygen concentrations in hot, dirty gases without sample conditioning. In addition, the moisture content of the gas has no effect on the analyzer.

Safety Precautions

Safety precautions include an automatic oxygen trip if the roaster blower trips out or if the roaster air rate decreases to a level where the oxygen concentration exceeds a safety limit of 26%. Oxygen injection may be manually switched off to any roaster in an emergency by means of an emergency cut-out switch in the acid plant control room. Also, the oxygen supplier is on twenty-four hour standby for technical or supply problems.

EFFECT OF OXYGEN

Zincor is limited in roaster concentrate feed rate by the off-gas system capacity. Dust handling systems cannot handle the quantity of calcine dust if the dust elutriation rate exceeds 85% of the feed at 10% over design feedrates.

Oxygen enrichment of the fluidizing air, for a given concentrate feed rate, causes a net reduction in freeboard gas velocity. The roaster feed rate may then be increased until the dust carry over reaches the capacity of the off-gas system. At Zincor, operating off-gas volumes at maximum concentrate feed rates remain less than originally designed, whereas SO2 and off-gas dust loading remain far in excess of design.

At the targeted calcine elutriation rate of 85%, oxygen enrichment of the fluidizing air has resulted in an increased feed rate to the roasters of 6.6 to 7.1 t/m2/day, a capacity that meets or exceeds design figures. Figure 3 illustrates the increase in roaster feed with oxygen enrichment at 100% running time.

90 91 92 93 94 95 96 97 98 99

Year

Figure 3 - Roaster Feed Rate at 100% Running Time

The effect of oxygen enriched fluidizing air on the operation and the equipment with time is discussed below.

424 LEAD-ZINC 2000

Although there are clear advantages in increased concentrate throughput at the roasters, there are also disadvantages. Calcine dust can become more "sticky" at high oxygen partial pressures in the off-gas, especially when a low roaster outlet gas temperature occurs. This has highlighted the need to restrict air in-leakage to a minimum and to maintain as high an off-gas temperature as possible.

Roasters

Oxygen enrichment of the fluidizing air can alter the heat balance of the roaster operation. In essence the cooling effect of the nitrogen is reduced and roaster bed temperatures are increased, causing a corresponding increase in roaster quench water. However, there is a net reduction in off gas volume that permits either an increase in feed rate or a reduction in the upward velocity of the gases in the roaster.

Roaster temperatures are controlled by using as many fluidized bed cooling coils as possible and by optimising both the moisture content of the roaster feed material and water injection to the bed. Roaster water injection must be kept to a minimum. Additional coils were installed during 1998/1999 and roaster water injection was reduced substantially. For example, in # 3 roaster the water injection rate decreased from 1.3 - 1.8 m3/h to 0.5 - 1.0 m3/h by increasing the number of coils from one double to three doubles. Water injection rates, however, still need to be reduced further.

Waste Heat Boilers

Since the introduction of oxygen enriched air to the roaster operation at maximum production, the prevailing off gas dust loading is significantly higher than design (85% elutnation compared with maximum 70% at design stage). Boiler bundle choke-ups have increased, particularly in the first and second positions. This has caused an additional down-time of approximately 0.6% per annum. A method to reduce or eliminate this problem was to increase the rapping stroke. The disadvantage of this method is that the bundle tubes eventually fail at the site where rapping occurs. An unacceptable amount of downtime would result from the replacement of the bundles. An alternative method of rapping is currently being investigated. A new trial bundle and rapping system will be installed during the annual shutdown in August 2000, and the trial bundle will be monitored. The correlation between oxygen enrichment and incidence of boiler bundle chokes is not fully understood.

Cyclones

Each roaster off gas system contains two cyclones in parallel. No established changes have occurred to cyclone availability since the introduction of oxygen enrichment.

ESPs and Hot Gas Fans

All four off gas system ESPs consist of two fields. In all cases the dust loading is higher than design. Down time for cleaning the ESP internals has doubled from approximately 0.25 to 0.5% per annum. Downtime for cleaning hot gas fan impellers has increased by 0.8% per annum. This equipment has been the focus of much attention since the inception of the oxygen enrichment programme. Factors such as dew point, particle size, and gas analysis have all been studied. Increasing cyclone efficiency and decreasing the moisture content of the gas is required for these units to be more effective.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 425

In 1998, the # 3 and # 4 ESPs were upgraded by increasing the spacing between the electrodes and replacing the starwires. This has reduced the frequency of cleaning and balancing the impellers from around once per ten days to approximately once per month because of high vibration levels.

Gas Scrubbing Circuit

Attention to this equipment was made necessary by the increased production throughput as a result of oxygen enrichment.

The cooling of the gas stream to below the dew point has been achieved by the implementation of the following programme. Auxiliary cooling towers for the cooling of water from the weak acid circuit were installed. The duty of the cooling pond pumps has been increased to meet the water design flow rate through the weak acid coolers. The optimisation of the cooling capacity of the weak acid plate heat exchangers has been achieved. Prior to these modifications, curtailment of the roaster feed rate was required during the hot summer months to allow cooling of the gas to below 40°C, but such curtailments are no longer required.

Zincor does not consider the advent of oxygen enrichment as the cause of the additional maintenance downtime in this plant area. Although the choking of humidifying sprays has increased, all maintenance can be carried out during normal stoppages.

Contact Plant

An outside company was approached to conduct a study to ascertain the changes required on the contact plant to accommodate the increased roaster throughput. The main recommendations arising from the study centred on the heat exchangers and the catalyst loadings. An extensive heat exchanger replacement programme, incorporating re-designed units, was realised from 1997 to 1999. Catalyst volumes were increased in the #1 plant by 43%.

Oxygen enrichment of the roaster fluidizing air, the planned elimination/reduction of air ingress and targeted control of off-gas oxygen partial pressure has resulted in an increased SO2 partial pressure at the converter. On occasions dilution air is allowed into the gas stream under controlled conditions. Prior to 1996, Zincor did not require the use of dilution air for SO2 control at the converter.

CONCLUSIONS

Oxygen enrichment of the roaster fluidizing air at Zincor has resulted in a 9 to 11% increase in zinc metal production capacity. Under Zincor conditions, this translates to 1 tonne oxygen per 1.4 tonnes dry concentrate. Optimum oxygen enrichment is in the range of 23.4 to 25.1% oxygen in air. At Zincor, using higher levels of enrichment is limited by constraints in gas handling.

Control of the off-gas temperature and off-gas oxygen partial pressure is critical for successful operation. Downtime has increased by less than 1.5% per annum attributable to oxygen enrichment. Finally, modeling of the roasting process for purposes of predicting and targeting process settings has provided a control tool for sustained operation with oxygen enrichment.

426 LEAD-ZINC 2000

ACKNOWLEDGEMENTS

The authors wish to acknowledge the contributions of the Afrox technical personnel and the Zincor Engineering Department.

REFERENCES

1. C.J. van Niekerk and C.C. Begley, "Zinc in South Africa", Journal of the South African Institute of Mining and Metallurgy. Vol. 91, No. 7, 1991, 233-248.

2. M.I. Shull, P. Ellis and A.S.E. Kleyenstüber, "Phase Chemical Investigation of Products from Roaster 1 and 2 at Zincor, Springs", Project No. 029942C. Mineralogy and Process Chemistry Division, Mintek, Randburg, South Africa, 23 June, 1995.

3. G.J. Martin and R. Khoun, "Mineralogical Observations of Roaster Feed and Roaster Feed Material", Mineralogical Reference No. ZC96/12, Gold Fields Laboratories Limited, 17 May, 1996.

4. C. MacLagan, R.D. Beck, A. Newall and E.H.O. Meyer, "Modelling of Fluo-solids Roasting at Zincor", Colloquium: Trends in Base Metals Smelting and Refining. The South African Institute of Mining and Metallurgy, Mintek, Randburg, South Africa, 24-25 April 1997.

5. A. Zanet, Private Communication, Union Miniere Group.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 427

METHOD FOR PROCESSING SILICEOUS ZINC ORES

S. Ikenobu Kamioka Mining & Smelting Co., Ltd.

1-1, Shikama, Oaza, Kamioka-cho, Yoshiki-gun Gifu, Japan

ABSTRACT

This paper describes a method for processing high-silicate zinc concentrates by hydrometallurgical zinc refining. Silica in concentrates reacts to form zinc silicate during roasting. The zinc silicate readily dissolves and slowly precipitates in sulfuric acid solutions during leaching. However, if proper operating parameters are not maintained, the colloidal silica may form an unfilterable gel. That is, according to the investigation results obtained, the silica concentration in the reactor has to be maintained at 5 g/1 or less (preferably 2 g/1 or less) by promoting the precipitation reaction. The proposed method has made it possible to precipitate silica in a form having excellent solid-liquid separation characteristics, without the need for facilities having a high filtration capacity, and to maintain a low silica concentration in the reactor by feeding a composition containing pre-adjusted silica contents.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

428 LEAD-ZINC 2000

INTRODUCTION

It is well known that the solid-liquid separation characteristics of residues become a problem when zinc concentrates with high silica content are treated in a hydrometallurgical process. Once treated improperly, the silica dissolved in the sulfuric acid solution turns to a gel which significantly reduces the solid-liquid separation characteristic of the residue. As a consequence, larger capacity solid-liquid separation facilities such as thickeners, filter presses and others are required. In some cases, the slurry may completely turn into a gelatinous form, which cannot be treated. For this purpose, a technology to settle the dissolved silica in a form easy for solid-liquid separation becomes necessary.

The process presented in this paper is designed for the effective treatment of high-silica zinc concentrates utilizing some of the properties of silica in sulfuric acid solutions.

EXPERIMENTAL

Basic Tests on Calcine Leaching

Silica in concentrates produces Zn2Si04 in the roaster following the reactions below.

ZnS + 3/202 -> ZnO + S02 (1)

2ZnO + Si02 -> Zn2Si04 (2)

The Zn2Si04 easily dissolves in sulfuric acid solutions. It is well known that the dissolved silica, if treated improperly, will turn to a gel and which will have an adverse effect on the solid-liquid separation characteristics. In order to identify the basic phenomenon, leaching of calcine was tested. The tests were conducted by extracting 100 g/L of calcine containing 20% Z^SiC^ for four hours at a temperature of 70° C under different pH conditions to investigate the solid-liquid separation characteristics and zinc extraction percentages.

The effect of the pH on the settling rate of the residues and the zinc extraction percentages are illustrated in Figure 1. The figure clearly shows that the parameters dramatically change in the area of pH 2.5. The consistency of the residues was investigated by X-ray diffraction analysis and a peak of Z^SiCU could be identified in the residues treated at a pH value of 2.5 or higher, whereas no Zn2SiÜ4 peak was observed in those samples treated at pH values below 2.5. The residues treated at pH values below 2.5 were further examined by electron microprobe analysis. Silica, which is deemed to be dissolved and settled, was found. The test results led to the conclusion that Zn2Si04 existing in the calcine dissolves at a pH value below 2.5 and settles as silica gel with poor solid- liquid separation characteristics.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 429

pH

Figure 1 - Effect of pH on the Settling Rate and Zinc Extraction

Behavior of Silica in Sulfuric Acid Solutions

In a sulfuric acid solution, silica in calcine is dissolved by the reaction shown in equation (3) and precipitates to an equilibrium concentration (5) in accordance with the reaction of equation (4).

Zn2Si04 + 2 H2S04 -> 2ZnS04 + Si(OH)4 (aq) (3)

nSi(OH)4 -> (Si02)n + 2nH20 (4)

The results of the performed tests show that the reaction of equation (3) is easily promoted. However, the reaction of equation (4) occurs slower than the leaching reaction, and the silica then exists in an over-saturated concentration as shown in Figure 2(a). One may conclude that as it becomes more saturated, the silica is preferentially precipitated in a gel form, resulting in an adverse effect on the solid-liquid separation characteristics. In order to quantify this over-saturation parameter, we defined the value with the highest content of silica in solution as "Max. Si02-Concentration".

430 LEAD-ZINC 2000

(a) Conceptual graph when (b) Conceptual graph when precipitation is slow precipitation is promoted

Actual change of silica concentration

Actual change of silica concentration

CJ ^ c o

Ή a o U £ n o o

.../.v.".: ".:".".".:".: ?. 4^τ*^—^ */ 1 ^^~" Λ Max. SiO 2 concentrarion

Jr reaction of i equation (3)(fast)

" ■ » ■ ^

reaction of * ~ ^ ^ equation (4)(slow)

4 *" reaction of equation Q)

0

' / !/Max.Si02 concentration^

■»

reactiS\of equation (4) (fast) X

Time(H)

Figure 2 - Relation Between S1O2 Concentration and Retention Time

Figure 3 shows the relation between the "Max. Si02-Concentration" and the settling rate of the precipitated residue. This figure demonstrates that the settling rate becomes poor as the "Max. SiCVConcentration" increases. Particularly, when the "Max. SiC>2-Concentration" reaches 6.0 g/1, the slurry turns into a gel. The results suggest that at a minimized "Max. S1O2-Concentration", good settling characteristics are achieved.

1111111111111111 iji 111 [ 111 ι μ 11

pH2.2, Τεπφ.9θΌ reaction time; 4hours

2 3 4 5 6 Max. SiCh-Concentration (g/1)

Figure 3 - Effect of "Max. SiC>2-Concentration" on the Settling Rate

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 431

Trial to the Improve the Solid-Liquid Separation Characteristic

The "Max. SiC>2-Concentration" is determined by the leaching rate of equation (3) and the precipitation rate of equation (4). As can be seen from Figure 2(b), the "Max. Si02-Concentration" decreases to low levels by accelerating the conversion rate of equation (4). Taking these facts into consideration, each parameter affecting the precipitation rate was investigated.

Setting the Most Suitable Treatment Conditions

The effect of the pH and the temperature on the precipitation reaction was investigated. The experiment was carried out by dissolving Zn2Si04 in spent zinc electrolyte at 70° C in order to obtain a S1O2 concentration of 5 g/1. After filtration of the residue, the solution was maintained at a specific temperature and then was adjusted to a specific pH value using ZnO (super grade reagent). The results are shown in Figure 4. The horizontal axis represents the reaction time and the vertical axis represents the S1O2 concentration. In can be observed that in order to increase the precipitation rate of silica in dilute sulfuric acid solutions, the pH should be raised. In addition, the elevation of temperature also significantly promotes the precipitation reaction of S1O2, even in the low pH range.

90°C ΓΓΓΤΤΠ

pHl.O

ρ Η 2 . θ \

Ό 1 ~ 3 ~4~ 5 ~6~ 7 T~ 0 l 2 3 4 5 6 7

Time (h) Time (h)

Figure 4 - Effect of pH and Temperature on the Rate of the Silica Precipitation Reaction

Addition of Seed

As was demonstrated, raising the pH value and elevating the temperature promotes the precipitation reaction of silica. For industrial applications, however, the consequential operating aspect of maintaining a narrow range of pH, as well as the economical consideration of the higher energy cost for heating the solution in order to obtain excellent precipitation rates, makes it difficult to justify the practical realization of the laboratory results. For these obvious reasons, an alternate way of leaching the silica-rich calcine was investigated. It might be expected that

432 LEAD-ZINC 2000

the reaction of equation (4), would be promoted once precipitated silica was added as a seed. The result is shown in Figure 5. It can be seen that the addition of seed indeed promotes the precipitation reaction causing the "Max. SiC>2-Concentration" to drop. The addition of seed has a positive impact on the precipitation process and permits a more effective treatment at low temperatures and over a wider range of pH values.

pH1.5,Temp.70 Ό reaction time ; 4hours recycled seed addition /acid soluble silica addition ratio

Time (H)

Figure 5 - Effect of Silica-seed on the "Max. SiC^-Concentration"

Further investigation revealed that the addition of seed is also effective at higher acid concentrations (H2SO4 : 60 - 120 g/1) such as those used in the hot acid leach process. These experiments were carried out by dissolving 10 g/L of soluble S1O2 at a specific sulfuric acid concentration and at a temperature of 90° C. Then, 10 g/L of insoluble silica was added as a seed. The results are shown in Figure 6. The horizontal axis is the sulfuric acid concentration and the vertical axis is the settling rate. As is obvious from the figure, the addition of seed results in residues with good solid-liquid separation properties regardless of the sulfuric acid concentration.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 433

10

c ε

(3j Pi to c

ιζι

I I | I I I | I I I |

Temp. ; 90Ό soluble Si02 ; lOg/L insoluble Si02 ; lOg/L

I I I 1 I I I I I I I I

100 80 60 40 20

H2SO4 Concentration (g/L)

Figure 6 - Effect of the H2SO4 Concentration on the Settling Rate

DISCUSSION

Application to Practical Operation

A study was performed to incorporate the obtained experimental results into practical operation.

The previous experiments indicated that the most important and critical factor for an efficient leaching and filtration process of silica-rich concentrates is the "Max. S1O2-Concentration" in the reaction vessel. The pH value, the temperature and seed addition are the only means to lower this concentration. Taking this into consideration, it can be assumed that a longer retention time in the vessel will lower the "Max. Si02-Concentration". On the other hand, a longer retention time requires a high plant capacity, which directly results in high capital expenditures. Therefore, pH control, temperature and seed addition are discussed in detail.

The proper pH is very effective for accelerating the S1O2 precipitation reaction, but at high temperature and in a high slurry environment, pH control is very difficult. During a 6 month pilot plant operation, however, we were successful in controlling the pH within a range of 0.1. This must be emphasized as a very important aspect because a difference of 0.1 in the pH value will greatly influence the solid-liquid separation characteristics, when no seed is added. In the pilot scale tests, we were able to treat calcine of 6.8% S1O2 content without aggravating the settling characteristics under operating conditions of no seed addition, at a

434 LEAD-ZINC 2000

temperature of 90° C and in the range of 1.8<pH<2.5. Nevertheless, the industrial application of the described technology under these conditions is limited because of the higher energy costs associated with the elevated temperature involved and the need to maintain the pH in a relatively narrow range. These circumstances make effective seed addition very attractive in order to achieve an improvement in residue precipitation.

Figure 7 shows a conceptual drawing of a single stage neutral leach process. When insoluble silica is added as seed from outside the circuit, the quantity of produced residue will increase. Therefore, in order to have effective utilization of the precipitated silica, the underflow from the thickener is partially recirculated as the additive. However, when the S1O2 concentration in the residue is low or when the net quantity of S1O2 is insufficient, the recirculated additive eventually becomes ineffective. To have silica consistently precipitated with good solid-liquid separation characteristics, the following condition has to be met and sustained:

(Net quantity of insoluble S1O2 in seed)/(Net quantity of soluble S1O2 as (Zn2Si04) /0.3 (5)

Spent Calcine electrolyte

ntet D D D

Recycled precipitated silica seed I Residue treatment process

Figure 7 - Example of Silica Seed Addition

Furthermore, it was found that the use of lead-silver residues (residues from the hot acid leach process), which have highly condensed silica, as a seed source is effective. This material can be employed over a wide acid concentration range and, therefore, is applicable to each process of residue treatment.

In the hot acid leach process, there is a blending of neutral leach residue containing precipitated silica. This silica is a residue produced at a pH value higher than that of the hot acid leach process. However, it was shown that the silica from the neutral leach residue does not re-dissolve in high acid concentrations, but acts as a seed and promotes the precipitation reaction. In other words, in the high temperature and high acid concentration leaching process, no special facility is required as long as the conditions do not go below those of equation (5). Even when the S1O2 concentration in the reaction vessel rises, the amount of underflow recirculated from the thickener of the hot acid leach process can be increased to maintain the above conditions.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 435

Having adjusted the most suitable conditions for the type of flocculant, the quantity of flocculant and the mixing ratio will further improve the solid-liquid separation. Together with the kind cooperation of Kurita Co., Ltd., we were able to find the most effective conditions for the seed addition in the previously described investigation.

CONCLUSIONS

By adopting some characteristics of silica precipitation in sulfuric acid solutions, we were able to achieve the efficient treatment of high silica-bearing zinc concentrate. This process can be applied in practical operations without any major changes in the conventional hydrometallurgical zinc refining process.

The industry has to face the need for treating more varieties of concentrates with different quantities of impurities. The process presented in this paper is able to treat high silica-bearing zinc concentrates without major changes in the plant facilities.

REFERENCES

1. I.G. Matthew and D. Eisner, "The Hydrometallurgical Treatment of Siliceous Zinc Ores", USPat.3.656.941, 18 April 1972, (EZ Co.of Australasia Ltd.).

2. S.P. Fugleberg and Poijarvi, "Hydrometallurgical Treatment of Soluble Silicate-bearing Zinc Materials", US Pat.4,148,862, 10 April 1979 (Outokumpu Oy).

3. H.L. Radino and Companhia Mercantil e Industrial Inga, "Process of Zinc Extraction from Zinc Ores Comprising Soluble Silicates by Means of Hydrometallurgy", Australian Pat.224.195, 31 May 1957.

4. F.J.J. Bodson, "Process for the Treatment of Material Containing Zinc and Silica for Recovering of Zinc by Hydrometallurgical Way", US Pat.3,954,937, 4 May 1976.

5. G.B. Alexander, W.M. Heston and R.K. Her, "The Solubility of Amorphous Silica in Water", Journal Physical Chemistry. 58, 1954, 453-455.

6. S. Ikenobu and K. Shimokawa, "Method for Processing Zinc Silicate Containing Zinc Crude Material", Aust.Pat.Application Au-A-36886/97, 5 September 1997 (Mitsui Mining and Smelting Co. Ltd.).

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 437

LEACHING AND PURIFICATION AT COMINCO'S TRAIL ZINC OPERATIONS

DJ. McKay, G. Sterzik, T.L. Salway and W.A. Jankola Cominco Limited P.O. Box 2000

Trail, British Columbia, Canada V1R 4L8

ABSTRACT

At Trail, British Columbia, Cominco's zinc leaching and solution purification processes include direct pressure leaching of zinc concentrates, leaching of fume from the lead smelter, leaching of calcine and electrolyte purification. These processes have been adapted to match the unique characteristics of Cominco's integrated zinc and lead operations. Recent modifications include changes to the purification process and iron-residue handling circuits. In 1997, a new five-compartment autoclave, complete with ancillary equipment, was installed in the pressure leaching plant. This paper discusses these changes, the current flowsheet and further improvements.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

438 LEAD-ZINC 2000

INTRODUCTION

The metallurgical facility at Trail, British Columbia was originally built to smelt copper and gold ores from the nearby Rossland mines in the late 1800s. Cominco Limited, incorporated in 1906, soon diversified into other metals and, many process technology changes and capacity expansions later, is now one of the world's largest zinc-lead smelters. The annual zinc production capacity is currently 290,000 tonnes, one of the highest-capacity zinc smelters in the world.

The implemention of KTVCET smelting technology over the last two years has mandated some changes in the leaching circuits of the zinc operations. As well, Cominco is working towards many strategic capacity, optimization and reliability enhancement projects with a view to producing more zinc in the future.

OVERVIEW OF ZINC OPERATIONS

Trail's Zinc Operations has been described well in previous papers (1-3). A brief review is presented here in order to facilitate discussion of the recently completed and planned changes to the leaching and purification circuits.

Cominco's zinc operation is based on standard roasting, leaching and electrowinning technology, but also incorporates direct pressure leaching of zinc concentrates. Recovery of zinc contained in the iron residues is achieved through KIVCET lead smelting and slag fuming. The zinc contained in the fume is recovered in an oxide leaching plant. The zinc-containing electrolyte from both the pressure and oxide leaching plants is processed through the roaster calcine leaching plant (generally called the sulphide leaching plant in Trail) purified and delivered to the electrowinning plant. The overall flowsheet is given in Figure 1.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 439

Gas

' ' '

Sulphur Gases

'

Sulphur Gas Handling

Sulr, ■

Products

Zinc Concentrates

1 1

Roasters

Calcine

P Ele

■ ■

Pressure Leaching

Pressure Leach Slurry

1 ' Leaching

& Purification

urified ctrolyte 1

Electrolysis &

Melting

R sfine &A

d Zinc loys

Oxide Electrolyte

_ Elemental Sulphur

Zinc Fume

i Oxide

Leaching

(Smelter)

Leach f Kesidues

(Smelter)

Figure 1 - Overview of Zinc Operations Processes

PRESSURE LEACHING

Process Description

Zinc pressure leaching has operated at Trail since 1980, and accounts for about 25% of Trail's overall zinc production. Direct leaching of zinc sulphide concentrates is done in an autoclave at 150°C and 1100 kPa using spent electrolyte, technical-grade oxygen and small additions of sulphuric acid to produce a zinc-rich electrolyte and elemental sulphur. The leaching chemistry is well described in the literature. The Trail plant (Figure 2) is a single train, straight-through operation, consisting of feed preparation, leaching, heat recovery, sulphur recovery and sulphur purification.

Feed preparation is composed of a feed receiving bin, a single ball mill and a single thickener. Historically, the concentrate was ground to a target dso of 25μπι, but since regrind projects were implemented at the Sullivan concentrator, the feed preparation system simply provides lump-free slurry at about 65% solids.

As outlined in a previous paper (3), the original four-compartment autoclave was replaced with a five-compartment autoclave in 1997. Technical grade oxygen is sparged into the first four compartments, and return acid is added to all compartments. With the new autoclave, the concentrate feed rate has increased from 18.5 t/h to 23 t/h.

440 LEAD-ZINC 2000

Heat recovery includes a flash tank to reduce the pressure from 1100 kPa to 101 kPa, with the flash steam being used to preheat the acid introduced into the first two compartments of the autoclave. The autoclave discharge slurry is further cooled to 80°C in water-cooled tanks.

Spent Acid

Zinc Concentrate

Water

ii Acid Heating

Feed Preparation

Hot/Cold Acid

Feed Slurry

Flash Steam

Pressure Leaching

Disch Slurry Cooling &

Heat Recovery

Cooled Slurry Sulphur

Recovery

Dirty Sulphur

Oxygen Sulphur

Purification

Pressure Leach

Slurry to Calcine

Leaching

Clean Sulphur to Sale

Residue to Roasters

Figure 2 - Pressure Leaching Flowsheet

Sulphur recovery consists of a hydrocyclone and two parallel flotation circuits. Normally, over 95% of the elemental sulphur is recovered in this circuit. Work is ongoing to improve the recovery of the finer sulphur size fraction. The sulphur concentrate is dewatered on a belt filter, melted and passed through a molten sulphur filter to remove unreacted minerals and ash to produce a marketable sulphur product.

General Operating and Maintenance Strategy

The plant enjoys 96% availability, with the vast majority of unavailable time attributable to planned shutdowns. The high availability is partly due to an aggressive maintenance plan including eight-hour monthly shutdowns and a 10-day annual shutdown. In addition, a long term in-house mechanical development program focused on improving the design of key pieces of equipment (such as pumps, pressure letdown valves and heating coils) has also provided large gains in equipment reliability. Strict adherence to operating standards has also helped to avoid extended downtime, such as not operating when oxygen supply, pressure and purity are below standard, batch leaching following any shutdown before starting up to continuous operation and maintaining the input acid strength above 160 g/L H2SO4.

Experience with Red Dog Concentrate

The plant was originally designed to treat 100% Sullivan concentrate, a high-iron (9%) marmatic zinc concentrate. Over the last 5 years, Red Dog zinc concentrate, a low-iron (4.5%) pyritic-marcasitic zinc concentrate, has also been very successfully processed with the Sullivan concentrate. The typical blend has contained 30% Red Dog feed.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 441

The Sullivan mine will be closing in 2001, creating an imperative for the treatment of Red Dog concentrate-based blends. The concern was that the current high production rates will not be possible with the lower contained iron and pyrite-marcasite mineralogy of Red Dog concentrate.

A two-week plant test was conducted in January-February 2000 to gather engineering data for the treatment of low-iron concentrates. Blends of Sullivan and Red Dog containing up to 100% Red Dog, were treated at various feed rates. The plant successfully treated 100% Red Dog concentrate at 18 t/h, yielding 96% zinc extraction, and up to 21 t/h was also explored. These results surpassed expectations based on previous in-house laboratory and plant-scale test work. In addition, the results were achieved without any operational optimization or equipment modifications. Some of the factors leading to this success included:

• Sparging and purging higher than typical oxygen volumes • Strictly adhering to the requirements for high (>95%) oxygen purity • Increasing the goulac dosage (triple the typical plant usage) • Ensuring that feed acid concentrations remained above 165g/L H2SO4 • Maintaining leach temperatures in A and B compartments above 145°C.

Iron levels in solution ranged up to 5 g/L (indicating significant pyrite leaching) compared to the typical 8-10 g/L under normal operations. The majority of the leaching is usually completed within the first two autoclave compartments, although under these low-iron conditions significant leaching also occurred in the third compartment. Because of the fine mineralogy of Red Dog concentrate and high goulac addition, elemental sulphur particle sizing decreased dramatically, posing problems for the flotation circuit.

Cominco is confident that with suitable optimization, Red Dog concentrate can be treated at the 22 t/h rate now reached with Sullivan-Red Dog blends. A major Cominco Research and plant technical effort is underway to reorganize, upgrade and optimize the flotation circuit to ensure recovery of the fine sulphur particles that result from treating Red Dog feed.

One of the options to increase zinc production in the future is to refurbish and restart the #1 autoclave and run two autoclaves in parallel. The successful plant test with Red Dog concentrate, combined with anticipated improvements in sulphur recovery, makes this pressure leaching option very appealing.

OXIDE LEACHING

Process Description

The oxide leach plant treats baghouse dust made in Lead Operations' slag fuming plant, and has been in continuous operation since 1930. The dust or fume contains principally zinc and lead, as well as a host of minor elements, including arsenic, antimony, cadmium, tin, aluminum, indium and germanium. In addition to recovering zinc and cadmium, the plant produces an indium-germanium preconcentrate which is processed in the downstream indium-germanium plant.

442 LEAD-ZINC 2000

The plant was originally a stand-alone zinc hydrometallurgical operation, with leaching, iron purification, zinc dust purification, a cadmium plant and a cellhouse for zinc electrowinning. With the introduction of fume dehalogenation leaching and a new cadmium plant over the last 20 years, plant flows have been integrated into other zinc circuits at Trail as described below. The oxide leaching plant also does miscellaneous services for Zinc Operations, including maintaining manganese levels in the electrolyte and recycling metallic zinc-containing materials.

The traditional oxide leach plant consisted of an acid leaching step, several neutral leaching steps and, until recently, a zinc dust purification step to remove cadmium, as shown in Figure 3. In the old scheme, most of the fume and various recycle streams were leached in spent acid in the acid leach step. The resulting acid slurry was thickened, and the lead-rich solid residues were filtered, repulped in water, combined with the residue slurry from the calcine leaching plant and pumped to the smelter.

Fume Fume Zinc Dust

Spent Acid

1. Ferric Iron

Acid Leaching

Neutral Leaching

Lead Residues to Smelter

Znc Dust Purification

Pre-concentrate

Electrolyte ' to Calcine

Leaching

Cadmium Residues to

Cadmium Plant

Figure 3 - Historic Oxide Leaching Process

The electrolyte from acid leaching advanced through neutral leaching where the majority of the minor elements (As, In, Ge and Al) were precipitated. A solution of ferric iron was also added to the electrolyte from acid leaching to assist minor element precipitation. A precipitate, or preconcentrate, was separated from the electrolyte in a thickener, filtered and then repulped and advanced to the indium-germanium plant.

The final overflow solution was then advanced to a small zinc dust purification circuit to cement cadmium, and the purified electrolyte was then forwarded to the calcine leaching plant.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 443

Flowsheet Modifications

With the start-up of the new lead smelter in Trail, fume characteristics have changed markedly, and the concentrations of lead, arsenic and other undesirable minor elements have increased while the concentration of zinc decreased through the initial period after the smelter startup. In 1999 a program was initiated to revise the flowsheet to cope with:

• High lead, high arsenic fume • Increased arsenic concentrations in electrolyte • Poor removal of impurities, as well as poor recovery of germanium from the zinc

electrolyte solution.

This program was implemented by changing some piping and the service of existing equipment.

Iron Distribution System

Ferric iron solution is now continuously added to both acid and neutral leaching. This has resulted in higher iron utilization and enhanced impurity precipitation efficiency. The addition of soluble iron to the acid leach promotes the precipitation of ferric arsenate, enhancing arsenic rejection to the residues and dramatically lowering the arsenic levels in the plant electrolytes. Previously, the arsenic concentration in the acid leach electrolyte was as high as 5 g/L. The arsenic concentration is now below one gram per liter.

Acid Leaching

The pH range of acid leaching and the iron concentration in solution are controlled in order to maximize the precipitation of arsenic from solution. It is presumed that iron-arsenic compounds are formed, which are not soluble in acid. The increased rejection of arsenic to the residue at this stage also helped to improve the efficiency of impurity metals (including germanium) removal in the subsequent neutral leaching step.

Modified Neutral Leaching

Changes in neutral leaching pH and the addition of iron solution were made to optimize the precipitation efficiency of impurity metals (including germanium) while minimizing the acid soluble content in the leach residue.

The above strategies (iron distribution, acid leaching and the modified neutral leaching) have collectively improved arsenic removal from the circuit, and hence improved the quality of the preconcentrate feed to the indium-germanium plant. With increasing experience in the operation of the new lead smelter, the fume quality has also improved. This further facilitates solution purification through iron precipitation as well as the production of indium and germanium.

The production of indium and germanium products is important, and as was expected with the start-up of the KIVCET furnace and a new slag fuming furnace, the production has been expanded. The associated indium-germanium plant has now the capacity to produce 60-70 tonnes of refined indium and 30-35 tonnes of contained germanium in germanium dioxide. Further expansion of germanium production is planned and will be implemented as necessary to match demand increases.

444 LEAD-ZINC 2000

Purification Decommissioning

The zinc dust purification circuit in the oxide leaching plant was originally commissioned as a purification system for an old oxide cellhouse, decommissioned about 20 years ago when the oxide and calcine leaching plant electrolytes were combined. However, the oxide purification circuit continued to operate with about one gram per liter addition of dry zinc dust, principally to cement cadmium in the oxide electrolyte prior to entering the calcine leaching plant. A portion of the purification residue was then forwarded directly to the cadmium plant. Some arsenic was also removed from solution.

There were several problems with the oxide electrolyte purification circuit. The arsenic concentration in the zinc fume from the smelter had increased since the start-up of KIVCET, resulting in an increased tendency for arsenic breakthrough to the oxide electrolyte. This, in turn, had led to an increased potential for arsine generation during zinc dust purification as well as in the storage tank for purification residues in the cadmium plant. In addition, the oxide purification equipment was in very poor working condition.

In mid-1999, the oxide electrolyte purification circuit shown in Figure 3 was shut down on an interim basis. The arsine outbreaks were essentially eliminated and the calcine leaching plant performance was not detrimentally affected. Therefore, the oxide purification process has been permanently decommissioned.

Equipment Upgrades

In addition to the low-cost piping and valve upgrades required to implement the new oxide leaching flowsheet, major equipment upgrades have also been pursued to modernize this very old plant. Aging Kelly pressure filters, installed many years ago for filtering acid leach residues, have been replaced with two Perrin semi-automatic plate and frame filters with a cake reslurry system. The newer of the two Perrins has proven capable of handling over 100 t/d of residue and producing a dry, well-washed cake. This has reduced electrolyte losses and improved cake quality, thereby easing downstream cake handling difficulties.

Three original wood-stave 65-m3 pachucas (two in the acid leach and one in neutral leaching) have been replaced with 80-m stainless steel pachucas. Although it may appear regressive, pachucas remained the best choice for the application, given plant geometric constraints and mixing requirements. Two wood-stave 550-m3 thickeners have been replaced with stainless steel thickeners. The wood drive and rakes for a third thickener have also been replaced with a stainless steel drive and rakes. The steel drive and rakes allow for higher torque and loading than the original wood parts. A new circuit, composed of two tanks and tie-in to filters, has been installed to treat additional internal feeds.

A Modicon analog control system has been installed and is slowly being implemented. The control system currently controls the Perrin residue filters and cake reslurry system. When the system is complete, it will represent a radical departure for a plant that has been completely manually operated for nearly 70 years.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 445

CALCINE LEACHING

Process Description

Cominco's calcine leaching circuit (also called sulphide leaching) combines the products from zinc pressure leaching and oxide leaching with the mainstream calcine leaching circuit. Process details have been well described in other papers (1-3).

The process begins with a weak acid leach (pH 1.5-1.8) followed by a pH bump to 3.8-4.0 for iron, silica, arsenic and antimony precipitation. The slurry is thickened, and the solid residues are filtered and washed to recover water-soluble zinc in two stages of American disk filters. This filtration step is being upgraded as described in a following section. These iron-rich residues are then repulped and combined with the residue slurry from the oxide leaching circuit and are pumped to the smelter where the slurry is filtered, combined with other materials and processed in the KIVCET furnace.

The solution overflowing the acid thickeners is further neutralized in the neutral leaching step with an excess of calcine to assure good quality electrolyte. The solids are separated in a thickener and are returned to acid leaching for recovery of zinc. The clear overflow electrolyte is forwarded to the purification circuit.

Zinc dust purification is done in three stages:

• Cold stage (50° - 55°C) for copper, cadmium and thallium • Hot stage (80° - 85°C) for cobalt and nickel • Polish stage (~80°C) for the elimination of any remaining impurities.

Each stage includes a filtration step. The residues from the cold stage are processed for cadmium recovery. Hot stage residues are releached to reclaim zinc, filtered and the final cobalt-nickel-containing cake is forwarded to the smelter. The polish stage solids are processed in the oxide leaching plant.

Following zinc dust purification, the hot solution is cooled to about 25°C in two stages of cooling. During cooling, gypsum is precipitated and is then separated from the zinc sulphate solution in two clarifiers. The zinc sulphate solution is forwarded to the electrolytic plant.

KIVCET Recycle Dust

The start-up of the KIVCET process has changed the deportment of several minor elements in the lead operations. Cadmium, for example, has tended to accumulate downstream of the boiler in the electrostatic precipitator (ESP) dust, and the leaching of chlorine from the fume in the fume dehalogenation leaching step has also decreased, leading to a buildup of chlorine in the zinc electrolyte. There is thus a strong incentive to bleed the cadmium- and chlorine-containing ESP dust from the smelter circuit to zinc operations to both manage cadmium and decrease the amount of chlorine reporting to fume. The chloride content of the dust mandates that it must be processed through the roasters as a method of introducing it to zinc operations.

Several attempts were made from late 1998 to early 1999 to add this material to the roasters. In all trials, the settling rates in the calcine leaching plant neutral thickeners increased

446 LEAD-ZINC 2000

significantly and the residues exhibited different characteristics (higher underflow density), leading to thickener failures. A third trial later in 1999 did not lead to a neutral thickener failure, but the changed settling characteristics mandated that the ESP dust again be taken off the roaster charge. In all three cases, a few days after the ESP feed was discontinued the settling rate in the neutral thickeners was highly variable. There was a carry-over of flocculant from the acid thickeners and the torque on the neutral thickener rakes increased.

This operating experience suggested that treating ESP dust through the roasters would not be technically feasible. However, several other factors may also have contributed to the neutral thickening problems, thus confounding the apparent effects of the ESP dust addition. For example, in the first two trials, the failures occurred the day following a plant shutdown. Thus, the electrolyte temperature was low, and the zinc concentration in electrolyte was also low. The negative effects of low electrolyte temperature and zinc concentration were confirmed in pilot tests in Cominco Research, and a fourth trial was run at the end of 1999.

In the fourth trial, up to several tonnes per day of ESP dust were carefully blended in the roaster feed over a 10-day period. Although increased settling rates and underflow densities were again observed in neutral thickening, these changes were successfully managed with close attention to flocculant addition and underflow pumping. Interestingly, thermogravimetric analyses (TGA) of selected solid samples revealed that the ESP dust did indeed alter the properties of the calcine, although it was not possible to explain why or how this occurred. In addition, the particle size of the neutral leach residues increased. At the time of submitting this paper, the ESP dust bleed to the roasters is being initiated on an indefinite basis.

Residue Blending

As shown in Figure 1, the KIVCET lead smelter processes the leach residues from both the calcine leaching plant, consisting principally of jarosite and ferrite, and the oxide leaching plant, consisting principally of lead sulphate. It had been observed that the ratio of sulphide-to-oxide residues could vary within wide limits, leading to potentially significant variations in the KIVCET feed composition. This contributed to challenges in setting the fluxing targets and was thought to be impacting on KIVCET chemistry.

In early 1999, an improved residue blending process was commissioned in the calcine leaching plant. In this process the oxide residues are directed to a new 35-m3 holding tank. From this tank, a predefined volume of oxide residue is advanced to a transfer tank, from which the combined sulphide-oxide residue mixture is pumped to the smelter. The volume to be pumped is set by the ratio of sulphide-to-oxide residue and the production rate of sulphide residue. To ensure consistent blending, this ratio set point is usually not adjusted. However, the calcine leaching and oxide leaching plants can, in consultation with the smelter, mutually agree to set point changes as a result of changes in the operating conditions.

The benefit of the residue blending was observed immediately as illustrated in Figure 4. The variation in both iron and lead assays was much less after the project was commissioned, and as a result, the frequent assays shown in the figure, used to aid in the setting of the fluxing targets, were actually discontinued.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 447

Figure 4 - Impact of the Residue Blending Process on the Variation in Residue Composition

Residue Filtration

The American disk filters used to filter the residues from calcine leaching are very old and require considerable maintenance effort to keep operating. In addition, cake washing is inefficiently achieved by repulping the cake and refiltering. The amount of repulp water is not always well controlled, and thus, can lead to excessive input of water to the zinc electrolyte. By the end of 2001, these filters will be decommissioned, and the acid thickener underflow will be sent directly to the smelter feed plant with the residues from oxide leaching and filtered on soon-to-be upgraded existing Lasta press filters. Operation of the upgraded Lasta filters will be modified to include a cake wash cycle for zinc recovery.

Experience with Straight-Through Purification

When the purification circuit at Trail Operations was upgraded in the early 1990s, a high-density, antimony-activated system was installed for cobalt removal (Figure 5). The expected advantages of this hot-stage purification system were:

low zinc dust consumption low operating temperature low residence time.

448 LEAD-ZINC 2000

Cold Stage Filtrate

Reagent

-£"

Zn Dust Zn Dust

Filtrate

Residues

W

Figure 5 - High-Density Hot Stage Purification System

In reality, the system was very complex to operate with multiple addition points for reagents, zinc dust and spent cellhouse acid. The circuit included a set of cyclones to allow a portion of the solids to be recycled to maintain the solids density at approximately 15 g/L in the tanks. To prevent excessive sanding-out in the purification tanks, the flow through the tanks was maintained by recycling, from the filters, a portion of the product filtrate. There were several problems with this circuit:

• Zinc dust consumption exceeded the target • There were frequent excursions of unacceptable cobalt levels in electrolyte - the

circuit was not stable • The complexity of the circuit made process control and troubleshooting of chemistry

very difficult - there were too many variables • The tanks sanded-out frequently and had to be taken offline for acid cleaning • The maintenance costs were high because of the additional pumping step required

for the cyclones.

Tests were undertaken using a straight-through purification circuit to determine if high quality electrolyte could be consistently produced (Figure 6). The first test, conducted in 1996, indicated the potential of this process. However, the existing circuit had insufficient residence time to produce the required quality of electrolyte at full production rates with the existing size of zinc dust. These tests revealed that either (a) additional tanks were required, or (b) the size of the zinc dust had to be reduced (increased surface area) before implementing a straight-through process.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 449

Cold Stage Filtrate

Reagent

Zn Dust

© θ Θ Filtrate

fV Residues

Figure 6 - Straight-Through Hot Stage Purification System

A second program, undertaken in 1997, utilized the latter approach and was very successful. In fact, the straight-through circuit has operated ever since that time. In the first two months of operation, the electrolyte quality improved, as did process reliability. Zinc dust consumption was lower than had been planned (Figure 7). However, years of acid cleaning had taken a toll on some of the purification tanks, which now require repairs, with the result that the circuit has generally been operated with only four of the five available tanks. To compensate, zinc dust additions have been increased beyond the target range once again. The quality of the electrolyte has been very good, with only a 0.7% occurrence of combined antimony and cobalt above the target values of 0.01 mg/L and 0.3 mg/L, respectively.

With straight-through operation well demonstrated, the focus in purification is now on equipment upgrading, such as replacing aging cold stage filters and the hot stage residue releach tanks. A project is just underway to define the purification process requirement in the calcine leaching plant to increase zinc production in the future.

450 LEAD-ZINC 2000

Figure 7 - Zinc Dust Consumption as a Percentage of Cathode Zinc

Managing Scale and Solids Agglomeration in the Cold Stage Purification

In the purification circuit, the most significant cause of lost production in 1997 and 1998 was the plugging of the cold stage equipment with scale dislodged from the sides of pipes and tanks, and with dense agglomerates, or "balls".

The first step to alleviating the plugging problem was the implementation of a preventative maintenance program to descale the pipes and tanks on a regular schedule. The maintenance program immediately had a significant benefit, but it did not eliminate the problem. The second step began by examining the morphology of the agglomerates to determine the cause of their formation. A literature review of research studies showed that cemented cadmium forms a microstructure consisting of two-dimensional, leaf-like dendrites. These dendrites are capable of interlocking, and are believed to be the mechanism leading to the formation of the agglomerates. These studies also showed that, when cadmium is cemented from a zinc sulphate solution also containing copper, agglomeration is reduced and the solids have a porous, more sponge-like structure. These studies support the observations from Trail Operations and other zinc plants that agglomerate formation is reduced during cadmium cementation if there is sufficient copper available in solution. Accordingly, a temporary copper addition system was put in place in the third quarter of 1999. For the six months starting in October 1999, there have been no production losses attributed to the plugging of the cold stage equipment with agglomerates. As a consequence, a small, permanent system for adding a copper sulphate solution to the feed to the cold stage will be installed.

Zinc Dust Addition

Cominco uses a water atomization process to produce the zinc dust needed for electrolyte purification in the calcine leaching plant. The zinc dust is delivered as a slurry in water from the melting plant to the leaching plant. There the zinc dust is decanted, repulped in purified electrolyte and distributed to the points of addition in the purification circuit. The

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 451

distribution lines require routine calibration of solids density and flow through the addition points to the purification tanks, particularly after descaling, with a view to achieving accurate zinc dust addition. The drift in these parameters between calibrations leads to an imprecise measurement of the actual zinc dust consumption.

A weigh tank system was successfully piloted and implemented in hot stage purification during the fall of 1999 to avoid these difficulties. In this system, the zinc dust slurry is discharged batch-wise into a conical tank suspended on load cells. Once a preset level (slurry volume) is reached, the slurry weight is recorded and then discharged to the purification tank. To facilitate accurate calculation of the zinc added per batch, the operator performs regular hand check measurements for electrolyte carrier density. In fact, this system auto-compensajes for variation in solids density and flow through the addition points. The overall daily consumption of zinc dust, determined by integrating the batch additions from the weigh tank, agrees within about 5% of the zinc dust produced at the melting plant. The weigh tank is equipped with alarms that allow the operator to diagnose problems not only with the weigh tank but also with other components of the zinc dust slurry distribution system. After a few commissioning challenges, the weigh tank has been operating trouble free for over six months.

OVERALL ZINC OPERATIONS ISSUES

Solution Volume and Sulphate Control

Management of water and sulphate in Zinc Operations is an important issue. The winter season in Trail is very wet and large amounts of precipitation can enter the circuit, driving the sulphate concentration down and increasing the circulating electrolyte volume by as much as 35%. The low sulphate concentration is equivalent to lower acidity in the return acid, and so additional concentrated acid must be added in pressure leaching to maintain leaching efficiency. The low sulphate concentration also decreases the gypsum removal efficiency. The inventory of water accumulated in the circuit during the winter is readily removed in the electrolyte cooling towers during the drier summer months, but the accumulated sulphate cannot be removed as easily. The result is that the sulphate concentration increases dramatically, leading to problems such as decreased current efficiency.

Since 1997, a detailed plan has been implemented each year to balance the water and sulphate inputs. The goal of the plan is to keep both the solution volume and sulphate concentration within target ranges while optimizing the additions of water and concentrated sulphuric acid. There are five plants.with specific action items that are tracked daily, with weekly conformance reporting. The adherence to these plans has significantly improved the control of the solution volume and sulphate concentration, although further refinements are possible.

CONCLUSIONS

Integrating zinc and lead smelters and operating three leaching plants (pressure leaching, oxide leaching and calcine leaching) presents challenging metallurgical issues. The never-ending optimization of any one operation generally mandates some modifications to another. This paper discussed many of these issues and how they have been addressed, and outlined some of the projects currently underway.

452 LEAD-ZINC 2000

ACKNOWLEDGEMENTS

The authors would like to thank Cominco Limited for permission to publish this paper.

REFERENCES

1. D.W. Ashman, O. J. Delong and W. A. Jankola, "Silica Control During Zinc Calcine Leaching at Cominco's Trail Operations", World Zinc '93, I. G. Matthew, Ed., The Australasian Institute of Mining and Metallurgy, Victoria, Australia, 1993, 217-226.

2. G. M. Belland, R. G. Pressacco and W. Van Beek, "Leaching Plant Flow and Level Control Optimization at Cominco's Trail Operations", Zinc and Lead '95. T. Azakami, Ni Masako, J.E. Dutrizac and E. Ozberk, Eds., Mining and Mineral Processing Institute of Japan, Tokyo, Japan, 1995, 599-611.

3. M. J. Brown, E. T. de Groot, M. G. Heximer, A. J. Karges, G. N. Masuch and C. M. Okumura, "Zinc Capacity Increase at Cominco's Trail Operations", Zinc and Lead Processing. J. E. Dutrizac, J. A. Gonzalez, G. L. Bolton and P. Hancock, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 41-54.

Chapter 6

Imperial Smelting Technologies

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 455

THE CONTINUING EVOLUTION OF THE IMPERIAL SMELTING PROCESS

R.W. Lee Imperial Smelting Processes

Stoke Orchard Cheltenham GL50 2LS, United Kingdom

ABSTRACT

The first prototype plant using the Imperial Smelting Process commenced operation in 1950. It is timely therefore to review the developments that have occurred with the process over the past half-century. A major advantage claimed initially for the process was its ability simultaneously to recover zinc and lead from bulk concentrates. This has remained a fundamental advantage, but it has been extended by the increasing treatment of secondary materials containing zinc and lead. This has been facilitated by the development of hot briquetting and tuyere injection. These techniques and their impact on the technology of the operation will be described. The other major advantage claimed for the process was its potential to achieve new levels of productivity in the pyrometallurgical smelting of zinc. This has been realised with operating furnaces producing up to 400 t/d of zinc. The technical developments that have enabled this scale of operation will be reviewed.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

456 LEAD-ZINC 2000

INTRODUCTION

The Imperial Smelting Process (ISP) for producing zinc and lead in a blast furnace was developed at Avonmouth, United Kingdom. After three years of work on an experimental furnace, the prototype production furnace was commissioned 50 years ago in September 1950 with a design capacity of 20 t/d. Its commissioning was described as good despite the fact that the first campaign lasted only 6 days during which an average of 4 t/d of zinc was produced. This No. 1 furnace was joined in 1951 by a No. 2 furnace with a capacity of 25 t/d of zinc. This latter furnace continued to operate until 1967 by which time it was producing about 90 t/d of zinc and 40 t/d of lead bullion.

The results obtained with the prototype furnaces inspired such confidence in the process that the first full-scale furnace was installed at Swansea, United Kingdom in 1959 with a design capacity of 150 t/d of zinc and 70 t/d of lead. Before the commissioning of this furnace, a number of smelting companies had become interested in the process. Trials were carried out for them on the prototype furnaces at Avonmouth and some became licensees of the process, which progressively found application throughout the world. Thirteen furnaces are now in operation as listed in Table I.

Table I - ISP Smelters in Operation Location Country Shaft Area

m Year

Started Max. Zinc Output, t/y

Max. Lead Output, t/y

Avonmouth Chanderiya Cockle Creek Copsa Mica Duisburg Hachinohe Harima Miasteczko Noyelles Godault Porto Vesme Shaoguan No 1 Shaoguan No 2 Veles

UK India Australia Romania Germany Japan Japan Poland France Italy China China Macedonia

27.2 21.5 24.2 17.2 19.3 27.3 19.4 19.0 24.6 19.0 18.7 17.2 17.2

1967 1991 1961 1966 1966 1969 1966 1979 1962 1972 1975 1996 1973

105,700 61,400 97,300 34,100 97,400 114,400 88,900 82,300 115,700 84,600 81,500 70,900 59,700

51,200 31,000 40,700 17,100 45,100 52,100 28,600 31,000 48,300 36,000 34,800 31,900 30,500

In 1999, production by these furnaces amounted to 1,020,000 tonnes of zinc and 429,000 tonnes of lead bullion, representing approximately 12% and 7%, respectively, of the world production of these metals.

The process has been well described in papers by Miyake (1) and Takewaki et al. (2) and is illustrated by Figure 1.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 457

Figure 1 - Diagram of the Imperial Smelting Process

458 LEAD-ZINC 2000

The major steps in the flowsheet are as follows:

• The feed consisting of zinc and lead sulphidic concentrates, oxidic materials and fluxes is desulphurised and agglomerated.

• The gas produced from the sintering process is cleaned and then is used to produce sulphuric acid.

• Sinter and other zinc and lead containing materials and coke are smelted with preheated air in an ISF blast furnace.

• Lead bullion and slag are tapped from the furnace and separated in a forehearth. • Zinc is produced as a gas in the furnace and is absorbed in molten lead in a splash

condenser. Zinc is then recovered by separation from the cooled lead in a cooling circuit external to the condenser.

• Zinc and lead produced in the ISF are refined as required to meet market requirements.

The success of the process has been due to its ability to treat unusual raw materials and to the continuing technical development of the process.

RAW MATERIALS

The ISP was originally developed as a straight zinc smelting process to replace horizontal and vertical retorts. However, during the work on the first experimental furnace, drosses were recycled to the furnace as part of its sinter charge in order to recover their metal content. The lead content of the sinter rose progressively and molten lead appeared at the slag tap-hole and could be separated and recovered. It was then apparent that the furnace could be a producer of lead as well as zinc. This ability of the ISP to treat materials containing both zinc and lead has remained its prime advantage.

Concentrates

The ISP consumes virtually all the world's production of bulk concentrates, materials that usually contain 45-60% of zinc plus lead and with at least 10% of each metal. The ability of the ISP to accept bulk concentrates was a major factor in determining the viability of the McArthur River mine in Australia. Tests over many years had been unable to determine a method of separating the zinc and lead in the ore with acceptable recoveries. However, much better recoveries could be obtained if a bulk concentrate was produced. Table II shows the results of flotation tests to produce selective and bulk concentrates, as given in the paper by Lee et al. (3).

Table II - Flotation Results on McArthur River Ore Recovery of Zn Recovery of Pb

Selective flotation 69.7% 39.6% Bulk flotation 85.0% 67.0%

The development of the mine to produce bulk concentrates and the need for assured outlets for a major part of its output led to the decision of Mount Isa Mines Ltd. to acquire the Avonmouth and Duisburg ISP smelters.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 459

Apart from bulk concentrates, the remainder of the concentrates fed to ISP smelters tend to be materials that are less acceptable to electrolytic zinc plants or to straight lead smelters. Thus ISP smelters can bid competitively for zinc concentrates with high lead, magnesia or iron oxide levels or lead concentrates containing appreciable quantities of zinc.

Secondary Materials

The ISP now accepts a wide range of secondary zinc materials including Waelz oxides, electrolytic zinc plant residues, galvanisers' ashes and non-ferrous flue dusts. These materials are mainly either incorporated in the sinter plant feed or are hot briquetted and charged directly to the ISF.

Materials that are fed via the sinter plant are mainly oxidic and there is a limit to how much can be accepted without a deterioration occurring in the performance of the sinter machine. This can be overcome to a certain extent by, for example, using clinkered Waelz oxide so that it becomes a replacement for recycled sinter fines. This means that less lump sinter output will need to be crushed for recycling so that the output of lump sinter increases.

Many ISP smelters are treating increasing quantities of electrolytic zinc smelter residues. Because of environmental concerns, dumping of residues even in lined ponds presents problems. Sending the residue to an ISP smelter can offer the electrolytic smelter a more economic solution to the residue problem than dumping it. The analysis of residues that are used can vary considerably but a typical analysis is as follows: 21% Zn, 6% Pb, 1.3% Cu, 28% Fe and 5% S

Electric arc furnace (EAF) dusts have become a significant source of feed for the ISP. These contain up to 35% Zn and environmental legislation in developed countries is discouraging the previous practice of dumping EAF dusts. In a number of countries the Waelz process to produce zinc oxides treats these dusts. In some cases these oxides are hot briquetted as a direct feed for the ISF. An alternative method of treating steel plant and other dusts is to inject them directly into the furnace through the tuyeres.

HOT BRIQUETTING

The hot briquetting process was originally developed by ISP Ltd. with the intention of providing an alternative to sinter as feed for the ISF. The ISP smelters at Duisburg, Hachinohe, and Porto Vesme operate hot briquetting plants. The Harima smelter receives hot briquettes from the Shisaka plant of its associated company. The Noyelles-Godault smelter receives briquettes from its associated company, Recytech SA. Other hot briquetting plants in Europe supply several ISP smelters with briquettes.

Figure 2 illustrates the flowsheet of a hot briquetting plant. The feed material is pelletised, preheated in a rotary kiln, and then briquetted in a roll press at about 400°C.

460 LEAD-ZINC 2000

Figure 2 - Diagram of a Hot Briquetting Plant

A typical feed for hot briquetting, as reported by Miyake et al. (4), is given in Table III.

Table III - Hot Briquetting Feed Material Material Waelz oxide Zinc ashes Lead sludge ISF dross Briquette fines Total

T/d 40 38 7 40 15 140

Zn(%) 57 67 37 26 49 49

Pb(%) 10 1-2 22 38 19 19

The size of the briquettes is typically 85mm x 45mm x 30mm or 60mm x 45mm x 30mm. The object of the briquetting process is to make briquettes that are sufficiently hard to withstand handling before they are charged into the ISF. They should also retain their structure while they are being smelted in the furnace.

The proportion of hot briquettes in the ISF feed is about 10% of the zinc charged at Hachinohe, Noyelles-Godault and Porto Vesme, but is usually around 30% at Duisburg. When significant quantities of hot briquettes are charged, particular attention has had to be paid to the furnace loading of alkalis and halides. Because these compounds have low boiling points, there is substantial volatilisation from the furnace to the condenser where they can have undesirable effects. Low melting point compounds are formed with zinc and lead oxides, which can result in the formation of hard accretions at the furnace offtake and condenser inlet. These are very difficult to remove. Further, the condenser dross can become pasty making it more difficult to remove from the condenser. This results in increased entrainment of condenser lead. Build up of dross on the rotors is another problem that can occur and it adversely affects the condenser efficiency.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 461

TUYERE INJECTION

Trials with oxygen enrichment of the furnace blast air were first carried out in the United Kingdom as long ago as 1964. In general, it is considered too expensive to use routinely, but a small amount of enrichment is used in several smelters. At Cockle Creek in 1975 oil was injected into the tuyeres and replaced up to 12% of the coke. There were no adverse results and oil injection continued until the rising cost of oil rendered it uneconomic. Subsequently, in 1986 the Harima smelter became the first ISF to inject solids on a consistent basis. Up to 6% of the lump coke charge was replaced by injected dry quench coke dust. This practice continued until November 1995.

In 1993, the Avonmouth smelter commenced the first long term programme to develop the injection of zinc-lead material through the tuyeres, a process for which a patent had been granted to ISP Ltd. in 1990 (5). The experience at Avonmouth has been described in the paper by Musson et al. (6). As this relates, there are two injection plants at Avonmouth. The first was developed to inject condenser pump dross and the second to inject secondary materials, principally EAF dusts. The dross injection plant has a single injection system that feeds eight tuyeres via an eight-way splitter valve. The plant for injecting secondary materials has four injectors each serving two tuyeres, and is illustrated by Figure 3. When injecting secondary materials, coal makes up about 25% of the total injected materials.

Figure 3 - Diagram of the Tuyere Injection Plant at Avonmouth

Further information about the operation of the plant has been provided by Gammon (7). Since the major shutdown in September 1998, the average monthly injection rates for secondary materials and dross have varied largely because of furnace problems. However, as much as 1360 t/m of secondary materials and 1630 t/m of pump sump dross have been injected. At the maximum rate, about 7% of the zinc in the furnace charge has been injected.

Control of tuyere injection is facilitated by the use of oxygen injection boosting the tuyeres showing low blast flow rates. When injection rates become higher, it is anticipated that oxygen will be used continuously at the rate of 2% enrichment for an injection rate of 1.5 t/h

462 LEAD-ZINC 2000

and at 4% enrichment for an injection rate of 2 t/h. At present, if the blast to a tuyere falls below 2000 Nm3/h, injection is turned off.

No negative relationship has been found yet between the injection rate and either zinc in slag or zinc condensation efficiency. It had been thought that carryover of unreacted EAF dust to the condensers might adversely affect the condensation efficiency. However, it was found that the carbon content of the pump sump dross rose when secondary material and coal were injected. This phenomenon was not observed at Duisburg and the larger coal size at Avonmouth was identified as the cause. Anthracite coal sized less than 3 mm was being used and it was estimated that a third of the coal was reaching the condensers unreacted. This problem has been eliminated by changing to pulverised coal sized 80 % <75 μπι.

The tuyere injection plant at Duisburg was installed in 1996. It is capable of handling 35,000 t/y of secondary materials and was described in the paper by Schneider and Schwab (8). Later developments with this plant are likely to be discussed by these authors in their paper "Zinc Recycling via Imperial Smelting Technology - Latest Developments and Possibilities" to be presented at this conference.

Compared with hot briquetting, tuyere injection would appear to have a number of advantages as a method of feeding secondary materials to the ISF:

The capital cost of the equipment is much less. The operating cost is less particularly since no preheating process is required and labour requirements are minimal - a single supervisor per shift at Avonmouth.

• Tuyere injection is an excellent method of treating fine dusts with the minimum environmental impact.

However, the rate at which materials are injected is significantly less than the rate at which hot briquettes are charged to the furnace. Briquettes, amounting to over 30% of the furnace zinc input, have been charged routinely at Duisburg. Thus far, the total of dross and secondary materials injected at Avonmouth has only reached 7% of the zinc input.

The raw materials costs for hot briquetting and tuyere injection are attractive and can be negative if the supplier would otherwise be facing dumping costs. If the smelter is at its limit for sinter production, both processes are very attractive because they offer the opportunity to increase metal output without proportional increases in operating costs.

PRODUCTIVITY

The majority of ISP smelters have achieved considerable increases in output since their installation. Some of the increases in capacity have been achieved by physical expansion of the furnace. However, many increases have been attributable to considerable increases in the intensity of operation. These have been permitted by:

• Improvement in sinter quality. This has meant a reduction in the sulphur content and closer control of factors such as the lime/silica ratio and sinter hardness.

• Increased blast preheat. The temperature has increased from about 650°C to up to 1100°C. The low calorific value of the furnace waste gas prevents temperatures above this level from being obtained. It has been demonstrated that an increase of 100°C in

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 463

preheat temperature is equal to an increase of some 3-4% in zinc output. Alternatively, a similar reduction in coke consumption can be obtained at the same zinc output.

• Improvements in blast distribution. Calculation procedures developed by ISP have been used with great success for determining the size of tuyere raceways and active zones. This has enabled the number of tuyeres, their diameter and spacing to be optimised.

• Improved cooling of the furnace to withstand the greater intensity of operation. The original design of the furnace was based on the lead blast furnace and used a refractory-lined upper shaft and water-cooled jackets for the lower shaft. The frequency of jacket failures led to the replacement of this design by the use of a shower-cooled casing, which has generally worked well.

• Improved refractories. The change from chrome-magnesite to alumina-chrome for the hearth refractories has eliminated the susceptibility of the hearth to hydration. Previously hydration had occurred on a number of furnaces and had resulted in hearth expansion. In some cases this had seriously affected the furnace stability and its operation.

• Improved tuyere design. Considerable attention has been paid to the design of the tuyeres and to water quality. In September 1998, Avonmouth installed cast-copper tuyeres following advice from British Steel that the heat flux was too great to use steel tuyeres. The life of the copper tuyeres has since been over a year and the simpler water circuit inside the tuyeres has enabled them to cope better with variations in water quality.

• Improved condenser design and operation. In designing condensers, water tank model work proved to be invaluable and has enabled the design of the lead splash rotor to be optimised.

The largest ISF in operation is at Hachinohe where the latest expansion was carried out in 1998. Information about the furnace and condenser before and after the expansion has been provided by Oshita (9) and is given in Table IV. The furnace now has a shaft area of 27.3 m2

and produced 114,000 tonnes of ISF zinc in 1999 compared with its original design shaft area of 17.2 m2 and capacity of 54,000 t/y zinc. The latest changes made at Hachinohe have actually decreased the blast intensity in the furnace so it might be expected that some further increase in output could be achieved, if sufficient feed were available.

To achieve a further step change in productivity at many ISP smelters, radical changes to the furnace operation will be required. It is considered that the best way to do this is by increased injection of feed materials, fuel, and oxygen through the furnace tuyeres. This more intensive operation of the furnace will require improved cooling. It may be necessary to redesign critical areas of the furnace and perhaps to take advantage of the experience gained in cooling copper smelting furnaces.

464 LEAD-ZINC 2000

Table IV - ISF Expansion at Hachinohe

PW + SHG zinc

Furnace zinc

Furnace lead

Slag

Sinter

Hot briquettes

Coke

Blast rate

Intensity in shaft

Intensity at tuyeres

LCV gas

Condenser efficiency

Hot coke/furnace zinc

Furnace shaft area

Tuyere level

Number of tuyeres

t/y

t/d

t/d

t/d

t/d

t/d

t/d

Nm3/h

Nm3/h/m2

m3/h/m2

NmVh

%

kg/t

m2

m2

Condenser size (internal) mm

Before Expansion

102,000

324

157

247

778

82

308

43,920

1,996

2,142

66,000

91.0

950

22.0

20.5

14

5.5m x 1.016m

After Expansion

112,000

355

172

294

833

130

344

49,980

1,830

1,923

75,600

91.0

970

27.3

24.4

16

6.0m x 1.096m

ENVIRONMENTAL PERFORMANCE

Environmental considerations play an increasingly important role in non-ferrous smelting. The ISP does possess an advantage in this context because it is a considerable consumer of wastes from other processes. Mention has already been made of the use of electrolytic zinc plant residues, steel plant dusts, galvanisers' ashes and car battery paste. Nevertheless, in the operation of the ISP itself, great attention has to be paid to the avoidance of pollution of the working atmosphere the external atmosphere, and receiving waters.

Working Atmosphere

Levels of 0.15 mg Pb/m3 are typically specified for the working atmosphere. Achieving this level is difficult in some working areas and requires the use of effective hooding and large ventilation volumes.

In ISP smelters, the measurement of die effect of the worker's exposure to lead is by measurement of lead in blood. A typical level at which workers legally have to be transferred away from the exposure to lead has been 70 μg/100 ml blood. However, a number of ISP smelters have been successful in reducing the transfer level to 50 μg/100 ml of blood and below. This has been achieved by attention to many factors in addition to plant ventilation.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 465

Two significant changes have been the introduction of air supplied helmets, and at some smelters, the banning of smoking. The Cockle Creek smelter became a non-smoking site in 1992 and the Avonmouth smelter in 1995.

Discharges to Atmosphere

Previously a typical limit for the total dust concentration in the discharged gas has been 50 mg/Nm3 with a lead content of 10 mg/Nm3. However, other restrictions, such as a total mass emission limit for lead for the site, have required the meeting of a much lower lead concentration in die discharged gas. Typically a discharge of less than 5 mg Pb/Nm3 is required, and this has necessitated the use of high efficiency bag filters wherever possible.

The discharge of gas containing compounds of sulphur has been under tight control for some years. Most of the ISP smelters use double-absorption acid plants that result in discharges containing less than 500 ppm SO2. Some of the smelters with older acid plants using single-absorption are introducing flue-gas scrubbing.

Discharge of Liquid Effluents

Limits on effluent discharged from ISP smelters have become progressively tighter. To meet them has required multi-stage treatment processes and frequently filtration of the effluent before discharge. The resulting costs of effluent treatment have focused attention on minimising the volume of effluent. At several smelters this is necessary in any case since payments to the authorities are required per unit of pollutant discharged.

SUMMARY

Since the operation of the first prototype furnace 50 years ago, the use of the Imperial Smelting Process has expanded considerably. In 1999, the thirteen operating furnaces produced over 1, 000,000 million tonnes of zinc and 429,900 tonnes of lead bullion. This amounted to about 12% and 7%, respectively, of the world's production of these metals.

The ability of the process to treat materials containing zinc and lead has remained its prime advantage. It consumes virtually all of the world's production of bulk concentrates. This production makes a significant contribution to the economics of mining zinc and lead ores. The flexibility of the process regarding feed materials has enabled it to become a substantial consumer of secondary materials containing zinc and lead. These are incorporated into the feed to the sinter plant, but are used also in the form of hot briquettes and are injected through the furnace tuyeres. Hot briquetting and tuyere injection offer the opportunity of using inexpensive raw materials and are particularly advantageous if the smelter is constrained by the capacity of the sinter plant.

The output of many ISP smelters has increased considerably since they were originally designed. Some of this increase has followed from increasing the size of the furnace but has also resulted from increasing the intensity of operation. This has required better sinter quality and improvements in the design of the furnace and condenser.

The ISP is being subjected to increasingly high standards for atmospheric and effluent discharges and hygienic conditions in the workplace. However, when considering its overall

466 LEAD-ZINC 2000

environmental impact, the process has the advantage that it consumes large quantities of waste from other metallurgical processes.

REFERENCES

1. M. Miyake, "Zinc-Lead Smelting at Hachinohe Smelter". Metallurgical Review of MMIJ, Vol. 12, No. 1,1995, 39-50.

2. M. Takewaki and H. Kubota, "Zinc-Lead Smelting and Refining at Sumitomo Harima Works", Ibid, 77-95.

3. R.W. Lee and D.J. Bishop, "Environmental Impact of the Imperial Smelting Process", World Zinc '93, I.G. Matthew, Ed., Australasian Inst. Mining Metall., Parkville, Australia, 1993,439-444.

4. M. Miyake, K. Kikuta, T. Oshita and F. Tanno, "Recent Operation at Hachinohe Smelter", Zinc & Lead '95. T. Azakami, N. Masuko, J.E. Dutrizac and E. Ozberk, Eds., Mining and Materials Processing Institute of Japan, Tokyo, Japan, 1995, 222-228.

5. P.J. Gabb and S.E. Woods, "Operation of Zinc Smelting Blast Furnaces", UK Patent. No. GB2197342B, 10 January 1990.

6. A.R. Musson, M.W. Gammon, I.M. Hitchens and S.P. Mathew, "The Development of Tuyere Injection Technology for the ISF at Britannia Zinc Ltd", Zinc and Lead Processing. J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton and P. Hancock, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 265-276.

7. M.W. Gammon, Private Communication, Britannia Zinc Ltd., 4 February 2000.

8. W.D. Schneider and B. Schwab, "Direct Zinc Recovery from Secondaries via the ISP Route", Zinc and Lead Processing. J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton and P. Hancock, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1998,229-242.

9. T. Oshita, Private Communication, Hachinohe Smelting Company, 31 January 2000.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 467

IMPURITY DISTRIBUTION IN THE ISP PROCESS AT THE HARIMA WORKS OF

SUMITOMO METAL MINING CO., LTD.

O. Kitamura and H. Kubota Harima Works

Sumitomo Metal Mining Co., Ltd. 364-4 Miyanishi, Harimacho, Kakogun, Hyogo, Japan

ABSTRACT

The Harima Works started its ISP operation in 1966 and been expanding its production capacity from 3,000 tonne Zn/month and 1,500 tonne Pb/month to 7,500 tonne Zn/month and 2,500 tonne Pb/month. It is not only treating zinc, lead and bulk concentrates but also secondary materials like Waelz kiln oxide from the Shisaka Works of Sumitomo Metal Mining, waste treatment sludges, industrial wastes and others. Its final products cover Prime Western Grade (PWG) zinc, Special High Grade (SHG) zinc, continuous galvanizing grade zinc, electrolytic lead, cadmium, indium, thallium and sulfuric acid. This paper briefly describes the smelter's historical and present operations. The key data for the current impurity control management and for the distribution of impurities such as copper, iron and arsenic are presented.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

468 LEAD-ZINC 2000

INTRODUCTION

The Harima Works started operation in 1966. The IS-technology, originally developed in the U.K., was used for the simultaneous smelting of zinc and lead. Based on the gained operating experience over 30 years, many improvements have been implemented in terms of process operation and mechamcal equipment, increasing production and efficiency. The progress in the production of zinc and lead is shown in Figures 1 and 2. In addition to the recovery of cadmium as a by-product, the production of 99.99% thallium was started in 1981 and that of 99.99% indium was commissioned in 1988. The progress of cadmium, thallium and indium production is shown in Figures 3, 4 and 5.

Figure 1 - Development of Furnace Zinc Production

Figure 2 - Production of Decopperized Bullion

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 469

Figure 3 - Refined Cadmium Production Figure 4 - Refined Thallium Production

Figure 5 - Refined Indium Production

PROCESS DESCRIPTION

Figure 6 shows the flowsheet of the Harima Works. Table I lists representative operating data for the different plant sections. The entire process consists mainly of a sintering process, a smelting process and a refining process. The major part of the zinc production is PWG zinc for hot-dip galvanizing. The PWG zinc production as shown in Table II is 6,900 t/month. The remaining 600 t/month of output is further purified and converted to zinc alloy by adding traces of alloying components such as aluminum, lead, etc. The zinc alloy is suitable for continuous zinc galvanizing and is usually supplied to the iron and steel industry. The lead produced is first decopperized and then cast into anodes. Further purification is conducted using Betts electrolysis. The purified lead is cast into 50-kg ingots, which are mainly shipped to battery makers.

470 LEAD-ZINC 2000

I Sintering machine

Furnace zinc

Zinc refluxer Ψ

Electric furnece

Φ COG zinc

T SO' MS

Lump, 5ΙΠΚΓ

~φ ψ \IT ISF

Casting hath

PWO zinc

Zinc briquette

Acid plant

Lead bullion

Refiner! learl

Snlnhiiric arid

| Lead electrolysis

Figure 6 - Process Flowsheet of the Harima Works

Table I- Typical Operating Conditions of the IS-process at the Harima Works

Sinter plant

ISF plant

Lead plant

Rich gas volume Plant availability SO2 gas content Sulfur burning rate

53,000 Nm3/h 97.5%

6.4% 1.6t/m2/day

Target blast rate Plant availability Coke consumption Overall condenser efficiency

34,000 Nm3/h 92.0%

950kg/tofzinc 91.0%

Zinc plant Number of columns SHG zinc production Butane gas consumption

2 columns 20 t/day 125 kg/t

Number of cells Current density Current efficiency Power consumption

62 cells 185 A/m2

97.0% 150kwh/t

Table Π- Assay of Impurities in PWG Zinc

Typical assay (%) Company standard (%)

Pb Cd Fe Sn Cu 1.05 0.05 0.020 0.025 0.010

*1.30 *0.10 *0.025 0.050 *0.030

As 0.002 ♦0.006

At the Harima Works, about 80% of the cadmium present in the raw material is recovered. In fact, about 75% of the cadmium present in the raw material is evaporated during the sintering process and is collected in the sinter dust. The sinter dust is treated with highly acidic process water to extract the cadmium for subsequent recovery in a distillation column. In addition, thallium and indium are extracted and separated from the cadmium prior to

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 471

distillation, and are individually recovered as pure products. The flowsheet for cadmium recovery is shown in Figure 7.

f Tl recovery

f In recovery

Sintering dust

±jt Leaching

Ύ

\LJI Stripping

Ψ \l/

Ψ CdCO,

w w Repulp

ψ Ψ

3 Ion-exchange |

Precipitation |

Cementaion |

Ψ Sponge cadmium

Ψ \l/ Melting

JL Distillation

JL Casting

Y Refined cadmium

Scrubbing water

NaCl

Na,CO,

H,S0 4

Zn powder

NaOH

Power

Figure 7 - Process Flowsheet of the Cadmium Recovery Circuit

472 LEAD-ZINC 2000

RAW MATERIAL PROCESSING

In addition to standard zinc and lead concentrates, the Harima Works also processes various mixed ores and secondary raw materials, which contain 10% to 20% lead. These feeds are considered difficult when the characteristic features of the ISP technology are utilized as the traditional zinc refining method. The secondary raw materials mainly consist of Waelz kiln oxide recovered from electric arc furnace dusts. The processing of the Waelz kiln oxide, obtained from the Shisaka Works of Sumitomo Metal Mining, started in 1977. The secondary raw materials also include zinc drosses from a hot-dip galvanizing process, zinc/lead-containing industrial wastes, precipitates from various wastewater treatment process, etc. Table ΠΙ illustrates the typical composition of the raw materials for the IS furnace at Harima. Figure 8 shows the history of the various raw materials processed based on their zinc content. Figure 9 shows the current distribution of zinc (top column) and lead (bottom column) in the different raw material groups. At the Harima Works we intend to permanently increase the production of zinc, and in parallel, to process more secondary raw materials until the limit of the desulphurization capacity in the sintering process is reached.

Table HI- Typical Assay of Treated Raw Materials

Zn Pb Chemical assay (%) S Fe Sn Cu Cd As

Zinc concentrate

Lead concentrate

Bulk concentrate

Secondary material

"aj

c

1 2 3 1 2 3 1 2 1 2 3

100,000

90,000 80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

0

49 .22 50.06 48.83

2.61 7.53

11.84 41.27 36.97 62.09 64.10 65.50

1.30 2.90 1.04

70.15 55.04 54.08 11.48 17.91 7.60 0.97 3.48

■ Sulfide Cone.

p™

■ ■ ■ u

™

■ ■ ■ L_T

30.96 11.52 30.10 8.67 29.59 10.63 17.75 5.40 22.94 10.90 19.96 5.20 25.24 5.52 26.85 7.64

0.30 4.50 0.44 6.68 0.01 2.15

0.01 0.01 0.01 0.03 0.01 0.01 0.01 0.01 0.12 0.10 0.01

0.16 0.21 1.02 0.26 0.17 0.06 0.79 0.74 0.17 0.08 0.27

Ξ Waelz Oxide D Other Secondary

-- ' 1 ■ ■ —II ■ I

"

I ■ _r u

0.19 0.17 0.11 0.04 0.17 0.06 0.11 0.08 0.02 0.03 0.01

0.08 0.01 0.04 0.03 0.71 0.04 0.12 0.09 0.01 0.01 0.01

Figure 8 - Zinc Recovery from Processed Raw Materials

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 473

Figure 9 - Constitution Ratio of the Raw Materials

INCREASED PROCESSING CAPACITY FOR IMPURITIES

The smelting process was modified in 1990 after 25 years of operation. The ISF shaft was completely changed and the cross-sectional area of the shaft was increased from 15.3 m2 to 19.4 m2. In 1996, the cross-sectional area of the furnace at the tuyere level was increased from 14.5 m2 to 16:6 m2 as part of a modification to increase the volume of process air in the furnace to suppress the impurity content of the furnace zinc. Figure 10 shows the relative increase of tin, copper and arsenic processed in 1992 and 1998, comparatively. With the higher production rate of zinc, the amount of impurities, and in particular the amount of tin, also increased. Because of a higher consumption rate of various kinds of secondary raw materials, the amounts of impurities will continue to increase in the future.

140

130

120

Ύ, no

100

90

80 1 1 1 1 Sn Cu Cd

V V

As

Figure 10 - Index of Impurity Input

474 LEAD-ZINC 2000

IMPURITY DISTRIBUTION

Tin

Long-term operating experience shows that the tin present in the raw materials does not evaporate in the sintering process and that most of it remains in the sinter, entering the ISP. The distribution ratio of tin in the ISF process is shown in Figure 11, and indicates that an average of 20% evaporates to the furnace gas and reports to the furnace zinc and eventually the PWG zinc. However, the evaporation ratio may rise to 30% with increasing shaft accretion or sudden changes in the composition, rate or uneven distribution of the furnace charge. Therefore, the shaft must be cleaned on a regular basis to remove the accretions formed and the amount of tin in the raw materials must be controlled in order to reduce the tin content in the product. At the Harima Works, because there is no extra process used to remove the tin present in the PWG zinc, special care is needed to control the tin content limiting it to 0.05% or lower in accordance with the specifications for PWG zinc. Tin present in lead is removed in the aeration process when the lead is melted in the anode casting kettle. As a result, the tin content of the lead anodes is maintained at 0.01% or lower. During electrolysis, tin precipitates together with the lead. Therefore, the electrolytic lead is subsequently melted and treated applying the Harris method before it is shipped as a product.

ISF slag 45

New Sn 100

Copper dross 25

Furnace zinc 20

De-copper lead 10

Figure 11 - Tin Distribution in the ISP Complex

Copper

In the ISP process, copper is one of the recovered impurities; it is present in the sinter at a content of 0.5-1.0%. As shown in Figure 12, the majority of the copper, like tin, remains in the sinter in the sintering process. In the smelting process, about 80% of the copper is dissolved in the lead bullion. In the decopperizing process, copper is removed from the bullion and is concentrated in the copper dross. As shown in Table IV, the solubility of copper in lead depends on the temperature of the lead. The copper present in the lead above the solubility limit may precipitate in different forms, but it mainly forms a matte. During actual production, an increased copper content in the raw material will cause the following problems: (1) The copper content in PWG zinc may increase. (2) The removal of the copper from the lead bullion may become more difficult. (3) Matte formation in the ISF may cause difficulties during slag tapping. At the Harima Works special attention is paid to problem (1). An increased copper

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 475

content in the furnace zinc is mainly attributed to the carry-over from the shaft caused by the blast air blown through the tuyeres. Therefore, the charge in the furnace must be controlled at a certain level. Moreover, the copper content in the sinter must be maintained below a certain level by means of raw material blending, etc.

m c ■■ NewCu ^ ^

ion L A ■V C %

Copper dross

ISF slag

De-copper lead

Furnace zinc

70

19

10

1

Figure 12 - Copper Distribution in the ISP Complex

Table IV -

Temperature (°C) 700 800 900 950

Copper Solubility in Pure Lead Solubility (%)

Pb Cu 98.0 2.0 95.5 4.5 92.5 7.5 86.0 14.0

Iron

As shown in Figure 13, most or all of the iron entering the ISP reports to the slag. However, the furnace zinc produced by the condenser contains 250-500 ppm of iron. This iron not only forms dross (hard zinc Zn-1% Fe) in the refining operation following the ISP process and thereby decreasing the yield of zinc, but also significantly affects the overall process. In 1990, the cross-sectional area of the shaft at the Harima Works was increased. At that time, no change was observed in the iron content of the furnace zinc. However, when the furnace hearth area was enlarged in 1996, the iron content in the furnace zinc increased and eventually affected the production capacity. As part of some corrective measures initiated, the declining angle of the charging bell gear was changed from 50° to 45°. Thus, more charge was loaded on the sidewalls of the furnace shaft. As a result, the iron content in the furnace zinc was reduced back to the original level shown in Figure 14. In addition, as shown in Figure 15, the ratio of CO/CO2 in the process gas was decreased as a result of this improvement. Although the exact details of that mechanism are still unclear, a theory can be provided which may explain both phenomena. During normal production, a part of the sinter loaded in the shaft is pulverized and carried over to the condenser. Part of the reduced iron present in the powder is mixed with the condensed lead and then is transferred into the furnace zinc during the cooling and separating process in the launder. By changing the bell angle, the sinter and coke were redistributed in the shaft. As a

476 LEAD-ZINC 2000

result, the degree of reduction in the top part of the shaft was decreased and the carry-over of reduced iron to the condenser was minimized. Consequently, the iron content in the furnace zinc decreased.

m c V* New Fe ^ ^ T 100 ΛΑ.

■V c tl

ISF slag 99.5

Copper dross 0.2

Furnace zinc 0.2

De-copper lead 0.1

0.0005

0 0.00045 c '" 0.0004 o c 0.00035

c 0.0003 >. 5> 0.00025 to

φ 0.0002

Figure 13- Iron Distribution in the ISP Complex

No. 17 campaign - o No. 18 campaign

c o B

1 B as

1 B

g E

CO

Number of operating months

§ B

Figure 14- Iron Assay in the Furnace Zinc

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 477

30,000 31,000 32,000 33,000 34,000 35,000

Average blast rate Nm/H

Figure 15- CO/C02 Ratio at the Stock-line Level

Arsenic

Arsenic, the most undesirable impurity, must be strictly controlled in terms of both the process operation and the product quality. For the process operation a large amount of arsenic may be combined with iron forming a speiss in the smelting process. The speiss not only causes difficulties during slag tapping but also may result in explosions during the water granulation process, representing a serious safety hazard around the furnace. According to user requirements, the arsenic content in PWG zinc must be controlled at equal to or less than 0.006% As for product quality reasons. In order to satisfy this stringent but self-imposed specification, an independently developed method is applied whereby molten sodium is injected into the metal to remove the arsenic from the furnace zinc. As shown in Figure 16, the furnace zinc contains about 10% of the total arsenic impurity content that is introduced into the process with the raw materials. The content, however, depends also on the reduction potential in the ISF. Figure 17 shows the relationship of the arsenic content in the furnace zinc to the ratio of FeO/Zn in the slag. Controlling the arsenic content in the sinter and adjusting the coke ratio used in the process can stabilize the arsenic content in the furnace zinc.

478 LEAD-ZINC 2000

m C m* New As ^ ^ f

100 L A ■■ c t§

ISF slag

Copper dross

Furnace zinc

De-copper lead

70

15

10

5

Figure 16- Arsenic Distribution in the ISP Complex

0.025

C? 0.02

.5 0.015 N

1 0.01 I .£ 0.005 >, a

0 en

<

y = -0.009X + 0.0584 R2 = 0.3969 (

('96/6 ~ '97/6) '

3.5 4.5 5 5.5

FeO/Zn ratio in ISF slag

Figure 17- Relationship Between the As-assay in the Furnace Zinc and the FeO/Zn Ratio in the ISF Slag

Table V lists the distribution ratios for tin, copper, arsenic and iron present as impurities in the PWG zinc, which is the major product of the Harima Works. Except for iron, the other impurities are expected to increase in the future. Therefore, in order to maintain the high quality of our PWG zinc, it is necessary to further strengthen the corresponding measures to control the impurity contents compared to existing competitive technologies.

Table V - Impurity Distribution in the Furnace Products

Furnace zinc De-copper bullion Copper dross ISF slag

Sn 20 25 10 45

Distribution (%) Cu Fe 1 0.2 10 0.1 70 0.2 19 99.5

As 10 5 15 70

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 479

CONCLUSIONS

The recycling of secondary materials as a raw material becomes more and more an important and dominant topic for the industry. The responsibility of zinc/lead smelting plants regarding the recycling of secondary materials is expected to increase in the future. Considering the fact that the Harima Works is one of a few coastal zinc/lead smelting plants located very close to metropolitan areas, we must make further contributions to the creation of a recycling society.

REFERENCES

1. C. Minami, K. Takei, T. Funahashi and H. Kubota, "Recovery of High Purity Thallium at Sumitomo Harima Works", Rare Metals '90. Kokura, Kitakyushu, 14-16 November, 1990,259-262.

2. H. Kadoya, Y. Kondo and H. Kawabata, "Improvements of Sinter Plant Operation at Sumitomo Harima Works", Zinc & Lead '95. T. Azakami, N. Masuko, J.E. Dutrizac and E. Ozberk, Eds., The Mining and Materials Processing Institute of Japan, Tokyo, Japan, 1995,213-221.

3. M. Mastuno, Y. Ojima and A. Kaikake, "Recent Development of EAF Dust Treatment Operation at Sumitomo", Zinc & Lead '95. T. Azakami, N. Masuko, J.E. Dutrizac and E. Ozberk, Eds., The Mining and Materials Processing Institute of Japan, Tokyo, Japan, 1995,432-441.

4. H. Kubota, T. Kusakabe, K. Takei and M. Takewaki, "Current Operation of Sumitomo Metal Mining's (SMM) Betts Lead Electro Refinery", Zinc and Lead Processing. J.E. Dutrizac, A.J. Gonzalez, G.L. Bolton and P. Hancock, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 353-366.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 481

OPTIMIZATION OF THE NEW JERSEY REFINING PROCESS

G. Hanko Department of Non-ferrous Metallurgy

University ofLeoben A-8700 Leoben, Austria

A. Lebleu Metaleurop Nord

1 rue Malfidano, BP1 62950 Noyelles-Godault, France

M. Sibony and J. Lecadet Metaleurop Recherche

1 ave Albert Einstein, BP120 78193 Trappes Cedex, France

ABSTRACT

The New Jersey distillation process is today the most important pyrometallurgical method for the production of zinc of very high purity. The flexibility of the pyrometallurgical zinc production process, especially its ability for recycling, makes it competitive, although it is energy intensive. A deeper knowledge of the New Jersey process is necessary to further increase its technical performance by ensuring its global safety. A project, integrated in the European research program EUREKA, has been started in order to define the process chemistry of distillation. Experimental measurements have been carried out with a column at the laboratory scale, and the results, linked with industrial measurements, support the development of a model based on material and thermal balances. The first conclusions indicate that the key parameters for process optimization are the column geometry and the distribution of the thermal losses.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

482 LEAD-ZINC 2000

INTRODUCTION

The New Jersey distillation process is the most important method for the pyrometallurgical production of zinc of high purity, although the process of distilling thermally produced zinc at atmospheric pressure at temperatures of 900°C to 1000°C is highly energy intensive. In 1998 Metaleurop started an Eureka project, supported by the French Ministry of Industry, with the objective of increasing the reliability and availability of the process by ensuring its global safety. The present study focuses particularly on the optimization of a cadmium column at Metaleurop Nord, Noyelles-Godault.

NEW JERSEY REFINING PROCESS

Basic Considerations

For the production of SHG zinc (special high grade, >99.995% Zn) the New Jersey Zinc Company designed a continuous two-step refining process consisting of parallel working lead and cadmium tray-columns. The processing steps depend on the impurites in the crude zinc, which commonly are iron, copper, lead and cadmium. The New Jersey refining process makes use of selective volatilization, selective condensation and of rectification. The selective volatilization and condensation are based on the differences in the boiling-point temperatures of the metals being separated. For rectification, countercurrent contact between the vapour and the liquid streams is decisive to enable intensive mass exchange.

A New Jersey refining unit consists of parallel working lead and cadmium columns. Each column has between 56 and 58 silicon carbide (SiC) trays, which are assembled in a way that the rising vapour and the descending molten metal experience a baffled flow. The outlets of the trays for the descending molten metal and the rising vapour are alternately arranged to establish conditions for rectification. Crude zinc is charged into the lead column where the less volatile elements such as iron, copper and lead are removed along with half of the zinc from the sump at the bottom of the column. Cadmium, the element with the lowest boiling point, and the remaining zinc are vaporised and condensed to molten metal. This distillate, containing about 0.3% - 0.9% cadmium, is fed into the cadmium column. The operation of the two columns is shown schematically in Figure 1.

Figure 1 - The Cadmium Circuit

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 483

Cadmium Column

The separation of zinc and cadmium occurs in the second processing step of the zinc refining process. Figure 2 describes a cadmium column by means of the geometric data of the cadmium column at Metaleurop Nord, Noyelles-Godault.

Pool of mcften Zinc

Figure 2 - The Cadmium Column

The 30 trays of the boiler are heated from the enclosing combustion chamber. The sump is attached to a horizontal outflow at the bottom of the boiler. The pre-heating section between the boiler and the feed section consists of 5 trays. The 19 trays following the feed section form the refluxing section; the condenser is situated next to the top of the column. The column is constructed of silicon carbide sections, sensitive to thermal cycling, thermal shock and mechanical vibration, which may lead to structural failure. To avoid thermal stresses on the columns, the feed rate and the feed temperature need to be kept constant. The distillate of the lead column is fed at an average temperature of 550°C into the cadmium column and is mixed with the descending molten metal of the refluxing section. In contact with the charged zinc, a part of the vapour arising from the boiler is refined by condensation. During its downward flow through the boiler, the molten metal is heated to volatilise the cadmium; a part of the zinc is also distilled. The rising vapour and the descending molten metal contact with each other and reflux refining occurs. The amount of boiling metal and its cadmium content increase with the height of the boiler. In the refluxing section, and also at the top of column, the downward flow of condensate causes the reflux refining effect. The resulting condensate is always richer in the higher boiling element zinc than the vapour from which the condensate originates. Thus, the vapour is enriched in cadmium until it reaches the condenser.

484 LEAD-ZINC 2000

MODELLING OF THE CADMIUM COLUMN IN THE NEW JERSEY PROCESS

Theoretical Considerations

A method to predict the quality of the refined zinc as a result of changing processing conditions is of great interest in practice. Only by theoretical calculations can the refining units be simulated, proving the operating parameters. The physico-chemical process occurring in the cadmium column allows the application of only the basic knowledge of distillation and rectification. The conditions for rectification are not completely fulfilled because equilibrium between the liquid and the vapour phases is not established in each tray.

Krupkowski and Fik (1) found a first mathematical model taking into account the following simplifying assumptions:

• Boiling and condensation can be regarded as two separate processes, which do not interfere each other.

• In the boiler the liquid phase is in equilibrium with its own vapour; no mass exchange with the vapour rising from the tray below occurs.

• In the refluxing section, the vapour is in equilibrium with its own condensate; no mass exchange with the molten metal descending from the tray above occurs.

• Because of the high turbulence of the two phases, no demixing happens.

The thesis of "parallel flux" describes the refining process as a fractionated distillation operation with subsequent fractionated condensation of the vapours; rectification is excluded by this model, which is illustrated in Figure 3.

Pre-heating Zone

Figure 3 - Model of Krupkowski and Fik

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 485

Krupkowski and Fik (1) found relationships for isothermal distillation and isothermal condensation. A calculation of the cadmium content of the distillate and of the refined zinc is possible; the feeding rate Gi, the reflux G2 and the outflow of the sump G3 are known from the mass-balance; the upward flow rate G4 of vapour entering the refluxing section and the flow rate G5 of distillate have to be calculated by using the thermal balance. This method of predicting the quality of the refined zinc is proved in practice and satisfys several sets of operating conditions.

Kamegai et al. (2) simulated the reflux refining process in the cadmium column by a model which is based on the material and thermal balances, and on gas-liquid equilibrium relationships. Therefore, the productivity and the compositions of the high purity zinc product and of the zinc-cadmium alloy could be estimated. To establish this model, the column is divided into four main sections: the boiler, the feed section, the refluxing section and the shaft condenser. For the boiler, the refluxing section and also for the shaft condenser, similar relations were used for the model. In the feeding section, the vapour of the boiler is in contact with the low temperature zinc feed and condenses. Therefore, the equations for the thermal balance and for the gas-liquid equilibrium relationship differ from the refluxing and the boiler sections. By connecting the process models for the shaft condenser, the refluxing, the feed and for the boiler sections the model is realised. This simulation program enables the calculation of the flow rates, of the Cd concentration of the molten metal, and of the vapour in each tray in the column. Thus the productivity and the Cd concentration of the SHG zinc can be computed.

New Mathematical Model of a Cadmium Column

Among design studies, experimental measurements support the development of a mathematical distillation model, which should mark a progress point in simulating the operating conditions. An improved New Jersey simulation should support the identification of the optimal operating conditions for industrial columns and also facilitate investigations on the geometry of the column. All the efforts for increasing the productivity of the New Jersey process, for improving the quality of the SHG zinc and for enriching the Zn-Cd alloy in cadmium are carried out in consideration of decreasing the operating costs. Krupkowski's thesis of "parallel flux" (1) excludes rectification in a cadmium column. The geometry and the number of trays is not considered in this model. Kamegai et al. (2) found a model, which calculates the flow rate and the composition of the gas and liquid phases. The equilibrium concept for determining the countercurrent separation efficiency is only a very rough estimation. Reflux refining efficiencies are introduced as a correction factor for the deviation from equilibrium.

The principles of the simulation program are presented in the following five-tray column example (Figure 4). The boiling and refluxing sections consist of two trays; the feeding section is situated in the middle of the column. The pre-heating zone (shown in Figure 2) is considered as a part of the refluxing section. All essential equations get explained by this example.

486 LEAD-ZINC 2000

Qp4

QP3

Qch2

Qchl

D :YD

£ ■®

V3IY3 "6—

U VXJ

VI !Y , ™ e —

®

® ©

Φ / Ä , XR

V4IY4 — s

Z.4 x, F,T,

V2 Y2 #

Ü&

QPS

Qp4

Qch2

Qch!

Figure 4 - Five-Tray Column

The simulation program of the distillation process must fulfill global relationships as well as tray-specific equations. Distillation can be predicted in the simplest way with the help of global material and thermal balances.

F = D + R

•F = y D D + x R R

QP=Qch •ΔΫΟ heating -Δ(2ν

(1)

(2)

(3)

In industrial plants, generally all essential heat inputs for calculating the absorbed heat QCh are recorded, so that the radiated heat Qp can be estimated by equation (3). The effective heat being absorbed or radiated by each tray is a function of the heat gradient. The circumstance of high heat radiation at the top of the column is also considered by this model.

As the vapour may not be in equilibrium with the molten metal in the tray, the Murphree efficiency η [-] is introduced as the ratio between the maximum possible change of the vapour composition under equilibrium and the actual change. The distillation coefficient d [-] describes the relation of the content between the gaseous and the liquid phases.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 487

Figure 5 - Murphree Efficiency

η = Yn Yn-1

Yn _ Yn-I

y*n=d-x„

(4)

(5)

The stage concept of the simulation program is described, as an example, by means of the feeding tray.

L4, XJ V3 / Y3

F ,TF ,XF

V2 , Y2 L3, X3

Figure 6 - Material and Heat Flow at the Feeding Tray

Mass balance: V2 + L4 + F = V3 + L3

Cd-mass balance: V2 · y2 + L4 · x4 + F · xF = V3 · y3 + L3 ■ x3

Equilibrium relationship: y * = d · x 3

v — v Murphree efficiency: η = —; -

y 3 - y 2

Thermal balance Qp3 = ΔΗ vap · (V2 - V3) - F · cp · (1180 - TF)

(6)

(7)

(8)

(9)

(10)

In the thermal balance ΔΗν3ρ [J/mol] is the latent heat of zinc vaporisation and cp [J/mol-K] is the specific heat of molten zinc. This calculation program determines a specific heating profile as well as measured values of the heat losses at each tray of the upper part of the column.

488 LEAD-ZINC 2000

Application of the Simulation Program

Operating data provided from the Zn refining department in Noyelles Godault enabled the comparison of the calculated results with industrially obtained values. Before carrying out a widespread operation study of the simulation program, the Murphree coefficients ηΓ and nb for the refluxing and the boiling sections must be calculated. An iterative computation sets the Murphee coefficients on ηΓ= 0.025 and r\b= 0.105. It should be noted that these tray efficiencies are very low, especially for those trays located in the refluxing zone.

The average operating conditions of Noyelles Godault applied for the simulation are shown in Table I.

Table I - Average Operating Conditions of Noyelles Godault

Cd-concentration of fed zinc [%]

Temperature of fed zinc [°C] Heat absorption of the boiler [kW]

Heat radiation of the section above the boiler

[kW]

Murphree efficiency r|b

Murphree efficiency ηΓ

0.70

541

933

713

0.105

0.025

In Table II, the calculated results concerning the productivity of the column and the Cd distribution are compared with the average processing data of Noyelles Godault.

Table II - Comparison of the Calculated Results with the Distillation Performance of Noyelles Godault

Processing of Noyelles Godault Calculated Results

Flow rate of distillate [t/d] 3.4 3.4 Cd concentration of the distillate [%] 16.1 16.3 Cd concentration of the refined zinc[g/t] 28 24

The calculated results correspond with the distillation performance of Noyelles Godault quite well. Representative results of the tray by tray calculation are shown in Figure 7 and in Figure 8. Figure 7 shows the flow rates of the vapour and the molten metal at each tray; the Cd concentration of the phases is given in Figure 8.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 489

22 B

20 Ξ

18 S

16 E

:E 'am)

60 80.

flow rate [t/d]

-.— ........ . —...... -_. 1E-02 1E-01 1E+00

% Cd [mass*]

D%Cdl |«%Cdv

._...! . .. 1E+01

Figure 7 - Flow Rate of Vapour and Molten Metal

Figure 8 - Cd Concentration of the Vapour and Molten Metal

In the refluxing section, the zinc condenses at the bottom of the trays and the condensate flows back into the boiling section. Because of the condensation of the zinc portion, the flow rate of vapour decreases with the height of the boiling section. The decrease of the vapour flow rate results in an enrichment of the Cd content of the gaseous phase.

Study of the Heat Distribution

The influence of heat absorption in the boiler and of heat radiation in the refluxing section on the refining-distillation performance is of great interest, not only with regard to the quality and quantity of the output but also with regard to the energy consumption and the service life of the column. For the parametric study, the Murphree coefficients ηΓ and r|b were regarded as constants. The heat contribution was changed in a way that Qp was set at a fixed value whereas the value of QCh was varied. To obtain an impression of the actual process values, the average operating data of Noyelles Godault are presented in Figures 9-12 which describe the influence of the heat profile on the flow rates of distillate and residue, respectively.

490 LEAD-ZINC 2000

14

12

»10

=5 8 ο Ο 6 0) D

*-»

2

84

. — - Qp=700kW

Qp=740kW

Qp=780kW

Qp=820kW

. ■

-

-

Nnvelles-Godault: Qch = 933 kW Qp = 713kW

D = 3.4 t/d

*1

y y

S" . . · " '

,--·' - - ■ ' ■

.-" .-* y

s ' '

S*

,

*s '

y

*^'

Qp=700kW

^ ·'' Qp=740kW

s- ' ^ ..••'Qp=780kW y ^y

•^ y ^ .-* ^ ^Qp=820kW

^ ^ y "

^ S ^ y '

> ^y ' ' ^ " " ^ ^ ^

.

Qch = 1000 kW Qp = 780 kW

D = 3.7 t/d

850 900 950 1000

absorbed heat Qch [kW]

1050 1100

Figure 9 - Influence of the Heat Distribution on the Flow Rate of the Distillate

80

78

5 · 76 g, 0 74 N

"g 72 c «= £ 7 0 o CC 68 a 2 66

64

62

60

Qp=740kW Qp=780kW ~^Qp=820kW

Qp=700kW

— ■ Qp=700kW

Qp=740kW

Qp=780kW

Qp=820kW

850 900 950 1000

absorbed heat Qc h [kW]

1050 1100

Figure 10 - Influence of the Heat Distribution on the Flow Rate of the Refined Zinc

Figure 9 and Figure 10 show a common tendency; a rise of the boiler temperature increases the outflow rate of distillate by a simultaneous reduction of the amount of refined zinc. The reason is the constancy of the material balance, which is consequently fulfilled for each heat contribution. The minimum heat input QCh for certain heat losses Qp is indicated as the intersection of the flow rate profile with the abscissa of Figure 9.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 491

■e 'S

1

100

90

i, 80

! 70

60

50

40

30

20

10

\ ■ \

ι·. § ■ ι> \ ii

a . ° \

\ \

\ \

* Qch = 933 kW Qp = 713kW

%CdD = 16.1 %

,

o \

II \

σ

\

l

\ — - Qp=700kW

£ Qp=740kW '

S'> : Qp=780kW ! 00 \ :

k § \ Qp=820kW ;

\ \

\ \

\ \ \ vv

950 1000

absorbed heat Qc [kW]

Figure 11 - Influence of the Heat Distribution on the Cd Content of the Distillate

o

— - Qp=700kW

Qp=740kW

Qp=780kW

Qp=820kW

Novelles-Cmlauli:' Qch = 933 kW Qp = 713kW

%CdR = 2.8 E-03 %

950 1000

absorbed heat Qch [kW]

Figure 12 - Influence of the Heat Distribution on the Cd-Content of the Refined Zinc

The absorbed heat Qch at a cadmium content of 100 % (shown in Figure 11) indicates the limiting value for a given heat loss. Because of the steep decrease of the Cd content profile, the range of high Cd output in the distillate is very narrow; the Cd content of 100 % is reduced to a Cd output of 10 % with an increase of QCh of 120 kW. The graphs of Figure 11 and Figure 12 reflect a contrary trend. A low processing temperature in the boiler improves the distillation with respect to high Cd contents in the distillate. However, with regard to low Cd contents in the refined zinc, low processing temperatures debase the refining effect.

It is a matter of optimizing the heat ratio between the boiling and refluxing sections. The temperature should be just high enough to vaporise almost all the cadmium of the zinc-cadmium alloy without vaporising too much zinc. If the temperature of the vapour is just above

492 LEAD-ZINC 2000

the boiling temperature of zinc, a gentle cooling of the refluxing section enables the condensation of the zinc. A gentle cooling of a non-superheated vapour could be realised in a low refluxing section without stressing the silicon carbide sections of the column.

EXPERIMENTAL MEASUREMENTS

In order to investigate more accurately the possible process optimization of an industrial column, a laboratory scale graphite column was constructed. The laboratory scale cadmium column is assembled from 13 graphite trays. The column can be divided into four main sections:

• The boiler section consists of 6 trays and includes the bottom tray with a horizontal outflow.

• The feeding section is connected with the feeding box. By aid of the feed regulation system, situated at the feeding tube, a constant feed rate can be established.

• The refluxing section consists also of 6 trays. The top of the column represents the connection with the condenser.

• The condenser is fixed next to the top of the column.

Description of the Trial

The experimental arrangement, shown Figure 13, consists of the model column, the graphite protecting tube, the steel protecting tube, nitrogen flush devices, the two zone furnace and a small little melting furnace. All exterior parts of the model column (condenser, inflow and outflow units) were supplied with additional electric heating elements to maintain the zinc in a liquid state. Thermocouples placed at the top, at the bottom and in the middle of the column indicated the actual temperature in these sections (Ttop, Toiler and Tmidd|e).

T— W J - ,

feedngt>ay7

*

; -

steel protecting tube

graphite protecting tube

VV^aJS"·

it'

Figure 13 - Experimental Arrangement

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 493

The graphite column was put in the furnace in such a way that the boiling section was enclosed by the first heating zone. During the trial, a nitrogen gas flow between the graphite and the steel protecting tube and between the steel protecting 'tube and the furnace, avoided the oxidation of the graphite. The designs for the refluxing trays and the boiler trays are given in Figure 14 and Figure 15, respectively. Figure 16 shows the actual graphite trays used in the laboratory column.

S £

nun M1gQx3

1 S>

^ I s

_225_

^ -22.

mmflm

8 9 5 MlgQ»3

_enfl_

Figure 14 - Design of the Refluxing Trays

4120x3-

Figure 15 - Design of the Boiler Trays

Figure 16 - Graphite Trays Used for the Laboratory Column

494 LEAD-ZINC 2000

A large amount of molten metal in the feeding box provided a constant inflow into the column. The cadmium concentration of the zinc feed was 0.4% + 0.07 % for all experiments. The temperature of the zinc feed was set at 550°C. The temperature distribution was set at Ttop=850oC, Tmiddie=900°C and Tboiier=920°. Samples of the condensate fixed on the trays of the column enabled a measurement of the Murphree efficiency after an experiment. The refining effect also occurs in the boiling section. Because of the charge of cold zinc feed, the vapour condenses and the content of the less volatile element zinc is decreased in the vapour.

The main operating conditions of the experiments are listed in Table III.

Table III - Comparison of the Experimental Results

_ „ , Measured Murphree-Power Output Feed Rate _,... .

Efficiency

Furnaceb Furnace,· ; r|boiier η reflux

Continuous Trial 1 2822 W 433 W 2.2kg/h 0.10 0.075

Continuous Trial 2 3296 W 657 W 4.5 kg/h 0.10 0.080

VERIFICATION OF THE EXPERIMENTAL MEASUREMENTS BY THE MATHEMATICAL MODEL

The aim of the following calculationsis the application of the mathematical model to the experiments which were carried out. The design of the column is incorporated into the model. The trial was realised with a two zone furnace, which heats the boiler and the refluxing sections. This experimental arrangement differs from the industrial performance of a refining distillation column. A detailed heat balance enables the simulation of the refining process. Because of the zinc feeding, the outflow of refined zinc and of distillate, the radiation at the top of the column, the outward heat conduction through the furnace wall, the cooling effect of the nitrogen flow and waste gas heat losses (3) occur.

The mathematical model was fitted to each experimental arrangement. For the simulation, the experimentally determined data such as the feed rate and the Murphree coefficient were used. The comparison of the experimental profiles with those which were computed demonstrates the effectiveness of the model, as is illustrated in Table IV.

Table IV - Simulation Data

Trial 1

Trial 2

Radiated Heat

1.64 kW

1.87 kW

Absorbed Heat

1.35 kW

1.50 kW

Results of Experiment

% Cd in Distillate

2.4

3.5

Results of Calculation

% Cd in Distillate

2.4

3.6

The following graphs (Figure 17 and Figure 18) show a comparison of the calculated and measured Cd contents of each tray for the first trial.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 495

'S 7 feeding tray

«•■exp.%Cdl ! —— calc.%Cdl

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5

% C d l

Figure 17 - Comparison of the Calculated and Experimentally Determined Liquid Cadmium

Content

Figure 18 - Comparison of the Calculated and Experimentally Determined Gaseous

Cadmium Content

The observed good accordance suggests that the experiment arrangement can be considered as a reliable tool to test different configurations of the column. Future trials will focus on the geometry of the column, taking into account the low efficiencies experimentally found, by studying the influence of the tray number in the heating and reflux zones, by modifying the tray shape and by changing the thermal profile of the heating zone.

CONCLUSIONS

The mathematical model was successfully used for the computation of the performance of the industrial cadmium column at Metaleurop Nord, Noyelles Godault and also for the trials on a laboratory scale. The Murphree efficiency and also the heat contribution along the height of the column are very sensitive parameters for the simulation program.

The results of the trials at the laboratory scale are promising in terms of either geometrical or process modifications to increase the safety of the distillation units, because the process is purely thermally controlled ("parallel flux" model) with a low tray efficiency. Further trials will investigate the possibility of decreasing the number of trays used in the refluxing

496 LEAD-ZINC 2000

zone, or modifying their geometry. This study is linked with the material optimization part of the "Eureka" project, to define solutions for increasing the global safety of the refining process.

NOMENCLATURE

L„ flow rate of molten metal leaving the tray n [mol/s] Vn flow rate of vapour leaving the tray n [mol/s] xn Cd mole fraction of molten metal of tray n [-] y„ Cd mole fraction or vapour of tray n [-] y* Cd mole fraction of the vapour equilibrated with the molten metal in tray n [-] D flow rate of distillate [mol/s] F feeding rate of molten zinc [mol/s] R outflow of refined zinc [mol/s]

Qch heat absorbed by the boiling section [kW] Qp heat radiated by the refluxing section [kW] d distillation-coefficient [-] η Murphree efficiency [-]

AHvap latent heat of vaporisation of zinc [J/mol-K] cp specific heat of molten Zn [J/mol-K]

^r'Qheai heat for heating the zinc feed material [kW]

AQvap heat for vaporisation [kW]

REFERENCES

1. W. Rudorff, "Über die Trennung von Zink and Cadmium nach dem New Jersey-Raffinationsverfahren", Erzmetall. Vol.15, 1992, 557 - 612.

2. T. Kamegai, H. Fukuyama, T. Fujisawa and C. Yamauchi, "Modeling of Column in New Jersey Reflux Refining Process for High Purity Zinc Production", Zinc & Lead '95. T. Azakami, N. Masuko, J.E. Dutrizac and E. Ozberk, Eds., The Mining and Materials Processing Institute of Japan, Tokyo, Japan, 1995, 90-99.

3. W. Heiligenstaedt, Wärmetechnische Rechnungen für Industrieöfen. 4. Auflage, Verlag Stahleisen M.B.H., Düsseldorf, 1966.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 497

A STUDY OF THE CHANGES IN THE PERMEABILITY OF THE SINTERING BED IN THE IMPERIAL SMELTING PROCESS

K. Kawanaka Niihama Research Laboratories Sumitomo Metal Mining Co., Ltd.

Niihama, Japan

Y.Mori Hyuga Smelting Co., Ltd.

Hyuga, Japan

ABSTRACT

The permeability changes during the sintering of the feed for the Imperial Smelting Process (ISP) were investigated using an experimental sintering pot, as well as with a mathematical model. The results obtained were as follows. The blowing resistance in the sintering bed changes as the sintering reaction proceeds. It increases at the beginning of the reaction when the temperature of the bed is still low. Then it decreases as the reaction continues, accompanied by an increase of the bed temperature. The increase of the blowing resistance in the bed during the sintering process mainly occurred in the drying zone, and was comparatively small in the reaction and cooling zones. Based on Erugen's equation, the change of the permeability during sintering is explained by the decrease of the apparent fractional void volume in the drying zone of the bed.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

498 LEAD-ZINC 2000

INTRODUCTION

The productivity of the sintering process is usually expressed with two indices; namely, the amount of sulfur removed and sinter production. One of the most important factors closely related to these indices is the permeability in the sintering bed.

The raw material for the sintering process is pellets made of sulfide concentrates and recycled sinter fines. The permeability depends considerably on the properties of the pellets. However, the properties and the permeability of the pellet bed change throughout the oxidation reaction in the sintering machine. The blowing resistance normally shows a considerable temporary increase during the sintering process. This would result in a decrease in the extent of sulfur removal and sinter production. Therefore, in order to keep the blowing resistance low enough to avoid this problem, about four times as much sinter fines (2 to 6 mm) as sulfide concentrate is recycled to the pelletizing step. If the blowing resistance of the bed can be kept low, the ratio of sulfide concentrate to recycled fines in the raw material could be increased and the productivity improved. However, the reasons for the changes in the permeability of the bed in the sintering process of the ISP have hardly been examined in detail so far. Accordingly, the authors investigated the reasons for the changes in the permeability by using an experimental sintering pot as well as a mathematical model. This paper describes the results.

THE TUBE EXPERIMENTS

The Experiment Method and Apparatus

An initial experiment using a 70-mm inside diameter stainless steel pipe as a reaction vessel was carried out in order to observe the change in the permeability of the bed caused by the progress of the sintering reaction.

In commercial sintering machines, the supply of heat from the outside is used only for the formation of the ignition layer. However, in a small-scale experiment most of the reaction heat generated is lost from the surface of the reaction tube, and the temperature does not rise enough to simulate the actual sintering process. On the other hand, it is expected that the change in the permeability of the bed observed in the sintering process is closely related to the changing temperature of the bed. Therefore, in the experiments it is necessary for the bed temperature to reach the same level as in commercial operations. In the present experiments the reaction pipe was heated in an electric furnace, and the outer wall of the pipe was kept at 800°C.

A schematic diagram of the experimental apparatus is shown in Figure 1. Raw material was mixed in the same ratio as for the commercial operation as shown in Table I, and was then pelletized by a 500-mm inside diameter pan pelletizer.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 499

Figure 1 - Schematic Diagram of the Tube Test Apparatus

Table I - Blending Ratio and Composition of the Samples

Recycled sinter fines Zinc sulfide concentrate Lead sulfide concentrate

Others

Blending Ratio

(%) 79.4 10.3 5.6 4.7

Zn 41.8 50.6 7.2

61.7

Composition (mass%) Pb

20.9 3.7

58.0 3.8

S 2.7 31.5 21.6 0.4

Fe 7.7 8.5 9.2 8.2

CaO 4.0 0.5 0.1 2.2

S1O2 2.5 2.9 0.4 2.7

The reaction pipe was heated to 800°C in advance. Raw pellets were charged into the preheated pipe to form a 40-mm thick bed and were kept for 30 minutes with a N2 gas flow, which was introduced into the pipe from the lower end. This allowed sufficient drying and preheating time of the bed for ignition. After that, additional raw pellets were charged on the ignition layer making the total bed thickness 300 mm. The N2 gas was then stopped and air was supplied at a constant rate of 501/min. The gas velocity, based on an empty column, was kept at about 0.2 m/s.

The air supply was cut off just after the temperature of the highest layer of the bed reached its maximum point and began to decrease. The blowing time was about 20 minutes and the blowing pressure was measured with a manometer every minute. A change in temperature inside the bed was measured by six sets of thermocouples (R) placed in the center of the reaction vessel.

500 LEAD-ZINC 2000

Results and Discussion

Change in the Bed Temperature

Changes in the temperature at each position in the bed are shown in Figure 2. The temperature in the ignition layer at the height of 25 mm decreased to about 400°C at the beginning of blowing and then increased to exceed 1150°C three minutes after blowing started. After it reached its maximum, cooling proceeded promptly and the temperature finally became lower than 300°C. As the outer wall of the reaction pipe was kept at 800°C, any additional rise in temperature was attributed to the endothermic reaction of the sulfide concentrate. Temperatures in the layers above 75 mm showed stagnation around 100°C just after the blowing started and then increased slowly followed by a subsequent rapid rise (1).

Figure 2 - Changes in the Bed Temperature

The early stagnation corresponds to the drying of the raw material. The subsequent rapid rise in temperature is attributed to the reaction of the sulfide. The temperature profile in the bed indicated that the drying and oxidizing zones moved from the bottom to the top as the reaction proceeded. Such temperature changes observed in these experiments were confirmed to be very similar to those in a commercial sintering bed (1), suggesting that the actual sintering process had been well simulated during the present experiments. Furthermore, the sulfur content of the experimental product shown in Table Π is low and at the same level as that of the commercial sinter. This also shows that the oxidation reaction is these experiments reaches the same level as in commercial operations.

Table Π - Sulfur Content in the Sinter S (mass%)

Present Experiments 0.2 ~ 0.6 Commercial Operation (HARDvIA) 0.5 ~ 0.7

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 501

Change in the Blowing Pressure

An example of the blowing pressure change during the sintering experiment is shown in Figure 3. The average temperature of the six points in the bed is also plotted in Figure 3 to examine the relation between the permeability and the average temperature of the bed.

1000

800

600

400

200

0

u o

3 sx E H

Figure 3 - Changes in Blowing Pressure with Time Obtained Using the Tube Test

The blowing pressure increased rapidly to 7.8 kPa about 3 minutes after blowing started, then it gradually decreased. It finally reached a stable value of about 2 kPa. In the early stages when the blowing pressure increased dramatically, more than half of the bed remained at room temperature and the average temperature did not even reach 350°C. In the latter stage when the bed temperature increased rapidly, the blowing pressure decreased in spite of the rise of the actual gas velocity. Since the amount of supplied air was kept constant, it is thought that the change in the blowing pressure is attributable to the change in the permeability of the bed. The blowing resistance increased rapidly when the temperature of the bed was comparatively low, and it decreased when the temperature increased. Therefore, the increase of the blowing resistance in the sintering bed of the ISP occurs in the low temperature zone, not in the high temperature zone. In the high temperature zone some materials having a low melting point might become molten, but this is not the main cause of the increased resistance as reported (2).

THE POT EXPERIMENTS

The Experimental Method and Apparatus

In order to determine the position in which the greatest increase of blowing resistance occurs, a large reaction pot was used. The pot was large enough for the continuous sintering reaction to occur, without the need for a supply of heat from the outer wall. A schematic diagram of the apparatus is shown in Figure 4. The reaction vessel was a steel cylinder in which a pellet bed having an outside diameter 260 mm and a height of 300 mm could be held. Five holes were made in the sidewall of the pot in a spiral arrangement for the inserting manometers and thermocouples used to measure the pressure gradient and temperature profiles in the bed. During the sintering

502 LEAD-ZINC 2000

operation, air was supplied through the wind box (W/B) under the reaction vessel and travelled upward through the bed. The pressure in the W/B was regarded as the total blowing pressure. The W/B was connected to an exhaust fan, and at the time of ignition, gas was absorbed downward from the top surface of the bed by the valve operation.

Figure 4 - Schematic Diagram of the Pot Test Apparatus

During the experiment, raw material was mixed in the same ratio as for the commercial operation as shown in Table I and was pelletized by a concrete blender (500 mm diameter by 530 mm length) whilst adding water. Raw pellets were charged into the pot to form an ignition layer of 40 mm thickness. Heating and ignition were carried out with liquid propane gas burners from the top of the layer. Additional pellets were charged to make the total bed thickness 300 mm, and air was blown from the bottom for specific periods of time. The blowing rate was 38.4 Nm3/h so that the gas velocity based on the empty column would be approximately 0.2 m/s.

Results and Discussion

The Temperature Change and Blowing Resistance Increase Inside the Bed

The change in temperature measured at each position inside the bed during the blowing is shown in Figure 5. In the pot experiment, heat was not supplied from the outside except for ignition. Therefore, the change in temperature mainly depended on the oxidation reaction and the heat transfer inside the system (2).

After blowing was started, the temperature increase and subsequent cooling occurred in the same way as in the small-scale test, in sequence from the lower layer to the upper part of the bed. The highest temperature exceeding 1300°C was observed at a height of 150 mm.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 503

Figure 5 - Changes in the Bed Temperature with Time Obtained by the Pot Test

Figure 6 shows the changes in the pressure loss, in each 50-mm layer of the bed. The loss increased in sequence from the bottom of the bed upwards. The maximum loss in the entire bed was about 5 kPa, which was observed 8 minutes after blowing started.

Figure 6 - Changes Pressure Loss with Time Obtained by the Pot Test

Profiles of the Temperature and the Pressure Loss in the Bed

The profiles of the temperature and pressure loss per unit bed height (ΔΡ/L) at 5, 7, and 9 minutes after the blowing started are shown in Figures 7 (a) - (c). The heat wave, the peak of the temperature profile curve corresponding to the most actively reacting zone, moved gradually from the bottom to the top as the reaction proceeded.

As shown in Figure 7 (a) the temperature at a height of 50 mm already exceeded 1000°C five minutes after the blowing started, while ΔΡ/L at the same point and time showed only a slight increase. It was the layer of comparatively low temperature in the forward side of the heat wave that showed the highest ΔΡ/L value. The same phenomena can be seen in Figures 7 (b) and 7 (c). This shows that the layer where the ΔΡ/L is maximum moves upward gradually with the ascent of the heat wave.

504 LEAD-ZINC 2000

ü o k. 3

4- · n o a. E .«

50 100 150 200 250 Height Above Grate Level (mm)

ι 50 100 150 200 250

Height Above Grate Level (mm)

E E

■ «

0 .

CL

<

O

!

50 100 150 200 250

Height Above Grate Level (mm)

Figure 7 - Pressure Loss and Temperature Profiles in the Sintering Bed at 5, 7, and 9 Minutes after the Start of Blowing

It is reported that, during the sintering of iron ore, ΔΡ/L reaches the maximum value in a layer of around 1000°C on the forward side of the heat wave (3). The combustion reaction of coke which is added as a heat source is considered to progress vigorously in this temperature zone (4). In

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 505

the present study, on the other hand, the highest ΔΡ/L was observed in the bed having a considerably lower temperature of 100 to 200°C. Since it is not expected that the sulfide concentrate used in the ISP would react rapidly at such a low temperature, the mechanism of the increase of the blowing resistance during ISP sintering is completely different from that of iron ore sintering.

Re-condensation of moisture, once vaporized in a lower drying zone, could be a cause of the blowing resistance increase in the low temperature zone (3). The temperature at which re-condensation occurs is lower than the temperature at which the ΔΡ/L increased in the present experiments. Moreover, no increase in the moisture content was detected in samples taken from each height position in the sinter cake. These sinter samples were obtained by stopping the blowing halfway through a sintering experiment. This suggests that the re-condensation of moisture does not cause a rise in the blowing resistance in the sintering operation of the ISP.

The layer of 100 to 200°C above the heat wave in the sintering part of the ISP is characterized as a zone where the pellets are dried and heated by the hot gas generated in the lower layer. Though the bed of this temperature exists on both sides of the heat wave, the drastic pressure loss was observed only in the bed on the forward side of the heat wave. Since the drying of the pellets proceeded in this area, the increase of the blowing resistance could be attributed to the re-powdering of the pellets accompanying their drying. Blocking of the air paths by fine particles generated by the re-powdering of the pellets accompanying their drying is considered to be a cause of the increased resistance in this region. Therefore, in order to minimize the resistance increase in the sintering bed and to improve the productivity, it is very important to produce pellets which do not significantly re-powder, even when they lose moisture in the drying and heating stages.

ANALYSES WITH THE MATHEMATICAL MODEL

Basic Equations

The extent to which the change of the blowing resistance could be explained by supposing the re-powdering of the pellets during sintering was examined. The factors influencing the change in the permeability of the bed were analyzed using a mathematical model capable of simulating the pressure loss in the bed.

Erugun's equation was also used in this study, which is widely employed in the analysis of iron ore sintering (5,6). Erugun's equation is as follows:

ΔΡ/L · gc = 150(1-ε)2/(ε3 · φ2) · (μ · Uo)/dp2 + 1.75(1-ε)/(ε3 · φ) · (G · Uo)/dp (1)

where: ΔΡ = pressure loss (Pa) L = layer height (bed height) gc = gravitational constant (kg«m/h2) γ = fractional void volume in bed (-) φ = shape factor (-) μ = viscosity of gas (kg/m»h) G = mass flow rate of gas (kg/m2«h) Uo= superficial gas velocity based on empty column (m/h) dp = particle diameter (m)

506 LEAD-ZINC 2000

The Classification of the Layer

The bed was classified into five zones according to the temperature and the position. The layer classification is shown in Table ΠΙ.

Table III - Character of the Zones

Zones

Row Drying Heating Reaction Cooling

Temperature (°C) RT

140>T>RT 600>T>140

Tmax > T > 600 Tmax>T

Forward/Backward of Heat Wave

F F F F B

Re-Condensation of Moisture

Re-condensation of moisture was not considered in this model because the existence of a layer in which water re-condensed during sintering had not been confirmed in the experiments.

Change in the Apparent Fractional Void Volume in the Bed

The fractional void volume in the bed is a property that should be defined from the actual measurements or the true specific gravity of the raw material and the bulk density of the bed. In this study, however, it was used as a fitting parameter to make the calculated pressure loss consistent with the observed value. The fitting was carried out as follows. Pellets of different moisture were prepared and packed in the reaction pipe used in the tube test. The pressure loss of the pellet bed during blowing was measured at each pellet moisture content at room temperature. The apparent fractional void volume, ε', was determined so that the calculated pressure loss given by equation (1) became the same as the measured value for each moisture content. This ε' value was assumed to be applicable to the pellet bed in the drying stage in the sintering process.

Gas Properties

Though the amount of gas changes during the oxidation of the sulfide concentrates in the sintering process, it was ignored in this model for simplification. However, the increase in the H2O gas caused by drying was taken into consideration. The value of the viscosity was evaluated with Wilke's equation for a mixed gas of N2-O2-H2O.

The Results of the Calculations

Comparison of the Calculated and the Experimental Results

A calculated pressure change is shown in Figure 8 in comparison with a measured value under the same conditions. In Figure 9 the calculated and observed profiles of ΔΡ/L for 5 minutes after the blowing was started are shown with the temperature profile. The calculated and the measured values agreed well with each other, and this suggests that the change of blowing resistance is well described with the present model.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 507

' Calculated Measured

10 15

Time (min) 20 25

Figure 8 - Calculated and Measured Blowing Pressure

Temp.(Measured) Pressure loss (Measured) Pressure loss (Calculated)

60

40

20

"a

a* <

50 100 150 200

Height Above Grate (mm)

250 300

Figure 9 - Calculated Pressure Loss in the Sintering Bed 5 Minutes after the Start of Blowing

The Influence of Each Variable on the Increase in the Blowing Resistance

The value of each variable in equation (1) at each height position 5 minutes after the start of blowing is shown in Figure 10.

The gas velocity (Uo), and viscosity (u) show a similar profile to that of the temperature. These two variables are comparatively low in the drying zone. The mass flow rate of the gas (G) hardly changes in the bed. There must be a positive correlation between these three factors and the blowing resistance in Erugun's equation. Therefore, a change in the amount and nature of the gas is not the main cause of the increase in the blowing resistance.

508 LEAD-ZINC 2000

50 100 150 200 250

Height Above Grate Level (mm)

x

Figure 10 - Calculated Profile of the Variables 5 Minutes after the Start of Blowing

In contrast, there must be the negative correlation between the fractional void volume in the bed and the blowing resistance. The apparent fractional void volume is minimized when the resistance is the highest in the bed of 100-200°C. In other words, the increase in the blowing resistance is easily explained with the apparent porosity introduced into the present model. However, the physical meaning of the apparent fractional void volume in the bed in this model is not clear as mentioned above. The re-condensation of the lead and cadmium volatilized in the high temperature part of the lower layer could be another reason for the blowing resistance increase in the bed at low temperatures. Further investigation is necessary to confirm this.

CONCLUSIONS

The permeability changes during sintering in the ISP were investigated with an experimental sintering pot, as well as with a mathematical model. The results indicated that the blowing resistance in the sintering bed changes as the sintering reaction proceeds. It increases at the beginning of the reaction when the temperature of the bed is still low. Then it decreases as the reaction continues, accompanied by an increase of the bed temperature. The increase of the blowing resistance in the bed during the sintering process mainly occurred in the drying zone, and was comparatively small in the reaction and cooling zones. Based on Erugen's equation, the change of the permeability during sintering may be explained by the decrease in the apparent fractional void volume in the drying zone in the bed.

REFERENCES

T. Nitta and K. Kikuta, "Improvement of Sinter Quality for the ISP," World Zinc'93.1.G. Matthew, Ed., Aus. Inst. Min. Metall, Parkville, Australia, 1993,445-451.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 509

2. H. Kitahara and O. Ogawa, "A Basic Study of the Oxidizing Roasting in the Sintering Process of ISP," Shigen-to-Sozai, Vol. 113,1997,53-59.

3. M. Wajima, Y. Hosotani, J. Shibata, H. Souma and K. Tashiro, "Condensing Phenomenon of Moisture in Sintering Bed and Its Effect on Bed Permeability," Tetsu-to-Hagane, Vol. 68,1982,1719-1727.

4. J. Higuchi and I. Muchi, "Mathematical Model of Sintering Bed," Tetsu-to-Hagane. Vol. 53,1967,1171-1173.

5. J. Shibata, M. Wajima, H. Souma and H. Matsuoka, "Analysis of Permeability and Characteristic Suction Gas Volume in Sintering Process," Tetsu-to-Hagane. Vol. 70, 1984, 178-185.

6. E. Kasai, J. Yagi and Y. Omori, "Mathematical Modeling of Sintering Process Considering Influence of Changes in Void Fraction and Apparent Particle Size in the bed on Pressure drop." Tetsu-to-Hagane. Vol. 70,1984,1567-1574.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 511

CHANGES IN THE PHYSICAL AND MECHANICAL PROPERTIES OF SiC TRAYS CAUSED BY AGEING IN ZINC REFINERY OPERATIONS

A. Piant, M. Fritz and M. Boussuge Ecole des Mines de Paris, Centre des Materiaux P.M. Fourt

B.P. 87, 91003 Evry cedex, France

M.-D. Dupuits Societe Frangaise de Ceramique

23 rue de Cronstadt, 75015 Paris, France

ABSTRACT

This paper summarises a part of the work that has been essentially carried out at the Ecole des Mines de Paris in the framework of an Eureka European programme funded by the French Ministry of Industry, in partnership with Metaleurop and Carborundum (subsidiary of Saint Gobain). It deals with the measurement of different properties (density, thermal linear expansion, elastic modulus, strength) of a SiC-based refractory constituting the trays that are used in the refining columns in the New Jersey zinc distillation process, especially in the boiling section. These parameters have been investigated to check the homogeneity of the as-received trays as well as to estimate the influence of ageing of the material under operating conditions. With this aim, specimens have been sampled in used trays having more than two years of operation. The as-received trays exhibited a good homogeneity, the strength being slightly lower in the direction in which they are compacted. Despite microstructural changes, most of the properties measured at room temperature are not significantly affected by ageing. High temperature tests involved thermal treated under vacuum to remove metallic zinc before testing at high temperature. An increase in temperature, as well as ageing, induces a decrease in the elastic modulus and strength.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

512 LEAD-ZINC 2000

INTRODUCTION

In the New Jersey process of zinc refining, zinc is distilled in columns made from numerous (typically 56 to 58) piled-up trays. Half of these trays are located in the lower, hottest part of the column, the boiling section. Fourty percent of the trays constitute the refluxing section in the upper part of the column, and ten percent the feeding intermediate section. An example of a tray, called a W-shaped tray because of its cross section, is represented in Figure 1. The denomination of the different areas, which will be used later, is reported in this figure.

Channel

Knee

Figure 1 - Cross-section of a W-type Tray

A tray, which is about 1.4x0.8x0.2 m, consists of a surrounding wall of 38 mm thickness, a peripheral channel, and a bottom of thickness 32 mm with a square hole of about 0.4x0.4 m that allows vertical zinc exchange between the trays. The columns are heated in their lower parts using gas burners, and are cooled in their upper part, to allow the refluxing. Service durations of several years are expected for such industrial systems, so that the study of degradation processes takes on a particular importance. The mode and the kinetics of the damage affecting the column depend essentially on two factors: the history of loading and the changes in the material properties induced by the operation. With this aim, the research program carried out by Metaleurop, Carborundum (Saint Gobain group) and Ecole des Mines intends to:

• Identify and quantify the loadings applied by the process to the trays, that may have an effect on the lifetime of the column

• Identify and quantify the changes in material microstructure and properties induced by ageing under operation conditions

• Estimate the relative influences using numerical simulation with finite element calculations.

This paper relates the intermediate results obtained in this framework.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 513

LOADINGS IMPOSED ON THE TRAYS IN SERVICE CONDITIONS

In the upper and colder part of the column, no significant damage is generally noted after operation, the only problem being a slow diffusion of liquid zinc through the walls of the trays and between the trays through the joining concrete. This leads to a very limited zinc loss outside the column, that can be managed. In the lower part, the trays are subjected to higher stresses, especially from thermal origin, that may lead to more extensive damage. Two different types of thermal loading can be noted, respectively associated with continuous operation and with transient conditions.

During continuous operation, a steady thermal gradient exists through the walls of the trays. The external surfaces of the trays may reach 1150°C, whereas the internal surfaces are maintained at 900°C by zinc distillation. Other stresses are caused by the introduction of the liquid zinc to be refined in the middle part of the column. The metal is introduced at a temperature varying between 500°C and 600°C within the column, so that several trays located under the zinc entrance are cooled (the normal distillation temperature is around 900°C), and their bottoms and channels are then submitted to an additional thermal gradients.

Transient thermal loadings are probably more aggressive for the trays. They can be associated with the maintenance and control operations (change of feed tube, adjustment of registers of the secondary air circuit, etc.) which require the stoppage of heating for some minutes. Other causes of stresses are the instabilities in the zinc flow entering within the column, that must be carefully avoided.

Some steady mechanical stresses also affect the trays, essentially stresses attributable to gravity. The weight of the filled-up column is the origin of the compressive stresses in the walls of the trays, that affect especially the lowest plates. The weight of molten zinc also generates a pressure on the bottom of the trays, which can be considered as a first approximation as a tray loaded in bending by a constant pressure. The analysis of the damage affecting the used columns indicates that none of these mechanical loadings plays a major role in the degradation of the trays. However, a sufficient mechanical resistance is absolutely necessary to insure a safe environment in anomalous working conditions.

RELATIONSHIPS BETWEEN FUNCTIONALITY AND THE BASIC PROPERTIES OF THE MATERIAL

The qualities that are expected from a material for good behaviour, reliability and durability under operating conditions are related to the basic thermal and mechanical properties of the material. Table I summarises the characteristics that a material must have to resist thermal and mechanical loadings.

514 LEAD-ZINC 2000

Homogeneity Density Thermal conductivity Thermal expansion Elastic modulus Strength Fracture toughness Microcracking resistance Viscoplasticity Slow crack growth resistance

High High

--

High High High High Low High

Table I- Relationships Between Functionality and the Basic Properties of the Material. Property Mechanical Loadings Thermal Loadings

High

High Low Low High High Low High High

This table reveals some contradictions between the functional properties expected from the material. For example, the development of microcracking within the material diminishes the stresses for thermal loadings (because it lowers the stiffness of the material), and then tends to avoid the creation of major cracks. However, microcracking also induces a decrease in thermal conductivity and in the strength of the material, that have a negative influence. Viscoplasticity (i.e., creep deformation) has a positive effect for thermal stress relaxation, but also a negative one by generating permanent deformations, damage and sometimes cracks in the material. A material operating well under service conditions reflects a complicated compromise between the numerous characteristics of the material.

The problem is complicated when considering ageing under operating conditions. Firstly, the durability of the material depends on these initial properties. Secondly, all these properties may change during ageing in operation. In this general framework, this paper will be especially devoted to the properties of the material and to the ageing effects on the microstructure and basic properties.

MICROSTRUCTURE OF THE MATERIAL

The material constituting the trays is produced by the Carborundum company (Saint Gobain group). The material is a SiC-based refractory (around 85% SiC), with an oxide-based intergranular phase. In this study, only specimens cut from actual trays, being more representative, have been considered. However, it is important to note that all the analyses presented here have been obtained on small pieces of material; because of the large amount of material involved in a column, some variability in the composition may be observed.

As-received Material

An optical microscopy examination, represented in Figure 2, revealed that the material is made from coarse silicon carbide grains of some millimetres size, bonded by an intergranular phase that also contains smaller SiC grains. Some porosity can be observed in the intergranular phase, as well as within the coarse SiC grains.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 515

Figure 2 - Optical Micrograph of the Material

An electron microprobe analysis was also performed, in order to reveal the location of different elements in the microstructure; the results are given in Figure 3. Silicon carbide coarse and fine grains can be easily identified because of their high silicon content. The composition of the oxide intergranular phase confirms the analysis made using X-Ray diffraction (silica and aluminum and magnesium oxides in smaller quantities). Small amounts of calcium and potassium have been also detected in this oxide phase, and some localised areas rich in iron and/or in chlorine can be detected in the intergranular phase.

An X-Ray diffraction analysis was carried out to identify the crystalline phases constituting the material. The silicon carbide (SiC) is in its moissanite form, whereas the intergranular oxide phase is essentially composed of silica.

516 LEAD-ZINC 2000

Figure 3 - Backscattered Microprobe Analysis of the As-received Material (Bright Areas Indicate a High Content)

Aged Material

In service operation, the material is submitted for a long time to high temperatures and contact with liquid or gaseous zinc. In order to examine the effect of this type of ageing on the microstructure of the material, specimens were cut from trays having more than two years of operation, coming from different levels in the column. In each case, different locations in the tray have been analysed in the Saint Gobain Ceramiques Industrielles Research Centre (C.R.E.E.) in Cavaillon, France. Chemical analyses were performed, as well as X-Ray diffraction analyses.

The ageing mode essentially consists of the infiltration of the material by zinc, and of possible changes in the structure of the intergranular oxide phase. In the upper, colder part of the column, the structure of the material seems not significantly affected by ageing. Only metallic zinc can be found, and it probably migrated through the porosity of the material. In this case, only the cristobalite form of silica was detected in the upper part of the column.

In the lower part of the column, an analogous phenomenon of zinc impregnation is noted for the bottoms of the trays which are maintained at a relatively low temperature by zinc distillation. Chemical reactions and more intense oxidation of the zinc complicate the degradation process for the walls of the trays, the external surface of which is submitted to higher temperatures in the combustion chamber. A high level of zinc oxide (ZnO) and the tridymite form of silica (S1O2) are found on the external surfaces. Metallic zinc is still sometimes observed in the material. Here, however, the simultaneous presence of zinc and

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 517

oxygen may lead to a reaction with the intergranular oxide phase to form zinc silicates (Zn2SiC>4) or aluminates (Z11AI2O4).

MECHANICAL TESTS AT ROOM TEMPERATURE

The trays are large parts having a rather complicated shape, and compaction of the material in some areas is not an easy task. This may result in an heterogeneity of the component, with weaker areas in the tray, and an eventual anisotropy in the mechanical properties. An experimental programme has been conducted to test the homogeneity of the trays. Beside this, tests have also been performed on specimens cut from used trays, in order to evaluate the effect of ageing on the material properties.

Measurements and Associated Methods

Parallelepipedic specimens of length 150 mm were cut from an as-received tray provided by the Carborundum company (Saint Gobain group), to test the as-received state of the material. Similar samples were taken from different trays coming from the used column. All specimens have squared sections corresponding to the thickness of the area from which they originated. Different measurements were performed on these specimens:

• Density was calculated as the ratio weight/volume, this latter being calculated from the dimensions measured with a calipher.

• Elastic modulus was measured using a self-vibration technique (Grindosonic). The specimen is placed on two supports with a 125 mm span, and its first vibration mode in bending is excited by a light shock. From the measured resonance frequency, the dimensions and the weight of the specimen, its elastic modulus can be calculated. Three measurements were performed on each specimen in order to check the reproducibility. Moreover, it was verified on one specimen that the elastic modulus measured with this method agrees well with that measured using strain gages.

• Strength has been measured using 3-poirtt bending tests with a 125 mm span.

Sampling of the Specimens

In the as-received tray, specimens of 39x39 mm cross section were cut from the walls of the tray, vertically (parallel to the pile-up direction of the trays in the column) and horizontally (parallel to a long side of the wall). They are respectively referenced as VW3 and VW1 (as-received material, walls in directions 3 and 1). Other specimens, of section 25x25 mm, referenced VK1 (knee in direction 1), were taken from the junction between the bottom of the tray and the channel.

Two different used trays were studied. The first was located at the bottom of the column, in the combustion chamber, (reference Bl) and the second under the zinc input in the column (reference II). In order to reveal an eventual asymmetry in the column, specimens were cut from different walls, parallel to a long side. During the tests, different surfaces (external aged in the combustion chamber, internal in contact with liquid or gaseous zinc or machined bulk material) were loaded in tension, to test the effects of ageing on the different surfaces. Finally, the bottoms of trays were also tested in two different configurations for the tensile face; that is, the surface which was in contact with the liquid zinc or that facing the gaseous zinc.

518 LEAD-ZINC 2000

Results

Table II summarises the obtained results.

Table II - Test Results on As-received and Used Materials (Standard Deviation in Brackets) Reference

Number of specimens Density RT Elastic mod. (GPa) RT Strength (MPa)

VW3 5

2.6 [0.02] 76 [5] 17 [31

VW1 10

2.6 [0.01] 104 [11] 27 [51

VK1 5

2.6 [0.02] 123 [8] 28 [41

Bl 26

2.8 [0.1] 117 [26] 24 [5]

11 32

2.9 [0.2] 96 [22] 23 [5]

Concerning the as-received tray, the density and the elastic modulus do not depend on the location of sampling. The only effect that can be noted is the influence of the direction of cutting from the walls; the strength and the elastic modulus are lower in direction 3. However, this observation has a limited influence on the functionality of the tray because, in normal service conditions, only compressive stresses are applied in this direction.

With respect to the scatter of the obtained results for the same conditions, no significant influence of the location of sampling and of the surface in tension was found for the used material. The results have then been grouped together in two columns in Table II. Operation always results in an increase in density, because of the impregnation by zinc. Another effect is noted on the elastic modulus that tends to increase because of the filling-up of the porosity by molten zinc. A strong increase in the scatter of the density and of the elastic modulus is associated with the use of the material, proving the heterogeneity of zinc impregnation. Finally, the strength, measured at room temperature, seems not to be significantly affected by the operation, but it is worth noting that the impregnation by solid zinc may strengthen the material.

TESTS AT HIGH TEMPERATURE

Bending tests were performed at high temperature in air on as-received and used material. All specimens were cut from the walls, along a long side of the tray (respectively VW1 and UW1 types for the as-received and used material). In the used material, the removal of the zinc impregnating the specimen was necessary to prevent the oxidation of the metal during testing. A first attempt was made using hydrochloric acid, which was unable to remove all the zinc from the bulk material. Another method, using vacuum at high temperature, was developed by Metaleurop Recherche in its Trappes laboratory. Specimens were slowly heated for four hours, and then maintained during six hours at 650°C.

All elastic moduli were determined using deflection measurements in 3-point bending tests at Societe Franfaise de Ceramique in Paris (specimen size 110x10x8 mm), whereas some of the strength measurements were performed at C.R.E.E. (Saint Gobain Ceramiques Industrielles research centre) in Cavaillon (specimen size 150x25x25 mm). Tables ΠΙ and IV summarise the obtained results.

For the used material as well as for the as-received one, an increase in temperature always induces a decrease in the elastic modulus (Table III). This observation is generally explained by a progressive weakening of the intergranular oxide phase. The damage created in

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 519

the intergranular phase by ageing is probably responsible for the decrease in stiffness induced during operation. However, it is worth noting that a decrease in elastic modulus is not necessarily negative for this application (see Table I). As a matter of fact, stresses induced by a thermal gradient are of the imposed strain type, and then tend to become lower when the elastic modulus decreases.

Table III - Elastic Moduli in GPa Measured in Bending (Averaged Values Obtained on 5 Specimens and Standard Deviation is Given in Brackets)

As-received Material Used Material after Zinc Removal RT 1100°C RT 1100°C

98 [12] 59 [5] 49 [3] 38 [2]

Whatever the test temperature, ageing reduces the strength of the material by a factor of two (Table IV). Changes observed in strength versus temperature are slightly more complicated. For the as-received material as well as for the used one, the strength measured at 1200°C is of the same order of magnitude as that measured at room temperature. At intermediate temperatures, the strength of the material is higher. The observed decrease in strength for temperatures between 1100°C and 1200°C can easily be explained by the weakening of the intergranular phase. It is worth noting that the refractory nature of the intergranular phase seems not to be significantly affected by ageing; in both cases, the strength decrease is observed between 1100°C and 1200°C. The increase from room temperature to 900°C and 1100°C is less clear and, without additional experiments, three assumptions can be made:

• A change in the crystalline form of the intergranular phase occurs • Cracks and flaws healing is induced by the heating performed under air; i.e., under

oxidising conditions • Some viscoplasticity affecting the intergranular phase and allowing a partial, local

accommodation of stresses.

Table IV - Strength in MPa Measured in Bending (Standard Deviation is Given in Brackets) As-received Material Used Material

after Zinc Removal Temperature (°C) 20 900* 1000* 1100 1200* 20 900* 1100 1200* Strength 30 [4] 43 [1] 44 [3] 54 [5] 34 [4] 14 [2] 36 [7] 24 [2] 16 [4] Specimen # 5 3 3 3 3 5 3 3 3

The symbol * indicates results obtained by Carborundum.

A comparison between the results obtained at room temperature after zinc removal (Table IV) with those given in Table II reveals an increase in stiffness and strength that is attributable to the solidification of the zinc impregnating the material. However, it is important to note that this strengthening effect does not apply under the operating conditions in the column, because the zinc is in liquid or gaseous form.

Finally, dilatometric measurements performed between room temperature and 1300°C showed that the linear expansion coefficient remains constant over this temperature range. For both materials, its value ranges between 4.7x10"6 and 4.9x10"6 °C~'. This observation and these

520 LEAD-ZINC 2000

values, which are very close to those observed on dense, pure silicon carbide, confirm that the dilatation is controlled by the skeleton of silicon carbide grains that constitutes the main part of the material.

CONCLUSIONS

This study constitutes the first part of a programme to investigate the degradation processes of columns that are used in the New Jersey distillation process of zinc refining. It has focussed on the different loadings acting on the trays under service conditions, the microstructure of the material, the homogeneity of the as-received trays, and the effects of ageing on the microstructure and on the mechanical properties of the silicon carbide-based material constituting the trays. These initial results will be completed along different research axes:

• Experimental data collection on the material, that can be used to perform accurate finite element numerical simulations of the trays. Two examples are statistical models for strength and constitutive equations of creep

• Microstructural ageing mechanisms of the material, and eventually, improvements which can be proposed for the refractory

• Consequences of ageing on the characteristics of the material • Numerical simulations in order to calculate the stresses induced in the tray under normal

and anomalous working conditions. These simulations will be used to optimise the design of the trays and the heating procedures for the columns.

ACKNOWLEDGEMENTS

The authors wish to thank Metaleurop Recherche (M. Sibony, J.C. Goimard, S. Sella), Metaleurop Nord (A. Lebleu, G. Vanhelle, Y. Le Quesne) and Carborundum, a subsidiary of Saint Gobain, (O. Marguin) for their active participation in this study, and the French Ministry of Industry for funding this project.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 521

DEVELOPMENT OF MECHANIZATION FACILITIES IN THE NON-FERROUS METALLURGICAL FIELD

Y. Sako, K. Muraguchi and K. Shibata MESCO, Inc.

2-10-5 Ryogoku, Sumida-Ku Tokyo, Japan 130-8531

E. Nomura Mitsui Mining & Smelting Co., Ltd.

1-11-1 Ohsaki Shinagawa-ku Tokyo, Japan 141-8584

ABSTRACT

In 1998, Mitsui Mining & Smelting Co., Ltd. (MMS) produced 70,000 tonnes of lead, 231,000 tonnes of zinc and 214,000 tonnes of copper. MESCO, Inc. has been in charge of the automation and mechanization of smelting facilities promoted by MMS as well as the construction of facilities for chemical production and electronics materials plants, by which MESCO has contributed to the development of MMS. Initially, MESCO developed a casting machine for refined lead in the early 1960's, and thereafter, it developed a casting machine for refined zinc, a stripping machine for zinc tankhouses and energy saving facilities. In the field of copper smelting, it has also developed automation and mechanization facilities.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

522 LEAD-ZINC 2000

GENERAL

Mitsui Mining & Smelting Co., Ltd. (MMS) is a general non-ferrous metal enterprise which originated its business from mining and smelting. At present, its spectrum of business activities covers not only the production of zinc, lead, copper and other materials but also it has extended its applications to the production of rolled copper products, synthetic chemical products, electric circuit materials (electrolytic copper foil), PVD (physical vapour deposition) materials, battery materials, ceramics, rare metals, rare earth and automotive components.

MESCO, Inc. is an independent engineering company listed in the Tokyo Chamber of Commerce which takes charge of the engineering services of MMS, and it has carried out the engineering for almost all of MMS's projects.

In the smelting business, MMS has 6 smelting and refining plants and they produce annually 70,000 tonnes of electrolytic lead, 133,000 tonnes of electrolytic zinc, 35,000 tonnes of reflux-refined zinc, 63,000 tonnes of distilled zinc and 214,000 tonnes of electrolytic copper.

OUTLINE OF THE SMELTING PROCESSES OF MMS

An outline of the lead, zinc and copper smelting processes are as described below.

Lead Production (Kamioka Plant and Takehara Plant)

Lead smelting is done in these two smelters and both of them use scrap batteries as the major material source. The smelting is by blast furnace and refining is done by electrolysis utilizing the largest cells in the world, developed by MMS.

Zinc Smelting (Hachinohe Smelter, Hikoshima Refinery and Kamioka Plant)

Zinc production is done in these three smelters. That is, one is for the production of distilled (reflux refined) zinc in an ISP smelter with one of the highest productivities in the world, another is electrolysis refining of MMS's own development and the last is electrolytic zinc production in a tankhouse, whose basic design was provided by Vielle-Montagne and detailed engineering was provided by MESCO. Further, MMS has long been engaged in the smelting and refining of zinc which indispensably requires treatment of residues. As a result, it has developed a special zinc residue treatment process which is practiced at the Miike smelter.

Copper Production (Tamano Plant)

This smelting plant adopts a flash smelting furnace improved by MMS and the refining plant adopts high current density electrolysis using periodical reversing current.

Figure 1 shows the relationship among the lead, zinc and copper smelters.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 523

| Copper GöiP L^J-Insorted Cower Recycled Cower Recycled Lud Inserted Leid Imported Ziec Recycled Zinc Owi Mine Zinc Steel Coecwtrtt» Miteml Matemi Coeceirinrte Coscestnte Miteritl Coscwtnte Dsit

SxisSsmo r^ Crude Leid ; [Crude Lead J LCiJ^Jii i iLJ j..Ctulie z i l ,c j ^ί Leidwf j

! Residue ί

PAC- Ψ . . _ ▼ T . ▼ [Refining| | Refluxin«| | Elec1roly»is| | Eleclrolytis| tloctrohriis Electro ly «is I Electror/tis| BecürtüySs] I Refhinf I — Γ

Electrolytic Electrolytic Distilled Refluxing Electrolytic I °°°°«Γ L»«< Zinc Zinc Zinc Smelt.

Figure 1 - Metallurgical Complex of MMS

HISTORY OF THE ENGINEERING BUSINESS

MMS was engaged in the development of mechanized facilities for smelting and refining as early as the 1960's. That is, in 1959, it developed a casting machine for lead which was a pioneer in this field, and in 1964, it developed a zinc stripping machine and a zinc ingot casting machine which were used in its own plants with good results. MESCO took charge of the marketing of this equipment, through which it has made improvements, developed new equipment, and made licensing and marketing agreements with overseas companies resulting in good sales promotion and high appraisals.

To mention some, in 1972, negotiations to supply mechanization facilities for zinc smelting and refining to foreign smelters started and contracts were signed successively. This opened the door for MESCO to export mechanization facilities, and thereafter, it has been successful in overseas marketing. Further, MESCO has established a relationship with Outokumpu of Finland for copper smelter facilities. In Japan, MESCO has delivered facilities not only to MMS but also to other companies which are in the same trade.

524 LEAD-ZINC 2000

Through the marketing of mechanization facilities, MESCO, jointly with Marubeni Corporation, was awarded a contract for the construction of the PASAR Copper Smelter and Refinery (production capacity: 138,000 t/y) in the Philippines in 1980 and the construction was completed in 1984. This project was successfully completed within the scheduled time and with satisfactory operational results. This success became a milestone for MESCO to promote an active copper business in overseas countries. Utilizing our own technology for the mechanization of copper smelting plants and with the collaboration of Copper Refinery Ltd. (CRL) of Australia, MESCO has supplied technology and constructed plants for copper electrolysis in foreign countries.

DEVELOPMENT AND PRESENT STATUS OF STRIPPING MACHINES FOR ZINC REFINERIES

Vacuum Pad Type

The type of stripping machine which was developed in 1964 uses vacuum pads to suck the zinc sheets deposited on both sides of the cathode and to strip them. This type had been in use at the Kamioka Plant of MMS with good results. It had been in use for about 10 years, but because it requires a flat and uniform zinc deposit surface, conversion to other systems was requested and thus the design was changed to a wedge type.

Wedge Type (Fixed Type)

The first wedge type stripping machine was completed in 1970. It replaced the vacuum pad type of Kamioka and at the same time, it was adopted in the Hikoshima plant constructed using the technology of MMS and exhibited good results. About 20 units of this type have been exported to Germany, Canada and other countries.

At the Hikoshima plant, this machine and the transport system of the cathodes from the cell are combined as one system, which is automated and mechanized. The whole operation, from stripping the deposited zinc to transporting the product zinc cathode to a melting furnace by a conveyor can be operated by a single person. High speed stripping of the deposited zinc from the cathode plate can be achieved at a rate of one cathode every seven seconds. Preliminary stripping, carried out by hammering, enables smooth and steady results with no particular strain on the cathodes. Very little damage is caused on the cathode plates by wedges and other items because of the above feature. An optional cathode polishing machine to go with the system is available depending in the upon customer's requirements. The major advantage of this system over other systems is the hammering operation, which is done prior to sending the cathode to the stripping section, which partially peels off the deposited zinc from the cathode surface. A wedge type device can then effect the smooth removal of the rest of the deposited zinc. In this way, less damage is caused on the cathode plates and a higher ratio of deposited zinc is stripped at a higher speed. Figure 2 shows the system of the wedge type stripping machine.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 525

Figure 2 - System of the Wedge Type Stripping Machine

Wedge Type (Movable Type)

In a normal newly constructed refining plant, the wedge type stripping machine (fixed type) can be utilized more effectively than the movable type. However, in an old facility like the Kamioka plant, the general layout is not suited for systemized mechanization and the wedge type (fixed type) stripping machine is not well utilized. Therefore, the Kamioka plant developed a movable multi-sheet simultaneous stripping machine in 1981 and used it until 1993. The outline of this machine is described below.

• The machine car travels on the ground along the side of the cell for easy access to the cell.

• Five cathodes can be stripped by special stripping knives at the same time. • Such facilities as cathode washing, stripping, detection of sticking cathodes, and

supplying a stand-by cathode are incorported in the car. • Stripped cathodes are automatically sent through a weigher to the stock yard by a chain

conveyor. • The stripping cycle is about 6.5 seconds per sheet, and at the Kamioka plant, about four

mea are required for the labor of cathode handling at the cell side and for stripping.

The Kamioka plant has been generally superannuated and in 1993, it introduced the super-jumbo process from Vielle-Montagne of Belgium; this plant is in operation at the present time.

526 LEAD-ZINC 2000

Figure 3 shows the movable type stripping system.

Figure 3 - Conceptual Drawing of the Movable Type Zinc Stripping Machine

DEVELOPMENT AND PRESENT STATUS OF THE COPPER STRIPPING MACHINE

The zinc stripping machine strips zinc sheets deposited on aluminum cathodes. Adopting this theory to strip deposited copper sheets from stainless steel cathodes is the basis of the ISA Process developed by CRL of Australia.

The zinc stripping machine adopts hammering for the preliminary stripping of the edge of the deposited zinc sheets. The copper stripping machine has been modified to employ flexing for this purpose utilizing the physical properties of the stainless steel mother blank and the deposited copper sheets. The ISA Process was developed in 1979 and since then, about 50 licenses have been granted. At present, the production of copper by the ISA Process accounts for approximately 4 million tons of copper per annum.

Features of ISA Process are explained below.

• Reduction in metal inventory by harvesting younger cathodes. • Consistently high cathode quality and current efficiency because of straight and vertical

cathode plates. • Increasing current densities, up to 330 A/m2 at present. • Variable copper cathode weights, between 5 and 14 days electrolysis. • Variety of cathode stripping machines, from 100-600 cathodes/h.

MESCO supplied stripping machines for the original ISA Process and has been contributing to the development of the ISA Process. In addition to mechanical technologies, MESCO has supplied the overall electrolysis technology based on the ISA Process. To mention some, MESCO supplied the basic and detailed engineering for a 200,000 tonnes per year copper electrolysis plant in Indonesia, and Carried out the basic engineering for a copper electrolysis plant in Australia.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 527

It is about 20 years since the ISA Process was developed and some refineries do not prefer the use of wax. Consequently, the ISA2000 Process was developed jointly by CRL, WENMEC and MESCO.

Figure 4 shows the outline of the ISA Process.

Blank Maintenance ♦

Anode — Press ► Electrorefming ► Washing/Stripping ► SUS mother blank

Anode scrap • > | Cathode

528 LEAD-ZINC 2000

Starting Sheet Electrorefining

Anode— Press

Polishing/Maintenance/Wax

Starter sheet electrolysis —► Washing/Stripping —i—► Mother blank

Cu starter sheet —ι

Ribbon setting/Cross bar setting/Press ^

Electrorefining

Anode scrap

Washing H Cathode

Figure 4 - ISA Process Machines for All Applications

DEVELOPMENT OF CASTING MACHINES

As early as the 1960's, MMS had been engaged in the mechanization of refinery equipment. For example, in 1959 a lead ingot casting machine was developed which was the first among other plants. In 1964, a zinc stripping machine and zinc ingot casting machines were developed and used in MMS plants with successful results. The history of development of zinc ingot casting machines is explained hereunder.

Small Slabs

The features of the small slab casting machine, Figure 5, are as follow. The entire operation, from molten zinc feeding to bundle strapping can be done by automatic operation with only one operator. This contributes greatly to save labor cost. An automatic dross skimming machine can be incorporated into the system depending on the customer's requirements. This eliminates the need to move the entire skimming machine as the molds are installed in line to meet the casting conveyor's movement. As the bundles are strapped with pressure, loosening of the bundles is kept to a minimum. The standard production rate is 20 tonnes per hour for casting 20 to 25 kg zinc slabs. Less manpower requirement, more consistent and better quality, smoother surface gloss and operation at higher speed are advantages over conventional processes. Moreover, as the stack of cast zinc slabs is securely strapped by the automatic handling system, loosening and breakage of strapped cargo which is in stock or in transit is eliminated.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James andA.H.-J. Siegmund 529

Figure 5 - Outline of Small Slab Casting Machine

Large Sized Zinc Ingot

Most zinc consumption is for the surface treatment of steel and most of this surface treatment is done by molten zinc galvanizing. With the spread and expansion of molten zinc galvanizing, a demand for large sized ingots, rather than for small ones, emerged. Further, in many cases, these ingots are alloyed beforehand to a composition required for galvanizing.

MESCO, taking advantage of its experience in the development of small slab casting machines, developed a casting machine for large-sized zinc ingots as shown in Figure 6. The outline is explained below.

• All processes - melting, volume control, casting, releasing from the mold - in this system are fully-automated. Mass production of large-sized ingots is made possible by this system with the highest degree of efficiency. The quality of the products is stable because of a cooling method meticulously devised to prevent the segregation of elements.

• The top heating system enables a fine finish on the surface of the ingots.

• A high rate of production is achieved by the continuous process.

• The following five major processes are carried out by the system to give good advantages over conventional systems which depend on a lot on manual operation.

530 LEAD-ZINC 2000

In the zinc metal melting process, the volume of molten zinc metal is continuously controlled, and a relatively small sized furnace can feed a volume equal to a large sized one. A fixed amount of molten zinc is fed constantly and is transferred to the next process automatically. An automated and continuous process is applied where certain alloying elements are melted and mixed with molten zinc metal to produce zinc alloy. In this process, the molten zinc alloy is poured into a mold through an automatic weigher. Also, the dross form the molten zinc alloy is skimmed and the molten alloy is cooled. The ingot is then released from the mold automatically.

Figure 6 - Outline of a Large-Sized Ingot Casting Machine

DEVELOPMENT OF A SPECIAL BOILER FOR ELECTRIC POWER GENERATION FROM THE WASTE HEAT RECOVERED FROM COOLING LAUNDERS

MMS carries out the simultaneous smelting of lead and zinc in a blast furnace utilizing the ISP Process. Zinc fume produced in the blast furnace is introduced to a condenser together with exhaust gas and the zinc is absorbed in molten lead. This molten lead, which has absorbed the zinc, is pumped to a cooling launder where it is cooled to separate lead and zinc for zinc recovery. In this cooling process for the molten lead, a vast amount of waste heat is generated, and previously, this waste heat was absorbed in the cooling water and discharged to atmosphere via a cooling tower. With the recent development of a special boiler, this waste heat is now recovered as super-heated steam and a power plant to utilize this super-heated steam has been constructed. This power plant currently has an average output of 3,300 kW and it is operating trouble-free.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 531

The outline of the facility and the project plan are shown in Figure 7, and are discussed below. The problem with heat recovery from molten lead, containing zinc, in the cooling launder is that the zinc in the molten lead corrodes almost all kinds of metals, and therefore, heat transmission pipes cannot come in direct contact with the molten lead to produce high temperature and high pressure steam. The cooling launder is a part of the zinc production facility and the temperature of the molten lead at the outlet of the cooling launder must be kept constant (430±10-) and it is necessary to control the heat volume absorbed from the boiler. It is necessary to prevent leakage of water circulating in the heat transmission pipes. Finally, the heat transfer has to be done in a limited space and it is necessary to adopt a boiler with good heat transmission.

For the above reasons, a special boiler was developed and the above problems were solved by this boiler. Taking into account the prevention of corrosion of the heat transmission pipes by molten zinc, control of heat absorption, prevention of leakage of the circulating water from the. pipes and the improvement of the heat transfer efficiency, the structure of boiler is made of protection casing, exterior low melting point alloy, inner casing, inner low meting point alloy and heat transmission pipe. The heat transmission pipes are protected from corrosion by zinc by several layers and breakage of the casing can be detected immediately from the level of the low melting point alloy. Further, the heat transmission pipes are supported at two points at the inlet and outlet to minimize thermal stress so as to prevent breakage of the pipes. Control of heat absorption from the cooling launder is done by separating the boiler into 40 small units (28 units of evaporator and 12 units of super-heater) which can be inserted into the cooling launder in units of two to vary the heat transmission area. As to the heat transmission efficiency, all gaps are filled with low melting point alloys to improve the overall heat transmission efficiency. Evaporators and super-heaters, formed as panels, are immersed into the molten lead in the cooling launder for heat exchange. During normal operation, 12 units of super-heaters are immersed up-stream of the cooling launder and 28 units of evaporators are immersed down-stream of the super-heaters. A circulation pump supplies water by forced circulation and this water passes through each evaporator and is sent to a steam drum. The steam generated in the drum is separated from the saturated circulating water. The separated saturated steam then passes through super-heaters to turn to super-heated steam which is sent to a steam turbine.

The outline of generator is given below: Boiler Type Forced circulation waste heat type

Max. evaporation volume 24.3 t/h Steam pressure (outlet of super-heater) 16 kg/cm2

Steam temperature (outlet of super-heater) 410°C Heat transmission area Evaporator 133.2 m2

Super-heater 57.1 m2

Turbine Type Mitsui-BBC condensing type Rated output (generation terminal) 8888 rpm Inlet steam pressure 14kg/cm Inlet steam temperature 400°C

532 LEAD-ZINC 2000

Generator Type Three phase alternate current synchronized type Rated output 4526 KVA Power factor 95% Voltage 3300V Frequency 50Hz Number of poles 4 Revolution 1500 rpm

Figure 7 - Layout of the Facilities

EXAMPLE OF THE MECHANIZATION OF A LEAD ELECTROLYTIC REFINING PLANT

At MMS, refining of crude lead produced from the blast furnace is done by electrolysis.The conventional system uses small size cells, but MMS succeeded in enlarging this facility and it is operating successfully. This electrolytic lead refining process features labor savings by the use of a highly mechanized and automated system. The process covers the anode casting machine, the starter sheet preparation machine, the cross rod slipping machine and the slime scraping machine. The Takehara and Kamioka lead smelters of MMS are presently equipped with the largest lead electrolytic cells in the world and the operation is carried out with the best of results.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 533

Cell size 5,000mm (L) x 1,300mm(W) x 1,600 mm (D) Anode size 1,150mm (L)x l,000mm(W)x 30 mm (T); weight 350-400 kg Cathode size 1,200mm (L) x l,080mm(W) x 0.8 mm (T)

The automatic electrode handling system, developed for use in lead electrolysis tankhouses and outlined in Figure 8 enables efficient electrode preparation and lead product finish. The entire operation from automatic anode casting, anode and starter sheet assembling, starter sheet assembling, scrap anode washing, cathode conveying and lead pig casting is included. All six operations mentioned above are automatically handled by three operators at the monitor, thus enabling considerable manpower saving. Anode and starter sheets, prior to being loaded on the cell, can be simultaneously assembled over the conveyor. Only one overhead crane is needed for handling the electrodes as both electrodes are loaded at the same time.

Figure 8 - Layout of the Facilities

Chapter 7

Zinc Electrowinning

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 537

REVIEW OF ENGINEERING AND CONSTRUCTION FACTORS IN BUILDING A ZINC CELLHOUSE

. G. Cooper, P. Mercille and M. F. Nasmyth SNC-Lavalin Inc.

455 Rene-Levesque Boulevard West Montreal, Quebec, Canada H2Z1Z3

ABSTRACT

The turn of the century sees considerable activity in the zinc industry for the replacement or expansion of existing cellhouses, and the building of new cellhouses as part of new zinc projects. Aside from the cellhouse technology itself, numerous engineering and construction issues play a major role in the success of a project for replacing or building new cellhouses. The purpose of this paper is to bring to light and to address the engineering and construction factors which go beyond the selection of a cellhouse technology and which remain of key importance in implementing such projects. In addition, some implications of these factors for project costs will be illustrated.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

538 LEAD-ZINC 2000

INTRODUCTION

The past three decades have seen important advances in the development of zinc cellhouse technology. From the original manually operated technology, a second generation of mechanized cellhouse equipment evolved, led by Union Miniere with their first jumbo cellhouse in 1969. Only 10 years later, Union Miniere again led developments with the introduction of their first super jumbo cellhouse (1). This was closely followed by Cominco's startup of their own automated, large electrode cellhouse (2). These developments are now being supplanted by a new generation of compact, highly optimized, and fully automated cellhouse technologies (3,4). Development has been stimulated by high labour costs and the need for improved working conditions for the employees. The problematical cellhouse numbers of an earlier era - thousands of cells, tens of thousands of electrodes handled per day, and hundreds of workers - have been cut by an order of magnitude. To a lesser extent, the need to reduce electricity costs has also played a role. The fully automated compact cellhouse is now the only practical choice for a new zinc plant, and usually is also the best choice for expanding an existing one.

This paper will not discuss the technical details or offer a comparison of the cellhouse technologies available. It is sufficient to say that they will all produce high quality zinc cathodes with a minimum of labour and electrical power when properly operated. Although equipment designs are still improving, further development of this fairly mature technology is likely to be a series of incremental improvements rather than large gains or radical changes. This situation is likely to continue for the foreseeable future.

We will instead discuss the issues beyond the design of the mechanical and process equipment, which can have significant effects on capital and operating costs. Some important differences between a greenfield and brownfield project are presented, followed by key considerations during site selection. The major components of capital cost are discussed, with emphasis on electrical power and the design of the cellhouse building. Finally, the factors that affect the project schedule are discussed.

GREENFIELD VERSUS BROWNFIELD

There are two distinct situations in which a company may be building a new cellhouse. The first is the common situation of a brownfield project at an existing plant. This is either a complete replacement of an old cellhouse, or an expansion of an existing facility.

The motivations for replacement of an old cellhouse are nearly always the same: reduce manpower for manual cathode stripping, improve workplace conditions for the employees, and reduce power costs. The need to expand capacity or improve product purity can be factors as well. It is also frequently the case that an old manually stripped cellhouse has high maintenance costs and may need major repairs to stay in operation.

Construction of a new cellhouse to add to existing capacity is a less common situation. Unless the existing cellhouse is reasonably new, the owner usually gives serious consideration to moving all the production to the new cellhouse for the reasons given above. As a minimum, the new facility is usually designed to allow expansion to full production capacity at a later date.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 539

The second scenario is the greenfield site for a completely new plant. In this case the cellhouse is part of a complete process plant which may also include a mine and concentrator.

There are some important differences in the approach needed by the project owners in the two situations, and therefore in the role of the engineering contractor.

In a greenfield project, the owner may or may not have experience in the zinc industry. The cellhouse will be a single component of a large project and will not receive undivided attention from the owner's team. The emphasis will likely be on a complete package which can be relied upon to start promptly on the scheduled date and perform according to the guarantee. This project does not have the fallback position of running an old cellhouse during commissioning difficulties.

The brownfield project owner's team is more likely to be appointed from within the ranks of the operating plant. This team is likely to have been looking at technology and equipment for many years, and will want to be involved in the detailed design process. It will also spend time preparing the existing workforce for the new plant by reviewing drawings and process control concepts with them, and by involving them in HAZOP studies. Conflict of priorities between the operating plant and the new project in this case can pose an added threat to the project schedule.

The environmental demands in the two scenarios can also be very different. The greenfield project will need to allocate more resources to activities like environmental permitting which the brownfield owner may be able to deal with in a matter of hours or days. Months may be spent on presentations to investors, governments and future neighbours to explain the impacts of the project on them and to assure them that the emission control equipment at the new plant will be the best in the world. A detailed environmental impact study and the necessary land use permits can take months or years to put in place. The impact of these activities on the project schedule cannot be ignored.

Another significant difference between the two scenarios is the interest rate which the project must carry. The replacement / upgrade project can borrow money at a relatively low interest rate since the project risk is limited to the construction and commissioning of the cellhouse only. In contrast, the greenfield plant pays an interest rate based on the risks involved from mine to market. These can include ore grade and quality, political risks, exchange rates, and construction / commissioning risks for the entire plant and also the market price risk associated with bringing new zinc production into the world's supply. The greenfield project will normally be borrowing money at a higher interest rate because of these factors.

Site Selection

In a greenfield plant and in retrofit situations, there is considerable flexibility in locating the cellhouse because the only physical connections to the leaching plant are the electrolyte supply and return pipelines. This allows the cellhouse to be remote from the other plant areas. The exception to this occurs when the re-use of the existing melting and casting facility "anchors" the new cellhouse to a limited range of locations. Even in this case the benefits of salvaging old equipment must be weighed together with other site selection criteria, outlined below.

540 LEAD-ZINC 2000

Soil Conditions

Finding good ground for cellhouse foundations is not usually a problem on a greenfield site. Generally the new site has been carefully selected to avoid poor ground conditions. It is possible, however, that the over-riding factors were the proximity to a port or concentrator, or other considerations than good ground. >

In a brownfield site, the conditions can be much more challenging. An old plant that has been modified and expanded over the years has probably already used up the best building sites. The new cellhouse could have to face the costs of building on inferior ground. Conditions that can render the ground inferior are :

• Sulfate Contamination : Sulfate contaminated ground is a possible legacy of an old plant which was not vigilant about controlling electrolyte leaks in the past. This can require the use of sulfate resistant concrete at best or, at worse, the excavation of large quantities of unsuitable ground.

• Toxic Residues : Contamination by toxic residues is another possible legacy of practices which are no longer acceptable. Removal and treatment of contaminated soil can seriously retard the construction schedule. It can be expected that environmental authorities will be involved, who will require a complete plan of action prior to excavation of the material and will monitor the site amelioration activity. The length of the delay will depend on the extent to which the owner and authorities are prepared to deal with the situation and act quickly.

• Archeological Interest: A situation with potential for even longer delays in construction is the discovery of an archeological site where the cellhouse is to be located. In this case, the owner may decide to look for an artifact-free alternative location for the building.

• Forgotten Foundations and Services : The brownfield project, in particular, must be prepared for unexpected discoveries of underground piping, other services, old foundations and other industrial archeology situations. These aspects are typically neglected in the preliminary engineering phases. When discovered during detailed engineering or construction they may translate into major added costs.

Prevailing Winds

The prevailing winds at the site should receive consideration for two reasons. The principal factor is the acid mist drift losses from the electrolyte cooling towers and also from the cells themselves in the case of an open walled cellhouse. These losses are minimized with modern cooling tower design that incorporates mist eliminators, and with good maintenance. A more serious short term emission can be caused by a power failure at the cellhouse that causes redissolution of the cathodes, generating larger amounts of mist. The effects of mist on the surrounding plant areas and off-site areas must be examined when locating the cellhouse. Also to be considered is the phenomenon of recirculation whereby contaminated air exhausted from the building is carried by the wind back to the air inlets of the building. Recirculation is influenced by building size and shape and by the location of the air inlets and outlets (5).

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 541

The second factor is the effect that an existing plant can have on the new cellhouse. The worst offender in this case is likely to be an old roaster that emits fugitive dust and sulfur dioxide fumes. The cellhouse construction project can experience delays if work areas are poorly located or if the construction workforce is not prepared for the situation.

Access

Access considerations usually center around construction needs, and around the needs of operating, maintenance and shipping activities that will come later. The guidelines for good access include allowance for:

• Construction material storage • Access for cranes and trucks • Subassembly areas, particularly for cells • Employee access to offices, control rooms, lunchrooms and changerooms • Avoidance of conflicts with existing operations.

Access for product shipping in particular should be looked at carefully. It is generally better to separate truck traffic that is carrying zinc products from in-plant traffic. Firstly, it is probable that the shipping driver is not a plant employee and may be unfamiliar with the plant. This can cause problems in an emergency situation as well as being a nuisance during normal operations. Secondly it is best to prevent exposure of outside trucks to possible contamination on the internal plant roads. The contamination can come from process residues, concentrates or other sources. Trucks can be washed down before leaving the property, but it is better to avoid contaminating the trucks in the first place.

CELLHOUSE COSTS

Cellhouse costs can be divided into three main categories: the cellhouse building, the mechanical systems, and the electrical systems. Figure 1 shows the approximate ranges in which these costs are expected to fall, based on SNC-Lavalin experience, as a percentage of the total direct cost. The great success of the developers of compact cellhouse technologies is immediately apparent. The working equipment of the cellhouse accounts for 80% or more of direct cost, with the building that houses the equipment accounting for the rest.

It is difficult to generalize, even to the extent shown in Figure 1, because of the many differences that distinguish one project from another. Electrical equipment costs in particular can vary widely. It is assumed that every project, whether new or retrofit, will require new transformer rectifier sets. However, a greenfield project faced with bringing in new power will incur a higher cost for electrical equipment and also a higher overall cost. Its mechanical equipment cost will be a smaller percentage of the total than that of a retrofit of an existing plant with power infrastructure in place. The spare capacity philosophy for the transformer rectifier sets can have an even bigger cost impact. In general terms, the new greenfield plant with 100% redundancy of transformer/rectifiers is likely to be the one with an electrical cost component of 25% and a mechanical cost component of 55% of the total direct cost.

542 LEAD-ZINC 2000

Figure 1 - Cellhouse Cost Breakdown

MECHANICAL EQUIPMENT

The mechanical equipment is fairly predictable from one project to another. The mechanical package is assumed to include concrete cells and cell tops, anodes and cathodes, electrolyte cooling towers, electrolyte circulation system, ventilation equipment, cranes, cathode stripping machines and anode cleaning machines. One might think that cost variations should only be those associated with equipment redundancy provisions, materials of construction preferences, and piping layout. However, one of the biggest cost impacts on a cellhouse project can be that of metal prices set on world exchange markets. Electrodes alone account for about 15 -18% of the total direct cost. Of this about one third is the cost of the metals Al, Pb, Ag and Cu. Add to this another 5 - 7% of the total direct cost for stainless steel equipment and piping, whose price responds promptly to changes in Ni and Cr prices. Add a further 15% of the total direct cost for transformer-rectifiers and bus bars, of which about one tenth is the cost of Al. It is quickly seen that price movement in these non-ferrous metals has a major impact on cellhouse equipment costs.

POWER SUPPLY

The electrical supply equipment is the second biggest equipment package after the mechanical equipment. The package includes the transformer/rectifier (T/R) sets and bus bars for providing DC power to the cells, the associated power factor correction equipment and harmonics filtration equipment, and the service transformers and circuitry. This equipment will typically comprise about 15 - 25 % of the cellhouse cost. The cost is heavily influenced by whether or not an electrical substation is included, and by the redundancy requirements for the T/R sets. A plant can choose to have no spare T/R capacity, partial redundancy, or full redundancy. Such cost savings must be weighed against the cost of lost production when a T/R set fails.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 543

The selection of the electrical equipment can play a major part in negotiating the price paid for electricity for the plant. The zinc cellhouse is a major consumer of electricity and may be the most important load on the distribution grid in the area. The characteristics of the rectifiers that affect the performance of the grid must be controlled much more closely than would be demanded of a minor consumer.

Outside of the plant boundaries, the security of the power supply must be reviewed. In co-operation with the power provider, redundant power sources, substations and transmission lines should be examined carefully and various scenarios analyzed for costs and benefits. Ideally, upgrading costs should be shared among all the beneficiaries of the upgrading. There is a danger, however, that the upgrading cost will be borne by the cellhouse project, either directly or indirectly. In the absence of a users' agreement, this becomes more likely.

Given today's technology and rectifier costs, the new plant will use thyristor rectifiers. Also, given the penalty in power costs for a low power factor, it will be necessary to install power factor correction banks. Similarly, the utility (and plant) will impose limitations on harmonic current levels. The impact on the electrical grid, and the impact on the plant of constraints imposed by the power utility, must be well understood by the engineer. It is important to obtain expertise in power system studies to ensure that the equipment is properly specified for the supplier and to negotiate knowledgeably with the utility to the future advantage of the owner.

It is generally preferable that the supply of the T/R sets, the power factor correction equipment and the harmonic filters all be sourced from a single supplier, and that the performance specifications be part of the contractual obligations. The design and supply of large rectifiers is now concentrated among several large multinationals and they are capable of undertaking sophisticated design, supply and, if necessary, installation contracts. They therefore can be provided with functional specifications requiring that they meet certain power system and plant conditions. However, if the specifications are not well written, properly evaluated and well managed, the equipment purchased may be less than optimal.

The most expensive power for any system is that supplied during the peak hours and peak periods of the year. There are also unforeseen power system emergencies which cost the power system dearly in purchasing additional capacity. An electrolysis process can assist the power system under these circumstances by quickly cutting power consumption. It becomes a question of trading off the loss of zinc production for a power price reduction. This capability should be used to gain the greatest possible benefit in negotiating electricity prices with the utility. The owner should also keep in mind that the net present value of future power bills will probably exceed the capital cost of the new cellhouse. This provides a strong incentive to give the power contract the closest scrutiny.

Installation of the electrical equipment involves a number of important and increasingly costly details such as:

• Ground grid • Transformer oil spill containment pit • Power supply routing - both high voltage and low voltage • Fire detection and protection • Provision for future expansion or other upgrading.

544 LEAD-ZINC 2000

Planning for the future in this area is very important, since a large part of the costs of upgrading or expanding the electrical equipment can be designed in or designed out at the early stages of the project. Plants have in some cases boxed themselves in by not planning a cost effective layout for the rectifier units, their associated power factor correction/harmonic filters and a properly arranged high voltage substation with good maintenance access.

THE CELLHOUSE BUILDING

A zinc cellhouse building is designed to accomodate its equipment: rows of concrete cells, cranes to the handle electrodes, machines to strip the cathodes and clean the anodes, and tanks and pumps to circulate electrolyte. In some cases the building also accomodates electrolyte storage tanks and supports the electrolyte cooling towers. This equipment is universally employed in new zinc cellhouses. Alternative cell concepts such as the fluidized bed cathode have surfaced from time to time but none has gained commercial acceptance. Since the current state-of-the-art cellhouse is universal, discussion of building issues will be confined to this well-known technology.

Building Cost

Building costs are typically 15% to 20% of the total capital outlay for a new cellhouse, depending on location. This would apply to a facility built on a new site and requiring minimal demolition. Building costs are taken to include excavation and civil works, all concrete except for cells and tanks, steelwork, and architectural finishes. The breakdown of the building cost for a concrete construction case is shown in Figure 2. This breakdown could be called typical were it not for the significant differences in the costs of acid resistant coatings. A plant that prefers an acid brick membrane in the basement could spend twice the 12% cost indicated in Figure 2.

Figure 2 - Building Cost Breakdown Example (Concrete Construction)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 545

Major Decisions

The building cost should be the smallest of the cost components of a cellhouse project after the mechanical and electrical equipment. Timely definition of the building design is of vital importance in ensuring that building costs do not escalate. The earliest construction activities are those related to the building; therefore, the layout and design of the building must be finalized very early after the project begins. The biggest decisions, apart from the selection of the actual process technology, concern the cellhouse building. These decisions can involve significant engineering time for comparison of alternatives. They therefore affect not only the direct cost of erecting the building but, if not in place early in the project, can lead to schedule delays, the most feared kind of project cost. For this reason the major building decisions, listed in Table I, should be taken before the final appropriation of funds.

Table I - Major Cellhouse Building Decision Checklist: Major Cellhouse Building Decisions That Can Impact on the Cost and Schedule of a Project

Will the cells sit on ground slab or will there be a basement? Where will the cooling towers be located? How many overhead cranes will there be? What will the crane span be? Will the structure be designed for an oversized crane to allow for cell erection? How much capacity will be allowed for solution storage? Will storage capacity allow for draining cell electrolyte in case of emergency? Will tanks and pumps be in the main building, outside, or in adjacent buildings? Will all tanks and pumps be located at ground level, or will some be on elevated levels? Will the tankage be in concrete, steel, or FRP? Where will the machine bays be located? Will the building be all concrete, all steel, or a combination of each? What kind of coatings and membranes will be used on the concrete and steel?

Cells on the Floor

Will the cells sit on ground slab or will there be a basement? The placement of electrolytic cells on the ground, slightly off the ground, or with a full height basement underneath has a significant impact on the building cost and the operational philosophy of the cellhouse.

Electrolytic cells have historically been mounted off the ground at full basement height. Some of the arguments for elevating the cells include the need to electrically isolate cells from spills in the basement, to protect workers from stray current, to adjust the level of the cells, to inspect for leaks and perform repairs, and to limit pumping costs by employing gravity flow. With the advent of polymer concrete cells and of synthetic cell liners, the frequency of leaks and the need for electrical isolation have been drastically reduced.

The type of ventilation system chosen for the cellhouse can also influence the positioning of the cells. For example, where air is drawn into the space below the cells and discharged through the roof, the open basement acts as a plenum for uniform air flow between the cells. This arrangement would not be required if, for example, the cellhouse is in a milder climate and is not totally enclosed or if other ventilation arrangements are selected.

546 LEAD-ZINC 2000

To our knowledge, elevated cells are universal in the zinc industry. Given the significant building cost saving (-10% of building cost for on the ground versus 2 m off the ground), the mounting of cells on or close to the ground should be explored for future cellhouses.

Location of Electrolyte Cooling Towers

Where will the cooling towers be located? The electrolyte cooling towers can be separated from the building, attached to the building, or located on the cellhouse roof. In current practice they are used for drawing contaminated air out of the cellhouse for acid mist control. In some cases, a restricted site area can dictate the location of the cooling towers. There are trade-offs to be made in terms of electrolyte pumping costs, support structure costs and piping costs. An easily overlooked consequence of locating the cooling towers on the roof is the added building height needed to accommodate larger roof beams. If steel construction is used, an added 3 m or more of height can be needed.

In northern climates there is the added complication of operating pipes and valving which are subject to extreme temperature variations during process starts and stops. Access for cleaning is also very important, particularly where weather conditions are extreme.

Cell and Crane Configuration

How many overhead cranes will there be, what will the crane span be and will the structure be designed for an oversized crane to allow for cell erection? The primary crane parameters, span, travel distance and capacity, are functions of the cellhouse technology used, the cellhouse capacity, the daily operating time for cathode stripping and the number of cell rows. These will be established in cooperation with the supplier of the cellhouse technology. It is also important to consider designing the crane rails to accomodate a temporary crane for installing the heavy cell pieces during plant construction. This would also be valuable in the unlikely event of an irreparable cell failure, and also to allow for future expansion. In this case the foundations, columns and beams would all be made larger. Since building foundations are the first activity in the construction schedule, these decisions must be made very soon after project initiation.

Tanks and Pumps

How much capacity will be allowed for solution storage, and will the storage capacity allow for draining the cell electrolyte in case of emergencies? Will the tanks and pumps be in the main building, outside, or in adjacent buildings? Will all tanks and pumps be located at ground level, or will some be on elevated levels? Will the tankage be in concrete, steel, or FRP? These questions, like the others in the list, are important because they affect building layout and foundation design. The most economical configuration can be to stack selected tanks vertically in order to limit building area. Access for maintenance must also be considered, however.

Machine Arrangement

Where will the machine bays be located? A cathode stripping bay and an anode handling bay need to be incorporated into the cellhouse. A smaller cellhouse with 2 rows of cells would locate the stripping machines at one end of the cell rows and the transformer-rectifiers and main bus bars at the other end. A larger cellhouse could have 4 rows with stripping machines at one

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 547

end, or two long rows with machines in the middle. The anode handling area can be more problematic. The most economical design will be to have the anode handling area accessible by the main overhead cranes and contained within the rectangular shape of the building. However, building layout constraints may dictate that the anode bay must be an appendage to the main building. This increases building costs.

Compared with a copper refinery, a zinc electrowinning cellhouse has less onerous machine requirements, largely on account of having fewer (larger) electrodes to handle. The copper business refers to its machine area as a machine "park" and has devoted considerable effort to find the optimum layout for it. It is almost invariably centrally located with rows of cells extending on either side. A recent project implemented a novel variation of this design, with a single circuit of cells in a C-shape around the central machine park (6). This design gave total flexibility so that any machine could serve all the cells, using two parallel overhead cranes. This design was claimed to be substantially cheaper than having a single long crane aisle.

Materials Decisions

Will the building be all concrete, all steel, or a combination of each? What kind of coatings and membranes will be used on the concrete and steel? Building materials costs must be evaluated on a case by case basis. The cost and availability of steel and concrete vary greatly around the world, as do labour costs for erecting buildings in one or the other. Both must be protected from attack by acid mist. Materials cost and labour cost for the application of coatings or membranes will again depend on location. Also, design lifetime of the project affects decisions on coatings. The following two examples give an idea of how choices were made on two projects.

Case 1 : Poor Steel Availability

This cellhouse was designed with cooling towers located adjacent to the main building. Concrete was readily available locally, but steel had to be imported. A decision was needed on whether to erect the structure (above the foundation) in concrete or in steel. The comparative features of the two options, including their relative direct costs, are shown in Table II. Concrete construction offered a substantial cost saving but had many disadvantages. Steel had many advantages but one overriding disadvantage, a longer total time for construction. In the end, the advantages of steel were judged to be insufficient to offset the cost advantage of concrete.

Table II - Comparison of Steel and Concrete Options for a Cellhouse Building (Case 1)

Option 1: Structural Steel Option 2: Structural Concrete Relative cost of cellhouse building : 109% Smaller members, giving larger working area

Lower engineering hours Longer lead time for procurement & shop fabrication Shorter construction time at site Smaller construction area needed Less interference with existing plant Flexible for future expansion Higher maintenance cost

Relative cost of cellhouse building : 100% Larger members, giving smaller working area Higher engineering hours Shorter lead time to begin construction

Longer construction time at site Larger construction area needed More interference with existing plant Difficult for future expansion Lower maintenance cost

548 LEAD-ZINC 2000

Case 2 : Good Steel Availability

This cellhouse was designed with cooling towers on the main cellhouse roof. It was determined that the structure would be all cast-in-place concrete up to the crane girder level. It remained to decide on the construction of the upper walls and roof. Three of the options considered and their relative costs are shown in Table ΠΙ. In this case, both concrete and steel were readily available, although the nearest supplier of precast concrete was located some distance away. All exposed concrete and steel were assumed to be treated with acid resistant coatings. Increased column and footing sizes for the building were factored in for the concrete options. Increased wall height on account of increased roof truss size was factored in for the steel option. The final cost differences were not large, as seen in column 2. A more rigorous comparison is shown in column 3, where repainting of all steel structures after 7 years is assumed, allowing this option to attain the same design lifetime as the concrete. The best option appeared to be cast-in-place columns and beams and precast roof panels. The ultimate decision was in favour of steel, not because of cost but because it could be erected faster than the concrete options.

Table ΠΙ - Relative Costs for Three Cellhouse Building Options (Case 2)

Option for Upper Walls and Roof Cost Relative to Cost Relative to Cost of Option 1 Cost of Option 1 with NPV of repainted steel after

7_y. Steel columns, purlins, trusses & roof 100% 100% deck CIP columns, beams & roof slab 105 % 103 % CIP columns & beams, precast roof 100% 99%

Fundamental Building Cost Factors

The fundamental design criteria that affect building costs include items like the current density and number of cell rows, that relate to the process that is accommodated inside the building, and to things like the minimum temperature and snowload, that relate to the environment outside the building. Table IV lists the fundamental design criteria for a zinc cellhouse, dividing them into those related to process requirements and those related to location.

Table IV - Fundamental Zinc Cellhouse Design Criteria that Impact on Building Cost

Fundamental Design Criteria Related Fundamental Design Criteria Unrelated to Process that Affect Building Cost to Process that Affect Building Cost

Electrode length Seismic zone Electrode width Distance to bedrock Number of electrodes per cell Minimum outside temperature Number of cells per row Maximum wind velocity Number of rows Maximum snow load Current density Current efficiency Fraction of time at full current Cathode plating time Anode cleaning frequency

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 549

Process Related Design Criteria

The interplay of process design criteria with cellhouse costs is an interesting subject. Indeed, the development of cellhouse technology over the last century, for zinc as well as for other metals, has been driven by just this interaction. The culmination of this effort is today's modern technologically compact cellhouse. The technology provider for a new cellhouse brings to the project a highly developed product. The process related fundamental design criteria in Table ΠΙ have been fine tuned and optimized to give the best balance of cost and performance. However, for each individual project there will be a range in which each variable can move to give the optimum design for that project. The customizing of these variables to suit an individual case may not have a major impact on building cost, but should be undertaken with an understanding of how cost is affected by each variable.

To simplify the discussion, a plant of fixed capacity is assumed. Because cellhouse capacity depends directly on certain fundamental design criteria it is useful to keep in mind the cellhouse capacity correlation:

C = A * C D * C E * f * 1.068 xlO"2 (1)

where: C = plant capacity in t/y Zn cathode A = total cathode area in the tankhouse in m2

CD = current density in A/m2 cathode area CE = current efficiency f = fraction of time at full current 1.068 x 10'2 = theoretical zinc deposition rate in t/A-y

Expressed in greater detail, Equation 1 can be written:

C = l * w * n * N * C D * C E * f * 1.068 xlO"2 (2)

where: 1 = length of the cathode in m w = width of the cathode in m n = number of cathodes/cell N = number of cells

The existence of this correlation for cellhouse capacity somewhat complicates the discussion. Many of the variables in Equations 1 and 2 are those that are also on the list that can affect building cost. For a cellhouse of fixed capacity, however, the variables in Equations 1 and 2 cannot vary independently. For instance, if the fraction of time at full cell current decreases, another variable, such as the number of cells, must increase if capacity is to remain unchanged. In evaluating impacts on building cost it is therefore necessary to look at the combination of variables that, while changing individually, maintains constant capacity. The pair of changes in the present example has the effect of adding to the building cost, because an increased number of cells requires a larger plan area. Such a provision is necessary for a plant that needs to limit power costs by operating at full current during off peak hours and by cutting back during peak hours.

Many fundamental design criteria have nothing to do with the capacity equation. Evaluating their impact on building cost is therefore easier. They can vary independently of other design criteria. For example, as the snow load increases, the roof costs increase.

550 LEAD-ZINC 2000

Amongst the variables listed in Table IV, the following deserve particular attention.

Current Density

Increasing the current density usually decreases the building size for a given zinc production. This assumes that the total cathode area is reduced, translating into fewer and/or smaller cells. Most modern cellhouses operate in the range of 400 - 500 A/m2. However, current densities ranging from <300 A/m2 to >750 A/m are encountered.

Although an increase in current density can result in a smaller building, other factors, mainly power requirements, dictate that newer cellhouses usually have design current densities in the range of 400 - 500 A/m2. As the current density is raised, the cell voltage rises. This translates into increased energy consumption per tonne of cathode and increased cooling tower duty because of higher heat generation in the cells (2). The optimum current density is site specific and is usually based on the cost of power rather than minimizing building costs. Plants that practice power peak shaving operate at low current density during peak hours and high current density during peak hours (7-9). Such plants will need more cellhouse area, not less, to minimize power cost.

Electrode Spacing

Reducing the electrode spacing decreases the building size for a given zinc production. The standard electrode spacing in modern compact cellhouse designs is 90 mm (center to center). Spacings as small as 80 mm have been reported. A 10 mm decrease represents an increase of approximately 12% in zinc production per cell, based on 125 cathodes per cell.

Reducing the electrode spacing also reduces power requirements. According to Honey et al. (10), the electrolyte IR drop accounts for about 14.5% of the total cell voltage at 90 mm spacing. A 10 mm reduction would theoretically correspond to a 1.6% power saving, which is supported by Figure 6 in the same paper.

Closer electrode spacing increases the potential for a low current efficiency because of electrical shorts and poor electrolyte mixing between the electrodes. A plant considering reduced electrode spacing will require high quality DC current and good cellhouse management to ensure electrode physical quality. Current efficiency monitoring, possibly on an electrode by electrode basis, may be required.

Electrode Shape

As the electrode length increases, not only does cell depth increase but so does the height needed above the cells for crane clearance. This has a double impact on wall height. However, for constant capacity, another factor will have to change to maintain the same total cathode area. This could be the electrode width, the number of cells or the number of electrodes per cell. A decrease in any of these would decrease the total plan area, offsetting the cost impact of wall height. The net effect of longer electrodes should be a reduction in building costs. However, increased electrode movement and deviations from optimum electrolyte flow place a practical limit on electrode length.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 551

Cathode Plating Time

Zinc electrowinning is usually conducted with a 48-hour plating cycle, but a few plants operate on cycles as low as 24 hours and as high as 72 hours. There is a strong incentive to extend the plating cycle, because this translates into reduced stripping machine time and crane time. Fewer or smaller machines and cranes, in turn, translate into reduced equipment and building costs. In general, a long plating cycle requires excellent control of electrolyte quality. The tradeoff for longer plating time may be larger electrode spacing or lower current efficiency.

PROJECT SCHEDULE

A complete cellhouse project will take two to three years from the initiation of a feasibility study to the start of zinc production. The main activities in the project schedule will be:

Feasibility study and appropriation of funds Engineering Fabrication of automated equipment Construction Commissioning

6-12 months 12 months 14 months 16 months 2 months

The critical path items on the project schedule will normally be the delivery of the automated equipment, delivery of the transformer/rectifier equipment and the construction of the electrolysis cells.

The importance of meeting the scheduled completion and start-up dates cannot be over emphasized. A late project causes embarrassment to both the owner and the project manager, but the impact is ultimately financial. A delayed construction project is a non-performing asset which is usually absorbing money at the same time as it fails to produce the expected zinc.

To demonstrate the effects of a delayed project, a pessimistic case of a 6 month delay will incur the following added costs, expressed as percentages of the total capital cost, assuming typical recent interest rates, capital and operating costs, and zinc market price:

Interest on borrowed capital Lost profits Interest on stockpiled concentrate Engineering/construction overheads Operating costs of an old plant Remedial work

4 % 2 % 1% 1-2% variable variable

Although these figures do not necessarily apply to all cases, they indicate that a project delay can have a serious effect on the total project cost. The sources of delay can vary widely, but the most common problems are mistakes and omissions made in the first few months, and changes in the scope of the project.

The replacement cellhouse has the fallback position of using the old facility in the event of a construction delay. This will provide the necessary zinc production, but it does come at a

552 LEAD-ZINC 2000

price. The obvious cost is the use of extra labour for stripping cells. The less obvious costs are the maintenance of an old facility at a time when maintenance personnel are required for other work in the new facility. Other costs are the delay in salvaging of the anodes and cathodes from the old cellhouse.

Similar arguments apply to the situation where the owner of the new plant is unable to make it run as expected. This is an instance where the value of having properly trained plant operators and competent technical help becomes clearly evident. An investment in training may never have an obvious payback, but failure to provide it can have clear and painful consequences

SUMMARY

Zinc electrowinning technology has reached a highly optimized state with the introduction of the compact, fully automated cellhouse. The challenge is to tailor the mechanical and electrical equipment and the cellhouse building to the unique demands of each project and still create a plant that functions optimally in all respects. Greenfield and brownfield projects need a different approach to project management as well as to site selection and environmental matters. Correct and early definition of the cellhouse building and electrical equipment are as vital as the selection of the technology for the project. The negotiation of the power contract is particularly important in laying the groundwork for long term financial performance. Finally, the project schedule requires vigilance and co-operation from all parties to avoid costly delays and changes of scope.

ACKNOWLEDGEMENTS

The assistance of a number of engineers in the Industrial Division of SNC-Lavalin in developing the concepts and examples presented in this paper is gratefully acknowledged.

REFERENCES

1. Y. de Bellefroid and R. Delvaux, "New Vieille Montagne Cellhouse at U. M. Balen Plant, Belgium", Lead-Zinc-Tin '80, J. M. Cigan, T. S. Mackey and T. J. O'Keefe, Eds., The Metallurgical Society of AME, Warrendale, PA, U.S.A., 1980, 204-221.

2. R. N. Honey, "Evolution of Electrowinning at Trail", Paper presented at the 13th Annual Hydrometallurgical Meeting of the Metallurgical Society of CIM. Edmonton, Canada, August 1983.

3. A. Caufriez, M. Dubois and E. Lejay, "Zinc Cellhouse, Concept and Practice at Union Miniere", Zinc and Lead Processing, J. E. Dutrizac, J. A. Gonzalez, G. L. Bolton and P. Hancock, Eds., The Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 337-352.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 553

4. Y. Lefevre and M. R. Pieterse, "The State of the Art and Constant Feasibility Upgrading of the Zinc Electrowinning Cellhouse of Asturiana de Zinc at San Juan de Nieva, Spain", Zinc and Lead Processing. J. E. Dutrizac, J. A. Gonzalez, G. L.Bolton and P. Hancock, Eds., The Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 367-377.

5. W. D. Baines, P. Brzustowski and J. A. Davis, "CFD and Physical Hydraulic Model Predictions of Contaminant Dispersion from Process Operations", Copper 95-Cobre 95. Vol. II - Mineral Processing and Environment. A. Casali, G. S. Dobby, M. Molina and W. J. Thobum, Eds., The Metallurgical Society of CIM, Montreal, Canada, 1995, 273-288.

6. J. Garvey, B. J. Ledeboer and J. M. Lommen, "Design, Startup and Operation of the Cyprus Miami Copper Refinery", Copper 99-Cobre 99. Vol. ΙΠ - Electrorefining and Electrowinning of Copper. J. E. Dutrizac, J. Ji and V. Ramachandran, Eds., The Minerals, Metals and Materials Society of AIME, Warrendale, PA, U.S.A., 1999, 107-126.

7. T. Yamada, H. Nagata, A. Hosoi, M. Kato and R. Togashi, "Recent Improvement of Electrolytic Zinc Production at Iijima Refinery", Zinc-Lead '95, T. Azakami, N. Masuko, J. E. Dutrizac and E. Ozberk, Eds., Mining and Materials Processing Institute of Japan, Tokyo, Japan, 1995, 589-598.

8. O. Yasuki and K. Watanabe, "Electrolytic Zinc Production at Annaka Refinery", Metallurgical Review of MMIJ. Vol. 12 (1), 1995, 51-62.

9. N. Ogata, "Electrolytic Zinc Production at Hikoshima Refinery", Metallurgical Review of MMIJ, Vol. 12 (1), 1995, 96-109.

10. R. H. Honey, R. C. Kerby and R. C. Legge, "Zinc Cellhouse Optimization", Zinc '85, K. Tozawa, Ed., Mining and Materials Processing Institute of Japan, Tokyo, Japan, 1995, 349-363.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 555

ASTURIANA DE ZINC EXPANSION AT THE SAN JUAN DE NIEVA PLANT FOR A ZINC PRODUCTION OF 440,000 TONNES PER YEAR

F. San Martin Asturiana de Zinc, S.A.

33417 San Juan de Nieva Castrillon, Asturias, Spain

F. Tamargo and Y. Lefevre Asturiana de Zinc Tecnologia San Juan de Nieva, Apdo. 178 33400 Aviles, Asturias, Spain

ABSTRACT

The San Juan de Nieva smelter was commissioned in 1960 with an electrolytic zinc capacity of 15,000 t/y. After a number of minor expansions, the capacity was increased to 320,000 t/y in 1991 and will reach 440,000 t/y by August 2001. The expansion involves the construction of a 900 t/day Lurgi roaster and a new acid plant. The leaching will be fully revamped and modernised, including the construction of a new building for residue treatment and for j arosite precipitation with a separate circuit for silver-containing concentrates. The project also includes the construction of a fully automated 137,000 t/y zinc cathode tankhouse fitted with 3.4 m2 cathodes designed by Asturiana de Zinc, and based on the existing tankhouse and introducing some improvements. New melting and casting facilities, also designed by Asturiana, will be constructed.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

556 LEAD-ZINC 2000

INTRODUCTION

In July 1999, Asturiana de Zinc (AZSA) decided to expand its electrolytic zinc facilities located in San Juan de Nieva by 120,000 t/y to reach a production of 440,000 t/y. This means a 37.5% increase over the present production capacity of 320,000 t/y and will assure that AZSA is one of the world's biggest producers (see Table I) and will make the San Juan de Nieva plant the largest in terms of production capacity.

Table I - Main World Zinc Producers

Position Producer Country Tonne/year No. of Plants

1 Pasminco Australia 651,000 5 2 Asturiana de Zinc Spain 440,000 1 3 Union Miniere Belgium 430,000 2 4 Cominco Canada 410,000 2 5 Noranda Canada 395,000 2

The project is being developed by AdZ Tecnologia, the technology division of Asturiana de Zinc, in close collaboration with the plant's production team. Except for the roasting and sulphuric acid areas, where Lurgi technology is being applied, the other areas will use Asturiana de Zinc proprietary technology. The budget for this project is 28,000 million pesetas (US$ 169 million), which means an investment cost of $1,400 per each additional tonne of zinc produced. The total estimated execution time is 27 months. The start of full production is expected to take place in November 2001. Between 70 and 75% of the equipment will be nationally manufactured. The civil work and erection will be locally or regionally contracted, with an average of 500 additional personnel subcontracted during construction. The plant outputs, before and after the expansion are given in Table II.

Table II - Effect of AZSA Expansion on Production Levels

Product

Zinc Sulphuric Acid Concentrates

Present Capacity (t/year)

320,000 497,000 596,000

After Expansion (t/year)

440,000 717,000 819,000

Increase

(t/y)

120,000 220,000 223,000

Increase

(%)

37.5 44.2 37

SCOPE OF THE PROJECT

In order to increase the production capacity of the plant, the following improvements are contemplated in the main production area:

• A new roasting and sulphuric acid facility • Modification of the present leaching installation to different stages for neutral leaching and

neutralisation • Building a new plant for the different acid leaching stages • Increase the residue filtration capacity • Expansion of the present purification installation • Automation of the leaching and purification processes

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 557

• Building a new totally automated cellhouse • Building a new melting and casting plant, including a new dross grinding installation and

the expansion of the zinc dust plant current capacity.

Additionally, a series of actions required to integrate the expansion project within the existing installations in non-productive areas are also contemplated. We describe hereafter the main characteristics of the new installations.

Storage of Concentrates

An increase in the capacity of the covered concentrate warehouse by 35,000 tonnes is planned.

Roasting Plant IV and Gas Cleaning

Storing and Feeding Systems

New concentrate belt conveyors between the warehouse and roasters No. Ill and IV will be installed with a capacity of 100 tonne/h. A concentrate storage hopper of 200 m3 size will also be built.

Fluid Bed Roster and Adjacent Installations

A roasting furnace with a nominal capacity for 900 tonne/day (grate area 123 m2), the same size as the one installed in plant III, will be built. An air blower with a capacity for 88,000 Nm3/h 2.500 mm.w.c. will also be installed.

Calcine Storage and Transportation Facilities

A calcine storage silo to store calcine with capacity for 6,500 tonne will be constructed, along with wet and pneumatic transport systems from the silos to the leaching plant for the calcine.

Thermal Recuperation Steam Boiler

A horizontal steam boiler for the forced circulation of gases to produce saturated steam, and having the characteristics of 50 tonne/h at 400 atm and 400°C including five evaporator bundles and two superheater bundles, will be built. A steam condenser of 40 tonne/h and the associated service installations will also be installed. A DM water plant of 90 m3/h capacity to feed the boiler will also be constructed.

Hot Gas Cleaning

Two cyclones as gas dedusters will be located at the boiler outlet; their diameter is 4,500 mm. Also one hot gas electroprecipitator with three fields, inclusive with ionisation electrodes, will be built.

558 LEAD-ZINC 2000

Wet Gas Cleaning and Mercury Removal

Four tubular wet electroprecipitators and service installations (two as first stage and two as second stage) will be constructed. Two towers for mercury removal from the gas, with their corresponding accessories, will be built according to AZSA design.

Sulphuric Acid Plant

The sulphuric acid plant has a nominal capacity for 900 tonne of monohydrate per day and it has been designed to comply with a maximum stack emission of 300 mg/Nm3 of S02.

Drying and Absorption Plant

The drying and absorption plant consists of an S02 blower (106,000 Nm3/h, 4,600 mm w.c), one drying tower for S02 (diameter 6,600 mm /11,500 mm high) supplied with a wire mesh filter on the top, and one intermediate absorption tower, (diameter 6,200 mm /15,500 mm high) supplied with a Candelas filter on the top.

Conversion Plant

One tray converter in four stages fabricated in AISI 304 H, (diameter 9,000 mm /17,000 mm high) will be built. The packing material INTALOX is placed on the trays to act as a support for the catalyser. Six gas-gas heat exchangers and one economiser to cool the gases before entering the final tower preheating system will be installed.

Water Cooling System

The closed circuit cooling of the industrial water by means of cooling towers will be utilized with a flow of 6,000 m3/h.

Leaching Plant

The future leachingplant is based, as is the present installation, on thejarosite process. The project contemplates two separate circuits; one to treat concentrates with high silver content and another one to treat the concentrates with low silver content. Both circuits are integrated after separation of the Pb/Ag residue.

Present Leaching Plant

The existing installation will be modified to be used as the neutral leaching and neutralisation stages, with independent circuits for concentrates with high and low silver contents.

Acid Leaching Plant

A new acid leaching plant will be built with an area of 2,650 m2 (106 x 25 m) and supplied with the following main equipment: 16 tanks with capacity of 300 m3, 4 thickeners of 21 m diameter and 4 thickeners with a diameter of 15 m. The tanks, fully designed by AdZ Tecnologia, will be manufactured in special stainless steel, without a coating. The thickeners will be build of concrete and provided with an anti-acid lining.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 559

Purification Plant

The present installation will be expanded by the construction of a 1620 m2 adjacent building, The equipment to be installed includes eight 340 m2 press filters, six 80 m2 press filters and eighteen 80 m3 tanks. The existing purified solution cooling facilities will be adapted to the new necessities by means of three 32 m2 cooling towers and two thickeners of 18 m diameter.

Filtration of Residues

Expansion of the existing facilities for jarosite filtration will take place with the incorporation of three horizontal belt filters for jarosite with 64 m2 filtration area and one horizontal belt filter for basic zinc sulphate with 64 m2 filtration area.

Electrical Facilities

The existing facilities need to be upgraded from the current consumption levels of 138 MWh (the equivalent to the consumption rate of a city with 400,000 inhabitants) to a consumption of 191 MWh afer the expansion. In order to comply with this requirement, the modification of the existing network, the expansion of the current SF6, and the creation of new substations in the roasting, leaching, purification and cellhouse areas is foreseen.

Compressed Air Facilities

The installation of three compressors with a capacity each for 6,000 Nm3/h at 7 kg/cm2 is anticipated.

Fuel Storage Facilities

The new installation will consist of two vertical tanks with a capacity for 200 m3 each, to store fuel, and one vertical tank with a capacity for 200 m3, to store gas oil.

Sulphuric Acid Transport and Storage Facilities

The existing facilities will be improved and one holding tank with a capacity for 10,000 tonne and one holding tank with a capacity for 15,000 tonne, to store sulphuric acid, will be built. Also, a second throughput pipe for acid transportation between the production and storage tanks, with a length of 1500 m, will be installed.

Effluent Treatment

The existing facilities need to be upgraded by the installation of three tanks with a capacity of 35 m3, one thickener with a diameter of 15 m and one silo to store lime with a capacity for 75 tonnes.

Auxiliary Services

Auxiliary services include a new electrical maintenance workshop, the upgrading of the existing general storehouse for spare parts and raw materials, and new changing rooms for staff and general washroom facilities.

560 LEAD-ZINC 2000

The Tankhouse

The tankhouse has been designed to reach a production of 137,000 tonnes of zinc cathode per year. The components have been designed for a potential maximum zinc cathode production of 190,000 t/y. The new "D" tankhouse is located next to the "C" cellhouse which was commissioned in May 1991. The space available is limited. The tankhouse footprint has been optimised at 0.02 m2 per tonne of cathode installed (design production). The supervision and housekeeping of the "D" and "C" tankhouses will be shared. This arrangement will reduce the manpower to 0.29, and preventive and predictive maintenance man-hours per tonne of zinc cathode produced to 0.09.

The new tankhouse will consist of two rows of 36 cells; each cell has 122 cathodes of 3.4 m2 area and a pitch of 90 mm. No major changes have been made to the electrode design because of the present good performance and excellent durability. The design of the cells has been improved to take advantage of the 10-year experience gained in the "C" tankhouse, and the design will reduce the cell cleaning time. The cathode stripping rate will be reduced to eight seconds per cathode. The noise level will be further limited because of a new concept of the zinc cathode sheet receiver and support structure. All the other concepts will remain the same. The anode conditioning area will be designed with two stages for dry and wet cleaning, and for flattening. The cycle time per anode will be from 20 to 35 seconds for conditioning. A semi-automatic anode handling device will be installed to reduce the time for inspection and repair.

The cells will be fed by external launders above the cell level. A single circulation tank on one side of the tankhouse will collect the overflow from the cells. Four acid cooling towers will be installed on the tankhouse roof. Their design has been optimised to improve the flow of air inside the towers. The housekeeping time will be further reduced by using rotating deminster systems.

The main automatic electrode handling cranes will use updated techniques to meet the requirements of the other equipment. This heavy duty electrode handling equipment will complete a cycle; i.e., handling one load each of electrodes out and in, in less than 250 seconds.

The existing control room will be used for both tankhouses, and will report the main production and quality parameters of individual cathode weighing, bundle weighing, temperature scanning, cell voltage drop, as well as other useful statistical measurements.

Two new transforectifier units feeding 132 kW will be installed. The DC regulation will be done from 110 to 300 V. The diode rectifiers are 120 kA per unit. The auxiliary power will be limited to 79.5 kWh per tonne/cathode.

The overall cellhouse operating philosophy will be maintained; there will be two independent operating rows, with one crane and one stripping machine per row. The optimum stripping time will be less than 7 hours per row, for half the cathodes per cell stripped. The capital cost for the "D" tankhouse, thanks to in-house engineering and the broad expertise and long experience of the engineering and production personnel, will be limited to 46,000,000 Euro. Construction will start in April 2000 and startup is scheduled for June 2001. The performance of the new tankhouse is expected to be in the same range as the "C" tankhouse. As a reference, in 1999, the "C" cellhouse produced 129,463 tonnes of cathode with a mean current efficiency of 93.2%, a mean kA consumption of 180.8, a mean voltage drop per cell of 3.37 V and a mean DC energy consumption of 3,006 kWh/tonne cathode.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 561

Melting and Casting

A new 24 t/h melting furnace will be installed in the new melting and casting area No. 3. A new 25 kg casting machine based on a new design will also be installed.

Zinc Dust

The test program, which is nearing completion, is expected to generate the full quantity of zinc dust for the total production of the plant. The dust will have a characteristic of Ο50=30μ with the size distribution shown in Figure 1. The expansion will also consist of revamping the existing zinc dust plant. Pneumatic transport will send the zinc dust to the various bins in the purification area.

Zinc Dust

Figure 1 - Particle Size Distribution of the Zinc Dust

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 563

ZINC AUTOSTRIPPING AT FALCONBRIDGE LIMITED KIDD METALLURGICAL DIVISION

J. Lenz and D. Ducharme Falconbridge Limited, Kidd Metallurgical Division,

PO Bag 2002 Timmins, Ontario, Canada P4N 7K1

ABSTRACT

Operation of the Kidd electrolytic zinc plant commenced in 1972 with a cellhouse capacity of 105,000 tonnes of zinc cathode. The original cellhouse layout consisted of 42 parallel rows for a total of 588 cells. As leaching capacity increased, the cellhouse was expanded to 630 cells. Zinc cathode was manually stripped from plant start-up until the development of a mobile automated stripping system in 1994. Machine development continued until a second unit was placed in production in 1996, from which point, 60 % of the cellhouse was being stripped with the automated system. The final phase of the project was implemented in 1999 with the commissioning of two more automated strippers. This paper describes the implementation of the automated stripping system and its impact on cellhouse productivity.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

564 LEAD-ZINC 2000

INTRODUCTION

The Kidd Metallurgical Division

The Kidd Metallurgical Division is an integrated metallurgical facility, located in Timmins, Ontario, Canada, which comprises a 10,000 t/d concentrator, a 145,000 t/y copper smelter and refinery, and a 150,000 t/y zinc plant. In addition to copper and zinc, the site produces 600 t/y of cadmium, 1200 t/y of lead contained in a lead/silver, residue, refined indium, and 100,000 kg/y of silver and 500 kg/y of gold, contained in anode slimes. Sulfur products produced are 580,000 t/y equivalent 100% H2S04 and 28,000 t/y of liquid S02. Total site complement as of March 2000 was 890 employees.

The Kidd Zinc Operations

The Zinc Operations, which commenced in 1972, comprise two 55 m2 Lurgi fluid bed roasters with a combined capacity of 750 t/d of dry concentrate, and a pressure leach autoclave with a 160 t/d concentrate capacity. The leach section treats the calcine, and pregnant solutions from the autoclave and indium plant through the oxidation, neutral leach, sodium jarosite precipitation, and high acid leach process steps. The impure solution, from the neutral leach, is purified through copper cementation, first stage purification with zinc dust and sodium arsenate, and second stage purification with zinc dust in a five-stage Outokumpu fluid bed reactor. The purified neutral solution is electrolyzed in a cellhouse, which will be described in greater detail below, and finally the cathode is melted and cast in a two slab line, one jumbo line, melting and casting plant.

The Kidd Cellhouse Design

The Kidd cellhouse comprises three rectifier circuits, each capable of operating with fourteen rows per circuit. Each row contains fifteen cells with forty cathodes and forty-one anodes. The total number of rows in the cellhouse is forty-four, although typical operation utilizes only forty-two rows, thus allowing two rows to be taken off-line for cleaning at all times. The rectifiers operate at 29,100 A and produce a current density of 540 A/m2. The plating cycle was increased from twenty-four to twenty-eight hours in 1994 to facilitate the implementation of automated stripping machines.

Prior to introduction of the autostrippers over the period 1995 to 1999, the cathodes were manually stripped. Cathode auto-stackers of Kidd design were introduced in 1976, and allowed considerable labour savings at that time.

TECHNICAL CHALLENGES

Although autostripping Offers clear benefits of reduced manpower and reduced repetitive motion-type injuries, the conversion of a conventional cellhouse to a fully automatic one, including a fixed base stripping machine, automatic hoists, and jumbo cells, is seldom economically viable. The approach used by Kidd was to keep as much of the existing cellhouse structure as possible, (cells and manual hoists) and to automate only the stripping and stacking functions. Kidd faced a number of challenges doing this. These are described below.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 565

Challenge #1, High Stripping Frequency

The cellhouse was designed for manual stripping, and as such, the cathodes are relatively small (1.35 m2). The small cathodes necessitate large numbers of cells (660) and large numbers of electrodes (26,400 cathodes and 27,060 anodes). The current density of 540 A/m2 is relatively high compared to state of the art jumbo cellhouses, which operate closer to 400 A/m2 (1). The high current density requires a short plating cycle of 28 hours relative to state of the art plants, which typically operate with plating cycles between 48 and 72 hours (1).

The large number of cathodes and the short plating cycle serve to increase the performance requirements of a stripping system designed for the Kidd cellhouse relative to a stripping system designed for a modern cellhouse, Table 1. In order to yield comparable productivity results, the Kidd stripping system must strip a greater number of cathodes per unit time. For instance, in the Kidd cellhouse, 51 cathodes must be stripped to yield one tonne of electrowon zinc. This is a substantially higher value than the 14 cathodes stripped per tonne of electrowon zinc in state of the art plants (1). This basic difference represents a formidable technical challenge in regards to designing a stripping system, which meets specified requirements for manpower reduction, while also meeting capital investment criteria.

Table I - Comparison of Plating Parameters at Kidd versus Modem Cellhouse Practice(l)

Current Density (A/m2) Plating Cycle (hours)

Cathode Plating Surface (m2) Cathodes Stripped per Tonne

Kidd Creek Cellulose

540 28

1.35 51

State of the Art Design

400 48 3.4 14

At Kidd, to produce 150,000 t/y of zinc, over 24,000 cathodes must be stripped each day.

Challenge #2 Space Limitations

Another technical challenge facing the project team was the space restrictions created by the cellhouse layout. The cellhouse layout is such that all forty-four rows are positioned parallel to one another. This layout necessitates a 200 meter long building containing a single aisle adjacent to all forty-four rows. All cellhouse traffic, including placement of stripping equipment, must fit within this 6 meter wide aisle while also allowing for fork truck passage and cell cleaning activity. There was simply no space for a conventional cathode conveying system to a fixed base stripping machine.

This design problem was overcome by mobile stripping machines, shown in Figure 1, which could be moved from row to row. This in itself now led to the design challenge of interfacing the mobile machines with the manual hoists.

566 LEAD-ZINC 2000

Figure 1 - The Kidd Mobile Automated Stripping Machine

CONCEPTUAL DESIGN

The design team faced a number of options before ultimately settling on a design using, 1) a cathode disk to enhance stripping knife penetration, 2) a side load rather than top load for the cathode bundles, and 3) moving the stripping head across the bundle rather than moving the cathodes to a stationary stripping head. The rational behind these choices is explained below.

Stripping Method

Three stripping concepts were considered and tested using prototype stripping rigs.

• "Top down" knife entry with no edge preparation,

• "Flex and rapping" to dislodge the zinc from the cathode plate, similar to the Kidd Process for removing plated copper from stainless steel permanent cathodes, and,

• "Cathode disk" to facilitate the stripping knife penetration.

The sentiment of the design team was to avoid using option 3, the cathode insulator disk, because of the additional development effort required to implement this technique, and the cost of conversion of all 26,400 cathodes in the cellhouse.

The results of stripping trials using full-size cathodes plated in the cellhouse are summarized in Table II. With the most important design criteria being simplicity and

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 567

reliability, the design team adopted, with some reluctance, the "cathode disk" option. This was to entail a significant development effort in order to bring the life of the disk up to the two-year cathode life expectancy.

Table II - Evaluation of Stripping Trials

Stripping Method

Number Stripped

Stripping Reliability

Issues and Comments

Top Down

Flex and Rapping

Cathode Disk

173

453

120

63% Variability of the electrolyte level created a difficult edge for knife penetration

58% The degree of flexing required to break the zinc aluminum bond often caused permanent deformation of the cathode plate

100% Prepared edge enabled reliable penetration of the knives under the zinc

Figure 2 shows the general layout of the stripping mechanism, the location of the cathode disk insert, and the travel arc of the stripping arm/knife assembly. Figure 3 illustrates the overall arrangement of the stripping machine and its associated conveyors.

STRIPPING

ARM — CARRIAGE

STRIPPING ARM —

Knife Assembly

Figure 2 - Stripping Head Assembly, shown Cathode with Disk Insert, and Travel of the Stripping Arm / Knife Assembly

Figu

re 3

- K

idd

Aut

o-st

ripp

ing

Mac

hine

Sho

win

g th

e St

ripp

ing

Mac

hine

Pro

per

on L

eft,

the

Aut

o-St

acke

r in

the

mid

dle,

and

the

Loa

d-ou

t Con

veyo

r on

the

Left

r m

> D

N 2 Π

o

o

o

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 569

Traversing versus Stationary Stripping Head

In some state-of-the-art stripping systems, the cathodes are conveyed through a stationary stripping head. The Kidd design team however chose to strip the cathodes in-situ using a mobile stripping head. The two primary reasons were:

1. The overall machine length would be substantially reduced, an important criterion on the crowded cellhouse aisle, and

2. The exchange times would be reduced, because the time waiting for the conveyer to reverse could be avoided.

Method of Loading

With the option of either top or side loading of the cathode bundle into the stripping machine, side loading was chosen for the following reasons:

• Side loading is faster and safer because the cathode rack is not hoisted beyond the operator's height, and the operator can guide the rack into position by hand,

• Side loading provides good operator access to manually strip the occasional problematic cathode,

• The operator can, at a comfortable working height, easily exchange damaged cathodes.

The key design elements of using a knife system introduced at a cathode disk, a movable stripping arm carriage, and a side load layout enabled the designers to meet the challenges of a high stripping frequency, within the tight space limitations of a manual cellhouse.

Electrical Design

Stripping Machine Automation and Control Systems

The heart of the system is a Quantum programmable logic controller driving two 20 HP hydraulic power packs used for stripping, conveying and machine mobility. The Quantum pic collects all discrete and analog inputs and outputs, and performs the closed loop PID and motion control functions. Three human-machine interfaces (HMO's) are provided for pushbuttons and graphics. One is located at each end of the stripper and one on the stacking conveyor for the operators. Modbus, Modbus Plus and Quantum's S908 I/O Bus communications are used to interconnect the various system components. A Curry Modpac device provides wireless Modbus-Plus communications between the pic and the plant Foxboro I/A distributed control system to integrate stripping machine operating data with the plant control system.

Power Conditioning

System power is provided via a cable connected to a 575 V three-phase source. A series of 12 receptacles located every 4 rows along the cellhouse allow a flexible umbilical cable connected to the machine to be limited to a 100-foot length. A 50KVA isolation transformer, surge and ground fault protection are used to condition the power and electrically isolate the

570 LEAD-ZINC 2000

machine from noise, problematic building grounding and strong dc electro-magnetic fields generated by the cellhouse 700 Volt 30,000 A bus. Carefully applied shielding and grounding technique's are the key to trouble free system operation.

Stripping

The stripper's two major functions include moving the carriage and operating the stripping arm. The carriage is driven by an electrical servo motor, drive and motion control module with optical encoder feedback to provide precision positioning of the stripping arm. The second motion control system uses a servo control valve and hydraulic cylinder to operate the stripping arm. A cylinder-mounted Temposonics magnetostrictive sensor provides feedback for precision control of knife placement. Inductive and photo proximity switches provide most of the discrete sensing functions.

Temperature control of the hydraulic fluid is used to provide consistent stripping times. Proper washing of the cathode is provided by a sequence of wash solenoids and water pressure control to spray nozzles.

Stacker Conveyor and Incline Conveyor

The cathode stacking box uses an ultrasonic level sensor. Associated logic increments the angle of the incline conveyor for proper stacking of the cathodes.

Machine Mobility

The machine is moved from row-to-row during its operation and occasionally to a repair bay by hydraulic motors and steering cylinders that provide power and direction at each corner of the stripping machine. A joystick located at each end of the stripper gives the operator 360-degree control of the machine direction.

Machine Safety Devices

Infra-red light curtains at the stripper loading station and stacking conveyor discharge points shut down the operation of the machine triggered by an operator. Pull-switches and emergency stop pushbuttons located at strategic points perform the same task manually.

Cathode Disk Design

The key design criterion is to withstand 500 stripping cycles or a minimum of 24 months of service. During development of the stripping machine, three concepts for the insulating disk were evaluated:

1. Thermoplastic injection moulded disk: The disk was moulded in a high capacity moulding process, and then mechanically fixed onto the punched-out section of the cathode.

2. Compression moulding: The disk is moulded directly onto the machined recess on the cathode sheet using a thermosetting plastic with silica fiber fillers.

3. Resin transfer moulding: The disk is moulded directly onto the machined recess on the cathode sheet using a liquid thermosetting resin with silica fiber fillers, as shown in Figure 4.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 571

2 MAX

CATHODE SIDE STRIP-

CATHODE SHEET MACHINED TO DEPTH SHOWN AND DRILLED TO ANCHOR RESIN TRANSFER MOULD COATING

(KN IFE STARTING SURFACE)

Figure 4 — Resin Transfer Moulded Cathode Disk

The thermoplastic injection moulded disk was quickly abandoned because of electrolyte penetration into the aluminum under the plastic. This caused rapid corrosion, and lead to the premature failure of the cathode disk. Both the compression moulding and transfer moulding resulted in robust cathode disks, with good resistance to electrolyte penetration. In 1999, Kidd purchased 10,000 compression moulded, and 4000 resin transfer moulded cathodes.

A number of other changes were also made to increase cathode life. The aluminum plate thickness was increased from 4.7 to 6.2 mm in 1995, then again to 7.0 mm thick in 1999 to take advantage of using scrap jumbo cathode plates from Canadian Electrolytic Zinc. The cathodes were also fitted with permanently attached polyethylene side strips. These required less manual attention than the previously used clip-on strips, thereby improving machine performance.

MACHINE PERFORMANCE

Stripping Project - Current Status

The chronology of events leading to the full conversion of the Kidd cellhouse to automated stripping is depicted in Figure 5.

Prior to automated stripping, 42 operators worked two dayshifts to strip the cellhouse. As a precursor to the stripping project, the cellhouse plating cycle was increased to 30 hours from 24 hours, thus allowing for a reduction of 10 positions in the stripping complement. Although the increase in the plating cycle was expected to reduce the current efficiency by 1 % and increase power consumption by 20 kWh/tonne, this was deemed to be attractive. In actual fact, routine maintenance in the cellhouse has had a much more significant impact on the process than the increase in plating time. Electrode maintenance and cell maintenance have remained a challenge driven mostly by manpower availability. The variations in current

I CATHODE

572 LEAD-ZINC 2000

efficiency and power consumption, depicted in Figure 6, are the direct result of cellhouse maintenance.

Figure 5 - Chronology of Automated Stripping

Figure 6 - Kidd Power Requirements and Current Efficiency

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 573

Stripping Operation

At present, the Kidd cellhouse is stripped exclusively with four mobile automated stripping machines. Manual stripping is no longer an option as the stacking equipment and the labor force are not available. The stripping complement totals 20 operators working 12 hour rotating shifts in five-person crews.

Although design and construction extended over four years, all four machines have a similar configuration with the only significant difference being the orientation of the stacking conveyor. The latter difference is necessary to allow the stripping operation to proceed at the extreme ends of the cellhouse.

Plating time is monitored through a Foxboro I/A DCS and the stripping machine operators are directed accordingly. The minimum plating time accumulated before a row is stripped is 700 kAh. The stripping equipment is moved from row to row by one operator and assisted by another to ensure safe travel through the restricted cellhouse aisle. The hoist arrangement has remained unchanged from the manual stripping era. Eight air driven hoists are in service at all times. The hoists are moved from row to row by means of a transfer bridge system.

Productivity

Figure 7 depicts the productivity gains achieved since the project's inception in 1993. It should be noted that the 1998 zinc production is the highest on record with 151,000 tonnes stripped for the year. This was achieved with 60% of the cellhouse output stripped by two automated stripping machines.

Automated stripping was fully implemented by the end of February 1999 following the conversion of the cathodes to insulating disks. At this point, the final manpower reduction was implemented and zinc production was now totally reliant on the automated stripping system.

Figure 7 - Annual Tonnes of Zinc Stripped per Stripping Employee

574 LEAD-ZINC 2000

Performance

The criteria defining acceptable performance and a successful project are as follows:

• One operator per machine

• Reduced injuries • Equipment availability of 87.5% • Stripping rate of 5.1 rows per 12 hour shift per machine • Maintain the plating cycle at 28 hours • Limit cathode damage to 300 per month per machine

Injuries related to stripping activities have been reduced as shown in Figure 8. Of the injuries still incurred, a reduction in repetitive strain injuries has been achieved.

Figure 8 - Total Stripping Related Injuries

Overall equipment on-line time has varied throughout 1999, this being the commissioning period for the final stage of the automated stripping project. Although machine availability was acceptable with units 1 and 2, a maintenance and operational learning curve was experienced. Machine availability is illustrated in Figure 9. It should be noted that zinc production in the first quarter of 2000 is on target for achieving 156,000 tonnes of cathode. The current rate, averaged over four automated strippers is over 20,000 cathodes per 24 hour period, and the evolution of the cathode stripping rate is presented in Figure 10.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 575

Figure 9 - Stripping Machine Availability

Figure 10 - Cathodes Stripped per Day

The plating cycle is being maintained at 28 hours as shown in Figure 11.

576 LEAD-ZINC 2000

1000τ

900

800-

700-

600

500

400

-Avgkarrphrsr -28HourZnc

/vvvvvvvvvvy/ Figure 11 - Average kAh of Rows Stripped

As demonstrated in Figure 12, an average of 278 cathodes per week are removed from service for all reasons, including damage from the automated stripping system.

Figure 12 - Cathodes Removed from Service on a Weekly Basis

SUMMARY

The Kidd automated stripper project has met all of the deliverables outlined in the original justification. The stripping system has proven to be reliable enough to accommodate production rates above the projections used for the original design. As a result, productivity at the Kidd electrolytic plant has improved with a positive impact on workforce occupational health and safety.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 577

ACKNOWLEDGEMENTS

The Kidd zinc stripping machine was successfully developed because of the close co-operation between plant operations and maintenance personnel, and the design and process engineering groups. Those deserving particular recognition for the machine development are: V. Robinson and D. Deluca, mechanical and electrical/instrumentation designers, respectively, G. Girard, maintenance technologist, and M. Boissoneault, process engineer.

REFERENCES

1. F. Tamargo and Y. Lefevre, "Concept and Operation of the New Conventional Cellhouse of Austuriana de Zinc at San Juan de Nieva, Spain", World Zinc '93. I.G. Matthew, Ed., Australasian Institute of Mining and Metallurgy, Parkville, Australia, 1993, 295-306.

2. World Zinc '93. I.G. Matthew, Ed., Australasian Institute of Mining and Metallurgy, Parkville, Australia, 1993, 612 pp.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 579

CELLHOUSE VENTILATION

J.A. Davis and J. de Visser DESOM Environmental Systems Limited

1211 Gorham Street Newmarket, Ontario, Canada L3Y 7 VI

ABSTRACT

Since the mid 1970's, development of zinc electrowinning (EW) facilities has out-paced its counterparts in the copper and nickel electrowinning industries in the design of efficient, compact, and automated production plants. Recent advances in process automation have provided opportunities to increase current efficiency, productivity and cathode quality, while reducing the footprint of the plant and the number of operating personnel working on top of the cells. Usually, the practice has been to allow a totally contaminated building structure, combining the plant ventilation system with the process cooling air. As a result, facilities employing this practice are finding difficult to meet their process design criteria and the stringent labor and environmental requirements. This paper provides an overview of the problems and disadvantages associated with these older process designs. It also describes new developments in ventilation systems, particularly building design developments and cost savings that are possible when improved ventilation systems are introduced into efficient process operations.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

580 LEAD-ZINC 2000

INTRODUCTION

In the development of EW facilities, the footprint of the facility provides a good indication of the operating efficiency. In general, the smaller the footprint for a given tonnage, the higher the possibility of having a plant operating at a higher efficiency. However, from a workplace health and safety perspective, higher process efficiencies do not always translate into better workplace environments. Higher productivity in ever more compact spaces can lead to more costly workplace environmental control systems. To place plant design developments into perspective, an historical review of plant designs is provided in the following section.

HISTORICAL REVIEW

Early 1900's

Process Design

Commercially viable zinc EW operations were developed in the early 1900's. These relied on manual retrieving and stripping of cathodes, using cascading electrolyte flows through successively elevated cells, primitive electrode contact arrangements, separation of cells into rows of cell pairs for electrical distribution, and cooling coils in the cells. Production facilities were large for the tonnages produced, yet, some of these plants operated until the mid 1970's.

Working Conditions

The sulfuric acid-based EW process emits gases, which, upon bursting at the surface of the electrolyte, eject acidic aerosol droplets into the working environment. In the early facilities, workers retrieved and stripped cathodes directly on top of the cells. Surfactants prepared from licorice and Arabic gum were used to reduce the surface tension of the electrolyte and to create a foam on the cell surface. Although the acidic aerosol emissions from the cells were reduced with the addition of surfactants, the foam tended to trap hydrogen, and the pyrotechnics that erupted when retrieving or replacing electrodes were truly impressive, making hearing and other protection gear mandatory.

Ventilation of these buildings, like their copper refinery cousins, was by gravity-induced flows. Roof openings were built into the design of the plant and wall openings around the periphery of the cell area brought some relief to the workforce.

Mid 1900's

Process Design

Early operations were replaced/rebuilt in North America during the 1970's. The more efficient designs provided higher productivity with improved electrolyte distribution, electrical contacts, electrolyte cooling and electrode handling systems. The resulting higher current densities, current efficiencies and productivity reduced the cost of producing zinc. These plants, however, were only an extension of earlier designs, and stripping was still a mechanically assisted manual operation.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 581

A more radical approach was taken to the design of EW plants in Europe. Large jumbo cathodes were developed. These required larger cells, massive cranes to retrieve the cathodes, and automated stripping machines to remove the zinc from the much larger cathode plating area. With these changes, the plant footprint was reduced and overall cost reductions were achieved because of the improved productivity associated with the automation of the material handling operations.

Working Conditions

The use of new surfactants reduced acid aerosol emissions. Alternative methods of ventilation were attempted to improve the in-plant working conditions. Gravity systems drawing fresh air from below the cells, gravity systems drawing fresh air from above the cells, mechanical systems supplying air from below the cells and mechanical exhaust systems removing air from the roof were all attempted in a number of combinations in different plants.

CONTEMPORARY APPROACHES TO CELLHOUSE VENTILATION

To provide a basis for discussion of the merits of the various designs, this paper will briefly describe and discuss the merits of three recent cellhouse ventilation systems in terms of ventilation environments, conditions within the cellhouse, building structure, automation required, and emission capture systems.

When dealing with cellhouse ventilation and its environmental consequences, the occupational exposure of personnel and the emissions from the cellhouse to the environment must be considered. It is prudent at this point to define the exposure and emission limits that regulate EW operations, with specific reference to sulfuric acid emissions.

Occupational Exposure to Chemical Substances

The regulations concerned with exposure limits to various substances vary around the world. However the Occupational Health and Safety Authority (OSHA) and the American Conference of Government Industrial Hygienists (ACGIH) set the most commonly applied regulations and guidelines used. With reference to sulphuric acid mist, the current OSHA/ACGIH requirements are:

TLV-TWA =1.0 mg/m3 (time weighted average) (1)

TLV-STEL = 3.0 mg/m3 (short term exposure limit) (2)

Emissions Limits

Again, these vary considerably from country to country, and in countries such as the United States, the limits vary from state to state. For sulfuric acid, the US Environmental Protection Agency (EPA) has set an emission limit for any given production facility at 10 tons per annum. However, in California and Arizona, this limit has been set at 7 tons per annum. In general, emissions from a facility are defined in two ways: namely total emissions per annum; and secondly, a border or fence-line concentration limits. Examples offence-line limits for sulfuric acid mist are:

582 LEAD-ZINC 2000

Ontario, Canada 100 μ§/πι3 (3)

California, USA 120μg/m3 (4)

Physical Modeling of Air Ventilation Systems at DESOM

In the early stages of each project, physical hydraulic modeling is conducted to establish the flow patterns and the expected acid concentrations within each building. For physical modeling, 3D- models of each building are created (e.g., see Figure 1). In the models, fresh water is used to simulate clean fresh air, and salt water is used to represent acid emissions from the EW cells. This combination allows accurate simulation of the conditions within each building, including those created by buoyancy-driven flows over the cell areas.

Figure 1 - Physical Hydraulic Modeling

Cooling Tower Ventilation System

Systems based on cooling towers rely on the cooling tower fans to provide ventilation for the building. Typically, cooling towers are located on the roof of the building with the point for the air exhaust located close to the center of the roof structure (Figure 2). Fresh air is provided through the basement of the building. From the basement, the air comes up through the walkways in and around the cell area, and then flows towards the central exhaust point. The acid mist is removed upwards through the roof structure of the building.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 583

Figure 2 - Location of the Cooling Towers

Center Aisle Ventilation System

In a center aisle system, the cooling towers are not used for ventilation. Instead, fans located at the center of the building are employed. Air is drawn from the sides of the building, and flows across the cells towards the central exhaust point (Figure 3). This system improves conditions over the cells, by providing a horizontal exhaust flow, which reduces acid contamination above the cells. This horizontal flow pattern provides an acceptably low level acid concentration at the worker-breathing zone (typically defined as 1.5 m above the cathodes).

Figure 3 - Ventilation Flow in a Center Aisle Ventilation System

The Cross Flow System

A standard cross flow ventilation system develops further the horizontal flow concept and provides ventilation to the tankhouse by generating a horizontal flow across the cell area of the tankhouse (see Figure 4 showing typical cross flow layout with flow patterns). The system has an inlet or baffle side and a fan or exhaust side. The fans located at the exhaust side of the tankhouse generate the air flow. The acid mist generated by the cells is removed by means of the cross flow of air.

584 LEAD-ZINC 2000

Figure 4 - Typical Cross-flow Layout with Flow Patterns

Typical Flow Patterns and Expected Exposure Levels within the Working Space (Modeling Results)

Results from modeling, show fundamental differences in the ventilation techniques used. In the cross flow system, the acid generated by the cells is transported in cross flow across the cells. The momentum of the cross flow system must be balanced with respect to the buoyancy forces generated by the cells. The predicted acid concentration profile shows an increase in concentration levels at increased distances from the fresh air inlet across the tankhouse, with the highest concentration being near the exhaust fans. Typical flow and acid concentrations are shown in Figure 5. With this system, the worker breathing-zone (normally defined as 1.5 m above the cell level) is kept acid-free for most of the cell area, with limited concentrations only being reached toward the fan side of the building. Under normal operating conditions, re-circulation of acid mist into the roof structure is kept to a minimum, with concentrations well below acceptable limits. Figure 5 illustrates a typical arrangement.

Figure 5 - Typical Flow and Acid Concentrations

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 585

Results for the center aisle system are similar to those of the cross flow system. The acid concentration profile is similar to that created by the cross flow system, except for the fact that the exhaust point is now at the center of the tankhouse, resulting in mirror image profiles on each side of the exhaust units.

For the cooling tower ventilation, the acid mist removal is upwards through exhaust points in the roof of the building. In this case, the flow is in the same direction as the buoyancy-driven flow generated by the cells. At first, this would appear to be of benefit; however, in this case, if the buoyancy generated flows are allowed to dominate, heavy re-circulation zones will form within the building. This is a result of having central exhaust points on the roof of the building. When these flow patterns are viewed down the long section of the building, as shown in Figure 6, it can be seen that, if the buoyancy forces are allowed to dominate, the roof structure becomes contaminated with acid mist. However, it should be noted that in this system, the primary objective is to keep the walkway areas clean and free from contamination. As can be seen in Figure 7, this is achieved by the flow of clean air from the basement. Over the cell area, the air is totally contaminated by acid mist, requiring personnel to wear protective gear when working over the cells.

Figure 6 - Recirculation Zones Formed within the Building, as a Result Of the Central Exhaust Points on the Roof of the Building

Figure 7 - The Walkway Areas are Clean and Free from Contamination Over the Cell Area, However, the Air is Totally Contaminated by Acid Mist

586 LEAD-ZINC 2000

Design Requirements Based on the Selected Ventilation System

As can be seen, the three systems described work in fundamentally different ways, and therefore have completely different flow and acid contamination patterns. These differences in the fundamental flow patterns of each system result in differing engineering requirements in terms of the design and construction of each tankhouse. The first of these differences is in the building structure. With the cross flow system, the building can be constructed using a conventional steel type structure (Figure 8), with only the base being formed by a concrete basin. With the cooling tower ventilation design, the entire building has to be constructed from acid proof concrete. The primary reason for this difference is in the acid concentrations in the upper levels of the structure. In the cross flow design, the acid concentrations are kept to a minimum height, whereas in the cooling tower system, the acid levels are distributed throughout the structure.

In the systems described, the center aisle of the cellhouse also had a concrete structure because the cooling towers were still installed on the roof even though they were no longer providing the primary means of ventilation. In cold weather climates there are additional problems associated with the cooling tower exhaust system. If the roof is not properly insulated, condensation will form in the roof structure, and this will result in acid rain forming in the tankhouse. Although this could happen in all systems, it is more likely to be a problem with cooling tower ventilation.

Figure 8 - With the Cross Flow System, the Building can be Constructed Using a Conventional Steel Type Structure

In all cases, use of modern, highly automated cranes for cathode and anode removal can be expected. In the cooling tower ventilation system, the crane would have to be made more corrosion resistant, as it would be moving through a more contaminated environment compared to that obtained in the cross flow ventilation system. This applies to all automated systems.

The next major difference in design, is the number of fans required to provide ventilation. In the case of the cooling tower system, the cooling tower fans provide cooling and ventilation.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 587

Although this reduces the number of fans required, it does place higher demands on these fans. This extra demand creates another problem. Cooling towers for zinc electrolytes rely on evaporative cooling, which is controlled by inlet conditions to the cooling tower. For a cellhouse using cooling tower ventilation, the inlet conditions to the cooling tower depend on the conditions in the tankhouse, which in turn changes the cooling capacity of the air entering the cooling tower. In other words, the increase in wet bulb temperature and relative humidity of the air with the addition of acid mist reduces the capacity of the cooling tower. For either the center aisle or the cross flow system, more fans are required; however, the use of ventilation fans reduces the cooling load on the cooling towers.

Control of Emissions

As regulations around the world become more stringent, the control of emissions from the cellhouse becomes more important. The key is to provide a simple, cost-effective solution to the problem. In the case of the cooling tower ventilation system, all the ventilation air passes through the cooling tower. In this case, the mist eliminator units in the tower must remove acid droplets contained in the air stream. The problem with this method is droplet size. Typically, the mist eliminator units used on cooling towers can effectively remove droplets 60 μπι or greater. However, the droplets generated within a cellhouse tend to be very fine and of the order of 10 μιη. This results in low mist removal efficiencies, and consequently, high emissions from the cooling tower. With the exhaust fan units used in the center aisle, an alternative cross flow system is required. DESOM has developed an effective, yet inexpensive system, which provides a high removal efficiency. The DESOM MistElim system (as shown for a wall fan in Figure 9) can meet most of the current requirements for emissions in both Canada and the U.S.A.

Figure 9 - The DESOM MistElim System

588 LEAD-ZINC 2000

CONCLUSIONS

With advances in process automation and the drive to reduce the size of EW plants, many techniques have been developed to deal with ventilation problems linked to the generation of acid mist. This paper reviewed three different approaches to address this problem. Each of these options has advantages and disadvantages. Some of the differences include:

• Building structure: concrete or steel • Cooling tower loads and performance • Number of fans and overall building air quality • Capture of emissions.

All of these factors need to be included in the design and construction of modern tankhouses to insure that the facility will not only be cost-effective, but also will be user and environmentally friendly.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 589

NEW WROUGHT Pb-Ag-Ca ANODES FOR ZINC ELECTROWINNING TO PRODUCE A PROTECTIVE OXIDE COATING RAPIDLY

R. D. Prengaman and A. Siegmund RSR Technologies, Inc.

2777 Stemmons Freeway, Suite 1800 Dallas, Texas, U.S.A. 75207

ABSTRACT

Lead anodes used in zinc electrowinning generally contain 0.5 - 1.0 % silver to reduce the rate of corrosion over the service life. For good quality zinc, the cathodes must contain less than 10 ppm lead. In order to reduce the lead contamination of the cathode, the lead anode must become coated with a protective layer of Pb02/Mn02. The formation of this layer may take as long as 30 - 60 days, during which time zinc production is substantially reduced and the cathodes are contaminated with lead. To improve the mechanical properties of Pb-Ag-anodes alloying elements such as calcium, strontium, barium and others are added. The production of cast lead-silver or lead-silver-calcium anodes often results in the formation of numerous holes, voids or laps in the anode surface that can initiate localized internal corrosion. Based on the experience as the major supplier of rolled anodes for copper electrowinning, RSR Technologies has developed an improved anode for zinc electrowinning. The anode consists of a rolled Pb-Ag-Ca-alloy with a controlled surface grain structure. It forms an adherent protective oxide coating within 2 - 3 days when placed in an electrowinning cell. The preferred calcium and silver contents of the alloy are 0.03 - 0.08 % and 0.3 - 0.4%, respectively. Rolling sheets significantly reduces the presence of internal porosity or laps. The lack of defects results in long anode life. The rolling also produces a uniform grain structure to which the corrosion product readily adheres, thus significantly reducing the conditioning period.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

590 LEAD-ZINC 2000

INTRODUCTION

Lead contamination of the cathodes in zinc electrowinning is minimized by the formation of a stable protective coating of Pb02 on the surface of the anode. The formation of the stable PbC>2 layer is a slow process on the normally utilized lead-silver anodes. Lead-calcium-silver alloys, investigated and applied in the past, also showed a very slow rate of formation of the protective Pb02 layer.

In the conditioning of the anodes to form a PbU2 layer, oxygen evolution on the anode surface causes small fine particles of PbC>2, which are not adherent, to be detached from the surface of the anode and become suspended in the electrolyte from where they contaminate the cathode with lead. The PbC>2 particles are reduced to metallic lead at the cathode leading not only to lead contamination, but also to a detrimental morphology of the zinc deposits which can lead to dendrite formation and short circuits. The lead contamination, poor morphology, and poor current efficiency are common in cells containing new anodes.

In addition to oxidation of the lead surface to PbC>2, a competitive oxidation process also takes place at the anode. New anodes rapidly oxidize MnCh from the manganese-containing electrolyte. The MnC>2 not only builds up rapidly on the anode surface but the manganese in solution can also react with the newly formed PbC>2 particles to produce even higher amounts of Mn02, further reducing the rate of formation of an adherent PbC>2 film. The result of the Mn02

deposition is the formation of large, soft, Mn02-PbC>2 layers which are not well attached to the anode. Oxygen evolution at the anode surface dislodges these layers into a fine suspension of Pb02/Mn02 particles. These particles can be conveyed in the electrolyte to the cathode where they are reduced to manganese ions whereupon they reform manganese sulfate. The reduction of large amounts of finely divided Mn02 at the cathode dramatically reduces the zinc current efficiency via Equation 1.

Mn02 + Zn + 2H2S04 -> ZnS04 + MnS04 + 2H20 (l)

Because the Μηθ2 sludge also carries Pb02 particles, contamination of the cathode with lead results from the spalling of the Μηθ2 layer (Equation 2). Only some of the MnÜ2/Pb02 layer contaminates the cathode. A large amount of sludge builds up in the cells and requires frequent cleaning.

Pb02 + 2Zn + 2H2S04 -> Pb + 2ZnS04 + 2H20 (2)

Eventually, as the anodic process continues, a hard, dense, protective layer of Pb02 is formed on the anode surface. Once this protective film has been formed, cathode contamination decreases and the amount of sludge generated by the anode decreases as well. This process (called conditioning) may take 30-60 days or more depending on the anode composition and current density (1). Because of the difficulty in conditioning anodes, operators of zinc cellhouses are very reluctant to replace an entire cell of used, conditioned anodes with new, unconditioned anodes. Operators will normally replace only one or two anodes per cell or try to condition the anodes prior to use in the cells.

Cominco in the late 1960's developed a method to precondition the anodes in separate cells outside the circuit (2). The process involves the creation of a hard dense Pb02 layer on the surface of the anode by oxidation of the anode at high current density in an acidic fluoride-containing electrolyte. This preconditioning process takes 8-12 hours. In this process the

Edited by I.E. Dutrizac, I.A. Gonzalez, D.M. Henke, S.E. lames and A.H.-I. Siegmund 591

PbC>2 layer is created without the interference of the rapidly forming MnC>2 layer. The anode corrodes to form a very hard resistant surface.

In the early 1990's, an alternative process to anode preconditioning was developed by Mintek (3). This process involves sandblasting the surface of the cast anode with a relatively coarse silica sand at high pressures above 500 kPa. This process roughens the surface of the anode to enhance the formation of a hard adherent PbC>2 layer. Sandblasting has been shown to reduce the time to form an initial corrosion layer on the lead anode surface. It also increases the adherence of the newly formed Pb02/MnÜ2 layers giving a conditioned anode within several weeks instead of several months, but raises environmental concerns for the manufacturer of the anodes.

As anodes have become larger, it has become much more difficult to cast the anodes to the preferred shape. Rolled lead-silver or rolled lead-calcium-silver sheets have been utilized alternatively in some tankhouses to form these jumbo anodes. The surfaces of the rolled anodes are much smoother than those of the cast anodes generating fewer places for the oxides to adhere. In addition, the texture is oriented in the rolling direction, producing a grain structure even more difficult to corrode and to form an adherent layer of PbC>2

RSR Technologies has developed a method to produce a surface on the anode, through controlled rolling and alloying, to which a hard, dense layer of PbC>2 adheres very quickly. In the rolling process a very fine, uniform non-elongated grain structure is developed which presents many fine grain boundaries. These grain boundaries are corroded by the anodic current causing PbC>2 to be produced at many uniformly dispersed sites. The large number of anodic sites cause the PbÜ2 to adhere to the anode surface at multiple locations, and permit the anode to be conditioned in the cell within several days, without the need for preconditioning or sandblasting.

CURRENT STATUS AND DEVELOPMENTS

Pb-Ag Anodes

Lead-silver anodes have been used for many years for zinc electrowinning. The anodes are produced by casting a lead - 0.5-1.0% silver alloy into a mold containing a copper bus bar and having the anode shape. The anodes have been produced in flat, ribbed, or perforated designs for the body of the anode.

In general the cast anodes have an oriented, dendritic grain structure. Such a structure is shown in Figure 1. The silver in the alloy is segregated as a silver rich phase at the interdendritic boundaries as well as at the grain boundaries. During the initial conditioning or corrosion of the cast lead-silver anodes, these areas where attachment of the PbC>2 corrosion product would normally occur most rapidly (the grain and interdendritic boundaries) are rendered most corrosion-resistant by the concentration of silver. Only the small, lead-rich region between the interdendritic boundaries can be corroded. There are, however, no boundaries for the PbC>2 corrosion product to attach itself to the anode surface. It is only when the surface becomes significantly roughened because of the spalling of the non-attached PbÜ2 particles and newly formed MnC>2 particles that the PbC>2 and MnC>2 particles finally form an adherent layer. The higher the silver content of the anode, the longer the process takes to fully form an adherent layer. The conditioning generally takes 30-90 days.

592 LEAD-ZINC 2000

Figure 1 - Structure of a Cast Lead - 0.75% Ag Anode (magnification 160x)

Cast lead-silver alloys are extremely corrosion-resistant, but in general have a life of only 2-4 years. The short life is in general related to the presence of casting defects within the body of the anode. The internal corrosion of these defects, as seen in Figure 2, can lead to the production of holes and uneven corrosion conditions in certain areas of the anode leading to bending, warping and short circuits.

Figure 2 - Defects in a Cast Pb - 0.75% Ag Anode (magnification 320x)

Rolled lead-silver alloys have been used to overcome the problems caused by porosity and casting defects. In this process a billet of lead-silver alloy is cast, rolled to the desired thickness, cut to shape, and joined to the copper bus bar.

Rolling the alloy also breaks up the original cast grain structure, produces a smooth surface with less corrosion, and facilitates the removal of the MnC>2 scale (4). One would think that rolling would produce a grain structure more compatible with rapid conditioning of the anode. The grain structure of a typical rolled lead-silver alloy is seen in Figure 3.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 593

Figure 3 - A Rolled Pb-Ag Anode (magnification 160x)

The structure consists of grains which are oriented in the rolling direction. The extension of the grains during rolling, however, merely bends the grains to the rolling direction and extends them while retaining the original cast-in segregation. The anode structure at the surface still has silver segregation at the grain and sub-boundaries, but the lead areas have been extended and elongated. Very few new grain boundaries have been created and the rate of attachment of the newly formed PbC>2 may not be significantly faster than in cast lead-silver alloys.

Lead-Calcium-Silver Anodes

Lead-silver alloys are extremely weak and bend easily in response to being struck by cathode blanks as they are removed or returned to the cells, during MnC>2 removal, or in response to variations in corrosion conditions caused by cast-in defects. Calcium has been added to lead-silver alloys to increase the mechanical properties. The calcium content added is normally 0.05-0.07% calcium; however, some researchers report the addition of significantly higher amounts of calcium (5,6,7). Increasing the calcium content to about 0.10% Ca results in higher mechanical properties such as tensile and yield strength and hardness. However, as the calcium content is increased above 0.10% Ca, the mechanical properties decrease because of the segregation of calcium within the anode as primary PbaCa.

Unfortunately, cast lead-calcium-silver alloys have the same problems of conditioning as lead-silver alloys. The calcium is primarily segregated upon casting to the lead-rich dendrites, while the silver is segregated primarily to the interdendritic region, as seen in Figure 4. Although the calcium slightly increases the rate of corrosion and the formation of PbC>2 in these anodes, the lack of attachment areas makes conditioning as difficult as for lead-silver anodes.

594 LEAD-ZINC 2000

Figure 4 - Grain Structure of a Cast Pb-Ca-Ag Anode

Sandblasted Lead Anodes

In an attempt to make lead-calcium-silver anodes more receptive to the formation of an adherent Pb02/Mn02 layer, lead-silver and lead-calcium-silver anodes have been sandblasted. This technique was developed by Mintek and is employed by Castle Lead (3) and some other manufacturers. The sandblasting disrupts the cast Pb-Ag or Pb-Ca-Ag structure and creates some new grain boundary surfaces, but roughens the structure for better adhesion of the PbCVMnC^ film, as seen in Figure 5. Conditioning times have been reported to be significantly reduced using this technique.

Figure 5 - Surface Grain Structure of a Cast Pb-Ca-Ag Sandblasted Anode (magnification 160x)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 595

Rolled Pb-Ca-Ag Anodes

RSR Technologies has developed a special rolling technique to produce a unique grain structure in wrought lead-calcium-silver alloys. The process produces a fine uniform grain structure both at the surface and throughout the rolled sheet. The fine grain structure has many new grain boundaries which serve as areas of attachment for the newly formed PbC>2. The structure is seen in Figure 6. The new grain boundaries are free of silver particles and can be attacked by the oxygen generated in the electrowinning process to rapidly produce an adherent PbC>2 layer. In tests, the anodes have produced an adherent layer of Pb02/Mn02 in as little as two days. The lead level of the zinc cathodes was low and a high current efficiency was obtained basically from the first day of operation.

Figure 6 - Grain Structure of a Rolled Pb-Ca-Ag Anode (magnification 160x)

In tests of up to one year duration, the corrosion rate of the anode was very low because of the improved grain structure, showed reduced cell voltage, and maintained its rigidity and high mechanical properties. Although the average life of these anodes for zinc electrowinning is yet unknown, rolled lead-calcium-tin anodes for copper electrowinning have given up to ten years service in certain tankhouses (8).

In addition, assembling the RSR rolled anode sheets to the copper hanger bar is carried out by soldering the sheet to the slotted bar. This forms an intermetallic compound resulting in a complete metallurgical bond between the sheet and the bar (8). Subsequently either casting or electroplating a layer of lead around the bar protects the bond as well as the bar. The bond is maintained for the life of the anode and always provides a high conductivity at the interface. In contrast, conventionally cast around systems generally increases in resistance during the anode life. Figure 7 shows the level resistance of RSR anodes after eight years of operation in copper electrowinning that results in significant power cost savings compared to competitive anodes showing an evident increase in resistance.

596 LEAD-ZINC 2000

Figure 7 - Resistance Between the Copper Bar and Anode Sheet as a Function of Service Life

CONCLUSIONS

Conditioning of lead anodes for zinc electrowinning has been a major problem. Rolled lead-calcium-silver alloys with controlled grain structure have been developed by RSR Technologies to enable the anodes to produce an adherent PbC"2/Mn02 deposit after a few days of operation. The reduced conditioning time produces less lead contamination of the zinc cathode and higher current efficiencies compared to conventional lead-silver or lead-calcium-silver anodes.

REFERENCES

1. E. R. Cole, Jr. and T. J. O'Keefe, "Insoluble Anodes for Electrowinning Zinc and Other Metals", U.S. Bureau of Mines Report RI-8531. U. S. Dept. of the Interior, Washington, D.C., 1981.

2. R. H. Farmer, "Anode Pre-Conditioning and Other Changes in Cominco's Electrolytic Zinc Operations", Proceedings of Extractive Metallurgy Division Symposium on Electrometallurgy, Cleveland, Ohio, American Institute of Mining, Metallurgical, and Petroleum Engineers, New York, U.S.A., 1968,242-250.

3. J. M. S. Rodrigues and E. H. O. Meyer, "Improving the Performance of Anodes for Zinc Electrowinning", EPD Congress 1996, W. Warren, Ed., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1995,161-180.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 597

4. Y. Lefevre and M. R. Pieterse, "The State of the Art and Constant Feasibility Upgrading of the Zinc Electrowinning Cellhouse of Asturiana de Zinc at San Juan de Nieva", Zinc and Lead Processing, J. E. Dutrizac, J. A. Gonzales, G. L. Bolton and P. Hancock, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 367-377.

5. H. Nozaka, K. Tozawa, and Y. Umetsu, "Improvement of Behavior of Lead Alloys as Insoluble Anode for Electrowinning of Zinc in Acidic Sulfate Solutions", Zinc and Lead '95. T. Azakami, N. Masuko, J. E. Dutrizac, and E. Ozberk, Eds., The Mining & Materials Processing Institute of Japan, Tokyo, Japan,1995, 384-393.

6. K. Koike, H. Watanabe, N. Masuko, "Energy Savings in Zinc Electrowinning with Modification of Lead Anode", Zinc and Lead '95. ibid, 373-383.

7. Y. Takasaki, K. Koibe, N. Masuko, "Mechanical Properties and Electrolytic Behavior of Pb-Ag-Ca Ternary Electrodes for Zinc Electrowinning", Lead-Zinc 2000. J.E. Dutrizac, J.A. Gonzalez, D. Henke, S.E. James, and A. Siegmund, Eds., The Minerals, Metals, and Materials Society, Warrendale, PA, U.S.A., 2000

8. R. D. Prengaman and A. Siegmund, "Improved Copper Electrowinning Operations Using Wrought Pb-Ca-Sn Anodes", Copper 99, Vol. III. J. E. Dutrizac, J. Ji and V. Ramachandran, Eds., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1999, 561-573.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 599

MECHANICAL PROPERTIES AND ELECTROLYTIC BEHAVIOR OF Pb-Ag-Ca TERNARY ELECTRODES FOR ZINC ELECTROWINNING

Y. Takasaki and K. Koike Faculty of Engineering and Resource Science

Akita University, Akita, 010-8502 Japan

N. Masuko Chiba Institute of Technology

Chiba, 275-0016 Japan

ABSTRACT

At the Zinc&Lead'95 symposium, one of the authors presented several modifications of Pb-Ag-Ca alloy anodes for the zinc electrowinning process, and also, the behavior of some organic additives in the electrolyte. In this study, the mechanical properties of the electrode materials composed by Pb-0.5 wt% Ag-(0~0.8) wt% Ca were studied by the tensile and Vickers hardness methods. Furthermore, experimental research on the influence of rolling and annealing treatments on the electrolytic behavior and mechanical properties of these alloys was carried out. The tensile strengths of the annealed Pb-0.5 wt% Ag-(0.1-0.8) wt% Ca alloys were the same or higher than those of the as-cast Pb-0.9 wt% Ag alloy. The anode potential using the as-cast Pb-0.5 wt% Ag-0.6 wt% Ca alloy decreased compared to the as-cast Pb-0.9 wt% Ag alloy. Moreover, the annealed Pb-0.5 wt% Ag - 0.6 wt% Ca anode after rolling showed a lower anode potential than the as-cast Pb-0.5 wt% Ag - 0.6 wt% Ca anode.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

600 LEAD-ZINC 2000

INTRODUCTION

The energy consumption of the zinc electrowinning process represents more than half of the entire electric power used in the extraction of zinc by hydrometallurgical methods. The equipment used in the electrowinning process consists mainly of anodes, cathodes and the electrolytic bath. Their influence on energy savings and on the effective electrodeposition of high purity metals is very important. In many electrowinning plants, efforts are directed toward decreasing the ohmic loss, increasing the current efficiency and improving the quality of electrodeposited metals. The Pb-(0.8~1.0 wt%)Ag anodes used in the conventional zinc electrowinning process have high formability, high resistance to corrosion and also the capability of lowering the content of lead in the electrodeposited zinc. The major disadvantage of this type of anode is the high overvoltage for oxygen on the anode surface. For this reason, one important objective of the experimental research on zinc hydrometallurgical extraction is to minimize the energy consumption of the electrowinning process.

At the Zinc&Lead'95 Symposium, Koike et al. (1) reported that the anode potential decreases when Ca was added to the conventional Pb-Ag alloy electrode material and also when alcohol was added as a depolarizing substance in the electrolyte.

In this study, the influence of rolling and annealing of the Pb-Ag-Ca alloy material on the electrolysis behavior and mechanical properties, considering power savings and the high efficiency of the zinc electrowinning, are reported. Furthermore, the behavior of these anodes in an electrolyte containing ethylene glycol as a depolarizing substance was studied.

EXPERIMENTAL METHOD

The alloys used as samples in the experiments were prepared by melting a mixture of electrolytic lead (99.9 wt%) electrolytic silver (99.9 wt%) and Pb-Ca master alloy in quantities corresponding to the target alloy composition. The mixture was melted in a graphite crucible at a temperature of 700 to 800°C. The alloys used as anodes were poured as bars with the dimensions of 10x10x100 mm, and the alloys used for tensile properties and Vickers hardness tests were poured as plates (100x140x3-7.5mm). The annealed samples were heated at 200°C using an electric furnace. The rolled samples were rolled at room temperature at rates of 20, 40 and 60%. The chemical composition of the samples was determined by ICP analysis. The anode with a surface area of lxl cm2 was fixed in an epoxy resin. The anode surface after resin solidification was polished with sandpaper (No.400 then No. 1000). Further, the anode surface was etched using an acetic acid-hydrogen peroxide-water solution; then it was ultrasonically cleaned in distilled water and finally washed with distilled water.

Before the anode potential measurements, the anodes were initially conditioned by electrolysis (chemical conversion treatment) applying a 160 g/1 H2SO4 electrolyte at 40° C and a current density of 50 mA/cm2 for 24 h in order to prepare a PbC>2 oxide layer at the anode surface. The distance between the anode and cathode in the chemical conversion treatment was 3 cm. After the chemical conversion treatment, the anode potential was measured in a 160 g/1 H2SO4 and 60 g/1 Zn electrolyte at a current density of 50 mA/cm2 for 1.0 h. The experiment temperature was 40° C and the anode-cathode distance was 3 cm. Another experiment with the same characteristics as above mentioned, but without Zn in the electrolyte and with no preliminary electrolysis, was performed for 20 days electrolysis time, the electrolyte being replaced every two days. The aluminum cathode used in the experiments had a surface area of

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 601

lxl cm2 and the reference electrode was Ag-AgCl (KC1 saturation) with a standard potential of +0.212 V (vs. NHE at 40° C).

The samples for the tensile tests were prepared according to the specifications of the Japanese Industrial Standard (JIS), No.l4B. The load-elongation curve was obtained for a crosshead speed of 1.0 mm/min. The hardness of the specimens was measured using the Vickers hardness method at 1 kgf for 15 s.

EXPERIMENTAL RESULTS

Electrolytic Behavior of Various Anode Materials

Electrolytic Properties of the As-cast Anode Materials

In this study, using anodes made from the as-cast material, the anode potential was measured over a current density range between 100 mA/cm2 and 5 mA/cm2. The current density during the experiment was gradually decreased from 100 mA/cm2 to 5 mA/cm2 using a decreasing step of 5 mA/cm2 at an interval of 5 min. The effect of the current density on the anode potential for several as-cast anodes is shown in Figure 1. The anode potential of the Pb-0.5 wt% Ag - Ca anodes containing equal to or more than 0.5 wt% Ca is lower than that of the Pb-0.99 wt% Ag. The lowest values of the anode potential in this study were obtained for the Pb-0.48% Ag - 0.88% Ca anodes and at a current density of 50 mA/cm2. The anode potential of this electrode was about 80 mV lower compared to that of the Pb-0.99 wt% Ag anode.

o «1.7 1 L <J 1.6 -

1.5

Pb-0.99%Ag

- ° - Pb-0.51%Ag-0.25%Ca

-+- Pb-0.54%Ag-0.50%Ca

- ° - Pb-0.48%Ag-0.88%Ca J I I I I I I L

20 40 60 80 100 Current density (mA/cm )

Figure 1 - Effect of the Current Density on the Anode Potential for Several As-cast Electrodes (Zn : 60 g/1, H2S04 : 160 g/1, 40°C)

The influence of the calcium content on the anode potential is shown in Figure 2. It can be seen that the increase in the calcium content of the anode leads to a decrease in the value of

602 LEAD-ZINC 2000

the anode potential. For the anodes containing more than 0.6 wt% Ca, any increase in the content of Ag shows no significant change in the values of the anode potential.

2.1

^2.0 >

a

a.

1

1.8

,ZP A

Δ *&**Η< X Pb-0.99%Ag Δ Pb-0.6%Ag-Ca

o Pb-0.4%Ag-Ca ■ Pb-0.7%Ag-Ca

A Pb-0.5%Ag-Ca , i I i I i I _L

0.2 0.4 0.6 0.8 Ca content (%)

1.2

Figure 2 - Effect of the Ca Content on the Anode Potential (Zn : 60 g/1, H2S04 : 160 g/l, 40°C, 50 mA/cm2)

Figure 3 shows the relation between the composition of the Pb-Ag-Ca anode and the anode potential at a current density of 50 mA/cm2. From this figure, it can be observed that any increase in the content of Ca and Ag tends to lower the value of the anode potential.

100 0.2 0.4 0.6 0.8 1.0

Agwt%-+

0.4 Anode potential

(V, vs NHE) • -1.99 ■ 1.99—2.02 ▲ 2.02-2.05 ♦ 2.05-

Figure 3 - Relation Between the Composition of Ternary Anodes and the Anode Potential (Zn : 60 g/l, H2S04 : 160 g/l, 40°C, 50 mA/cm2)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 603

Electrolytic Properties of Rolled and Annealed Anodes

The lead alloys including calcium undergo strain as a result of aging, and on one occasion, the anode was deformed after 20 days of electrolysis time and its hardness increased. For this reason, the strain in the anodes used in the experimental research was removed by an annealing operation. Also, the influence of rolling of the anode material on the electrolysis was investigated.

Figure 4 shows the relation between the current density and the anode potential for various types of anode. The anode potential of Pb-0.52 wt% Ag - 0.50 wt% Ca anodes annealed after rolling is lower than that of the Pb-0.99 wt% Ag conventional anode and that of the same chemical composition of as-cast anode. However, the anodes annealed without rolling showed high values of the anode potential compared to that of the as-cast anodes for current densities over 70 mA/cm2. The values of the anode potential for the anodes rolled and not annealed were high compared to these of the as-cast anodes, and for this reason, these values were not plotted. It can be observed that the rolled Pb-Ag-Ca anodes must be further annealed in order to obtain a decrease in the anode potential during the electrolysis process.

20 40 60 80 Current density (mA/cm2)

100

as-cast Pb-0.99%Ag

as-cast Pb-0.52%Ag-0.50%Ca

annealed Pb-0.52%Ag-0.50%Ca

annealed after rolling Pb-0.52%Ag-0.50%Ca

Figure 4 - Effect of the Current Density on the Anode Potential for Several Electrodes (Zn : 60 g/1, H2S04 : 160 g/l, 40°C, annealed at 200°C for 48 h, rolling rate 40%)

20 Day Electrolysis Experiments

The anodes used for the 20-day electrolysis experiments were made using a Pb-0.50 wt% Ag - 0.62 wt% Ca alloy. The anodes were rolled at a rate of 20, 40, and 60 % and were annealed at 200° C for 13,48, 96, 144, and 192 h. After each electrolysis experiment, the anode surface was investigated by SEM, and the lead oxide species were identified by X-ray diffraction.

604 LEAD-ZINC 2000

Figure 5 shows the relation between the electrolysis time and the bath voltage for some types of anodes. For comparison, this figure also shows the variation with time of the bath voltage for a Pb-0.9 wt% Ag anode annealed for 48 h at 200° C. As shown in Figure 5, the bath voltage of the anode made from the ternary alloy material annealed for 48 h at 200°C decreased the bath voltage, and after two days electrolysis time, the bath voltage was 50 mV lower compared to that of the Pb-0.9 wt%Ag anode. Moreover, it can be observed that the difference between the bath voltage corresponding to the Pb-0.9 wt%Ag and the ternary alloy anodes increases with increasing annealing time.

annealed at 200°C for 48h

W-K-*

4 8 12 16 Time of electrolysis (day)

-*- not rolled Pb-0.9%Ag (annealed at 200°C for 48h)

-*- not rolled Pb-0.50%Ag-0.62%Ca

-*- rolling rate 20% Pb-0.50%Ag-0.62%Ca

-o- rolling rate 40% Pb-0.50%Ag-0.62%Ca

- ° - rolling rate 60% Pb-0.50%Ag-0.62%Ca

Figure 5 - Bath Voltage versus Time of Electrolysis for Several Anodes (H2S04 : 160 g/l, 40°C, 50 mA/cm2)

Figure 6 shows that the bath voltage decreases with increasing annealing time. For example, the bath voltage of the anode annealed for 192 h at 200°C was 170 mV lower compared to the bath voltage of the Pb-0.9 wt%Ag anode. Moreover, this figure shows that the rolling rate has a small influence on lowering the bath voltage.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 605

48 96 144 Annealing time (h)

192

—Π— not rolled Pb-0.50%Ag-0.62%Ca

—♦— rolling rate 20% Pb-0.50%Ag-0.62%Ca

—O— rolling rate 40% Pb-0.50%Ag-0.62%Ca

rolling rate 60% Pb-0.50%Ag-0.62%Ca not rolled Pb-0.9%Ag

Figure 6 - Bath Voltage versus Annealing Time (H2S04 : 160 g/l, 40°C, 50 mA/cm2,20 days electrolysis time)

Figure 7 shows the SEM images of the anode surface after an electrolysis time of 20 days. From Figure 7, it can be observed that the anode surface of the Pb-0.9 wt%Ag anode is rough compared to the Pb- 0.50 wt%Ag-0.62 wt%Ca anode. In this experiment, no significant influence of the rolling rate or the annealing time on the quality of the anode surface was observed.

Pb-0.9%Ag not rolling annealed at 200°C for 48h

Pb-0.50%Ag-Ö.62%Ca rolling rate 40% annealed at 200°C for 48h

Figure 7 - SEM Images of Anodes Surfaces after 20 Days Electrolysis Time (H2S04 : 160 g/l, 40°C, 50 mA/cm2)

606 LEAD-ZINC 2000

Figure 8 shows the X-ray diffraction patterns of the surface of some types of anode used in the 20-day electrolysis experiments. For the Pb-0.9 wt% Ag anode, the diffraction detects ß-PbC>2 peaks as well as a large peak of a-Pb02- For the Pb-0.50 wt% Ag - 0.62 wt% Ca anode material, p-Pb02 diffraction peaks were obtained and these peaks become sharper with increased annealing time.

a:a-PbCfc ß:ß-PbO> Pb-0.9%Ag annealed 200°C 48h

JiJ^u_-A^ Pb-0.50%Ag-annealed 200°C 48h

Nl

I Pb-0.50%Ag-annealec 200°C 144h

W U _ J W J V 20 30 40 50

i (degree Cu Ka) 60

Figure 8 - The X-ray Diffraction Patterns for Pb-0.9%Ag and Pb-0.50%Ag-0.62%Ca Anodes after 20 Days Electrolysis (H2S04 : 160 g/l, 40°C, 50 mA/cm2)

Anode Structure Observation

The base structure of the Pb-1 wt% Ag and Pb-0.52 wt% Ag - 0.65 wt% Ca anodes was observed using the EPMA observation method. Before the applying the EPMA observation techniques, the samples were etched in an acetic acid- hydrogen peroxide-water solution. Figure 9 shows the SEM images of the base structure of the as-cast Pb-1 wt% Ag and Pb-0.52 wt% Ag - 0.65 wt% Ca alloy materials. These SEM images show that the Ag-rich phase is distributed at the grain boundaries for both types of anode materials.

In the Pb-1 wt% Ag alloy, the Ag-rich phase is fine and almost homogeneously dispersed at the grain boundaries. For the Pb-0.52 wt% Ag - 0.65 wt% Ca alloy, the Ag-rich phase is scattered at the grain boundaries. In this experiment, it was observed that the content of

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 607

Ag in the Ag-rich phase is higher for the Pb-0.52 wt% Ag - 0.62 wt% Ca alloy compared to the Pb-1 wt% Ag alloy.

Moreover, in the ternary alloy a phase that seems to be CaPb3 was observed at the grain boundaries, and the amount of this phase increases with increasing Ca contents in the alloy. In the annealed samples, the CaPb3 phase was observed at the grain boundaries. In this alloy type, the Ag-rich phase was observed as round granules at the grain boundaries, and the silver concentration in the Ag-rich phase increased with any increase in the annealing time. From the Pb-Ca phase diagram (2), it can be observed that, at room temperature, Ca is hardly dissolved in lead (<0.01%), and for this reason, the CaPb3 phase formed with excess Ca. The excess dissolved Ca in the lead grains during the pouring operation rapidly precipitates at the grain boundaries during annealing. After formation of this CaPb3 phase, the Ag content in the Ag-rich phase increases because the CaPb3 is an intermetallic compound, and the affinity of lead for Ca seems to be stronger than its affinity for Ag. Therefore, one factor that may lead to a lower anode potential is the catalytic effect of the Ag-rich phase separated at the grain boundaries.

On increasing the content of Ca in the anode alloy, a product that seems to be mainly an oxide was observed at the grain boundaries on the anode surfaces after 3 h of electrolysis. With the increase in the content of Ca in the anode material, the thickness of the oxide layer increases after 20 days electrolysis time; therefore, the presence of Ca in the anode was considered to lead to a rapid covering of the anode surface by the PbC>2 oxide layer. Also, the reaction of the oxide layer is closely related to the presence of the CaPb3 and Ag-rich phases at the grain boundaries.

Figure 9 - SEM Images of Several Alloys: (a) Pb-1.00 %Ag, as-cast, (b) Pb-0.52 %Ag-0.65 %Ca, as-cast, (c) Pb-0.52 %Ag-0.65 %Ca, annealed at 200°C for 13 h (not rolled), (d) Pb-0.52

%Ag-0.65 %Ca, annealed at 200°C for 48 h (not rolled)

608 LEAD-ZINC 2000

Effect of the Addition of Ethylene Glycol

At the Zinc&Lead'95 Symposium, it was reported (1) that the bath voltage decreased on adding alcohol to the electrolyte. In this study, the influence of the addition of ethylene glycol to the electrolyte was investigated.

2.2

2.1 -

> 1.9

1.7

1.6

1.5

. . £2'—' C§ O No additive Π A Δ V ■ - Pb-0.99%Ag

D 0 D 4

-56 6 $8

No additive

Pb-0.99%Ag

» = Pb-0.5I%Ag-0.25%Ca

k= Pb-0.50%Ag-0.68%Ca

> - Pb-0.48%Ag-0.88%Ca

0.2mol/l ethylene glycol

D - Pb-0.99%Ag

0 = PI>-0.51%Ag-0.25%Ca

Δ - Pb-0.50%Ag-0.68%Ca

0 - Pb-0.48%Ag-0.88%Ca

20 40 Current density

60 80 (mA/cnV)

100

Figure 10- Relation Between the Current Density and Anode Potential (H2S04 : 160 g/l, 40°C, annealed at 200°C for 100 h)

The relationship between the current density and anode potential for several anode types using an electrolyte containing 0.2 mol/1 ethylene glycol is shown in Figure 10. This figure shows that the characteristics of the electrolyte were improved because the anode potential decreased after the addition of the ethylene glycol in the electrolyte. The experimental values of the anode potential measured in an electrolyte containing 0.2 mol/1 ethylene glycol were 100~150 mV lower than the values in the electrolyte without ethylene glycol, for all types of anode. No influence of the anode composition on the anode potential values was observed, for annealing the rolled anode materials, in electrolytes containing ethylene glycol.

Figure 11 shows the effect of the addition of ethylene glycol in the electrolyte on the relation between the electrolysis time and the anode potential and generated gas quantity. The quantity of the generated gas on the anode surface was measured using a gas collector tube installed over the anode. The gas collector tube was kept at 40° C, and for the calculation of the gas quantities, the measured values were converted to the standard state (0° C and latm). It is clearly shown that the addition of ethylene glycol significantly contributed to the decrease in the anode potential and generated gas quantity, because both rapidly decreased when ethylene glycol was added at the middle of the electrolysis experiment.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 609

Ζ2·1

-jj^jjj^g^JHJH^-π^^^>D-c>D^IH>c>o■

120 Electrolysis time (min)

180

Ι - ε ^ > [ > αα Λ = Η > Μ - { ^

i t i t a ^ i i t i ^ 60 120

Electrolysis time (min)

I60g/IH;SO.

160g1 HJSO. + 0.2mol/l ethylene glycol

ethylene glycol added after 60min

180

Figure 11 - Effect of Addition of Ethylene Glycol in the Electrolyte (as-cast Pb-0.20wt%Ag-0.59wt%Ca, 40°C, 50 mA/cm2)

It could be demonstrated that the anode does not dissolve by the addition of ethylene glycol in the electrolyte because the anode potential did not rise, the content of lead in the electrodeposited Zn was low and the surface of the anode did not show any changes.

The theoretical value of the generated oxygen on the anode surface, using the following reaction, was calculated to be 1.74 ml/lOmin (at 0°C and latm) but this value is large compared to the experimental value.

H20 - 2H+ + l/202 + 2e" (1)

As reported in the previous study (1), in an electrolysis cell using an active catalyst electrode, carbon dioxide was generated when alcohol was added in the electrolyte. For the addition of ethylene glycol in the electrolyte, assuming the following chemical reaction, the quantity of carbon dioxide can be calculated.

C2H602 + 2H20 - 2C02 + 10H+ +10e" (2)

The calculated carbon dioxide quantity using reaction (2) was about 1.39 ml/10min (at 0° C and latm), but this value differs from that obtained by the experimental measurement. Because in this experiment the Pb base electrode material did not act as an active catalyst electrode, the probability of the occurrence of reaction (2) is very small.

The steps of the ethylene glycol oxidation mechanism (3) are: formation of glycolic aldehyde, glyoxal or glycolic acid, glyoxylic acid, oxalic acid and finally, formic acid is obtained. For this reason, in the next experiment, the effect of adding glyoxal and oxalic acid in

610 LEAD-ZINC 2000

the electrolyte was investigated. The electrolyte used in the experiment was prepared using an extra pure reagent grade of ethylene glycol, glyoxal and oxalic acid.

Figure 12 shows the relation between the concentration of each of the additives and the anode potential, and also illustrates their influence on the quantity of the generated gas. It can be seen that the anode potential and the quantity of generated gas decreased with increasing concentrations of each of the additives.

It can be noted that the reaction mechanism and also the kind of product resulting after adding ethylene glycol in the electrolyte are still unknown and that the possibility of regenerating a depolarizing substance is still under study.

3 1 1-2.1 —■ □— Ethylene glycol

Figure 12 - Relation Between the Concentration of Each Additive and the Anode Potential and the Quantity of Generated Gas (as-cast Pb-0.52%Ag-0.51wt%Ca, H2S04 : 160 g/l, 40°C, 50

mA/cm2)

Mechanical Properties of Pb-Ag-Ca Alloys

Tensile Tests

The strengths of the Pb-Ag-Ca alloys were experimentally examined using tensile and Vickers hardness tests. The results were compared to the data for a conventional Pb-Ag alloy used as anode material for zinc electrowinning.

Figure 13 shows the relation between the Ca content in the Pb-0.5 wt% Ag - Ca alloys (as-cast or annealed at 200° C for 48 h) and the tensile strength and yield strength. The tensile strength and yield strength of the as-cast Pb-0.5 wt% Ag - Ca alloys are stronger compared to those of the conventional Pb-0.9 wt% Ag alloy and the annealed Pb-0.5 wt% Ag - Ca alloy. The tensile strength increases with increasing Ca content in the Pb-0.5 wt% Ag - Ca system alloy to

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James andA.H.-J. Siegmund 611

about 0.1 wt%. This result agrees fairly well with the data reported by Petrova et al. (4). They reported an increase in the Brinell hardness of the Pb-0.5% Ag-Ca alloys at about 0.05 mass% Ca, then the hardness decreased with increasing calcium content in the alloy. The reason for this behavior is probably related to the solubility of Ca in Pb. From optical microscope observations it can be concluded that the structure of Pb-0.5 wt% Ag - Ca alloys with calcium contents below 0.1 wt% resembles that of the Pb-0.9 wt% Ag binary alloy. When the content of Ca in Pb-0.5 wt% Ag - Ca increased over 0.1 wt%, the CaPb3 phase was observed at the grain boundaries. In addition, from the SEM observations of the specimen fracture surfaces after the tensile tests, ductile fracture was observed and the precipitate at the grain boundaries seemed to be the origin of the rapture. From the above result, it is considered that the tensile strength increases with increasing contents of Ca in the Pb-0.5 wt% Ag - Ca alloy for a Ca content less than about 0.1 wt%. An increase in the Ca content in the alloy higher than 0.1 wt-% leads to a decrease in the tensile strength as a result of the precipitation of the CaPb3 phase. However, the tensile strength and yield strength of both as-cast and annealed Pb-0.5 wt% Ag -Ca system alloys were the same or higher in comparison with those of the conventional Pb-1 wt% Ag alloy.

60

50

40

«>

^ 30

•3

£

20

10

o

0° 0 o

o

*A &

w

o Tensile strength

Δ Yield strength:

• Tensile strength

A Yield strength:

o°o

0°<9 ° °

Sg.

A

"i

1 !

0 8 0 °

• Δ

Δ

: as cast

as-cast

: annealed

annealed o

° o o o <gb o

• *

~&f ä" "f as-cast Δ Pb-0.9%Ag Δ tensile strength

T as-cast Pb-0.9%Ag yield strength

1 , 1 ,

0.2 0.4 0.6 0.8 1.0 Ca content in the Pb-0.5wt%Ag-Ca alloy (%)

Figure 13 - Tensile Strength and Yield Strength versus the Ca Content in the Pb-0.5wt%Ag-Ca Alloy (annealed at 200°C for 48 h)

Figure 14 shows the relation between the annealing time of rolled Pb-0.42 wt% Ag -0.65 wt% Ca alloy material and the tensile strength and yield strength. After annealing at 200°C for 48 h, the measured values were stable. The ternary alloy annealed for 48 h after rolling showed tensile strength and yield strength values almost the same or higher than those of the Pb-0.9 wt% Ag alloy. However, when the annealing time was long and the rolling rate increased, the yield strength tended to decrease.

612 LEAD-ZINC 2000

48 96 Annealing time

Figure 14 - Tensile Strength and Yield Strength for the Pb-0.47wt%Ag-0.63wt%Ca Alloy (annealed at 200°C)

Vickers Hardness Tests

Figure 15 shows the Vickers hardness test results for various alloys. The increase in the Ca content of the Pb-0.5 wt% Ag - Ca alloy was considered to increase the Vickers hardness of the alloy. The rolling rate increased the hardness of the Pb-0.5 wt% Ag - Ca alloy, but after annealing, the hardness decreased. Because the Vickers hardness shows a constant value after annealing at 200° C for 48 h, it was considered that the strain in the anode material could be removed by annealing at 200° C for 48 h. The Pb-0.5 wt% Ag - Ca alloy annealed at 200° C for 48 h after rolling showed the same or better mechanical properties relative to the conventional Pb-1.0wt%Ag alloy.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Pb-1.05%Ag

Pb-0.54%Ag-0.30%Ca

Pb-0.48%Ag-0.58%Ca

Pb-0.52%Ag-0.97%Ca

Pb-0.53%Ag-0.57%Ca rolled 20%

Pb-0.48%Ag-0.55%Ca rolled 40%

Pb-0.47%Ag-0.59%Ca rolled 60%

48 96 144 192 Annealing time (h)

240

Figure 15 - Vickers Hardness versus Annealing Time (annealed at 200°C)

CONCLUSIONS

The increase in the contents of Ca and Ag in Pb-Ag-Ca anodes tends to lower the anode potential. The anode potential of Pb-0.52 % Ag-0.50 % Ca anodes annealed after rolling is lower compared to that of the as-cast and Pb-0.99 % Ag conventional anodes. The values of the anode potential for the rolled anodes and those not annealed were high compared to the as-cast anodes. The difference between the bath voltage corresponding to the Pb-0.9 wt% Ag and the Pb-0.50 wt% Ag-0.62 wt% Ca alloy anodes increases with increasing annealing time, and the rolling rate has a small influence on lowering the bath voltage.

In the Pb-1 wt% Ag alloy, the Ag-rich phase is fine and almost homogeneously dispersed at the grain boundaries, and in the Pb-0.52 wt% Ag-0.65 wt% Ca alloy, the Ag-rich phase is scattered at the grain boundaries. The content of Ag in the Ag-rich phase is higher for the Pb-0.52 wt% Ag - 0.62 wt% Ca alloy compared to the Pb-1 wt% Ag alloy.

For the as-cast and annealed samples, the CaPb3 phase was observed at the grain boundaries. The amount of this phase increased with increasing Ca content in alloy. In the annealed alloy, the Ag-rich phase was observed as round granules at the grain boundary, and the silver concentration of the Ag-rich phase increased for any increase in the annealing time.

The Pb-0.2 wt% Ag - 0.59 wt% Ca anode potential at a current density of 50 mA/cm2 in an electrolyte containing 0.2% ethylene glycol decreased about 140 mV compared to the Pb-1 wt% Ag anode in an electrolyte with no depolarizing substance. No differences between the

614 LEAD-ZINC 2000

anode potential values of as-cast, annealed and annealed-after-rolling anode materials in the electrolyte containing ethylene glycol were observed. The anode potential and the quantity of gas generated decreased with increasing concentrations of each of the depolarizing additives.

The tensile strength and yield strength of both as-cast and annealed Pb-0.5 wt% Ag-Ca system alloys were the same or higher in comparison with the conventional Pb-1.0 wt% Ag alloy. The increase in the Ca content in the Pb-0.5 wt% Ag-Ca alloy increased the Vickers hardness of the alloy. The rolling rate increased the hardness of the Pb-0.5 wt% Ag-Ca alloys but after annealing, the hardness decreased.

REFERENCES

1. K. Koike, H. Watanabe and N. Masuko, "Energy Saving in Zinc Electrowinning with Modification of Lead Anode," Zinc & Lead '95, T. Azakami, N. Masuko, J.E. Dutrizac and E. Ozberk, Eds., The Mining and Materials Processing Institute of Japan, Tokyo, Japan, 1995, 373-383.

2. M. Hansen, Ed., Constitution of Binary Alloys, Second Edition, McGraw-Hill Book Company, New York, U.S.A., 1958, 405-406.

3. An editorial committee of KAGAKU DAIJITEN, KAGAKU DAIJITEN. Vol.1, Kyoritsu Syuppan, Tokyo, Japan, 1960, 908.

4. M. Petrova, Z. Noncheva, Ts. Dobrev, St. Rashkov, N. Kounchev, D. Petrov, St. Vlaev, V Mihnev, S. Zarev, L. Georgieva and D. Buttinelli, "Investigation of the Processes of Obtaining Plastic Treatment and Electrochemical Behaviour of Lead Alloys in Their Capacity as Anodes During the Electroextraction of Zinc, I. Behaviour of Pb-Ag, Pb-Ca and Pb-Ag-Ca Alloys," Hvdrometallurgv. Vol. 40,1996,293-318.

Chapter 8

New Developments in Lead and Zinc

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 617

EQUILIBRIUM VERSUS KINETICS IN LEAD REFINING

T.R.A. Davey METACON Consulting Services

Unit 220, 57 Gloucester Ave Berwick, Victoria, Australia 3806

ABSTRACT

It is often important in the study of a process to determine whether the limitations are governed by equilibrium or kinetic considerations. Sometimes it is possible to devise a design of the process or the equipment so that the apparent. limitations set by equilibrium considerations are overcome. In this paper, detailed consideration is given to the lead refining processes of sulfiir-decoppering and of vacuum dezincing to illustrate this, and possible implications for other refining processes are indicated.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

618 LEAD-ZINC 2000

INTRODUCTION

It used to be generally taught that thermodynamics could indicate that a reaction could not proceed, but could not indicate that a reaction would proceed; it might be possible, but only at a speed so slow that in practice it would not proceed. Thus it would seem to be anomalous if a reaction proceeds in a direction away from the thermodynamically determined equilibrium. Examples of reactions proceeding away from the equilibrium concentrations are by no means rare in petrochemicals, because there are many instances of organic chemicals which react slowly. Thus, when an overall process involves more than one possible reaction, some fast and some slow, intermediate products may be formed, which later are consumed, at least in part.

Most pyrometallurgical processes are performed at temperatures high enough to ensure that all the possible reactions proceed very fast, but occasionally this is not so, and therefore the equilibrium condition of the system may not be optimal for the yield of a desired product of reaction. This is the case for the removal of copper from primary lead by reaction with elemental sulfur. It is highly desirable to know and understand this, in order to be able to specify the optimum conditions for the conduct of the process - in this case, known as decopperising, or fine decoppering or sulfur drossing.

When apparently only one reaction is to occur, it may still be of advantage to cause a second reaction to take place, so as to produce a more favorable outcome. This is not necessarily done by introducing additional reagents, but can possibly be done by suitable design of the reaction vessel. This is the case for the removal and recovery of zinc from lead after the desilverising operation. The operation is performed under vacuum, so that it will proceed at reasonable speed at relatively low temperatures, and is known as vacuum dezincing.

The material in the following sections contains no new data, and is not presented in great detail. The reader is referred to the original sources for this. The presentation here is intended to point up the necessity of being fully aware of the implications imposed by kinetic considerations in processing, and the dangers of seeking explanations in terms of equilibria when no such equilibrium is involved. The figures here are given merely to make the text more understandable. They have all been taken from previous publications, and again the full details may be found in the original papers.

DECOPPERING LEAD WITH SULFUR

The Industrial Process

Primary lead, produced from ore-concentrates in a blast furnace or other high temperature reaction vessel, commonly contains in solution a large number of impurity elements, chief among them being copper, sulfur, oxygen, arsenic, antimony, tin, bismuth, nickel, iron and noble metals. As soon as the lead is cooled, some of these impurities are thrown out of solution and the generated solid crystals, together with entrained liquid lead, form a dross. Beside a minor amount of oxides, the main constituents of the dross are sulfides, arsenides or antimonides of copper, and hence the cooling operation is termed copper drossing. When the lead has cooled nearly to its melting point (a little under 327°C) the copper content will have fallen to within the range 0.02-0.06% Cu, depending upon the amounts of residual sulfur, arsenic and antimony. Figure 1 shows the copper solubility, for a range of temperatures, as dependent upon the lead's content of other relevant solutes. In the case of many primary lead producers, especially those smelting predominantly sulfide ores, not excessively high in

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 619

arsenic or antimony, the copper content of the lead after hot drossing to near its freezing point will be about 0.06% Cu. It is desirable to reduce the copper content to no more than a few hundredths of a percent before proceeding to the subsequent stages of lead refining, in order to reduce complications caused by the presence of significant amounts of copper in the by-products. This further reduction of copper content can be effected by stirring elemental sulfur into lead just above its melting point, and the process is generally termed "sulfur drossing", or "fine decoppering". (The names of Hülst or Colcord (1) have also been associated with this process.) It is possible in this way to reduce the copper content of the lead to about 1 ppm (0.0001% Cu) experimentally, and very good plant practice can produce results of 0.001-0.002% Cu. Thus the process starts at the equilibrium value (0.02-0.06% Cu, depending on the amounts of S, As and Sb in the hot- drossed lead) and proceeds away from this, to about 0.001-0.002% Cu.

Figure 1 - Solubility of Copper in Lead as a Function of Temperature and Composition. Note the Values of 0.02 - 0.06% Cu at the Freezing Point of Lead. See Reference (2)

The sulfur dross, being skimmed near the freezing point of lead, contains a great deal of entrained metallic lead, which is partially oxidised. Some of this entrained lead may be removed by continued stirring at a higher temperature. The dross then becomes "dry" or powdery with a lesser amount of entangled metallic lead. However, some copper thereby redissolves in the lead, and it is not unknown for the lead's copper content to return to a figure of nearly 0.06% Cu again and to render the sulfur drossing operation quite ineffective. Another practice, which was formerly in vogue, was to continue stirring the dross, with the addition of sawdust, but without raising the temperature. This had the effect of reducing the degree of oxidation of the dross particles, and further decreased its lead content.

620 LEAD-ZINC 2000

An alternative to the use of sulfur is the use of a mixture of sulfur and pyrites, when the temperature used may be somewhat over 330°C, otherwise the process is the same as described above, but reversion of copper into the lead is not then a problem.

The Facts

The literature teems with misinformation about this industrial process, particularly regarding the imagined effect of other impurities in the lead on the effectiveness of the decoppering process. In fact, using sulfur alone as the reagent, either of the elements silver or tin must be present for the process to decopper lead to low values (10-30 ppm, or 0.001-0.003%Cu). None of the other elements often present is necessary for successful fine decoppering, although their presence in the lead during the previous stage of hot copper dressing may have lowered the starting level of copper, and so made it easier to achieve low final figures. Very high antimony values (several percent in lead) make it difficult to reach low final copper figures.

About 0.1% of granular sulfur suffices as the reagent, and a great excess merely causes more lead to enter the dross as PbS. Continuing stirring for more than a few minutes after adding the sulfur causes copper to be redissolved in the lead. This reversion of copper is also promoted by raising the temperature before the dross is skimmed. When Ag is the element promoting fine decoppering, the achievement of low final copper figures is more difficult if the temperature is above 330°C, but if tin is the promoting element, then rather higher temperatures enable the decoppering to proceed more quickly. The dross skimmed contains essentially PbS and CU2S, and sometimes some CuS, as well as entangled metallic lead, which is partly oxidised to PbO.

When using pyrites plus sulfur (4:1 is usually the optimum proportion) there is no reversion of copper, and the use of much higher temperatures is beneficial (more rapid elimination to lower final figures). Again, antimonial lead is more difficult to decopper than soft lead. When adding pyrites alone, no appreciable decoppering occurs, whether or not Ag or Sn is present. This is surprising, since pyrites contains free, or labile, sulfur.

In the absence of Ag or Sn, sulfur alone can decopper lead only if repeated, large doses of sulfur are stirred in, but the large amount of lead sulfided makes the practice of no commercial significance.

A Curious Phenomenon

In 1948, from study of ternary phase relations, the author realised that the limit of removal of copper from lead by stirring with sulfur should be only a little less than 0.06%Cu. It resulted in the equilibrium between liquid lead containing copper, solid PbS and solid CU2S, according to the equation:

Cu2S + Pb, -> 2CuPb + PbS (1)

Mr. G.M.Willis, then Reader in Chemical Metallurgy at the University of Melbourne, became interested in this phenomenon. Using the then recently published data of Kleppa (3) for activities in the Cu-Pb system, he confirmed that the theoretical limit for copper removal down to the ternary eutectic point of the Pb-Cu-S diagram is indeed 0.05% Cu (4). Discussion of this with Professor F.D. Richardson of Imperial College, London, produced the suggestion that kinetic effects could be playing a significant role (5,6), since the system started at equilibrium, and proceeded away from it. Subsequent investigations by many workers have confirmed this

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 621

beyond doubt. Despite this, numerous investigators have attempted, without any success, to find an interpretation in equilibrium terms rather than in kinetic considerations.

Summary of the Reactions Occurring

When Stirring with Sulfur

When a number of reactions can occur simultaneously, some fast and some slow, then conditions are provided for the temporary occurrence of non-equilibrium conditions (7,8,9) and this is the case here:

Pbi + Si -> PbS fairly slow (2)

(very slow in presence of Ag or Sn)

Cupb + Si -> CuS very fast (3)

CuS + Cupb -> Cu2S fast (4)

2CuS + Pb, -> Cu2S + PbS very slow (5)

Cu S + Pb, -> Cupb +PbS slow (6)

Cu2S + Pb, ->2CuPb + PbS slow (1)

Where suffix Pb denotes dissolved in liquid lead, suffix 1 denotes liquid, and all other constituents are crystalline solids.

When Stirring with Pyrites and Sulfur

In this case (10) it is not necessary to postulate that movements away from equilibrium occur. The additional reactions have been shown to be:

FeS2 + Pb -> FeS + PbS vanishingly slow (7)

FeS2 + Cupb -> CuFeS2 very fast (8) (if elemental S present)

CuFeS2 + 4CuPb + 2S, -> Cu5FeS4 fast (9)

CuFeS2 + Pb -> Cu, + FeS + PbS vanishingly slow (10)

Cu5FeS4 + 3Pb -> 5Cu, + FeS + 3PbS vanishingly slow (11)

Experimental Difficulties

Many investigators have not attempted to simulate the industrial procedure, but have sought to explain the speeds of reaction on the basis of diffusion rates, and have paid attention to defects in the PbS. Their results have sometimes led to conclusions that either have no connection with the industrial process, or worse, have been quite misleading, as their predictions of the effects of impurities run counter to factual experience.

622 LEAD-ZINC 2000

Some of the confusion that exists about the process is probably largely due to the fact that the plant operation depends very much on the skill of the workman. The sulfur must be thrown into the vortex, at a regular rate over about 10 minutes, or else some floats on the surface of the lead, and is lost by burning. The sulfur must be of a suitable size; if too fine, it bums before hitting the lead surface. If it is too coarse, it does not dissolve completely before emerging again on to the lead surface, where the liquid sulfur bums. The lead temperature must be maintained close to the freezing point. The stirring must be continued for just a short time (a few minutes) following sulfur addition, before the dross is skimmed. The results also depend upon the knowledge of the supervisor that "drying" of the dross by very prolonged stirring or by raising the temperature will cause re-solution of copper into the lead. Similar difficulties may occur with laboratory tests, and have led to numerous false conclusions in the published literature.

When stirring is stopped to take samples of lead or dross during a test, it is important not only to avoid contamination of the samples, but also to provide an experimental regime that permits stirring to be resumed in an identical manner to what went before. Figure 2 shows an apparatus, which has been found to fulfil these conditions very well. It was first used by Pereira and Campos (11) and gives excellent results, because of the possibility of keeping the conditions standard throughout a test, and sampling reliably. Without this inbuilt sampling device, it is usually necessary to remove the stirrer to obtain a lead sample. To put back the stirrer to the identical position is very difficult and sometimes the subsequent lead flow pattern is sufficiently different to give inconsistent results, so that the elimination curves are unreliable. What appears to be a further improvement in design appeared in a recent Russian publication (12), showing the sampling rod passing through a hollow shaft carrying the stirrer blades. This would cause only negligible disturbance of the lead flow pattern during stirring.

Figure 2 - Apparatus: Stirring Pot Incorporating Sampling Device (10)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 623

In each of the tests reported here a constant amount of lead was used. It contained about 0.05% Cu. With another added element to determine whether such an impurity had any effect on the decoppering, stirring was interrupted for one minute to take samples of lead or dross. The metal temperature was held constant within 2° C of the target value. Several stirring speeds were tested. Full details are given in References (10,13). Sulfur particles in a narrow size range were added in a steady stream right at the start, or as several increments at constant intervals of time.

Experimental Results

Figure 3 shows the results of tests conducted to determine the amounts of silver or tin required to ensure that lead can be decoppered with sulfur to around 0.0015 Cu, at 330°C, within about 10 minutes. The lead contained about 0.05% Cu initially, and varying amounts of tin or silver, but no significant amounts of other impurities. Increments of 0.02% S were added each minute, and lead samples were taken after 5 minutes and after 10 minutes. It will be noted that 0.02% Ag suffices to permit 0.1 % sulfur to decopper to 0.001% Cu, even within 5 minutes, and the results are not significantly different for higher amounts of silver. Sn behaves differently. With 0.05 % Sn present, 0.001% Cu is reached after 10 minutes stirring, and a total addition of 0.2% S. On the other hand, 0.001% Cu is reached with 0.10% Sn after 5 minutes stirring, and the addition of 0.1% S in total.

0.0001 0 0.02 O.Oi» 0.06

PERCENT SILVER ADDED

006

0.02

^ 0.01

g 0.006

^ 0.002

0.001

0.0006 0.0001

0,0002

-

-

1 1

10MIN>\^

1 1

1 1

~ ^ \ S MIN.

1 1

-

-

-

-

—

-

0.02 O.Ot 0.06 O.t

PERCENT TIN ADDED 0.1

Figure 3 - Effect of Silver or Tin on the Effectiveness of Decoppering with Sulfur, See Text and Reference (13)

Figure 4 shows a small selection of the results obtained using the apparatus of Figure 2. The points on the elimination curves are omitted for the sake of clarity. Reference 10 gives detailed results and a description of the experimental technique. Briefly, this was to hold a standard amount of lead, containing about 0.05% Cu, plus other elements as required, at a constant temperature, and to stir in the reagent (sulfur or sulfur/pyrites mixture).

624 LEAD-ZINC 2000

^ 0.002

\

K 1 \

, \

\ \ - v. \

-I

1 1

NoAg.Hulfipl.SAddm™

~~"~"- 5.5XSb

. - » . .

^ ^ ^ ^

^ ^ ^

1 I

1

— p

P

i » —

THARD J LEAD S LOW Ag

U.0W Sn

«^———L

-

P

P

0.04% Ag

No Ag

0.002% Ag

20 30

STIRRING TIME, MINS.

Figure 4- Decoppering Soft Lead (with or Without Ag), Hard Lead (Ag-free and Sn-free, with Varying Amounts of Sb) with Sulfur or Sulfur plus Pyrites (Curves Denoted by P),

For Various Temperatures and Rates of Stirring, See Reference (10)

The triangular points on some of the curves show the time at which the dross changed in appearance from silvery-metallic to a dark greyish or blackish powder, which floated on the surface and was no longer drawn into the vortex. Microscopic examination of the dross showed that there was not a sudden transition, despite the optical appearance, but a gradual increase in the proportion of oxidised material. Thereafter, the dross was no longer wetted by the lead, and could no longer remove further copper, but reversion was still possible.

When stirring with sulfur alone, reversion of copper occurs when silver or tin is the "catalyst" retarding the reaction between sulfur and lead. When the reagent is a mixture of pyrites/sulfur (in the ratio 2:1 - 4:1) rather than S alone, no reversion of copper into lead occurs. Blast furnace lead can be rapidly decoppered to very low values, and the operation is faster at temperatures up to at least 450°C. This procedure does not require the presence of either Ag or Sn. With soft lead, when Ag is present, values as low as 1 ppm Cu were achieved experimentally. Hard lead, with no Ag or Sn, could be decopperised to 0.01%Cu in the presence of 2.3% Sb, or to 0.02%Cu with 5.5% Sb. Thus increasing Sb contents make it impossible to decopper to very low values using a moderate quantity of S alone, and difficult to decopper even with S plus FeS2- Using repeated treatments can, however, continue to remove copper to much lower values.

When no silver or tin is present, repeated additions of sulfur alone produce some decoppering, but the PbS formation is excessive, and the procedure has no commercial significance.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 625

Historical

It is not proposed to detail here all the individual steps and contributions which have led to the present understanding of fine decoppering, especially since the references cited contain several quite extensive surveys, and the reader is especially directed to references (14, 15a, 16). The following outline illustrates the unwillingness of many investigators to recognise or accept the implications that we have to deal with kinetic phenomena, and it is futile to search for an explanation in terms of thermodynamic equilibria.

Following the Colcord patent (1) in 1923, the previous practice of decoppering by using zinc began to be replaced by sulfur treatment. This had the advantage of removing copper separately from gold and silver, and of producing much less total dross. This was the point being made by George (17) in the (mistakenly) oft-cited reference to CuS in his publication in 1930. George was dealing with the replacement of zinc by sulfur as the decoppering reagent, to reduce the amount of skimmed material to be retreated. He referred in passing to the formation of PbS and CuS, which were then stirred with rising temperature to produce powdery dross. The point he was making was that a sulfide dross, with relatively little entrained lead, was produced, instead of the metallic zinc-copper mush with a high lead content. Presumably Tafel (18) was also just guessing that CuS was the copper sulfide concerned, as was Richardson (5) in 1951. In 1931 Dice, Oldwright and Brighton (19) made the first extensive microscopic examination of copper drosses, and showed indubitably that the copper sulfide in the dross is Cu2S.

In 1933-1934 extensive plant and laboratory tests at Port Pirie (20) established that 0.1% of sulfur is the optimum, and that temperatures much above the freezing point reduce the effectiveness of copper removal. It was also shown that stirring with PbS instead of S does not decopper lead. Numerous investigators in both plant and laboratory have confirmed this. Prolonged stirring, after the sulfur addition ceases, causes re-solution of the copper from the dross. In 1954 Cunningham (21) at Port Pirie showed that blast furnace lead could not be successfully decoppered in the absence of silver. Kleinen (22) in 1940 had already shown that a tin content of at least 0.13% was required to decopper antimonial lead containing 15-23% Sb, but because of the war this was not generally known outside Germany until much later. Blanks and Willis (23) in 1958 established that silver slows the reaction (2) of lead with sulfur, and thus favors reaction of sulfur with copper rather than lead. They also believed that tin appears to increase the rate of reaction of lead with sulfur, indicating that its promotion of the decoppering process must be due to another factor. It was attributed to the likely formation of a double sulfide of copper and tin, which would lower the activity of copper in the dross. On the other hand, Davey and Happ (13) considered that their results showed that tin, like silver, probably acted catalytically to slow the reaction of Pb and S. This was proven by direct observation to be the case by Clark et al. (9).

The first observation of CuS as a (transient) constituent of copper dross was made in 1956 by McClincy, a PhD student in Golden, Colorado, and Professor Schlechten, his department head at the Colorado School of Mines, gave colloquia in Australia on this topic. McClincy and Larson (24) showed that reaction (3) is very fast, so that CuS is always formed initially. Clark (9), Worner (25) and Pollard (26) all confirmed the early formation of CuS, but as it is rapidly consumed by reaction (4) it is never a major constituent of the final dross, which consists of CU2S, PbS, entrained Pb, and some PbO. Although several investigators have paid a great deal of attention to the transient existence of CuS, there seems to be no reason why the process should not proceed equally well if there were no formation of CuS.

626 LEAD-ZINC 2000

Clark (7) believed that PbS formed first and then reacted with Cu in solution to produce first CuS and then CU2S. After Willis and Davey (8) pointed out the impossibility of this, he published the now accepted explanation (9): initially, Cu and Pb compete with each other to form PbS and CuS. The CuS then reacts with Cu to form CU2S. The formation of PbS by any of the reactions above is undesirable but unavoidable. The function of Ag is to slow the reaction of Pb with S, as a sort of reverse catalyst, and the function of Sn is probably similar, but the formation of Cu-Sn compounds complicates matters somewhat (13). None of the other impurities present in lead has any noticeable effect on fine decoppering, but their presence beforehand may have reduced the level of copper present in the lead entering this operation.

Theoretical Mechanisms - Unsolved Mysteries

McClincy and Larson (24) correctly stated in 1969 that there is no point in looking for an explanation of decoppering in terms of thermodynamics, because the explanation lay in kinetics. Unfortunately, many investigators have ignored this recommendation. Pin and Wagner (27) attempted an explanation in terms of cation vacancies in PbS providing places for Cu ions, thus reducing their activity. They showed that the number of defects in doped PbS decreases in the order Bi, Sb, undoped, Sn, Ag, and claimed that this is the order of decreasing effectiveness in practice of these impurity elements. In fact, it is the reverse. Other similar attempts to link point defects in PbS to explain decoppering by lowering the CU2S activity are likewise pointless. The presence of crystals of Q12S in the dross shows that the activity of Q12S is unity, and has not been lowered at all, and in any case it is easily calculated that it would be impossible to generate sufficient PbS to dissolve it all.

Belis et al. (28) decoppered lead by stirring with sulfur-deficient copper sulfide, with Cu/S ratios varying from 1 to 1.98 (chalcocite to diginite). All decoppered to a minimum and then reversion brought the copper in lead back to the solubility limit for CU2S. Although they did not say so, this sounds the death knell for all the sulfur-deficiency explanations for decoppering to low levels.

Krysko (15b) suggested the formation of stromeyerite, a double sulfide of Cu and Ag, to reduce the activity of Cu, followed by its oxidation from the dross floating on the surface, to account for its non-appearance in the dross product. Once again, it could not possibly have a large enough effect.

A plausible explanation given by Terry et al. (29, 30) for the effect of silver is that, above a critical content of silver, the diffusion coefficient of lead in the lead sulfide layer decreases, because of a change of the PbS from n-type to p-type. They explain decoppering as the formation of a layer of solid PbS around the liquid sulfur particles, followed by rapid diffusion of copper, but slow diffusion of lead or sulfur, through this PbS layer. The CuS, and later CU2S, form against the sulfur surface. Reversion occurs when the sulfur and CuS are depleted, with the diffusion of Cu outward from CU2S into the liquid lead. A mathematical model was stated to give reasonable agreement with experiment. Although this model fails to account for the fact that PbS grains in the dross are frequently surrounded by the copper minerals, it does have the merit that the explanation is not self-evidently infeasible. It addresses the problem in kinetic, and not equilibrium, terms.

That the use of pyrites together with sulfur is much more effective in decoppering, than sulfur alone, was discovered (reportedly accidentally) at the Bixby refinery of ASARCO in 1970 (32); it is then unnecessary to have silver or tin present for effective decoppering. Elimination of the copper is faster, a higher temperature can be used if desired, and much lower final copper contents are attainable. In this case the copper enters into the pyrite grains as

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 627

chalcopynte or bornite, but a puzzling aspect is that this does not occur to any marked extent unless elemental sulfur is also stirred in (10). Possibly the sulfur acts in some way to improve the wettability of the pyrite grains by the liquid lead. It is curious that the labile sulfur content of the pyrite cannot act in the same way as elemental sulfur.

VACUUM DEZINCING

This case is essentially quite different from the foregoing, since it is obvious that one is dealing with a kinetic process. Nevertheless, there have been several misconceptions in the literature, which arise because the full implications of the kinetic aspects have not been recognised. These implications result from the fact that a condensation stage has been introduced close by the evaporation stage, with almost negligible resistance to vapor flow between them. The only resistance to be considered is that offered by residual inert gas to the diffusion of the zinc atoms.

The Industrial Process

Silver is almost invariably removed from lead by the addition of zinc, which forms an intermetallic Ag-Zn compound. After removal of this there remains about 0.54-0.60% Zn in the lead, depending upon the details of the desilverising process. The lead-zinc solution is at a temperature just above the melting point, and to remove the zinc by distillation, the lead must be heated considerably. Distillation at atmospheric pressure would require such an excessive temperature as to be uneconomic, and therefore the process is conducted under vacuum. Mild steel vessels can be operated to a temperature of about 600°C, and with suitable design can be made robust and vacuum-tight.

The proposal for vacuum dezincing came from Kroll (32), but the first practical development came in 1956 at the St. Joe refinery in Herculaneum, Missouri (33). Figure 5 shows the design schematically. Later developments (34) introduced several improvements similar to features introduced in the design of the continuous (actually semi-continuous) vacuum dezincing plant developed at Port Pirie, shown in Figure 6(a). These improvements to the batch plant include a change in direction of stirring, to produce more surface turbulence, so that a shorter treatment time or lower operating temperature was required to effect a given reduction in zinc content of the lead. The "bell" was replaced by a flat water-cooled top sealed to a water-cooled extension of the kettle rim, and the vacuum line entered through this extension, so that it did not need to be broken between charges, thus reducing the possibility of leaks. The stripping of zinc from the flat top was much safer and less arduous than from inside the bell; the area of lead "wetting' the kettle wall was much greater than with the original design, reducing the stress of heating.

All of these plants have a water-cooled condensing surface, producing what is referred to later as a "cold condenser", designated "CC". Evaporating and condensing surfaces are approximately equidistant throughout, but are not equal in area. The harmonic mean (reciprocal of the mean of the reciprocals) is the effective surface area for distillation (37). The vacuum line leaves the vacuum chamber near the point of the lowest zinc partial pressure, making the still "self-pumping".

628 LEAD-ZINC 2000

VACUUM LINE WATER COOLED TOP

ZINC DISTILLATE

LEVELLINGSCREW

LEAD KETTLE

BELL

FURNACE BRICKWORK

DRAFT TUBE

IMPELLOR

Figure 5 - Batch Vacuum Dezincing Plant, as at St. Joe, Herculaneum, See Reference (34) The Distance Between the Evaporating and Condensing Surfaces is About 30 cm

A Mechanical Vacuum Pump Produces a Residual Pressure of About 30 Microns Hg A Lead Temperature of 600°C and Stirring Time of 5 hours Reduced Zinc rom 0.56%

to 0.05% Zn in the Dezinced Lead.

A further development at the BHAS, Port Pirie (shown in Figure 6b), and later applied also at Imperial Smelting in the UK, was to use a condenser design which caused the condensing surface to remain at the melting point of zinc. It was later referred to as a "liquid zinc condenser", and designated as LZ. The unit was designed so that the heat entering the condenser (by radiation plus the latent heat of condensation of zinc) produces a temperature gradient in the solid zinc condensate to bring its surface to 420°C, the melting point of zinc. Small variations in the flow rate or temperature of the cooling water have no adverse effect on this automatic control.

This design of a plant was piloted at Port Pirie, South Australia. It was later used on a full scale in the UK at Avonmouth and Swansea. An alternative design was to replace the vertical falling film by a spiral launder, both in the pilot plant at Port Pirie, and later in the full-scale plants in the United Kingdom. This resulted in the evaporating surface being horizontal, and the vapor flow being nearly horizontal, and hence no longer normal to the evaporating surface. This has implications for the theoretical rate estimations, as discussed below.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 629

LIQUID ZINC CONDENSER COLD CONDENSER

Figure 6(b) - Continuous Vacuum Dezincing Plant With Liquid Zinc Condenser, See References (36,37) Lead Flow and Zinc Recovery Both Continuous

Figure 6(a) - (Semi)-Continuous Vacuum Dezincing Plant at B.H.A.S. Port Pirie, S. Australia - See Reference (35) Lead Flow is Continuous, Vacuum Broken for 10 Minutes every 16 Hours For Condenser Change and Zinc Removal

Distillation Rate

There have been many theoretical determinations of the rates of distillation under vacuum, but none that appeared to be applicable to vacuum dezincing when its development was commenced at Port Pirie in 1946. Thus it was felt to be desirable to develop the theoretical side (37) simultaneously with the practical development, as a guide to understanding and possible later application to optimising of the process. Figure 7 illustrates the concept of the distillation process which was developed. It was necessary to discard some faulty ideas or misconceptions, which derived from implicit notions associated with equilibrium, but not kinetic conditions. Carman (38), for example, assumed that the partial pressure of the vapour of the condensing species is equal to its partial pressure in the condenser. It is not, unless the condensation rate is zero. Richardson (6) assumed that the measured vacuum is equal to the distilling species, which it is not, but is instead the partial pressure of the inert atmosphere. Warner (40) assumed the partial pressure of zinc to be constant across the distillation space, which is not correct unless the distillation rate is zero.

The correct state of affairs in a vacuum still is depicted in Figure 7, showing the total gas pressure at any point as the sum of the inert gas and the distilling species (in this case zinc). Figure 8 illustrates how it is important to place the vacuum off-take, for optimum benefit of the vacuum. It has been calculated that, for the conditions prevailing in the practical vacuum dezincing process, the distillation rate is 2 - 20 times faster for a self-pumping than for a reverse-pumping still, with zinc contents ranging from 0.60% to 0.05% in the lead.

630 LEAD-ZINC 2000

Pressure

y \ — _ _ \

Pi ^

(Ti)

/ Evaporating / Surface

P

" P I T ^ ^ ^ ^ ^ .Condensing

/ Surface

<P2(T2)

Vacuum Line

t PN2 ^ ^ - ^ X Pc V

Distance

Figure 7 - Schematic Representation of Zinc Distillation through a Nitrogen Atmosphere See Reference 36

f PRESSURE

cu

o Q . «X.

>

P

P z n -

P N j _

N Ι Λ ■Z.

Q

VACUUM

, - ' " " V

DISTANCE-

PRESSURE

TO VACUUM PUMP

Pz„

DISTANCE "

VACUUM LINE

TO VACUUM " ] y PUMP

Figure 8 - Self-pumping and Reverse-pumping Stills (a) P = pZn +pN2 = p2 + V (b) P = pZn + pN2 = 0 + pi

See Reference 42 Pi

In failing to recognise fully the difference between kinetic and equilibrium conditions, when developing an expression for the distillation rate, the author first made the mistake of assuming that the pressure of a gas is the same in all directions. Of course it is not, when the gas has a large net velocity in one direction. Thus it was necessary to devise a correction to be applied when the gas has a net velocity which is a sizeable fraction of the mean molecular velocity. This enters into the expressions for evaporation and condensation, as shown below.

It is not the intention to repeat here the derivations which were published so long ago (37), but merely to emphasize that errors of conception arise when one does not discard completely the notions that pertain only to equilibrium, but not to kinetic, conditions. Some other errors of conception were committed or corrected at around that time (41,42,43).

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 631

As Figure 9 shows, when the net vapour flow away from an evaporating surface is normal to the surface, this gives rise to one set of conditions. As zinc atoms can leave the surface at any angle, the net mean component of molecular velocity normal to the surface is obtained by integrating over all angles. When the vapor flow is not normal to the surface, the changed geometry requires a different correction to be applied. Failure to add this further correction gives rise to errors by calculating too large an evaporation rate. As can readily be seen in Figure 9 (b), some of the evaporating zinc atoms are actually moving away from the condensing surface. Attempting to compensate for this error, Womer (40,44) introduced the well known and well understood "accommodation coefficient" at a value much less than unity. The accommodation coefficient has a respectable ancestry when used to account for the failure of complex molecules to condense on impact with the condensing surface, because of the effect of alignment. It should not, however, be used as a. fudge factor, "that must be considered as an unknown variable" (45), as it is well established that the accommodation factor is unity for monatomic species, and dealing with clean surfaces. The effective distillation rate was calculated roughly by the writer to be reduced to about a third of the expected rate from an evaporating surface parallel to the condensing surface (44). No fudge factor was then needed to reconcile the theoretical and practically observed rates.

Distilling Gas

LEGEND

v = net velocity of distilling gas

c - mean molecular velocity pmmv - projected mean molecular

velociry of evaporating atoms

Figure 9 - Vacuum Distillation - Theoretical (a) Vapor Flow Normal to the Evaporating Surface (b) Vapor Flow not Normal to the Evaporating Surface

For a complete account of the following derivations, the reader is referred to References (36,37,43,44). Expressions were derived for E, evaporation, M, migration across the distillation space, and C, condensation, and these were equated to give the overall distillation rate. Assuming the diffusion of zinc across the distillation space to be isothermal, the following relations apply:

632 LEAD-ZINC 2000

£ = a(pe-piW(M/2nRT,) (12)

M=bln(V/(V + p 2 -p i ) (isothermal) (13)

C = a(pc-p2)V(M/2nRT2) (14)

Putting E = M=C, and rearranging:

pe =V + c +(2E/a) - Ve_E/b (15)

where a, b and c are constants derived from physical constants of the system, or dimensions of the still (see Reference 37 or 38).

If the diffusion of zinc across the distillation space be adiabatic (which is more likely), a more complicated version of Equation (15) results:

E/a - c = (pe- E/a)0'2 (pe - E/a - V)08 (16)

Evaluations of Equations (15) and (16), applying values of the constants which would be appropriate for the vacuum dezincing of lead in practice, showed that there was no significant difference between the two. It also showed that the exponential term in Equation (15) could be neglected, without introducing an error of more than 1%, so that Equation (15) simplifies to:

E = 0.5a (pe- V - c) [gm/sec/cm2] (17)

If the partial pressure of zinc in lead can be approximated by a linear function of w, the weight percent zinc, as:

pc= xw + y,

then, replacing pe by w, and integrating between initial and final zinc concentrations, we have:

100At/B = (2/ax) In [(XWJ + y - V - c)/( xwf + y - V - c)] (18)

When the terms V and c in Equation (17) are placed equal to zero (complete vacuum and very cold condenser so that no back-evaporation occurs from the condenser) it reduces to:

£ = ape/2 (19)

This is just half of the correct expression for molecular distillation, E = a pe. The reason for this error is that, in allowing for back-condensation of zinc from the vapor to the evaporating surface, it was tacitly assumed that the vapor was at rest. When the vapor is moving away from the evaporating surface with a net velocity close to the free molecular velocity, its backpressure will be reduced - in the limit, to zero. Otherwise the pressure will be reduced in proportion as the velocity of molecules striking the evaporating surface is reduced from the projected free molecular velocity, c [cm/sec]. The pressure of a gas at rest is p oc c . Thus the forward pressure of a gas moving with a projected mean net velocity v [cm/sec] of zinc atoms away from the evaporating surface is (1+ r^p, and the backward pressure is (1- r)p, where r = v/c. The appropriate changes may be made to the evaporation and condensation terms in equations (12) and (14) above. A first approximation may be made to v, from the uncorrected calculation of E. Then a second approximation can be made to E, a further

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 633

approximation to v, and so on. Evaluations for several sets of conditions (before correction) are shown in Figure 10, together with one curve showing the effect of adding the correction.

0 0.5 10 1.5

A.tVB, cm2 sec/gm »—

Figure 10 - Zinc Elimination Curves - One Showing Effect of Correction Temperature 560° C; L = 30 cm, Various Vacua, Cold Condenser, CC, or Liquid Zinc Condenser, LZ, at

Zinc Melting Point, 420° C, See Text and Reference 2

CONCLUSION

Although it is very important to recognise that the results of a process may be governed by kinetic rather than equilibrium considerations, it is difficult for some metallurgists to accept all the consequences of this. The importance may be more significant in other cases, where it is not recognised, than for the examples given here. It could, for example, enable refining processes to be devised which are unthinkable based on present knowledge.

ACKNOWLEDGEMENTS

Miss Ann-Maree Ahern of Ausmelt Limited, Dandenong, Victoria, rendered considerable assistance in the preparation of the manuscript for printing, and Mr. Jimmy Wong, also of Ausmelt, prepared the figures, modified from previous publications.

634 LEAD-ZINC 2000

REFERENCES

1(a) F.F. Colcord, "Treating Lead Bullion", US Patent No. 1,386,503 (1921).

1(b) G.P. Hülst, "Purifying Antimonial Lead Alloys", US Patent No. 1523,980 (1925).

2. T.R.A. Davey, "The Physical Chemistry of Lead Refining, " Lead-Zinc-Tin '80, J.M. Cigan, T.S. Mackey and T.J. O'Keefe, Eds., The Metallurgical Society of AIME, Warrendale, PA., U.S.A., 1980,477-507.

3. O.J. Kleppa, "Approximate Thermodynamic Data from the Systems Copper-bismuth, Copper-lead and Copper-thallium", J.Am.Chem.Soc, Vol. 74,1952, 6047-6051.

4. G.M. Willis, private communication, 1948.

5. F.D. Richardson, private communication, 1951.

6. F.D. Richardson, "Thermodynamic Principles in the Refining of Metals", Paper 6A in Symposium Physical Chemistry of Metallic Solutions and Intermetallic Compounds. National Physical Laboratory, Teddington, Middlesex, England, 1958, 6A1-23.

7. I.S.R. Clark, "Mechanism and Kinetics of the Decoppering of Lead with Sulfur", (in German) Erzmetall. Vol. 21, 1968, 497-502.

8. G.M. Willis and T.R.A. Davey, Contribution of Discussion to Ref.7, Erzmetall, Vol. 22, 1969, 382-383.

9. I.S.R. Clark, L.A. Baker and A.E. Jenkins, "Decoppering of Lead Alloys with Sulphur", Parts 1 and 2, Trans. Instn. Min, and Met.. Section QMineral Processing and Extractive Metallurgy. Vol. 81, 1972, C195-203 and Vol. 82, 1973, Cl-9.

10. T.R.A. Davey, G. Jensen and E.R. Segnet, "Decoppering of Lead and Hard Lead by Stirring with Sulphur", in Australia - Japan Extractive Metallurgy Symposium. Sydney, Australia, 1980,301-312.

11. E.F.S. Pereira and V.F. Campos, " Kinetics of the Decoppering of Lead", (in Portuguese) Metallurgia, Brasil, Vol. 32, 1976, 85-91.

12. M.P. Smirnov, G.V. Beletskii and G.A. Bibebina, "An Examination of the Macrokinetics of the Fine Decoppering Process for Lead," The Soviet Journal of Non-ferrous Metallurgy. (English Translation of Tsvet. Met.), Vol. 9, 1970, 19-21.

13. T.R.A. Davey and J.V. Happ,"The Decoppering of Lead, Tin and Bismuth by Stirring with Elemental Sulphur", Proc. Aust. Inst. Min, and Metall.. No. 237, 1971,23-31.

14. G.M. Willis, "Refining Metals with Sulphur", Proc. Aust. Inst. Min, and Met.,No. 237, 1971,11-22.

15. W.W. Krysko, "The Decoppering of Lead" (in German), Erzmetall, (a) Vol. 17, 1964, 285-291 and (b) Vol. 20, 1967, 214-217.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 635

16. H.A Simpson and F. Lawson, "Mathematical Model for Fine Decoppering of Lead, " Trans. Instn. Min, and Met.. Section C: Mineral Processing and Extractive Metallurgy. Vol. 94, 1985, C83-C88.

17. W. George, "Metal Stocks and Intermediate Products in Lead Smelting and Their Reduction", Metall und Erz, Vol. 27, 1930, 605-612.

18. V. Tafel, Textbook of Metallurgy (in German), K. Wagenmann, Ed., Hirzel, Leipzig, 1953.

19. G.M. Dice, G.L. Oldright and P.B. Brighton, "Drosses in Lead Smelters", Trans. ATME, Vol. 121, 1936, 127-159.

20. Various Authors, B.H.A.S. Experimental Station Research Reports (Unpublished), 1931-1934.

21. D.A. Cunningham, "The Removal of Copper from Lead by Sulphur Treatment", B.H.A.S. Technical Report R/l 184 (Unpublished), 1956.

22. R. Kleinert, "On the Decoppering of Hard Lead" (in German), Metall und Erz Vol. 45, 1945, 18-20.

23. R.F. Blanks and G.M. Willis, "Equilibria Between Lead, Lead Sulphide and Cuprous Sulphide and the Decoppering of Lead with Suphur", Physical Chemistry of Process Metallurgy.. G.R. St.Pierre, Ed., 1961, 991-1028.

24. J. McClincy and A.H. Larson, "Removal of Copper from Lead with Sulfur", Trans A.I.M.E.. Vol. 245, 1969, 193-196.

25. H.W. Worner, Private Communication, as quoted in Reference 8 above, 1969.

26. D.M. Pollard,"Phases Formed During the Sulphur Treatment of Lead", Trans.I.M.M.. Vol. 88, 1979.C131-132.

27. C. Pin and J.B. Wagner, "The Removal of Copper from Liquid Lead by Lead Sulfide Containing Controlled Atomic Defects", Trans Met. Soc. A.I.M.E, Vol. 227, 1963, 1275-1281.

28. F. Belis, W. Verwimp and J.R. Roos, "Decoppering Lead with Sulfur. Influence of Deficient Sulfides", Metallurgical Transactions, Vol. 7B, 1975, 149-150.

29. B.S. Terry, C.L. Harris and D.G.C. Robertson, "The Decoppering of Liquid Lead Bullion by Elemental Sulphur Additions, Part 1: The Sulphidation of Lead by Elemental Sulphur." Trans. I.M.M.. 102, 1993, C57-C62.

30. S. Terry, C.L. Harris and D.G.C. Robertson, "The Decoppering of Liquid Lead Bullion by Elemental Sulfur Additions. Part 2: A Mechanism and Model for Decoppering," Trans. Instn. Min, and Met.. Section C: Mineral Processing and Extractive Metallurgy. Vol. 102, 1993, C63-C69.

31. W. Gibson, Refinery Supt., Bixby, ASARCO, Private Communication, 1970.

636 LEAD-ZINC 2000

32. W.J. Kroll, "The Vacuum Distillation of Metals", Metal Industry. Vol. 47, 1935, 155.

33. W.T. Isbell, "Vacuum Dezincing in Lead Refining", Trans. A.I.M.E..Vol. 182, 1947,186.

34. T.R.A. Davey, "Pyrometallurgical Methods for Refining Metals", Chapter 12 in Techniques of Materials Preparation and Handling. Part 2 , Interscience Publishers, N.Y., 1968, 549-581.

35. T.R.A. Davey and K.C. Williams, "Continuous Vacuum Dezincing Plant at the B.H.A.S. Pty. Ltd., Port Pirie", Proc. Aust. Inst. Min, and Met., No. 180, 1956, 1-11.

36. T.R.A. Davey, "Distillation Under Moderately High Vacuum, Illustrated by the Vacuum Distillation of Zinc from Lead - Theoretical", Vacuum. Vol. 12,1962, 83-95.

37. T.R.A. Davey, "Vacuum Dezincing of Desilverized Lead Bullion", Trans. A.I.M.E.. Vol. 197,1953,991-997.

38. P.C. Carman, "Molecular Distillation and Sublimation", Trans. Faraday Soc. Vol. 44, 1948, 529.

39. F.D. Richardson, "Principles Underlying Refining Processes", Principles of Extraction and Refining of Metals. Inst. of Metallurgists, London, 1951, 83.

40. N.A. Warner, "Kinetics, of Vacuum Dezincing", Symposium on Advances in Extractive Metallurgy. London, Instn. of Min. and Met., London, 1967, 317-332.

41. V.A. Pazhukhin and E.E. Loukashenko, "Limitations of the Davey Equation"(in Russian) U.S.S.R. J.Phys.Chem.. Vol. 34, 1960, 2224.

42. V.A. Pazhukhin and E.E. Loukashenko, Contribution to discussion on Reference 37, Vacuum. Vol. 14, 1964, 227-228.

43. T.R.A. Davey, Author's Reply to Reference 42, Vacuum, Vol.14, 1964, 228-230.

44. T.R.A. Davey, Discussion to Reference 40. Symposium on Advances in Extractive Metallurgy. London, Instn. of Min. and Met., London, 1967, 417-420.

45. C.F. Harris, J.F. Castle and J. McNish, "Vacuum Dezincing", I.S.P. Conference 1967. Imperial Smelting Corporation, London, 1967, 852-887.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 637

DIRECT ZINC SMELTING IN AN IRON OXYSULFIDE BATH

R.-Q. Li Kvaerner, Metals E&C Division

12657 Alcosta Blvd., San Ramon, CA 94583, USA

J.G. Peacey Noranda Inc.

240 Hymus Boulevard, Pointe Claire, Quebec, Canada H9R 1G5

P. Hancock Noranda Inc., Brunswick Smelting Division

Belledune, New Brunswick, Canada EOB 1G0

ABSTRACT

A pyrometallurgical process for the direct recovery of zinc from zinc concentrates and zinc/iron residues has been proposed and tested extensively by Noranda. The process consists of smelting bone-dry zinc containing materials (sulfide concentrates and secondary zinc/iron materials) in a molten iron oxysulfide bath to volatilize metallic zinc into a SC -free offgas. Sulfur contained in the feed materials is fixed as an iron oxysulfide matte for disposal. Thus, this process not only is capable of treating zinc sulfide concentrates and secondary zinc materials simultaneously, but also eliminates the need of sulfuric acid production. Detailed thermodynamic analysis and experimental test work are described in this paper.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

638 LEAD-ZINC 2000

INTRODUCTION

Bath smelting has the advantages of achieving high intensity and low dust carryover. For zinc recovery, molten oxide slag, sulfide matte and metallic baths have all been proposed as a smelting medium. These reaction media have limited flexibility to treat different materials and also offer limited degrees of freedom to determine the reaction conditions. Slag and carbon-saturated iron baths cannot be used directly to fume zinc from sulfide concentrates. Furthermore, copper baths cannot be used to fume zinc from oxides. The slag composition for zinc fuming must be controlled to a very narrow range to maintain the fluidity of the slag. Utilization of a molten iron bath for zinc fuming requires extremely low oxygen potentials, a fact which makes it difficult to effectively use the combustion heat of carbonaceous materials.

In contrast, the iron oxysulfide bath provides the capability to treat different materials simultaneously and offers a large degree of freedom in determining the required operating conditions. Because of the complete miscibility of FeS and FeO in an iron oxysulfide bath, both sulfide and oxide materials can be readily dissolved in this bath and can be further reacted with other species such as iron, oxygen and carbon. For zinc processing, zinc sulfide concentrates, zinc oxide calcine and secondary zinc materials can all be smelted inside a single bath. Thermodynamically, the Fe-S-0 system can be liquid at temperatures as low as 920°C over a large range of oxygen and sulfur potentials. This provided flexibility to determine the operating conditions such as temperature, CO (g)/C02 (g) ratio of the gas phase, and the solubility of sulfides in the bath. Kinetically, higher thermal conductivity, higher mass transfer rates, higher fluidity and lower viscosity have made the iron oxysulfide bath a unique reaction medium to achieve higher smelting intensities compared with the conventional slag fuming process (1).

In a previous publication (1) the process description and the capability of recovering zinc directly from zinc sulfide concentrates and zinc secondary materials in an iron oxysulfide bath were discussed. It is the goal of this paper to outline the thermodynamic analysis of the patented process and present laboratory test results.

PROCESS CONCEPT AND THERMODYNAMICS

Process Concept

Figure 1 illustrates the main features of the direct zinc smelting process. The bath smelting process can be described as follows:

• Zinc sulfide concentrate is smelted with iron bearing secondary materials and coke in a molten Fe-S-0 bath at ~1350°C.

• Zinc-laden gas is formed in a reducing CO/CO2 atmosphere (CO/CO2 ~ 3-4) and metallic zinc is absorbed in a lead splash condenser.

• The Fe-S-0 melt is tapped from the primary vessel and disposed of as a solid residue in an environmentally acceptable manner.

• Crude zinc from the lead splash condenser is refined separately to produce SHG zinc.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 639

Figure 1 - Schematics of the Noranda Direct Zinc Smelting Flowsheet

In this process, zinc in the feed materials is volatilized into a zinc-laden gas free from sulfur or sulfur dioxide, while sulfur in the feed is fixed in a Fe-S-0 matte by the addition of a proportional amount of iron-containing materials (e.g., EAF dust and zinc plant leach residue). Oxygen associated with the oxides of zinc and/or iron is removed by carbon to form CO/CO2 gas while refractory oxides such as S1O2, CaO, MgO and AI2O3 report to a disposable slag phase. Hence, this process is capable of treating zinc sulfide and zinc secondary materials and their mixtures. The recovery of sulfur in a Fe-S-0 melt prevents the formation of SO2 gas, making condensation of zinc in a conventional lead-splash condenser a relatively straightforward option. In addition, sulfur (or FeS) acts as a flux for iron and other high melting point refractory oxides and therefore eliminates the requirement for a silica or lime flux which otherwise would be needed. The existence of oxygen in the oxysulfide matte permits combustion of petroleum coke with air or oxygen-enriched air inside the Fe-S-0 bath to supply at least part of the heat required by the process. The products of combustion dilute the zinc-laden gas to a desired level, say 10-20%.

Because of the extremely high ZnS activity coefficient in the Fe-S-O matte, a low residual zinc (<lwt%) in the Fe-S-0 melt can be easily achieved allowing the Fe-S-0 matte to be disposed in an environmentally acceptable manner.

The main reaction of the direct zinc smelting process may be written as:

FemaBe + ZnSmat,e = Zn(g) + FeSraaKe (1)

Fema,,e + ZnOmat,e = Zn(g) + FeOma«e (2)

For zinc sulfide concentrate, sulfur is fixed by reduced iron present in the Fe-S-0 matte while zinc is volatilized. For the oxide feed, zinc oxide is reduced by iron present in the Fe-S-0 matte. From the Fe-S phase diagram (1), the sulfur partial pressure in equilibrium with molten stoichiometric FeS matte is in the range of 10"2 to 1.0 atm depending on the temperature, which is obviously too high to avoid the back reaction:

640 LEAD-ZINC 2000

i s 2 ( g ) + Zn(g)=ZnS(s) (3)

Fortunately, unlike the Cu-S system, molten Fe and FeS are completely miscible and the sulfur potential decreases with increasing iron content in the Fe-S matte. Thus, the sulfur potential can be controlled to any desired low value by varying the Fe/S ratio in the Fe-S matte. Basically, the iron required for fixation of the sulfur can be in any form, such as scrap iron, iron ore, steel plant dust and even zinc plant leach residues. When iron oxides are used, reductants such as coal or coke are required to produce reduced iron. The overall reaction of direct zinc smelting is:

Fe203(s)+ ZnS(s)+ C=> Zn(g)+ FeS™«. + CO/C02(g; (4)

The presence of oxygen in the system leads to the formation of a Fe-S-0 ternary matte because of the miscibility of FeO with FeS (2,3). The oxygen content in the Fe-S-0 matte depends on the oxygen potential or CO/CO2 ratio of the system.

Activity Coefficient of ZnS in Fe-S-O Matte

The ZnS in the FeS matte shows a positive deviation from Raoult's law. The activity coefficient of molten zinc sulfide in molten ZnS-FeS matte was estimated as 5, yzns=5, from the optimized FeS-ZnS phase diagram (2). Zinc sulfide in molten metallic iron and iron oxide should also display a positive deviation from Raoult's law. Therefore, a 7zns

>5 is expected in the Fe-S-0 matte. In fact, the ZnS activity coefficient in an Fe-S-O-Pb matte has been reported (2) to be about 10, a value which was confirmed by our experimental work to be described later in this paper. Consequently, when the activity of molten ZnS in the Fe-S-0 matte is maintained to less than 0.05, the expected zinc sulfide solubility in the Fe-S-0 matte should be less than 1 mole %.

Fe-S-O Matte Composition Under Defined Zinc Smelting Conditions

The main components of the present smelting system are Zn-Fe-S-O. If the Zn-Fe-S-0 matte is in equilibrium with a gas phase containing zinc, there are four degrees of freedom of the system. The four independent variables can be defined as:

• Partial pressure of Zn in the gas phase, Pzn

• Activity of ZnS in the matte (representing the targeted solubility of zinc in matte), azns • Temperature of the matte, T • Oxygen potential of the system which is represented by the CO/CO2 ratio.

Thus, the system is uniquely defined given Pz„, azns, T, and the CO/CO2 ratio. For practical purposes, the following conditions are defined:

• Partial pressure of zinc, Pzn 0.1 to 0.2 arm; • Activity of the ZnS in the matte, azns 0.05 to 0.1 • CO/CO2 ratio above 2 • Temperature ~ 1350°C

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 641

An azns=0.05 to 0.1 corresponds to a 1-2 wt% zinc solubility in the matte because the activity coefficient, yzns in the matte is about 10 as mentioned earlier. Because the Zn solubility in a Fe-S-0 matte is very low under these conditions, the matte can be considered a Fe-S-0 ternary system whose composition is defined at a fixed temperature, oxygen potential and sulfur potential. Based on the reported relationship between the Fe-S-0 matte composition, temperature, oxygen and sulfur potentials (2), Fe-S-0 matte compositions under the described smelting conditions have been estimated see Table I. At 1350°C and a CO/C02 ratio between 3 and 4, azns=005 (or less than 1% Zn solubility in matte) requires the composition of matte of 6-7% O and Fe/S 2.7-2.9 (weight basis). The proposed matte smelting composition is shown in Figure 2 (6).

Wright h r c M fey««n

Figure 2 - Fe-S-O Liquidus Projection in Weight Percent (The Marked Area is the Targeted Smelting Matte Composition)

642 LEAD-ZINC 2000

Table I - Fe-S-0 Matte Compositions (in wt%) Under PZn= 0.1 atm and azns= 0.05

CO/C02 ratio

2

3

4

6

8

T,°C

P(S2) P(o2) 0% Fe% S%

Fe/S

P(02) 0% Fe% S%

Fe/S

P(o2) 0% Fe% S%

Fe/S P(02) 0% Fe% S%

Fe/S

P(02) 0% Fe% S%

Fe/S

1300 1.36e-05 5.03e-ll

9.50 69.70 20.80 3.35

2.23e-ll 8.20

69.80 22.00 3.17

1.26e-ll 6.40 69.60 24.00 2.90

5.58e-12 5.00

69.50 25.50 2.73

3.14e-12 3.50

69.70 26.80 2.60

1350 6.09e-05 1.89e-10

9.30 69.50 21.20 3.28

8.38e-ll 7.40

69.00 23.60 2.92

4.71e-ll 5.90

68.90 25.20 2.73

2.10e-ll 4.20 68.80 27.00 2.55

1.18e-ll 3.00 68.80 28.2 2.44

1400 2.49e-04 6.53e-10

9.00 69.00 22.00 3.14

2.90e-10 6.60 68.40 25.00 2.74

1.63e-10 5.00

67.80 27.20 2.49

7.26e-ll 3.50

67.50 29.00 2.33

4.08e-ll 2.40 67.70 29.90 2.26

1450 9.33e-04 2.10e-09

8.00 68.72 23.28 2.95

9.35e-10 5.90 67.85 26.25 2.58

5.26e-10 4.20 67.36 28.44 2.37

2.34e-10 3.00

67.00 30.00 2.23

1.32e-10 2.10 66.96 30.94 2.16

Effect of Temperature on the Solubility of ZnS in Fe-S-O Matte

Assuming that zinc lost to the Fe-S-0 matte is mainly due to sulfidic dissolution, the effect of temperature on the solubility of zinc in a given Fe-S-0 matte can be evaluated from the thermodynamic relationship described by reaction (1). Thus, for every 50°C increase in temperature, a 40% reduction in ZnS solubility under a given matte composition can be expected.

Formation of an Iron Silicate Slag in Equilibrium with Fe-S-O Matte

Although the proposed process is primarily carried out in a Fe-S-O matte phase, formation of slag cannot be avoided if the feed materials contain significant amounts of S1O2 and other refractory oxides. This can be seen from the FeS-FeO-Si02 phase diagram (2) shown in Figure 3. When the FeO-FeS matte is in the FeS rich area, say > 60% FeS, the solubility of Si02 in the FeS-FeO matte is limited to less than 2 wt%. The FeS-FeO-Si02 system has a miscibility gap which separates a FeS-FeO rich matte phase from a slag phase. The critical point is at 43 wt% FeS, 50 wt% FeO and 7 wt% S1O2. Experiments indicated that an increase in temperature does not result in a change in chemical composition of the two liquids. It was also found that the addition of G12S reduces the mutual solubility between matte and slag, resulting in a more complete separation of the two liquids (10).

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 643

Figure 3 - Portion of the FeS-FeO-SiC>2 Phase Diagram

As shown in Figure 3, a matte phase having a composition of 72.9% FeS, 27.42% FeO and 0.16% S1O2 is in equilibrium with a slag phase having a composition of 17.90% FeS, 54.82% FeO and 27.28% Si02 with a melting point of about 1150°C. Point "B" having a composition of 72.4% FeS, 27.4% FeO and 0.2% S1O2 is in equilibrium with point "A" which is in equilibrium with solid S1O2 and FeSi04. Point Ei in Figure 3 is the binary eutectic of S1O2 and Fe2Si04 in the FeO-Si02 system (2). If a matte phase is richer in FeS than matte "B" in Figure 3, formation of a low melting point fayalite slag becomes impossible. Thus, to form a fluid fayalite slag phase, the oxygen potential in the direct zinc smelting should not be allowed to drop to very low values which can produce a matte richer in FeS than matte "B" of Figure 3.

Under direct zinc smelting conditions, if the CO/CO2 is controlled above 3 at temperatures above 1300°C, the Fe-S-0 matte as shown in Table 1 can be approximately considered as a FeO-FeS binary system. Consequently, the composition of slag can be estimated from Figure 3 and is presented in Table II.

644 LEAD-ZINC 2000

Table II - Estimated Slag Compositions in Equilibrium with FeS-FeO Matte Under Direct Zinc Smelting Conditions, PZn=0.1 atm and 1350°C

1 2 3

4*

azns

0.05 0.05 0.1

0.1

CO/

co2 3 4 3

4

FeS 64.9 69

69.8

FeO 33.3 27

29.7

Fe

1.8 4

1.5 aFeO=0.37 at 1350°C 74.8

»FeO=

22.5 2.7

0.29 at 1350°C

Matte, wt% Fe 69

68.9 68.4

67.8

O 7.4 5.9 6.6

5

S 23.6 25.2 25

27.2

Si02

~ 2 < 1 < 1

< 1

FeS 20 19 18

17

Slag, wt% FeO 55 52 53

51

Si02

24 29 31

32

* This slag may have reached silica saturation since the oxygen potential is lower and the FeS/FeO ratio in matte is higher than the matte "B" in Figure 3.

Effect of Refractory Oxides on the Slag Composition Under Direct Zinc Smelting Conditions

Yazawa's experimental study revealed that the addition of CaO and AI2O3 to the slag had negligible influence on the composition of the matte, but reduced the solubility of the matte in the slag (10). Since an addition of 6% CaO or AI2O3 into a slag which was in equilibrium with a 71% FeS-27% FeO matte reduced the sulfur content of the slag from 6.5 wt% to ~3 wt%, it is expected that the sulfur content in an iron slag containing CaO and AI2O3 should be lower. It is therefore feasible to analyze the effects of various refractory oxides on the compositions of the sulfur-containing slag based on the sulfur capacity concept which is reliable for slags containing low amounts of sulfur, say less than 3 wt%.

Pelton et al. (2) proposed a mathematical model to predict sulfide capacities of multi-component slags from thermodynamic activities of the oxide components, with no adjustable parameters. They were able to combine their model with their slag solution database that predicts activities of the components of the slag required to estimate sulfide capacity. Thus, using F*A*C*T* (2) and its powerful slag database, it is possible to estimate the sulfur-containing slag compositions under direct zinc smelting conditions.

In industrial operations, refractory oxides contained in the feed raw materials, such as coke, zinc sulfide concentrate, and oxidic zinc secondary materials, consist mainly of S1O2, CaO, AI2O3 and MgO. Thus, the slag in the direct zinc smelting process can be defined as a FeO-Si02-CaO-Al203-MgO-ZnO-S system. Assuming that this slag is in equilibrium with a FeS-FeO-Fe matte and zinc laden gas, the system has 3 phases and 8 elements, and therefore, 7 degrees of freedom (f= 8 + 2-3 = 7). The seven independent variables can be defined as:

• Temperature, say 1350 GC • Zinc partial pressure in gas, say Pzn = 0.1 atm, • Zinc solubility in matte, say azns = 0.1 • Oxygen potential, say CO/CO2 = 3 or 4 • CaO/Si02 ratio • AI2O3/S1O2 ratio • MgO/Si02 ratio

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James andA.H.-J. Siegmund 645

Since it is certain that the solubility of refractory oxides in the FeS-FeO-Fe matte is very low under direct zinc smelting conditions, the activities of FeS, FeO and Fe in the matte are negligibly influenced by the slag chemistry; i.e., the activity of FeO, FeS and Fe in the system is determined mainly by the matte-gas two phase equilibrium. Therefore, in the absence of a thermodynamic database for the matte, the slag composition in equilibrium with an Fe-S-O matte can be determined by examining the equilibrium between the zinc-laden gas and the slag by using F*A*C*T* as long as the FeO activity determined by the matte-gas equilibrium is accounted for the activities of FeO(l) for matte No.3 and No.4 in Table Π have been estimated as 0.37 and 0.29 at 1350°C from measured thermodynamic data for the Fe-S-O ternary system (9).

Fe-Zn-Q-S-Si02-CaO Slag

Figure 4 shows the estimated slag compositions at 1350°C, PZn = 0.1 arm, a^s = 0.1 and CO/CO2 = 3. The predicted matte composition is presented in Table II. From Figure 4, it is clear that the concentrations of FeO, ZnO and FeS decrease with increasing CaO/Si02 ratio. The ZnO content in the slag decreases from 3.4 mole% at CaO/SiO2=0.4 to less than 1 mole% at CaO/Si02=l .2. Interestingly, the ZnS content in slag was almost negligible.

1360C.P(Zn>*.1atn,a(Zn8)-ai,CQCO2-3»de(FeOH).37 MgO8K)2-0,AI2O3BICC-0

M U U M U U U U U 1 1.1 12 14

CaO/Si02 Molar Ratio

Figure 4 - Si02 - Containing Slag in Equilibrium With Matte and Zinc Laden Gas

Fe-Zn-Q-S-Si02-CaO-AbOj Slag

When coke is used, it is unavoidable to have a certain amount of AI2O3 in the slag. Figure 5 shows the estimated slag compositions at 1350°C, Pa = O.latm, a&s = 0.1 and CO/CO2 = 3 and AI2O3/S1O2 = 0.2. Similarly to CaO, the addition of AI2O3 to the slag reduces the FeO and ZnO content in the slag, but has only minor effects on the S1O2 and CaO contents of the slag.

646 LEAD-ZINC 2000

1360C,P<ZnH).1»«nr>,a(ZnS)-0.1,CCVCO2«3«xle(FeOH).37 MgO/S02-0,AI203/SI02-0.2

MS

I

0.1 0.2 0.3 0Λ 0Λ 0.6 0.7 0Λ 0.» 1 1.1 1.2 1.3;

CaO/Si02 Molar Ratio

AI203

CaO

CaS

Fe203

FeO -»-FeS

Si02 -«-ZnO

Figure 5 - AI2O3 Containing Slag in Equilibrium with Matte and Zinc Laden Gas

Fe-Zn-Q-S-Si02-CaO-MgO Slag

Although MgO is chemically similar to CaO, addition of MgO to an iron fayalite slag always has a negative effect lowering the melting point of the slag. It is almost impossible to form a high fluidity FeO-Si02-MgO slag. CaO has to be added when MgO is present.

Figure 6 shows the estimated slag compositions at 1350°C, Pzn = 0.1 atm, az„s = 0.1 and CO/CO2 = 3 at MgO/Si02= 0.2. The addition of MgO to the slag reduces its FeO and S1O2 contents significantly, but has only a minor effect on the concentration of ZnO, contrary to the observation in the presence of CaO and AI2O3.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 647

§

1380C, PtaiHklakn, a(ZnS)0.1. COXXS «8 wd «<FeO)"0.37 MpO/aOMA ΛΒ0ΜΙΟ&0

U U M U U U U M U 1 1.1 12 14

CaO/SiOi Molar Ratio

Figure 6 - MgO Containing Slag in Equilibrium With Matte and Zinc Laden Gas

Fe-Zn-0-S-Si07-CaO-MgO-Al,03Slag

Figure 7 shows the estimated slag compositions at 1350°C, Pzn = 0.1 atm, az„s = 0.1,

MgO/Si02 = 0.2 and Al203/Si02 = 0.2 at CO/C02 = 3.

1350C, P(Zn)-0.1atm. a(ZnS)-0.1, CO/C02 -3 and a(FeO)=0.37 MgO/SK>2-0.2, AI2O3/SIO2-0.2

1—r

β.1 01 0Λ M U M 0.T OJ 04 1 1.1 1J! 1J| ZrtO

CaO/SiQi Molar Ratio

Figure 7-The S1O2-AI2O3-MgO Containing Slag

648 LEAD-ZINC 2000

EXPERIMENTAL

Laboratory scale tests were designed and carried out to demonstrate the feasibility of the proposed process and to verify the thermodynamic predictions, namely:

• Zinc solubility in the matte as a function of the process parameters • Possibility of fixing the sulfur present in the raw material as a Fe-S-O matte • Feasibility of using iron oxide as the sulfur-fixing agent • Possibility of smelting oxidic zinc materials in the proposed process • Feasibility of combusting carbon with air or oxygen enriched air inside the Fe-S-O matte • Purity of the condensed metallic zinc product • Reaction kinetics.

The experimental system consists of a graphite crucible of 12-cm diameter placed inside an induction furnace. Five to seven kg of solid mixture consisting of iron ore, metallic iron and pyrrhotite were charged into the crucible and melted under a nitrogen gas atmosphere to give a pre-designed Fe-S-O matte composition. Once the matte reached the target temperature of 1250-1400°C, a solid mixture consisting of iron ore/metallic iron, zinc concentrate, zinc oxide and coke was fed to the Fe-S-O bath through a submerged lance made of a 1-cm diameter mullite tube. The temperature of the bath ranged between 1250 to 1400°C. The feed rate of the solid mixture was controlled with a screw feeder. Feed rates were in the range of 20 g/min to 120 g/min. In most tests, feed rates were around 50 g/min. The reaction took place inside the matte to volatilize zinc. In general, it took only a few minutes to attain stable zinc fuming (by observation), which indicates a low zinc solubility in the Fe-S-O matte.

During the tests, samples of the volatilized zinc were taken by condensing it on a water-cooled steel pipe. The CO/CO2 ratio and the SO2 composition were monitored to make sure there was no SO2 in the gas phase. Samples of Fe-S-O matte were taken at intervals using a cold steel rod. Both condensed zinc samples and matte samples were chemically assayed. More than ten successful tests were carried out to examine the following parameters:

• Temperature • Sulfur potential (Fe/S ratio in the matte) • Partial pressure ofZn in the gas phase • Oxygen potential (or CO/CO2 ratio) • Metallic Fe feed vs Fe203 feed • N2 injection vs air injection. • Zinc sulflde concentrate feed vs zinc oxide feed

The compositions of the raw materials were also analyzed and are listed in the Table ΙΠ.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 649

Table III - Chemical Compositions of the Raw Materials (in wt%)

Chemicals Zn Fe S Pb O C Si02+CaO+Al203+MgO 33.8 2 2.0

30 2.0 Zinc Concentrate Iron ore Metallic Iron Zinc Oxide Charcoal Pyrrhotite

51.3 10.5 68 100 80

71

20 90

29

Table IV summarizes the experimental conditions and the results for the eleven tests. During the campaign, almost one hundred samples were taken to assay the matte and the condensed zinc samples. Each of the samples was chemically analyzed to determine its As, Cd, Fe, Pb, S and Zn contents, whereas the S1O2 and O contents were analyzed only for some of the samples. The temperature and compositions of the matte and condensed zinc samples reported in Table IV are average values. Because the total assay for both matte and condensed zinc added to more or less than 100%, the contents of the matrix elements (iron for matte and zinc for condensed zinc sample) were calculated by subtracting the minor element chemical analysis content from the 100% total and these are shown in Table IV as "balanced iron' and "balanced zinc'. The Fe/S ratios for the matte samples reported in this table were calculated on the basis of the "balanced iron" and analyzed sulfur wherever the "balanced iron" was available.

The zinc solubility in the matte is affected by the partial pressure of zinc and the temperature under a given matte composition. To compare the zinc solubility from one test to another, this must be normalized to a standard zinc partial pressure, say 0.1 atm, and a temperature. The normalized zinc solubility at 0.1 atm zinc partial pressure was calculated by dividing the measured zinc solubility by the corresponding zinc partial pressure and by 10, (Zn wt%) / (10 PZn). As discussed in the above section, an increase of 50°C in temperature causes a 40-60% decrease in zinc solubility for a given matte composition. Thus, the normalized zinc solubility at Pzn= 0.1 atm and experimental temperature was converted to that at 1350°C by considering the effect of temperature. These zinc solubility data is summarized in Table IV as normalized zinc solubility at Pz„ = 0.1atmandl350oC.

The distribution ratios of the minor elements, As, Cd and Pb, were calculated by dividing the average content of these elements in the condensed zinc phase by that in the matte phase.

To compare the elemental distribution ratio between the condensed zinc and the matte from the different tests, these were also normalized to Pzn = 0.1 atm for all tests by assuming: (normalized distribution ratio at 0.1 atm zinc partial pressure) = (measured distribution ratio * Pa, / 0.1) as listed in Table IV. In general, the scatter of the estimated distribution ratio is large. This is not surprising if we recognize the many sources of variation: different temperatures, different degrees of splashing caused by different feed rates and gas flow rates, analytical errors and even different matte compositions.

Nevertheless, it is clear that Cd and Pb are almost completely volatilized because the average distribution ratios of Pb and Cd are 250 and 23, respectively. It is also certain that the majority of As stays in the matte because it has a distribution ratio of less than one. In most of the tests, the Fe/S ratios of the condensed zinc samples are very close to those of the matte, indicating that the sulfur in the condensed zinc samples comes mainly from splashing.

Tab

le IV

- E

xper

imen

tal C

ondi

tions

and

Res

ults

O

N

o

Tes

t No,

E

stim

ated

CO

/C0 2

rat

io

Pyrr

hotit

e In

itial

Ir

on o

xide

m

atte

fr

on

g

Tot

al

1 N.A.

5317

843

868

7028

5317

843

868

7028

6718

0 282

7000

4 8 5758

0 1242

7000

N.A.

3798

582

620

5000

2 3016

1446

586

5048

4 4721.5

1398

927.5

7047

3798

582

620

5000

5317

843

868

7028

10

3868

61

3 63

2 51

13

Fe

S O

%w

t 74

.22

21.9

4 3.

6

74.2

2 21

.94

3.6

72.1

7 27

.83

0

76.1

5 23

.85

0

74.2

5 22

.03

3.49

73.5

1 17

.33

8.59

74.2

2 19

.43

5.95

74.2

5 22

.03

3.49

74.2

2 21

.94

3.6

74.2

2 21

.94

3.6

Feed

rate

Fe

ed

mat

eria

ls

g/m

in

50

19.5

51

.1

47.8

62

.2

52.5

52

.5

126

42.9

50

Z

inc

cone

Ir

on o

re

Met

allic

Iron

C

harc

oal

Zin

c ox

ide

Wt%

55.8

8.

7 35

.5

0 0

44

47.9

0 8.1 0

51.4

39

.9

0 8.7 0

44.7

45

.4

0 9.9 0

56

8.4

35.6

0 0

38.6

55

.2

0 6.2 0

40.9

51

.8

0 7.3 0

44.7

47

.1

0 8.2 0

42.3

44

.5

0 13.2

0

0 0 0 19

.23

80.7

7 N

2

Air

C

alcu

late

d pa

rtia

l pre

ssur

e of

Zn

Tem

pera

ture

of b

ath

NL

/min

Atm

°C

10

0 0.

346

1345

20

0 0.

066

1350

.4

8 0 0.

237

1324

.3

8 0 0.

197

1324

.8

8 0 0.

452

1327

12

0 0.

355

1324

8 0 0.

23

1329

8 0 0.

228

1325

14

0 0.

266

1306

0 6.4

0.18

4

1311

0 12

0.30

8

1287

av

erag

e m

atte

as

say

duri

ng

test

As

Cd

Fe

Pb

S Zn

Si0

2

O

Bal

ance

d Fe

ppm

Wt%

397.

2 5

75.4

2 0.

11

26.5

7 0.

83

<0.5

2.

69

69.7

6

390.

67

4.87

77

.5

0.05

27

.46

0.29

<0

.5

2.91

69

.25

596.

6 20

.46

62.2

0.

21

28.1

4 5.

06

<0.5

4.

07

62.4

6

652.

75

20.2

1 69

.08

0.14

23

.95

1.32

<0

.5

5.45

69

.08

557.

8 9.

69

73.7

8 0.

12

22.0

8 1.

17

<0.5

3.

51

73.1

1

608.

5 28

.21

73.7

3 0.

17

22.1

8 0.

87

<0.5

3.

51

73.2

7

697.

9 10

.21

70.4

6 0.

12

19.0

2 1.

41

<0.5

8.

61

70.7

6

489.

71

7.21

68

.63

0.11

19

.41

1.01

<0

.5

6.51

72

.92

533.

8 52

.81

65.3

5 0.

16

21.4

8 6.

29

<0.5

9.

09

62.9

2

411.

87

6.16

75

.69

0.11

28

.04

0.9

<0.5

3.

84

67.0

6

4.93

70

.85

0.00

49

25.4

6 2.

72

<0.5

68.2

1 Fe

/S ra

tio

Nor

mal

ized

wt%

Zn

at P

zn=0

.1 a

tm

2.64

0.

24

2.52

0.

44

2.22

2.

14

2.89

0.

67

3.31

0.

26

3.31

0.

25

3.73

0.

62

3.77

0.

44

2.94

2.

37

2.4

0.49

2.

69

0.88

LEAD-ZINC 2000

Tes

t No.

N

orm

aliz

ed w

t%Z

n at

Pzn

=0.1

atm

an

d 13

50°C

Tab

le I

V -

Exp

erim

enta

l C

ondi

tions

and

Res

ults

(C

ontin

ued)

12

3 4

5 6

0.23

0.

44

1.63

0.

51

0.2

0.19

0.

49

7 0.

34

8 1.54

9

0.33

10

0.

5

Ave

rage

as

say

of

zinc

sa

mpl

e

As

Cd

ppm

25

.31

140.

3 98

.67

1008

.8

82.2

3 25

66.7

54

.42

1516

.7

233.

5 13

52.4

47

.07

1009

.6

555.

8 80

8.2

25.2

7 87

7.49

Fe

Pb

S Zn O

wt%

3.16

2.

55

1.46

10

0.6

20.8

9 0.

95

7.27

70

.78

6.71

5.59

1.

51

2.8

5.39

8.25

1.

33

2.99

87

.43

14.4

7 1.

99

4.58

71

.67

6.15

7.32

1.

41

2.29

78

.13

3.67

6.2

1.39

4.

68

65.7

2 3.

67

1.31

1.

27

0.57

92

.86

4.97

Fe

/S

2.17

2.

87

1.89

2.

69

3.16

3.

31

2.16

2.

57

Bal

ance

d Z

n 92

.81

64.0

7 89

.84

87.2

7 72

.66

85.2

1 91

.79

Dis

trib

utio

n ra

tio

As

0.06

0.

25

0.14

0.

08

Of

elem

ents

bet

wee

n C

d 28

.06

207.

07

125.

45

75.0

4 co

nden

sed

zinc

and

p

D

23.0

3 19

.7

7.11

9.

57

mat

te

0.33

0.

1 1.

04

0.06

13

2.45

13

9.97

15

.31

142.

34

16.8

6 13

.19

8.56

11

.1

Nor

mal

ized

A

vera

ge

Stan

dard

E

lem

ent

Nor

mal

ized

Dis

trib

utio

n ra

tio a

t Pz„

= 0

.1 a

tm

dist

ribu

tion

ratio

be

twee

n zi

nc a

nd

mat

te

devi

atio

n 0.

29

244.

69

23

0.22

74

.28

8.74

As

Cd

Pb

0.17

13

7 13

.0

0.33

29

7 16

.8

0.16

14

7 18

.8

0.77

30

4 38

.7

0.22

31

9 30

.1

0.11

26

1 20

.4

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

652 LEAD-ZINC 2000

Test-1, Metallic Iron Feed - Effect of Temperature

This test was designed to examine the technical feasibility of sulfur fixation using metallic iron and to investigate the influence of temperature on zinc solubility in the Fe-S-0 matte. A 7,028g solid mixture consisting of 5,317 g pyrrhotite, 868 g metallic iron and 843 g iron ore was charged into a graphite crucible and melted. This would give an Fe-S-0 composition of 74.22 wt%Fe, 21.94 wt%S and 3.6 wt%0. When the temperature reached 1250°C, solid feeds with a composition of 55.8% zinc concentrate, 8.7% iron ore and 35.5% metallic iron were injected into the molten bath at a feed rate of 50 g/min by 10 NL/min of nitrogen carrier gas. The estimated zinc partial pressure was 0.346 atm when the system reached equilibrium. The purpose of adding iron ore in this test was to keep some of the oxygen (-3.6%) in the matte.

Table IV shows the results of the tests in detail. Test-1 was done in two different temperature ranges, 1250-1300°C for the first 30 minutes and 1400-1430°C for the second half of the test. The zinc content in the Fe-S-0 matte increased as the test proceeded, until at 30 minutes, it reached its maximum, 1.37 wt%, which is considered the saturation value. In the second half of the test, temperature was increased to 1410-1430°C and the zinc solubility decreased to -0.5 wt% which basically is the same as that present in the feed. It is worth noting that the zinc concentration in the matte before the test should be very small because there was no zinc added to the system at the beginning. The reported value might be due to a chemical analysis error. If this were true, the zinc recovery into the gas phase was almost 100% for the second half of the test at a temperature >1400°C. Figure 8 shows the changes in zinc solubility during the test.

1.4

- 1.2

1

0.8

0.6

- 0.4

0.2

u 00

a *-? a N

Figure 8 - Zn Solubility in the Fe-S-0 Matte During Test-1

The present test was done at Pzn= 0.34 atm. It is expected that zinc solubility in the Fe-S-0 matte would be even lower if the zinc partial pressure was controlled at 0.1 atm.

Test 1 successfully demonstrated the feasibility of the process to use metallic iron to fix sulfur. This test indicates that the activity coefficient of ZnS in matte is very large because the zinc solubility is about 1 wt% even at Pzn = 0.34 atm and 1350°C, whereas the normalized zinc solubility in the matte is only 0.24 wt%.

1,500

1,200 20 40 60

Time (Minutes)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 653

Test-2, Iron Ore Feed, at Lower Feed Rates and Higher Gas Injection Rates

This test was designed to examine the feasibility of sulfur fixation using iron ore instead of metallic iron and to find the equilibrium zinc solubility in Fe-S-O matte. Therefore, this test was carried out under a high gas flow rate and with a low solid material feed rate. Initial matte preparation was done exactly as in the previous test. The molten bath temperature was controlled to around 1350 °C. A solid mixture with 44% zinc concentrate, 48% iron ore and 8% charcoal was injected into the bath at a feed rate of 19.5 g/min with 20 NL/min nitrogen carrier gas to give Pz„ = 0.066 atm.

There is no apparent variation of the zinc content before, during and after the test, indicating that the zinc content in the matte is negligible. The condensed sample has a higher Fe and S content, but a lower metallic zinc content than in Test-1. This may be attributable to the carryover of solid feed materials, such as iron ore because the gas flow rate is higher than in Test-1. This also explains the higher Fe/S ratio of the condensed sample than that of the matte. Furthermore, it is obvious that sulfur in the condensed zinc sample results from splashing and solid carryover rather than from sulfur volatilization. Otherwise, the Fe/S ratio in the condensed zinc sample should be lower than that in the matte.

It is also interesting to note that the normalized zinc content in the matte is around 0.4%wt, which is somewhat higher than that in Test-1; this is due to the lower Fe/S ratio in the matte as explained later.

Test-3, Effect of the Fe/S Ratio

This test was carried out to investigate the effect of the Fe/S ratio on the solubility of zinc in the Fe-S matte. The initial matte consisted of 6,718 g pyrrohite and 282 g iron which corresponds to 72 %Fe and 28 %S in the matte. The molten bath temperature was controlled to around 1325°C. The feed material consisted of 51.4% zinc concentrate, 39.9% iron ore and 8.7% charcoal. The feed mixture was injected into the bath at 51.1 g/min with 8 NL/min nitrogen. The zinc partial pressure was calculated to be 0.237 atm.

The solubility of zinc is much higher than in the previous tests because of the lower Fe/S ratio. This agrees with the theoretical analysis. Zinc solubility is proportional to the square root of the sulfur potential, whereas the sulfur potential is controlled by the Fe/S ratio: the higher the Fe/S ratio, the lower the sulfur potential. Nevertheless, the "normalized solubility of zinc" in matte at Pzn = 0.1 atm is less than ~2 %wt even at a Fe/S ratio of 2.22.

Test-4, Iron Ore, Effect of Nitrogen Flow and the Fe/S Ratio

Test-4 was similar to test-3 except that it was done at a higher Fe/S ratio. The initial matte consisted of 5,758 g pyrrhotite and 1,242 g metallic iron and had a composition of 76 %Fe and 24 % S. The matte composition does not add to 100% from its Fe, S and Zn contents. Therefore, its oxygen content was calculated and was found to be higher than in the previous tests. Clearly, the zinc solubility in the matte is lower than in test-3 because of the higher Fe/S ratio. However, it is also higher than that in tests 1 and 2 even though the Fe/S ratio in test-4 was higher than in test-1 and this may be linked to the higher oxygen content in comparison with that of tests 1 and 2.

654 LEAD-ZINC 2000

Test-5, Effect of the Nitrogen Flow Rate

This test was designed to examine the effect of the partial pressure of Zn on the solubility of zinc in the matte. Nitrogen flow rates were changed from 8 NL/min to 12 NL/min under otherwise identical experimental conditions. It was found that the solubility of zinc in the matte decreased with increasing nitrogen flow rate. However, the normalized zinc solubility in the matte remained constant (within experimental error). This is in agreement with theory, indicating that the solubility of zinc in the matte is in equilibrium with the zinc-laden gas.

Tests-6 and 7, Effects of the Oxygen Content of the Matte

Tests 2-5 were carried out at an oxygen content range of- 4% O, corresponding to CO/CO2 ~ 8. Tests 6 and 7 were done at a higher oxygen potential of the matte to explore the influence of the oxygen potential on the zinc solubility. Test-6 was carried out at CO/CO2 = 2. As shown in Table IV, the initial matte consisted of 73.51% Fe, 17.33% S and 8.6% O. The average matte composition was Fe/S = 3.73 and 8.61% O. As expected the normalized average zinc solubility in the matte was 0.49%, which is higher than that in test-5.

Test-7 was done at CO/C02=4. The average matte composition is Fe/S=3.77 and 6.51% O. The zinc solubility is lower than that in Test-6 because of the lower oxygen content of the matte or the lower oxygen potential.

Under a given Fe/S ratio, the oxygen potential has a twofold influence on the solubility of zinc in the matte: 1) the higher the oxygen potential (or the higher the oxygen content in the matte), the higher the sulfur potential and the higher the ZnS solubility; 2) the higher the oxygen potential, the higher the ZnO solubility in the matte. Thermodynamically, the main zinc loss in the matte is as ZnS(l) rather than as ZnO(l) and therefore an increase of the oxygen potential should be considered as an increase of the sulfur potential under a fixed Fe/S ratio.

Chemical analysis of the condensed zinc samples showed similar assays as those reported in the above tests, indicating that the effect of the CO/CO2 ratio on zinc condensation is minimal.

Figure 9 summarizes the influence of the oxygen content on the zinc solubility for Tests-5, 6 and 7 having a Fe/S ratio between 3.3-3.7. It is seen that the solubility of zinc is directly proportional to the oxygen content of the matte.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 655

0.55

S 0.5

S 0.45

S °-4

£ 0-35 ■S 0.3 | 0.25

ä °·2

0.15

O wt% in Fe-S-O Matte P(Zn) = 0.1 atm, 1350C, Fe/S = 3.3-3.7

Figure 9 - Effect of Oxygen on the Zinc Solubility in the Matte

From the oxygen contents reported in Table IV for Tests 6 and 7, an excellent agreement between the thermodynamically predicted oxygen contents of the matte was found. This was obtained from the initial matte composition (Table IV), and the measured data, shown as the average matte composition for different CO/CO2 ratios. Consequently, the theoretical analysis was successfully verified. For example, at CO/C02=2, Fe/S=3.3 and 1350°C, thermodynamics predicts ~9 wt% O and azns=0-05 in comparison with the measured 8.61 wt% O and 0.49 wt% Zn in Test-6. Furthermore, the experimental data confirmed that the activity coefficient of ZnS(l) in the matte is about 10.

Test-8, Effect of the Feed Rate

This test was performed at a very high feed of rate, 126 g/min. Although a high solids carryover and a poor contact between the solid particles and the molten bath were experienced because of the high feed rate and top feeding rather than submerged feeding, the normalized zinc solubility in the matte is still less than 2 wt% Zn, indicating better process kinetics in comparison with slag fuming.

Test-9, Iron Ore Feed, Effect of Air Injection

In tests 1-8, nitrogen gas was used as the carrier gas for injecting zinc concentrate and for diluting the zinc-laden gas. Test-9 was carried out to demonstrate the feasibility of using air instead of nitrogen. As shown in Table 4, air works as well as nitrogen. Not only was a very low solubility of zinc achieved, but also the condensed zinc sample has a very high purity if the splashed matte is excluded. Again, the Fe/S ratio of the condensed zinc sample is almost the same as that of the matte. The SO2 concentration in the zinc-laden gas was monitored during the test and was less than 50 ppm.

656 LEAD-ZINC 2000

Test -10, Zinc Oxide Feed

Test-10 was done to demonstrate the technical feasibility of volatilizing zinc by feeding zinc oxide into a pre-melted Fe-S-0 matte. A solid mixture consisting of 81% ZnO and 19% charcoal was injected into the bath at a feed rate of-50 g/min by air. The zinc solubility of the matte was about 0.9 wt% Zn at Pzn=0.1 atm and 1350°C. This test successfully demonstrated the feasibility of treating zinc oxide materials in the Fe-S-O matte by using carbon as a reductant and oxygen as the diluting gas.

The Activity Coefficient of ZnS

Table V summarizes the experimental results of the tests with low oxygen content in the matte; i.e., tests 1-5,9 and 10. The matte composition and normalized zinc solubility at Pzn=0.1 atm and 1350°C as reported Table ΠΙ are repeated here. The zinc solubility has been converted into mole fractions to calculate the activity coefficient. The sulfur potentials at 1450 and 1200°C are taken from the reported activity diagrams for Fe-S-0 matte (9), from which the sulfur potentials at 1350°C are estimated. The activities of ZnS in the matte were calculated from: azns

=

K*Pzn*(Ps2)°5, where K is the equilibrium constant for reaction Zn(g)+'/2S2(g)=ZnS(l) and was estimated to be 64 at 1350°C. The activity coefficient of ZnS in the matte was estimated (Table IV). Although the estimated activity coefficient of ZnS varied from 7 to 22, it has an average value of 12.6 and is very close to the reported value, 10 (8). Considering the many possible sources of error in the tests, it is not surprising that there is a large scatter in the estimated ZnS activity coefficient. Nevertheless, the experimental data were in close agreement with the theoretical predictions.

Table V - Analysis of the Experimental Data for Tests 1-5, 9 and 10

Test Measured 0%wt

Matte Fe%wt :omposition pe /s

Zn%wt* Zn mole fraction

log(PS2) at 1450 °C log(PS2) at 1200 °C log(PS2) at 1350 °C ZnS activity, aZnS

Activity coefficient of ZnS in matte

Average activity coefficient of ZnS

1 2.69 69.76 2.64 0.23

0.0015

-4.30 -5.60 -4.77 0.026 17.27

2 2.91 69.25 2.52 0.44

0.0029

-4.20 -5.50 -4.67 0.030 10.06

3 4.07 62.46 2.22 1.63

0.0103

-3.00 -4.30 -3.47 0.118 11.42

4 5.45

69.08 2.89 0.51

0.0033

-4.40 -5.70 -4.87 0.023 7.04

12.64

5 3.51 73.11 3.31 0.20

0.0014

-4.70 -6.00 -5.17 0.017 12.08

9 3.84 67.06 2.40 0.33

0.0022

-3.75 -5.05 -4.22 0.050 22.67

10 3.60 68.21 2.69 0.50

0.0033

-4.30 -5.60 -4.77 0.026 7.92

*. Normalized zinc solubility at Pzn=0.1 atm and 1350°C.

Figure 10 shows the zinc solubility obtained experimentally as a function of the Fe/S ratio. The zinc solubility decreases with increasing Fe/S at the lower Fe/S ratios. The scatter in the lower zinc solubility range is large. This might be attributable to analytical errors, because even for the initial "blank" Fe-S-O matte, the analytical results vary from 0.05 wt% Zn to 1.0 wt% Zn.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 657

1.8 ■

1.6

1.4

1.2

t9 °-8

0.6

0.4

0.2

0 2

Figure 10 - Measured Zinc Solubility as a Function of the Fe/S Ratio in Matte

CONCLUSIONS

The experimental program described in this paper confirmed the concept of the process and the thermodynamic analysis. That is, the sulfur in the feed materials can be fixed as a Fe-S-O matte using metallic iron or iron ore while both zinc oxide and sulfide materials can be simultaneously smelted in a single reaction medium. The combustion of carbon with oxygen inside the matte has been proved to be feasible.

Iron oxysulfide matte is an ideal medium for perform metallurgical reactions. This occurs firstly because of the degrees of freedom the oxysulfide matte phase provides, which makes the simultaneous smelting of sulfide and oxide materials in a single vessel and the combustion of carbon inside a matte phase feasible. It also occurs because of the intrinsic physical and chemical behavior of the system: low melting point and viscosity, high chemical activities of the chemical species (yzns=10 in matte vs. γζηο=0-2 in slag!), excellent reaction media for reduction, oxidation and sulphurization and self-fluxing of the refractory oxides. Despite the attractive characteristics and well-understood chemistry of the Fe-S-O matte system, its potential application to pyrometallurgical processing has not been well recognized. The thermodynamic analysis and its confirmation by extensive laboratory tests successfully demonstrated the feasible application of the oxysulfide matte to pyrometallurgical zinc processing.

ACKNOWLEDGEMENT

I Test 3, 4.07%O

Zn% normalized to Pzn = 0.1 atm, 1350C

Test 10,3.6 %0 Test 2, 2.91 °/4o ■ ■ Test 4, 5.45 %0

■ Test 9, 3.84 %0

Test 1, 2.69 %0

2.2 2.4 2.6 2.8 3 3.2 3.4

Fe/S Ratio

The authors would like to thank Noranda Inc. for permission to publish this paper.

658 LEAD-ZINC 2000

REFERENCES

1 R. Li, M. Zamalloa and P. Hancock, "Zinc Fuming from Iron Oxysulfide Melts", Challenges in Process Intensification. C. A. Pickles et al., Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1996,24-29.

2 R. Li, J. G. Peacey and P. J. Hancock, "Direct Smelting of Zinc Concentrate and Residues", The Howard Worner International Symposium on Injection in Pvrometallurgy. M. Nilmani and T. Lehner, Eds., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1996, 107-121.

3 R. Li and J. Peacey, U.S. patent. 5.443.614; Canadian Patent. 2.151.791.

4. Y.-Y. Chuang, K.-C. Hsieh and Y. A. Chang. Met. Trans.. Vol. 16B, 1985, 277-285.

5. E. M. Levin, C. R. Robbins and H. F. McMurdie, Phase Diagrams for Ceramists. The American Ceramic Society, Fig. 1902,1964.

6. V. Raghavan, Phase Diagrams of Ternary Iron Alloys, part 2, ASM, 1987, 197-208.

7. Y. Dessureault, "Modelisation Thermodynamique du Smeltage du Plomb Dans un Haut Foureau", Ph. P. Thesis. Ecole Polytechnique, Montreal, Quebec, Canada, 1993.

8. V. Rajakumar and M. W. Nagle, Pvrometallurgy for Complex Materials and Wastes. N. Nilmani, T. Lechner and W. J. Rankin, Eds., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1994, 233-247.

9. H. Shima and A. J. Naldrett, Economic Geology, Vol. 70, 1975, 960-970.

10. A. Yazawa and M. Kameda, Technol. Rep. Tohoku University. 16. Sendai, Japan, 1953, 40-58.

11. INCRA Series on the Metallurgy of Copper, "Thermodynamic Properties of Copper-Slag Systems". Vol. 3.1974. 97.

12. A. D. Pelton, G. Eriksson and A. Romero-Serrano, Met. Trans.. Vol. 24B, 1993, 817-825.

13. A. D. Pelton and M. Blander, Met. Trans., Vol. 17B(4), 1986, 805-815.

14. C. W. Bale, W. T. Thompson, A. D. Pelton, G. Eriksson, P. K. Talley and J. Melancon, Proceedings of the 2nd International Symposium on Computer Software in Chemical and Extractive Metallurgy. Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1993, 73-86.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 659

THE INFLUENCE OF NEW TECHNOLOGY AT SULPHIDE ORE MINE SITES ON METALS PRODUCTION AND RECOVERIES, WITH ITS COMMERCIAL

SIGNIFICANCE

H. Fletcher ZINCOR Ltd.

65 Wickersley Road Rotherham, S. Yorkshire, U.K. S60 3PX

P. Gray Zinc Metallurgy Ltd.

2, The Avenue, Backwell, Bristol, U.K. BS48 3NB

ABSTRACT

Process and engineering development of the Warner Zinc Process over the past two decades has proceeded alongside the world-wide collection and analysis of operating and cost data of current processing practices from sulphide ores to refined metals. Study of these data shows that recoveries are dependent on mineralogical characteristics of the sulphide matrix and point to the inadequacy of differential flotation as a tool for preparing feed for smelters. The paid-for return at the mine on the combined metal values in zinc rich sulphide ores rarely exceeds 80% and increasingly runs as low as 50-60%. Surveying the possible alternative ways of improving these performances leads to the conclusion that the Warner Process is the only technology in sight that could meet a target of 90% paid-for recovery and that its development has now reached a point where it can be a fully practicable process for all mixed sulphide ores.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

660 LEAD-ZINC 2000

INTRODUCTION

In 1985 in a paper presented at the Complex Sulphides Symposium, Gray (1) analyzed how the mineralogical character of sulphide ores related to metallurgical results by reference to a range of 12 Australian deposits and their concentrators. A measure for the metallurgical difficulty of each ore was proposed by determining the proportion of the Total Contained Value (TCV) in the run-of-mine ore that was recovered and paid for in the mine products. The TCV is a summation of all the contained metals of value in the ore (for the ores examined, this was Zn, Cu, Pb, Ag, andAu); thepaid-for values are those for which the mine receives f.o.b. payment in concentrate sales. Over the 12 operating mines, the proportion of TCV actually received in payments ranged from 94% down to 49%. The proportion clearly declined as the complexity of the sulphide mineral matrix increased. The highest percentages were for the copper ores with few by-products that receive payment for a high proportion of the copper content of the concentrates. Not surprisingly, the Zn/Pb mines with the highest payment recoveries were at Broken Hill with their coarse grained mineral matrices. Apart from Broken Hill, payments were 75% or less for the zinc-based ores.

The data were also examined to see what proportion of the lost value went to tailings and what was unpaid. The values are given in Table I. For the ores of simpler mineralogy, the losses of unpaid values are dominant; for those of complex mineralogy, the losses in tailings are much the greater. In 5 of the 12 mines the total lost value ranges from SAUD 66 to 105 per tonne of ore (at the metal prices of the time in Australian dollars) which was considerably more than the total TCV of the less rich ores. This is a tantalizing financial prospect; recovery as saleable product of only half these lost values could transform the financial situation of the mines.

Subsequently to publishing these data, one of the authors collected similar data from many more mines worldwide - especially those operating on zinc-rich sulphide deposits. The pattern was very similar to that for the Australian mines. The more recent the mine the more complex is likely to be its mineralogy in relation to flotation separation. Some of this worldwide data is given in Table II.

At present day metal prices, the average TCV for run-of-mine ores of this type is around SUSI40 per tonne. Of this, the world average paid recovery in mine products is probably no more than SUS80 per tonne of ore. At a annual mine zinc production of 7.5m tonnes per year, the attainable gross worldwide annual loss of revenues must be not less than SUS 5,000,000,000 to 6,000,000,000.

This paper emphasizes the need of the industry supplying zinc and allied metals to extend itself by adopting new technologies with the following targets:

• To increase recovery from ore to metal supplies by 20 to 40% to avoid waste of already established reserves

• Extend the reserves in the earth's crust by similar percentages • Allow the cut-off grade of existing reserves to be lowered • Promote further the treatment of sulphide ores in an environmentally friendly way with

minimal emissions at mine sites

To the metallurgist it is a challenge: there must be a better way.

Tab

le I

- D

istr

ibut

ion

of V

alue

Los

s fr

om O

re in

Min

e Pr

oces

sing

(fro

m r

efer

ence

No

1)

Ore

type

%

of t

otal

ore

val

ues

lost

in c

one

taili

ngs

% o

f tot

al o

re v

alue

s no

t pai

d fo

r in

con

es.

Tot

al lo

sses

%

Cle

an C

oppe

r or

e 4

Pyri

tic c

oppe

r or

e 11

C

oars

e hi

gh g

rade

Zn-

Pb-A

g or

e 6

ditto

with

hig

her P

b &

Ag

4 H

igh

grad

e Z

n-Pb

-Ag

5 Fi

ne g

rain

ed Z

n-Pb

-Cu-

Ag-

Au-

Pyr

13

Low

gra

de Z

n-C

u-A

g 9

Ver

y H

igh

grad

e Z

n-Pb

-Cu-

Ag-

Au-

Pyr

15

Fine

gra

ined

com

plex

Zn-

Pb-A

g 18

C

ompl

ex h

igh

grad

e Z

n, p

yriti

c 25

C

ompl

ex Z

n-Pb

with

hig

h A

g 34

U

ltra

fine

gra

ined

Zn-

Pb-A

g 37

2

2

12

14

14

15

22

16

14

11

11

14

6

13

18

18

19

28

31

31

32

36

45

51

Thi

s ta

ble

show

s ho

w th

e ph

ysic

al l

oss

of v

alue

s in

to ta

iling

s fr

om d

iffer

entia

l flo

tatio

n ri

ses

stee

ply

as th

e co

mpl

exity

of

the

min

eral

mat

rix

incr

ease

s. T

he s

econ

d co

lum

n sh

ows

a fa

irly

unifo

rm l

evel

of

unpa

id-f

or v

alue

s in

con

cent

rate

s fo

r th

e or

es

of t

he Z

n-Pb

-Ag

type

. It

is m

uch

low

er f

or t

he c

oppe

r or

es a

s th

e st

anda

rd r

ate

to p

aym

ent

for

copp

er i

n co

ncen

trat

es i

s su

bsta

ntia

lly h

ighe

r th

an it

is fo

r zi

nc.

The

com

bine

d lo

sses

(th

ird c

olum

n) r

ise

to f

orm

idab

le a

mou

nts

for

the

mor

e co

mpl

ex

ores

.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

Tab

le I

I - S

umm

ary

of M

ine

Rev

enue

s (E

xclu

ding

Min

e an

d Sm

elte

r C

osts

) fo

r O

pera

tion

with

Sta

ndar

d T

echn

olog

y M

etal

Pri

ces:

Zn,

$U

S 11

40/to

nne;

Pb,

480

; Cu

1720

: Ag,

$U

S 0.

146/

gram

me;

Au,

8.1

7 M

ine

No

1 2 3 4 5 6 7 8 9 10

Run

-of-

Zn %

0.

96

10.5

6.

8 9.

7 7.

8 6.

3 2.

6 5.

1 6.

7 14

.0

■min

e or

e Pb

%

4.

1 6.

0 - 0.

8 1.

4 6.

9 0.

2 2.

3 1.

8 3.

0

! as

say

Cu

%

0.15

0.

95

2.25

0.

43

1.2

- 2.13

0.

5 0.

7 0.

5

Ag

g/to

nne

- 103

5 220

178

162

18

128

93

102

Au g/to

nne

- 0.75

- - 3.1 - 1.2

- 0.3

-

Ton

nage

m

.tonn

es/y

r

1.0

0.43

2.

4 1.2

0.

333

2.9

1.7

0.36

5 0.

315

0.13

6

TC

V

of o

re

$US/

tonn

e or

e 3.

6 18

8.0

123.

4 15

6.5

169.

1 12

9.8

79.8

97

.5

114.

1 20

0.3

Rev

enue

cur

rent

ly

real

ised

in c

ones

$U

S/to

nne

ore

28.2

12

6.2

96.6

10

6.3

121.

3 89

.0

48.3

67

.0

76.8

15

8.1

Prop

ortio

n of

ore

T

CV

cur

rent

ly re

alis

ed

as re

venu

e %

83

.9

67.1

78

.3

67.8

71

.7

68.6

60

.5

68.7

67

.3

78.9

T

his

tabl

e su

mm

aris

es th

e ac

tual

rev

enue

ear

ning

s of

a n

umbe

r of

cur

rent

act

ive

min

es. T

he te

n m

ines

are

div

erse

in o

re a

ssay

and

in s

ize;

a

dive

rsity

that

cou

ld b

e m

irro

red

and

exte

nded

man

y tim

es o

ver f

rom

oth

er m

ines

aro

und

the

wor

ld.

The

last

two

colu

mns

sho

w th

e ex

tent

of

curr

ent

reve

nue

loss

whi

ch m

ight

be

reco

vera

ble

by o

ther

mea

ns.

It i

s th

e ch

alle

nge

to t

he m

etal

lurg

ist

to f

ind

a be

tter

way

. T

he a

bsol

ute

loss

es o

f re

venu

e ra

nge

up t

o ab

out

$US5

0 pe

r to

nne

of o

re (

Min

e N

o 4)

whi

ch e

quat

es t

o a

loss

of

reve

nue

of $

60m

per

yea

r. T

he m

ines

w

ith th

e bi

gges

t lo

sses

are

thos

e w

ith th

e m

ost c

ompl

ex o

res

and

thos

e w

ith th

e hi

ghes

t con

tent

of p

reci

ous

met

als.

LEAD-ZINC 2000

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 663

ALTERNATIVE ROUTES

The physical limitations of differential flotation cause the losses and, as the data collected show, the recoveries are getting worse. From this data bank, alternative ways by which the treatment of the sulphides might be improved were reviewed and the possible methods are discussed below.

Improved Differential Flotation

An obvious improvement would be to develop current practice. Flotation is probably the most thoroughly researched process in all extraction metallurgy and has led to advances in grinding methods and the extent of grinding, classification, flotation equipment, bubble creation, flowsheet design, reagent formulation and pre-conditioning by time, temperature or atmosphere and on-line control. The use of finer and finer grinding to increase liberation in the complex matrices and specialized flotation cells such as the Jameson cell is now common.

Rougher flotation to separate sulphides from non-sulphides is evidently not too difficult for the ores of volcanogenic origin because the boundary between the sulphide matrix and host rock is usually well defined. For the class of deposits of sedimentary origin (e.g. MacArthur River and Century in Northern Australia) it can be more difficult, as sulphides and gangue minerals are intimately mixed.

Separating one sulphide from another (cleaner stages) can be much more difficult because complete liberation of all the individual sulphides is virtually impossible. Furthermore the sulphides contain elements which are chemically bound (e.g., Fe, Mn and Cd in zinc sulphide) and are therefore outside the scope of separation by mineral processing.

For the concentrator, the market target is to prepare a product which, as nearly as possible, satisfies the smelter's specification (pure value mineral if possible) and at least to make a product that is better than that any competitor has to offer. In practice, the miner has to try to minimize the treatment charges, penalties and deductions negotiated with the smelter.

Despite the enormous effort put into improving flotation, it is clear that the trend is for concentrates to become less and less clean or recoveries have to be lowered to meet smelter specifications. One might particularly note that the separation of pyrite from other sulphides is an ever-pressing need as iron disposal by smelters has become a more environmentally sensitive issue. Differential flotation relies on a scientific paradox. In ores of volcanogenic origin the sulphide mass has formed in a combination of chemical and thermal reactions. In liberation and flotation the aim is to unravel what nature has raveled by purely physical means.

In viewing concentrator data from around the world and from all types of ores, it is impossible not to conclude that differential flotation is not winning and probably will fall behind in its ability to handle the more complex ores to stem the ever mounting losses of mined values. The gap between the potential value in the ores and what is realized is just too great to hope that a technique that has already so long been in use can now reduce it further markedly. Indeed there are many instances of known deposits of promising wealth content that are currently non-viable because flotation cannot recover an economic product from them.

664 LEAD-ZINC 2000

Hydrometallutgical

Because leaching largely destroys the mineral species, it is an attractive alternative to explore, as it is likely to be little influenced by the ore mineralogy. The technology of separating elements from each other in solution by crystallization, precipitation, ion exchange or solvent extraction has advanced massively in recent times.

The main process for recovering zinc from sulphide concentrates is, of course, the hydrometallurgical roast-leach-electrowin or pressure leach-electrowin approach. This process is not readily adaptable to increasing amounts of non-zinc elements in the feed. In the 1980's the Low Contaminant Jarosite Process (2) was developed at the Electrolytic Zinc Co. of Australasia in Tasmania and this might have been developed further to enable plants to be fed with less pure sulphide concentrates. Faced with the rising social pressure to improve the disposal of iron residues, however, the company had to pursue other routes to meet those pressing requirements.

If one considers the possible use of a similar route for treating sulphide ore or bulk concentrates at or near the mine to maximize recoveries, there are some fundamental difficulties that must be overcome.

• Conditions to get the best possible leach efficiency for Zn and Cu in an acidic solvent will also give a high level of iron in the solution. Iron removal from sulphate solution is possible, but it is not easy and gives rise to waste products that are not readily disposed of. For ores containing massive pyrite, as many do, the iron problem becomes formidable.

• Lead sulphide is converted to insoluble lead sulphate. This together with silver, gold and insoluble gangue minerals (mostly silicates) will form a leach residue. The contamination of sulphate and silicate is likely to yield a residue of less than 30% Pb. Unless rich in silver, it is rarely attractive to treat at or transport from a mine location.

• There is no known economic way of recovering zinc from a sulphate solution as a saleable product, except by electrolysis. Electrolytic plants require about 4000 kWh per tonne of cathode zinc and it is therefore essential that power be available at low cost. At or near mine sites, power is rarely available at less than 10 US cents per kWh. One has then either to consider transporting ore or a low-grade bulk concentrate to sites where low cost power is available or to devise a different route for recovering a saleable zinc concentrate from solution. Crystallization or precipitation of a zinc-rich oxide salt is chemically possible but the economics are not attractive because of high reagent costs; such products will have to be converted to metal elsewhere so that the selling price of the zinc at the mine has to be similar to that in sulphide concentrates, albeit at higher recoveries.

There are numerous permutations and combinations of different leaching and recovery techniques, but in our own investigations we did not see any real hope of an economic improvement on flotation.

There are other leaching agents that can be considered; e.g., chloride, cyanide and ammoniacal solutions. These too were given due examination but the above three difficulties applied to all of them in some degree with generally higher reagent costs to be added.

Figu

re l

- C

urre

nt E

cono

mic

Situ

atio

n fo

r Po

lym

etal

lic O

res

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

666 LEAD-ZINC 2000

High Temperature Processing

At temperatures of 1000° C or higher, almost all mineral characteristics are destroyed. If the solids are melted or volatilized they are entirely destroyed. Separation processes by phase separation or distillation can be used.

These principles are used to process multi-metal sulphide concentrates in the Imperial Smelting Zn/Pb blast furnace. Why could not the ISF be developed further to treat ores or low-grade bulk concentrates at or near the mine? The limitation to this development is also that of most forms of smelter; e.g., flash furnace, Zn, Pb or Fe blast furnace and submerged combustion furnace (e.g. the Ausmelt type). In all these processes the slag is formed at an oxygen potential such that the slags take up at least 5-7% Zn. This occurs because the oxidation (of carbon or sulphide) for heat generation is in the same chamber as the reduction of oxidized mineral to metal. As a result, the feed must have a not too high gangue to metal ratio or the losses of zinc and other values in slag will be prohibitive. Feeding an ISF with roasted low grade bulk concentrates would result in very poor recoveries; with ore much worse. Those who make and sell Zn/Pb bulk concentrates to ISF smelters are all too aware that ISF concentrate terms are not attractive.

It was concluded that none of the known smelting routes could be adapted to smelt zinc rich sulphide feed with a high slag fall and that, therefore, they would not be suitable for raising the recoveries of value from polymetallic ores.

IS THERE A BETTER WAY THAN DIFFERENTIAL FLOTATION?

By 1990, after a lot of study, it was concluded that a better way of treating the polymetallic sulphides was not in prospect. We were also forced to conclude that the gap of lost values was widening and was more likely to get still wider as the mineralogical complexity of the ores processed became worse. Strategically, the industry might be able, for a time, to select, and find, further deposits with "easy" mineralogy but with the demand for metals increasing this could be no more than a short term palliative.

In Figure 1, the current economic situation is shown in diagrammatic form. The chart shows performance for a representative polymetallic ore. The cost data are presented in an unfamiliar way: all the costs (vertical axis) are in $US per tonne of ore ex mine. The bars on the left of the chart are concerned only with revenue from the values in the mine (Zn, Cu, Pb, Ag, Au) and those to the right are all of the costs (i.e., mining, concentration, concentrate reduction to metals as represented by treatment and penalty charges paid by the mine, and concentrate transport from mine to smelter). The figures in the chart do not include capital charges, which are very site specific.

This chart does reasonably well represent the current plight of the industry - a small positive margin but one that is too easily eroded by a change in metal prices and one that inhibits modernization of the plant.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 667

THE WARNER PROCESS

During the 1980's, Professor Noel Warner at the University of Birmingham studied a new approach to the conversion of mineral sulphides to metals and published a number of papers on his work, (3,4,5). A paper summing the process concept and commercial potential was presented by Gray (6).

Warner's proposal was to use the properties of molten copper-rich mattes to act both as a medium for reduction to metal and for transferring heat. He showed that the following well-known reactions can all be carried out in molten matte at temperatures of 1150 to 1300°C.

ZnS + 2 Cu = Zn + Cu2S (1)

Cu2S + 0 2 = 2 Cu + S02 (2)

PbS + 2 Cu = Pb + Cu2S (3)

FeS + 1.5 02 = FeO + S02 (4)

The product of reaction 1 is zinc vapor; of reaction 2 is molten copper metal; of reaction 3 is lead metal; of reaction 4 iron oxide for slagging with silica. Reaction 4 is strongly exothermic; reactions 1 and 3 are endothermic.

Of these four essential reactions, 2 and 4 require oxidizing conditions, whilst 1 and 3 require neutral (non-oxidizing but not necessarily reducing) conditions. The chamber where reactions 2 and 4 are carried out must, therefore, be separate from that for 1 and 3. This division immediately establishes a new principle in sulphide metallurgy where zinc and lead metals are produced in contact with slags in non-oxidizing conditions. This has the promise of making zinc recovery largely independent of the zinc to gangue ratio in the feed. This is a most significant forward step because it means that low-grade sulphide bulk concentrates containing most (say, 90+%) of the values in the ore with much gangue can be efficiently converted to metals - something that is not currently possible by any processing route.

Warner proposed, therefore, a two-compartment smelter with the matte medium circulating continuously from one to the other. In the non-oxidizing chamber, the feed is incorporated into the matte and zinc and lead are released as metals according to equations 1 and 3. This chamber needs an input of heat that can be generated in the other chamber from the exothermic reactions between iron sulphide and oxygen. The surplus heat generated raises the matte temperature so that it is carried as sensible heat to the other chamber. Heat balance analysis showed that it was possible to make the combinations of reactions autogenous. A large purchase of expensive energy - an unavoidable feature of all current zinc technology - was thus no longer required. This opened the way to operating a smelter close to and fully integrated with a mine concentrator, a desirable feature for maximizing revenue from ore.

The engineering of such a system poses a number of novel challenges. Warner proposed that, to circulate the molten matte, a low pressure 'pump' be used based on the well known RH de-gasser of the steel industry, albeit at a much lower temperature than for steel refining. An elevation of 30 cm or so is sufficient to maintain the necessary circulation through the two chambers. The composition and temperature of the matte are close to those for mattes handled in copper smelting and for which the properties have long been established. The circulating copper matte would not be consumed and would, in fact, accumulate copper metal if the feed

668 LEAD-ZINC 2000

contained more copper as sulphide than would be lost in the iron silicate slag after slag cleaning. Condensation of zinc vapor is also a technology that has been practiced for all thermal zinc metal processes for centuries.

Once Warner had established from detailed test and computational work that the process was fundamentally feasible and asked no engineering questions that would not be solvable, it was possible to make an assessment of what the value of such a process could be for the processing of polymetallic ores.

DEVELOPMENT OF THE PROCESS PLANT ENGINEERING AND DESIGN

An attempt was made on the campus of the University of Birmingham to build and operate a small-scale rig with continuous circulation of matte and feed of zinc sulphide concentrate. The difficulties of achieving a heat balance at this scale with disproportionate losses through the walls together with satisfying the local requirements for health and safety on a smelting rig on a campus in a large city were too formidable to conquer. The rig did confirm that matte circulation and zinc volatilization did occur as predicted, but was not able to contribute much to the quantitative data on which a practical plant could be designed and built.

It was calculated that to build a plant in which the heat balance would be maintained would require a scale of around 25,000 tonnes per year of zinc metal output. The next step, therefore, was to carry out detailed design and costing studies for a production plant of this size. A further design and cost study was then carried out for a prototype plant of about 25% of the capacity of the production plant. This would be thermally self-sufficient but not economically so. This could be used as a demonstration plant and test-bed for parcels of different feed material.

The studies were carried out under contract by KSLE (Kilborn SNC Lavalin Europe) in 1997/8 aided by specialist input from ISP Ltd. on the zinc condensation section and VacMetals on the RH matte pump. The designs took into account all process and ancillary plant equipment for a green field site plant.

The design study established two important economic points:

• The process plant does not have any features that are not well known already in standard copper smelting, steel making or zinc smelting practice; temperatures are similar to those of copper smelting and refractory requirements are not abnormal. Ancillary plant such as the lead splash condenser, sulphur dioxide capture, oxygen making, metal casting, slag disposal, feed storage and blending, feed preheating and services distribution are entirely standard. The risk from catastrophic failure; e.g., a breakout from a furnace hearth or collapse of roof or sidewalls, should be low so long as the known practices for design are adhered to. Failures in ancillary equipment even on established plants are not unknown but are seldom catastrophic nor very costly.

• The capital for a production plant was estimated excluding the site acquisition cost. The total cost will, of course, vary somewhat from location to location but it was clear that it would be less than for an equivalent operation using present day technology (e.g., differential flotation followed by separate and distant smelting and refining of clean zinc, lead and copper concentrates). The standard plants would be replaced with a bulk concentrator, a single smelter and refineries. The capital cost comparison is not quite so

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 669

favorable to the single new smelter if the differential concentrator can still find existing amortized smelter capacity to utilize. The design study also showed clearly that the operating costs would be distinctly less than in current practice.

The Warner Process applied to polymetallic sulphides would lead to lower cost processing and may in some circumstances be justifiable for that reason alone. Its major cost attraction however is the increased revenue that comes from higher overall recoveries and the increase in unit product prices at the mine.

THE SMELTER PRODUCTS

Zinc

The natural zinc metal product is Prime Western (Zn4) grade since contact with lead is bound to give a product saturated in lead at the freezing point. However, it needs to be borne in mind that as a smelter of bulk concentrates it should recover, in payable form, over 90% of the zinc in ore. The metal price would be discounted for grade by no more than 10%. In contrast, processing based on differential flotation recovers zinc in concentrate (paid for at about 50% of LME price) and possibly payment for no more than 60-70% of the zinc in ore. The revenue difference at the mine is very large. Prime Western has a market though a declining one compared with the higher grades. As with Imperial Smelting, zinc refining is necessary to upgrade Zn4 to Znl or Zn2 grade metal.

Lead

Lead will be recovered as bullion and, as from a lead smelter, it will go to a refinery.

Copper

Copper metal is the product of the oxidation of the copper matte. The more copper that is in the feed the better are the economics since co-smelting of copper is low-cost. Tapped copper will have to be refined to give high-grade copper and to recover precious metals. The furnace product should be cast into anodes.

Precious metals

Precious metals are recovered by refining copper and lead in the normal way. As copper and lead metals are the only products of the furnace, which can dissolve significant amounts of precious metals, the recovery from smelter feed should be at least 95%.

Sulphur

For location at or near a mine site, the disposal of sulphur dioxide is rather a different proposition than for smelters located remotely from the mine. Marketing of sulphuric acid is unlikely from most mine sites but safe disposal by reaction of SO2 with calcareous rocks to form insoluble sulfates is an option since often the host rock to polymetallic sulphides is carbonaceous. There is no single answer to this question except that, whatever the location, the sulphur must be safely captured and disposed of and that this does not demand any novel technology but will incur some cost penalty.

670 LEAD-ZINC 2000

Iron

Iron sulphides are a major source of fuel to the process. The oxidation of iron sulphides to form ferrosilicate slag does not pose any major technical problem. This process can probably dispose of iron more easily and profitably than can any existing process route. This is a most useful feature because many of the polymetallic sulphides are Very pyritic and mineral separation of pyrite from the other sulphides results in losses of values.

Contaminants

In a smelting process some of the contaminants that are a serious concern in electrolytic zinc plants; e.g., Mn, Fe, Co, Ni, Ge, Hg are of minor consequence. The distribution of arsenic, antimony and bismuth cannot be fully defined until a prototype plant is operated, but a matte and an iron-rich slag should provide a good outlet for them.

ECONOMIC EVALUATION OF THE WARNER PROCESS

There are as many different evaluations as there are ore types and no two of those are the same. The only meaningful way of valuing is to calculate for each specific location and ore, and compare the economy with that for a standard processing route. This is something which has been done for many actual cases.

A general comparison can be developed from Figure 1. In Figure 2 the projected operating economics for revenue and costs are shown for the same ore as that used in Figure 1 and using the same units. Comparing the right hand side of the charts, there is a reduction in all costs except the mining. The total operating cost is reduced from $US 67 to $US 48 per tonne of ore processed.

The major change is in mine revenue which is increased from $US 77 to $US 119 per tonne of ore processed; this is due to the large improvement in recoveries and the higher added value.

The overall profit margin has increased from $US 10 to $US 71 per tonne of ore treated. In this case, a mine producing, say, 800,000 tonnes of ore per year would have an increased trading profit margin of $US 48.8m per year - a sixfold increase on that with standard practice. . Over 70% of this improvement is due to the increase in revenue or, to put it another way, to reducing the wastage of mine values.

The improvement in the paid for recoveries of each of the five values in the ore are compared in Table III.

I Ω σ a 00

en E

Figu

re 2

- Pr

ojec

ted

Ope

ratin

g E

cono

mic

s fo

r th

e Pr

opos

ed W

arne

r Pr

oces

s

672 LEAD-ZINC 2000

Table III - Paid-for Recovery of Metals in Polymetallic Ore Ore TCV Standard Practice Warner Practice

$US/tonne $US/tonne % recovery $US/tonne % recovery paid-for paid-for

Zn Cu Pb Ag Au Total

66 28 25 12 6

135

49 7 11 7 2 77

75.0 25.0 47.8 58.3 33.3 57.0

60 23 21 11 4

119

90.9 82.1 91.3 91.7 66.7 88.1

The minor element values - in a zinc ore these are Cu, Ag and Au - are responsible for a substantial part of the improvement. Copper and gold may be hosted in pyrite and become more readily recoverable. All three elements are sold in products (copper anodes and lead bullion) in which they are paid for at a high percentage of content. The greater the content of these three elements in the feed material the more advantageous economically the Warner route becomes. It might be practicable to treat a feed in which the Cu:Zn ratio is 1.0;· the feed then being composed most probably of a blend of low grade zinc concentrates and copper concentrates. Sulphide deposits of both metals are often located in the same region.

There are other likely economic benefits from using this route, but at this stage, they cannot be costed or included into a general illustration. Some of the more important ones are:

• The reserves of a deposit may be revised upwards by a reduction in the cut-off grade or by incorporating sections of the deposit as mineable ore.

• Ore prospects included on the assets of a company but currently valued as 'non-viable' may be up graded to 'book assets'. This may particularly apply to deposits which have contaminants (e.g., Mn or Co) too high for the production of zinc concentrates destined to be treated by the electrowinning route.

• Metal production is brought back from the location of smelters to the mine location. The increase in added value may be advantageous for both the mining company and the host country.

• Tailings for ground or underground disposal should contain fewer sulphides and thus be less reactive. Most of the iron in ore, as pyrite, will be disposed of as iron silicate slag that is relatively unreactive. Sulphur may be disposed of into surface dumps as unreactive sulfates at some disposal cost but using practices that are well established already.

To broaden the picture of the economic scope of the process, some summary figures for various actual ores and mines that have been evaluated are appended.

In Table IV, the data given in Table II are projected to the revenue targets that we now consider are realistic for each of these mines using the Warner Process system. The projected improvements in revenue will, of course, be further increased by the cost savings possible in operations. As one would expect, the mines that stand to gain the most are those with the more complex ore mineralogy. Mine No. 9 is on an ore that contains significant amounts of cobalt in

Tab

le IV

- S

umm

ary

of M

ine

Rev

enue

s (E

xclu

ding

Min

e an

d Sm

elte

r C

osts

) fo

r Pr

ojec

ted

Ope

ratio

n w

ith th

e W

arne

r Pr

oces

s M

etal

pri

ces:

Zn

$US

1140

/tonn

e; P

b 48

0; C

u 17

20; A

g $U

S 0.

146/

gram

me;

Au

8.17

/gra

mm

e

Min

e N

o.

1 2 3 4 5 6 7 8 9 10

Ore

TC

V

$US/

tonn

e or

e 33

.6

188.

0 12

3.4

156.

5 16

9.1

129.

8 79

.8

97.5

11

4.1

200.

3

Cur

rent

Act

ual,

see

Prop

ortio

n of

T

CV

rea

lised

%

83

.9

67.1

78

.3

67.8

71

.7

68.6

60

.5

68.7

67

.1

78.9

Tab

le II

Pr

ojec

ted

real

izat

ion

from

ore

$U

S/to

nne

ore

31.5

16

8.4

113.

6 13

7.2

154.

4 12

0.4

72.7

89

.4

95.9

19

1.9

Proj

ecte

d da

ta fo

r W

arne

r Pr

oces

s Pr

ojec

ted

real

izat

ion

as p

ropo

rtio

n of

or

e va

lue

%

93.7

89

.6

92.1

87

.7

91.3

92

.7

91.1

91

.7

84.0

95

.8

Proj

ecte

d U

p-lif

t in

reve

nue

$US/

tonn

e or

e 3.

3 42

.2

17.0

30

.9

33.1

31

.4

24.3

22

.4

19.1

33

.8

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

674 LEAD-ZINC 2000

its sulphides. Cobalt has a serious effect on tankhouse operation in zinc plants and consequently this mining company can sell its zinc concentrates only in small parcels and with heavy penalties on their cobalt content. Cobalt should be efficiently lost in the smelter slag. A similar situation will apply to ores containing high levels of manganese.

CONCLUSIONS

The Warner Process is not yet a production route for treatment of zinc-rich sulphides. As the developers, we have had to realise that the cost and risk of designing, building and operating a prototype plant purely as a demonstration of practicability of the projected metallurgy are too great to implement. We have, therefore, concentrated the speculative funding on detailed design studies of the plant engineering and extensive studies of the economics in real situations. This has been justified by the repeated comfort of finding that the fundamental metallurgy is wholly within well-established parameters and that the plant engineering is not outside ofthat in common use around the world already. Following on these studies, we can now confidently project performance and economic targets for diverse applications. This is not to say that we have all the answers to many detailed questions yet - we do not. We do believe that we can now advise a mining company or consortium of companies to invest in a prototype unit to advance its business prospects of profitable trading and investment return from a few years hence and thereafter. As we are not exploration experts, we can only tentatively put forward the view that the possibilities of a rewarding return from investing in this sort of process are better and the technical risks lower than that in looking for new and profitable ore deposits.

Our studies of the treatment of polymetallic sulphides have now spanned several decades. We cannot see that the future prospects for the production industry can be much better than breaking even over rises and falls in metal prices with the present technology. It is over-due for a fundamental renewal to regain the financial attraction it once had. We believe that the Warner Process is the route for new investment now.

ACKNOWLEDGEMENT

The authors wish to acknowledge the help and advice in preparation of this paper from Dr. Rod Sinclair of Saitis Nominees Pty Ltd, Melbourne.

REFERENCES

1. P.M.J. Gray, "Metallurgical Characterisation of the Zinc-Lead-Copper-Silver-Gold-Pyritic Ores " Complex Sulphides, Processing of Ores, Concentrates and By-products. A.D. Zunkel et al., Eds., TMS-AEViE, Warrendale, PA, U.S.A.. 1985, 759-772.

2. I.G. Matthew, C.J. Haigh, and R.V. Pammenter, " Jarosite-type Compounds and Their Application in the Metallurgical Industry", Hvdrometallurgy Research. Development and Plant Practice. K. Osseo-Asare and J.D. Miller, Eds., TMS-AIME, Warrendale, PA, U.S.A., 1983, 553-568.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 675

3. N.A. Warner, "Direct Smelting of Zinc-Lead Ore", Trans. Institution of Mining and Metallurgy, 1983, Sect C, C147-152.

4. N.A. Warner, "Towards Polymetallic Sulphide Smelting", Complex Sulphides. Processing of Ores. Concentrates and By-products, A.D. Zunkel et al., Eds., TMS-AIME, Warrendale, PA, U.S.A., 1985, 847-865.

5. R.K. Hanna and N.A.Warner, "Process Requirements for the Direct Condensation of Both Zinc and Lead as Metals in Polymetallic Smelting of Zn-Pb-Cu Sulphides", Proc. Non-Ferrous Smelting Symposium: 100 Years of Lead Smelting and Refining in Port Pirie. The Australasian Institute of Mining and Metallurgy, Melbourne, Australia, 1989, 227-236.

6. P.M.J. Gray, "The Warner Zinc Process", World Zinc '93. I.G. Matthew, Ed., The Australasian Institute of Mining and Metallurgy, Melbourne, Australia, 1993, 483-489.

Edited by J.E. Dutrizac. J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 677

RECOVERY OF ZINC AND CADMIUM FROM LEAD SMELTER FURNACE DUSTS AT MET-MEX PENOLES BY A SOLVENT EXTRACTION PROCESS

I. Sofia Fernandez del Rio Met-Mex Penoles S.A. de C. V.

Av. Metalurgica # 550 Col. Metalurgica Torr eon, Coahuila, Mexico 27370

ABSTRACT

This paper describes a process for the treatment of the dusts generated during the operation of blast and reverberatory furnaces in the lead smelter, to remove zinc, cadmium, arsenic and halides from the circuit. The process consists of a Pefloles' developed leaching/purification stage and a solvent extraction process developed by Lurgi. The plant capacity is 5,000 tonnes Zn/year. The feed to solvent extraction contains 8 g/1 Zn. Impurities such as copper, iron, arsenic and cadmium are removed by conventional cementation processes. The cadmium product is 99.99 % pure. The solvent extraction process consists of two extraction stages, one washing stage and two stripping stages. D2EHPA is used as the extraction agent and 25% ammonia as the neutralizer. The rich-zinc solution is sent to the electrolytic zinc plant for zinc recovery. The raffinate produced, which is a mixture of ammonium sulfate and chloride, is used for agricultural applications.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

678 LEAD-ZINC 2000

INTRODUCTION

The furnaces operation at the Met-Mex Penoles lead smelter generates 1,700 tonnes/month of baghouse dusts ( dust from the blast and reverberatory furnaces ), which have been recycled in the process, causing a saturation of non desirable elements such as Zn, Cd, As and halides.

A solvent extraction plant, designed for the treatment of these dusts, began operation in September 1998, with a capital investment of about 9 million US dollars, and it has the following operating, economic and environmental advantages for Met-Mex Penoles.

• Operating advantages: The bleed of impurities from the lead circuit allows more metallurgical flexibility in the process for treating " dirty " lead concentrates.

• Economic advantages: The removed impurities are recovered as saleable products: zinc sulphate solution, metallic cadmium and ammonium salts for agricultural applications.

• Environmental advantages: The process allows more efficient materials handling because rehandling and airborne dusts are decreased.

The process for treating the baghouse dusts consists of leaching the dusts with a sulphuric acid solution, generating a lead and silver residue as a by-product, which is returned to the lead smelter, and a zinc sulphate solution, that is purified to remove Pb, Fe, Cd and As. Zinc extraction is carried out by a solvent extraction process based on Lurgi technology. The zinc-rich solution obtained in the solvent extraction plant is sent to the electrolytic zinc plant for zinc recovery and an ammonium sulphate and chloride salt is obtained from the raffinate for agricultural applications. The plant has a capacity of 5,000 tonnes Zn/year.

RAW MATERIALS AND PRODUCTS

Raw Materials

Dust from baghouse Number 3 ( PCS 3 ) is a by-product generated in the blast furnaces during sinter melting. Dust from baghouse Number 4 ( PCS 4 ) is generated in the reverberatory furnaces during the treatment of the dross coming from the refining kettles. The compositions of the dusts are given in Table I.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 679

Table I - PCS 3 and PCS 4 Chemical Analyses

Element PCS 3 PCS 4 (%) (%)

Zinc (Zn) Cadmium ( Cd ) Arsenic (As)

Lead (Pb) Iron (Fe)

Chlorine (Cl ) Fluorine ( F ) Sodium (Na)

Potassium (K )

10.59 23.82 1.06

39.10 0.14 4.32 0.52 0.48 2.82

15.6 1.11 8.80

30.20 0.07 0.44 0.27 8.07 0.70

Main Products According to the Design Basis

Zinc sulphate solution, which is sent to the zinc plant for zinc recovery as cathodic zinc is the main product. Its production is 5,000 tonnes Zn/year. The composition of the zinc sulphate solution is given in Table II.

Table II - Zinc Sulphate Solution Chemical Analysis

Element Analysis Units

Zinc (Zn ) Cadmium (Cd) Arsenic (As)

Lead (Pb) Iron (Fe)

Chlorine (C l ) Flourine ( F )

Free acid

95 128

<1.0 <2.0 <2.0 7.0 2.4 140

g/i ppm ppm ppm ppm ppm ppm g/1

Cadmium of 99.99 % purity is made. A production of 700 tonnes is expected for the first year, decreasing to 240-360 tonnes in the following years, according to the cadmium inputs in the lead concentrates. The composition of the cadmium is presented in Table HI.

680 LEAD-ZINC 2000

Table III - Cadmium Chemical Analysis

Element Cadmium (Cd)

Zinc (Zn) Arsenic (As)

Lead (Pb) Iron (Fe)

Thallium (Tl)

Analysis 99.99 3.15 <1.0 33.67 3.64 2.0

Units %

ppm ppm ppm ppm ppm

Ammonium salts ( ammonium sulphate and chloride, mainly ) are made for marketing. The production is 9,650 tonnes/year. The composition of these salt is summarized in Table IV.

Table IV - Ammonium Salts Chemical Analysis

Element

Zinc (Zn) Cadmium (Cd) Arsenic (As)

Lead (Pb) Iron (Fe)

Sodium (Na ) Potassium ( K ) Chlorine (Cl) Fluorine ( F )

Ammonium (NH4+ ) Sulphate (S04

2")

Analysis

0.64 2.15 <2.0 0.31 <0.2

1.2 3.17 6.34 0.83 25.07 64.30

Units

ppm ppm ppm ppm ppm

% % % % % %

Lead and silver residues, which are recycled to the lead smelter, are also produced. Their compositions are presented in Table V.

Table V - Lead and Silver Residues Chemical Analyses (Dry Basis)

Element

Zinc (Zn ) Cadmium (Cd) Arsenic (As)

Lead (Pb) Iron (Fe)

Chlorine (Cl)

PCS 3 Analysis (%)

1.605 7.45 1.01

52.80 0.085 0.880

PCS 4 Analysis

(%) 3.49 0.29 2.40

49.40 0.15 0.13

H2S

0 4

>

PCS

3

JL

V

PCS

3 L

EA

CH

ING

Ψ

Lea

d an

d si

lver

resi

due

To

Lea

d Sm

elte

r W

ater

SrC

O,

KM

n04

H,0

,

PUR

IFIC

AT

ION

Pb/F

e/A

s R

esid

ue to

L

ead

Smel

ter

Zn

pow

der

CA

DM

IUM

C

EM

EN

TA

TIO

N

Ψ

<r

RA

FFIN

AT

E

CO

NC

EN

TR

AT

ION

Raf

fina

te

conc

entr

atio

n

Na 2

S

CA

DM

IUM

D

EN

SIFY

ING

SOL

VE

NT

E

XT

RA

CT

ION

Wat

er '

HE

AV

Y

ME

TA

LS

RE

MO

VA

L

Wat

er

(ste

am)

<r

V

V

NaO

H ►

WA

SHIN

G

—>

Met

alic

su

lphi

des

to

Lea

d Sm

elte

r

SPR

AY

D

RY

ER

J A

mm

oniu

m

salts

Spen

t el

ectr

olyt

e fr

om

Tan

khou

se

Loa

ded

Org

anic

STR

IPPI

NG

ME

LT

ING

AN

D

CA

DM

IUM

R

EFI

NIN

G

NH

4CI

Free

O

rgan

ic

Ψ

Cd

slab

V

Dro

ss to

L

ead

Smel

ter

Wat

er

Ψ

Ric

h el

ectr

olyt

e to

Zin

c Pl

ant

Figu

re 1

- Flo

wsh

eet f

or th

e D

ust T

reat

men

t Pro

cess

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund

682 LEAD-ZINC 2000

LEACHING STAGE

The general flowsheet for the dust treatment operation is given in Figure 1. The objective of the leaching stage is to dissolve the zinc, cadmium, arsenic and halides contained in the PCS 3 and PCS 4 dusts, using a sulphuric acid solution and generating a lead and silver residue with the lowest levels of the mentioned elements. Zinc and cadmium from the dusts pass into the solution during the leaching process, as zinc and cadmium sulphates, whereas arsenic is dissolved as arsenic acid. Lead remains in the leach residue as lead sulphate mainly.

The physicochemical properties and characteristics of PCS 3 and PCS 4 are considerably different, so that their treatment is carried out separately and in a batch mode. At present, only the PCS 3 dust is treated using the following operating conditions:

• Leaching reaction time: 30 minutes • Temperature: 70-75 °C • Final solution pH: 1.0-1.5 • Particle size: 100 % - 20 microns.

The extraction efficiencies in the leaching stage are:

• Zinc: 90- 93 % • Cadmium: 40-45% • Arsenic: 0.07-0.19% • Chlorine: 90-95 % • Fluorine: 90-95 %.

PURIFICATION STAGE

The filtrate from the leaching stage is purified by the addition of strontium carbonate, hydrogen peroxide and potassium permanganate, to cement lead, iron and arsenic. The pH for cementation is in the range from 2.0 to 3.5. Impurities tend to redissolve at lower pH's. When the reaction is over, the suspension is pumped to a filter press. Other important operating parameters in the purification stage are:

• Reaction time: 30 minutes • Solids concentration: 15 g/1 • Particle size: 5.2 microns (average).

The precipitation efficiencies are:

• Zinc: 0.0 % • Cadmium: 0.87 % • Lead: 98 % • Iron: 99.0 % • Arsenic: 99.0 %.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 683

CADMIUM CEMENTATION STAGE

The filtrate from purification is sent to the cementation reactor, where zinc powder is added to cement cadmium. Cementation consists of the reduction of metallic ions in solution to their elemental state by the action of a more electropositive metal:

Cd+2 + Zn->Cd + Zn+2 (1)

When the reaction period is over, the generated sponge cadmium is separated from the solution by decantation. The operating conditions in this stage are:

• Reaction time: 30 minutes • pH: 2.0-3.5 • Cadmium recovery: 99.0 %.

MELTING AND CADMIUM REFINING STAGES

The cadmium sponge goes through a briqueter with the objective of compressing the sponge cadmium to eliminate residual moisture and to avoid cadmium oxidation during the melting stage. Cadmium briquets are melted in kettles, in the presence of sodium hydroxide to remove the contained zinc. The zinc-free cadmium is sent to a refining kettle, where thallium removal is carried out by adding ammonium chloride. Refined cadmium goes to the casting area to make cadmium slabs.

SOLVENT EXTRACTION

The solvent extraction process consists of two extraction stages, one washing stage and two stripping stages.

The extractant agent is di-2-ethylhexyl phosphoric acid. The extraction grade depends on the pH of the operation, in a range between 1.5 and 5. In the lower range, the solvent extraction capacity is relatively low, compared to that obtained at pH = 5. However, it is not possible to operate at high pH's because of the presence of iron and other cations in the zinc-rich solution which precipitate as hydroxides, causing troubles in the phase distribution.

The reactions carried out are:

2 D2EHPAH + ZnS04 -» (D2EHPA)2Zn +H2S04 (2)

H2S04 + 2NH3 -* (NH4)2S04 (3)

6 D2EHPAH + Fe2(S04)3 -> 2(D2EHPA)3Fe + H2S04 (4)

2 (D2EHPA)NH4 + ZnS04 ■* (D2EHPA)2Zn + (NH4)2S04 (5)

684 LEAD-ZINC 2000

The sulphuric acid released during the extraction ( equation 2 ) has to be neutralized by adding 25% ammonia, in order to achieve the optimum loading of the solvent and to keep the pH in the range required ( 2.5 - 3.0 ) (Figure 2).

e o a a X

^5

90

80

70

60

50

40

30

20

10

0 0 1 2 3 4 5 6 7 8

pH

Figure 2- Extraction of Metals by D2EHPA from Sulphate Solutions (from Henkel Corporation)

EXTRACTION STAGES

The extraction of zinc from the zinc-rich liquor, coming from the cadmium cementation stage, is carried out in two steps, using D2EHPA and kerosene (12 % and 88 % by volume, respectively) as the extractant. The organic / aqueous ratio is 1.5: 1.

The zinc concentration in the feed stream is 8 g/1 and extractions higher than 98% are achieved. The generated aqueous phase is the raffinate and it is an ammonium sulphate and chloride solution. Ninety percent of the raffinate is recycled to act as dilutant to adjust the zinc concentration of the zinc-rich liquor to 8 g/1. The balance of the raffinate goes through a coke coalescent filter, in which the organic drops, carried along with the raffinate, are recovered. Finally, the aqueous phase is sent to the concentration stage.

WASHING STAGE

The objective of this stage is to put the loaded organic in contact with water, to remove the halides physically trapped by the organic in the extraction stages. The organic / aqueous ratio in this stage is 1:1. The washed organic is sent to the stripping stage and the water is recycled until the concentration of chlorides builds up to 4 g/1, at which point it is discharged to the leaching area.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 685

STRIPPING STAGES

The loaded organic from the washing stage is sent to two countercurrent stripping steps, in which it is put in contact with sulphuric acid to strip the zinc. As is observed in equation (2), the equilibrium is moved to the left with high acid concentrations. Therefore, D2EHPA-Zn salt is decomposed, regenerating the D2EHPAH in its free acid form and zinc sulphate that is dissolved in the sulphuric acid. The D2EHPAH is recycled to the extraction stage and the zinc sulphate solution is sent to the electrolytic zinc plant, passing first through adsortion filters to prevent the organic solvent from entering the zinc processing plant.

The sulphuric acid used in this stage is spent electrolyte from the zinc plant tankhouse, and it loads from 50 to more than 90 g Zn/1 in the solvent extraction plant and reintegrates it into the electrolytic process.

RAFFINATE CONCENTRATION

The raffinate from the solvent extraction plant is pumped to a basin with two evaporative towers, from which is sent to a heat exchanger. Most of the hot raffinate is recycled to the top of the towers and about 66 % of the volume of water is evaporated, by cooling with an air flow in a countercurrent mode. The concentrated raffinate is sent to the heavy metals removal stage. The recycled raffinate flow in the tower is determined by the concentration of the hot raffinate.

HEAVY METALS REMOVAL

The objective of this stage is to remove the traces of heavy metals contained in the concentrated raffinate, so that the ammonium salts comply with commercial specifications. The purification process consists of the precipitation of metallic sulphides by adding sodium sulphide to the concentrated raffinate. The generated suspension is filtered, leaving a mixture of ammonium sulphate and chloride concentrated solution, that is pumped to the fertilizer plant (fertlizer operations). The solid residue recycled to the lead smelter. The operating conditions are:

• Reaction time: 30 minutes • Concentration of Na2S solution: 20 % by weight • Operating pH: 5-6.

The precipitation efficiencies are:

• Zinc: 97-98 % • Cadmium: 99-100% • Lead: 98.5% • Arsenic: 99-100%.

686 LEAD-ZINC 2000

CONCLUSIONS

Met-Mex Penoles, by introducing solvent extraction technology, was able to achieve flexible operations for treating lead concentrates, a reduction in secondary materials recycling and economic benefits because of additional zinc recovery as well as a new agricultural product.

The operational results show that the extractions and recoveries in the different stages are according to the design parameters. Current operation is about 40 % of the installed capacity. Some adjustments are under review to achieve the design plant capacity.

The solvent extraction process offers operating, economic and enviromental advantages, already mentioned in this paper, that make this technology very atractive for its integration into Met-Mex Penoles processes.

AKNOWLEDGEMENTS

Thanks are due to Lurgi Metallurgy GmbH personnel for the solvent extraction technology and to Penoles' Research and Technological Development Center, for its collaboration in the development of new technology for the leaching and raffmate areas.

REFERENCES

1. I. Almaguer, Cadmium Plant's Project Book. Met-Mex Penoles, Torreon, Mexico, 1995.

2. Lurgi Metallurgy GmbH, Basic Engineering Zinc Solvent Extraction Plant. Met-Mex Penoles S.A. de C.V., 1995.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke. S.E. James and A.H.-J. Siegmund 687

THE KIVCET TREATMENT OF POLYMETALLIC FEEDS

L.V.Slobodkin OAO "Kazzinc"

1 Promyshlennaya Street, 492020 Ust-Kamenogorsk, Kazakhstan Republic

Yu.A. Sannikov, Yu. A. Grinin, M.A. Lyamina, V.A. Shumskij and N.N. Ushakov Vniitsvetmet Institute

1 Promyshlennaya Street 492020 Ust-Kamenogorsk, Kazakhstan Republic

ABSTRACT

This manuscript describes the Kivcet treatment of lead-bearing materials having high contents of copper and zinc. The tests were carried out on bulk concentrates from the Zyrianovsk Mining and Concentrating complex, "Kazzinc" JSC (Kazakhstan) with the following composition: Pb 37.6%, Cu 10.4% and Zn 6.8%. Two properties of the concentrate were under investigation: their calorific value and their desulphurization rate. By adding flux, the charge properties were brought close to the properties of high-grade lead sulphide concentrate. The charge was melted with high productivity. To ensure high copper recovery, Waelz residue was added to the charge. The residue contained more than 20% carbon and about 30% iron metal. The residue was partially replaced by coke that facilitates high copper recovery into the matte. The smelting achieved a production rate of 20 t/h with one burner. Lead recovery into the lead bullion was 88.8%; copper recovery into matte was 92.3%; zinc recovery into slag was 75.5%; and zinc recovery into the fume was 5.1%.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

688 LEAD-ZINC 2000

INTRODUCTION

A distinctive feature of the Kivcet process is its environmental friendliness, ability to run with high metals recovery and gaseous sulphur dioxide recovery. Originally, the Kivcet process was designed as a process for sulphide concentrate smelting. With that feed excellent process and cost attractive results were achieved.

Oxidized lead dusts, lead sulphate residues and jarosite-process residues were the second type of feed treated by the Kivcet process. For that type of feed, the Kivcet unit was improved allowing the achievement of high process and economic values of the smelting as well. The running of the Trail lead smelter (Cominco Ltd., Canada) has confirmed this situation. However, the raw materials base of the lead industry goes beyond pure sulphide concentrates and oxidized lead middlings. Large amounts of concentrate are produced from polymetallic ores, by putting them through selective concentration operations, and facing considerable losses of metals and high expenses for the resulting selectivity.

Owing to the foundation laid in the Kivcet technology, work was done in the institute of VNIItsvetmet (The Eastern Institute for Non-ferrous Mining and Metallurgical Research) to extend the raw materials range for the Kivcet process to involve complex lead-copper-zinc ores and concentrates. This presentation looks at the possibility for a further advancement of the Kivcet process with the aim of treating a complex high-copper-lead-zinc feed. Thus, an attempt has been made to make the Kivcet process a versatile technology for the treatment of sulphide, oxidized and complex lead-copper-zinc raw materials.

RESULTS

The lead-copper-zinc bulk concentrate produced at the Zyryanovsk mining concentrator owned by OAO "Kazzinc", having the following composition: lead 37.6%, copper 10.4% and zinc 6.8%, was employed for the tests.

Since the lead smelter owned by OAO "Kazzinc" has ten years of experience in smelting lead-zinc sulphide concentrates by the Kivcet technology, a method of comparing physical and chemical properties of the bulk concentrate to the characteristics of typical lead-zinc raw materials has been developed. This method was used to decide on the feed composition and process operating conditions. Two major features of the bulk concentrate that affect the results of the smelting in the Kivcet unit were examined: the calorific value and the desulphurization rate.

Through changing proportions of the fluxes, a charge was made up whose calorific value corresponded closely to that of the lead-zinc feed that provided the best smelting results achieved in the Kivcet furnace. The feed turned out to be the one whose composition can be varied over a rather wide range: lead 22-42%, zinc 5.0-7.6%, copper 8-10% and calcium oxide 0.3-0.8%. The calorific value of the feed ranges from 600-660 kcal/kg, which is similar to the case of a lead-zinc sulphide mixture.

Once the charge was made up to generate the calorific value desired, the opportunity was taken to compare the desulphurization rate of the charge to that of the lead-zinc concentrate charge. As expected, the duration of the desulphurization process turned out to be nearly the same. If half of the lead-copper-zinc concentrate is substituted for lead sulphate middlings (for example, lead residue or lead dust), the desulphurization rate noticeably increases. This can be explained by the fact that the rate of the exchange reactions between sulphides and sulphates,

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 689

proceeding under the conditions of very fast heating in the oxygen stream, considerably exceeds the oxidation rate of the pure sulphide feeds under similar conditions. Therefore, the variant of the combined smelting of lead-copper-zinc sulphide concentrates with oxidized lead materials resulted in an increased desulphurization rate. Consequently, the specific production rate of the Kivcet unit increased.

The influence of the desulphurization degree on the recovery of copper to matte and the matte composition was investigated. As one might expect, a too high desulphurization rate caused the enhanced solubility of copper in lead, and the copper recovery to matte decreased. The best results were realized for a desulphurization rate within the range of 80-83%. The proportion of copper passed to matte ranged from 87 to 94%.

The possibility of making a high-copper matte using a copper-rich feed was also investigated. With this objective, the copper drosses produced in the decoppering of lead bullion were added into the charge and then it was smelted in the Kivcet unit. Almost all the copper went to the matte, making it rich in copper. However, much of the lead sulphide passed to the matte together with the copper. To remove lead from the matte, metallic iron, as the Waelz slag magnetic fraction produced by magnetic separation, was added to the charge. Additional tests showed that the best results on copper recovery to matte could be achieved when metallic iron, along with drosses, was added to the charge in an amount equal to the dross weight. The iron additions decreased the lead in matte by a factor of as much as three times.

Besides the metallic iron, a Waelz slag produced from lead residues is advantageous to add into the feed of the Kivcet process. The size of the Waelz slag should be minus 20 mm; its moisture content should be no more than 3%>. The composition of the mixed feed using Waelz slag is shown in Table I.

Table I - Chemical Composition of the Feed Made-up of the Polymetallic Material and Waelz Slag

Charge Composition, % Waelz Slag Composition, % Pb 22.4-42.0 0.2 Zn 5.0-7.6 0.1 Cu 5.2-8.0 3.0-3.5 Fe 12.0-17.0 28.5-29.0 S 21.6-26.5 5.0 Si02 8.0-10.0 11.0-12.0 CaO 0.3-0.8 3.0-3.6

_C 20.0-25.0 Calorific value of the feed is 600-760 kcal/kg

The need to add Waelz slag to the charge is dictated by two circumstances. First, Waelz slag is a good reductant since it contains free unreacted carbon; second, iron contained in the Waelz slag is also a good reductant and serves to remove lead from the matte. Thus, a fundamental possibility to smelt a lead-copper-zinc feed achieving satisfactory process values was found in the course of the laboratory tests.

At the same time, the Vlltsvetmet specialists made a statistical analysis of the laboratory experimental and industrial results on the smelting of various lead feeds, including a high-copper and valuable metals feed (Cu 3-10%, Au 50-100 g/t, Ag 2,000-5,000 g/t). Basic factors crucial for the quality of the smelting products, and the non-ferrous and noble metals distributions, were identified. These tests provided for the development of a statistical model of

690 LEAD-ZINC 2000

oxygen-flash smelting of any type of lead feed. This model explained the behavior of the feed constituents on the basis of the parametrical relationships. The model also makes it possible to find out the best conditions for oxygen-flash smelting of lead feeds with respect to the minimum losses of valuable metals in the slags depending on the feed composition or to choose the compositions of the feed to get high-grade smelting products (matte and lead bullion) with minimum allowable losses of the valuable metals in the slags.

The technological regulations for the industrial application were developed on the basis of laboratory and simulation studies done in the VNIItsvetmet together with the specialists of OAO "Kazzinc". They comprised the following metallurgical processes:

• Feed preparation and drying • Feed smelting in the Kivcet furnace and smelting products tapping • Gas cooling and de-dusting • Gas recovery.

Feed Preparation and Drying

The feed is prepared by mixing, grinding and drying of the following materials (in weight %): a lead-copper-zinc bulk concentrate 78-80%; iron concentrate or metal chips 2.5-3.0%; copper drosses 2.5-3.0%; limerock 8-9%; quartzite ore 7-8%. The calorific value of this kind of charge (feed) is in the range from 600 to 660 kcal/kg.

The feed prepared in this way is delivered from the bin to the burner unit and it is weighed on a steady basis into the burner. Waelz slag under 20 mm in size is continuously weighed as it is conveyed from a separate bin to the burner unit.

Feed Smelting

Process oxygen (96-98% O2) is injected and Waelz slag is continuously added together with the feed to the burner unit. The mixture of feed and Waelz slag in the oxygen flow arrives at the smelting shaft. There, it is ignited on exposure to the high temperature of the gases. It is then oxidized and autogenously smelted in suspension owing to the heat generated from the oxidation of the sulphides and extra fuel. A finely dispersed oxide smelt is produced. In the lower part of the smelting shaft the flame temperature is as high as 1350°C.

Since the Waelz slag particles are coarser than the other feed particles, they are not burnt away in the flame but are only heated while they fall to the surface of the slag bath. The Waelz slag forms a carboniferous checker on the surface of the Kivcet slag bath. Carbon from the coke checker is steadily consumed to reduce the lead oxide entering in with the Waelz slag. The coke checker is always carbon-saturated. Fine droplets permeate through the coke checker and undergo a selective reduction; the lead is reduced to metal, a major portion of the zinc remains in the slag as zinc oxide and the copper forms mostly copper matte. Copper drosses and metallic iron reach the matte smelt, react with the reduced lead sulphide and upgrade the quality of the matte. Gold, silver, antimony and bismuth accumulate in the lead bullion.

The products (lead bullion, matte and slag) pass from the smelting shaft to the electric thermal zone of the Kivcet unit and separate there by density. The various phases are tapped from the furnace through the tap-holes arranged at different levels.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 691

Lead bullion is sent to refining, matte is forwarded to converting and slag is passed to fuming. The results of the smelting operation are given in Table II.

Table II - Results of Polymetallic Feed Smelting with Waelz Slag Added

One burner rate Metals recovery

Lead to lead bullion Copper to matte Zinc to slag Zinc to fumes

T/h

% % % %

Chemical composition of the lead bullion Lead Copper

Chemical composition of the matte Copper Lead Zinc Iron Sulphur

Chemical composition of the slag Lead Zinc Copper

% %

% % % % %

% % %

20

88.8 92.3 75.5 5.1

96 2.1

35.4 11.4 4.3 20.0 20.0

1.2 11.4 0.4

Gas Cooling and Dusting

The dust-laden gases have a high SO2 contenet and are cooled in the waste-heat boiler to obtain high-quality steam. Then these gases are de-dusted in an electrostatic precipitator. Since the SO2 concentration in these gases runs as high as 40%, they are diluted by ventilation gases before being sent to sulphuric acid production.

CONCLUDING REMARKS

Investigations to find a possible measure to treat lead-copper-zinc feeds in the Kivcet unit have been carried out in the institute of VNIItsvetmet. The results of the investigations show that a polymetallic feed as challenging as lead-copper-zinc concentrate can be smelted by the Kivcet method, obtaining rather good results. This conclusion is validated by the pilot-scale smelting at the industrial Kivcet unit at the Ust-Kamenogorsk lead smelter owned by OAO "Kazzinc". Lead is recovered to lead bullion, copper is converted to copper matte and zinc reports to the slag. Metallic iron and copper drosses produced during lead bullion refining will upgrade the quality of the matte if added into the feed to be smelted. Sulphate products (for example, lead dusts and residues or jarosite process residues) added into the polymetallic feed mixture distinctly increase the specific production rate of the Kivcet unit. Slag-forming constituents should be kept in the following weight ratio in the feed: FeO : S1O2 : CaO = 1: (0.7-0.75): (0.4-0.5).

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 693

A MATHEMATICAL MODEL FOR THE ELECTRIC FURNACE IN THE KIVCET PROCESS

G.S. Hanumanth and G.A. Irons Department of Materials Science and Engineering

McMaster University 1280 Main Street West

Hamilton, Ontario, Canada L8S4L7

ABSTRACT

A three-dimensional mathematical model for the electric furnace of Cominco's Kivcet process was developed to compute the electric field intensity, power dissipation, temperature and velocity distributions in the furnace. A number of simulations of the electric furnace operation were carried ομΐ for various electrode immersion depths and electrode currents. The results showed that the current penetrated the bullion layer, generating high current densities in a layer approximately 0.3 m thick located in the slag/bullion interface region. An increase in the electrode immersion depth increased the furnace power as well as slag temperatures, and generated more vigorous bubble stirring and thermal homogeneity in the slag. For a given applied current, the computed power was in reasonable accord with the measured power input. The calculated temperatures were in rough agreement with the tapping temperatures of slag and bullion.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

694 LEAD-ZINC 2000

INTRODUCTION

The most widely used process for the recovery of lead from concentrates is the two-stage sinter machine-blast furnace process. The first stage of this process is the sinter-roasting of the lead concentrates, following which the roasted concentrates are conveyed to the blast furnace where they are reduced to metallic lead by the added coke. This two-stage process has been found to be thermally inefficient, and limited in its capability to process low grade ores (1). These drawbacks have encouraged efforts to develop new processes, based on the direct smelting of lead concentrates, which are thermally more efficient and also capable of handling low grade or mixed concentrates. The stricter environmental regulation in force nowadays is an additional factor that has motivated the development of new processes. Further, there is an economic incentive to process lead concentrates and zinc residues in the same smelter (2). These efforts have resulted in the development of the Kivcet process which has become a feasible alternative to the well-established sinter machine-blast furnace process for lead smelting. The Kivcet smelter consists of a reaction shaft where oxygen is used to flash smelt and desulphurize lead and zinc raw materials mixed with flux and coke, primarily while in suspension. A bed of unreacted coke is allowed to build up on the surface of the slag pool at the bottom of the shaft. As the slag being produced in the shaft trickles through this bed of coke it is subjected to further reduction. This bed of coke is commonly called the "coke-checker," and plays an important role in determining the overall rate of lead recovery. An attached electric furnace serves as a reduction unit where lead and zinc oxides remaining in the slag from the reaction shaft are further reduced by the coke added to the slag surface. The reduced lead metal settles into the bullion, whereas zinc fumes from the slag surface and is subsequently recovered. Slag and bullion are periodically tapped from the furnace.

Cominco Ltd. has adopted the Kivcet process for its lead and zinc smelter in Trail, British Columbia. The work reported in this article is a preliminary attempt to understand the fundamental aspects of the transport phenomena in the Cominco electric furnace. Relatively more attention has been paid by previous workers to the processes and reactions occurring in the smelter shaft of the Kivcet process (3,4), whereas there is little detail available on the electric furnace side of the process. Fundamental questions relating to the generation and transport of current and power in the furnace have not been addressed in a quantitative manner thus far. These include considerations such as the distribution of current between slag and bullion, and the effect of electrode geometry, immersion depth and the physical properties of the slag and bullion such as viscosity and electrical conductivity. A good understanding of these is crucial to optimize energy utilization, tapping temperatures of slag and bullion, reduction rates and furnace integrity. Previous work by Irons and coworkers (5,6) on the Falconbridge six-in-line electric furnace has clearly demonstrated the usefulness of mathematical modelling as a strategy to control furnace operation for the case of nickel smelting. Following this approach, as a first step towards understanding the complex processes in the Kivcet electric furnace, a three-dimensional mathematical model of the full scale furnace has been developed. This model was formulated in two parts: 1) an unsteady state model involving the Maxwell's equations to compute the time-averaged power distribution resulting from the flow of A.C. current in the furnace and, 2) a steady state model based on the flow conservation equation for the computation of temperature and velocity fields in the furnace. The coupling between the two parts of the model was achieved by inputting the power calculated in the first part of the model to the energy conservation equation as a heat source term. The model was run to simulate a number of different operating conditions of the full scale furnace; the simulations examined the influence of carbon monoxide bubbling caused by electrode consumption as a result of reactions with the slag, and variables such as electrode immersion depth and currents, slag and bullion electrical conductivity and viscosity. The results of the simulations discussed in this article

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 695

include the electrical power distribution in the furnace, as well as slag and bullion velocity and temperature distributions.

THE MATHEMATICAL MODEL

Electrical Power

The electrodes of the furnace are powered by a 3-phase A.C. supply. Because the three electrodes are located in line, the current flow in the bath is asymmetrical. Hence, all the three phases have to be modelled to calculate the total power input to the furnace. The mathematical model for electrical power is developed from the Maxwell's equation for magnetic flux density, B(7):

V-B = 0 (1)

This equation is reformulated in terms of the magnetic vector potential, A, in order to minimize the number of unknown variables to be solved to calculate the magnetic and electrical quantities of interest. By this approach, the Maxwell's equation (1) is transformed into the following expression (8):

V2A - μσ A = μ Jo (2) ot

where J0 is the electrode current density, and μ and σ are magnetic permeability and electrical conductivity, respectively. The electric field intensity, E, and the current density distribution, J, are related to A by the following expressions:

E = - — (3) δί

J = σ Ε (4)

In this model, the current due to convective motion of the melt has been neglected, because the magnetic Reynolds number for this system is of the order of 10"7, which means that the charge transported by convection is much smaller than the diffusive current. The electric field intensity, E, can be used to calculate the local heat release rate in the slag or bullion by Joule heating according to the equation:

P = σ Ε2 (5)

where P is the power intensity (W/m3).

Momentum and Heat Transport

The flow of heat and momentum in the electric furnace was assumed to occur under steady state conditions, so that only steady-state solutions were sought. Further, the flow was mainly laminar, but where it was turbulent the conventional k-ε model (9) was applied. Hence, the steady-state transport equations for momentum and heat were solved in three dimensions. The generalized steady-state three-dimensional equation for a conserved variable, q>, is:

696 LEAD-ZINC 2000

Α ( ρ ί / φ ) + Α ( ρ κ φ ) + Α ( ρ ^ ) + | ( ν ) + | . ( ν ) + | ( ν ) = 5ψ (6)

where φ represents the variables U,V,W,T,k and s,J+is the diffusion flux, and St is the source term.

The source term for heat transfer is the electrical power which is obtained by solving Equations (2-5) prior to the flow calculations. The solutions were obtained for one complete cycle of the A.C. current (60 Hz), and the power distribution was time-averaged over the complete cycle for each grid to yield the heat source for the flow calculations.

The source term for the vertical momentum is made up of two components. The first is the buoyancy attributable to the natural convection of the liquid, and the second is the buoyancy of the carbon monoxide bubbles generated by the reduction of oxides in the slag by carbon in the electrode:

Sw, = gpß(r-rre/) + pmag (7)

where a is the void fraction and pm is the mixture density of the slag and gas around the electrode:

Pm = (l-a)ps + apg (8)

From the measured electrode consumption rate (70 kg/h) the total carbon monoxide flow rate was calculated to be 0.161 m3/s, or 0.054 m3/s per electrode. The resulting agitated zone around the electrodes was assumed to extend to three grid layers and, from continuity considerations, void fractions of 0.02,0.02 and 0.01 were assigned to the first, second and third grid layer around each electrode, respectively. This void fraction distribution results in a slip velocity of approximately 0.5 m/s, which corresponds to a bubble size of about 5 mm which is in reasonable accord with observation. The remaining source term is the electromagnetic force, also called the Lorentz force, equal to (J x B). This force has been shown to be of the order of 1 percent of the natural convection and bubble-driven forces for large-scale smelting furnaces (6), such as the Kivcet furnace. Therefore, this force was neglected in the model.

Geometry of the Computational Cell

The computational cell was chosen to be one half of the full scale furnace partitioned at the centre by a longitudinal vertical plane. A schematic diagram of the computational cell is shown in Figure 1, where the three electrodes are denoted by El, E2 and E3. The computational cell was 12 m in length, 2.5 m in width and 1.8 m in height. Because of symmetry, all the quantities in the other half of the furnace are a reflection of those in the computational cell. A non-uniform body-fitted 12x89x14 grid system was used for the computations with finer grids at the boundaries. Tapping of bullion and slag is the only asymmetrical aspect of fluid flow; however, in the model bullion and slag were assumed to be tapped continuously from the end wall.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 697

Figure 1 - A Perspective View of the Computational Cell

The furnace has three in-line semi-graphitic carbon electrodes with a 3-phase electrical supply, designed to supply heat for the reduction reactions and to heat the slag and bullion. The electrodes have a diameter of 1 m with an inter-electrode spacing of 2.8 m, and are immersed in the slag to a depth ranging from 0.2 to 0.5 m. The thicknesses of the slag and bullion are 0.8 m and 1.0 m, respectively. Since the electrodes are in permanent contact with the slag, they react with the oxides of the slag to produce carbon monoxide gas which generates convection around the electrodes. These reducing reactions result in the erosion of the electrodes at a rate of about 70 kg/h.

The electric furnace is separated from the smelter shaft by a water-cooled partition wall. Slag and bullion produced in the shaft enter the furnace through a gap between the partition wall and the furnace. The nominal feed rates to the furnace are 22 and 16 t/h of slag and bullion, respectively. Coke is added to the slag surface at a rate of 0.4 t/h to reduce the remaining lead and zinc oxides. Lead settles into the bullion through the interface, whereas zinc fumes from the surface and is recovered subsequently. Slag and bullion are assumed to be continuously tapped from tap holes in the end wall.

Boundary Conditions

At the walls of the furnace the conventional non-slip boundary condition was used for momentum transfer, whereas constant heat fluxes, based on plant data, were prescribed at the walls for heat transfer. Details of the furnace boundary conditions for power and flow computations are provided in the following sub-sections.

Inlets and Outlets

The magnetic vector potential, A, was assumed to be zero at the inlet and outlet planes, which is equivalent to the prescription of zero current across these planes.

698 LEAD-ZINC 2000

The inlet velocities of slag and bullion were calculated from their feed rates. In order to maintain steady state conditions in the furnace, the same amounts of slag and bullion were assumed to exit the furnace from the end wall, neglecting the flow of coke and zinc whose flow rates together amounted to about 2 percent of the total. These velocities were small, each of the order of 100 μητ/s, compared to the velocities caused by the rising bubbles (0.15 m/s).

Slag Surface

For the electrical power model, the slag surface was taken to be a plane of symmetry; therefore, the gradient of A perpendicular to this plane was set to zero.

Since all the coke is added to the slag surface and tends to form a freely floating layer, the endothermic reduction reactions were assumed to occur uniformly over the entire slag surface. Accordingly, the heat sink term was calculated for these reactions and was used as a constant heat loss flux at the surface. The heat sink for the whole furnace amounted to 2.2 MW.

Bottom of the Furnace

This boundary was assumed to be a ground plane with zero magnetic vector potential. A constant heat flux of 5.5 kW/m2 was prescribed for the bottom, which is the measured cooling rate for this brick wall. For the momentum transfer, the bottom wall was regarded as a non-slip wall.

Side Wall of the Furnace

Since there was no current flow in the refractory walls, the magnetic vector potential, A, was set equal to zero. For heat transfer, a constant heat flux boundary condition, equal to the measured heat loss flux, was specified for this wall. The heat loss fluxes at the side wall for the water-cooled copper panels in the bullion and slag were 31.3 kW/m2 and 1.75 kW/m2, respectively. The conventional non-slip boundary condition was used for momentum transfer.

Slag/Bullion Interface

In the Falconbridge furnace (5,6) it was found that the interface between slag and matte was rendered immobile, probably by the presence of solid chromites. In the Cominco electric furnace, semi-solid precipitates have also been found to be present at the slag/bullion interface; therefore, as a first approximation, the interface between slag and bullion was considered to be immobile. Consequently, both sides of the interface were treated as non-slip walls. With this boundary condition, heat was transferred across the interface only by conduction. Therefore, continuity of the heat flux across the interface was prescribed. Similarly, for electrical power computation, continuity of magnetic vector potential, A, at the interface was prescribed.

PHYSICAL PROPERTIES

The physical properties of the slag and bullion at the feed temperature of 1150 ° C are listed in Table I. Relationships for the temperature dependence of density and viscosity were developed from the plant data provided by Cominco Ltd. Because of a lack of information, the magnetic permeabilities of both slag and bullion were taken to be equal to the value for free space.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 699

Table I - Physical Properties of the Slag and Bullion1·2

Property Slag Bullion

p (kg/m3) 381Ö 9620 ß(K') 8.3 xlO"5 1.2x10^ Cp(J7kg/K) 1295 138 v(m2/s) 2.6 xlO"4 9.9 xlO"8

k(W/m/K) 1.5 28 o(mho/m) 150 7.9 xlO5

μ (H/m) 1.257 x 10"* 1.257 x 10"6

1. The density-temperature relationship was p=p0exp[-ß(T-1423)];

where p0 is the density at the feed temperature of 1423 K.

2. The viscosity-temperature relationship was vs.ag= v0 exp[0.094-0.0239(T-1423)]; Vp,,;=voexp[1.2301-0.0027(T-1423)];

where v0 is the viscosity at the feed temperature of 1423 K.

COMPUTATIONAL RESULTS

The mathematical model was solved by using the commercial Computational Fluid Dynamics (CFD) code PHOENICS. The first step was to solve for the unsteady-state magnetic vector potential, A, (Equation 2). This equation was solved for one complete cycle of the A.C. current (60 Hz) using a time step of 850 μ s. To account for the electrode skin effect where the current becomes concentrated close to the electrode surface, the cross-section of the electrode was divided into two annuli, of which the outer annulus, with a width of 0.29 m, was assumed to carry 67 percent of the total current, and the inner annulus carried the remaining current. After obtaining the solutions for A, the routine use of the Equations (3-5) yielded the power distribution for each grid. This power distribution was time-averaged to calculate an average power for each grid, which served as the heat source term for the steady-state energy equation.

Computations of power, temperature and velocity distributions were made for two cases where the electrode immersion depths and currents were varied. In the first case, the three electrodes were immersed in the slag to a uniform depth of 260 mm, and the currents were 24.5, 24.7 and 21.7 kA for the electrodes El, E2 and E3, respectively. In the second case, the immersion depth was increased to 540 mm, and the currents were increased to 27.3,29.1 and 25.4 kA for El, E2 and E3, respectively. In order to represent the 3-D effects, the results are shown in two mutually perpendicular planes: (1) the longitudinal symmetric plane of the furnace (Plane-I), and (2) a cross-sectional plane (Plane-Π) which extends from the symmetric plane to the side wall, intersecting the middle electrode E2 at the centre.

Figures 2a and 2b show a selection of the power distribution contours in the Planes I and Π, respectively, for the 260 mm immersion case. Power generation was most intense in a thin layer roughly 0.3 m thick located in the slag/bullion interfacial region. In this layer, the regions directly beneath the electrodes experienced the highest power intensities. For example, the maximum

700 LEAD-ZINC 2000

power intensity, of the order of 10 MW/m3, occurred in the layer below the electrode E2. By contrast, near electrodes El and E3 the power intensity decreased to the order of 0.01 MW/m3. Power intensity was even lower, of the order of 10"4 MW/m3, in the regions far from the electrodes and near the inlet and tap holes.

\ I gi I \5.0E-4 \ 0.01

1 I E2 \l SLÄG \l E3 M

Λ io / 00' \ Z_ 1.0E-4 1.0E-4I

0.01

0.1 .0.1

BULLION

(a)

(b)

Figure 2 - Calculated Power Distributions (MW/m3) for the 260 mm Immersion Depth: Plane I and (b) Plane Π

Figures 3a and 3b show the calculated flow fields for the 260 mm immersion case. The predominant flow direction in the slag was upwards from the interface to the top and back down in a loop. This flow was driven by both natural convection and the buoyancy generated by the carbon monoxide gas produced by the electrode-slag reactions. Hence, it was strongest around the electrodes, of the order of 0.1 m/s, and decreased to the order of 0.01 m/s away from the electrodes. In addition to the principal recirculatory flow loop around the electrodes, a second weaker recirculatory loop formed below the electrodes due to the temperature difference between the electrode and the interface. By contrast, the bullion was relatively quiescent, except for a weak flow loop near the side wall, apparently in a counterintuitive direction (Figure 3b). More work is required to establish whether this loop is a real physical phenomenon or an artefact of the mesh geometry.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 701

■> 0.1 m/s

(a)

E2

< : ■ ' ■

r \ \ s.

' SLAG ■

"BULLION

•

»-0.1 m/s (b)

Figure 3 - Calculated Flow Fields for the 260 mm Immersion Depth: (a) Plane I and (b) Plane II

The temperature contours, which are coupled to the flow patterns, showed a marked difference between the slag and bullion (Figures 4a and 4b). However, they are in the range of typical plant tapping temperatures which are 1570 to 1620K for the slag, and 1320 to 1370Kfor the bullion. The slag was thermally more homogeneous than the bullion which displayed a stratified temperature distribution. This observation conforms to the strong convection that occurred in the slag and the relatively quiescent conditions in the bullion. Because convection was strongest around the electrodes, the temperature distribution was also more homogeneous there, ranging from 1460 to 1500 K in the bulk of the slag in Plane I, and from 1420 tol470 K in Plane Π. On the other hand, the bullion temperatures exhibited a stable stratified pattern with temperatures ranging from 1300 to 1550 K in Plane I, and from 1200 to 1600 K in Plane Π. This stratified temperature pattern further reinforced the stability of the bullion against convection.

Figure 4a - Calculated Temperature Distributions (K) for the 260 mm Immersion Depth: (a) Plane I

702 LEAD-ZINC 2000

Figure 4b - Calculated Temperature Distributions (K) for the 260 mm Immersion Depth: (b) Plane Π

As Figures 5a and 5b show, the general pattern of power distribution in the furnace for the deeper 540 mm immersion case was very similar to the 260 mm immersion case, although both the magnitude of the power intensity and the bullion penetration depth increased with immersion depth. At 540 mm immersion depth, the electrodes carried 12 to 20 per cent more current than at 260 mm immersion, thereby generating higher power intensity in both slag and bullion. The resulting total furnace power, which is equal to the sum of the power in each grid, increased from 3.7 MW for 260 mm immersion to 5.7 MW for 540 mm immersion.

(a)

Figure 5 - Calculated Power Distributions (MW/m3) for the 540 mm Immersion Depth: (a) Plane I and (b) Plane II

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 703

An increase in electrode immersion depth caused a large increase in slag velocity around the electrodes (Figures 6a and 6b); the maximum velocity increased by nearly 50 per cent to 0.15 m/s. In contrast, the velocities in the bullion were smaller suggesting the existence of a more stabilizing temperature gradient, compared to the 260 mm immersion case. However, although the velocity magnitudes were different, the velocity patterns were similar in both cases. Because of the increased velocities, the temperature contours of the slag and bullion also changed noticeably. For instance, Figures 7a and 7b show that increased convection rendered the slag thermally more homogeneous. For the deeper immersion, the furnace temperatures were higher, ranging from 1400 to 1700 K in the slag, and from 1300 to 1700 K in the bullion, which is consistent with the higher power input. Despite the higher temperatures in the deeper immersion case, the slag was thermally more homogeneous because of the more efficient redistribution of the heat by convection. At the same time, the bullion exhibited a remarkable horizontal temperature stratification across the entire furnace (Figure 7a), which rendered it practically motionless. A similar temperature gradient was observed in Plane Π for the bullion (Figure 7b).

-»· 0.1 m/s (a)

> 0.1 m/s

Figure 6 - Calculated Flow Fields for the 540 mm Immersion Depth: (a) Plane I and (b) Plane II

704 LEAD-ZINC 2000

(a)

Figure 7 - Calculated Temperature Distributions (K) for the 540 mm Immersion Depth: (a) Plane I and (b) Plane II

Figure 8 is a comparison of the computed and measured furnace power for different combinations of immersion depth. The measurements were provided by Cominco Ltd. The maximum and minimum immersion depths for the plant data were 300 mm and 500 mm. This compares with 260 and 540 mm for the computations. An exact match of immersion depth with the plant data could not be achieved because the electrode depth in the model can be specified only in discrete steps of size equal to the grid spacing. Therefore the closest possible immersion was selected for the computations. A reasonable agreement was observed between the calculated power and measured values for all immersions, except the deepest immersion of 540 mm. In this case, the model yielded a value of 5.7 MW, in comparison to the measured value of 3.11 MW. It is interesting to note that although the higher calculated power reflected the change in the total current carried by the electrodes, which increased by about 16 per cent for the 540 mm immersion compared to the 260 mm case, the measured values were insensitive to this large change in current.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 705

I I I

Measured Power (MW)

Figure 8 - A Comparison of the Calculated and Measured Power Input to the Furnace

DISCUSSION

A key inference that can be drawn from the power distribution results is that a significant portion of the current tends to flow in a layer of approximately 0.3 m thickness located in the slag/bullion interfacial region, with the highest current density, occurring below the middle electrode (E2). As a result, the power distribution in the furnace was nonuniform, and natural convection was most vigorous around the middle electrode (E2) where the maximum power intensity occurs. Because of this nonuniform power generation in the furnace, slag convection caused by thermal buoyancy and bubble stirring assumes a central role because of its importance in homogenizing the slag temperature.

As the immersion depth increases so do the current and power to the furnace. This result has important implications for the safe operation and integrity of the furnace, because it can lead to dangerous levels of bath overheating. In the simulations where the bath cooling rate and throughput were constant, lowering the electrodes in the slag led to higher slag and bullion temperatures as the results demonstrate. At the same time, convection in the slag became more vigorous because a larger electrode surface became exposed to the slag, leading to higher rates of reduction and carbon monoxide generation. Although convection in the slag reduces the risk of bath overheating by redistributing the heat, in some cases it might not be sufficient to maintain a safe bath temperature. An even greater risk is that, since electrical conductivity of the melt increases with temperature, the temperature rise may accelerate at an ever increasing rate, quickly leading to a runaway situation in the furnace. To avoid this situation, the cooling rates at the walls must be recalculated and accordingly adjusted to remove the excess heat. In its present form, the model treats electrical conductivity as constant and hence cannot predict bath overheating accurately. So, for a more rigorous model, the temperature dependence of the electrical conductivity must be incorporated and Equation (2) solved iteratively for every temperature field.

706 LEAD-ZINC 2000

The homogenization process was not as effective in the bullion, and the bullion remained thermally stratified as the results demonstrate (Figures 4 and 7). According to the model the interface is a motionless plate, which means that it acts a barrier to the transfer of momentum from the slag to the bullion. In the absence of convection, heat conduction becomes the predominant method of heat transfer in the bullion. There may be regions near the side walls of the furnace where convection currents develop because the side walls are water-cooled. However, this is a local effect, and is unlikely to influence the overall temperature distribution in the bullion. This is convincingly illustrated by Figures 4a and 7a where thermal stratification of the bullion is clearly noticed; further, at the deeper immersion, the thermal stratification became more strongly established and spreads to the whole furnace. The mechanism that establishes this thermal stratification develops because the slag/bullion interface and the water-cooled furnace bottom wall form a differentially-heated cavity filled by the bullion. With deeper immersion, the temperature difference of the cavity increased and consequently a more stable thermal gradient became entrenched. Since this gradient opposes natural convection which would have aided the homogenization of the bullion, the alternative method to homogenize the bullion is to change the properties of the interface to permit momentum transfer across the interface. A major reason for the rigidity of the interface between the slag and bullion is the precipitation of semi-solid substances like magnetite (commonly called smag in plant terminology) in the slag. In the plant, this can be reduced by maintaining proper temperature control and eliminating cold spots in the slag. In the model, the rigid plate at the interface will have to be replaced by a mobile interface that permits momentum transfer to the bullion. Such a model is expected to provide more realistic simulations of the flow and temperature distributions in the bath. In addition to a mobile interface, two other features can be incorporated to further improve the model: first, a better way to calculate electrical variables; and second, a procedure to calculate the power factor which would give more accurate values of power input to the furnace. The latter may indeed help to explain the large difference in the calculated and measured power which was observed for the 540 mm immersion, because the power factor is likely to decrease as the electrodes approach the bullion.

CONCLUSIONS

The findings of this study can be summarized as follow. The electrical current penetrates the bullion, generating high power intensities in a layer approximately 0.3 m thick located in the slag/bullion interfacial region. The furnace power input increases with the electrode immersion depth, and the slag and bullion temperatures also increase with the electrode immersion depth. Thermal homogeneity of the slag improves with electrode immersion depth because of the more vigorous bubble stirring.

ACKNOWLEDGEMENTS

The authors are very grateful to Dr. J. Dableh for his help in developing the electrical model of the furnace. The many fruitful discussions with Dr. G.G. Richards and Mr. A. Hall of Cominco Ltd. on all aspects of the electric furnace operation are greatly appreciated. The financial support of CANMET and Cominco Ltd. for this project is also gratefully acknowledged. Finally, the authors would like to thank Cominco and Snamprogetti for reviewing the final paper.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 707

NOMENCLATURE

A Magnetic vector potential [Volt-sec/m] B Magnetic flux density [Weber/m2] Cp Specific heat [J/kg/K] E Electric field intensity [Volt/m] g Gravitational acceleration [m/s2] J Current density [Ampere/m2] P Power intensity [W/m2] U Velocity in X-coordinate direction [m/s] V Velocity in Y-coordinate direction [m/s]

W Velocity in Z-coordinate direction [m/s] T Temperature [K] k Thermal conductivity [W/m/K]

also Turbulent kinetic energy [m2/s2] ε Energy dissipation rate [m2/s3] a Void fraction μ Magnetic permeability [Henry/m] σ Electrical conductivity [ohm"1 m"1] β Coefficient of thermal expansion [K1] v Kinematic viscosity [m2/s]

REFERENCES

A.G. Matyas and P.J. Mackey, " Metallurgy of the Direct Smelting of Lead," Journal of Metals. Vol. 28. 1976,10-15.

D.W. Ashman, " Pilot Plant and Commercial Scale Test Work on the Kivcet Process for Cominco's New Lead Smelter," J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton and P. Hancock, Eds., The Metallurgical Society of CIM, Montreal, Canada, 1998, 783-794.

K.B. Chaudhuri and G. Melcher, "Comparative View on the Metallurgy of the Kivcet-CS and other Direct Lead Smelting Processes."CIM Bulletin. Vol. 78, 1978, 126-130.

4. Y.I. Sannikov, M.A. Liamina, V.A. Shumskij and Y.A. Grinin, "A Physical and Chemical Description of the Kivcet Lead Flash Smelting Process," CIM Bulletin. Vol. 91,1998,76-81.

5. Y.Y. Sheng, G.A. Irons and D.G. Tisdale," Transport Phenomena in Electric Smelting of Nickel Matte: Part I. Electric Potential Distribution," Met. Mater. Trans. B. Vol. 29B, 1998, 77-83.

6. Y.Y. Sheng, G.A. Irons and D.G. Tisdale," Transport Phenomena in Electric Smelting of Nickel Matte: Part II. Mathematical Modelling," Met. Mater. Trans. B. Vol. 29B, 1998,85-94.

7. J. Szekely, Fluid Flow Phenomena in Metals Processing. Academic Press, New York, U.S.A., 1979, 178-203.

8. J. Dableh," Formulation for the Electromagnetic Field Problem in an Electrical Furnace," Internal Report. April 1998, LRT Technologies Inc., Mississauga, Ontario.

9. B.E. Launder and D.B. Spalding, " The Numerical Computation of Turbulent Flows," Computer Methods in Appl. Mech. Eng.. Vol. 3, 1974, 269-289.

Chapter 9

New Zinc Processing Technologies

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 711

ATMOSPHERIC LEACHING OF ZINC SULPHIDE CONCENTRATES USING REGENERATED FERRIC SULPHATE SOLUTIONS

C. J. Ferron Lakefield Research Limited

185 Concession Street P.O. Box 4300

Lakefield, Ontario, Canada KOL 2HO

ABSTRACT

Most of the world's zinc metal is produced via the roast- leach-electrowinning (RLE) process. The process is well established, but most of the sulphur in the sphalerite concentrates reports in the acid, which has to be stored and/or sold. There are significant incentives to develop a simpler process with much reduced acid generation. Ferric ion is a well known oxidant for most sulphides, in particular sphalerite, and it converts the sulphide to elemental sulphur. Based on stoichiometry, however, the amount of ferric ion needed to achieve acceptable zinc extractions is prohibitive for downstream zinc recovery from the leach solutions. The process suggested here makes use of less-than-stoichiometric ferric sulphate additions because the ferric ion is regenerated by sparging SO2/O2 mixtures. This regeneration can be effected in-situ in the leach vessel or outside the leach vessel. An example of the application of this new process is presented for a Canadian sphalerite concentrate assaying 60% Zn. The effects of temperature, ferric concentration and ferric ion regeneration are presented. Options to recover zinc from the leach solution are also briefly discussed.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

712 LEAD-ZINC 2000

INTRODUCTION

The large majority of the world's zinc is produced by the RLE (roast-leach-electrowinning) process. The process is well established but during the roasting step, most of the sulphide sulphur in the sphalerite concentrate reports as SO2 gas, which must be captured and converted to sulphuric acid, which has to be sold or stored. Therefore, there is a compelling incentive to develop processes which do not transform sulphide sulphur into acid. One such process is the medium temperature pressure oxidation process first commercialised by Cominco in Trail, B.C. During the pressure leaching step(s), most of the sulphide is oxidised by oxygen to elemental sulphur, according to the reaction:

ZnS + H2SO4 + I/2O2 -> ZnS04 + H20 + S° (1)

To compensate for the low efficiency of gaseous oxygen as an oxidant, medium temperatures (150-160°C) and oxygen overpressures are required (1).

Atmospheric processes have also been studied. As examples, the FCL process (Ferric Chloride Leach) extensively investigated and developed by CANMET (2), and a nitrate-based process (3) could be cited. The ability of ferric ions to oxidise various sulphide minerals, in particular sphalerite, has been known for years, and this was the basis for the FCL process. One of the perceived difficulties of the FCL process is the use of a chloride medium, and there are several incentives to operate in a sulphate system, not the least of which is the fact that the sulphate system is extremely well established.

Dutrizac has reviewed (4) the dissolution of zinc sulphide minerals in acidified ferric sulphate solutions. The results suggested that the important parameters involved are temperature and ferric ion concentration, and that the dissolution rates are rapid, at least near the solution boiling point. The reaction involved when using ferric sulphate solutions can be written as:

ZnS + Fe2(S04)3 -> ZnS04 + 2FeS04 + S° (2)

Based on stoichiometry alone, the reaction indicates that to dissolve 20 g/L Zn, at least 34.2 g/L Fe3+ is required. Taking into account the excess reagent needed to drive the reaction, it is obvious that the final leach solution would be very difficult to treat with present technologies. A reoxidation of the ferrous ion is required to reuse the same iron atom. In the case of the chloride system, an obvious oxidant is chlorine. In sulphate systems, ferrous sulphate could be reoxidised to the ferric state using oxygen, according to the reaction:

2 FeS04 + H2S04 + Vi 02 -> Fe2(S04)3 + H20 (3)

The kinetics of this reaction are fairly slow, however, and a faster but cheaper oxidant is desirable. The new process presented here proposes to reoxidise ferrous sulphate using S02/02

mixtures, as suggested by the reaction:

2FeS04 + S02 + 0 2 -> Fe2(S04)3 (4)

The ability of S02/02 mixtures to oxidise iron and various other ionic species is well known (5), and a similar process for the reoxidation of ferrous ions during the leaching of copper sulphide minerals has been recently published (6).

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 713

A study has been presented on the reoxidation of ferrous ions during the ferric ion leaching of sphalerite, using SO2/O2 mixtures (7), but it was a somewhat fundamental study, under conditions too distant from industrial application. Direct leaching of sphalerite with SO2/O2 mixtures has also been proposed (8).

The object of this paper is to describe a process developed to treat a Canadian zinc concentrate. The first part of the paper examines the ferric sulphate leaching of the concentrate. The second part examines the effects of ferrous oxidation by SO2/O2; this reoxidation can be effected in-situ, i.e., during the leach, or ex-situ in a separate vessel. The third part briefly discusses options to treat the resultant leach solution and recover the zinc from it.

DESCRIPTION OF THE SAMPLE TESTED

The sphalerite tested was a flotation concentrate produced in a Canadian operation. Typical assays are presented in Table 1. The sphalerite was high grade, and contained very little iron.

Table I - Analysis of the Sphalerite Flotation Concentrate

Element Assay % Zn 60.75 Pb 1.13 Fe 0.40 S 28.3

Cd 0.27

EXPERIMENTAL RESULTS

Ferric Sulphate Leaching of the Sphalerite Concentrate

A series of experiments was conducted to examine the response of the sphalerite concentrate to ferric ion leaching. The parameters varied were retention time, temperature and the initial ferric ion concentration.

Proof-of-Principle

A few leach tests were conducted using ferric ion, either as chloride or sulphate, as the oxidant. Typical results are summarised in Table 2 and in Figure 1.

714 LEAD-ZINC 2000

Table II - Typical Leach Results Obtained on the Sphalerite Concentrate. No Grind, Temperature 95°C, 6 hours, 5% Solids

Test# g/L Ferric Medium % Zn Ext'n A B C D

60 104 112 105

Chloride Chloride Sulphate Sulphate

98.8 99.6 97.7 97.3

400

Figure 1 - Ferric Ion Leaching of the Sphalerite Concentrate. Curve 1: 100 g/L Ferric Ion (as Chloride); 95°C, - 5% Solids. Curve 2: 60 g/L Ferric Ion (as Sulphate); 95°C, 5% Solids

The results indicate that, although the chloride system has better kinetics, both systems produce zinc extractions greater than 97% in 6 hours.

A sample of partially leached sphalerite from the ferric sulphate system was examined under the microscope. The photomicrophotograph is presented in Figure 2. Although not perfectly clear, the photograph shows a layer of reaction product around the partially leached sphalerite. Most likely, this layer is mostly composed of elemental sulphur.

Effect of Ferric Ion Concentration

The next series of tests examined the effect of varying the initial ferric ion concentration, all other parameters being kept constant. All tests were conducted at 80°C. The results are illustrated in Figure 3.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 715

Figure 2 - Photomicrophotograph of Partially Leached Sphalerite (Sulphate System)

3 I

30 i

25 -ί

20 i

15 ■'

10 1/ 5 r n L U Ψ

0

. ^ ^ ^ j

1 ■ 1 1

100 200 300 400

Time (min)

-60g/L

-30g/l

Figure 3 - Effect of the Ferric Ion (as Sulphate) Concentration on Zinc Extraction. 80°C, 5% Solids

716 LEAD-ZINC 2000

As shown in reaction (2), zinc dissolution parallels ferric ion consumption, and once the ferric ion is consumed, the leach reaction stops. In fact, residual ferric ion concentration is a direct measure of the extent of the reaction, and redox potentials could be used to monitor zinc extractions, as suggested by Figure 4.

1

600

500

400

300

200

100

0

-60g/L;

-30 g/L,

100 200 300 400 Time (min)

Figure 4 - Pulp Redox Potential as a Function of the Initial Ferric Ion Concentration. 80°C, 5% Solids

Effect of Temperature

Five additional tests were conducted to examine the effect of temperature on the zinc extraction kinetics. The results are presented in Figure 5. All tests were conducted with an initial concentration of 60 g/L Fe3+ (as sulphate).

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 717

100 200 300

Time (min)

400

-80°C "95°C -60°C "40°C "25°C

Figure 5 - Effect of Temperature on Zinc Extraction Kinetics. 60 g/L Fe + Initial, 5% Solids

Kinetics of the Ferric Sulphate Leaching of Sphalerite

In order to determine the mechanism of the process, extraction data were plotted as a function of the square root of time. Typical results are presented in Figure 6.

5 10 15 20

Square Root of Time (min)

■80°C

■60°C

■25°C

Figure 6 - Extraction of Zn as a Function of the Square Root of Time

718 LEAD-ZINC 2000

The data indicate that the process follows, at least in the initial stages, a parabolic law of the type: extraction = k Vt, where t is time and k is the rate constant.

Arrhenius Plot

The data presented in Figure 4 were plotted as a function of the square root of time, and the various slopes of the resulting lines, the rate constants, were plotted against 1000/T, where T is the absolute temperature in degrees Kelvin. This Arrhenius plot is presented in Figure 7. Using the Arrhenius relation k=k° exp(-E°/RT), the apparent activation energy E° for the process was calculated to be 36.4 kJ/mol (or 8.7 kcal/mol). In the temperature range investigated (25-80°C), the correlation was excellent, with a R2 of 0.999.

i -

y = 1 S.399 - 4.3744X RA2 = 0.999

1 1 1 1 1 1 1 1 ■ 1 ■—

2.8 2.9 3.0 3.1 3.2 3.3 3.4

1000/K

Figure 7 - Arrhenius Plot for Ferric Ion Leaching of Sphalerite. Initial 60 g/L Fe .3+

Regeneration of Ferric Sulphate Using S02/02 Mixtures

The regeneration of ferric sulphate using SO2/O2 mixtures is known and has been described in the literature (9, 10). The reaction proceeds as indicated in equation (4). Typical experimental results are illustrated in Figure 8.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 719

120.0

100.0 V, No pH Control

pH Control

100 200

Time (min)

300

Figure 8 - Oxidation of Ferrous Sulphate Solutions using SO2/O2 Mixtures. Temperature: 60°C

As described elsewhere in more detail (9), the reoxidation of ferrous sulphate using SO2/O2 mixtures depends on several parameters such as temperature, the % SO2 in the gas mixture, pH and the presence of some catalytic ions in solution. In less than four hours, 100 g/L of ferrous ion were reoxidised, corresponding to an average production of more than 25 g Fe3+/L/h. With pH control, in one hour, more than 40 g Fe3+/L were produced, before the reaction stopped.

The Regenerated Ferric Sulphate Leach Process (RFSL)

The process described in this paper combines the two reactions discussed,in the prior sections; i.e., the ferric sulphate leach of sphalerite and the reoxidation of the ferrous sulphate generated during the leach, to be able to continue the leach.

The regeneration of the ferric sulphate can, in practice, be effected in two ways. Firstly, it can be effected in the same vessel where the leach is operating; this option is called the direct, or in-situ, process. The second option is to effect the regeneration of ferric sulphate in a vessel different than the leach vessel, on the solution only; this option is named the indirect, or ex-situ, process.

720 LEAD-ZINC 2000

The In-Situ Regenerated Ferric Sulphate Leach (RFSL) Process

In this process, schematically described in Figure 9, the leaching of sphalerite and the regeneration of the ferric ion occur in the same vessel within the pulp containing the leach solution and the sphalerite. In this scenario, reactions (2) and (4) occur simultaneously to produce the overall reaction 5.

ZnS + Fe2(S04)3 -*■ ZnS04 + 2FeS04 + S°

2FeS04 + S02 + 02 -*■ Fe2(S04)3

ZnS + S02 + 02 -> ZnS04 + S°

(2)

(4)

(5)

Obviously, both reactions 2 and 4 will proceed with their rate constants k2 and Lt, with the overall rate constant ks = k2 x Lt.

so2/o2

H2S04 -

ZnS

Fe 3+

■>■ Leach pulp ► To Zn recovery from solution

Figure 9 - Schematic of the In-Situ Regenerated Ferric Sulphate Leach of Sphalerite (RFSL)

The in-situ RFSL process was applied on the sphalerite concentrate described in this paper. Figure 10 illustrates the kinetics of Zn extraction as a function of time with and without the use of the S02/02 mixture, when starting with a 30 g/L ferric ion solution, at 80°C.

Using ferric sulphate without regeneration, the leach reaction starts following the parabolic law, but then slows because ferric ion has been consumed and the driving force to leach sphalerite has decreased. After 6 hours, 49% of the zinc has been extracted, and the shape of the curve indicates that the reaction has practically stopped.

When regenerating the ferric sulphate in-situ with S02/02 mixtures, the initial extraction after 1 hour is the same, but the reaction does not slow down as much as without regeneration. The leaching proceeds at an acceptable rate to the point where after 6 hours it has reached 82% Zn extraction, and the shape of the curve indicates that dissolution is still proceeding.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 721

25

0 100 200 300

Time (min)

Figure 10 - Example of the Application of the In-Situ RFSL for the Sphalerite Concentrate. Temperature: 80°C, Initial Ferric Ion: 30 g/L, 4 % Solids

The Ex-Situ RFSL Process

In this scenario, the two main reactions (2) and (4) occur in separate vessels, as schematically illustrated in Figure 11. The number of leach stages can vary according to the iron concentration in the circuit. From an engineering point of view, one would try to use as few leach stages as possible, but that implies beginning with high concentrations of iron in the circuit. This makes the subsequent recovery of zinc from solution with the present state-of-the-art technology very difficult. Alternatively, the iron has to be sacrificed and considered a consumable, an assumption which would affect the costs.

722 LEAD-ZINC 2000

ZnS

SO2/O2

Leach pulp ► To Zn recovery from solution

Figure 11 - Schematic of the Ex-Situ Regenerated Ferric Sulphate Leach of Sphalerite (RFSL)

Discussion

General - Ferric Ion Leaching of Sphalerite

The results illustrated in Figure 1 confirm that ferric ion is an effective oxidant to leach sphalerite and, although ferric sulphate is slower than ferric chloride, it still provides acceptable extraction rates in a medium which is fully compatible with present industrial practice. The photomicrograph presented in Figure 2 and the parabolic law response of the system (Figure 6), at least during the initial part of the leaching process, indicate that the rate determining step is diffusion through a barrier of the elemental sulphur reaction product. Later, an additional phenomenon occurs and the rate is controlled by the ferric concentration.

The effect of temperature is obvious from Figure 5. From the Arrhenius plot presented in Figure 7, an apparent activation energy of 36.4 kJ/mol (8.7 kcal/mol) is calculated, consistent with control by diffusion through a solid reaction product, when excess reactant (ferric ion) is present.

Calculating the reaction rates from the data of Figure 3, it is observed that:

k+ = k° + a [Fe3+] = 8 + 0.124 g/L ferric (g Zn /L/min05) (6)

where [Fe +] represents the initial concentration of ferric ion in solution between 10 and 60 g/L Fe3+.

Edited by JE. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 723

Ferric Regeneration with SO?/Q? Mixtures

A more detailed investigation of the rate expression for ferric sulphate regeneration with SO2/O2 will be presented later. Based on these results and data already published, the reaction rate for ferric ion regeneration with SO2/O2 mixtures can be expressed as:

rate = k[S032"]a[S02]b[H2S04]c exp [— ] (7)

RT

where [] expresses concentrations in g/L for SO32", H2SO4, and % SÜ2(v/v) for SO2 gas; the coefficients a, b, c being 1.06, 2.1 and -2.5, respectively, and E the apparent activation energy being 6.7 kJ/mol (1.6 kcal/mol)

The RFSL Process

The process is a combination of two sub-processes which can be operated simultaneously (in-situ) or in sequence.

At first glance, it would appear that the in-situ method offers technical advantages over the ex-situ process, since it would require less equipment. The fact, however, is not as simple as it would first appear. The two sub-processes (ferric leaching-ferric regeneration) have optimum values of some parameters (in particular temperature) which are not necessarily the same. For temperature, as an example, it has been shown that the higher the temperature, the better the rate for the ferric ion leaching of sphalerite. This is not true for ferric ion regeneration, where it has been shown (11) that the temperature effect on the oxidation rate passes through a maximum around 80°C, because of the decreased gas solubility at higher temperature. Because of this situation, there could be some advantages in using the ex-situ process, since it will result in a much easier control strategy.

Zinc Recovery from RFSL Process Solutions

Iron-zinc separation from solutions has always presented a challenge to the hydrometallurgist. In industrial practice, iron and zinc are separated by selective precipitation of an iron (III) precipitate from zinc. Various processes are or have been practised commercially depending on the form of the final iron precipitate (jarosite, goethite, para-goethite and hematite). Apart from the inherent difficulties of each of these processes, iron is lost in those systems; it cannot be recycled and therefore becomes a consumable, and this is likely cost prohibitive in the process considered here.

Iron-zinc separation by solvent extraction has been suggested (12), and although some of the results are quite promising, there has not yet been an industry acceptance of the idea, probably because of the fear of organic making its way through the tankhouse. Recently, a new reagent had been proposed to effect the Fe/Zn separation by SX, and the initial results were quite promising (13); however, the reagent was later abandoned due to stability problems (14). Zinc has been and is presently electrowon from solutions generated by solvent extraction using DEHPA as the extractant (15), and therefore, the acceptance of solvent extraction in zinc metallurgy is widening. In fact, several alternatives are available to the metallurgist to effect the iron-zinc separation using solvent extraction.

724 LEAD-ZINC 2000

Extraction of Zinc

In this scenario, zinc is selectively extracted from iron, which is left behind and can be re-oxidised/recycled to the leach, or can be bled from the circuit. A typical reagent would be DEHPA, which is relatively selective for zinc over iron provided the iron is in the ferrous state. The ferric ion still present in the system will be co-extracted strongly and can only be stripped with a very acidic solution (6N HCl for example). As an alternative to acidic regeneration, iron can be removed from the loaded organic using reductive stripping, with Zn powder as the reductant (16).

Extraction of Ferric Ion

In this scenario, ferric ion is selectively extracted from iron-zinc solutions (17,18, 19) prior to Zn recovery, for example using MEHPA.

CONCLUSIONS

The test program described in this paper has confirmed that ferric sulphate is a suitable leaching agent for a Canadian sphalerite concentrate. Experimental data indicated that the leaching proceeded according to a parabolic law of the type: Zn extraction = kVtime, at least during the initial portion of the leach. Under this scenario, the leach mechanism is controlled by a diffusion process through a resistance layer of reaction product. This mechanism has been confirmed by SEM photographs of partially leached sphalerite grains and by the Arrhenius activation energy measured at 36 kJ/mol (8.7 kcal/mol) between 25 and 80°C. The parabolic rate constant k was shown to vary linearly with ferric concentration in a relation k=k° + 0.18 [g/L Fe3+], between 10 and 60 g/L Fe3+.

The novel process proposed here (RFSL) involves regeneration of the ferric ion using SO2/O2 mixtures. The regeneration can be effected directly during the leach (in-situ process), or in a separate vessel on a clear liquor (ex-situ process). In-situ regeneration tests indicated a 33% absolute increase in zinc extraction in 6 hours under otherwise similar conditions. Ex-situ regeneration of ferric sulphate was successful and proceeded according to a well established mechanism.

The challenge for the hydrometallurgist resides now in effecting a good iron-zinc separation from the leach solution. Several solvent extraction options appear feasible but need to be demonstrated and integrated within the global process.

REFERENCES

1. D.B. Dreisinger and E. Peters, "The Oxidation of Ferrous Sulphate By Molecular Oxygen Under Zinc Pressure-Leach Conditions", Hydrometallurgv. 22,1989,101-119.

2. W.J.S. Craigen and Canmet/MSL Staff, "The CANMET Ferric Chloride Leach Process for the Treatment of Bulk Base Metal Sulphide Concentrates", MSL Division Report. MSL 89-67, June 1989.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 725

3. R.W. Adams, P. Mangano, E.G. Roche and S.J. Carpenter, "Direct Leaching of Zinc Concentrates at Atmospheric Pressure", Lead-Zinc 90, T.S. Mackey and R.D. Prengaman, Eds., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1990, 351-372.

4. J.E. Dutrizac and R.J.C. MacDonald, "Ferric Ion as a Leaching Medium", Minerals Sei Eng.. Vol. 6, No. 2,1974, 59-100.

5. E.A. Devuyst, A. Mosoiu and E. Krause, "Oxidising Properties and Applications of the SO2-O2 System", Hydrometallurgy Research, Development and Plant Practice. K. Osseo-Asare and J.D. Miller, Eds., The Metallurgical Society of AIME, Warrendale, PA, U.S.A., 1982, 391-403.

6. C.J. Ferron, "New Atmospheric Leach Process for Copper Sulphide Ores and Concentrates", Copper '99 Cobre '99, Volume IV, The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1999,151-165.

7. F. Nami, "The Kinetics of Zinc Sulphide Leaching by Oxygen, Sulphur Dioxide and Ferrous Sulphate", Ph.D. Thesis. Columbia University, New York, NY, U.S.A., 1985.

8. R.W. Adams and I.G. Matthew, "Leaching of Metal Sulphide Concentrates at Atmospheric Pressure using SO2/O2 Mixtures", Proc. Australasian Institute of Mining and Metallurgy, No. 280,1981. 41-53.

9. C.J. Ferron, D.O. Kwateng and P.F. Duby, "Kinetics of the Precipitation of Goethite from Ferrous Sulphate Solutions using Oxygen - Sulphur Dioxide Mixtures", paper presented at the TMS Meeting. New Orleans, Feb. 1991, 165-177.

10. E. Krause, "The Oxidation of Ferrous Sulphate Solutions by Sulphur Dioxide and Oxygen", Ph.D. Thesis. University of Waterloo, Waterloo, Ontario, 1988.

11. B.L. Tiwari, J. Kolbe and H.W. Hayden Jr., "Oxidation of Ferrous Sulphate in Acid Solution by a Mixture of Sulphur Dioxide and Oxygen", Met. Trans. B. Vol. 10B, 1979, 607-612.

12. D. Flett and A.J. Monhemius, "Solvent Extraction for Iron Control in Hydrometallurgy.", Iron Control and Disposal, J.E. Dutrizac and G.B. Harris, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1996, 331-356.

13. D. Cupertino et al., "Iron Control in Zinc Circuits: the Role of Highly Selective Zinc Extractants", Iron Control and Disposal. J.E. Dutrizac and G.B. Harris, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1996, 369-379.

14. J.M. Hearne, R. Haegele and R.D. Beck, "Hydrometallurgy Recovery of Zinc from Sulphide Ores and Concentrates", Zinc and Lead Processing, J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton and P. Hancock, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 765-780.

15. G.D. Noguera, Tecnicas Reunidas, Personal Communication, Madrid, Spain, 1998.

726 LEAD-ZINC 2000

16. C. Chang, H. Gu and T.J. O'Keefe, "Galvanic Stripping of Iron from Solvent Extraction Solutions from Zinc Residues Leaching", Iron Control and Disposal. J.E. Dutrizac and G.B. Harris, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1996, 417-428.

17. V. Lakshmanan, N. Ramie and B. Monzyk, "Evaluation of N-alkylhydroxamic Acids for Selective Iron Separation from Zinc Process Liquors for High Purity Iron Products", Iron Control and Disposal. J.E. Dutrizac and G.B. Harris, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1996, 357-367.

18. F. Delmas et al., "Novel Highly Efficient Selective Extractants for Iron in Zinc Hydrometallurgy", Iron Control and Disposal. J.E. Dutrizac and G.B. Harris, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1996, 381-393.

19. G. Demopoulos, R. Molnar and L. Rosato, "Bench scale and Mini-Plant Investigations on the Selective Removal of Iron From Zinc Process Solutions by Solvent Extraction", Iron Control and Disposal. J.E.Dutrizac and G.B. Harris, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1996, 395-416.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 727

COMPARISON OF DIRECT PRESSURE LEACHING WITH ATMOSPHERIC LEACHING OF ZINC CONCENTRATES

K.R. Buban, M.J. Collins, I.M. Masters and L.C. Trytten Dynatec Corporation

Metallurgical Technologies Division 8301-113 Street

Fort Saskatchewan, Alberta, Canada T8L 4K7

ABSTRACT

Both pressure leaching of zinc concentrate, in an autoclave, and atmospheric leaching, in stirred tanks, have been employed in the expansion of zinc production at plants employing roast-leach-electrowin technology. These direct leaching methods result in the conversion of the sulphur in the feed to the elemental form, rather than to sulphur dioxide, decoupling zinc production from acid production. Atmospheric and pressure leaching methods are also readily implemented over a wide range of production rates. In choosing the best technology for expansion, capital and operating costs, zinc recovery, byproduct quality and the compatibility of recycle streams with the existing refinery must be compared. Leach testwork with several zinc concentrates using both methods has been performed by Dynatec. Recent results are presented, along with a discussion of their implications for process selection.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

728 LEAD-ZINC 2000

INTRODUCTION

The direct pressure leaching of zinc sulphide concentrates in spent electrolyte with oxygen, at elevated temperature and pressure in continuous horizontal multi-compartment autoclaves, has been proven to be a commercially viable route for expanding zinc production at zinc refineries employing roast-leach-electrowin technology. In 1981, the world's first commercial zinc pressure leach facility commenced operations at Cominco's zinc refinery at Trail, British Columbia, Canada (1). This was followed in 1983 by a pressure leach plant at the Kidd Creek (now Falconbridge Timmins Division) zinc refinery in Timmins, Ontario (2). A third plant was installed at the Ruhr Zink operations in Datteln, Germany and commenced production in 1991 (3). However, this pressure leach plant was shut down in 1994 following a decision by Ruhr Zink to decrease zinc production. In 1997, Cominco commissioned a new, larger pressure leach autoclave to meet expansion objectives at their Trail zinc refinery (4). A fourth pressure leach plant employing Dynatec's zinc pressure leach process was commissioned in Flin Flon, Manitoba in 1993. In this operation, two-stage countercurrent leaching obviates the need for roasters (5). Table I summarizes the zinc pressure leach facilities integrated with roast-leach-electrowin refineries.

Table I - Zinc Pressure Leach Plants Integrated with Roast-Leach-Electrowin Zinc Refineries Company

Cominco Falconbridge Ruhr Zink Cominco

Location

Trail Timmins Datteln Trail

Startup Year 1981 1983 1991 1997

Zinc Input Design, t/y

30 000 20 000 50 000 75 000

Zinc Input Actual, t/y

48 000 -25 000

55 000 n.a.

Autoclave Status

Not in service In operation Shut down In operation

Many other potential integration applications have been investigated at Dynatec, including the treatment of low-grade zinc and bulk zinc-lead or zinc-copper concentrates.

Lower temperature and atmospheric pressure direct leaching of zinc sulphide concentrates has also been investigated in considerable detail (6), and there are two commercial facilities in operation, integrated with roast-leach-electrowin operations. The first is located at the Korea Zinc refinery in Onsan, Korea, and the second, at Outokumpu's zinc refinery in Kokkola, Finland (7). Similar to the pressure leach plant installations described above, both of these facilities were installed to expand zinc production. Table II summarizes the commercial zinc atmospheric pressure leach facilities.

Table II - Zinc Atmospheric Leach Plants Integrated with Roast-Leach-Electrowin Plants Company Location Startup Year Zinc Input Zinc Input Status

Design, t/y Actual, t/y Korea Zinc Onsan 1994 160 000 150 000 In operation Outokumpu Kokkola 1998 50 000 n.a. In operation

Both pressure leaching and atmospheric leaching processes convert the sulphide sulphur in the feed zinc concentrate to elemental sulphur, rather than to sulphur dioxide, decoupling the

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 729

expansion in zinc production from sulphuric acid production. In choosing the best technology for expansion, the zinc producer must compare capital and operating costs, zinc recovery, byproduct recovery and quality, and the compatibility of process streams with the existing refinery.

PROCESS CHEMISTRY

Pressure Leach

The zinc pressure leach process depends upon a simple net reaction.

ZnS + H2SO4 + 0.5 0 2 -> ZnS04 + H20 + S° (1)

This reaction is slow in the absence of an oxygen transfer agent, and dissolved iron is effective. The process is more correctly depicted by two reactions.

ZnS + Fe2(S04)3 -> ZnS04 + 2 FeS04 + S° (2)

2 FeS04 + H2S04 + 0.5 0 2 -> Fe2(S04)3 + H 20 (3)

There is usually sufficient acid soluble iron in zinc concentrates to supply the needs of the leach. The oxidation reaction of iron in pyrrhotite (FeySs) or iron in sphalerite ((Zn,Fe)S) is similar to that of zinc in sphalerite.

The extent of pyrite (FeS2) oxidation depends upon a number of leaching conditions. Under strongly oxidizing conditions and at high temperatures, oxidation of pyrite will result in sulphate generation.

2 FeS2 + 7.5 02 + H20 -> Fe2(S04)3 + H2S04 (4)

2 FeS2 + 7.5 02 + 4 H20 -> Fe203 + 4 H2S04 (5)

With lower oxygen availability, lower temperature and higher acid concentration, oxidation of pyrite may result in some elemental sulphur production.

FeS2 + 0.5 02 + H2S04 -> FeS04 + H20 + 2 S° (6)

Copper, usually present as chalcopyrite (CuFeS2), also leaches.

CuFeS2 + 0 2 + 2 H2S04 -> CuS04 + FeS04 + 2 S° + 2 H20 (7)

Galena (PbS) reacts readily during the leach to form lead sulphate at higher sulphuric acid concentrations, or lead jarosite at lower sulphuric acid concentration.

PbS + H2S04 + 0.5 0 2 -> PbSC-4 + S° + H 20 (8)

PbS04 + 3 Fe2(S04)3 + 12 H 20 -> PbFe6(S04)4(OH)i2 + 6 H2S04 (9)

730 LEAD-ZINC 2000

Usually, much less than 10% of the nonpyritic sulphide sulphur in the concentrate is oxidized to sulphate sulphur during the pressure leach.

MS + 2 0 2 -> MS04 (10)

Where M represents Zn, Pb, Cu or Fe.

Hydrolysis reactions remove iron from solution with the formation of hydrated ferric oxide or hydronium jarosite and result in the regeneration of some sulphuric acid.

Fe2(S04)3 + (x+3)H20 -> Fe203xH20 + 3 H2S04 (11)

3Fe2(S04)3 + 14H20 -+ 2 H3OFe3(S04)2(OH)6 + 5 H2S04 (12)

The zinc pressure leach process is amenable to treating a variety of feedstocks, including low-grade, high-iron zinc sulphide concentrates, bulk Pb-Zn concentrates, and zinc ferrites or other residues from the integrated zinc refinery. Zinc ferrite is not produced in the autoclave and high zinc recovery is maintained in the pressure leaching of low-grade concentrates with high iron contents. Dilution of the concentrate by pyrite may be desirable if it increases zinc recovery in the concentrator, but excessive amounts of pyrite would complicate elemental sulphur recovery from the pressure leach residue. Silicates are essentially inert to the zinc pressure leach process and concentrates with high silica content are readily treated.

Atmospheric Leach

In the atmospheric leach process, a high iron concentration is required and this is provided by an appropriate recycle of iron precipitates. Upon contact of the iron oxide residue with the spent electrolyte, which contains sulphuric acid, ferric iron is dissolved as follows, depending on the species of iron oxide.

2 FeOOH + 3 H2S04 > Fe2(S04)3 + 4 H20 (13)

Fe203 xH20 + 3 H2S04 > Fe2(S04)3 + (x+3) H20 (14)

2 FeOHS04 + H2S04 > Fe2(S04)3 + 2 H20 (15)

Sphalerite reacts with the dissolved ferric sulphate to produce soluble zinc sulphate and elemental sulphur, with the iron being reduced from the ferric to ferrous state, as in reaction (2).

Under oxidizing conditions the ferrous iron is reoxidized to ferric iron as in reaction (3); however, this reaction is extremely slow at low oxygen partial pressure.

If the oxidation potential of the solution is low, sphalerite can also be attacked to a limited extent by sulphuric acid with the liberation of hydrogen sulphide according to the following reaction.

ZnS + H2S04 > ZnS04 + H2S (16)

As the acid concentration in the leach solution decreases, ferric iron may precipitate by reactions (11) or (12), or by reaction (17) below.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 731

Fe2(S04)3 + 2 H20 > 2 FeOHS04 + H2S04 (17)

TEST RESULTS

Comparative pressure leach and atmospheric leach tests and residue treatment tests were carried out in the Dynatec laboratory in Fort Saskatchewan in 1999 and 2000. The procedures and results are summarized in the following sections.

Leach Test Procedure

The pressure and atmospheric leach tests were performed in a 3.8 L titanium autoclave equipped with baffles, dual axial impellers and a water cooled condenser. The operating conditions are presented in Table III. Samples were taken throughout and at the completion of the leach tests.

Table ΙΠ - Conditions in the Pressure and Atmospheric Leach Testwork

Type of Leach Duration, h Temperature, °C Pressure, kPa (gauge) Impeller Tip Speed, m/s Oxidant Vent, L/min

Pressure 2

150 1100 4.7

Oxygen Gas 0.5

Atmospheric 24 95 50 4.7

Oxygen Gas 0.5 for first 8 h, 0 for last 16 h

Flotation Test Procedure

The leach test residues were subjected to froth flotation. The solids were repulped in a blend of leach solution and water to achieve a solids content in the slurry of approximately 8 to 10% solids, and were subjected to flotation in a Denver Model D-12 laboratory flotation machine. The slurry was mixed in the cell for 2 minutes prior to introduction of air, with the agitation rate at 1 500 rev/min in both the mixing and flotation steps of the operation.

The flotation operation was a combined rougher-scavenger circuit. The first concentrate (rougher) contained about 80% of the floatable solids. The second concentrate (scavenger) was collected until the froth was essentially barren.

Feed Materials

Three zinc concentrates were used in the comparative leach testwork. The chemical analyses of the concentrates are given in Table IV.

732 LEAD-ZINC 2000

Table IV - Chemical Analyses of Zinc Concentrates Used in the Testwork

Element Cu, % Fe Pb S Zn

Concentrate A 0.1 4.6 3.4

29.6 56.7

Concentrate B 1.2 8.7 0.2

30.6 53.7

Concentrate C 0.2 9.9 1.3

32.8 52.7

Zinc Extraction in the Leach Testwork

Pressure leach tests were performed on the three zinc concentrates. The zinc pick-up to solution was greater than 80 g/L in all three leach tests. The zinc extraction rates are summarized in Table V.

Table V - Zinc Extraction Results from the Pressure Leach Testwork

B C~~ 337 37.7

83.6 96.5 97.1 99.1 98.4 99.5 99.0 99.5 99.5 99.4

The particle sizes of the zinc concentrates fed to the pressure leach tests were all in the range of 90% passing 32 to 38 μπι. Greater than 98% zinc extraction was achieved with all three concentrate feeds within 60 minutes. The third concentrate was very reactive under pressure leach conditions, achieving 99.1% zinc extraction after only 30 minutes. The three concentrates were also subjected to atmospheric leaching. The atmospheric leach tests results are summarized in Table VI.

Table VI - Zinc Extraction Results from the Atmospheric Leach Testwork

Zinc Concentrate Particle Size D90, μηι Zinc Extraction, %

15 min 30 min 60 min 90 min 120 min

A 32.3

89.2 97.8 99.7 99.9 99.7

Zinc Concentrate Particle Size D90, μπι Zinc Extraction, %

2h 8h 24 h

A 43.7

22 45 67

A 32.3

35 66 85

A 11.5

58 80 96

B 33.7

n.a. 62 83

B 23.5

31 43 83

B 14.6

55 84 95

C 37.7

25 80 95

Zinc extractions of about 95% were achieved with each of the concentrates within 24 h under atmospheric conditions in the batch tests. However, regrinding was necessary to further reduce the particle size of Concentrates A and B to achieve 95% zinc extraction. The zinc extraction rates obtained from Concentrate A with varying particle size and different modes of leaching are compared in Figure 1.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 733

is x ω o c N

100

80

60

40

20

£*** - -

D A

Δ /S~

O

Pressure Leach Dqn = 32.3 μηι

""^Atmospheric Leach Don = 43.7 μπι

"^ Atmospheric Leach Don = 32.3 μηι

~ ^ Atmospheric Leach Don =11.5 μιη

12 Time, h

16 20 24

Figure 1 - Zinc Extraction; Concentrate A

Zinc Recovery in Flotation

The residues from the atmospheric and pressure leach tests were floated and the flotation solids were analyzed for zinc. Based on these assays, and the weights and assays of the feed to the leaching step, the overall deportment of zinc to the flotation tailings was calculated and is presented in Table VII. The overall deportment of zinc to tailings was in the range of 0.8 to 2% in the atmospheric leach tests, and in the range of 0.1 to 0.5% in the pressure leach tests.

Leach Type Concentrate Zn Extraction, % Total Zn to Tails, %

Atm A

67 2.0

Table VII - Zinc Deportment

Atm Atm Atm Atm Atm A A B B B

85 96 83 83 95 1.1 1.0 0.9 1.5 0.8

Atm C

95 0.8

Press A

99.7 0.1

Press B

99.5 0.2

Press C

99.4 0.5

Solution Composition

Final solutions from both pressure and atmospheric leach tests were submitted for chemical analysis, and the results are presented in Table VEIL The concentrations of cobalt, nickel and selenium in the final solutions were not significantly different. Arsenic, antimony and tellurium levels were about the same or lower in the solution generated in pressure leaching (the arsenic concentration was substantially lower in the pressure leach liquor for

734 LEAD-ZINC 2000

Concentrate A). Both germanium and thallium were not consistently lower or higher in the solution generated from either type of leach. The iron concentration in solution, which must be precipitated in an iron rejection circuit, is significantly higher in the atmospheric leach solution.

Table VIII - Final Solution Analyses from Pressure and Atmospheric Leach Testwork

Concentrate Type of Leach Final Sol'n Analysis

Fe,g/L H2SO4 Zn Sb, mg/L As Co Ge Ni Se Te Tl

A Press

7.8 45.5

159 6.0

32, 16 4.6

11 0.1

<0.5 3.7

A Atm

25.9 51.2 94 16.0

508 14 2.6

13 0.1 0.8 1.2

B Press

3.6 28.1

128 2.7

1150 8 0.05

30 0.2

<0.5 <0.1

B Atm

28.8 43.4 93 4.4

1020 8 0.11

31 0.2 0.9 0.2

C Press

3.5 40.3

133 1.4

448 6 1.5

22 0.1

<0.5 1.1

C Atm

28.7 44.8 91

3.5 694

7 2.5

21 0.1 0.9 0.8

Byproduct Quality

The sulphidic portions of the leach residues were subjected to sulphur melting and hot filtration. The arsenic and mercury contents of the elemental sulphur byproducts originating from the pressure leach treatment of the concentrates were consistently lower than the arsenic and mercury contents of the elemental sulphur originating from atmospheric leaching. The selenium content was about the same or lower in the elemental sulphur originating from pressure leaching compared to atmospheric leaching.

CAPITAL AND OPERATING COSTS

Basis

An order of magnitude comparison of the economics of atmospheric and pressure leaching of sphalerite concentrates has been conducted for a hypothetical zinc refinery expansion of 40,000 tonnes of slab zinc per year. For the expansion, it is considered that the concentrate storage, leach residue handling, solution purification, electrowinning, and cathode melting and casting areas are all identical. The atmospheric leach plant has been assumed to be oxygen-sparged, with ammonia addition for precipitation of iron as ammonium jarosite. The sparging of air in atmospheric leaching was considered; however, the economics favoured the use of oxygen.

The hypothetical case is for a sphalerite concentrate of 50% zinc and 8% iron. The relevant economic data (US$) used are shown in Table IX. The indirect costs and contingency for the capital cost are each estimated as 25% of the total direct costs.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 735

Table IX - Operating Cost Data (US$)

Item Value Zinc Concentrate, Grade 50 wt% Zn ($/t) $ 167 Power Cost ($/MWh) $30 Steam Cost ($/t) $5 Ammonia Cost ($/t) $300 Maintenance Materials:

Pressure Leach 6% Atmospheric Leach 3%

Two hypothetical concentrates have been used for the study: a fast-leaching concentrate and a slow-leaching concentrate. Assumed characteristics of the two concentrates are given in Table X.

Table X - Batch Leach Times for Hypothetical Zinc Concentrates

Leaching Type Concentrate 1 Concentrate 2 Atmospheric (to 95%) 5.0 24

Pressure (to 99%) 05 1.0

It should be noted that none of the concentrates tested by Dynatec yielded 95% zinc extraction in less than 24 hours in an atmospheric leach. Based on the results for the different concentrates presented in this paper, the extraction in the atmospheric leach is assumed to be 95%, compared with 99% extraction in the pressure leach. The overall zinc recoveries are assumed to be 1% lower in each case. The difference in zinc recovery between pressure and atmospheric leaching is 1,600 t/y, which must be supplied with additional zinc concentrate at a cost of $0.53 million per year.

Further, it has been assumed that the concentrate would require grinding to 90% passing 44 μπι for pressure leaching, and 90% passing 20 μπι for atmospheric leaching. The additional grinding in a vertical stirred mill is estimated to cost $0.66/t concentrate, or $56,000 per year.

Concentrate 1 Results

The atmospheric leach circuit for Concentrate 1 is estimated to consist of five 5.6 m diameter by 7.7 m tall tanks, equipped with 75 kW motors. A 60 t/d oxygen plant is included. The overall installed capital cost for this circuit is estimated at $7.07 million.

The pressure leach circuit for this concentrate is estimated to consist of one 2.7 m diameter by 16 m long autoclave with six 45 kW agitators, plus feed solution and slurry pumps and agitated flash and conditioning tanks. A 45 t/d oxygen plant is included, as is a seal water circuit. The overall installed capital cost for this circuit is estimated at $8.82 million.

The atmospheric leach circuit, by virtue of the large amount of heat lost with the vent gases, requires a substantial steam addition to maintain the energy balance. The comparison of the operating costs is shown in Table XI. The operating costs of the two circuits are roughly comparable, excluding the steam.

736 LEAD-ZINC 2000

Table XI - Operating Cost Comparison for Concentrate 1

Operating Cost Item Atmospheric Leach Pressure Leach Power $313 000 $261 000 Maintenance $141 000 $353 000 Reagents $195 000 $76 000 Additional Grinding $56 000 Steam $267 000 $21 000 Total $972 000 $711000

Comparison of the two circuits shows that the pressure leach circuit costs $1.75 million more to install, however it can be operated for $0.26 million per year less. Adding on the additional zinc recovery value of $0.53 million per year gives a net earnings increase for the pressure leach facility of $0.80 million per year. Using an annual cash flow approach, the pressure leach plant is shown to be an attractive investment.

Concentrate 2 Results

The atmospheric leach circuit for Concentrate 2 is estimated to consist of seven 8.4 m diameter by 11.1m tall tanks, equipped with 150 kW motors. The overall installed capital cost for this circuit is estimated at $8.78 million.

The pressure leach circuit for this concentrate is estimated to consist of one 3.0 m diameter by 20 m long autoclave with six 75 kW agitators, plus feed solution and slurry pumps and agitated flash and conditioning tanks. The overall installed capital cost for this circuit is estimated at $9.53 million.

The comparison of the operating costs is shown Table ΧΠ. The operating costs of the pressure leach circuit are considerably less than the operating costs of the atmospheric circuit.

Table XII - Operating Cost Comparison for Concentrate 2

Operating Cost Item ospheric Leach $449 000 $176 000 $195 000 $56 000

$267 000 $1 143 000

Pressure Leach $303 000 $381 000 $76 000

-$21 000 $781 000

Power Maintenance Reagents Additional Grinding Steam Total

Comparison of the two circuits shows that the pressure leach circuit costs $0.75 million more to install, however it can be operated for $0.36 million per year less. Adding on the additional zinc recovery value of $0.53 million per year gives a net earnings increase for the pressure leach facility of $0.90 million per year. Using an annual cash flow approach, the pressure leach processing of Concentrate 2 is also an attractive investment.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 737

CONCLUSIONS

This test program has demonstrated that higher overall zinc extraction can be achieved from pressure leaching in comparison to atmospheric leaching. Based on the extractions obtained in the leach testwork and subsequent recoveries in flotation, a higher overall recovery of zinc is possible from a process employing pressure leaching as compared to atmospheric leaching.

Solutions from both types of leaching were submitted for limited minor impurity analyses. There were no major differences, in terms of minor impurities, between the solutions generated from atmospheric or pressure leaching. The higher levels of iron in the atmospheric leach solution will obviously have to be precipitated at some point in the process. More efficient iron precipitation and a higher level of zinc pick-up are benefits of pressure leaching.

The impurity content of the elemental sulphur filtrate produced from the pressure leach residues was slightly lower than that produced from the atmospheric leach residues. The elemental sulphur produced from the pressure leach residues contained less arsenic and mercury and about the same level of selenium in comparison to the elemental sulphur produced from the atmospheric leach residues.

For the zinc concentrates tested by Dynatec an expansion of an existing roast-leach-electrowin plant using pressure leaching offers economics that are favourable in comparison to those of an atmospheric leach.

ACKNOWLEDGEMENTS

The authors thank the management of Dynatec Corporation for permission to publish this paper.

REFERENCES

1. M.T. Martin, and W.A. Jankola, "Cominco's Trail Zinc Pressure Leach Operation", CM Bulletin. Vol. 78, No. 876, 1985, 77-81.

2. B.H. Johnston and B.N. Doyle, "Start Up and Operation of the Kidd Creek Zinc Sulphide Pressure Leaching Plant", Minerals and Metallurgical Processing. February 1986,1-7.

3. E. Ozberk, M.J. Collins, M. Makwana, I.M. Masters, R. Pullenberg and W. Bahl, "Zinc Pressure Leaching at the Ruhr-Zink Refinery", Hvdrometallurgy. Vol. 39,1995, 53-61.

4. M.J. Brown, E.T. deGroot, M.G. Heximer, A.J. Karges, G.N. Masuch and CM. Okumura, "Zinc Capacity Increase at Cominco's Trail Operations", Zinc and Lead Processing. J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton and P. Hancock, Eds., Canadian Institute Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 41-54.

LEAD-ZINC 2000

M.J. Collins, E.J McConaghy, R.F. Stauffer, G.J. Desroches and B.D. Krysa, "Starting up the Sherritt Zinc Pressure Leach Process at Hudson Bay", JOM. Vol. 46, No. 4,1994, 51-58.

S.P. Fugleberg, A.B. Farviken, "A Method for Leaching Zinc Concentrate in Atmospheric Conditions", PI Patent No. 100806 B, February 27, 1998.

H. Takala, "Leaching of Zinc Concentrates at Outokumpu Kokkola Plant", Fortschritte in der Hydrometallurgie. GDMB, 1998,129-138.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 739

TREATMENT OF SECONDARY ZINC OXIDES FOR USE IN AN ELECTROLYTIC ZINC PLANT

S.S. Chabot and S.E. James Big River Zinc Corporation

2401 Mississippi Avenue Sauget, Illinois, U.S.A. 62201

ABSTRACT

A recent survey of major steel companies worldwide revealed the importance of recycling zinc units from electric arc furnace (EAF) dust, mill scale, blast furnace dust and other zinc containing streams within the steel industry. Many processes have been developed and installed in response to this challenge. The most successful approach involves the production of a crude zinc oxide and its sale to primary zinc producers to recover the metal values. The problem with this process is the high concentration of halides in the crude oxide. This type of material cannot be processed directly in an electrolytic zinc plant because of the corrosion problems it will cause in electrolysis. In 1999, Big River Zinc installed a zinc oxide receiving and washing plant to minimize the impact of halides in its electrolysis circuit. A description of the washing plant and the initial operating results are presented in this paper.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

740 LEAD-ZINC 2000

INTRODUCTION

Many sources of secondary zinc oxide are available on the market (1). Some come from the direct production of zinc products, such as the skimmings from furnaces and baghouse dust, whereas others originate from steel recycling. The main source of zinc from recycling steel is in the form of electric are furnace (EAF) dust. For years, the zinc contained in the EAF dust ended up in landfills because of a lack of processing facilities. In recent years, parts of the steel industry investigated EAF dust as a possible feed source for their plant by reducing the iron and removing the zinc and other impurities it contained. This created a high zinc product that needed an outlet.

Some producers attempted to make zinc metal products directly from EAF dust in their processes (2). The increased value of the transformed zinc by-product. Although this is appealing, it does not take into account the difficulties involved with the transformation. As a result, some producers ended up installing equipment that did not produce a zinc product of a high enough quality to meet the market demand. The low-grade zinc metal product obtained could be sold only at scrap value. This eliminates any economic advantage of reprocessing the zinc.

Other producers took the approach of marketing the zinc oxide and approached Big River Zinc (BRZ) as a possible outlet. As shown in Table I, the zinc oxide is similar in nature to the zinc oxide in the calcine produced by roasting zinc concentrates. Most of the impurities found in the oxide, such as lead, cadmium and copper, are also typical constituents of zinc concentrates. The biggest difference is the presence of chloride and fluoride. Big River Zinc saw these requests as opportunities to assist the steel and the zinc industries in eliminating waste. Big River Zinc strongly believe that business associations between the steel and zinc industries will help reduce, if not eliminate, the need to landfill EAF dust. With investments in the right technologies, we believe that the steel industry can economically recover the iron units from the EAF dust. At the same time, the zinc industry can use the zinc derived from the EAF dust profitably, helping to conserve our natural resources, while simultaneously assisting the steel industry to solve one of its major problems.

Table I - Chemical Composition of Big River Zinc Calcine and Ameristeel Crude Zinc Oxide

Element

Zinc Lead Cadmium Copper Iron Sodium Potassium Chloride Fluoride

BRZ Roaster Calcine (wt %)

67.8 1.73 0.66 0.34 2.16 0.02 0.1

<0.01 O.001

Ameristeel Crude Zinc Oxide(wt %)

64.02 4.88 0.145 0.034 0.46 2.24 2.18 5.37 0.21

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 741

OXIDE PLANT DESIGN

The overall criteria were to install a facility that would allow Big River Zinc to process crude zinc oxide through its electolytic zinc plant. The limitations for processing crude zinc oxides have been described previously (3). The relatively high levels of chloride and fluoride pose significant problems.

Big River Zinc investigated two processes to dechlorinate the oxides. The first process used a water wash accompanied with a pH adjustment to minimize zinc losses. This method offered a known technology because Big River Zinc provided assistance to ZTT in Caldwell, Texas, to refine their similar process. Tests showed that washing, followed by neutralization with soda ash, succeeded in reducing the chloride content to less than 1% in the final product. The other process investigated was thermal dechlorination (4). It is known that thermal dechlorination can remove over 90% of the chlorides from roaster feeds. A few tests indicated promising results for this technique but the decision was taken to pursue washing as the process of choice.

Overall Plant Design

The first step in the plant design was to acquire all the necessary permits relevant not only to the construction of the plant but also for the classification of the material. Big River Zinc worked with the various government agencies to remove the hazardous waste stigma from the crude zinc oxide derived from the EAF dusts. The efforts paid off in May 1999, when Big River Zinc received an Adjusted Standard from the Illinois Pollution Control Board declaring that zinc oxide derived from EAF dust is a product, and not a derived hazardous waste, when recycled at the BRZ zinc refinery.

The main wash plant design criteria were selected to keep the construction and operating costs low to ensure the viability of the project. The other main factor involved the handling of the fine crude zinc oxide in a reliable and environmentally responsible manner. This prompted the selection of pneumatic shipping of the oxide to the plant site. This insures that the oxide is transported in an airtight vessel and that the unloading can be done cleanly into our storage silo. Also, shipping in bulk containers requires a minimal amount of labor at both the shipping and receiving sites.

We elected to wash the oxide using a batch process. This allows for greater process flexibility and enables us to segregate material that might contain impurities that need to be fed at a reduced rate to the plant. The batch system is also easy to automate and does not require constant operator intervention. The overall washing process is best described by the flowsheet shown in Figure 1.

742 LEAD-ZINC 2000

TO ATMOS.

BAGHOUSE

TO COLLECTING CONVEYOR

BUCKET ELEVATOR

MIXING CONE

\

1)1 ^ il l ■ ιΤι

SUPER SACKS STORAGE

παα DQD QDD

VENT BWSJCR

XHEPA lirr — 'JLTER

VENT ( , .

Ö HEPA I RAIL CAR

FILTER

PNEUMATIC SYSTEM

§m

00 v Y v OA.

LOAD CELLS

QUAD SCREW CONVEYOR

COLLECTING CONVEYOR

*m PR

LAROX FILTER AIR- m

» J J S7vma iif&

DRAIN

FILTER CAKE i FIITRATS

TO CONCENTRATE STORAGE BUILDING

PEABODY TANK

FILTRATE m w

-fkTER

Figure 1- Flowsheet of the Big River Zinc Oxide Washing Plant

The plant design was carried out on an accelerated schedule. To simplify the construction, we opted to build the facility in phases. The first phase represents the core of the plant. It contains a pneumatic track unloader coupled to a storage silo, a reactor tank and a Larox filter. This approach enabled us to start up quickly and to provide a good way to determine the compatibility of the washing plant with the rest of the facility. Phase 1 was completed in June 1999 and provided the capacity to wash up to 22,000 tons per year of oxide.

Phase 2 of the process was completed in October 1999. Its main goal was the addition of a pneumatic rail unloading station and a second oxide silo. The additional storage capability ensures that there is always enough capacity to unload the railcars as they arrive in Sauget. The plant capacity remains unchanged with the completion of this phase.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 743

Phase 3 of the project depends upon the supplies of zinc oxide. We have plans to add a facility to handle oxide delivered in supersacks. This expansion would be aimed to take material from smaller zinc oxide producers. The versatility of the supersack is well known but, because of the amount of oxide we are prepared to treat, a good infrastructure is required to handle a large number of supersacks.

Integration of the ZnO Washing Plant at Big River Zinc

The integration of the washing plant into the Big River Zinc process revolves around the following factors:

• Washing requires large amounts of water • Washing generates large amounts of chloride-rich water • Washing requires handling large volumes of material, such that easy truck

access, easy rail access and close proximity to the roasters for further processing is required.

With these criteria, we opted to build the plant as an extension of the wastewater treatment plant. This plant is located at the north end of the site. The location makes water disposal easy because of the proximity to the water treatment plant. Also, the site-wide water recovery project channels the recovered storm water to the wastewater plant. This should make recycle water available to wash the oxide and will eliminate the need to purchase fresh water.

The plant design allows for the future construction of a conveyor to carry the washed oxide directly to the concentrate storage shed. At present, the washed oxide falls into a dump truck that the operator empties once or twice a shift. Big River Zinc decided to delay the installation of a conveyor until the process has been proven and the plant production reached a higher level. We are also evaluating possible means of adding the washed oxide directly to the leach circuit.

Where to Add the Oxide Into the Circuit

The final chloride and fluoride levels in the washed oxide determine where it can be added in the zinc plant. If the levels were low enough, the best location to add the oxide would be in the leach department. However, because of the variability in the halide level in the incoming materials and their different levels of washing, we opted not to feed the material directly into the leach circuit. Instead, we decided to feed the washed oxide to the roasters. This has the advantage of:

• Providing additional halide removal (>80% for fluoride and >90% for chloride) • Eliminating the need for an addition point in the leach plant • Reducing the heat load in the roasters by displacing cooling water injected into the

roasters with moist, non-sulfur bearing, cool material.

Using the roasters as the addition point carries its own problems. If too much of the halides is eliminated in the roasters, we will need to increase the scrubber solution bleed to minimize corrosion. Corrosion will also occur in the gas handling system, especially where the temperature falls below the condensation point of hydrochloric and hydrofluoric acids. The heat balance difficulties in the roasters is attributable to the increased feed rate. Heat removal achieved by the addition of moist oxide offsets the use of direct injection water in the roasters. The composition of the oxide material also poses a problem as roaster feed. Because the oxides

744 LEAD-ZINC 2000

do not contain any sulfur, we must insure that they are well blended with the concentrate. Without good blending, the roaster operation will become irregular and overall output could decrease. Also, some of the oxide on the market contains large amounts of lead, copper, or even metallic zinc. These elements can cause problems with roaster operation if their concentration becomes excessive. Another possible problem could arise if all the washed oxide is very fine and reports to the waste-heat boilers. The problem of increased fines is well documented in the zinc industry when plants convert to finer concentrates (5). We are in the process of updating the boiler rapping system to increase the boiler cleaning ability. Although this upgrade was required because of the age of the original equipment, the addition of finer feed, both oxide and possibly concentrate, mandates that this improvement be made now. The first phase of the improved rapper system will be installed in September 2000.

Limits on Impurities

The capacity for "undesirable" elements at Big River Zinc falls into two categories. The first category involves elements such as sodium and potassium. These elements do not affect the equipment and they do not directly affect the process. The problem with these elements is that the process used at Big River Zinc does not eliminate them. As a result, they build up in the electrolyte resulting in a lower zinc solubility. This increases the production cost because of the need to treat more solution to produce the same quantity of zinc.

The other category of impurities includes elements that are detrimental to the process or equipment. This category includes iron, lead, chloride and fluoride. Iron is required in the process as a purification agent. However, it also generates residue that cannot be treated on site. Iron exits the Big River Zinc plant in the lead-silver concentrate. If the iron content is too high, the value of the concentrate is reduced, resulting in a cash flow reduction. Lead, as stated earlier, will cause problems in the roasters if its concentration is too high. In the case of chloride and fluoride, their major impact is equipment corrosion. As with sodium and potassium, the leach and purification circuit does not eliminate chloride and fluoride. When the chloride or fluoride contents increase, we observe increased corrosion rates in the equipment, especially in the cellroom aluminum cathodes.

Chloride and Fluoride Capacity

The two elements that require the most attention during the washing process are chloride and fluoride. This, as explained earlier, arises from the inability of the circuit to eliminate them and the resulting corrosiveness resulting from their buildup in the circuit. To better understand how important this is, we must look at the chloride and fluoride balance for the electrolyte circuit. Figure 2 shows the chloride distribution throughout the circuit.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 745

230,000 r>3/year City 30 ng/L

Voter 7 , 6 -tons/year Cl <42 Ibs/doy)

Cellroon

Return Acid -200 ng/L

10 tons/year <55 Ibs/doy)

Cl

Leach Plant

13.4 tons/yeor Cl (74 lbs/day)

Lead-Silver Concentrate

2.7 tons/year Cl <15 lbs/day)

ZSM <Zinc Sulfate Monohydrate)

1.5 tons/year Cl (8 lbs/day)

Zinc Sulfate Solution

Figure 2 - Chloride Balance in Big River Zinc's Solution Circuit Assuming 200 mg/L of Chloride in the Cell Acid

The balance assumes a maximum allowable chloride concentration of 200 mg/1. This level is a compromise between using chloride-rich feed material and corrosion throughout the plant, especially on the aluminum cathodes in the cellroom.

Chlorides are not effectively bled from the circuit. In fact, the daily chloride loss occurs through solution loss throughout the plant. The largest chloride loss is the solution left in the silver-lead concentrate. This method of eliminating chloride is reduced by our attempts to wash the soluble zinc and cadmium contained in the cake. Recovering the soluble zinc and cadmium also recycles the chloride back into the plant. Zinc sulfate monohydrate (ZSM) production and the marketable zinc sulfate solution provide for a direct bleed of solution from the circuit. Although more efficient at eliminating chloride, the relatively small volume of ZSM and marketable zinc sulfate solution limit their potential for chloride removal. Other means of eliminating chloride are present but they are not significant. As a result, the total leach plant chloride input is limited to approximately 55 pounds per day. This translates into an allowable 1100 pounds per day of EAF dust-derived zinc oxide, which contains about 5% chloride. Removing the chloride from the material first enables the treatment of significantly more material, as shown in Figure 3.

746 LEAD-ZINC 2000

15,000 tons /year

5.8 ·/. Cl

870 "tons/year

(4800 lbs/day)

\f "

1 tons /year C5.5 lbs/day)

Cl

32,000 nVyr 30 riQ/L Cl C i t y

V a t e r ci

Oxide Washing

Plant

14,700 ions /year 1.3 '/. Cl

680 tons /year Cl

191 tons/year Cl (1,050 lbs/day)

9 "tons/year Cl

(50 lbs/day) Roasters <Assun* 95X Cl

r*J«ctloo>

Leach Plant

182-tons/year Cl

Sulfuric Acid Plan-t

Wastewater Trea"tnen-t

Plan-t

862 tons /year Cl

Anerican Bo-t-ton

Figure 3- Allowable Chloride in the Zinc Oxide Washing Plant to Maintain 200 mg/L Chloride in the Cell Return Acid

As noted previously, the roasters eliminate 95% of the chloride. This increases the daily permissable intake of crude zinc oxide, containing 5% chloride, to 22,000 pounds, or 11 tons, per day. Adding the washing process ahead of the roaster and lowering the chloride content below the design concentration of 1% increases the chloride capacity to 55 tons of crude zinc oxide per day. With the present operation averaging 0.35% chloride, the capacity for crude zinc oxide, based on chloride limitations, has increased to 155 tons per day (56,000 tons per year).

Fluoride

Fluoride is much harder to quantify in the system because of its low level. Big River Zinc is very sensitive to the fluoride concentration in the electrolyte. Because we do not condition the surface of the cellroom cathodes, we are limited to 10 mg/1 in the cell electrolyte. If we were able to brush the cathodes after stripping, we could tolerate up to 50 mg/1 of fluoride.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 747

Fluoride elimination occurs in a similar fashion to that of chloride. This means that the maximum daily fluoride input is about 3 pounds. As for chloride, the roasters eliminate fluoride but to a lesser extent. Tests showed that the roasters eliminate about 95% of the fluoride in the feed material. This allows us to feed the leach plant with 55 pounds of fluoride daily or 27,500 pounds (13.75 tons) of crude zinc oxide containing 0.2% fluoride. It becomes obvious that, if we want to process the 155 tons per day of crude zinc oxide allowable from a chloride standpoint, the wash plant must produce a washed oxide with a fluoride content of 0.015%.

The allowable fluoride input to the leach circuit at Big River Zinc might have increased because of the addition of a gypsum removal system in November 1999. The removal of gypsum from the circuit might create the right conditions to precipitate calcium fluoride. This would create a good output for fluoride and would lower the required washing efficiency for fluoride in the oxide washing plant.

Plant Start-up

The oxide washing plant started up in May 1999 and we could not have hoped for a better start-up. One of the major concerns was materials handling. We knew that the material can be difficult to move once it is compressed or wetted. The silo design, which includes a quad-screw discharge, two vibrating panels and air cannons, delivered a reliable flow of oxide to the ventilated conveyors leading to the reaction tank. The other problem expected was the difficulty of wetting the oxide. The design included a mixing cone, that uses recycled solution from the tank as a wetting agent. This design has worked flawlessly. The plant produced in-specification material from the first day.

Plant Operation

The washing plant effectively removes chloride and fluoride. Figures 4 and 5 show the chloride and fluoride assays obtained since the plant went into operation.

M

jr

w 1-Jun-99 21-Jul-99 9-Sep-99 29-Oct-99 18-Dec-99 6-Feb-00 27-Mar-00

Figure 4- Chloride Concentration in the Washed Zinc Oxide

748 LEAD-ZINC 2000

Wun-99 21-Jul-99 9-Sep-99 29-Oct-99 18-Dec-99 6-Feb-OO 27-Mar-OO

Figure 5- Fluoride Concentration in the Washed Zinc Oxide

Of course, like every start-up, we experienced some problems. The first problem encountered was related to the nature of the silo discharging equipment. We discharge the silo using a quad-screw conveyor, which removes the oxide at a pre-determined volumetric rate. The silo conveyor then discharges into a long ventilated single screw conveyor that brings the oxide to a bucket elevator. When the silo is full, the contained oxide compresses. The silo discharge equipment continues to feed the same amount of oxide to the screw conveyor. The action of the screw on the oxide undoes the compaction that occurred in the silo, resulting in a much larger volume of oxide. This results in an excessive volume of oxide traveling through the screw and causing it to back-up. Speeding up the screw conveyor solved this problem, but it moved the difficulty to the bucket elevator. Luckily, the elevator design was oversized enough to allow us to increase its speed and resolve the problem.

A more troublesome problem involves the operation of the Larox filter to produce the washed oxide. We selected the Larox filter because of its automation and the performance observed in a laboratory unit. The filter has proven reliable mechanically and does not require continuous operator attendance. However, the cake produced has a higher moisture content than originally expected. This high moisture is explained by the fine nature of the oxide. When we selected the filter, the secondary oxides expected as the feed to the plant were much coarser, filtered easily, and produced a cake that was dry enough to be readily handled. The Larox performance on the finer oxide needs to be improved.

We attempted to add concentrate to the washed oxide prior to its filtration. The concentrate is much coarser than the zinc oxide and we hoped it would act as a filtration aid. These additions resulted in a dryer cake, about 16% to 18% H2O. This is an improvement, but the cake is still sticky, and is hard to handle. The solution will require the installation of

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 749

additional equipment to blend this sticky oxide with concentrate to ensure a proper mixture going to the roasters. The selection of the equipment has proven difficult because of the nature of the material and the possible need to partly dry the oxide before blending it for the roaster feed.

Future Improvements

Feeding Washed Oxide Directly to the Leaching Plant

Feeding the oxide to the roasters has advantages and disadvantages, as described earlier. If we were able to remove enough chloride and fluoride from the crude zinc oxides, the need to use the roaster as a final dechlorination step would be eliminated. Additional work is required to discover the operating conditions required to reliably remove the halides to the desired level. Once we discover the right conditions, an economic analysis will be done to see if the implementation of this new technique is desirable.

Handling Crude Zinc Oxide in Supersacks

Big River Zinc realizes that many smaller zinc users produce crude zinc oxide that must be recycled. Most of these producers use supersacks to handle the oxide because they do not generate enough oxide to justify the equipment required for shipment using pneumatic trucks or railcars. To effectively service this market, Big River Zinc is planning the construction of a supersack unloading facility that will feed into the existing plant. This will provide a good outlet for the smaller oxide producers and will greatly help to close the zinc industry recycling loop.

Filter Improvement and Washed Oxide Handling

As discussed earlier, the washed oxide produced is very sticky, making it hard to handle in ordinary equipment. We are working at different levels to solve this problem. Present efforts address the filter operation, including the filter cycle, cloth selection, process parameters and blending equipment.

CONCLUSIONS

Big River Zinc has demonstrated that it can help close the recycling loop for zinc and provide a service to the steel industry by effectively recycling crude zinc oxide recovered from EAF dust, basic oxygen furnace (BOF) dust and other sources of crude zinc oxide. Preventing large quantities of zinc-rich material from reaching landfills is an important first step.

The construction of a washing plant at Big River Zinc represents a major step in the recycling of crude zinc oxide. The plant, originally designed to handle only oxide arriving in pneumatic trucks, was expanded to receive material in delivered pneumatic railcars. The next step will include the ability to handle material in supersacks. Big River Zinc's entry into the crude zinc oxide market now provides the North American steel industry with an additional viable market for recycling crude zinc oxide.

Big River Zinc will continue to expand its crude zinc oxide washing facility to accommodate different sources of crude zinc oxide, and thereby, further increase its critical role in zinc industry stewardship.

750 LEAD-ZINC 2000

REFERENCES

1. P. B. Queneau, S.E. James, J.P. Downey and G.M. Livelly, "Recycling Lead and Zinc in the United States", Zinc and Lead Processing. J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton and P. Hancock, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 127-154.

2. D.K. Xya and C.A. Pickles, "Extraction of Non-ferrous Metals from Electric Arc Furnaces Dust", Waste Processing and Recycling in Mineral & Metallurgical Industries III, S.R. Rao, L.M. Amaratunga, G.G. Richards and P.D. Kondos, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 221-245.

3. G. Houlachi, A. Condy and R.W. Stanley, "Review of Zinc Recycling", Zinc & Lead '95. T. Azakami, N. Masuko, J.E. Dutrizac and E. Ozberk, Eds., The Mining & Materials Processing Institute of Japan, Tokyo, Japan, 1995, 417-431.

4. C. Mattich, K. Hasselwander, H. Lommert and A.N. Beyzavi, "Electrolytic Zinc Manufacture with Waelz Treatment of Neutral Leach Residues", Zinc and Lead Processing. J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton and P. Hancock, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 561-578.

5. M.J. Brown, E.T. deGroot, M.G. Heximer, A.J. Karges, G.N. Masuch and CM. Okumura, "Zinc Capacity Increase at Cominco's Trail Operations", Zinc and Lead Processing. J.E. Dutrizac, J.A. Gonzalez, G.L. Bolton and P. Hancock, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1998, 41-54.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 751

UPCOMING ZINC MINE PROJECTS: THE KEY FOR SUCCESS IS ZINCEX SOLVENT EXTRACTION

M.A. Garcia, A. Mejias, D. Martin and G. Diaz Tecnicas Reunidas, S.A. R&D Division

C/Sierra Nevada 16, Pol. Ind. San Fernando II, San Fernando de Henares, Madrid 28830, Spain

ABSTRACT

During 1997 and 1998 two successful zinc feasibility studies, Skorpion and Sanyati, were carried out by Tecnicas Reunidas based on the modified ZINCEX® technology. The technical viability of each project was proved during more than 600 h of continuous running in a pilot plant. The Skorpion zinc project would recover 150,000 t SHG Zn/y from an oxidized zinc ore of a Namibian mine. The economical study showed that the plant would have the lowest production costs in the world. The Sanyati zinc project would treat a bleed from an existing copper refinery in Zimbabwe to recover 5,000 t SHG Zn/y. In the first case the electrowinning cellhouse was supplied by Union Miniere and in the second by Tecnicas Reunidas. This article highlights the pilot plant results and the advantages of the proposed process over conventional zinc production practices.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

752 LEAD-ZINC 2000

INTRODUCTION

All over the world there are many oxidized, silicated and carbonated zinc deposits that have never been exploited mainly because conventional technologies, developed for zinc sulfides, showed significant technical difficulties when applied to these ores.

The British company Reunion Mining Pic (RM) performed a techno-economical feasibility study for the Skorpion mine (Namibia) during 1997 and 1998. At that time, RM entered in contact with the R&D Division of Tecnicas Reunidas S.A. (TR). The objective was to apply the modified ZINCEX® process (MZP) to the treatment of Skorpion ore to produce SHG zinc. One of the most important stages of the feasibility study was the piloting of key sections of the process; namely, leaching, solvent extraction and electrowinning.

As presented in this paper, the Skorpion feasibility study was a complete success, and nowadays this project is close to becoming an industrial reality. After this first project, TR has been working on two other feasibility studies, and presently there are other similar projects with different degrees of development in which TR is involved.

BACKGROUND OF THE MODIFIED ZINCEX® PROCESS

Based oh the original ZINCEX process developed by TR during the 1970's, two industrial plants (8,000 and 11,500 tonnes of zinc per year) were built in 1975 and 1979. These plants were successfully commissioned and each of them operated for about 15 years (1-3).

During the eighties and the beginning of the nineties, the original ZINCEX process was refined and simplified (4,5), resulting in the modified ZINCEX® process, MZP. The MZP was experimentally confirmed and demonstrated within an European Union research project called "Recuperation of Zinc from Secondary Materials" that was carried out in collaboration with other European companies. In the early 90's the application of the MZP to the treatment of electric arc furnace dusts produced by members of the Elansa consortium of Spanish steel companies was explored (6-12).

The ZINCEX® process was further improved with additional industrial design data, scale-up factors, materials, etc. through 15 years of operating demonstration plants, prototype installations and industrial models of the critical equipment in the R&D Centre of TR.

A small, industrial plant for the treatment of 2,000 t/y of spent batteries was built in the surroundings of Barcelona in 1996. This plant was commissioned in 1997 using the MZP to recover zinc and manganese from these batteries. Nowadays, this plant is in full operation, being the only known plant in the world successfully applying a hydrometallurgical process to recycle materials contained in spent domestic batteries (13). This plant is a good example of how the MZP can produce high quality zinc electrolytes from pregnant leach solutions (PLS) containing a high level of impurities. In the battery treatment plant, the PLS feed to solvent extraction (SX) contains up to 30 g/L of chlorides and 50 g/L of manganese, as well as minor impurities such as copper, cadmium and nickel. The presence of these impurities does not impair the performance of the SX circuit, as the quality of the pure zinc-loaded solution is suitable for electrowinning. Thus, the larger industrial implementation of the MZP has become a larger possibility that is being confirmed by development of the Skorpion zinc project.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 753

PROCESS DESCRIPTION

The MZP is based on three steps, which are shown in Figure 1 and are discussed below.

Figure 1 - Conceptual Block Diagram of the Modified ZINCEX® Process

Leaching

The zinc raw material at the proper particle size is mildly leached at atmospheric pressure with dilute sulfuric acid at a controlled pH at about 50°C and for the required residence time. The reaction mixture is purified by precipitation with limestone and/or lime at controlled pH to remove some impurities, mainly iron, silica and aluminium. The leach residue together with the precipitated gypsum and hydroxides are thickened and filtered. A zinc PLS and a leach residue are produced.

Solvent Extraction

The PLS containing zinc and impurities is sent to the extraction stage of a SX circuit where an organic phosphoacid solution in kerosene is used as the organic extractant, specifically Di-2-Ethyl Hexyl Phosphoric Acid, D2EHPA, is recommended. Here, the zinc is transferred selectively to the organic phase and the acid released by the D2EHPA, equivalent to the extracted zinc, is transferred to the aqueous phase (acid raffinate). The acid raffinate is recycled back to the leaching step, except for a small fraction that is bled off to avoid impurities building up and to control the water balance in this first aqueous closed circuit.

The zinc-loaded organic solution is washed with acidified water to remove aqueous entrapment and traces of co-extracted impurities in a washing/scrubbing stage. The cleaned zinc-loaded organic is stripped using spent electrolyte to produce an extremely pure zinc-loaded electrolyte.

Electrowinning

The zinc-loaded electrolyte is fed to the electrowinning circuit, where zinc is plated on aluminium cathodes. The electrowinning, melting and casting units are almost conventional technologies, producing special high-grade zinc, SHG, (99,995 %) commercial slabs.

754 LEAD-ZINC 2000

Treatment of Oxide Ores

Direct treatment of oxide, silicate and carbonate zinc ores is difficult if "conventional" zinc technology is used (leaching, purification by means of cementation with zinc powder and zinc electrowinning). There is only one case (Padaeng, Thailand) where conventional zinc flowsheets were successfully modified to treat zinc silicate ores. This was possible given the high-grade zinc ore treated (30% Zn) and the fact that this plant was able to solve the solid-liquid separation problems.

When the ores have high levels of leachable fluorides, chlorides, magnesium, etc., conventional flowsheets are not easily applied. Even if they could be applied, conversion costs would be very high. This is one of the reasons that the MZP was selected for the Skorpion project. Some of the important characteristics of the MZP technology are:

• It is not necessary to finely grind the ore. A particle size with Deo between 200 and 500 μπι is enough to obtain good zinc recoveries.

• The ore is directly leached. Neither ore flotation nor roasting is needed. • Very low-grade zinc ores can be treated; e.g., with 5% zinc. • Because of the flexibility of the process, and with no large additional modifications, it is

possible to treat different zinc minerals that could appear during the life of the mine. • Because of the leaching conditions and use of dilute solutions, solid-liquid separation is

more feasible than when conventional technology is applied to these ores. • It is possible to work with high levels, 1 o 2 g/L, of elements such as Ni, Co, Cd and Cu

in the PLS with no contamination problems in the electrowinning circuit. Also, there are no chloride, fluoride, calcium and magnesium contamination problems in electrowinning. This results in less corrosion, scaling, maintenance and operating problems. The SX zinc technology is a physical barrier to impurities because there are two different aqueous circuits separated by the organic circuit: the leaching-extraction and stripping-electrolysis circuits.

UPCOMING ZINC MINE PROJECTS WITH THE USE OF MODIFIED ZINCEX® TECHNOLOGY

There are some upcoming mine projects which are currently at different stages of development. In some of them, TR has been involved developing complete flowsheets, whereas in others, TR has been asked to incorporate its SX technology in a broader circuit. A brief summary of the status of each major project is provided in the following sections.

Skorpion Zinc Project (14,15)

Reunion Mining approached TR in 1997. After a series of preliminary laboratory tests to confirm that the MZP could be applied to the Skorpion ore, a pilot plant was commissioned. The total duration of the laboratory-pilot plant program was 8 months.

The selected flowsheet for the pilot plant was the MZP with a variation in the bleed treatment. Basic zinc sulfate precipitation was used instead of zinc removal by an additional SX depletion stage. A simplified block diagram of the MZP pilot plant flowsheet is shown in Figure 2.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 755

The main specifications of the pilot plant were the following:

• Production capacity 7.5 kg/day of zinc. • Operation type Continuous (24 h/day), 5 Runs • Total piloting time About 800 h • Electrodeposition cycle 48 h

Figure 2 - Simplified Block Diagram of the Skorpion Project

One of the most important requirements of the pilot plant campaign was to successfully demonstrate the capability of operating with 48-h electrowinning cycles. This need arose because zinc cathode stripping had to be done automatically in the industrial plant according to Union Miniere (UM) parameters. Another important issue was testing the process with a level of impurities in the circuit higher than expected in the industrial plant. After calculating the impurity levels by means of a simulation tool, the PLS produced in the pilot plant was continuously doped with Cu, Ni, Cd, Co, Cl", F", Na, etc. to test the selectivity of the SX step. In Table I, the main components of the ore and the concentration of these elements in the PLS feed to the SX circuit are shown.

Table I - Ore and PLS Typical Compositions in Skorpion Zinc Pilot Plant (1998) Zn Fe Si AI Cu Cd Ni Co _ F CT_

0.2 7

The pilot plant was also used to perform tests with samples produced in-situ using equipment from different suppliers. This was used, for example, to determine the best type of filter and to obtain adequate design parameters. This information was used in the feasibility study. The results obtained during piloting can be summarized as follows:

Ore % PLS g/L

8-14 30-40

2-3 <0.01

22-27 <0.1

4-6 <0.5

0.1-1 2.9

0.1 1.1

0.3 0.7

0.005 0.1

756 LEAD-ZINC 2000

Leaching efficiency Overall zinc recovery SX Zn selectivity Current efficiency in EW Energy consumption (kwh/t Zn) A.C Zinc quality SHG (99.99%)

Zinc morphology Main reagent consumptions

>95% >93% Very high >91% <3,100 Always achieved. Certified by independent LME laboratory Excellent for melting H2S04

CaC03

Ca(OH)2

HC1 Extractant

1.5t/tofZn 0.9t/tofZn 36kg/tofzinc 12kg/tofzinc 1 kg/t of zinc

The high performance of the process was proven in the pilot plant campaigns, illustrated in Figures 3 and 4. Performance was demonstrated by a high zinc recovery, low power consumption, and the high quality of the SHG Zn produced. Analyses of zinc plates produced in the Skorpion pilot plant on May 9, 1998 and certified by Union Miniere are:

Zn 99.995 %

AI Cd <5ppm <lppm

Cu <5ppm

Fe <2ppm

Pb Sn Tl 12ppm <lppm <5ppm

Figure 3 - Skorpion Pilot Plant at Tecnicas Reunidas R&D Centre (1998)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 757

Figure 4 - Zinc Cathode Stripping During the Skorpion Zinc Pilot Plant Campaign at TR (1998)

Based on these results, the feasibility study for a plant with a capacity of 150,000 t/y of zinc showed that the Skorpion zinc project would be the most profitable zinc project in the world. The following figures taken from a DCF financial study back up the above statement:

• Total capital expenditure (mine, refinery, infrastructures): about $300 US million. • IRR for the base case: 25%. • Zinc production cost: about 25 c/lb of zinc.

Presently, the Skorpion zinc project is owned by Anglo Operations Ltd. (AOL), a 100% subsidiary of Anglo American Pic, which is performing a new drilling campaign to better assess the ore reserves. AOL has also built and operated a pilot plant, with TR assistance as technology supplier, at its facilities during the year 2000. The flowsheet used in the Spanish pilot plant has been tested. Piloting in South Africa confirmed the good results obtained in 1998 in the TR pilot plant and demonstrated the robustness of the process. Therefore, it is likely that the launching of the industrial project would happen as this paper is presented in October, 2000.

Sanyati Zinc Project (16)

Sanyati is a copper mine located in Zimbabwe. The project was developed during 1998 with a flowsheet for zinc SX and EW, similar to the one described above. The feed was a solution containing copper and zinc. This solution was a PLS from an existing copper heap leaching operation, but because of the concentration of zinc in the ore, the zinc concentration in the copper PLS was building up to levels that made recovery of the zinc by the MZP feasible and profitable.

The Sanyati zinc project is smaller compared to Skorpion, only 5,000 t/y of zinc production, but it is a very good example of how the MZP can be used to recover zinc from a very impure PLS and how it can be integrated with other processes. The flowsheet was piloted in continuous operation for about 600 hours, and the main results were similar to those obtained during the Skorpion pilot plant campaign. The results obtained during piloting can be summarized as follows:

758 LEAD-ZINC 2000

• Overall zinc recovery >99% • SX selectivity Confirmed • Current efficiency in EW 93% (24 h cycles, manual stripping) • Zinc quality SHG (99.99%) minimum Always achieved. Certified by

independent LME laboratory • Zinc morphology Excellent for melting • Main reagents consumption: H2SO4 0.13t/tofZn

CaC03 0.9 t/t of Zn HC1 9kg/tofzinc Extractant lkg/tofzinc

The feasibility study showed that the installation of the MZP, integrated into the existing plant, would be profitable while solving the zinc accumulation problem.

Other Upcoming Zinc Projects

There are various other mine-zinc projects, all of them based on vertical integration of the refinery and mine where the MZP could be successfully applied. These projects are currently at different stages of development.

Shaimerden Zinc Project

The deposit is located in Kazakhstan; it is a mixture of zinc oxides, silicates and carbonates with zinc contents of about 20%. The companies involved are the Zinc Corporation of Kazakhstan and Ennex International Pic (Ireland).

The feasibility study for a 100,000 t/y zinc plant was finished in March 1999. This study was optimized by Union Miniere Engineering (UME) and Tecnicas Reunidas in 1999 and at the beginning of 2000. The revised feasibility study shows that the economics of the project are very attractive.

At the same time and after completion of the engineering study, UME and TR have submitted a lump sum turn key proposal for construction of the whole plant. The companies involved in the project are studying the proposal to proceed with financing the project.

Lanping Project

This deposit is situated in Yunnan, China. The company which has the rights for mine development is Yunnan Lanping Non-Ferrous Metals Ltd. This mine is considered one of the biggest zinc oxide mines in the world. Its ore is a mixture of zinc oxide and sulfide. Zincox Pic, with the technological support of UM and TR, presented a preliminary feasibility study and project development proposal for a 300,000 t/y zinc plant. The Yunnan Lanping company is evaluating that proposal as well as an alternative preliminary feasibility study carried out by Billiton SA Ltd. using bioleaching for the zinc sulfide ore and atmospheric leaching for the oxide ore. In both leaching processes, the use of zinc SX for zinc recovery from the blended PLS's has been considered, for which TR has provided assistance to Billiton.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 759

Hellver Tailing Retreatment Project

These tailings are located in Australia and are the result of former exploitation of the Hellyer mine. The owners are Dominion Mining Ltd. and Western Metals Resources Ltd.

The project consists of the re-treatment of existing tailings to recover gold, silver and zinc using pressure leaching and conventional technologies for gold and silver recovery, as well as TR's zinc SX and EW for SHG zinc production. The zinc plant will have a capacity of 25,000 t/y and presently Dominion is leading a preliminary feasibility study of the project which looks attractive.

Angouran Zinc Mine

The Angouran mine is a very well known mine which has been exploited only at low capacities. The Angouran deposit is located in northwest Iran and is comprised of oxidized zinc ores, mainly carbonates and silicates.

The company holding the rights to the mine is Iran Zinc Mines Development Co. (IZMDC). After some prospective works, a project to build a plant to produce 100,000 t/y of zinc is being defined. An international competitive tender was completed in early 2000. Zincox Pic, with the technological support of UM and TR, presented a preliminary feasibility study and project development proposal for that capacity. IZMDC has not decided which company is going to carry out the feasibility study.

Jabali Project

The mine is located in Yemen and Anglo American Pic and Zincox together with a local partner own the rights to the mine. The initial idea was to install a zinc refinery with a capacity of 65,000 t/y of zinc. A preliminary study has been performed by Zincox showing encouraging results. Further studies will be undertaken in the near future when more adequate economical and political circumstances are in place.

PRESENT AND FUTURE OF ZINCEX59 TECHNOLOGY

In this paper the origins, initial steps, development and present status of the modified ZINCEX® technology have been compiled. The process offers an array of possibilities that can be used to treat widely available, non-exploited oxidized zinc ores. TR will continue looking worldwide for zinc projects with similar characteristics. Moreover, TR is exploring the integration of zinc SX with other existing or emerging technologies, like pressure leaching or bioleaching, to enlarge the field of application of the MZP.

The policy of TR is continuously to explore all new possibilities to apply the MZP to any zinc material different than oxidized or carbonated zinc ores, and zinc secondaries. An example would be the treatment of zinc sulfides that are difficult to process via conventional roast-leach-electro winning technology.

760 LEAD-ZINC 2000

CONCLUSIONS

Beside previous successful industrial applications for secondary raw materials, the MZP has been demonstrated as a robust, reliable, and suitable technology capable of treating primary zinc ores containing between 5 and 30% zinc. This has been proven by several pilot plant campaigns and feasibility studies.

The Skorpion zinc project is currently proceeding based on the application of the MZP. It is a very promising project with a low operating cost, given the integration of the refinery and the mine. This project will likely become an operating industrial plant in a couple of years. The authors hope to have another paper in the next Lead-Zinc conference to describe the first five years of operation of the Skorpion zinc plant.

Finally, we would like to anticipate that the implementation of the MZP will cause a breakthrough in the zinc mining business and technology. In consequence, we foresee a future in which zinc SX plants will be used all over the world to allow the treatment of ores that were previously discarded as untreatable and valueless materials.

ACKNOWLEDGMENTS

The authors, on behalf of TR, thank all the companies and their technical and managerial staff for trusting TR to successfully demonstrate the benefits of ZINCEX® technology.

REFERENCES

1. E.D. Nogueria, J.M. Regife and A.M. Arcocha, "Winning Zinc Through Solvent Extraction and Electrowinning", Engineering Mining. Vol. 180, No. 10,1979,92-94.

2. E.D. Nogueira, J.M. Regife and P.M. Blythe, "Zincex - The Development of a Secondary Zinc Process", Chemistry and Industry. No. 2,1980,63-67.

3. E.D. Nogueira, J.M. Refige and M.P. Viegas, "Design Features and Operating Experience of the Quimigal Zincex Plant", Paper presented at the 111th TMS Annual Meeting. Dallas, Texas, 1982,59.

4. D. Martin et al., "Process for the Production of Electrolytic Zinc or High Purity Salt from Secondary Zinc Raw Materials", USA Patent. No. 4401531,1983.

5. E.D. Nogueira, J.M. Regife, D.M. San Lorenzo and G.D. Nogueira, "Using Zinc Secondaries to Feed an Electrowinning Plant", Zinc' 85. K. Tozawa, Ed., The Mining and Metallurgical Institute of Japan, Tokyo, Japan, 1985,763-781.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 761

6. G. Diaz and D. Martin, "The Modified Zincex Process: The Clean, Safe and Profitable Solution to the Zinc Secondaries Treatment", Paper presented at the 3rd European East-West Conference on Materials. Symposium E: Recycling of Materials in Industry. Strasbourg, France, 1992.

7. G. Diaz and D. Martin, "The Modified Zincex Process: The Clean, Safe and Profitable Solution to the Zinc Secondaries Treatment", Paper presented at the International Solvent Extraction Conference (ISEO. London, UK, 1993.

8. G. Diaz and D. Martin, "Modified Zincex Process : The Clean, Safe and Profitable Solution the Zinc Secondaries Treatment", Resources. Conservation and Recycling. Vol. 10,1994,43-57.

9. G. Diaz, D. Martin and C. Lombera, "Zinc Recycling Through The Modified Zincex Process", Paper presented at 6th International Conference on Recycling Lead and Zinc into the 21st Century. Madrid, Spain, 1995.

10. G. Diaz, D. Martin and C. Lombera, "An Environmentally Safer and Profitable Solution to the Electric Arc Furnace Dust", Paper presented at the 4th European Electric Steel Congress. Madrid. 1992.

11. G. Diaz, D. Martin and C. Lombera, "Zinc Recycling Through The Modified Zincex Process", Paper presented at the 2nd International Conference on the Recycling of Metals. Amsterdam. (The Netherlands). October 1994. the 5th European Electric Steel Congress, Paris, France, June 1995.

12. Tecnicas Reunidas R&D Division, "Planta para Recuperacion de Cine y Plomo de Oxidos de Acerias Electricas", Report 6236. Clients: Almagrera/Siderinsa, 1992, Elansa, 1993.

13. D. Martin, M.A. Garcia, G. Diaz and J. Falgueras, "A New Zinc Solvent Extraction Application: Spent Domestic Batteries Treatment Plant", Paper presented at the International Solvent Extraction Conference (ISEO. Barcelona, Spain, 1999.

14. Ticnicas Reunidas R&D Division, "Modified ZINCEX® Process Applied on Skorpion Zinc Ore",Report ITR/P-4501/1998 and ITR/P-4694/1998. Client: Reunion Mining, Namibia, 1998.

15. Reunion Mining PIC Report, "Skorpion Zinc Project Feasibility Report", Report November 1998.

16. T&nicas Reunidas R&D Division, "SHG Zinc Plant by the Modified ZINCEX® Process" and "Modified ZINCEX® Process Applied on Sanyati Copper PLS", Report ITR/P4696/1998. Client: Munyati Mining, 1998.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 763

THE GALVANIC STRIPPING TREATMENT OF ZINC RESIDUES FOR MARKETABLE IRON PRODUCT RECOVERY

J. Antonio Barrera-Godinez Departamento de Ingenieria Metalurgica

Facultad de Quimica, UNAM C. U., Mexico D. F. 04510, Mexico

J. Sun and T.J. O'Keefe Graduate Center for Materials Research

University of Missouri-Rolla Rolla, Missouri, U.S.A., 65409-1170

S. E. James Big River Zinc Corporation

Sauget, Illinois, U.S.A., 62201

ABSTRACT

The solvent extraction separation of iron from a zinc sulfate medium using DEHPA was investigated. After loading, the ferric ions were reduced in the organic phase using metallic zinc. The ferrous ions produced could be easily stripped into a sulfate solution with pH values in the range of 1.4 to 2.0. The small, laboratory scale pilot system used in this study allowed for continuous cycling of the organic and strip aqueous solutions. It was possible to establish and evaluate a series of steady state conditions under selected operating paiameters. The effects of processing parameters such as solution chemistry, flow rates, zinc metal loading and pH on iron recovery and efficiency were also evaluated. Under proper conditions, it was possible to remove about 5 g/L iron from the organic phase in one stage and to produce a concentrated iron sulfate strip solution containing ferrous ion concentrations in the range of 90 to 130 g/L. Preliminary feasibility tests were also made on electrowinning metallic iron from the strip solutions and crystallizing ferrous sulfate as a means of making a marketable iron product.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

764 LEAD-ZINC 2000

INTRODUCTION

The significant quantities of impurities that must be removed from zinc processing streams in the roast-leach-electrowin production of the metal present a major technical and economic problem. These impurities are usually removed either by a chemical precipitation method or by cementation using zinc metal as the reductant, depending on the nature of the elements present. Of all the impurities associated with zinc production, iron has proven to be one of the most troublesome. Over the past decades, a number of novel iron precipitation processes were developed and successfully used to deal with the iron. The jarosite, goethite, paragoethite and hematite processes (1,2) are the most prominent and effective of those being used commercially; however, one drawback in their use is the large volume of iron residue generated. The residues pose a possible environmental problem (3). Consequently, it would be desirable to find an alternative method that gives a more acceptable waste product or in which the iron is recovered in a marketable form.

In a recent comprehensive survey article on the recovery of iron from zinc-sulphuric acid processing solutions, the use of solvent extraction and ion exchange was reviewed (4). Although some of the technologies evaluated showed promise, it was concluded that when both the extraction and stripping aspects are considered, no commercially proven system is available. In particular, the need to generate a "useful" iron strip solution was cited as a critical need. The primary objective of the experimental work presented in this paper is to evaluate the use of galvanic stripping of iron from an organic solvent as a possible means of addressing the significant challenge of producing a concentrated solution for the recovery of an iron by-product.

Galvanic stripping (5), as the name implies, uses an electrochemical driving force to cause the reduction of a cation. The unique aspect of this process is that the reactions are conducted directly in the organic phase using a solid, reactive metal as the reducing agent. The reactions taking place are comparable to the simple displacement reactions observed in standard aqueous cementation processes. The main difference is that in galvanic stripping an organic solvent can be directly utilized. Even though the organic solutions have very limited electrolytic properties such as conductivity, it was found that there is sufficient localized electrochemical activity to allow short-range displacement reactions to occur. There are two primary types of spontaneous cathodic reactions that can occur during galvanic stripping. One involves a redox reaction in which there is a partial reduction of a cation, such as Fe37Fe2+, Ce4+/Ce3+ or Sn47Sn2+, or the cation may also be completely reduced to the metallic state or to a gas, as for hydrogen.

In this application to recover iron from a neutral leach zinc residue, a DEHPA extractant is used in conjunction with metallic zinc as the active metal to reduce the ferric ion. Although DEHPA is an excellent solvent for the separation of ferric ion, the subsequent stripping step is difficult. Conversely, the stripping of ferrous ion from DEHPA is easily accomplished, even with dilute acid solutions in the range of pH 1.5 to 2.0. However, the concentration of ferrous ion that could be obtained using galvanic stripping was not known. Consequently, one of the primary objectives of this study was to determine if a relatively concentrated iron solution could be produced.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 765

GALVANIC STRIPPING

Overview

The galvanic stripping reactions involved in separating iron from a zinc sulphate electrolyte using solvent extraction can be assumed to comprise the following half cell reactions. The major anodic and cathodic steps are:

Zn2*R2 + 2/T + 2e Anodic Step (1) org

Fe2*R? + RHnrv Cathodic Step (2) org *

giving a net reaction of:

Zn + 2Fe3tR3 - Zn2*R2 + 2Fe2'R2 (3) org org org

Other competing cathodic reactions can occur, and these cause an increase in zinc usage which adversely affects the process efficiency. Examples of these are hydrogen gas evolution or oxygen reduction. Other metal cations or impurities that load into the organic may also be reduced depending on the processing parameters and the electrochemical nature of the impurities.

As indicated by reaction (3), one mole of zinc will theoretically reduce two moles of ferric ion to ferrous ion to give an ideal stoichiometry number of one. The actual value obtained in practice is always higher because of the non-productive side reactions, in particular hydrogen evolution. Because of this, one important economic aspect of the galvanic stripping process is to identify the operating conditions which minimize the stoichiometry number.

The reactions occurring in the organic phase are spontaneous and the reaction direction is dictated by the potentials of the metal/ion systems. As in the case of aqueous cementation, the more noble ion in the solution will be reduced by a more active base metal. Though these potentials are relative to the particular organic and metal/ion in question, the ordering of the organic systems studied to date is similar to the aqueous electromotive force series. There are also two process variations, simultaneous and separate galvanic stripping, which have been described previously (6). Simultaneous stripping was used in the continuous laboratory pilot study described here.

Once galvanic stripping has been shown to be feasible using batch type screening tests, the next step is to determine the reaction kinetics and the effects of various operating parameters on process efficiency. Some influential factors which have been identified include temperature, metal reductant and surface area, strip solution pH, A/0 ratio, reaction time, organic extractant and diluent. This study had the goal of establishing the parameters which provide the optimum iron recovery, reaction rates and reaction stoichiometry in a continuous operation mode. Various methods to recover the iron in a marketable form were investigated using the strip solution generated in the continuous run.

Zn + 2RHorg

Fe3*R* + e

766 LEAD-ZINC 2000

EXPERIMENTAL

Simultaneous galvanic stripping was used in the continuous flow tests, but some batch tests were conducted to supplement these studies. Zinc dust was the only metal used as a reductant in the continuous runs, but scrap iron has also been used successfully in other batch experiments. Alternative stripping solutions based on acidic chloride media may also be used, but sulfate was employed in the majority of the work.

Continuous Flow Process

The schematic flowsheet arrangement used for the continuous flow system consisted of one-stage loading and one-stage simultaneous galvanic stripping with pH control, as shown in Figure 1. The temperature was held constant at 40 °C and the flow rates remained constant during any given steady state condition. Each stage involved a mixer and a settler, 160 and 200 ml effective volume, respectively. The mechanical agitation in the range of 1500 rpm in both mixers was sufficient for aqueous and organic phase mixing. In the galvanic stripping reactor, the agitation suspended the zinc particle shot relatively well, but better solid/liquid contact would be advisable. All the mixers and settlers werejacketed glass beakers with polypropylene lids. Teflon rigid tubing and valves were used to deliver and draw fluids from the various beakers with Cole Parmer calibrated peristaltic pumps. Suitable flexible inert tubing was used to transfer the fluids between the mixers, settlers and flasks. Temperature control was accomplished using a controller immersed in a separate bath which pumped water through the jacketed beakers. The acidity of the ferrous sulfate product stream was continuously monitored by a pH controller which in turn activated a peristaltic pump to feed acid to control the strip solution pH at the desired value.

Figure 1 - Bench Scale, Simultaneous Galvanic Stripping Flowsheet Arrangement

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 767

Process Streams

A zinc neutral leach residue was dissolved in sulfuric acid to produce the high-iron concentration solution used in the tests. The resulting leach solution, supplied by Big River Zinc Co., contained 8.76 g/L iron, 130 g/L zinc, 40 g/L sulfuric acid with Cd, Cu, Co, Ni, As, Ge, Ca, Pb, Sn, Tl and Sb as trace impurities. Aqueous solution (Al) was contacted with the organic phase (02) during the loading stage giving a low iron solution (A3), which would be returned to the purification circuit for standard zinc recovery, and the iron loaded organic (04).

Pure phosphoric acid bis(2-ethylhexyl) ester, (DEHPA), supplied by Albright & Wilson, was diluted to 30 vol% using a commercial diluent, SX-12 supplied by Phillips Mining Chemicals, which contained 22 wt% aromatics, 42 wt% napthenes and 36 wt% paraffins. A 50 g/L H2S04

solution was used to wash the organic phase for 30 minutes at 40° C before use. Then the organic was mixed with deionized water for 15 minutes. This solution was continuously recycled as the organic phase in the galvanic stripping process and was designated as stream 04 after loading and as stream 02 after stripping (Figure 1).

Two stripping solutions were prepared from deionized water and reagent grade sulfuric acid of either 400 or 250 g/1 H2S04 concentration. This solution made up stream Al 1 which was fed directly into the galvanic stripping reactor with the return strip solution A9. A high acid concentration was utilized to produce a more concentrated iron solution and the low concentration was utilized to study the effect of the pH.

After stripping, the A10 aqueous exit solution containing mainly ferrous sulfate, was split into the product stream, A12, and the return stream, A10, going into the galvanic stripping reactor. Therefore streams A9, A10, and A12 have the same chemical composition but different volumes and flow rates. The iron was allowed to build up in these streams until a steady state was reached.

Reductant

Zinc shot containing 99.7 wt% zinc was used as the reducing agent. The particles were spherically shaped and in the range of -8+12 mesh. The surface area of these particles is approximately 4.1 cm2/g.

Procedure

The system chemistry was allowed to build up rather than starting with synthetic solutions in the vessels. Measured volumes of the fresh organic phase were initially placed into the mixers and settlers of both the loading and galvanic stripping beakers. The Al solution was added into the mixer for loading. Similarly, the reactor and settler of the galvanic stripping step were filled with measured volumes of fresh organic and sulfuric acid solution. The pH electrode receptacle was filled with acidified water and the pH controller set to the desired value. The system was purged with nitrogen for 15 minutes before adding the reductant to the stripping mixer and starting the pumps. A nitrogen flow rate of approximately 20 ml/min was maintained while the system reached the steady state. The experimental conditions were: temperature 40° C, organic flow rate of 2 ml/min, a reductant addition giving an equivalent surface area of 164 cm2, an aqueous to organic ratio (A/O) of 1.0 during loading and 2.0 during stripping.

768 LEAD-ZINC 2000

Stream Analyses

Organic and aqueous samples were taken at intervals as steady state was approached. Iron and zinc concentrations in the organic and aqueous samples were estimated at the University of Missouri-Rolla using a model 2501 Portaspec X-ray fluorescence portable analyzer. Atomic absorption spectroscopy and inductively coupled plasma analyses were made by Big River Zinc to determine more accurate values, including trace element concentrations in the organic and aqueous solutions.

Iron removal from the organic and the amount of zinc present in the organic were calculated based on the sample concentrations. The stoichiometry number, defined as the mass ratio of reductant actually consumed to reductant stoichiometrically required to reduce the iron, was also determined. An approximate mass balance was established for each sampling time, and this allowed an estimation of the galvanic stripping process rate. This parameter represents the net amount of iron reduced and stripped from the organic phase, per unit of surface area of zinc reductant per unit of time and is expressed in g Fe/cm2/min. In this continuous flow galvanic stripping set-up, the rate of zinc particle dissolution was assumed to be small enough to assume that the reductant surface area does not change significantly during the experiment.

The experiments were designed to determine if the process could be run in a continuous mode while producing results similar to those established in the batch tests. Of particular interest are the rates of iron removal related to the solution chemistry, solution concentrations, overall process efficiency, the distribution of zinc and impurities and the concentration of iron that could be obtained in the final product stream.

RESULTS AND DISCUSSION

An overview of some general data generated and the results obtained during the experiments is given in Table I.

Table I - Average Response Variables for Continuous Galvanic Stripping of Iron Room temperature pH value Iron stripped (removed) Stoichiometric index Zinc consumption Iron removal rate Iron concentration in product stream Zinc concentration in product stream Product flow rate

IRON LOADING (ZINC STRIPPING) Iron loaded Zinc stripped Iron concentration in leach stream after loading

% n.a. g/hr

μg Fe/cm2/min

g/i g/i

ml/hour

% %

Φ

1.5/2.1 55/25

1.35/4.1 0.53 74.15

90/120 .04/.02

6/8

44 89.3 4.9

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 769

The results indicated that iron removal was affected by the aqueous pH value. Lower pH values can be considered if higher iron removal is required. However, the zinc concentration in the product stream will increase as the pH of the strip solution decreases. Even though the iron removal varied widely in these experiments, high stripping percentages (83%) have been consistently obtained in previous tests. The zinc present in the organic is dependent on the stoichiometry number, and the concentration in the product stream can be relatively low, depending on the strip solution pH. In general, the process is capable of separating iron from zinc very effectively since the zinc contained in the return organic phase 04, is stripped back into the A3 solution during reloading. The values of the stoichiometry number and the reductant consumption obtained suggest that increasing the pH enhances the reductant consumption, and this was not expected. The additional zinc dissolution is probably due to a relative increase in hydrogen ion reduction compared to iron reduction, which may be adversely affected by the pH increase.

The iron removal rate is lower as the pH and iron content increase in the strip solution, and this is possibly due to the lower stripping power or an equilibrium condition. Since DEHPA is an acid extractant, its loading and stripping equilibria are controlled by the acidity in the system. Also, these equilibria are affected by the iron activity, which depends on its concentration. The high aqueous iron concentration probably displaces the iron stripping equilibrium to lower pH values. This forces the system to require a higher free H+ concentration in order to achieve the same stripping power that a higher pH produces at a lower iron concentration. Thus, the greater, the iron concentration in the strip aqueous phase, the more difficult it is to use this solution to remove iron from the organic phase, while still keeping the zinc concentration low. However, the rate values obtained are typical of those determined in previous tests, and additional studies are needed to quantify these interactive relationships more precisely.

The chemistry of the product solution must be controlled by manipulation of several parameters whose values are based on the final desired concentrations of iron and zinc in the process streams. The maximum iron concentration attainable is proportional to the acid and iron concentrations in the strip solution. The zinc concentration in the product stream can also be manipulated by the acidity of the stripping process, since the zinc-DEHPA loading-stripping equilibria are also pH dependent.

When run continuously, the loading of iron from for Al into 02 is strongly dependent on the stripping conditions. The unshipped iron remaining in the returning organic phase limits the uptake of additional iron.

Another test was run using 70% more reductant at a pH = 1.80 and under the same conditions as the previous tests. The percent iron removal doubled, the zinc concentration remained low, but the stoichiometry number increased. In summary, the results obtained are very dependent on the operating conditions and a compromise among these parameters and the proposed use of the ferrous solution must be made.

An example of the chemical compositions of the various streams is given in Figure 2, where the values shown are those after the sixty-two hours of running.

770 LEAD-ZINC 2000

2 ml/min 8.76 g Fe/I 130gZn/l

A1

Loading Mixer

2 ml/min 5.12 g Fe/I 5.24 g Zn/I

4ml/min

> Loading Settler

04 02

Zn 0.52 g/hr

2 ml/min 4.92 g Fe/I 133 gZn/l

A3

2 ml/min 11.12 g Fe/I 0.58 g Zn/I

> ' >r

Galvanic Strip

Reactor

6ml/min Strip

Settler

4 ml/min 89.75 g Fe/I 0.415 g Zn/I 0.098 ml/min

250 g/l H2S0,

A9

A10

A11

A12

4 ml/min 0.098 ml/min 92 g Fe/I 92 g Fe/I 0.425 g Zn/I 0.425 g Zn/I

Figure 2 - Chemical analyses and flow rates after 62 hours continuous processing

IRON RECOVERY

As mentioned previously, one desired outcome of this study was to determine the composition of the final strip solution generated in a continuous galvanic stripping operation. Obtaining a correlation between the concentration of ferrous sulfate and the pH needed for effective stripping was of particular interest. Previous batch tests had shown that concentrations of ferrous ion in the range of 90 g/L could be obtained using pH values of 1.7 to 2.0. The pH value was important because of the need to minimize zinc stripping, which could be an undesirable impurity in some application. If the zinc impurity content in the final solution is not critical, then the process efficiency can be improved.

The relationships between strip pH and iron/zinc recovery are illustrated in Figures 3 and 4. In this particular set of tests, it is assumed that about 65% of the ferric ions were reduced to the ferrous state. The reduced iron was completely stripped into an aqueous phase up to a pH of about 2.0 and then the value decreased to 40% at a pH of 2.7, as shown in Figure 3. As expected, the zinc concentration was also sensitive to pH and the amount stripped was minimized when the pH was about 2.0 or higher, as seen in Figure 4.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 771

70

65

60

55 s ■fi «> 50 t o

45

40

1.0

Test Conditions Org. Phase: Loaded BRZ sample

Diluted H2S04 solution at different pH values 438 g/L Fisher Zn Shot 40 C

Aq. Phase: Reductant: Temperature: A/O: Reaction Time:

I IS minutes

I

1.5 2.0

pH 2.5 3.0

Figure 3 - The Efficiency of Fe2+ Stripping as a Function of the Aqueous Phase pH

Figure 4 - The Efficiency of Zn2+ Stripping as a Function of the Aqueous Phase pH

Water Treatment

One possible, but limited, use of the concentrated ferrous sulfate solution is in water treatment. The solutions must be relatively pure, and although the strip assays are promising, a few

772 LEAD-ZINC 2000

potentially troublesome impurities were identified. These included tin, indium, copper, cadmium and lead, which were mostly in the range of 10 mg/1. Arsenic and zinc were the most prevalent impurities being about 100 mg/1 and 130 mg/1, respectively, and are probably the most harmful. Extensive chemical analyses were made by Big River Zinc on the various aqueous and organic streams in the continuous galvanic stripping circuit. The analyses provide a good perspective for the evolution of the impurity distribution with time, and more complete analyses after 62 and 101 hours of running, including impurities, are given in Table II.

Table II - Impurity Distribution After 62 and 101 Hours of Testing

Sample No.

A1-62 Al-102 A3-62 A3-101 A12-62 A12-101 02-62 O2-101 04-62 O4-101

A l : A 3 : A12 :

Concentration (g/L)

Zn Fe 130 129 133 124 0.43 0.25

5.24 5.71 0.56 0.54

BRZ

8.76 8.96 4.92 4.96

92 131

5.12 5.64

11.12 7.76

Cu 1160 1160 1168 1156 2.75 3.00

2.40 0.72 0.48 0.38

Co 33 33 32 31 0.33 0.30

0.015 0.004 0.005 0.008

Ni 52 52 51 50 0.70 0.75

0.03 0.01 0.01 0.01

high iron leach solution Raffinate solution Aqueous solution after stripping

Electrolytic Iron

Concentration (m

As 38.0 40.1 36.0 33.8 39.5 112.3

0.62 0.81 5.26 4.45

Pb 8.4 12.1 83.0 11.2 42.5 10.0

2.1 1.9 2.3 1.7

04: 02:

Ge 14.70 16.10 15.40 14.80 1.45 3.05

0.03 0.06 0.14 0.16

g/L)

In 5.40 5.90 6.30 3.70 0.13

<0.02

47.50 39.80 67.50 29.20

Sn 2.60 2.70 0.53 0.25 2.63 0.50

66.80 54.10 88.80 39.70

Loaded organic Organic after

Cd 1730 1720 1730 1700 13.25 6.75

0.44 0.51 0.74 0.17

stripping

Tl 1.21 1.32 1.28 1.31 0.20 0.13

0.033 0.053 0.017 0.018

Preliminary tests were made to determine the feasibility of using the ferrous sulfate solution as a feed for iron electrolysis. The experiments were conducted in a jacketed beaker using a stainless steel cathode blank with 1 cm2 plating surface. The lead anode was bagged to prevent extensive mixing of any ferric ions generated.

Initially, the short-time deposits were made using synthetic, pure electrolyte to provide a comparative reference. The conditions were in the ranges given below:

Fe2+ (as sulfate) pH Temperature Current density Agitation Time

70 to 100 g/L 1.5 to 2.2 40 - 60° C 25 - 50 mA/cm2

Magnetic stirrer One hour

Current efficiency 82 to 97%

The deposit made from synthetic electrolyte with 89% current efficiency, shown in Figure 5, had a reasonably good morphology. The electrolytic iron deposit was not very ductile, but had a good crystalline structure and was free of dendrites.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 773

Figure 5 - Electrolytic Iron Deposited From Synthetic Solution

The deposits made directly from the generated strip solution with a concentration of ferrous ion of 100 g/L at pH 2.0,40° C and 30 mA/cm2 were very porous and fine grained. The morphology suggested the presence of an organic, as seen in Figure 6. Another test was made using a similar solution after treatment with activated carbon to remove any entrained organic. The deposit, shown in Figure 7, was much better in terms of its morphology and appearance, and it compared favorably with the deposit made from the synthetic electrolyte. This particular deposit was plated at a current efficiency of 82% which is at the low end of the values obtained.

Figure 6 - Electrolytic Iron Deposited From Galvanic Strip Solution

Overall, the results were promising in that they demonstrated that direct electrolytic iron production was feasible without extensive solution purification. Of course, in itself, the process of electrolytic iron is challenging and more extensive studies are needed to define the technical and economic possibilities for this method of recovery.

774 LEAD-ZINC 2000

Figure 7 - Electrolytic Iron Deposited From Activated Carbon Treated Galvanic Strip Solution

Ferrous Sulfate Crystallization

The ferrous ion concentration of the strip solution at the end of the continuous run carried out at 40°C and at low pH was in the range of 130 g/L.

When the aqueous solution was cooled to room temperature, ferrous sulfate crystallized until the aqueous solution concentration decreased to about 113 g/L. The precipitated product was relatively pure, as indicated by comparing the chemical analyses, shown in Table III, of the initial solution and the solid ferrous sulfate crystals. The distribution coefficient between the aqueous and solid phases for the impurities is also very favorable.

Table III - Analysis of the Precipitated Ferrous Sulfate Crystals Strip Solution FeS04 Crystals

Element mg/L Wt% Fe(g/L) Zn (g/L) As Ce Cu Ni Co Sn Sb Ge In

113 3.6

22.9 1.9 1.0 .13 .06 .65 .05 .71 .09

22.80 0.17 0.003

< .0001 .0001

< .0001 < .0001

.0001 < .0001 < .0001 < .0001

As with the other proposed methods of recovery, these data are very preliminary and a better quantitative correlation between the concentrations and the operating parameters (temperature, pH, A/O ratio, etc.) must be obtained. Evaluating other methods of crystallization

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 775

using more conventional hydrolysis-based processes should also be beneficial. In the continuous work, only sulfate solutions were used for stripping. Acid chloride stripping is also feasible, as reported in the following section, and may even be preferred in some instances. Other anion systems might also be evaluated, depending on the desired form of the final iron product.

IMPURITY REMOVAL

Although the focus of these studies was on iron removal using a redox reaction, complete reduction of a number of other impurities using galvanic stripping has also been demonstrated. The metal reductant may be varied, depending on the impurities and conditions, but aluminum, zinc, iron, copper, etc. have all been used to remove more noble cations. Examples of cation impurities removed include cadmium, lead and copper and for by-product recovery silver, gold and palladium. It is also important to recognize that organic extractants other than DEHPA can also be used. Batch tests to demonstrate electrochemical activity have been made not only on standard extractants, such as TBP, LIX and Aliquat, but also on newer organics such as the Eichrom Dipex Extractant.

Even though galvanic stripping works reasonably well for bulk impurity removal, such as iron, its use in recovering smaller concentrations of components may be even more attractive. If a particular impurity or valuable by-product can be suitably concentrated using solvent extraction, galvanic stripping offers a direct and simple means of extracting the metal in a solid condensed form. This would eliminate the need to find a suitable strip medium plus reduction process for complete recovery.

ACID CHLORIDE STRIPPING

Although not used in the continuous tests, some batch tests were made to show the use of HC1 as the stripping agent. Similar experiments were also made using H2S04 to allow for a comparison. However, no extensive evaluation of impurity distributions was made.

The test conditions were: 30% DEHPA with SX-12 containing 5 to 6 g/L Fe3+ and 0.4 to 1.4 g/LZn2+ loaded from BRZ leach solution using 25 g/L atomized zinc powder (-60 +100 mesh) at 40°C with an A/O ratio of 1. The experiments were made using a Burrell agitator at a 5 setting and various initial aqueous pH values and reaction times.

In Table IV, the data show that HC1 is an effective stripping agent. The stoichiometry number is in the range of two. The lower pH value again seems preferred and iron removal of about 80% is possible. The value decreases substantially as the initial pH is increased.

776 LEAD-ZINC 2000

Table IV - Effect of the pH of the HCl Strip Solution

Test No.

0314-1

0315-1

0314-2

0315-2

0315-3

Initial

0.77

0.77

1.10

1.10

1.34

pH Final

2.03

1.95

2.20

2.21

2.45

Time (Min)

15

15

15

15

15

Iron Removal Ol

Fe

5.10

5.10

5.10

5.10

5.10

(g/L) Zn

1.40

1.40

1.40

1.40

1.40

02 Fe

1.05

1.05

2.36

2.33

3.70

(g/L) Zn

5.68

5.68

5.55

5.48

5.01

Rem (%)

79

79

54

54

28

These initial tests were done for 15 minutes, and a series was also made by reducing the reaction time to 10 and 5 minutes. The iron removal did not change with time, indicating that the values obtained are related to some type of equilibrium and that more favorable kinetics with lower retention times are possible.

Based on these results, additional comparative tests were made at 5 minutes using either HCl or H2S04 in the strip solution. The results are given in Table V. Nearly 90% iron removal after only 5 minutes is obtained with a pH of 0.7, regardless of the acid used. The HCl was less effective when the pH increased to 1.10, but additional tests are needed to verify this trend. In general, HCl appears to be viable as a stripping agent and comparable to H2S04 in most aspects.

Table V - Effect of pH and Acid Type on Fe Removal

Test No.

0321-1

0321-2*

0321-3*

0321-4

Initial

0.70

0.70

1.10

1.10

pH

Final

1.38

1.12

1.88

2.20

Time (Min)

5

5

5

5

Iron Removal

01(g/L)

Fe

5.10

5.10

5.10

5.10

Zn

1.40

1.40

1.40

1.40

02 (g/L)

Fe

0.66

0.54

1.17

2.33

Zn

4.52

3.84

5.79

5.25

Rem (%) 87

89

77

54

*H2S04 Stripping solution used

CONCLUSIONS

Using the galvanic stripping process, it is possible to remove iron from an acidic zinc sulfate leach electrolyte and to produce a ferrous sulfate solution having an iron concentration as high as 130 g/L. A zinc concentration as low as 250 mg/1 is possible, but the pH of the strip

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 777

solution is critical. If the purpose is to produce a greater Fe-Zn separation, these results indicate that the pH should be in the range of 1.45 to 1.75.

It was found that using more solid zinc or increasing the reductant reaction surface area is one of the main factors affecting the rate and degree of iron removal. The retention time and overall rate of removal of iron will be strongly dependent on the system engineering. Any features which improve the relative solution/reductant contact will have a positive impact on the overall iron recovery. The stoichiometry number is always higher than theoretical, but values in the range of 2.0 or less are routinely obtained.

The acidity of the aqueous strip solution is a very strong variable whose value should be carefully selected in order to effect a compromise among the iron removal, the zinc concentration in the iron product, the stoichiometry of the process, the process rate and the iron-zinc separation index. As demonstrated, it is possible to separate iron from zinc sulfate electrolyte and recover a concentrated, potentially usable form of iron. It is interesting to contemplate the use of galvanic stripping for the removal of other less concentrated impurities or valuable by-products from zinc processing solutions or from other hydrometallurgical streams, either in conjunction with iron removal or separately.

Overall, the data obtained were promising, but some of the limitations of the galvanic stripping that were identified require additional evaluation. On the positive side, it was demonstrated that running in a continuous mode is possible and that a concentrated ferrous sulfate solution can be produced. A number of by-product uses appear to be potentially possible and include water treatment chemicals, electrolytic iron or a precipitated or crystallized solid iron salt.

ACKNOWLEDGMENTS

The assistance provided by the Great Plains Rocky Mountain HSRC, Kansas State University Project 94-05 is gratefully acknowledged.

REFERENCES

1. V. Arregui, A.R. Gordon and G. Steintveit, "The Jarosite Process - Past, Present and Future," Lead-Zinc-Tin '80. J.M. Cigan, T.S. Mackey and T.J. O'Keefe, Eds., The Metallurgical Society of AIME, Warrendale, PA, U.S.A., 1979, 97-123.

2. J.E. Dutrizac, "The Physical Chemistry of Iron Precipitation in the Zinc Industry," Lead-Zinc-Tin '80. J.M. Cigan, T.S. Mackey and T.J. O'Keefe, Eds., The Metallurgical Society of AIME, Warrendale, PA, U.S.A., 1979, 532-564.

3. N.L. Piret and A.E. Melin, "Impact of Environmental Issues on Iron Removal Process Evolution in the Zinc Industry," Hvdrometallurgv Fundamentals. Technology and Innovation. J.B. Hiskey and G.W. Warren, Eds., Society for Mining, Metallurgy and Exploration, Littleton, CO, U.S.A., 1993,499-520.

778 LEAD-ZINC 2000

4. P.A. Riveros, J.E. Dutrizac, E. Benguerel and G. Houlachi, "The Recovery of Iron From Zinc Sulphate-Sulphuric Acid Processing Solutions by Solvent Extraction or Ion Exchange." Min. Pro. Ext. Met. Rev.. Vol. 18,1998,105-145.

5. T.J. O'Keefe, "Method for Stripping Metals in Solvent Extraction," U.S. Patent. No. 5,228,903, 20 July 1993.

6. C. Chang, H. Gu, T.J. O'Keefe and S.E. James, "Galvanic Stripping of Iron from Solvent Extraction Solutions from Zinc Residue Leaching", Iron Control and Disposal. J.E. Dutrizac and G.B. Harris, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1996, 417-428.

7. L.M. Chia, M.P. Neira, C. Flores,.and T.J. O'Keefe, "Overview of Galvanic Stripping of Organic Solvents in Waste Materials Treatment," Extraction and Processing for the Treatment and Minimization of Wastes. J.P. Hager, B.J. Hansen, J.F. Pusateri, W.P. Imrie and V. Ramachandran, Eds., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1994, 279-292.

8. C. Flores and T.J. O'Keefe, "Gold Recovery From Organic Solvents Using Galvanic Stripping", Proceedings of the Separation Processes Symposium. M. Misna, Ed., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1995, 187-202.

9. L.M. Chia and T.J. O'Keefe, "Stripping Lead From D2EHPA by Direct Displacement Reactions Using Metallic Zinc," Proceedings of the Treatment and Minimization of Heavy Metal-Containing Wastes Symposium. The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1995, 29-42.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 779

SEPARATION OF IRON FROM A ZINC SULPHATE ELECTROLYTE BY COMBINED LIQUID-LIQUID EXTRACTION AND ELECTRO-REDUCTIVE

STRIPPING

K. Verbeken, M. Verhaege and E. Wettinck Ghent University

Laboratory of Non-ferrous Metals Technologiepark 9

B-9052 Zwijnaarde, Belgium

ABSTRACT

Iron (III) can easily and selectively be removed from acidic sulphate solutions with organo-phosphoric acids like Di-2-Ethyl-Hexyl-Phosphoric Acid [D2EHPA]. Stripping of the iron (III) loaded organic, however, is difficult unless highly concentrated inorganic acid solutions are used. In this paper results are shown for the enhanced stripping of iron from loaded D2EHPA by electro-reduction. The loaded solvent is contacted with a 2 M H2SO4 solution, which is continuously recycled in an electrochemical reactor where iron (III) is reduced to iron(II). As iron (II) is not extracted by D2EHPA, the equilibrium for iron (III)-D2EHPA is shifted towards the aqueous iron (Il)-sulphuric acid solution. The electro-reduction unit is a vessel with separated catholyte and anolyte chambers. The catholyte is the actual stripping solution where iron (III) is reduced to iron (II) on a stainless steel electrode. The anolyte is a 2 M H2SO4 solution where oxygen is evolved on a lead alloy electrode. A sintered glass plug separates the compartments. With this unit, enhanced stripping of iron from the solvent is possible under mild acid conditions [2 M H2SO4], and a concentrated ferrous sulphate solution can be obtained. The results shown have mainly been obtained by studying the parameters affecting the electro-reductive stripping of a D2EHPA loaded organic phase in a batch reactor. From an iron (IH)-saturated 20 vol. % D2EHPA solvent containing ~ 12.5 g/1 Fe (III) over 90% of the iron can be stripped in one stage with a 2 M H2SO4 solution at 50°C. It is shown that in this way it should be possible to obtain a concentrated ferrous sulphate solution which can be processed downstream to recover a pure iron product.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

780 LEAD-ZINC 2000

INTRODUCTION

Iron is an important impurity in many process solutions. The environmentally friendly removal of iron and its downstream valorisation has attracted the attention of many researchers in industry and universities. In the last decade, much effort has been dedicated to the removal of iron from zinc sulphate electrolytes in the hydrometallurgical processing of zinc feeds (1-7). Similar problems of iron control, however, occur in other acidic sulphate-based process solutions in hydrometallurgy, in electroplating [conditioning and control of iron (III) in iron-zinc co-deposition baths (8)] or in the treatment of spent zinc galvanizing pickle liquors (9). Numerous extractants or extractant mixtures for the selective removal of iron from acidic sulphate solutions have been proposed.

Organophosphoric extractants like D2EHPA and M2EHPA have been extensively studied at the Technical University of Delft, The Netherlands, (2) and at McGill University, Canada (3,4,10). These extractants are very satisfying as far as loading, selectivity towards zinc and stability are concerned. The drawbacks of these extractants are related to anion carry-over, kinetics, and poor stripping behaviour. For some of these reasons, other extractants have been suggested recently. The N-alkyl hydroxamic acids [Monsanto] are able to separate iron and to obtain iron bisulphate by stripping with a mixture of sulphuric and phosphoric acid (5). A consortium of European institutes developed a series of amino-ethylene-phosphonic acids [EU-type], which showed high selectivity of Fe over Zn and faster kinetics than D2EHPA, but the stripping efficiency is lower (6).

Russian workers suggested isododecyl phosphetanic acid [IDDPA] for the extraction of iron from sulphuric acid solutions (11). Stripping efficiency is high, but as this molecule is a rather weak acid, the Fe/Zn selectivity could be a problem. An octylphenyl acid phosphate, OPAP, [Albright & Wilson] has recently been added to the list of potential selective extractants for iron removal from zinc process solutions. Stripping with 4-6 N HC1 yields a highly concentrated iron chloride solution (7). The OPAP is an equimolar mixture of mono- and di-alkyl phosphoric acid esters. Other extractant mixtures have been studied mainly to improve the extraction kinetics or the stripping efficiency. Demopoulos studied a mixture of D2EHPA and KELEX 100 (12). Stripping of iron was only possible with concentrated sulphuric acid [~ 500 g/1 ] which affected the stability of the solvent. Hirato suggested the use of a D2EHPA-TBP mixture to improve the stripping characteristics of iron with hydrochloric acid (13). Chen worked on mixtures containing various commercially available extractants. It was concluded that a mixture of di-2-ethylhexyl phosphonic acid [HEHEHP] and tri-iso-octyl amine [TIOA] could be used to separate iron from zinc. The iron was stripped with H2SO4 solutions at pH ~ 0.7 (14).

It is clear from the above-mentioned literature that efficient stripping of iron from a loaded solvent, resulting in an iron strip solution with a high iron concentration, is a major problem with most of the proposed extractants or extractant mixtures. Accordingly, to enhance the stripping efficiency "reductive stripping" has been proposed.

REDUCTIVE STRIPPING

To facilitate stripping of iron from organophosphorus-type extractants, several techniques have been proposed. All of them are based on the fact that Fe(II) , in contrast to Fe(III), is not extractable. Thus the introduction of a suitable reductant in the stripping unit will

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 781

shift the equilibrium from Fe(III)-organic to Fe(II) -aqueous. This technique is referred to as "reductive stripping" or "galvanic stripping". Demopoulos and Gefvert performed stripping under hydrogen pressure (12). Japanese researchers at the University of Kyoto autoclaved an Fe(III)-D2EHPA-organic phase under sulphur dioxide at pressures ranging from 100 to 300 kPa (15). Stripping of iron from D2EHPA and Versatic Acid solutions with L-ascorbic acid or with iron powder was proposed by Shibata et al. (16). O'Keefe and co-workers also followed the approach of using solid metals to reduce the Fe(III) in the organic phase. The addition of Zn or Zn-Pb-alloys to the stripping reactor promoted the stripping efficiency to a great extent (17,18). All these chemical reductive stripping methods involve high pressures that require autoclaves or produce slurries that have to be filtered.

The use of electro-reduction in a separate electrochemical reactor, connected to the aqueous loop of the stripping section, is an interesting alternative. Leiby and Bricker described a process for the recycling of sulphuric acid from lead acid batteries (19). To remove iron from the battery acid, a D2EHPA-KELEX 100 mixture [Sherex] was used. Electro-reduction cells are integrated into the stripping units. The electrochemical cells were constructed with cathode and anode compartments separated by an anion selective membrane. Long term experiments revealed a loss of extraction power of the solvent, and KELEX had to be added periodically to restore the system. Electro-reductive stripping in a separate loop is also mentioned by Principe and Demopoulos (4). A highly concentrated iron chloride strip is obtained. No details on the electro-reduction unit were given.

Electro-reduction deserves further attention as, by applying this technique for iron reduction and stripping, neither chemicals, reductive metals or alloys, nor an autoclave is needed. Because D2EHPA is frequently referenced as the standard for iron extraction from acidic sulphate solutions, especially for Fe/Zn separation, we decided to study some aspects of the electro-reduction of Fe (Ill)-loaded D2EHPA.

EXPERIMENTAL

Chemicals, Preparation of Solutions and Analysis

All chemicals used to make up the aqueous phases were of analytical grade. The D2EHPA was obtained from Johnson and Matthey Gmbh. Shellsol T [Shell], an aliphatic diluent [aromatics < 0.5 vol % ] , was used throughout the experiments. For the reductive stripping experiments, a batch of "Fe-loaded" organic solution was prepared. The solvent contained 20 vol. % D2EHPA and was loaded with iron by two subsequent contacts with an aqueous solution containing 70 g/1 ferric sulphate at pH 0.5 at an organic to aqueous phase ratio of 1 to 2. This loaded organic contained 12.79 g/1 iron. For the first series of experiments, no zinc was added to the system. From the literature study it was clear that, under the extraction conditions used, the full loading capacity of the organic phase was being utilized, and with strip solutions of 1 or 2 M H2SO4, only a minor transfer of zinc would occur.

The total iron, ferric iron and ferrous iron concentrations in the aqueous phase were analysed as mentioned by Dewit (8). The iron content in the organic phase was determined by stripping a sample with 6 N HC1 using an aqueous to organic phase ratio of 30. In this way the iron was stripped quantitatively from the solvent. The iron content of this solution was then analysed by AAS.

782 LEAD-ZINC 2000

Electro-reduction Unit

The electro-reduction unit is shown schematically in Figure 1. It contained two cylindrical vessels separated by a porous glass sheet. The volume of the catholyte, which was a water-jacketed, baffled vessel [4 baffles] containing a woven stainless steel electrode [active surface 200 cm2], was 500 ml. The volume of the anolyte was 350 ml. The anode, with a surface area of 53 cm2 was a Pb-alloy sheet [0.11%Ca, 0.80%Sn and 0.02% Al]. A power unit with variable tension between 0 and 30 V was used as the DC source. The terminal voltage and cell current were measured continuously during the experiments. At regular time intervals, samples were taken from the reduction unit and were analysed for their ferric and ferrous concentrations.

Figure 1 - Schematic Presentation of the Electro-reduction Unit

RESULTS

Electro-reduction

Before electro-reductive stripping was investigated, a series of experiments was done to define the ferric reduction power and the kinetics of the above-mentioned unit as a function of temperature and electric power. The catholyte [350 ml] was a 2 M H2SO4 solution with 11.83 g/1 ferric iron. The impeller speed was 330 rpm. The results are shown in Figure 2.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 783

Ό ω u 3

ω

o Έ n-,

100.0 90.0 80.0 70.0 60.0 50.0 40.0 30.0 20.0 10.0 0.0

- » - 2 M 1 0 V 2 0 ° C

- » - 2 M 2 0 V 2 0 ° C

- ± - 2 M 10V50°C

- X - 2 M 2 0 V 5 0 ° C

6 12 18 24 30 36 42 48 54 60 66 72

Time (min)

Figure 2 - Percentage of Ferric Iron Reduced as a Function of Time

During the experiments, the terminal voltage was kept constant at 10 or 20 V. For the curves in Figure 1, this meant an average current of, respectively, 0.595 A and 1.378 A for 20°C and of 0.705 and 1.677 A for the experiments carried out at 50°C. The maximum amount of ferric ion reduced was 91.2 % [20 V, 1.677 A, 50 °C, final ferric concentration 1.04 g/1 ]. The reduction rate of the ferric iron is the highest, as would be expected, at 20 V and 50 °C. For the linear part of this curve (between 18 and 42 minutes), the reduction rate is 81.7 mg/min at a current of 2.4 A. The current efficiency ranges between 90% at the beginning of the experiment and 85 % at the moment the plateau in the curve is reached. Then the current efficiency drops slowly to 50 %, indicating simultaneous hydrogen reduction, which was also visually observed.

Electro-reductive Stripping

Simultaneous stripping and reduction were carried out as follows. A volume of 250 ml of loaded organic containing 12.78 g/1 of Fe(III) was contacted in the reactor with an equal volume of strip solution [1 M or 2 M H2SO4]. The impeller was put in the aqueous phase in order to have the organic dispersed into the strip solution; this was essential to carry out electro-reduction in the cell. The impeller speed was 850 rpm. In this way the emulsion was dispersed equally in the vessel. It is known from the literature that the stripping kinetics of iron are rather slow; thus, we waited 30 minutes before switching on the reduction power. During that time equilibrium without reduction was established. At regular time intervals a sample was withdrawn from the cathode compartment. The sample was analysed for its total iron, ferrous and ferric concentrations. Some results are shown in Figure 3.

784 LEAD-ZINC 2000

12

- IM 20V 20°C i

-1M20V 50°C

-2M20V 20°C

-2M20V 50°C

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (min)

Figure 3- Total Iron Concentration (g/1) in the Strip Solution as a Function of Time

The equilibrium concentration of iron without electro-reductive [thus after 30 minutes of contact time] is 2.88 and 5.82 g/1 for the 1 M H2S04 strip at, respectively, 20°C and 50°C. For the 2 M solution, this is 5.48 and 7.26 g/1. It is obvious from these results that the stripping kinetics without any electro-reductive action are considerably faster at 50 °C than at 20 °C. At the end of the experiment, the iron in the strip phase increased to 5.52, 7.84, 7.56 and 10.08 g/1, respectively. Thus, enhanced stripping with electro-reductive action is observed which is clearly more effective at the higher temperature. It can also be seen that, after the current is switched on, it takes some time before the iron concentration in the strip liquor starts to increase. This indicates that, before iron is stripped from the organic phase, the ferric iron in the strip phase has first to be converted to ferrous iron below a critical level. This is also shown in Figure 4 which depicts the total iron, ferric and ferrous concentrations as a function of time for the 2 M, 20 V, 50°C experiment.

■Fe3+ (g/1)

■Fe2+(g/l)

"Fetol(g/l)

20 40 60 80 100 120

Time(min)

Figure 4 - Iron Species Concentration (g/1) in the Cathode Compartment as a Function of Time

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 785

From this figure it can be seen that, after equilibrium is reached [30 min] and electro-reduction is started, no further immediate increase of iron in the strip solution is obtained until the ferric concentration is lowered by electro-reduction to less than 3 g/1. Once this value is reached, further stripping occurs, and this in a final total iron concentration in the strip solution of 10.08 g/1. With electro-assisted stripping, the stripping efficiency can be increased from 56.8% to 78.8 %. The remaining iron concentration in the organic phase after electro-reduction is 2.7 g/1. The electrical current in the cell remained constant at ~ 1.4 A. The overall current efficiency ranges from 90 % to 80 % at the beginning of the experiment, but drops to 50 % at the end of the test.

Build-up of Iron in the Strip Phase

An option is to integrate an electro-reduction unit into the aqueous flow of an iron strip SX-unit with recycling of the strip solution until a high build-up of iron is obtained. To find out the influence of the initial ferrous concentration in the strip solution on the electro-reductive stripping efficiency, a series of experiments was performed identically with the above-mentioned electro-reductive stripping experiments, but with various initial concentrations of ferrous sulphate. The temperature was 50°C and voltage was set to 20 V. The results are shown in Table I.

Table I - Influence of the Initial Ferrous Concentration on the Stripping Efficiency 2 M H2S04, 50°C, 20 V

Initial Organic: Fe-saturated 20 vol % D2EHPA / Shellsol T

Initial Ferrous Concentration (g/i)

0 10 20 30 40

Percentage Ferric Stripped

86.03 85.75 84.87 85.65 83.48

From this table it may be concluded that the initial ferrous concentration has practically no influence on the stripping efficiency. Strip solutions containing -50 g/1 of ferrous iron can be obtained.

DISCUSSION

As already mentioned, stripping of ferric loaded D2EHPA solvents is well documented. The stripping behaviour is ruled by a complex combination of various parameters. One of them is the pH value of the aqueous solution from which the iron was extracted. Yu and Chen (20) reported that iron extracted from high pH [1.5-2.1] solutions in the form of hydroxy-ferric ions can be stripped with dilute sulphuric acid. If extracted from low pH [-0.5] solutions in the form of unhydrolysed ferric ions, however, the iron can only be stripped with concentrated sulphuric acid. Another parameter affecting the stripping efficiency is the D2EHPA concentration of the solvent. Hogeweg (21) showed that the stripping efficiency [O/A = 1] with 4 M nitric acid at 50 °C of a ferric saturated D2EHPA solvent decreased from 71.3% to 41% and 25.6% for, respectively, 5, 10 and 20 vol. % D2EHPA solutions. The same tendency was observed with

786 LEAD-ZINC 2000

sulphuric acid strip solutions (20). As our option was to investigate iron removal from acidic zinc sulphate solutions [pH <0.5] having a high iron concentration, the solvent requires the highest possible extractant concentration. Unfortunately, this delivers a ferric loaded solvent with the worst stripping properties.

Although it is not always possible to exactly compare studies, as some experimental conditions may differ slightly, Table II shows the results for the stripping of Fe(III) from D2EHPA solutions obtained by various methods, including the present results.

Table II - Comparison of the Stripping Efficiency and Kinetics of Ferric Loaded D2EHPA Solvents by Various Methods

D2EHPA Cone. 30 vol. % 20 vol. % 30 vol. 30 vol. % 10 vol. % 30 vol. % 10 vol. % 20 vol. % 20 vol. % 20 vol. % Sulphuric

Organic loading 20.7 g/1 2.12 g/1 5.58 g/1 0.279 3.5 g/1 17.06 g/1 6.83 g/1 11.01 g/1 12.79 g/1 12.79 g/1

Strip*

1.5 M 1.5 M 2M 4M 0.2 M 2M 2M 1.8 M 2M 2M

Reductant

S0 2

so2 Fe L-ascorb. Zn-alloy None None None None Electro

acid concentration. O/A = 1.

Temperature

70 °C 70 °C 25 °C 25 °C 25 °C 25 °C 25 °C 20 °C 50 °C 50 °C

Stripping efficiency 60% 50% 92% 55% 92% 40% 70% 52 . 1 % 56.7 % 78.9 %

Kinetics

60 min 120 min 120 min 120 min 30 min 20 min 20 min 30 min 30 min 120 min

Ref.

15 15 16 16 17 20 20 21 This paper This paper

From this table it is clear that the galvanic stripping of iron from ferric loaded solvents with high D2EHPA concentrations is most effective using solid metal reductants such as iron, zinc or its alloys. Close control of the operating parameters, however, is necessary. The oxygen content of the system and the pH of the strip solution must be carefully monitored. Some oxygen is necessary to initiate the reaction, but an excess causes re-oxidation of the Fe (II) and increases the stoichiometric metal usage. Low pH causes the dissolution of the metal stripping agent. The oxygen concentration in the strip solution also affects the electro-reduction. Electro-reduction, however, has the main advantage that the introduction of metallic phases in the system is avoided because electric power creates the reductive driving force. Continuous treatment is possible and a high iron concentration in the strip phase can be obtained. The efficiency of stripping and the kinetics in electro-reduction have to be improved, and this may be easily realised by an appropriate cathode material selection and the use of an anion selective membrane as the electrolyte separator. A cathode material exhibiting a higher overpotential for hydrogen reduction should result in lower residual or "equilibrium" ferric concentrations in the strip solution. In turn, this should lead to more efficient stripping and faster kinetics, the driving force of the stripping reaction being increased. The use of an anion selective membrane would avoid the diffusional carry over of iron from the catholyte to the anolyte compartment, as was detected in some of our experiments. Taking into account this iron transfer for the 20 V, 50°C experiment, for instance, leads to an overall stripping efficiency of 91.2 %.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 787

CONCLUSIONS

It has been shown that electro-reductive treatment of a ferric loaded D2EHPA solvent enhances the stripping efficiency. Removal of iron from an acidic zinc sulphate solution with a high iron content, such as a hot acid leach solution in zinc hydrometallurgy or Zn-Fe galvanic co-deposition baths, has not been implemented because of the bad stripping characteristics with dilute H2SO4. Electrochemical treatment has the potential to adequately solve this problem and to allow the selective removal of iron from a D2EHPA solvent loaded.

REFERENCES

1. U. Kerney, "Solvent Extraction of Iron from Sulphuric Acid Liquors of Roasted Ore", Ertzmetall Vol. 43, N° 5, 1990, 195-199.

2. G. Van Weert, T. van Sandwijk and P. Hogeweg, "Solvent Extraction of Ferric Iron from Zinc Sulphate Solutions with D2EHPA-Investigation of Nitric Acid as Stripping Agent", EPD Congress 1998. B. Mishra, Ed., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1998, 245-266.

3. G.P. Demopoulos, R. Molnar and L. Rosato, "Bench Scale and Mini-pilot Plant Investigations on the Selective Removal of Iron from Zinc Process Solutions by Solvent Extraction", Iron Control and Disposal, J.E. Dutrizac and G.B. Harris, Eds., Canadian. Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1996, 395-416.

4. T. Principe and G.P. Demopoulos, "Solvent Extraction Removal of Iron from Zinc Process Solutions Using Organophosphorus Extractants", EPD Congress 1998, B. Mishra, Ed., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1998,267-287.

5. V. Lakshmanan, N. Rathie and B. Monzyk, "Evaluation of N-alkyl Hydroxamic Acids for Selective Iron Separation from Zinc Process Liquors for High Purity Iron Products", Iron Control and Disposal, J.E. Dutrizac and G.B. Harris, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1996, 357-367.

6. F. Delmas, M. Ehle, R. Koch, C. Plazanet and K. Ujima, "Novel Highly Efficient Selective Extractants for Iron in Zinc Hydrometallurgy", Iron Control and Disposal. J.E. Dutrizac and G.B. Harris, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1996, 381-393.

7. F.T. Principe and G.P. Demopoulos, "Separation and Concentration of Iron from Zinc Process Solutions", Journal of Metals. Vol. 51, N°12, 1999, 34-35.

8. K. Dewit," Elektrodepositie en Aanwendingsgedrag van ZnFe-deklagen op Staal", PhD Thesis. University of Ghent. Belgium. 1998, 127-141.

9. A. Prior, "Method of Processing Acidic Fe-containing Solutions", US Patent. N° 5.051.186, 24 September, 1991

788 LEAD-ZINC 2000

10. G. Demopoulos and G. Pouskouleli, "Solvent Extraction of Iron(III) from Acid Sulphate Solutions by Mono(2-ethyl-hexyl) Phosphoric Acid", Canadian Metallurgical Quaterlv, Vol. 28, N° 1,1989,13-18.

11. A. Kravchenko, G. Miroyevsky, M. Loseva and V. Travkin, "Solvent Extraction of Iron from Sulphuric Acid Solutions with Phosphororganic Reagents", Tsvetnve Metallv. Vol. 34,N°3, 1993, 10-13.

12. G. Demopoulos and D. Gefvert, "Iron(III) Removal from Base-Metal Electrolyte Solutions by Solvent Extraction", Hvdrometallurgv, Vol. 12,1984, 299-315.

13. T. Hirato, Z. Wu, Y. Yamada and H. Majima, "Improvement of the Stripping Characteristics of Fe(III) Utilizing a Mixture of Di(2-ethyl-hexyl) Phosphoric Acid and Tri-n-butyl Phosphate", Hvdrometallurgv. Vol. 28, 1992, 81-93.

14. J. Chen, Y. Shuqiu, L. Huizhou, M. Xiquan and W. Zhichun, "New Mixed Solvent Systems for the Extraction and Separation of Ferric Iron in Sulphate Solutions", Hvdrometallurgv. Vol. 30, 1992,401-416.

15. H. Majima and T. Izaki, "Reductive Stripping of Fe(III)-Loaded D2EHPA with the Aqueous Solutions Containing Sulfur Dioxide", Met. Trans. B, Vol. 16B, 1985, 187-194.

16. J. Shibata and M. Sano, "Stripping of Fe(III) from D2EHPA with Mineral Acid in the Presence of Reduction Agents", Metallurgical Review of MMIJ. Vol. 6, 1989, 105-114.

17. M. Moats, C. Chang and T.J. O'Keefe, "Recovery of Zinc from Residues by SX-galvanic Stripping Process", Third Int. Svmp. on Recycling of Metals and Engineered Materials. P. Queneau and R. Peterson, Eds., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A.,1995, 545-562.

18. C. Chang, H. Gu and TJ. O'Keefe, "Galvanic Stripping of Iron from Solvent Extraction Solutions from Zinc Residue Leaching", Iron Control and Disposal, J.E. Dutrizac and G.B. Harris, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1996,417-428.

19. R. Leiby, M. Bricker and R. Spitz, "The East Penn Process for Recycling Sulfuric Acid from Lead Acid Batteries", Third Int. Svmp. on Recycling of Metals and Engineered Materials. P. Queneau and R. Peterson, Eds., The Minerals, Metals, Materials Society, Warrendale, PA, U.S.A., 1995, 311-319.

20. S. Yu and J. Chen, "Stripping of Fe(III) Extracted by Di-2-ethylhexyl Phosphoric Acid from Sulfate Solutions with Sulfuric Acid", Hvdrometallurgv. Vol. 22,1989, 267-272.

21. P. Hogeweg, "The Application of Nitric Acid as a Stripping Agent in the Solvent Extraction of Ferric Iron with D2EHPA from Zinc Sulphate Solutions", PhD Thesis. Delft University, The Netherlands, 1996.

Chapter 10

New Electrowinning Technologies for Lead and Zinc

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 791

NEW CLEAN TECHNOLOGIES TO IMPROVE LEAD-ACID BATTERY RECYCLING

C. Frias, M. Garcia and G. Diaz Tecnicas Reunidas, S.A (R&D Centre)

Sierra Nevada 16, 28830 San Fernando de Henares Madrid, Spain

ABSTRACT

Advances in hydrometallurgy promoted by Tecnicas Reunidas are providing increasingly simple and clean means for controlling the entire lead recycling chain. Used in parallel with pyrometallurgy, these processes allow furnace temperatures to be reduced and fumes and atmospheric pollution to be minimised. Furnace slags are digested, and residues (mainly gypsum) are non-toxic and convertible into marketable products. In addition, the global economy of the process is substantially improved by reducing the operating cost, increasing the lead recovery above 99% and obtaining a 99.99% pure lead product. These new PLACID and PLINT processes may provide the cleanest and healthiest practicable means for recycling lead from batteries.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

792 LEAD-ZINC 2000

INTRODUCTION

This paper describes advances of the PLACID and PLINT technologies that seem to be opportune for the secondary lead industry. They have all been identified by recent and current work in the Research and Development Centre of Tecnicas Reunidas S.A. (TR) at San Fernando de Henares near Madrid in Spain.

As is usual in hydrometallurgy, lead is extracted in the PLACID process by electrowin-ning. However, although electrowinning is generally acceptable for the extraction of valuable metals, the high capital cost of the electrolytic system appeared disadvantageous to established smelters when compared with the current (and possibly the future) market price of lead. Electrowinning is superior for plants producing more than about 20,000 tonnes a year of lead. One smelter, which witnessed the operation of the PLACID pilot plant at TR's R&D Centre, nevertheless suggested that it might be possible to process the electrolyte by precipitating a pure lead compound which could be fed into a furnace for decomposition or reduction. TR has since worked on this concept for three years, leading to the definition of the PLINT (PLacid INTermediate) process described below.

The confidence acquired by this development has caused TR to review its conception of the scope of hydrometallurgy for lead processing. It is no longer the aim to promote the best hydrometallurgical process and evaluate it against conventional standards, but to design processes that take best account of the constraints within which users operate. The emergence of this confidence can be seen in the chapters that follow.

To complete this presentation, the economics of a selected base case are discussed and results and conclusions are included.

OBJECTIVES

Existing battery recycling technologies have important deficiencies that need to be revised and adapted to meet the most restrictive environmental regulations and to reach a sustainable growth model for the secondary lead industry in the new century. The most relevant shortcomings of currently applied technologies are as follow:

• Many technological inconveniences are due to the usual pyrometallurgical processing of battery paste and its associated lead sulphate content. This sulphate produces sulphurous gases that require costly capture and neutralisation techniques. The thermal treatment of battery paste generates lead fumes that need to be retained by using electrostatic precipitators and/or bag filters.

• Conventional techniques for battery paste desulphurisation based on sodium hydroxide or sodium carbonate are expensive and produce large volumes of undesirable sodium sulphate solution, which requires large scale handling or represents a waste difficult to dispose.

• For producing refined lead, the addition of sodium hydroxide and sodium nitrate to the refining kettles is a common practise. This technique generates large amount of drosses containing soluble sodium salts.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 793

• In addition, sodium carbonate is a quite usual flux for lead smelting giving leachable sodium slags, which are very difficult for disposal.

• Lead extraction from battery paste by smelting requires the addition of fluxes and a high temperature (1100 °C or above) for the decomposition of the lead sulphate. The inconvenience that such a high temperature entrains is well understood. On the contrary, smelting operations would be substantially simplified if sulphate-free lead oxide materials could be treated.

The proposed new processes intend to deal with all these mentioned deficiencies in a reasonable and efficient way with the following objectives. The PLACID and PLINT processes intend to complement and improve existing battery recycling factories, but not to replace them. Conventional melting processes would treat the metallic lead components from the batteries together with lead scraps, while battery paste would be sent for hydrometallurgical processing. The latter would minimise sulphurous gases and lead fumes. No paste desulphurisation step is required. Lead sulphate is dissolved in brine and the sulphur is removed in the form of gypsum. Gypsum is an inert residue and, eventually, could be converted into a commercial by-product. Pure lead is produced when PLACID or PLINT technologies are applied. Conventional lead refining in kettles is not needed. The generation of sodium-based drosses and leachable sodium slags is avoided. The small amounts of produced fume, drosses and slags are internally recycled to the hydrometallurgical line. No toxic solid residue is generated. Both processes run in a closed brine circuit. The water balance is controlled by evaporation, if necessary. No liquid effluent exists either. When PLACID or PLINT processes are used in parallel with a lead smelting plant, both alloyed and pure lead qualities are produced. Alloyed lead would be converted to grids or other battery components, whereas pure lead would be ideal for new battery paste manufacturing. In this way, the secondary lead chain would be closed in an efficient manner, ensuring a sustainable growth model without additional needs for primary lead. Another important advantage of obtaining pure lead may be the extended life of the batteries, reducing significantly the annually recycled tonnage of lead. For instance, by increasing the average service life of an automotive battery in Europe by one year, the annual recycling rate of batteries would be reduced by approximately 20%, causing positive effects on the environment, energy savings and sustainability.

ELEMENTS OF PLACID AND PLINT PROCESS DESIGN

The design of the new proposed processes is conceptually very simple, easy to understand and very efficient from chemical and energetic points of view. Those processes include a series of properly combined unit steps.

Hydrometallurgy and Pyrometallurgy

Despite focussing on hydrometallurgy as its speciality, TR is not contemptuous of pyrometallurgy and fully recognises the advantages of processing materials in furnaces. It simply considers hydrometallurgy cleaner, more exact, and more easily controllable for treating battery pastes, lead fumes and drosses. Pyrometallurgy is ideal for metallic lead component treatment, but presents many disadvantages when it is applied to battery pastes processing.

The proposed hydrometallurgical processes are cleaner in themselves, and in combined cycle plants they can be used to reduce furnace temperatures, thus reducing the energy

794 LEAD-ZINC 2000

demands. In addition, they increase the productivity of any existing pyrometallurgical plant and reduce the environmental impact of smelting.

It should be remarked that the aim of the PLACID and PLINT technologies is to complement existing battery recycling plants, yielding the best synergy and complementary aspects of both the 'pyro' and 'hydro' lines.

Leaching

This is the initial step in which all accessible soluble lead in the feedstock is selectively dissolved. Feed materials to hydrometallurgical processes may vary, including battery paste (the principal feed material) together with lead fume, drosses and slags. Also a small portion of lead sulphide concentrate could be fed. Even more, old slag deposits and lead contaminated soils could also be treated.

Lead dissolution efficiency is very high. In a PLACID-Pyro or PLINT-.Pyro combined process, the overall measured extraction rate was 99.5%, because the leaching process extracted the available lead from slags and other products already rejected by the furnace operation.

The leachant is dilute hydrochloric acid in a brine solution, which has the ability of dissolving lead oxides and lead sulphate in an efficient manner. In the two final campaigns of the PLACID pilot plant operation, leaching efficiency from representative battery paste/fume mixtures was 99.4 to 99.7% after treating more than fifteen tonnes of feed materials. The main reactions involved are:

PbO + 2 HC1 + 2 NaCl -> PbCl4Na2 + H20 (1)

Pb + Pb02 + 4 HCl + 4 NaCl -> 2 PbCl4Na2 + 2 H20 (2)

PbS + 4 Pb02 + 8 HCl + 12 NaCl -> 5 PbCl4Na2 + Na2S04 + 4 HzO (3)

PbS04 + 4 NaCl -» PbCUNaj + Na2S04 (4)

Na2S04 + 2 HCl + Ca(OH)2 -> CaS04 + 2 NaCl + 2 H20 (5)

In the case of the PLACID process, hydrochloric acid is regenerated in the electrolytic cells, so lime is the only consumable. When the PLINT process is applied, make-up acid addition is necessary, and sulphuric acid can replace hydrochloric acid because of its lower cost.

Sulphate Removal

A battery paste desulphurisation step is not required, avoiding the generation of undesirable sodium sulphate solution. The separated paste from the battery breaker is directly fed to the hydrometallurgical processes. Drying is not required.

In the case of sulphur present as lead sulphate, the treatment with lime (about the cheapest material suitable for this purpose) forming gypsum is preferred, which subsequently is removed by filtration. TR has developed techniques whereby pure gypsum can be produced in any of its three morphologies (hydrated, hemi-hydrated and anhydrous) to suit market demands. Therefore, the obtained gypsum residue could be converted to commercial gypsum products, if commercially feasible. The real benefit in this is not having to pay for its disposal.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 795

Purification

The purification technique is very simple and involves injecting lead powder into the leachant or electrolyte to enable an electrochemical reduction of metal impurities more noble than lead. The purification reaction is:

MeCln + n/2 Pb° -> n/2 PbCl2 + Me0 (6)

"Me" means any metallic impurity such as Cu, Bi, Sn, Ag, As, Sb...

The leachant is then filtered to remove the cement. The efficiency of this process is good enough to produce a pure electrolyte that ensures a 'four-nines' lead final product. Of course, the concentration of impurities in the electrolyte does not translate directly to the concentration of impurities in the lead because of competitive electrodeposition. The metallic lead content of the cement is above 90%. The cement is re-melted and absorbed into secondary alloyed or other commercial grades of lead in the factory.

Lead Extraction Steps

Electrowinning

The PLACID electrolytic cell is the core of this technology. The electrolyte for the two electrodes, the anode and the cathode, is different, and the anode and cathode are separated by a membrane that is permeable only by protons (H+). On the cathode, lead chloride is stripped of its lead atom, leaving two chloride atoms, which are negatively charged. These negatively charged chloride atoms combine with protons passing through the membrane from the anode to reform hydrochloric acid that is returned to the leaching bath for reuse. The main electrochemical reactions are as follows:

Cathodic: PbCl2 + 2 e" ->· Pb° + 2 Cl~ (7)

Anodic: H20 - 2 e" -> 2H+ + '/2 02 (8)

Global: PbCl2 + H20 -> Pb° + 2 HC1 + Ά 02 (9)

The design of this cell is unusual: instead of depositing lead on metal plates, the electrolysis process deposits lead as dendrites or sponge, which are subsequently shaken off and collected on a conveyor belt. Immediately after leaving the electrolyte, the dendrites are pressed to extract the liquid and to form platelets of pure lead, which can then be conveyed to a kettle for casting into ingots.

There is no special virtue in the conventional practice of depositing the lead on plates: the plating process must be interrupted periodically while the plates are replaced, and there is no application for lead in the form of discrete thin flat plates. In contrast, there are important cost and convenience advantages in depositing lead as dendrites or sponge. The applied amperage, for example, can be increased by a factor between 4 and 10, greatly reducing the number of electrolytic cells that must be provided for a given throughput of lead product. Moreover, the entire extraction process can be run continuously without interruption. Labour for cathode stripping and replacement is not needed.

796 LEAD-ZINC 2000

Electrowinning is a capital-intensive process mainly because of the electrical transformers and rectifiers needed. Although direct operating costs are lower than for pyrometallurgical extraction processes, the amortisation costs are not insignificant. Electrowinning is, therefore, most advantageous for large plants.

It is strongly recommended that advantage be taken of the cost reduction obtainable by the usage of the combined cycle generating plant. Electricity would be used for electrowinning and motors while waste heat from the process can be used to maintain leachant temperatures at appropriate levels (typically about 80° C) and utilised for drying and evaporation where required.

Pure Lead Oxide Smelting

In the case of the PLINT process, a pure lead hydroxide (oxide) or lead carbonate concentrate is produced from the purified pregnant solution. Various reagents or a combinations of them can be used depending on local price, availability and process constraints.

PbCl4Na2 + Ca(OH)2 -> Pb(OH)2 + 2 NaCl + CaCl2 (10)

PbCL,Na2 + 2NaOH->Pb(OH)2 + 4NaCl (11)

PbCUNa2 + Na2C03-»PbC03 + 4NaCl (12)

This concentrate should be smelted in a furnace or a kettle. Pyrometallurgical treatment of this lead concentrate would be quite simple and efficient because only the addition of some reducing agent is required. A low smelting temperature would be sufficient and the amount of energy that must be provided to heat a charge is low in comparison to present operations. The overall time required for smelting this charge is substantially reduced. The existing 'pyro' facilities would have some over-capacity to increase actual production.

PLACID AND PLINT PROCESSES

The PLACID Process

Much about this process has already been described in previous sections. The conceptual block diagram is shown in Figure 1.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 797

PASTES

Lime ■

I Sulphate Removal

T Inert Residue

(Gypsum)

Electrowinning

I Purification

.ead ▼ Lead Powder

Melting and Casting

J 99.99 LEAD INGOTS

Lead Cement (Bi, Cu, As, Sb)

Figure 1 - PLACID Process, Conceptual Block Diagram

The leachant is dilute acid brine, and the desulphurisation sequence is quite interesting. The lead sulphate reacts with salt to form lead chloride and sodium sulphate. The sodium sulphate then reacts with hydrochloric acid and lime to yield gypsum, as well as simultaneously generating reforming salt ready for reuse as the leachant. Because hydrochloric acid is regenerated in the electrowinning section, lime is the only component consumed.

Different feed materials would also be suitable for this process besides battery paste; lead fume, drosses and slags, or even a small fraction of lead sulphide concentrate.

Pilot plant development in the laboratories of the R&D Centre was carried out in several campaigns (above 1000 hours total operation), during which 10 tonnes of pure electrolytic lead were produced. The energy consumption was 0.9 kWh/kg Pb. Representative samples of electrolytically recovered lead had purities above 99.99% Pb, containing 3 ppm Cu, 6 ppm Sb, 2 ppm As, 1 ppm Sn and 2 ppm Bi. Those data likely could be improved in a continuous industrial plant.

The PLACID process could be perfectly integrated if it were employed in parallel with a pyrometallurgical smelter. In this way, any lead fumes, drosses and slags from the pyrometal-lurgical line can be passed to the leaching bath of the PLACID line and the cements from the purification step could be fed into the furnace. Important gains would be obtained from the environmental, process efficiency, product quality and economical points of view.

The PLINT Process

A conceptual block diagram is shown in Figure 2. Comparing the PLACID and PLINT block diagrams, it becomes obvious that the only difference in principle is in the substitution of a precipitation step for electrowinning. In the subsequent furnace, the lead hydroxide product is first decomposed and then reacted with hard coal to obtain pure lead. The process takes place at a temperature much lower than that required by current smelting processes. Because the leaching and purification processes are unchanged, the leaching efficiency of this process and the purity of the produced lead should be the same as in the PLACID process.

798 LEAD-ZINC 2000

The PLINT process needs the addition of acid make-up, so sulphuric acid can replace hydrochloric acid because of its lower cost. Any sulphuric acid is converted to gypsum.

It is possible to combine a PLINT process line with a pyrometallurgical line, creating a system similar to that in the PLACID-Pyro process. However, to achieve lead purities of 99.99+% it is necessary to have separated, dedicated kettles or furnaces.

PASTES'

Lime '

H2S04

Leaching

1 Sulphate Removal

i Inert Residue

(Gypsum)

—

Lime

Lead Precipitation

t Purification

Low Temperat Smelting

Melting and Casting

\ f 1 99.99 LEAD INGOT

Lead ▼ Powder Lead Cement

(Bi, Cu, As, Sb)

Figure 2 - PLINT Process, Conceptual Block Diagram

The development of the new PLACID and PLINT processes has proceeded to a stage where a demonstration plant is needed.

ECONOMIC EVALUATION

PLACID Process

The selected base case for the final feasibility study included the following conditions:

The new PLACID plant would be annexed to an existing battery recycling plant. Overall lead recovery of the combined PLACID-Pyro plant was estimated to be above 99.5%.

Lead production would be 20,000 t/y electrolytic lead, with a purity of 99.99%. As feed materials 27,500 t/y of battery paste and 2,500 t/y of fumes and slags were considered.

The energy consumption was optimised. After having performed a detailed study about electricity cost optimisation, it was decided to integrate the process with a co-generation energy plant using natural gas to supply electricity and heat. It is a very attractive option, reducing substantially the operating costs.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 799

The basic engineering of the base case has been developed in detail, performing material and heat balances. Operating costs were estimated based on the previous information. The main consumption figures are shown in Table I. The most important cost factor is natural gas for the co-generation plant.

Table I - PLACID Process Consumption Figures Material

Lime Sodium Hydrosulphide

Process Water Natural Gas

Unit kg/tPb kg/tPb m3/t Pb

Mcal/t Pb

Value 143 1.5 2.2

1,625

An operating cost break down at current prices in Spain is depicted in Table II in US dollars at the prevalent exchange rate, 170 PTA per USD. A 10% contingency factor has been allowed. As a worst case scenario, a residue disposal cost has been included, assuming that the gypsum was not sold.

Table II - PLACID Operating Cost Break Down Cost Units $/t Pb

Consumable 30.8 Reagents 9.2

Residue disposal 20.0 Labour 36.0

Maintenance 16.3 Contingency 12.0

TOTAL 124.3

The estimated total investment cost is approximately 19.0 Million USD, based on good practice engineering standards with a +25% accuracy factor. This includes the co-generation plant investment. The high contingency factor makes this study fairly conservative.

The PLINT Process

A similar base case has been selected for the PLINT process, according to the following:

• The new PLINT plant would be annexed to an existing battery recycling plant. Overall lead recovery of the combined PLINT-Pyro plant would higher than 99.5%.

• Production would be about 23,000 t/y of lead concentrate (oxide or carbonate), containing 20,000 t/y of lead being recoverable with a purity of 99.99%. As feed materials 27,500 t/y of battery paste as well as 2,500 t/y of fume and slags had been considered.

In a similar approach to PLACID, the basic engineering of the PLINT base case has been developed in detail. Operating costs were estimated based on material and energy balances. Main consumption figures are presented in Table III.

LEAD-ZINC 2000

Table III - PLINT Process Consumption Figures

Materials Unit Value Lime kg/t Pb 520

Sulphuric Acid kg/t Pb 320 Lead Powder kg/t Pb 40 Process Water nrVtPb 1.5

Electricity MWh/t Pb 0.1

Operating costs are presented in Table TV in US dollars at the prevalent exchange rate, 170 PTA per USD. A 10% contingency factor has been allowed. The higher cost corresponds to reagent make-up consumption. As a worst case scenario, a residue disposal cost has been included, assuming that the gypsum was not sold. The low temperature smelting cost is not included in this estimation.

Table TV - PLINT Operating Cost Break Down Concept $/tPb

Consumable 32.1 Reagents 40.2

Residue disposal 20.0 Labour 20.5

Maintenance 9.5 Contingency 13.0

TOTAL 135.3

The estimated total investment cost is approximately 7.1 million USD, based on good practice engineering standards and with a +25% accuracy factor. The high contingency factor also makes this study fairly conservative.

Profitability

Based on the above operating and investment costs, several DCF (discounted cash flow) calculations have been carried out. An IRR (internal rate of return) in the range of 16% to 28% has been worked out, which indicates that both the PLACID and PLINT processes are economically very attractive.

CONCLUSIONS

When PLACID or PLINT hydrometallurgical processes were implemented in parallel with any existing pyrometallurgical battery recycling plant, the environmental, technical and economical aspects of the integrated plant could be substantially improved. The proposed new processes are simple, efficient, very flexible and profitable, which make them ideal for being adapted to the customer's needs and local requirements.

The final evaluation study of both the PLACID and PLINT processes, applied to a selected base case producing 20,000 t/y of lead with a purity of 99.99% indicated very satisfactory results, which facilitate their industrial implementation in the short term. The

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 801

availability of this technology has opened a window of opportunity for the secondary lead industry, and inquiries are invited.

ACKNOWLEDGEMENTS

The European organisations that have developed the PLACID Process into the Brite-Euram Programme wish to express their gratitude to the European Community for its encouragement, support and partial funding of the project.

REFERENCES

1. "Proceso PLACID: Recuperacion del Plomo de Pastas de Plomo", Internal Report No. ITR/P4618/005/1988.

2. "Proyecto de Planta de Produccion de 25.000 t/a de Plomo por el Proceso LEADCLOR", Internal Reports No. ITR/P4628/029/1988 and ITR/P4650/017/1991.

3. D. Martin and G. Diaz, "Hydrometallurgical Treatment of Lead Secondaries and/or Low Grade Concentrates; The PLACID and the LEADCLOR Processes", Conference organised by ILZSG on Recycling Lead and Zinc-The Challenge of the 1990's, Rome, Italy, 1991,315-336.

4. R. D. Prengaman, "Recovering Lead from Batteries", Journal of Metals, Vol. 47, 1995, 31-33.

5. G. Diaz, "Lead Recycling", Letter Journal of Metals, June 1995, 3-4.

6. G. Diaz, C. Frias, L.M. Abrantes, A. Aldaz, K. van Deelen, and R. Couchinho, "Lead-Acid Battery Recycling by the PLACID Process - A Global Approach", TMS Third International Symposium on the Recycling of Metals and Materials, Point Clear, Alabama, 1995,843-856.

7. G. Diaz and D. Andrews, "Placid - A Clean Process for Recycling Lead from Batteries", Journal of Metals. Vol. 48, 1996, 29-31.

8. G. Diaz and D. Andrews, "Placid Lead for Batteries", Batteries International, April 1996,73-74.

9. C. Frias, M.A. Garcia and G. Diaz, "Industrial Size Placid Electrowinning Cell", Aqueous Electrotechnologies: Progress in Theory and Practice, D.B. Dreisinger Ed., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1997, 101-113.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 803

ELECTROWINNING OF LEAD BATTERY PASTE WITH THE PRODUCTION OF LEAD AND ELEMENTAL SULPHUR USING BIOPROCESS TECHNOLOGIES

M. Olper and M. Maccagni ENGITECsrl

Via Borsellino e Falcone, 31 20026 Novate Milanes, Italy

C.J.N. Buisman and C.E. Schultz Paques Bio Systems B. V.

P.O. Box 52 8560 AB Balk, The Netherlands

ABSTRACT

The key point of the CX-EWS process is the conversion of lead compounds, contained in the active mass of spent batteries, into lead sulphide. The obtained lead sulphide is leached with ferric fluoborate electrolyte to dissolve lead and oxidize the sulphur from sulphide to the elemental form. The lead-rich solution is fed into a diaphragm electrolytic cell depositing lead and regenerating ferric fluoborate. The introduction of sulphate reducing bacteria technology in the sulphidization step of the battery paste improves dramatically the economics of the CX-EWS process, reducing the costs of chemicals, avoiding gypsum disposal and minimising the unit operations of the process. This paper describes in detail this technology change with reference to the economics of the process.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

804 LEAD-ZINC 2000

INTRODUCTION

The CX breaker (1) for the separation of the components of spent lead acid batteries and the thermal recovery of metallic lead is well known and is considered "state of the art" technology in this field. More than 20 plants are operating all over the world and requests for the construction of new ones are frequently received. The Engitec CX process is a mature technology that provides answers to the two most important problems related to spent lead acid batteries:

• The environmental impact • The recovery and the relevant reuse of the battery components.

Increasing environmental pressures during the last 20 years, on both primary and secondary lead facilities, have stimulated the industry to develop new and more environmentally friendly technologies. As an alternative to improving the traditional pyrometallurgical processes, hydrometallurgical and electrochemical technologies have also been investigated to a greater extent and have been developed. Compared to the thermal processes, particularly the electrochemical processes have some characteristics that make them very attractive, such as:

• High selectivity because the electrochemical deposition is selective for the metal to be produced.

• Wastes reduction as only small amounts of chemicals and reactants are used, resulting in minimised waste production.

• In an electrochemical cell, O2 and H2 are usually produced, and these are environmentally acceptable gaseous emissions.

• A more friendly workplace because the ambient working conditions in a cellhouse are free of the dust emissions usually found in pyrometallurgical smelters.

• The electrochemical plant is completely modular because its productivity is proportional to the number of cells, and therefore, it is relatively easy to increase the productivity by simply adding more cells.

• Economics are favourable because the electrochemical plants, as described in this paper, do not require big capacities to operate in an economically viable way.

The need for a different approach to overcome the environmental constraints of pyrometallurgical processes led to the CX-EWS technology, which is the direct electrorefining of metallic lead (grids and poles) obtained from the CX breaker. It is based on fluoborate technology, and when introduced into primary and secondary lead operations, can eliminate all the drawbacks present in lead production.

After about ten years of intensive research from the bench scale to pilot demonstration plants, performed by Engitec in co-operation with the most important world lead producers, this technology is ready for commercialisation.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 805

CX-EWS PROCESS

With the traditional approach of recovering lead from spent lead acid batteries using pyrometallurgical methods, and adopted by large and small sized refining companies, it may become more and more difficult to meet the stringent environmental standards introduced. Even with massive investments, some of the plants applying conventional technologies are still far away from satisfying the statutory requirements related to the release of dust and SO2 gas into the atmosphere, to the lead levels in micro and macro climates, and to the Pb, Cd and As contents of the solid waste emissions.

During the last years, an extensive amount of time and energy has been consumed developing an alternative hydrometallurgical method to handle the paste and to ensure the continuation of its efficient recovery. This is of vital importance if the dumping of waste batteries is to be avoided. Several processes (2), as shown in Table I, have been published in the past but none of these reached industrial size operations.

Although each process differs in its approach, the processes to recover lead from battery paste via electrowinning have several problems:

• Desulphurization of the PbSC>4 converting it into a soluble species • Reduction of Pb02 to PbO • Prevention of PbC>2 deposition on the anode • Use of an insoluble anode which does not deteriorate in use • Recovery or disposal of the sulphate • Use of materials that can withstand corrosive acidic and alkaline solutions • Preventing the escape of anode and/or cathode generated gases or mists • Treatment of the insoluble sludge after leaching.

Until recently, our process excluded the metallic components of the battery, namely the grids and poles. Like all the other processes, we focussed the investigations mainly on the paste. The metallic components that normally consist of lead alloys, with antimony or other metallic impurities, represent 25 - 30 % of the battery weight and are usually re-melted and refined employing a thermal process. Because extensive efforts have to be taken to eliminate the hygienic and environmental problems associated with the re-melting of oxide-coated lead scrap and the thermal refining of impure lead alloys, we considered it indispensable to extend the principle of electrolytic refining it to these types of metallic scrap.

Tab

le I

- L

ead

Past

e H

ydro

met

allu

rgic

al R

ecov

ery

Proc

esse

s

Past

e tr

eatm

ent

Ele

ctro

lytic

sys

tem

Pr

oces

s D

esul

phur

isat

ion

PbC>

2 re

duct

ion

Ele

ctro

lyte

C

once

ntra

tion

Cur

rent

Pb

C>2

type

(g

/1)

dens

ity

prev

entin

g Pb

Fr

ee

(A/m

2 ) ag

ent

acid

Ano

de

Not

es

RSR

Bur

eau

of M

ines

Eng

itec

CX

-EW

Gin

atta

AA

S

PbS

0 4

slur

ry

Na,

Li,

K, N

H4+

carb

onat

e N

a, L

i, K

, NH

4+

carb

onat

e

Na 2

C0 3

, NaO

H

Ele

ctro

chem

ical

di

ssol

utio

n

= =

S0 2

in

alka

li Pb

po

wde

r

Pb

pow

der,

H20

2

S0 2

in

alka

li S

0 2

H2S

iF6/H

B

F 4

H2S

iF6

HB

F 4

HB

F 4

(NH

4)2S

0 4

Na 2

S0 4

/ H

2S0 4

70-

200

> 15

0

100-

120

>50

150-

250

100

350

400

As

> 0.

5 g/

1

Poly

phos

phat

es

1-2

g/1

Hig

h C

D.

resi

dual

H2O

2

Co

> 0.

2 g/

1

PbC

Vm

esh/

gr

aphi

te

Pb0 2

/Ti

Pb0 2

/Ta/

Cu

Pt/T

a/C

u

Gra

phite

Flui

dise

d be

d ca

thod

e in

a m

embr

ane

cell

Big

am

ount

of

Pb p

owde

r re

cycl

ing.

Lea

d ph

osph

ate

prec

ipita

tion.

Ano

de f

ailu

re

Pb s

pong

e an

d (N

H4)

2S0 4

prod

uctio

n C

ompl

icat

ed c

ell.

Low

ef

ficie

ncy

of th

e m

embr

ane

PLA

CID

Pb

po

wde

r C

hlor

ides

M

embr

ane

cell

Big

am

ount

of P

b po

wde

r re

cycl

ing.

Pb

spon

ge

prod

uctio

n. C

ompl

icat

ed

cell.

Hig

h ch

lori

de

corr

osio

n

r a > D

Z o

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 807

ENGITEC EXPERIENCE: THE EVOLUTION THAT LED TO A NEW PHILOSOPHY TO RECOVER BATTERY SCRAP

ENGITEC has had experience dealing with battery scrap for more than thirty years. In that time, several systems have been developed which efficiently and economically reclaim virtually every component of a spent battery. The activity of ENGITEC has been increasingly motivated by the requirements of the industry to meet the strictest environmental protection regulations. The chronological summary of ENGITEC's development of lead-acid battery reclamation systems is as follows:

1968 - 1982

• System for crushing and separating battery components including paste drying • Rotary furnaces for paste and grids smelting • Process for polypropylene separation, cleaning and the production of granules.

1982 - 1987

• Improvement of the technology for the separation of battery components (paste, grids and poles; polypropylene, ebonite, separators)

• On-line desulphurization of paste to reduce SO2 emissions and solid waste disposal, and to improve the smelting efficiency

• Production of detergent grade anhydrous Na2S04 • Separation and clean-up of ebonite for use as a fuel in conjunction with gas fired burners • Furnace for continuous, low temperature, melting and drossing of grids and poles.

1987-1993

• Electrowinning of desulphurized paste (CX-EW Process (3)) to eliminate the high temperature smelting process with its production of fume and slags

• Electrolysis of the Na2SÜ4 solution from paste desulphurization, allowing the recycle of NaOH and the production of battery grade sulphuric acid

• The complete electrochemical system (4) to recover Pb and the other components of battery scrap, and introducing the continuous melting of grids and poles.

Currently

• A new way of treating the battery paste to produce Pb electrochemically through the CX-EWS process. The foundation of this novel way of treating battery scrap was the series of technical problems described in Table I, and the recollection that the treatment cost is very relevant if we want to introduce a new technology.

808 LEAD-ZINC 2000

CX-EWS PROCESS DESCRIPTION

After some years of extensive research, a new and innovative modification of the previous CX-EW process is presented. This new process, as shown in Figure 1, is based on a combination of two innovative technologies: THIOPAQ® (developed by PAQUES Bio Systems B.V.) and FLUBOR® (developed by Engitec).

Lead paste THIOPAQ®

T H2/co2

PbS FLUBOR8

T s°

-► Lead cathodes

Figure 1 - Combined THIOPAQ -FLUBOR* Process

The reduction of lead sulphate is the main reaction of THIOPAQ technology (5). It is a unique process as it is normally not possible to reduce SO42" at ambient conditions without the presence of an anaerobic microorganism. The bacterial reduction of sulphate to sulphide is an electrochemical conversion. In order to supply the electrons needed for the conversion reaction, a reductant (electron donor) needs to be added. Examples of suitable electron donors are ethanol and various fatty acids. For large-scale applications (more than 2.5 t of H2S produced per day), hydrogen gas is the preferable reductant. Hydrogen gas can be produced on site by cracking methanol, or by steam reforming of natural gas or liquefied petroleum gas (LPG). These processes convert hydrocarbons to hydrogen gas and carbon dioxide, which are both fed to the bioreactor. The carbon dioxide is used for biomass growth and to provide buffering capacity which keeps the pH at acceptable levels.

The choice of the type of reductant is dependent not only on the sulphur load to be processed, but also on the geographical location of the installation, reagent availability, cost, etc. When ethanol is used, the operating cost is higher than with hydrogen gas, but the investment cost is lower because no reformer is required.

The FLUBOR® process is an electrochemical process for electrowinning lead from impure Pb metal and/or PbS based raw materials. This process is based on a ferric fluoborate leaching medium which is used to dissolve the Pb. The generated fluoboric electrolyte is fed to the cathodic compartment of an electrolytic cell, divided into two compartments by a diaphragm, where Pb is deposited. In the anodic compartment ferric fluoborate is regenerated, and is sent to the leaching reactor closing the electrolysis circuit.

The entire process, however, starts with the crushing of the batteries and the separation of all the components in a CX breaker plant. A rough composition of battery scrap, from data collected during several years of monitoring of CX plants, is reported in Table II.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 809

Table II - Battery Scrap Composition

Metallic Pb (grids and poles)

Lead paste -s

rpbso4 Pb02

PbO Metallic lead

L Other Polypropylene cases Separators, hard rubbers, etc. Sulphuric acid (about 15 %)

50-60 % ^ 15-35 % 5 -10 % 2 - 5 % 2 - 4 % ^

25 %

>- 38 %

5 % 10 % 22 %

The paste obtained during the CX operation can be treated applying the CX-EWS process, which is based upon the following main operations.

Paste Sulphidization

This step is a direct bio-sulphidization reaction and it converts all the lead-containing salts and/or oxides in the paste into PbS by means of anaerobic bacteria. Compared to previously studied systems, this technology improves the required conversion reaction in terms of the technical and economical aspects. The overall process can be simplified and described by the following reactions:

PbS04 + 4H2 (bacteria) -> PbS + 4 HzO

PbO + 4H2 + H2S04 (bacteria) -> PbS + 5 H20

PbQ2 + 5H2 + H2S04 (bacteria) -> PbS + 6 H20

(1)

(2)

(3)

Fluoborate Leaching of the Sulphidised Paste

This consists of leaching the precipitated PbS with a ferric fluoborate solution coming from the anodic compartment of the electrolytic cell which is divided by a diaphragm. Lead is dissolved and an elemental sulphur-based residue is generated. The metallic Pb present in the paste is leached simultaneously. The grids and poles fraction is leached separately but also with ferric fluoborate solution. This promotes the distinguishable production of refined lead and antimony lead alloys. The overall leaching reaction for both feed materials is the following:

PbS + 2Fe(BF4)3 -> Pb(BF4)2 + 2 Fe(BF4)2 + S°

Pb + 2Fe(BF4)3 -»· Pb(BF4)2 + 2 Fe(BF4)2

(4)

(5)

Electrolysis in the FLUBOR Divided Diaphragm Cell

The purpose of carrying out the electrolysis in the FLUBOR divided diaphragm cell is to produce Pb cathodes in the cathodic compartment and to oxidise the ferrous ion to ferric ion in the anodic compartment, thereby regenerating the ferric fluoborate leaching solution. The electrolysis reactions are the following:

810 LEAD-ZINC 2000

Cathodic:

Anodic:

Pb(BF4)2 + 2e" -> Pb + 2 BF4"

2Fe(BF4)2 + 2BF4" -> 2 Fe(BF4)3 + 2e

(6)

(7)

Overall: Pb(BF4)2 + 2Fe(BF4)2 Pb + 2Fe(BF4)3 (8)

All the steps are outlined in the conceptual flowsheet given in Figure 2.

Battery electrolyte (H2S04)

PbS

Natural

■

Fluoboric leaching reactor

Depleted Pb, Fe3+ rich solution Pb cathodes

Diaphragm divided cell

Figure 2 - Battery Paste Treatment: Conceptual Flowsheet

Looking at the problems evident in other electrowinning processes previously mentioned, we can emphasise the following:

• The solubility of PbS (Ks = 10"27) is much lower than that of PbC03 (Ks = 10"8) or of Pb(OH)2 (Ks = 10"10). The conversion to PbS is practically 100 %, compared to the lower conversions obtained during the conventional desulphurisation process; for example, 90 % conversion for the process with Na2C03 and 95 % for the one with NaOH.

• The Pb02 is reduced directly during the sulphidization of the paste without any expensive reactant, and it does not require a dedicated unit.

• No Pb02 is deposited on the anode because Fe oxidation occurs at a very low potential and is far from the oxygen evolution and the Pb02 deposition potentials.

• The anode used in the electrolytic cell consists of ordinary graphite and is long lasting, because no oxygen is evolved. This prevents any graphite oxidation to CO and C02.

• No mists are generated because no gas evolution occurs at either of the electrodes. • The energy consumption is lower compared to electrowinning with 0 2 evolution.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 811

• The diaphragm cell is similar in design to the one employed in Ni electrowinning which uses a polypropylene cloth as a diaphragm.

As was described above, this new CX-EWS process answers most of the technical and commercial questions that arose from the application of the other processes. An additional advantage is the possibility to obtain a complete electrochemical recovery of lead, producing high grade lead from grids and pole instead of lead alloys.

DEVELOPMENT STATUS OF CX-EWS PROCESS

The CX - EWS process, as previously described, consists of three main steps. The development status of each single step is defined below.

Paste Sulphidization to Convert all the Paste Components to PbS

After completion of the bench and laboratory pilot plant tests, this process step was tested in the desulphurisation reactor of the CX plant of Tonolli Canada. Batches of 8 tonnes were used in order to define all the reaction parameters. Sodium sulphide was used as the sulphidization reactant and a sodium sulphate solution was produced. The results of these tests on an industrial size reactor confirmed the performance obtained during the laboratory investigations. Taking advantage of these tests, PAQUES Biosystem B.V., a Dutch company that has significant expertise in biological sulphate reduction, applied its experience to achieve the same results in a safer and environmentally more friendly way.

Biological sulphate reduction has already been applied successfully on an industrial scale in the metallurgical industry since 1992 (5). Analogous to the "in situ" production of PbS from PbS04, this technology is applied on an industrial scale at the Budelco zinc refinery in the Netherlands where ZnSCU is converted "in situ" to ZnS. A paper about this project will be presented during this conference.

Leaching and Electrowinning of the Lead Sulphide Obtained from the Battery Paste

Twenty tons of lead sulphide, produced during the test run in Canada, was fed into a pilot plant in our laboratory. The capacity of the pilot plant was approximately 50 kg/day of electrolytic lead. The operation of the pilot plant lasted six months and provided all the data necessary to design an industrial demonstration plant.

A second pilot plant, based on the same technology, is currently operating in the research laboratory of The Doe Run Company at Viburnum, Missouri, U.S.A., treating primary lead sulphide concentrates with very promising results.

In both applications, the freshly precipitated lead sulphide is much more reactive than lead sulphide of mineral origin. The biogenic reduction of the lead compounds contained in battery paste into lead sulphide, and the integration of the lead sulphide into the CX-EWS process made the overall process very powerful and safe. With the fluoborate technology, it is now possible to process spent acid batteries by employing only a hydrometallurgical/electrochemical process and avoiding any pyrometallurgical operation, as

812 LEAD-ZINC 2000

shown in the conceptual flowsheet in Figure 3. The electrolytic loop is also capable of refining grids and poles together with the lead sulphide coming from paste sulphidization.

Spent lead acid batteries

Grids and

poles

Sbc

1

■ »

CX breaker

' Fluoboric leachi ng

i ί

r

sment

1 12 r

Paste

1 ' , H

r Bio-

sulphidisation

v Fluoboric leachi ng

Electrolytic refining FLUBOR Process

P 3 catho

1 de s99, 99°/ Ό

i k

Steam reforming

^ Steam ^

^

2/co2

▼ S°

Figure 3 - Electrochemical Lead Recovery from Spent Lead Acid Batteries

ECONOMICS

Studies were carried out to determine the investment and operating costs for processing batteries based on the new CX-EWS technology. Subsequently, these figures were compared with similar data for other available technologies. As an example of a competitive plant, a pyrometallurgical operation applying a breaker, paste desulphurisation, rotary furnace and thermal refinery considered. The plant size chosen for this comparison was 50,000 t/y of treated batteries producing 30,000 t/y of metallic Pb.

The estimated operating cost for the conventional pyrometallurgical plant is reported to be between 11 and 15 eVlb of Pb, averaging 13 0/lb (6). It was verified that, for the considered plant size, the operating cost would be about 11 0/lb. The operating costs for a plant with the same capacity and based on the CX-EWS technology, as shown in the flowsheet in Figure 3, are the following:

Unit $/t of Pb

Total 167

eVlbofPb Reforming/Bioreduction - THIOPACT 28 1.3 Pb electrolysis - FLUBOR® JJ39 63_

7.6

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 813

The investment cost for a conventional pyrometallurgical plant is about U.S. $11,000,000, whereas the cost for the CX-EWS technology is about U.S. $15,000,000. The investment cost difference, however, is amortised within a relatively short period considering the significant savings in the operating cost.

CONCLUSIONS

We hope enough information was provided to generate some interest about the potential and capabilities of fluoborate technology for processing lead, as a viable alternative to the traditional pyrometallurgical processes in terms of economical and environmental advantages.

REFERENCES

R.H. Reynolds, E.K. Hudson and M. Olper, "The Engitec CX Lead Acid Battery Recovery Technology," Lead-Zinc '90, T.S. Mackey and R.D. Pregaman, Eds., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1990, 1001-1021.

R.D. Prengaman, "Recovering Lead from Batteries", Journal of Metals, January, 1955, 31-33.

M. Olper, "CX - EW Process: A comprehensive Recovery System for Lead Acid Batteries," Recycling Lead & Zinc: The Challenge of the 1990's, Rome, Italy, 11-13 June 1991, 79-90.

M. Olper, "A Full Electrochemical Approach in Processing Junk Batteries," EDP Congress 1993, The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1993,959-966.

H. Dijkman, C.J.N. Buisman and H.G. Bayer, "Biotechnology in the Mining and Metallurgical Industries: Cost Savings through Selective Precipitation of Metal Sulphides." Copper 99 - Cobre 99, Volume IV, D.B. Dreisinger et al., Eds., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1999, 113-126.

F. Labro, "Welcome Address," Environmentally Friendly Lead and Zinc - The Challenge of the Recycling Millennium, Toronto, Canada, 25-29 May 1998, 17-26.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 815

PERFORMANCE OF A CONVENTIONAL CELL DESIGN FOR ZINC CHLORIDE ELECTROWINNING

C. Allen Noranda Inc., Technology Center

240 Hymus Boulevard Pointe-Claire, Quebec, Canada H9R-1G5

ABSTRACT

A series of zinc chloride electrowinning tests was carried out in a laboratory scale cell to evaluate the effect of temperature, current density, electrolyte composition, gas sparging and additives on the performance of the cell with respect to current efficiency, voltage, specific energy consumption and deposit morphology. The anode material was DSA for chlorine evolution and the cathode was made of aluminum. At 400 A/m2 current density, a current efficiency of 90.9%, a voltage of 3.4 V and an energy consumption of 2.78 kWh/kg were observed. Air sparging was found beneficial but was not pursued for health and safety reasons. Various additives in the same chemical family as tetrabutyl ammonium chloride (TBAC1) were tested. Both TBAC1 and tetrabutyl ammonium bromide were found to be equally effective. Attempts to carry out the electrolysis at 1200 A/m2 were not as successful.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

816 LEAD-ZINC 2000

INTRODUCTION

In the last few years, Noranda took advantage of recent expertise acquired in chloride metallurgy to recover zinc from low value residues or complex sulphides. The zinc chloride electrowinning unit is one of the required steps, that if successful, would permit the elaboration of a viable process without any need for a change of the facilities used for zinc extraction with conventional electrowinning in sulphate media. It would also allow for the recovery of chlorine. The work reported here concerns only the zinc electrolysis unit. A number of studies on zinc chloride electrowinning were reported in the late seventies and the eighties. MacKinnon, who did extensive work on chloride electrometallurgy at CANMET, identified TBAC1 (1) as the best additive to provide a smooth dendrite-free zinc deposit. He also provided some insight (2,3) into the effect of several impurities during zinc chloride electrolysis. However, in his study, sparging of air under the cathode was an essential part of the cell design. Thomas and Fray (4) identified some additives with promise at high current density; their experiments made use of sparging under the cathode and lasted for one hour only. Majima et al. (5) put more emphasis on understanding the fundamental behaviour of the zinc chloride solutions. Noranda felt that because of industrial hygienic concerns, such a design could not be implemented at an industrial scale. This paper summarises the work carried out at Noranda Inc., Technology Centre in the late nineties to operate a zinc chloride electrowinning cell at high current density without any gas sparging.

EXPERIMENTAL

Figure 1 shows a drawing of the cell used for testing as well as the laboratory scale assembly used for the experiments. The cell was made of one inch thick CPVC. A Terrylene diaphragm was maintained in place with a CPVC frame having a variable opening. Two DSA" anodes from ELTECH specific for chlorine evolution were used with an aluminium cathode. The surface of the cathode was carefully prepared by polishing with 600 grit sand paper, washed with 5% HC1 and then with distilled water. A specific active area was achieved by applying electrochemical tape on the area not to be plated. The three-compartment cell was held together with titanium bolts. Each compartment was insulated with a rubber gasket to prevent leaks. The cell was placed in a thermostatic bath. Solution was fed through a Teflon heat exchanger submerged in the bath maintaining a stable temperature. Tests were carried out using synthetic solutions. Anolyte and catholyte were collected in separate vessels. The net outlet of solution was through the anolyte to prevent chlorine bubbles from migrating to the cathode. Chlorine evolved at the anode was continuously removed with a vacuum pump. The anolyte was scrubbed by air sparging and the chlorine was captured in a NaOH-containing solution. The overflow from the catholyte compartment of the cell was collected in a vessel, where the catholyte, depleted in zinc, was enriched with a zinc feed solution and was recycled on both sides of the cathode. Additives were prepared daily and were added continuously to the catholyte by means of a peristaltic metering pump and a timer. Zinc deposits were peeled from the cathode, washed, dried and weighed. Four cells were connected in series. The system was connected to a data acquisition unit to monitor temperature, current and voltage every 10 minutes for the duration of the test. Voltage, current and temperature were averaged for the test. The Faradic current efficiency and specific energy consumption were calculated.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 817

Figure 1 - Cell Used for the Laboratory Work on Zinc Chloride Electrowinning

Typical conditions for the electrowinning are shown in Table I, together with the ranges for the variable conditions. These conditions were selected according to work already published in the literature.

Table I - Typical Operating Conditions for 6-h Electrowinning Tests Current density Temperature Cell solution Cathode material Cathode size Organic additive

Anode material

Anode frame Anode-cathode space Cell membrane Zinc bite

200-600 A/mz

30-50 °C 0-0.25M HC1, 30-70 g/L Zn, 0-0.5M NaCl aluminium 87 mm x 100 mm x 2 sides tetrabutyl ammonium chloride hydrate 98% (TBAC1) #34,585-7 Sigma-Aldrich, 10-40 mg/L Eltech DSA® coated with EC-300 on a titanium substrate of 0.2 cm with an active surface of 50.8 x 101.6 mm opening of 70 mm x 100 mm 34 mm Terrylene™ membrane 5-25 g/L

818 LEAD-ZINC 2000

RESULTS

Preliminary Work

Screening tests

A series of electrowinning screening tests was performed on synthetic solutions of zinc chloride electrolyte to evaluate the effect of 8 parameters on the three responses of cell voltage, current efficiency and zinc deposit morphology after 6 hours testing time. These tests were carried out using a design of experiment method with a two-level design for each factor. This series of tests was planned to obtain an indication of the magnitude of the effect of the eight factors on the electrolysis. This design of experiment approach was used to minimize the number of experiments required to study this number of factors, with a high degree of precision and a minimum amount of time. The results are shown in Table II.

Table II - Results of 6-h Screening Tests Factor [HC1] [TBAC1] [NaCl] [Zn] Temperature Gas sparging Current density Zinc bite

Cell voltage major

major major

major

Current efficiency

major

major major major

Morphology major

major major major

Power Major

Major Major

Major

Major effects were observed for the hydrochloric acid concentration, temperature, sparging under the cathode and current density. These effects seemed to warrant more work to optimise the parameters. Before proceeding, however, it was necessary to establish the best way of producing a synthetic solution of high quality suitable for electrowinning over a longer time period.

Synthetic Solution Preparation

A problem was rapidly identified when the electrolysis tests were operated for a period of 24 hours. The morphology of the deposit became dendritic with long dendrites at the edges of the exposed cathode area. To identify if an impurity in the reagent were the cause, different methods of making the synthetic solution were tested. Technical zinc chloride (Anachemia AC-9939T) was subjected to a cementation stage with zinc dust to remove inorganic impurities from the solution. The zinc deposit was still dendritic as can be seen from the first column of Table III. The same solution was then subjected to treatment with activated carbon to remove organic impurities. Again the zinc morphology was dendritic (second column of Table III). Special High Grade zinc shots were dissolved in hydrochloric acid. The dissolution process was long, but the resulting solution was of good quality and the zinc deposit was smooth. Reagent grade zinc oxide (Anachemia AC-9952) was mixed with HC1 to produce zinc chloride solution. This was the preferred way of preparing the zinc chloride solution as the ZnO readily dissolves and gives a very clean solution. The zinc morphology was smooth when that solution was used for electrolysis.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 519

Table III- Effect of the Type of Preparation of the Synthetic Solution on the Electrolysis Performance

Reagents Treatment

As Ca Cd Co Cu Fe Ge K Mg Mn Ni Pb S Sb Se Tl Current Density [TBAC1] Current Efficiency Cell Voltage Specific Energy Morphology

Effect of the Frame

mg/L mg/L mg/L mg/L mg/L mg/L μΕ/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L A/m2

mg/L % V KWh/kg

ZnCl2

Cementation

<5 529

0.002 0.013 <0.25 0.56 4.84 677 151 2.36

0.089 0.125 43.4

0.110 <5

0.214 200 15

37.3 2.73 6.01 bad

ZnCl2

Cementation/ Activated Carbon

<5 511

0.004 0.026 <0.25 11.4 4.50 663 146 2.35

0.137 0.153 46.6 0.044

<5 0.173 200 15

37.6 2.75 5.98 bad

Zn° + HC1 None

<5 0.64

0.180 0.009 <0.25 4.70 2.55 <50 0,13 0.14

0.107 21.8 7.12

0.029 <5

0.021 400 40

91.9 3.13 2.79

Smooth

ZnO + HC1 none

<5 0.79

0.198 0.002 <0.25 <0.25 2.27 <50 0.11

<0.05 0.053 8.54 6.06

0.027 <5

0.149 400 40

95.7 3.10 2.65

smooth

A frame was used to maintain the diaphragm in place in front of the anode. The size of the opening was changed as described in Figure 2. In Figure 2A the exposed area of the cathode is the same size as the anode frame area. Even though this anode is smaller than the cathode and never changes, the frame works as a diffuser for the current. The resulting effect is a more elevated current at the edges of the cathode than at the center. In Figure 2B the anode frame area has been slightly reduced and the edges of the cathode are subjected to a lower current density. In Figure 2C the frame are has been dramatically reduced and the edges are subjected to the same current density as the center of the cathode. Frame B was selected for the optimisation tests.

820 LEAD-ZINC 2000

( A ) ( B ) ( C )

Figure 2 - Effect of the Frame on the Morphology of the Zinc Deposit. A) Cathode and Anode Frame Opening are the Same Size, B) Cathode is Slightly Larger than the Anode Frame Opening, and C) Cathode is Much Larger than the Anode Frame Opening

Optimisation Work

For the optimisation work, the deposition time was increased to 24 h, which would normally give a good indication of the performance at the industrial scale. The previous testing period was so short that the effect of additives could not be properly demonstrated. The level of additives was then included in the optimisation work. Current density was definitely to be tested on a 24-h time period, considering the major impact it had on all responses. A new factor, not tested before, was the catholyte recirculation rate. A new series of optimisation tests with a five level central composite design of experiments was initiated to obtain more precise results on the effects of four factors ([HC1], [TBACl], current density and catholyte recirculation flowrate) on zinc electrowinning. The temperature was fixed at 35°C, the concentration of zinc chloride in solution was maintained at 50 g/L and the concentration of sodium chloride was fixed at 0.25 M.

Current Efficiency

The current efficiency was mainly a function of the TBACl concentration and current density. The relationship is quadratic with an optimum for the TBACl concentration between 30 and 40 mg/L, as may be seen in Figure 3. Figure 3 is a contour plot of the current efficiency response as the two major factors, current density and the TBACl additive are varied.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 821

50

DESIGN EXPERT Plot

Actual Factors: X = CD. Y= TBAC1

Actual Constants: Flowrate 63 mL/min HCI = 0.35 M

40

30

10

._

Current Efficiency

100.0

9JLQ

.85-0

- "" 8Qi> 75-0

95 0

»u.u

_ - -

200 250 300 350

Current Density

400 450

Figure 3- Effect of [TBAC1] (mg/L) and Current Density (A/m2) on Current Efficiency (%)

The effect of the TBAC1 is thought to be two-fold:

• The strong polarising agent increases the hydrogen overvoltage for some impurities, such as antimony.

• The organic additive coats the growing crystals at the surface of the cathode where the current density is higher. The additive blocks the site for new zinc to be deposited, producing a more even surface.

An increase in the TBAC1 concentration increases the current efficiency whereas an increase in current density resulted in a lower current efficiency, as can be deduced from the following equation:

C.E. = 77.94 - 0.056 CD +2.1 lfTBACl] - 0.028 [TBAC1]2 (1)

Where C.E. is the current efficiency in % and CD is the current density in A/m2. It can be calculated from Equation 1 that it is possible to obtain 95% C.E. at 400 A/m2 with the addition of 40 mg/L TBAC1. A few cyclicvoltammetry tests were carried out to demonstrate that there is no degradation of the additive over time. Cyclicvoltammetry curves on the freshly made solution exactly matched those from a solution made one month earlier.

822 LEAD-ZINC 2000

Cell Voltage

The cell voltage varied from 2.99 to 4.3 Volts for the design of the experiments. The model for the cell voltage was not reliable, as the correlation coefficient was 0.52. However, a plot was made to evaluate the effect of TBACl concentration on cell voltage at a fixed current density, hydrochloric acid concentration and sodium chloride concentration. This plot (Figure 4) shows a net increase in cell voltage with increasing concentration of TBACl. In fact, an increase of 14 mV per mg/L of TBACl added may be expected.

0 10 20 30 40 50 60

[TBACM (mg/L)

Figure 4- Effect of TBACl on Cell Voltage at 340 A/m2, [HCl] 0.08M, [NaCl] 0.25M

The effect of TBACl combined with the effect of current density will cause an increase in cell voltage. The TBACl polarised the surface of the cathode, and in fact, increased the overvoltage. The only way to overcome this effect is by the addition of HCl or another support electrolyte (such as NaCl) that will increase the conductivity of the electrolyte. This is discussed below.

Power Consumption

The power consumption was most affected by the TBACl concentration of the solution and by the current density as shown in Figure 5. A maximum in the TBACl concentration efficiency may be seen between 30 and 40 mg/L. The model shows a quadratic relationship between current density and [TBACl]. In fact the model may be asymptotic after the maximum shown, but this type of relationship is not found in the software used to analyse the data. There is a trade off between current efficiency and cell voltage. As the [TBACl] is increased, the cell voltage increases. The result will be higher power consumption. If the [TBACl] is not high enough, the current efficiency will be lower, increasing the power consumption. The desirable response is for the power consumption to be as low as possible, as it affects the operating costs of the process.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 823

Power

DESIGN EXPERT Plot

Actual Factors: X = CD. Y = TBAC1

Actual Constants: Flowrate63mL/min HC1 = 0.12M

2.5

,--"-""""

3.0

3.5

4.0

4.5

3.5

/

5.0_

Current Density

Figure 5 - Effect of Current Density (A/m2) and [TBAC1] (mg/L) on Power Consumption (kWh/kg)

An increase in the current density results in an increase in power consumption since the voltage of the system is increased. There might be a way to reduce the power consumption by increasing the acid and NaCl concentrations. The actual relationship for power consumption is as follows:

Power = 3.20 + 4.445 x 10 "3 CD. - 0.095 [TBAC1] + 1.299 x 10"3 [TBAC1]2 (2)

If the cellhouse is operated at 400 A/m2 and at least 30 mg/L TBAC1, a power consumption of 3.5 kWh/kg or lower can be expected.

Deposit Morphology

Morphology was arbitrarily defined on a scale of 1 to 10, the lowest number being the best; i.e., smoother, dendrite-free and no powdery areas. The desirable response for morphology is to be low, preferably below 3. We expected the morphology to be a function of all the factors, each of them affecting the morphology in a different way. Additives should produce an intermediate overpotential as to randomise the deposition of zinc and to coat the areas on the cathode where the deposition is too fast and slow down the deposition elsewhere. The flow of solution should affect the rate of mass transport by convection. The HC1 is believed to increase the conductivity. The model obtained for the morphology from the design of experiments has a coefficient of correlation of 0.65. All factors were found to have an effect, the main one being the HC1 concentration, followed by the TBAC1 concentration.

The relationship between the factors tested is the following:

Morphology = 5.95 + 0.14 CD. - 0.051 Flowrate + 0.22 [TBAC1] 11.74 [HC1] -4.97 x 10J [TBAC1]2 (3)

824 LEAD-ZINC 2000

From equation (3), we may deduce that an increase in [HC1] leads to an improvement in the morphology. According to the model, an increase in [TBAC1] was beneficial on morphology. We expected an optimum or a plateau for TBAC1. Figure 6 shows a contour plot for the effect of hydrochloric concentration and TBAC1 on the morphology at a fixed current density, NaCl concentration and recirculation flow. From this figure, it may be seen that the TBAC1 is effective and that a high concentration will not be detrimental to the morphology.

50

DESIGN EXPERT Plot

Actual Factors: *" X = HC1 Y = TBAC1

30 Actual Constants: CD. 388 A/m* Flowrate 82mL/min

20

10

0.0Ö Ö!Ö7 0~Ϊ4 02Ϊ 0.28 Ö35

HO

Figure 6- Effect of [HC1] (mg/L) and [TBAC1] (mg/L) on the Morphology at 388 A/m2, [NaCl] 0.25 M, Flowrate 82 ml/min

It is possible to achieve a good morphology at 400 A/m2 by increasing the HC1 content of the solution above 0.35M and by using at least 30 mg/L TBAC1.

Another series of 24 tests was conducted under continuous operation aiming to determine whether it was possible to obtain good Faradic performance and a smooth morphology by operating the cellhouse at a current density of 1,200 A/m2. Operating at such a high current density is vital to the viability of the process since electrolysis costs would account for a major part of a refinery project. The study was carried in two phases:

Addition of reagent in combination or in massive doses, at 400 and at 1200 A/m2. Variations in the composition of the solution (HC1, NaCl).

Effect of Additives at 400 A/m2

Tests E4 to E14 were conducted to find the best additive to optimise zinc electrowinning. Quaternary ammonium salt solutions were used for these tests. Additives

Morphology T~^r

6

8 7

5 4

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 825

TBACl1, TBAC1-S2, MTBCl3, TMC14 and TBABr5 were tested in concentrations of 200 and 400 mg/L and in a 2M NaCl solution. Table IV gives the results for those tests. The best additives found were MTBCl and TBACl for their excellent results in terms of specific energy consumption and deposit morphology. These additives were more performing at a high dosage (400 mg/L). The TBABr, at a concentration of 400 mg/L, also gave good results in terms of current efficiency and morphology. However, with the TBABr, a yellow deposit tends to form in contact with the electrolyte. For this reason, the TBABr should not be used to avoid obstructing the solution circulation piping. The TMC1 additive, at the concentration tested, was not appropriate. The zinc deposit morphology obtained with this additive was dark grey, with very long dendrites along the sides. This deposit was very soft and crumbled easily.

Effect of additives at 1200 A/m2

Tests 15 to 24 aimed at assessing zinc electrowinning at 1,200 A/m . A higher current density normally results in an enhanced effect of the impurities. It was expected that the performance of the cell at 1200 A/m2 would be much worse than its performance at 400 A/m2. For this series of tests, combinations of the best additives identified in the preceding section were prepared. Two new additives; i.e., Arabic gum and gelatine, were also tested. The results are given in Table IV. Overall, the results obtained with regard to specific energy consumption were average. In three cases, the current efficiency was above 90%. However, the deposit morphologies were not acceptable, being very rough in two out of three cases. In several instances, the tests lasted less than 24 hours because of short circuits caused by large dendrites forming at the edges of the zinc deposit. As expected, cell voltages were higher than those measured at a current density of 400 A/m2, by one to three volts. The large amount of additives required to produce smooth deposits and the magnitude of the cell voltage are two indicators that the performance of this type of design for high current density operation is unsatisfactory.

Effect of fNaCll in the Supporting Electrolyte

Table V gives the results obtained in these tests with 2 M and 3.4 M NaCl at a current density of 400 A/m2. The addition of NaCl had a positive effect at all levels; it improved the current efficiency by 2%, decreased cell voltage by 250 mV, and improved the morphology. Sodium chloride acts as a support electrolyte and appears to better distribute the current across the cathode surface.

Table V- Effect of NaCl Concentration

Test

E2 E5 .

[Zn]Ceii

(g/L) 32.0 56.1

[NaCl]

(M) 3.4 2

CD.

(A/m2)

408 398

C.E.

(%) 98.97 96.78

C.V.

(V) 3.186 3.451

Energy

(kWh/kg)

2.639 2.924

Morphology

3 4

Duration

Test (h)

23.62 24.00

' TBACl: tetrabutyl ammonium chloride manufactured by Aldrich 2 TBACI-S : tetrabutyl ammonium chloride manufactured by Sachem 3 MTBCl: methyltributyl ammonium chloride manufactured by Sachem 4 TMCI: tetramethyl ammonium chloride manufactured by Sachem 5 TBABr: tetrabutyl ammonium bromide manufactured by Sachem

Tab

le IV

- R

esul

ts o

f th

e T

ests

on

the

Effe

ct o

f th

e A

dditi

ves

Tes

t [Z

n]ce

||

(g/L

)

Typ

e of

A

ddit

ive

Res

ults

at

400

A/m

2

E5

E4

E7

E9

E8

E6

E14

E12

E13

Ell

47.6

38.5

42.3

43.6

41.8

43.9

38.8

39.0

48.0

39.3

TB

AC

1 (2

00)

TB

AC

1 (4

00)

TB

AC

1-S

(200

)

TB

AC

1-S

(400

)

MT

BC

1 (2

00)

MT

BC

1 (4

00)

TB

AB

r (2

00)

TB

AB

r (4

00)

TM

C1

(200

)

TM

C1

(400

)

Res

ults

at

1200

A/m

2

E16

E18

E15

E17

E19

E20

51.3

46.0

4.5

45.0

44.8

44.3

TB

AC

1 (4

00)

TB

AC

1 (6

00)

TB

AC

1 (4

00)

MT

BC

1 (4

00)

TB

AC

1 (4

00)

TB

AB

r (4

00)

Ara

bic

Gum

(20

)

Gel

atin

e (1

0)

CD

.

(A/m

2 )

398.

1

395.

1

403.

6

400.

6

403.

6

398.

1

400.

8

401.

5

400.

8

401.

5

1179

1173

1179

1173

1174

1174

C.E

.

(%)

96.7

8

100.

40

100.

13

100.

21

100.

57

99.8

3

89.6

8

100.

24

45.4

0

96.7

8

55.7

5

99.0

0

90.2

1

96.8

7

80.4

2

82.5

2

C.V

.

(V)

3.45

1

3.40

9

3.47

4

3.51

1

3.41

0

3.37

0

3.41

0

3.53

3

2.56

3

3.33

4

4.35

2

6.51

6

5.06

4

5.56

8

5.21

3

5.40

8

Pow

er

(kW

h/kg

)

2.92

4

2.78

4

2.84

4

2.87

3

2.78

0

2.76

8

3.11

8

2.89

0

4.62

9

2.82

5

6.40

1

5.39

7

4.60

2

4.71

3

5.31

5

5.37

4

Mor

phol

ogy

4 2 3 3 4 3 5 2 9 9 6 6 7 6 9 9

Dur

atio

n

Tes

t (h

)

24.0

0

23.9

8

23.8

2

24.0

0

23.8

2

24.0

0

23.9

5

24.0

0

23.9

5

24.0

0

24.1

0

6.98

24.1

0

6.98

5.97

5.97

r en

> O

N

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 827

Another supporting electrolyte, NH4CI, was tested. This product would decrease cell voltage, thus increasing the electrical conductivity in the electrolyte, as for the case of NaCl. Several papers have been written on the efficiency of this product for electrowinning in a chloride solution. Tests were performed at high current density, and one of them combined 1.0 M NaCl with 0.5M NH4C1 (test E-24). The results are given in Table VI.

Test

E22 E24

[Zn]ce||

(g/L)

66.0 37.2

Table VI - Results for Two Supporting Electrol Additive C D . C.E. C.V. Energy

(A/m2) (%) (V) (kWh/kg)

NH4C1(1.0) 1132 40.84 1.719 3.450 NH4C1(0.5) 1147 94.60 4.863 4.215 +NaCl(1.0)

ytes Morphology

3 3

Duration

Test (h)

23.93 6.50

In test E22 where only NH4CI was used, the current efficiency was very low. The deposit morphology was superior to that obtained in the tests where the supporting electrolyte was NaCl. This result is only an artifact since corrosion is most likely responsible for the smooth morphology. Test E-24 lasted only 6 hours because of a small explosion caused by the high concentration of hydrogen. This would be unacceptable in an industrial operation. This explains the high current efficiency in this test since the duration of the test strongly affects current efficiency. Also, a solid of unknown nature was produced on the walls of the cell in the anolyte compartments and on the anodes themselves. Therefore NH4C1 as a supporting electrolyte should be rejected.

CONCLUSIONS

The Performance of the laboratory scale cell for zinc chloride electrowinning at 400 A/m2 was very good. The TBAC1 and MTBC1 are the best additives at a current density of 400 A/m2 and for a NaCl concentration of 2 M. Table VII summarises the results for these experiments and compares them with those obtained from zinc sulphate electrowinning at the laboratory scale and at the industrial scale.

Table VII - Comparison of Performance Factor This work NTC Sulphate Cell Sulphate Industrial Current Density (A/m2) Temperature (°C) Deposition Time (hr) Cathode area (m2) [Zn]0 (g/L) Zn Bite (g/L) Media [Acid] Additive C. E. (%) C.V. (V) Power Consumption (kWh/kg)

400 40 24

0.0102 50 15

Chloride 9

TBAC1 40 mg/L 91.9 3.13 2.79

450 38 24

0.0018 55 7

Sulphate 180

Gelatine 80 g/t 93.0 3.2

2.82

450 35-40 24-72

3.2 50

5-10 Sulphate 160-200

Gelatine, glue, MPC 90-93

3.4 2.8-3.3

828 LEAD-ZINC 2000

The TBACl concentration affected current efficiency. An optimum for the TBACl concentration was found between 30 and 40 mg/L. Cell voltage showed a net increase with increasing content of TBACl. In fact, an increase of 14 mV per mg/L of TBACl may be expected. The TBACl is effective for improving the smoothness of the zinc deposit, and a high concentration will not be detrimental to the morphology. The presence of acid had no influence on the current efficiency. An increase in HC1 concentration improves the morphology. An increase in flowrate was found to improve the morphology but had no effect on the other responses. An increase in current density, form 400 to 1200 A/m2, resulted in a lower current efficiency and a higher energy consumption. The morphology was adversely affected by an increase in current density. Increases from 2 to 3.4 M NaCl concentration in the solution gave a lower cell voltage and lower power consumption with a slight improvement in the morphology.

For the 1,200 A/m2 electrowinning tests, the results are not as encouraging. Deposit morphology is not acceptable in the presence of 2 M NaCl. Most of the tests were performed at reduced hydrochloric acid concentrations when it was determined that its presence was deleterious. Even massive doses of additives were not sufficient to increase the current efficiency to the required 90% level. The best additive for electrowinning in a chloride solution was TBACl, but at 1,200 A/m2, the addition of 600 mg/L of TBACl did not sufficiently improve the process efficiency. Several tests had to be terminated because of short circuits.

Ammonium chloride (NH4C1) displays possibilities that seem difficult to exploit. The NH4CI appears to provide a better zinc deposit morphology at a current density of 1,200 A/m2. During small-scale tests, however, several small explosions were observed. The explosion risk could possibly be reduced by operation at a lower current density, but this defeats the purpose of operating at high current density.

The tests at high current densities showed that there was little latitude to produce a compact and smooth zinc deposit at high current efficiencies. The morphology of the deposit was very rough with powdery areas and long dendrites. The massive doses of additives did not have a sufficient effect to allow the operation at a high current density. It is evident that zinc electrowinning in a chloride solution, operated in a conventional manner without gas sparging under the cathode, did not provide sufficiently interesting results to pursue work on this design. A new approach has been evaluated. Its goal is to demonstrate the feasibility of using the spouted bed electrode (6) to produce zinc from a zinc chloride solution.

ACKNOWLEDGMENTS

The author wishes to thank Dr. Pierre Claessens and Dr. Ole Morten Dotterud for their kind review and Noranda Inc., Technology Center for permission to publish this work.

REFERENCES

1. D.J. MacKinnon and J.M. Brannen, "Evaluation of Organic Additives as Levelling Agents for Zinc Electrowinning from Chlorides Electrolytes", Journal of Applied Electrochemistry. Vol. 12, 1982, 21-31.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 829

2. DJ. MacKinnon, J.M. Brannen and V.l. Lakshmanan, "Zinc Deposit Structures Obtained from Synthetic Zinc Chloride Electrolyte", Journal of Applied Electrochemistry. Vol. 9, 1979, 603-613.

3. D.J. MacKinnon, J.M. Brannen and R.M. Morrison, "Zinc Electrowinning from Aqueous Chloride Electrolyte", Journal of Applied Electrochemistry, Vol. 12, 1982, 39-53.

4. B.K. Thomas and D.J. Fray, "The Effect of Additives on the Morphology of Zinc Electrodeposited from a Zinc Chloride Electrolyte at High Current Densities", Journal of Applied Electrochemistry. Vol. 11, 1981, 677-683.

5. H. Majima, E. Peters, Y. Awakura, S.K. Park and M. Aoki, "Electrical Conductivity of Acidic Chloride Solutions", Metallurgical Transactions B, Vol. 19B, 1988, 53-58.

6. V. Jiricny, A. Roy and J. W. Evans, "Spouted Bed Electrowinning in the Recovery of Zinc from Scrap Galvanized Steel", EDP Congress 1998. G.W. Warren, Ed., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1998, 411-426.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 831

SPOUTED BED ELECTROWINNING OF ZINC FROM CHLORIDE ELECTROLYTES

J. W. Evans and A. Roy Department of Materials Science and Mineral Engineering

University of California Berkeley, California, U.S.A. 94720

C. Allen Noranda Inc., Technology Center

Pointe Claire, Quebec, Canada H9R 1G5

ABSTRACT

A spouted bed electrode (SBE) has been used, in laboratory experiments, to electrowin zinc from zinc chloride electrolytes with the objective of determining the suitability of this electrode for commercial application. Zinc was successfully electrowon at current densities ranging up to 4,38 lA/m2, which are an order of magnitude larger than those used at present in the zinc industry. Current efficiencies, for the cell with the SBE, have been as high as 90% and DC energy consumption as low as 2.8 kWh/kg zinc deposited. The performance of the cell has been measured as a function of current density, anode type, cell diaphragm, addition of tributyl ammonium chloride and purging of the electrolyte (to remove dissolved chlorine).

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

832 LEAD-ZINC 2000

INTRODUCTION

This paper describes the application of the spouted bed electrode (SBE) to the electrowinning of zinc from chloride electrolytes. There have been several previous investigations into zinc electrowinning from chloride electrolytes (1-12); none used the SBE. One difficulty in electrowinning zinc from chloride electrolytes is that the deposited zinc is frequently not dense, smooth and adherent as expected in electrowinning from conventional acid sulfate electrolytes (8). A characteristic of some investigations is that they have entailed additives or gas sparging to improve the deposit morphology. Additives affect the nucleation and growth of the zinc whereas gas sparging enhances mass transfer at the cathode. Improved mass transfer means that the electrolysis occurs at a smaller fraction of the limiting current, which usually minimizes formation of dendritic deposits. It will be seen below that gas sparging of the catholyte has the additional benefit of increasing current efficiency in the SBE.

The previous investigation that is closest to the present one is that of Tuffrey, Jiricny and Evans (5). Those investigators used a cell with a fluidized bed electrode (FBE) to study zinc electrowinning from chloride electrolytes. Both the FBE and the SBE are comprised of particles but recent work at Berkeley has suggested that the SBE is superior. Some of the results of Tuffrey et al. were encouraging. Current efficiencies were close to 90% in some experiments at current densities of 4,000-8,000 A/m2; electrical energy consumptions were 3-5 kWh/kg Zn in this range of current densities for some electrolytes. The means to achieve a comparable, or better, cell performance, with good zinc deposits, was an objective of the present investigation.

APPARATUS AND PROCEDURE

The cell used in this investigation is seen in Figure 1. Its cathode consists of a spouted bed of zinc particles that grow as zinc is deposited on them. The particles lie in two regions on either side of the central "draft tube". Catholyte is pumped into the cell through the nozzle at the bottom and most of it flows up the draft tube to exit the cell at the top. Particles fall through the gaps at the bottom of the cell into the catholyte stream and are swept up the draft tube. At the top of the draft tube the particles fall out of the catholyte stream and drop onto the top of the slowly descending bed of particles on either side of the draft tube. The particles thus circulate repeatedly in the cell as they grow and their relative motion prevents their adhering. At the back of the cell is a plate of aluminum (the current feeder) which makes electrical contact with the particles. The particles are separated from the anode side of the cell by a porous material (the diaphragm) and if the permeability of this material is low, the anolyte and catholyte can have different compositions. Daramic® from Daramic SA, France (0.6 mm thick) or from Daramic Corp. of America (0.28 mm thick) were used as the diaphragm in this work. In the present investigation, the anodes used have been DSA® anodes from Eltech Corp. The anodic reaction is the evolution of chlorine but two types of DSA anode were tried, one intended for oxygen evolution (although chlorine evolution occurs in the cell) and the other intended for chlorine evolution.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 833

78

Catholyte Inlet

Diaphragm Anode Compartment p c * Anode

ALL DIMENSIONS ARE IN MM

Figure 1 - Schematic Diagram of the Spouted Bed Cell

The system in which the cell functioned is shown in Figure 2. Catholyte and anolyte were pumped separately to the cell from separate temperature controlled reservoirs. The electrolytes from the cell flowed back to their reservoirs, the anolyte containing the chlorine gas (as bubbles and in solution) evolved at the anode. The chlorine bubbles were disengaged from the anolyte in the anolyte reservoir and this chlorine stream flowed to a vessel where the chlorine was absorbed in a solution of potassium hydroxide. The absorber was connected to a water jet ejector to place a slight vacuum on the absorber and entrain any residual chlorine in the water passing into the laboratory sink. As discussed below, the phenomenon that lowers current efficiency in zinc electrowinning from chloride electrolytes is transport of chlorine to the cathode and its subsequent reaction. That reaction is either the chemical attack of the zinc by the chlorine (resulting in zinc chloride) or the electrochemical reduction of the chlorine; either reaction lowers the measured current efficiency. Transport of chlorine bubbles to the spouted bed of zinc particles is precluded by the diaphragm and separate electrolyte loops, but chlorine is soluble in the electrolyte and transport of dissolved chlorine through the diaphragm is still possible. This transport could be minimized by sparging of the anolyte with air (thus reducing the amount of chlorine in solution) and provision was made for purging the anolyte of chlorine in this way. A second concept was to sparge the catholyte with nitrogen (again reducing the amount of dissolved chlorine but this time on the cathode side of the cell) and provision was made for this too. Not shown in Figure 2 are the power supply and monitoring computer. The latter tracked the cell voltage and current throughout an experiment.

834 LEAD-ZINC 2000

Anolyte Tank

To PC via A/D Converter

Γ Catholyte Tank

To Sink

Tube Immersed in Anolyte for Bubbling

Figure 2 - Schematic Diagram of the Experimental Set-up for Spouted Bed Electrowinning of Zinc from Chloride Solution

The electrolytes used in this investigation were prepared from reagent grade hydrochloric acid, zinc oxide and sodium chloride. Particles were cut wire particles from Abrasive Materials and were approximately cylinders 1.45mm in diameter by 1.45mm long. Compositions and other experimental parameters are given in the figure captions below. The experimental procedure entailed the assembly of the cell, preparation of solutions (dissolution of zinc oxide in hydrochloric acid solution in the case of the catholyte), pouring the solutions into the reservoirs, bringing the reservoirs up to temperature (38°C) and adding a weighed amount of particles to the cell (800 g). The water jet ejector, nitrogen to purge the anolyte and current were turned on and then, as quickly as possible, the pumps were turned on and the flowrate of the catholyte adjusted to bring about satisfactory spouting. Cell current and voltage were monitored until completion of the experiment. Experiments were run for 30 A-hours. The electrolytes were not replenished during that period so that some change in composition occurred; typically the zinc content of the catholyte would drop from an initial 30g/L to approximately 22 g/L at the end of the experiment. After the current was turned off, the zinc particles were removed from the cell as quickly as possible, washed, dried and re-weighed. Current efficiency results below are based on the weight increase due to zinc deposition.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 835

EXPERIMENTAL RESULTS AND DISCUSSION

Replicate experiments showed that the current efficiency could be measured to a precision of better than ±2 %. The precision of the voltage and current measurements is much higher (±0.2% for the voltage and ±0.1 % for the current).

Effect of Cell Current

Figure 3 gives the results of a representative experiment; this experiment was carried out at a current density of 2190 A/m2 . The current density here is the current per square meter of diaphragm. Because the dimension of the cell perpendicular to the diaphragm is only a few centimeters, this current density gives a good impression of the intensity of electrowinning achievable in the SBE. For example, by comparing the current density of Figure 3 with conventional electrowinning of zinc, which is usually carried out at about 500 A/m2, it is seen that the intensity in the SBE is high. Because a chlorine evolving DSA anode was not initially available, the first several experiments were carried out using an anode intended for oxygen evolution, although the anodic reaction was still chlorine evolution.

DSA(02) Anode, 2190A/mA2

II . 4.5

4

3.5

3

- * *■ * * * * * A

-*—Vol tage - Energy for Electrowinning CE

15 20

Charge Passed, Ah

Figure 3 - Voltage, Current Efficiency and Energy Consumption for a Representative Experiment. Oxygen Evolving DSA. 38°C. Catholyte 30 g/1 Zn (as ZnCl2), 7 g/1 HC1 and 50 g/1

NaCl

836 LEAD-ZINC 2000

The current efficiency shown in Figure 3 is one determined by weighing the zinc before and after the run. It is therefore a value averaged over the run and an assumption of constant current efficiency is implicit in plotting it as shown. The measured cell voltage drops at the start of the run (probably because of the wetting of the diaphragm) but is thereafter nearly constant. The energy consumption curve in Figure 3 is calculated from the measured cell voltage and current efficiency.

The results of operating at lower and higher current densities are shown in Figure 4 (20 Ah results); they are consistent with the concept that the loss of current efficiency is due to chlorine, dissolved in the electrolyte, finding its way into the cathode chamber. There it would re-dissolve the zinc by chemical reaction or a portion of the cell current would be diverted to its electrochemical reduction. It is likely that this mechanism is governed by the transport of chlorine into the cathode chamber (e.g., by diffusion through the diaphragm) and is probably roughly independent of current. At low cell current, the fraction of current diverted to chlorine reduction (or the current equivalent to the zinc re-dissolved) is therefore large, and the current efficiency poor. At higher current densities this inefficiency is much smaller as a fraction of the cell current; i.e., the current efficiency, is higher. Note that, as expected, the cell voltage rises monotonically with current density but that electrical energy consumption reaches a minimum of 3.8 kWh/kg at approximately 2000 A/m2 because the current efficiency increases significantly with current density at low current density. Beyond that current density, gains in current efficiency, on increasing current density, are more than offset by increases in cell voltage.

Double Diaphragm

To test the notion of current efficiency loss by chlorine transport through the diaphragm, and to seek higher current efficiencies, a few experiments were carried out with the cell fitted with two layers of Daramic. Figure 5 summarizes the results (at 20 Ah). The current efficiency was indeed improved by the double diaphragm, comparing this figure with Figure 4. The effect on current efficiency is stronger at lower current density. However, the double diaphragm also increases the cell voltage so that the improvement in electrical energy consumption is marginal. The electrical energy consumption is now a minimum of approximately 3.6 kWh/kg at about 1500 A/m2. Although these experiments with two layers of Daramic yielded only small improvements in cell performance, they were a clear demonstration that the diaphragm is responsible for a significant part of the cell voltage and, therefore, it may be possible to reduce cell voltage using other materials.

TBAC1 Additions

Tributyl ammonium chloride, added to the electrolyte, has resulted in improved cell performance in conventional cells used for chloride electrolytes and this additive was tried at Berkeley in the SBE at the level of 10 ppm in the catholyte (with no TBAC1 addition to the anolyte). Results are summarized in Figure 6 and should be compared with the results from additive free electrolyte in Figure 4. At low current density the TBAC1 improved the current efficiency considerably. However, at high current the effect was smaller than the precision of the current efficiency measurements. It is seen that the improvement in current efficiency at low current density lowers both the minimum energy consumption (to 3.2 kWh/kg) and the current density for that minimum.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 837

0)

Cell Voltage -Standard Experiment

1000 2000 3000 4000 5000

Current Density A/m2

c ° > 5 öi 5 Φ Ε £ 4 = = S HI » i 3

8 2

EEC -Standard Experiment

1000 2000 3000 4000

Current Density A/m2

5000

I 100

80

60

UJ 40

Current Efficiency - Standard Experiment

1000 2000 3000 4000 Current Density. A/m2

5000

Figure 4 - Cell Voltage, Current Efficiency and Energy consumptions at Different Current Densities (730, 1022, 1461, 2190, 2920, 3651 and 4381 A/m2) Old 22 mm Cell, Al Current

Feeder, DSA (02) Anode, 0.6 mm Daramic [30 g/1 Zn, 7 g/1 HC1 and 50 g/1 NaCl Catholyte] 38°C Temperature

Anode Design for Chlorine Evolution

For the remaining experiments, the DSA made by Eltech Corp for chlorine evolution was used. Figure 7 should be compared to Figure 4 to judge the effect of this change. The first (expected) result observed is a reduction in cell voltage by something of the order of half a volt, when using the anode designed for chlorine evolution. Also there is an unexpected increase in current efficiency, at least for the lowest current density.

838 LEAD-ZINC 2000

CELL VOLTAGE - TWO 0.6mm DARAMICS

1000 2000 3000 4000

Current Density, A/m2

5000

TWO 0.6mm DARAMICS

5000

Current Density, A/m2

o> 6 r 5.5

5

Ö .2 4 · 5

4 3.5

- 1 3 ü

\i

I! 0

EEC - TWO 0.6mm DARAMICS

1000 2000 3000 4000

Current Density, A/m2

5000

Figure 5 - Cell Voltage, Current Efficiency and Energy Consumptions at Different Current Densities for Two 0.6-mm Daramics (1022, 1461, 2920 and 4381 A/m2) Old 22 mm Cell, Al

Current Feeder, DSA (02) anode, 0.6-mm Daramic [30 g/1 Zn, 7 g/1 HCl and 50 g/1 NaCl Catholyte], 38°C Temperature

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 839

CELL VOLTAGE - TWO 0.6mm DARAMICS

1000 2000 3000 4000 5000

Current Density, A/m

CE - TWO 0.6mm DARAMICS

5000

Current Density, A/m

EEC - TWO 0.6mm DARAMICS

1000 2000 3000 4000

Current Density, A/m2

5000

Figure 6 - Cell Voltage, Current Efficiency and Energy Consumptions at Different Current Densities for 10 ppm TBACL (1461, 2190, 2920, 3651 and 4381 A/m2) Old 22 mm Cell, Al Current Feeder, DSA (02) Anode, 0.6 mm Daramic [30 g/1 Zn, 7 g/1 HCL and 50 g/1 NaCl

Catholyte], 38°C Temperature

840 LEAD-ZINC 2000

CELL VOLTAGE - 10ppm TBACL

O 80

1000 2000 3000 4000 5000

Current Density, A/m2

CE-10ppm TBACL

1000 2000 3000 4000 5000

Current Density, A/m2

EEC -10ppm TBACL

1000 2000 3000 4000

Current Density, A/m2

5000

Figure 7 - Cell voltage, Current Efficiency and Energy Consumptions at Different Current Densities for DSA(Cl) Anode (1022, 2190 and 4381 A/m2), Old 22 mm Cell, Al Current

Feeder, DSA (Cl) Anode, 0.6 mm Daramic [30 g/1 Zn, 7 g/1 HCL and 50 g/1 NaCl Catholyte] 38°C Temperature

Catholyte Purging

If, as appears certain, the loss of current efficiency is due to chlorine finding its way, in solution into the cathode chamber, then purging of the catholyte should improve the current efficiency by stripping the chlorine out of this electrolyte. This idea was tested by bubbling

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 841

nitrogen at 850 cm per minute through the catholyte reservoir. The anolyte was also purged as usual (at approximately 100 seem of air). The results are summarized in Figure 8 and comparing this figure with Figure 7, it is seen that the catholyte purging improves the current efficiency, particularly at low current density. Comparing Figure 8 with Figure 4, the purging of the catholyte and the use of the DSA for chlorine evolution bring substantial benefits. At about 1,500 A/m2 the electrical energy consumption is well below that of conventional electrowinning and might even be acceptable (depending on local power costs) at 3,000 A/m2 or higher.

Cell Voltage- Catholyte Purging

1000 2000 3000

Current Density, A/m2

4000

CE - Catholyte Purging

1000 2000 3000 4000

Current Density, A/m2

EEC - Catholyte Purging

1000 2000 3000

Current Density, A/m2

4000

Figure 8 - Cell Voltage, Current Efficiency and Energy Consumptions at Different Current Densities for DSA(Cl) Anode and Catholyte Purging (1461, 2190 and 2920 A/m2) Old 22 mm

Cell, Al Current Feeder, DSA (Cl) anode, 0.6 mm Daramic [30 g/1 Zn, 7g /l HC1 and 50 g/1 NaCl Catholyte], 38°C Temperature

842 LEAD-ZINC 2000

One experiment was carried out where the catholyte was purged but the anolyte was not. The current efficiency was identical to that of an experiment where both electrolytes were purged. It appears then that, from the viewpoint of current efficiency maximization, anolyte purging may be unnecessary when the catholyte is purged.

Thinner Daramic Diaphragm

The Daramic diaphragm used in most of this work was 0.6 mm thick. One experiment was carried out, at 4381 A/m2, with Daramic that was 0.28 mm thick. As expected, the voltage was significantly lower with the thinner diaphragm. Previous work (with acid sulfate electrolytes) has shown that the potential drop across the diaphragm is a major contributor to cell voltage. Somewhat surprisingly, the current efficiency was the same for the two diaphragms. The net result was pleasingly low energy consumption averaging less than 4 kWh/kg Zn at this very high current density of 4381 A/m2) when using the thinner diaphragm.

The Nature of the Zinc Deposits

The deposits produced in these experiments had a metallic appearance (a matte silvery color) to the naked eye and were adherent. Figure 9 shows two scanning electron micrographs of the surfaces of particles onto which zinc has been deposited in the SBE at a current density of 2920 A/m2 from an electrolyte containing 10 ppm TBAC1. The deposit is rough on a microscale but is similar to the deposits produced in conventional deposition of zinc from acid sulfate electrolytes. The hexagonal platelets of zinc seen in conventional electrowinning are apparent in the lower micrograph. The deposit appears to be free from gross porosity.

CONCLUSIONS

The conclusions from the work are as follows. A cell with a spouted bed of zinc particles can be used to electrowin zinc from zinc chloride electrolytes. In such applications, the spouted bed electrode has shown current efficiencies over 90% and electrical energy consumptions as low as 2.8 kWh/kg zinc deposited. The cell can be operated over a range of current densities (730 to 4,381 A/m2 have been the range of the present study with current efficiency and energy consumption optima in that range). Zinc deposits are metallic in appearance and are free of porosity under the SEM. Although some improvement in cell performance is achieved by using the TBAC1 additive, the improvement is small at high current densities. The cell consumes less energy when an anode (DSA) intended for chlorine evolution is used. The material of the diaphragm separating the anode and cathode sides of the cell (Daramic in the experiments to date) is important in determining both the current efficiency and the cell voltage; lower cell voltage and energy consumption were evident when a thinner diaphragm was used. Considerable gains in cell performance can be obtained by minimizing the transport of chlorine from the anode to the cathode side of the cell, for example, by purging chlorine from the catholyte.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 843

Figure 9 - SEM Micrographs of Zinc Particles, 2920 A/m2 with 10 ppm TBACl

844 LEAD-ZINC 2000

REFERENCES

1. E.D. Nogueira, L.A. Suarez-Infanzon and P. Cosmen, "The Zinclor Process: Simultaneous Production of Zinc and Chlorine," Zinc '83, Proc. 13th Annual Hydrometallurgical Meeting of CIM Metallurgical Society, Edmonton, Alberta, 21-26th

August 1983, Section 7, pp. 1-22.

2. D.J. MacKinnon, J.M. Brannen and R.M. Morrison, "Aspects of Zinc Electrowinning from Aqueous Chloride Electrolyte," Chloride Metallurgy, P.D. Parker, Ed., TMS-AIME, Warrendale, PA, U.S.A., 1982, 1-19.

3. B.K. Thomas and D.J. Fray, "Chloride Hydrometallurgy of Zinc Residues," Chloride Metallurgy, P.D. Parker, Ed., TMS-AIME, Warrendale, PA, U.S.A., 1982, 21-41.

4. B.K. Thomas and D.J. Fray, Trans. Inst. Min. Metall., C (1982) 105.

5. N.E. Tuffrey, V. Jiricny and J.W. Evans, "Fluidized Bed Electrowinning of Zinc from Chloride Electrolytes," Hydrometallurgy, Vol. 15, 1985, 33.

6. D. MacKinnon, J.M. Brannen and R.M. Morrison, "Zinc Electrowinning from Aqueous Chloride Electrolyte," J. Appl. Electrochem., Vol. 12, 1983, 39.

7. M. Sider, and D.L. Piron, "The Effects of Metallic Impurities and 2-butyne-l,4-diol on Zinc Electrowinning from Chloride Solutions," J. Appl. Electrochem., Vol. 18, 1988, 54.

8. T. Hirato, S. Inamine, S. Ukai and Y. Awakura, "Fundamental Studies of Zinc Electrowinning from Aqueous Chloride Electrolytes at High Current Densities," Zinc & Lead '95, T. Azakami, N. Masuko, J.E. Dutrizac and E. Ozberk, Eds., The Mining and Materials Processing Institute of Japan, Tokyo, Japan, 1995, 360-369.

9. D. Buttinelli, C. Lupi, E. Beltowska-Lehman and A. Riesenkampf, "Recovery of Zinc from Zinc Tankhouse Bleed Liquors by a SX/EW Process," Extraction Metallurgy '89, Institution of Mining and Metallurgy, London, 1989,1003-1016.

10. M. Sanchez Cruz, F. Alonso and J.M. Palacios, "Nucleation and Growth of Zinc Electrodeposits on a Polycrystalline Zinc Electrode in the Presence of Chloride Ions," L Appl. Electrochem., Vol. 23, 1993, 364.

11. H. Majima, E. Peters, Y. Awakura and K. Tsugui, "Fundamental Studies on Chlorine Behavior as Related to Zinc Electrowinning from Aqueous Chloride Electrolytes", Metall. Trans. B, Vol. 21, 1990, 251.

12. H. Majima, E. Peters, Y. Awakura, S. K. Park and M. Aoki, "Electrical Conductivity of Acidic Chloride Solutions," Metall. Trans. B, Vol. 19, 1988, 53.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 845

THE EFFECT OF MICROSTRUCTURE ON THE ELECTROCHEMICAL BEHAVIOR OF LEAD-SILVER ALLOY ANODES DURING ZINC

ELECTROWINNING

S. Jin and E. Ghali Department of Mining and Metallurgy, Laval University,

Ste-Foy, Quebec, Canada, G1K 7P4

G. St-Amant LTEE, Hydro-Quebec,

600 Avenue de la Montagne, C.P. 900, Shawinigan, Quebec, Canada, G9N 7N5

V. Cloutier G. Houlachi Noranda Inc., CEZinc Division Noranda Technology Center

860 Cadieux Boulevard, 240 Hymus Boulevard Valley field, Quebec, Canada, J6S 4W2 Pointe Claire, Quebec, Canada, H9R 1G5

ABSTRACT

During the casting of lead-silver anodes, variations in cooling rates result in the formation of an uneven microstructure. The slowly cooled areas (SCA) have a coarser microstructure than that of the remaining part (called "the general zone"). The effect of anode microstructure on the electrochemical behavior was investigated using galvanostatic, potentiodynamic and electrochemical impedance techniques. The results show that the resistances for the two steps of the oxygen evolution reaction on the slowly cooled areas of Pb-Ag anodes are higher than that on the general zone. The double layer capacitance for the first step of the oxygen evolution reaction (OER) and the pseudocapacitance of the second step of OER on SCA are smaller than those on the general zone. The potentiodynamic results show that the amount of lead dioxide formed on the SCA is greater than that of the general zone of the anode. The potential decay curves after 30 minutes of galvanostatic polarization at 45 mA/cm2 show that the corrosion potential of the SCA is 200 mV lower than that of the general zone. This can lead to perforation because of the formation of a galvanic cell with a small anode and a large cathode.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

846 LEAD-ZINC 2000

INTRODUCTION

A lead-silver alloy containing 0.5 - 1% Ag has been used as the insoluble anode in the electrolytic production of pure zinc for many years (1). During the casting of the Pb-Ag alloy anodes, a more slowly cooled region is formed in the upper center of the anode plate. Industrial practice shows that the slowly cooled area is much more susceptible to corrosion leading to the perforation of the anodes. Although this is a serious problem, there is a lack of information about the mechanism of the localized accelerated corrosion. In this work, we attempt to compare the electrochemical behavior of the two parts of the anode and to elucidate the corrosion mechanism using several dc and ac electrochemical methods.

EXPERIMENTAL

Materials and Sample Preparation

Samples of industrial anodes were used in this study. The dented anode plates were cut into small pieces of ca. 10 x 10 x 10 mm. A cross-section surface with the substrate and two side teeth was used as a working electrode. The anode cubes were welded to a plastic insulated copper wire and were cast in acrylic resin. The aluminum cathode sheet was cut to cuboids of ca. 20 x 15x5 mm. The electrical contacts were made with plastic insulated copper wires and the pieces were cast in acrylic resin. Before being introduced into the electrolytic cell, the working surfaces of the anodes and cathodes were ground with SiC abrasive paper to 600 grit, washed with de-ionized water and wiped immediately with tissue paper. The working surface area of each electrode was measured accurately. The electrolyte was made from A.C.S. grade chemicals and was prepared using double-distilled water.

Experimental Setup

The electrolytic cell was a one liter double walled cell containing 800 ml of electrolyte heated by the flow of thermostated water in the double wall (38 ± 0.5°C). The supporting electrolyte contained 180 g/L H2SO4 and 60 g/L Zn2+. The electrolyte was magnetically stirred during the experiments. The anode and cathode were mounted in a suitable Teflon holder and the distance between them was fixed at 2 centimeters. The reference electrode was the mercurous sulfate electrode (MSE): Hg, Hg2S04/sat.K2S04 (0.636 V vs. SHE). A saturated K2SO4 salt bridge was used to keep the reference electrode close to the cathode. The experimental setup for dc measurements was an EG&G Princeton Applied Research 273 potentiostat/galvanostat controlled by an IBM computer. The software SOFTCORR M342 was used for the acquisition of the polarization data. The potentiodynamic experiments were carried out using a potential sweep rate of 1 mV/s. The oxygen overpotentials at 45 mA/cm were determined by subtracting the ohmic drops (obtained by the ac impedance technique) and the calculated reversible potentials (1.264 V vs SHE) from the measured electrode potential values. In order to simulate industrial conditions, some experiments were also carried out using an acidic zinc sulfate solution with 5 g/L Mn2+ addition in the form of MnS04. The electrochemical impedance measurements were carried out using a Solartron 1255 HF Frequency Response Analyzer and a Solartron 1286 Electrochemical Interface over the frequency range from 10 kHz to 0.001 Hz. High amplitude sinusoidal signals were used to overcome the interference of the oxygen bubbles on the anode surfaces. An ac signal of 30 mA amplitude (rms) was used for the potentiostatic mode and one of 30 mA for the galvanostatic mode at 45 mA/cm2. The software Zplot was used for the acquisition and treatment of the

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 847

impedance data. The Nyquist plots of the impedance data were analyzed using a nonlinear least squares fitting program developed by Macdonald (2).

RESULTS AND DISCUSSION

Microstructures of the Samples

An electrode made of the SCA and an electrode made of the general zone were polished and examined by optical microscopy. Their microstructures are shown in Figure 1 and Figure 2, respectively. It can be seen from Figure 1 that the SCA has very evident dendntes and the interdendritic spaces are very large. The silver segregation is obvious and the segregated silver is concentrated in several limited interdendritic spaces. Figure 2 shows that there are no obvious dendrites in the general zone. Besides, though the silver segregation is observed in the general zone, the segregated silver is dispersed in the homogeneous eutectic mass. From the point of view of corrosion, the lead in the SCA is much more susceptible to electrochemical corrosion.

Figure 1 - Microstructure of the Electrode Made from the Slowly Cooled Area (SCA)

Figure 2 - Microstructure of the Electrode Made from the General Zone

848 LEAD-ZINC 2000

Overpotential for Oxygen Evolution

Because of the microstructural differences between the two zones in the lead-silver alloy anode, the two electrodes have different electrochemical activities for oxygen evolution. Figure 3 illustrates the overpotentials for the oxygen evolution reaction on the SCA and on the general zone in the zinc electrolyte at 45 mA/cm2. It is observed that after 15 hours of galvanostatic polarization, the potentials of the two electrodes become stable and that the SCA has an overpotential that is 30 mV higher than the overpotential of the general zone. This means that the electrochemical activity for oxygen evolution on the SCA is lower than that on the general zone. In other words, the SCA exhibits a higher resistance to the electrolytic current than the general zone. This observation was confirmed by electrochemical impedance studies.

LU g8 0 _ General zone

>-g «50 -

640 - -

620 - —

600 I 1 1

0 5 10 15 ELECTROLYSIS TIME, h

Figure 3 - Overpotential for Oxygen Evolution on the SCA and the General Zone of the Lead-Silver Alloy Anode in Zinc Electrolyte at 38°C and 45 mA/cm2

Charge Transfer Resistance for the Anodic Process

Figure 4 shows the Nyquist plots of the electrochemical impedance for the oxygen evolution reaction on the SCA and on the general zone. The electrochemical impedance measurements were carried out using the potentiostatic mode; i.e., a constant potential was imposed on the anodes and a small alternative signal was added to the constant potential. The impedance of the alternative signal was measured as a function of frequency. It can be seen from Figure 4 that the Nyquist plot for the SCA has a bigger semicircle than that for the general zone. According to the oxygen evolution reaction mechanism (3), this semicircle should consist of two smaller semicircles corresponding to the two electrochemical steps of the oxygen evolution reaction:

H2Oad -► OHad + H+ + e (1)

OHad -> Oad + H+ + e (2)

2 0 a d H > 0 2 (3)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 849

The third step is not an electrochemical process but a chemical one. The inductance in the Nyquist plots is probably related to adsorbed oxygen. An equivalent circuit shown in Figure 5 is suggested to simulate the electrochemical process of the oxygen evolution reaction. In Figure 5, Ri and R2 correspond to the charge transfer resistance of the first and second steps of the oxygen evolution reaction, respectively, and CPEi and CPE2 are constant phase elements related to the double layer capacitance (Cdi) for the first step and the pseudocapacitance (Cp) for the second step, respectively. Rs is the ohmic resistance. It consists of two parts: the electrode surface oxide layer resistance and the electrolyte resistance between the electrode surface and the salt bridge tip. Since the second part does not change under certain conditions, the change of Rs should be related to the thickness and properties of the surface layer. The calculated resistance and capacitance values are given in Table 1.

-3

-2

E υ

I -1

o Psl

-

" -

-

SCA

0 ° ° ° 0 0 0

0 -

° ^" ~~~\ °„ r- S ^ \ °

Sr \ O a \ a S General zone \ %

i 1 δ f 1 fit

) £ ό^

1 , 1 . 1 .

Figure 4 - Nyquist Plots of the Electrochemical Impedance on the SCA and the General Zone of the Lead-Silver Alloy Anode in Zinc Electrolyte at 38°. Polarization Potential was 1.345 V

vsMSE

CPE,

Figure 5 - Equivalent Circuit Proposed For Fitting the Experimental Data of the Electrochemical Impedance Measurements of the Oxygen Evolution Reaction on Lead-Silver

Alloy Anodes

850 LEAD-ZINC 2000

Table 1 - Comparison of the Parameters of the Equivalent Circuit of the Oxygen Evolution Reaction on the Two Electrodes Measured by Electrochemical Impedance. Potentiostatic

Mode: E = 1.345 V vs MSE Superposed by an Alternating Signal of 5 mV Amplitude

Electrode

SCA General zone

Rs (Ω cm2)

0.3129 0.2948

Ri (Ω cm2)

2.5747 1.9403

R2

(Ω cm2)

0.9616 0.6412

C<ji

(mF/cm2)

41.93 48.00

Cp

(mF/cm2)

99.66 264.60

It can be seen from Table 1 that the surface layer resistance of the SCA electrode is higher than that of the general zone by 6%; the charge transfer resistance for the first step (R|) is ca. 2.6 Ω cm2, whereas the Ri value of the general zone is only 1.9 Ω cm2. As for R2, the SCA is 50% higher than the general zone. On the contrary, the double layer capacitance of the general zone electrode is larger than that on the SCA. The big difference is observed in the pseudocapacitances which are related to the efficiency of the first step; i.e., the production rate of OHad. Obviously, the general zone has a higher OHad production rate and the second step on it is also easier than that on the SCA. So, the impedance for oxygen evolution on the SCA is much larger than the impedance on the general zone, and the general zone exhibits a better electrochemical activity.

It should be noted that in Figure 4, each Nyquist plot has an inductance at low frequencies. This inductance is related to the number of Oaa on the anode surface. The calculations show that the inductance on the SCA anode is 308.9 H/cm2, whereas that on the general zone anode is only 190.5 H/cm2. This means that the adsorbed oxygen atoms on the SCA anode have more difficulty to combine with each other to form oxygen molecules than those on the general zone anode. It is known that the Oad is a strong oxidant; it can diffuse across the oxide layer and attack the metal substrate to form more oxide. This accelerates the corrosion rate of the anode.

According to the results of the electrochemical impedance measurements, it can be expected that if a constant potential is imposed on the anodes and the produced current is recorded as a function of time, the chronoamperogram of the SCA should be below that for the general zone. Figure 6 shows the potentiostatic polarization curves. It is true that the current density on the general zone is ca. 10 mA/cm2 higher than that on SCA when a potential of 1.345 V vs MSE (1.981 V vs SHE) is used.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 851

tu <r

3 »

"T Γ

Γ""™-™

I I Γ

General zone

'■" ' " ' '" U_

SCA

J I I L 6 8 10 12 14

TIME, h

Figure 6 - Potentiostatic Polarization Curve of the SCA and the General Zone of Lead-Silver Alloy Anodes in Zinc Electrolyte at 38°C and 1.345 V vs MSE

Reduction Behavior of the Lead Dioxide on Lead-silver Alloy Anodes

A potential sweep in the cathodic direction was conducted after galvanostatic polarization at 45 mA/cm2 for 16 hours. The reduction curves are shown in Figure 7. The continued, broken and dotted lines represent the general zone, the SCA and pure lead anodes, respectively. It can be seen from Figure 7 that the shape of the reduction curve of the SCA anode is similar to that for pure lead anodes after 16 hours of oxygen evolution reaction at 45 mA/cm2. Their cathodic reduction peaks have the same potential, whereas the general zone anode has a reduction peak in a more positive potential range. The pure lead anode has the highest cathodic reduction peak; the SCA anode has a medium one, whereas the cathodic peak of the general zone anode is the lowest. This means that the lead dioxide layers formed on the SCA anode and the lead anode have a similar structure and properties. This is probably one of the causes of the fact that the slowly cooled areas are less corrosion resistant.

852 LEAD-ZINC 2000

Ui a. a. 3 O

Pure lead SCA General zone

1580 1590 1600

POTENTIAL, mV vs. SHE

Figure 7 - Potentiodynamic Polarization Curves of the Pure Lead Anode (Dotted Line), the SCA Anode (Broken Line) and the General Zone Anode (Continued Line) After 16 Hours of

Galvanostatic Polarization at 45 mA/cm2 in Zinc Electrolyte at 38°C with Magnetic Stirring and Nitrogen Bubbling. Potential Sweep Rate: 1 mV/s

Anode Potential Decay

Figure 8 illustrates the anode potential decay curves after 30 minutes of oxygen evolution on the SCA and the general zone anodes at 45 mA/cm2 in zinc electrolyte. It is observed that the corrosion potential of the general zone anode is 200 mV higher than that on the SCA anode after 8 hours of potential decay. In fact, after one hour of potential decay, the two anodes have a potential difference of about 200 mV. Since the general zone area is much larger than the SCA, and because the potential of the former is much higher than that of the latter, a serious corrosion cell is formed when the current of the electrolytic cell is cut. That is, a corrosion cell with a large cathode (general zone) and a small anode (SCA) is created. With all types of corrosion cells, this one is the most dangerous. The anode in this type of corrosion cell will suffer severe localized corrosion and the anode area will be inevitably perforated. This situation is unavoidably encountered in the zinc industry during cell maintenance, power cuts and electrode treatment.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 853

ID T c/> » > ■ >

j

< 1-7 III 1-o 0.

2.0

1.8

1B

14

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Figure 8 - Electrode Potential Decay Curves of the SCA and the General Zone Anodes after 30 Minutes of Oxygen Evolution Reaction at 45 mA/cm2 in Zinc Electrolyte at 38°C and with

Stirring

CONCLUSIONS

The following conclusions can be drawn from this study. The microstructure of the slowly cooled areas of the lead-silver anode is less homogeneous than that of the general zone. This results in a decrease of the electrochemical activity for oxygen evolution. When the current is cut, the corrosion potential on the SCA is much more negative than that on the general zone. Since the surface area of the general zone is much larger than that of the slowly cooled areas, the most dangerous corrosion cell is formed.

ACKNOWLEDGEMENTS

The authors are grateful to Le Ministre des Ressources Naturelles du Quebec and Hydro-Quebec for their financial support and interest. The authors acknowledge General Smelting of Canada Ltd. for their assistance in preparing the electrode materials, and Noranda Inc. CEZinc for their assistance and valuable contributions during the evaluation of the Pb alloys.

LEAD-ZINC 2000

REFERENCES

V.A. Ettel and B.V. Tilak, "Electrolytic Refining and Winning of Metals", in Comprehensive Treatise of Electrochemistry, B.E. Conway, J.O'M. Bockris, E. Yeager, S.U.M. Khan and R.E. White, Eds., Prenum Press, New York, NY, U.S.A., Vol. 2, 1983, 327-520.

J.R. Macdonald, Complex Nonlinear Least Squares Immittance Fitting Program LEVM, Version 1989, Department of Physics and Astronomy, University of North Carolina, Chapel Hill, NC, U.S.A., 27599-3255.

Q. Zha et al., An Introduction to the Kinetics of Electrode Processes, Kexue Chubanshe, Beijing, China, 1987,369.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 855

EFFECT OF POLYMER ADDITIVES ON ZINC ELECTROWINNING

T. Ohgai, H. Fukushima and N. Baba Department of Materials Process Engineering, Kyushu University

6-10-1 Hakozaki, Higashi-ku Fukuoka 812-8581, Japan

T. Akiyama Department of Industrial Chemistry, Kyushu Sangyo University

2-3-1 Matsukadai, Higashi-ku Fukuoka 813-8503, Japan

ABSTRACT

The electrodeposition behavior of Zn and the morphology of the deposited Zn were studied in electrowinning solutions containing polymer additives such as gelatin and polyethylene glycol (PEG). The cathode potential was shifted in a less noble direction and the polarization resistance for Zn deposition increased with increasing concentrations and molecular weights of both gelatin and PEG added to the electrolytic solutions. However, when the molecular weights of these additives exceeded about lxlO4, the cathode was depolarized and the polarization resistance for Zn deposition decreased with increasing molecular weight. The crystal grain size and the (002) orientation of the deposited Zn became smaller when the concentration and molecular weight of the additives increased, and the cathode was polarized.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

856 LEAD-ZINC 2000

INTRODUCTION

Certain polymers such as glue, bean cake and waste pulp fluid have been generally added to conventional zinc electrolytes to maintain the high purity as well as the quality of deposits in the hydrometallurgical production of nonferrous metals such as Zn, Cu and Pb (1). However, it has been empirically demonstrated that these polymer additives are not only degraded during electrolysis to diminish their effectiveness but also exert a harmful influence on the purity of the deposited metal and on the conductivity of electrolyte. Many studies have been made so far on the effect of additives on the electrodeposition of metals, but few mechanisms accounting completely for their role have been proposed (2,3,4).

In this study, the effects of gelatin and polyethylene glycol (PEG) on Zn deposition potential, and on the morphology and crystal orientation of the deposits were examined to elucidate the behavior of polymer additives in Zn electrowinning in relation to the structure of the polymers.

EXPERIMENTAL PROCEDURE

Electrolytic solution was prepared by dissolving ZnO (0.765 mol/L) in distilled and ion exchanged water. The concentration of free sulfuric acid was 1.53 mol/L. Gelatin and PEG were added to the electrolytic solution. Aluminum and platinum were used as the cathode and anode, respectively. Electrolysis was conducted in unagitated solutions under galvanostatic conditions at 40°C.

The adsorption behavior of gelatin and PEG on the cathode was evaluated by the measurement of the cathode potential and AC impedance during electrolysis. A saturated Ag/AgCl electrode was used as a reference electrode to monitor the cathode potential during electrolysis. To obtain Cole-Cole plots, the frequency dependence of the AC impedance and phase difference were measured by a frequency response analyzer at 500 A/m2 (±3mA sine wave, 10"2 ~104 Hz, 7 points / decade).

The initial surface appearance of the zinc deposited at 500 A/m2 for 90 kC/m2 (3 minutes) was observed by SEM. The structure of the deposited Zn was analyzed by X-ray diffraction analysis (XRD) and the crystal orientation index was calculated by Wilson's equation (5).

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 857

RESULTS AND DISCUSSION

Structural Formulas and Adsorption Capacity of Polymer Additives

Gelatin is a straight chain structure formed by peptide linkage of amino acids such as glycine (Gly), proline (Pro) and alanine (Ala). An example of a structural formula of gelatin is schematically shown in Figure 1(a). Oxygen atoms in the polymer act as the radicals for adsorption by utilizing their lone pairs of electrons. According to the literature, the adsorb behavior is quite different between high and low molecular weight polymers (6) because higher molecular weight polymers have many radicals and adsorb tenaciously (7). Therefore, with an increase in the molecular weight, the number of radicals for adsorption in a given straight-chain molecule, and hence the adhesion capacity of the polymer, also seems to increase exponentially. On the contrary, it can be predicted that the adsorption capacity of a polymer rapidly decreases when the molecular weight is decreased by their decomposition.

It is known that hydrolysis and oxidation cause the degradation of gelatin. Hydrolysis proceeds by severing peptide linkages (C-N). Therefore, gelatin added to an electrolytic solution may be hydrolyzed to decrease its molecular weight, depending either on the acidity of the electrolytic solution or the pH change caused by hydrogen and oxygen evolution in the vicinity of the electrodes. On the other hand, the oxidization of gelatin is reported to proceed by the formation of weak links in the main chain because of the attack of oxygen radicals. Therefore, it seems that gelatin added to an electrolytic solution is easily oxidized and degraded by dissolved oxygen generated at insoluble anodes during long-term electrolysis.

On the basis of the mechanism described above, the time-dependence of the effectiveness of polymer additives in practical electrowinning can be discussed by the decomposition, hydrolysis and oxidation into polymers of lower molecular weight.

On the other hand, PEG, which is a polymerized product of ethyleneglycol (HOCH2CH2OH), has a straight chain structure similar to gelatin and has many oxygen radicals for adsorption at constant intervals in each molecule (Figure 1(b)). In contrast to gelatin, PEG is said to be stable and difficult to degrade (8). In addition, PEG with various molecular weights is readily available. Therefore, PEG is used as a reference additive for gelatin.

Effects of the Concentration and Molecular Weight of Gelatin and PEG on Zinc Deposition

The effects of gelatin (mean molecular weight: 2x104) and PEG (mean molecular weight: 1.54xl03) concentrations on the cathode potential and on the polarization resistance for Zn deposition are shown in Figures 2(a) and (b). Coverage of the additives on the Zn deposition sites seems to increase with an increase in the concentration of gelatin and PEG because the deposition potential of Zn polarizes in a less noble direction and the polarization resistance for Zn deposition increases.

858 LEAD-ZINC 2000

Figure 1 - Structural Formulas of Gelatin (a) and PEG (b)

-0.80

υ -0.95

X Free O PEG(1.54x103) • Gelatin(2x105)

(b)

-L J L Free 0.001 0.01 0.1 1 Free 0.001 0.01

Gelatin & PEG Concentration (g/L)

0.70

a 0.65 a

0.60 o

0.55 K

0.50 10

Figure 2 - Effects of Gelatin and PEG Concentration on the Cathode Potential (a) and on the Polarization Resistance (b) at 500 A/m2 in Zn Electrowinning Solution

As predicted previously, the effectiveness of gelatin gradually disappears because the molecular weight of gelatin decreases with an increase in electrolysis duration. The effects of the molecular weight of gelatin (0.1 g/L) and PEG (0.1 g/L) on the cathode potential and on the polarization resistance for Zn deposition at 500 A/m2 are shown in Figure 3(a) and (b). As the molecular weights of gelatin and PEG increase, the overpotential and the polarization resistance for Zn deposition first increased, but then decreased when the molecular weight exceeded lxlO4.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 859

UJ" -0.80 I

> > ~ -0.85 n c o 'S °- -0.90 a Ό o £ «■»

n U -0.95

t(a)

~ ·1 -

—

"~ 1

1

• N .

1

1

X Free O PEG • Gelatin

V o^cf**

\ί^ν· ο β ^ · 1 I

(b)

1 1 %-i<—1_

S - N ^

Cr /

°f 1 1 J · /

j ^ ^ 1

w ^ν ~ V

\

• \ >

1 1

m \ -\

»

Free 102 103 104 105 Free 102 103

Molecular Weight of PEG & Gelatin

10« 105

0.70

a 0.65 S

- 0.60 S

- 0.55 H

o a.

0.50 10β

Figure 3 - Effects of Molecular Weight of Gelatin and PEG on the Cathode Potential (a) and or the Polarization Resistance (b) at 500 A/m2 in Zn Electrowinning Solution

In Figure 3, the concentrations of gelatin and PEG were kept constant at 0.1 g/L Therefore, the electrolytic solution contains a smaller number of longer chain as the molecula weight increases. When the molecular weight of the polymers is less than 1x10 , almost all thi oxygen radicals are utilized for adsorption to inhibit Zn deposition effectively. As a result, thi Zn deposition potential shifted in a less noble direction and the polarization resistance for Zi deposition increased. In contrast, when the molecular weight of the polymers exceeded lxlO4

the number of effective radicals decreased because of the entanglement of the polymei Consequently, the Zn deposition potential depolarized and the polarization resistance decreased

Morphology of the Deposited Zinc

It is well known that the electrodeposition of Zn proceeds with a low overpotential am that the electrodeposits consist of large platelets packed together. On the other hand, th addition of adsorbent or complexing agents to the electrolytic solution increases the depositioi overpotential, resulting in finer crystal platelets. Figure 4 shows the SEM images of th initially deposited Zn from the solution containing gelatin (mean molecular wei ght2xl05) ani PEG (0.1 g/L). The electrodeposits obtained from the additive-free solution are composed o hexagonal platelets approximately 3x10"6 m in diameter, which are characteristically observo in hexagonal crystals (Figure 4(a)). In the solutions containing gelatin, on the other hand, th grain size of the deposited Zn becomes smaller as the additive concentration increases (Figur 4(b), (c)). Furthermore, in the solutions containing PEG, the grain size of the deposited Zn als· becomes smaller. This effect of PEG on the morphology of the deposits was most prominent e the mean molecular weight of 4000, and the most significant polarization was simultaneousl observed. Thus, the polymer additives in Zn electrowinning cause smooth deposits b polarizing the deposition potential of Zn.

860 LEAD-ZINC 2000

Figure 4 - Effects of Gelatin and PEG on the Initial Surface Appearance of Zn Deposited from Zinc Electrowinning Solutions at 500 A/m2 for 3 Minutes

Also, the X-ray diffraction patterns of the Zn deposited from solutions containing gelatin and PEG at 500 A/m2 for 30 minutes were examined. The orientation was determined by Wilson's equation (5). The results are shown in Figure 5(a). The electrodeposit from the solution containing no additives showed a predominant orientation of (002) which is the most close-packed plane of the hexagonal close packed structure. This result corresponds to the SEM images shown in Figure 4(a). On the other hand, the (002) orientation of the deposited Zn from the solution containing PEG approached unity as the molecular weight increased, and the morphology of the deposits became of the unorientated dispersion type. Thus, PEG adsorbs on the active sites for Zn deposition and inhibits the predominant growth of the (002) plane.

In the solutions containing gelatin, the morphology of the electrodeposits was also of the unorientated dispersion type. Consequently, the (002) orientation index of the deposited Zn depended on the overpotential for Zn deposition as shown in Fig.5(b). This result can be explained well by the Pangarov's two-dimensional nucleation theory (9) that the crystal orientation of the deposited metals is determined by the deposition overpotential of the metals.

Suppression of the Harmful Influence of Impurities by Gelatin and PEG

The authors have already reported that impurities such as As, Sb and Ge adsorb on the cathode and catalyze hydrogen evolution because of the shift of the cathode potential in a more noble direction (10). Therefore, the cathode current efficiency is significantly lowered even by a trace amount of these impurities. However, it can be expected that certain additives, which adsorb on the cathode in preference to the impurities, suppress the harmful influence of the impurities by maintaining the cathode at the desired potential. Additives such as glue and gelatin are used in conventional Zn electrowinning as a cathode polarizer. Therefore, the suppression effect of gelatin and PEG on the harmful influence of the impurities was investigated.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 861

o o

0

X Free 0 PEG (0.1 g/L) • Gelatin (0.1 g/L)

' " I i n i i i i i l I I M I I I I I I Ml

PEG' 4 x 1 0 J

I I I I 1 I I I I 1 I I I I I

Free 102 103 104 105 -0.80 -0.85 -0.90 -0.95 Molecular Weight of PEG & Gelatin Cathode Potential (Vvs. NHE)

Figure 5 - Effects of Molecular Weight of PEG and Gelatin (a) and Cathode Potential (b) on the (002) Crystal Orientation of Zn Deposited from Zinc Electrowinning Solutions at 500 A/m2

for 30 Minutes

Figures 6(a) and (b) show the effect of additives on the current efficiency for Zn deposition and the cathode potential during electrolysis at 500 A/m2 for 30 minutes in the Sb-containing solutions. The mean molecular weights of gelatin and PEG were 2xl04 and 1.54χ103, respectively. When the Sb concentration increased, the current efficiency drastically decreased from 90 to 40 %. This decrease in the current efficiency is caused by the catalytically increased rate of hydrogen evolution and, therefore, the cathode is significantly depolarized. Thus Sb acts as a depolarizer in Zn electrowinning. On the other hand, gelatin and PEG act as a polarizer to cancel the depolarization by Sb, as shown in Figure 6(b) and the current efficiency for Zn deposition was restored to about 80 %.

Figure 6 - Effect of the Sb Concentration on the Current Efficiency for Zn Deposition (a) and on the Cathode Potential (b) in Gelatin or PEG-containing Zinc Electrowinning Solution

862 LEAD-ZINC 2000

CONCLUSIONS

The effects of polymer additives such as gelatin and PEG on Zn electrowinning have been investigated. The results obtained are as follows. Gelatin and PEG act as a polarizer to shift the deposition potential of Zn in a less noble direction. The degree of polarization depends on the molecular weight of the additives and there exists an the optimum molecular weight exhibiting the prominent polarization effect. The electrochemical impedance analysis revealed that the polarization resistance for Zn deposition increased when the concentration and molecular weight of the additives increased. The morphology of the initially deposited Zn obtained from the additive-free solution was composed of hexagonal platelets with a size of approximately 3 xl0~6m in diameter, and orientated with the (002) plane predominant. In the solution containing gelatin and PEG, the grain size of the deposited Zn was smaller and the (002) orientation index decreased to unity. This means that the morphology of the deposits became of the unorientated dispersion type. Additives in the electrolytic solutions overcame the depolarization effect caused by hydrogen evolution by Sb to maintain the cathode at the desired potential. An addition of gelatin to the solution containing Sb effectively prevented the current efficiency for Zn deposition from decreasing.

REFERENCES

1. C. L. Mantell, Electrochemical Engineering, McGraw-Hill Book Company, Inc., New York, NY, U.S.A., 1960, 142-167, 185-192, 210-224.

2. R. C. Kerby, H. E. Jackson, T. J. O'Keefe and Yar-Ming Wang, "Evaluation of Organic Additives for Use in Zinc Electrowinning," Metallurgical Trans. B, Vol. 8B, 1977, 661-668.

3. M. Karavasteva and St. Karaivanov, "Electrowinning of Zinc at High Current Density in the Presence of Some Surfactants," J. Appl. Electrochem.. Vol. 23, 1993, 763-765.

4. X. Tang, Pu Yu, T. J. O'Keefe and G. Houlachi, "Characterization of Antimony-Gelatin Additives in Zinc Sulphate Electrolytes Using Impedance Analysis," Aqueous Electrotechnologies : Progress in Theory and Practice. D. B. Dreisinger, Ed., The Minerals, Metals and Materials Society, Warrendale, PA, U.S.A., 1997, 115-125.

5. K. S. Willson and J. A. Rogers, "Orientation, Crystal Structure and Appearance of Nickel Deposits from a Watts Bath Containing Coumarin," Tech. Proc. American Electroplaters Society, Vol. 51, 1964, 92-95.

6. A. Takahashi and M. Kawaguchi, J. Society of Rubber Industry, Japan, Vol. 60, 1987, 231-239.

7. Surface Modification Technology, Sangyo-Gijyutsu Service Center, Tokyo, Japan, 1993,604-607.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 863

8. M. Dobashi, S. Yoshihara, T. Shirakashi and T. Yokota, "Influence of Additives in Electrolytic Solution on Mass Change of Zinc Electrodeposited Films - Investigation by Quartz Crystal Microbalance Measurement," The Journal of the Surface Finishing Society of Japan, Vol. 48, 1997, 330-336.

9. N. A. Pangarov, Electrochem. Acta, Vol. 7, 1962, 139 ; Vol. 9, 1964, 721.

10. T. Akiyama, T. Ohgai and H. Fukushima, "Effect of Impurities in Electrowinning Solutions on the Critical Current Density for Zinc Deposition," Zinc & Lead '95, T. Azakami, N. Masuko, J. E. Dutrizac and E. Ozberk, Eds., Mining and Materials Processing Institute of Japan, Tokyo, Japan, 1995, 343-351.

Chapter 11

Environmental Aspects of Lead and Zinc Production

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 867

EFFLUENT TREATMENT AT THE PASMINCO CLARKSVILLE ZINC PLANT

S. Subhawong Pasminco Clarksville Zinc Plant

Clarksville, Tennessee, U.S.A., 37041

ABSTRACT

Following the startup of the zinc plant in Clarksville, Tennessee, in 1978, the original effluent treatment plant did not perform as designed. The sludge, produced in a conventional lime neutralization treatment, settled poorly and began to fill the permitted lined impoundment within two years, well short of the original design capacity. Another sludge pond had to be constructed. This led to a new process modification of the original design. Although the design is still based on the conventional lime precipitation of heavy metals, the current effluent treatment plant has evolved into a hydrometallurgical circuit, capable of producing good quality water without generating sludge. Undesirable elements are bled from the zinc circuit and are rejected through the effluent treatment operations. Wallboard grade gypsum is produced, instead of sludge, and all the gypsum is sold to cement and wallboard plants. This operation also recovers and returns more than 1,500 tonnes of zinc annually to the 105,000 tonne/y zinc plant.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

868 LEAD-ZINC 2000

INTRODUCTION

The Clarksville zinc project was a joint venture between the New Jersey Zinc Company, a wholly owned subsidiary of Gulf + Western Industries Inc., and Union Zinc Company, a wholly owned subsidiary of Union Miniere SA. The plant, commissioned as a new greenfields zinc manufacturing facility in November 1978, was designed to recover zinc metal in a conventional electrowinning circuit after the roasting, leaching and purification stages (1). Though originally rated to produce 81,650 tonnes (90,000 ST) of SHG zinc/year, the plant gradually increased its capacity to about 105,000 tonne/year. The company went through a series of ownership changes from a wholly owned subsidiary of Union Zinc in 1984, to Savage Resources in 1994 and most recently to Pasminco in 1999.

The original effluent treatment plant did not performed as designed and was shut down only three months after the startup. A simplified treatment system was put into operation on a temporary basis. Although the system produced good quality water initially, poor settling characteristics of the neutralized sludge required additional impoundment volume. A new process design of the effluent treatment system had to be developed. The concept of an effluent treatment plant, based on a conventional lime neutralization process, with zero sludge discharge was investigated. This paper outlines the initial problems of the original design, which led to three separate process development stages, and consequently reviews the implementation of the current effluent treatment plant.

ORIGINAL DESIGN

In the original design, shown in Figure 1, the plant effluent was categorized into two separate systems. High acidity streams, which generated high volume of sludge reported to the non-recyclable system. Other low acidity streams, which generated low volume sludge, reported to the recyclable system, where the treated water could be recycled back to the process water storage tank in the zinc plant. Descriptions of these two systems are summarized below:

The Non-Recyclable System

This system consisted of low volume, high acidity streams such as the magnesium bleed electrolyte from the cellhouse and the Venturi scrubber water from the roaster gas cleaning operation. Typical compositions of these two streams are shown in Table I.

Table I - Flow Rates and Composition of Input Streams to Non-Recyclable Unit

FlowirrVh Zn, g/1 H2S04, g/1 Mg, g/1 Cl, g/1 Mg Bleed Electrolyte 4 25 240 12 0.3

Venturi Scrubber Water 10 1 20 < 0.2 2.0

The mixture was neutralized with lime to a pH of 10 to precipitate the heavy metals. After adding the polymer, the slurry was settled in a 26 meter thickener. The thickener overflow was adjusted to a pH of between 7 and 8 with sulfuric acid and was discharged to the Cumberland River via a permitted National Pollutant Discharge Elimination System (NPDES)

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 869

outfall. The thickener underflow was pumped to the permitted non-recyclable pond, where additional settling would increase the solids content in the slurry to about 25 %. Based on this design, the lined non-recyclable pond of 34,800 cubic meter capacity was projected to have a life of at least 10 years. The supernatant from the pond was discharged with the thickener overflow, via the acidification tank, to the NPDES outfall. The total flow rate through this system was designed to be 400 cubic meters/day.

The Recyclable System

All other plant effluents, consisting of high volume but lightly contaminated flows, such as boiler blow-down, vacuum pump seal waters and anode cleaning water, were collected in another permitted pond. This lined pond in the recyclable system had a capacity of 37,400 cubic meters. Water from the recyclable pond was pumped into a separate lime treatment circuit, where a final pH of 10 was maintained. Since the water was lightly contaminated, the neutralized sludge reported to a 7 meter reactor clarifier. The underflow of the reactor clarifier discharged into the 26 meter thickener in the non-recyclable system. The overflow pH of the reactor clarifier was adjusted to between 7 and 8 with sulfuric acid, before passing through a sand filter. This water from the sand filter was either discharged to the river, via the permitted NPDES outfall or recycled back to the process water storage tank. Design flow through the recyclable system was 670 cubic meters/day.

START-UP PROBLEMS AND SHORT-TERM SOLUTIONS

Within days of the start-up, the non-recyclable system encountered severe corrosion problems, because of the high halide content in the Venturi scrubber water. The lime neutralization circuit did not have sufficient capacity to maintain the target pH of 10. The treatment rate had to be decreased in order to achieve the designed water quality. The neutralized sludge settled poorly in both the non-recyclable and the recyclable systems. In addition, both systems experienced serious gypsum scaling problems. As a result, an average discharge through the NPDES permit was limited to only about 70 cubic meters/day during the first few months of operation. After unsuccessful attempts to operate the effluent treatment plant at design flow rates and within target parameters, it was concluded that the original effluent treatment plant could not keep up with the volume of the plant effluent that had to be treated. An alternate short-term solution had to be implemented quickly. The original treatment plant was then shut down in March 1979, after only three months of operation.

Since the non-recyclable and the recyclable ponds were constructed adjacent to the third pond, which was permitted for plant residue collection, a simplified treatment system was put into operation. Under this alternate short-term plan, as shown in Figure 2, all the plant effluents were collectively discharged into the lined residue pond of 144,400 cubic meter capacity. The supernatant from this pond was pumped into two 40 cubic meter reactors in series, where the final pH of 10 was maintained with lime slurry. The neutralized slurry then discharged into the non-recyclable pond, which was utilized as a primary settling pond. Liquid from the non-recyclable pond overflowed into the recyclable pond for an additional polishing step before being discharged to the river, via the permitted NPDES outfall. This arrangement increased the effluent treatment rate from 70 cubic meters/day to about 500 cubic meters/day.

870 LEAD-ZINC 2000

Recyclable System Non-Recyclable System

Lime Bleed Electrolyte and Lime

\r v

pH 10

Reactor Clarifier

IT Acid

pH7-l

Sand Filter

Recycle to Zinc Plant or Discharge to River

Recyclable Pond Non-Recyclable Pond

Other Plant Effluents

Figure 1 - Original Design of the Effluent Treatment Plant at the Clarksville Zinc Plant

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 871

All Plant Effluents

Lime

Main Residue Pond

oo Recyclable

Pond

pH 10

Non-Recyclable

Pond

River

Figure 2 - Simplified Effluent Treatment Process at the Clarksville Zinc Plant (1979)

Although the modified effluent treatment plant had sufficient capacity and allowed the treated water to be discharged within the NPDES parameters, the neutralized sludge still settled poorly in the non-recyclable pond. The solids content in the sludge was found to be only about 5 % as compared to the original design of 25 %. As a result, the non-recyclable pond began to fill rapidly and solids overflowed into the recyclable, or polishing pond. Soon the water quality began to deteriorate. The sludge had to be dredged from both the non-recyclable and the recyclable ponds into the main residue pond within two years of operation. This in turn reduced the life of the residue pond. A new sludge holding pond had to be constructed and was put in operation in October 1980. As more sludge was dredged into the new pond, it became apparent that either more sludge holding ponds would have to be built or a more efficient process would have to be developed.

PROCESS DEVELOPMENT

A new process development was initiated to reduce or to eliminate the sludge production. Research efforts are categorized into three following areas.

Gypsum Recovery Circuit

In an electrolytic zinc plant, magnesium contained in the zinc concentrate becomes soluble and would accumulate in the electrolyte. Different techniques have been employed, such as the selective zinc precipitation process and the releaching of concentrates, to reject magnesium from the zinc circuit (2,3). The original design for magnesium control at the Pasminco Clarksville Zinc Plant was to strip the zinc content in the electrolyte from 55 g/1 to about 25 g/1 and then to bleed the stripped electrolyte, approximately 100 m3/day, to the

872 LEAD-ZINC 2000

effluent treatment plant. When the pH of this stripped electrolyte was raised to 10 in the lime treatment process, a large quantity of sludge was generated because of the high content of metals and free acid, shown in Table I.

A bench scale feasibility test was conducted to neutralize the acid with limestone at pH 4 and to remove the gypsum prior to the lime treatment step at pH 10. The objective was to reduce the sludge volume by removing the gypsum before it became contaminated with metal hydroxide precipitated in the lime treatment step. Experimental data from both bench and plant scale tests showed that the gypsum, recovered as a filter cake, was of commercial grade and was found to pass the Toxicity Characteristic Leaching Procedure (TCLP) tests. The neutralized filtrate was then combined with other plant effluents and the mixture reported to the hydroxide precipitation step. By installing this gypsum recovery circuit, a 50 % reduction in total sludge quantity was achieved and a marketing program for the gypsum filter cake was initiated.

High Density Sludge (HDS) Process

In a parallel effort to reduce the sludge volume, an investigation was made to increase the solids content in the sludge. After reviewing several existing commercial-scale operations, a decision was made to select the High-Density Sludge (HDS) process (4,5). A pilot scale study was conducted at Cominco's Technical Research Centre, Trail, British Columbia, in 1981. Under the assumption that only the free acid in the magnesium bleed electrolyte would be neutralized, synthetic effluent solution of the following composition was prepared for the pilot plant.

Table II - Synthetic Plant Effluent Feed to the Pilot Plant

Zn H2S04 Mg Mn Cd Fe++ Fe Cl mg/1 2000 2500 1200 80 30 6 272 234

Rather than a straight through neutralization, the process employed a seeding technique by mixing the milk of lime with the recycled sludge before it entered the neutralization reactor where it was mixed with the synthetic effluent, as shown in Figure 3.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 873

Lime Slurry

Synthetic Effluent

v v

Pre-Mix Tank

Flocculant

Neutralization pH 10

Treated Effluent

Recycle

Purge

Figure 3 - High Density Sludge (HDS) Process Pilot Plant

Milk of lime was added to the pre-mix tank to maintain a pH of 10 in the reactor. The slurry was then flocculated and settled. A portion of the settled slurry was recycled back to the pre-mix tank to maintain a desired recycle ratio of old solids to new solids. The optimum ratio was found to be between 5 and 7 to 1. This seeding technique generated a denser sludge, which settled to about 23 % solids as compared to the sludge of 5 % solids without the HDS recycle technique.

A High Density Sludge circuit, designed to treat 2,400 mVday of plant effluent, was installed at the Pasminco Clarksville Zinc Plant in December 1981. Typical analysis of the sludge is shown in Table III.

Table III - Typical Analysis of Neutralized Sludge

Zn Cd Pb Fe Mn Ca Mg % Dry Basis 11.4 0.17 <0.05 1.19 0.67 16.38 3.16

Metals Recovery Circuit

Although the amount of sludge was successfully reduced in both quantity and volume through the implementation of the gypsum recovery circuit and the HDS process, a feasibility study was made to convert the remaining sludge into commercial grade gypsum. This concept

874 LEAD-ZINC 2000

would completely eliminate the sludge generation from the effluent treatment plant. An analysis of the high-density sludge in Table III showed that there was an economic incentive to recover zinc in the sludge. Virtually all of the zinc and other metal hydroxides became soluble in the bench scale leaching experiment. The soluble metals could be returned to the zinc plant, after a filtration step, as a zinc sulfate solution. The test data, however, showed that more magnesium would be recycled back as well and that this would lead to a magnesium accumulation in the zinc circuit.

For the proposed zero-sludge concept to be feasible, a technique had to be developed to minimize the amount of magnesium precipitating with the zinc in the lime neutralization step. This led to another bench scale experiment with an objective to produce good quality effluent within the NPDES permit parameters, while minimizing magnesium precipitation in the HDS circuit. The effects of the neutralization pH on the solubilities of zinc and magnesium were investigated, and the results are shown in Figure 4. The purpose of the study was to find an optimum pH where most of the magnesium could be separated from the zinc. This optimum pH was found to be in the range of about 8.7, where approximately 80 % of the magnesium remained soluble while soluble zinc was reduced to about 1 mg/1. Soluble cadmium levels, however, were found to be too high for the NPDES permit.

o C/3

1000

10

0.1

6 7 8 9

Neutralization pH

10 11

Zn ■Mg

Figure 4 - Bench Scale Lime Neutralization Study

Since the solubilities of metal sulfides were much lower than those of the metal hydroxides, an investigation was made into the combination of lime, followed by a sulfide precipitation process (6,7). By adding a small dosage of sodium sulfide solution to the neutralized slurry, at pH of 8.7, bench scale tests showed the sulfide reaction to be fast and the soluble cadmium was preferentially precipitated. Although the sulfide precipitate was extremely fine, it was flocculated and settled easily. Typical bench-scale data, conducted in 1982, are depicted in Figure 5. In plant practice, the sodium sulfide solution was added to the neutralized slurry in the launder reporting to the 26 m thickener, prior to flocculant addition.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 875

After a settling step, both zinc and cadmium in the supernatant were found to be acceptable within NPDES parameters.

0.6

0.5

0.4

ε 0.3 -o U

0.2

0.1

0

0 2 4 6 8 10 12 14 16

Sulfide Dosage, mg/1

Figure 5 - Bench Scale Test on Soluble Cadmium

These modifications in the HDS circuit generated a sludge with a low magnesium content. The sludge was filtered, followed by a leaching and another filtration step. Zinc and other heavy metals were recovered in the filtrate and were returned to the zinc plant circuit, whereas the filter cake was recovered as commercial grade gypsum. The concept of operating an effluent treatment plant with zero-sludge production was successfully implemented at the Pasminco Clarksville Zinc Plant in September 1983.

CURRENT EFFLUENT TREATMENT PLANT

The current effluent treatment plant was installed in 1991, based on the design described above. One of the modifications was to utilize the HDS sludge filter cake, produced and stockpiled during the process development period prior to September 1983, in place of limestone in the neutralization of the magnesium bleed electrolyte. The process was also simplified to combine two gypsum filtration stages, before and after the HDS circuit, into one single operation. This was accomplished by pumping the neutralized electrolyte slurry directly to the HDS circuit, as shown in Figure 6.

876 LEAD-ZINC 2000

Lime

Plant Effluent

v v

Pre-Mix Tank

Hydroxide Precipitation

pH8.5

NaSH

Polymer

Acid Neutralization

pH3-4

Stockpiled HDS Cake

or Limestone

Gypsum to Market

Zinc Solution to Zinc Plant

Figure 6 - Current Effluent Treatment Plant at the Clarksville Zinc Plant

The current effluent treatment plant produces good quality water, acceptable for discharge under NPDES parameters, while recovering more than 1,600 tonnes of soluble zinc annually, as shown in Table IV. This represents an additional zinc recovery of about 1.5 % to the Pasminco Clarksville Zinc Plant.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 877

Table IV - Amount of Zinc Recovered from the Effluent Treatment Plant

1996 1997 1998 1999 Tonnes Zinc Recovered 1,696 1,870 1,984 1,800

GYPSUM MARKETING

The gypsum marketing program was initiated in 1980 when the bench-scale, followed by plant-scale tests, successfully produced commercial grade gypsum by neutralizing magnesium bleed electrolyte with limestone. At that time the Clarksville Zinc Plant was managed by the Gulf + Western Natural Resources Group which also owned the Marquette Cement Company. The first plant test on 40 tonnes of synthetic gypsum was conducted at the Marquette cement plant in Ogelsby, Illinois, in 1981. This led to several additional plant tests at other cement plants located within the economic shipping distance of the Clarksville Zinc Plant. Although the cement quality was not affected during the tests, handling problems were encountered, because of the high moisture of the gypsum. Some cement plants modified their gypsum handling system in order to use the gypsum with 15-20 % free moisture. Eventually, all of the gypsum produced was sold to cement and wallboard plants.

ACKNOWLEDGEMENTS

The author appreciates the contributions of colleagues in the Technical, Engineering and Operations departments whose joint efforts made this a successful project from inception through implementation. The author is also grateful to Pasminco management for permission to publish this paper.

REFERENCES

L. A. Painter, et al., "Jersey Miniere Zinc: Plant Design and Start-Up". Engineering and Mining Journal, July 1980, 65-88.

I. G. Matthew, O. M. G. Newman and D. J. Palmer, "Water Balance and Magnesium Control in Electrolytic Zinc Plants Using the E. Z. Selective Zinc Precipitation Process", TMS Paper Selection A-79-15, 1979.

W. S. Schmittroth, "The Removal of Magnesia and Lime from Dolomite-Type Concentrate", Presentation at the Annual Meeting of the AIME, New York, February 16-20, 1958.

W. J. Kuit, "Mine and Tailings Effluent Treatment at the Kimberly, B.C. Operations of Cominco Ltd.", CIM Bulletin, December 1980, 105-112.

R. V. Typliski and G. J. Labarre, "Wastewater Treatment at Hudson Bay Mining and Smelting Co., Limited", CIM Bulletin, June 1980, 99-103.

878 LEAD-ZINC 2000

6. Control and Treatment Technology for the Metal Finishing Industry - Sulfide Precipitation. EPA Summary Report (EPA 625/8-80-003), April 1980.

7. R. W. Adams, "Pilot Plant Treatment of Waste Liquors at Risdon", Australian Institute of Mining and Metallurgy Conference, Tasmania, May 1977.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 879

DEVELOPMENT, TESTING AND FULL-SCALE OPERATION OF A NEW TREATMENT METHOD FOR SELENIUM REMOVAL FROM ACIDIC EFFLUENTS

G.A. Monteith Noranda Inc., CEZinc Division

860 Cadieux Boulevard Valleyfield, Quebec, Canada, J6S 4 W2

G. Houlachi Noranda Inc., Noranda Technology Center

240 Hymus Boulevard Pointe-Claire, Quebec, Canada, H9R 1G5

M. Pineau and M. Laliberte SNC-Lavalin Inc.

455, Rene-Levesque Boulevard West Montreal, Quebec, Canada, H2Z 1Z3

ABSTRACT

Conventional physico-chemical treatment methods, either hydroxide or sulfide precipitation, can hardly achieve a high removal of selenium from the liquid effluents which can be generated in some metallurgical operations. In weak acid solutions, selenium exists mainly as anionic species (selenite Se03-2 and selenate SeCV2) which, therefore, cannot precipitate as metal hydroxide or metal sulfide compounds. At the same time, neither selenite nor selenate form stable precipitates with the common cations (Ca, Mg, Fe, etc.) that are typically used in water treatment. At Noranda Inc., CEZinc Division in Valleyfield, Canada, the introduction of new discharge requirements for effluent toxicity initiated a R&D program which resulted in the full-scale implementation of a new treatment method for selenium removal. The goal of this paper is to present this new method which is now being patented by CEZinc. The process chemistry will be presented along with pilot-scale developments and full-scale operational results.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

880 LEAD-ZINC 2000

INTRODUCTION

Zinc concentrates contain various levels of selenium. Part of this selenium is released in the off-gas from the roasting operation. Following treatment of the roaster off-gas at an acid plant, the ultimate fate of the selenium is the weak acid effluent from the acid plant.

The USEPA has recently begun the process of revising its water standard for selenium. This may indicate future needs for further removal of this element from liquid effluents which discharge into the environment. Renner (1) reports a recent workshop hosted by the EPA in May 1999 on the effects of selenium exposure to aquatic life. As part of this workshop, 15 documented cases were cited of selenium poisoning which affected fish and wildlife in the United States. Recent field studies on the effect of selenium pollution were also cited which reveal reproductive impairment and deformities caused by high levels of selenium in waters. Selenium behaves like mercury because it bioaccumulates through the chain food. In the current debate on selenium in the environment, some academics and scientists are asking for the halving of the current water standard; this is reported to be strongly opposed by mining and other industries. Others are questioning whether the standard should continue to be based on the concentration of selenium in the water, or whether it should be based on a direct measure of selenium in sediments or fish tissues. Although some conclusions to this debate are expected this year, the ultimate outcome may certainly be increasing future pressure for a higher removal of selenium from liquid effluents.

From that perspective, the joint work by Noranda Inc., CEZinc Division and its parent research agency, the Noranda Technology Center (NTC), brings a useful contribution to the search for better treatment technologies for selenium removal. The engineering company, SNC-Lavalin Inc., also contributed to this work, in particular in the areas of engineering, cost estimating and full-scale implementation.

The purpose of this paper is to present the results of the work carried out by Noranda Inc., CEZinc Division and NTC which was aimed at bringing into full-scale operation a treatment method capable of achieving a high removal of selenium from acidic effluents. The new treatment method is currently being patented under US and Canadian patents laws.

Sources of Selenium in Liquid Effluents

In typical zinc operations, the source of selenium in liquid effluents is the weak acid bleed from the acid plant. Selenium is most often found as a selenide solid solution impurity in a sulfide ore, as is the case for the zinc sulfide concentrates being used in the CEZinc plant in Valleyfield. During roasting of the zinc sulfide concentrate, the sulfur is oxidized to sulfur dioxide:

ZnS + 3 / 2 0 2 -» ZnO + S0 2 [1]

At the same time the selenide component, nominally present as zinc selenide, is oxidized and selenium dioxide is formed:

ZnSe + 3 / 2 0 2 -> ZnO + Se02 [2]

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 881

A small part of the sulfur dioxide (equation 1) is also transformed into sulfur trioxide during roasting, although no further oxidization of the selenium dioxide (from Se(IV) to Se (VI)) is possible under the conditions encountered in the roaster. As a result, selenium in the off-gas essentially exists as Se(IV).

The sulfur dioxide-rich gas produced in the roaster is sent to a sulfuric acid plant. At the Noranda Inc., CEZinc Division, three acid plants are operated. This gas is first washed with water to remove soluble impurities that would otherwise be found in the sulfuric acid. This wash water is saturated with sulfur dioxide and contains the following impurities:

• Sulfuric acid produced by hydration of the sulfur trioxide • Zinc chloride • Iron, as ferrous chloride • Mercury, mainly as mercurous chloride • Cadmium mainly as chloride • Copper, mainly as cuprous chloride • Ammonium chloride

Additionally, the selenium dioxide in equation [2] reacts with water to form the soluble selenous acid as given in Dean (2):

Se02 + H20 -> H2Se03 [3]

This reaction is very favorable and for this reason, the major part of the selenium contained in the zinc sulfide concentrate is found in the scrubbing liquor as selenous acid, which in turn dissociates into the selenite anion, SeC^"2.

The scrubbing liquor is recirculated at the scrubber. A bleed of the scrubbing liquor is required to control the build-up of sulfuric acid, selenium, mercury and other species. This is the so-called weak acid bleed. The volume of this bleed is relatively small. However, it is concentrated since it results from a recirculation loop at the scrubber. Historical data from the Valleyfield operation show that the average concentration of selenium in the weak acid bleed is in the order of 50 mg/1 with peaks up to 180 mg/1; the average mercury concentration is 30 mg/1 (SNC-Lavalin (3)). The concentration of selenium in such a bleed first depends on the specific selenium content of the zinc concentrate being processed. It is known that some zinc ore deposits are characterized by high contents of selenium. Unless adequate treatment technologies are available for the removal of selenium from the acid plant effluent, future development of such deposits may encounter environmental obstacles.

Conventional Available Treatment Technologies

It is known from experience and from publications such as Kuit (4), USEPA (5) and SENES (6) that the best available treatment technologies for treating acidic liquid effluents in the metal mining industry are:

• Lime precipitation, where the most well known application is the high density sludge (HDS) process

• Sulfide precipitation

882 LEAD-ZINC 2000

• Two-stage lime/sulfide precipitation.

It is also known that the above treatment technologies can hardly achieve selenium removals below 1 mg/1 in the treated effluents (not considering dilution by other process effluents). For liquid effluents with initial selenium concentrations in the range of 1 to 10 mg/1, the USEPA (7) treatability database reports selenium removal performance in the range of 1 to -3 mg/1. No performance results are reported by the treatability database for liquid streams containing selenium concentrations above 10 mg/1.

The limitations in removing selenium are as follows. As opposed to many metals and metalloids in acidic wastewaters, selenium exists as an anion, that is SeCV2 (selenite) or SeCV2

(selenate). As such, selenium cannot be precipitated by hydroxide (OFT) or sulfide (S ) to form metal hydroxide or metal sulfide precipitates. At the same time, neither selenite nor selenate form stables precipitates with the common cations (Ca, Mg, Fe, Na, etc.) which are used in wastewater treatment.

At the Noranda Inc., CEZinc Division, a new HDS (lime treatment) facility was recently commissioned to treat the liquid effluents from the zinc process. This facility, however, could not be expected to achieve a high removal performance for selenium.

In the typical lime treatment process, lime reacts with the weak acid bleed as follow:

H2S04 + Ca(OH)2 -> C a S 0 4 4 + 2H20 [4]

S0 2 (aq) + Ca(OH)2 -> C a S 0 3 1 + H 2 0 [5]

CdCl2 + Ca(OH)2 -> Cd(OH)2 i + CaCl2 [6]

2CuCl + Ca(OH)2 -> 2CuOH i + CaCl2 [7]

FeCl2 + Ca(OH)2 -> Fe(OH)2 i + CaCl2 [8]

ZnCl2 + Ca(OH)2 -> Zn(OH)2 4- + CaCl2 [9]

Copper and iron ions also undergo a slow but steady oxidation with air in the reactor and in the decantation pond. This ultimately results in the presence of cupric oxide, ferric oxide and hydroxide in the sludge produced by these reactions.

The selenous acid in the weak acid bleed also reacts with the hydrated lime:

H 2 Se0 3 + Ca(OH)2 -> CaSeOj + 2H20 [10]

The calcium selenite thus formed (equation 10) is, however, soluble in water to about 500 mg/1. Oxidation of calcium selenite to calcium selenate does not proceed to any significant extent, and even if it did, calcium selenate is even more soluble in water. Selenite salts of most common metals are also relatively soluble. Therefore, the treatment of the weak acid bleed by lime precipitation cannot achieve high removal levels of selenium.

Dilution is not an alternative either. Mixing of the weak acid bleed with other effluents from the zinc plant (for the joint treatment at the central HDS facility) would result in lower concentrations of selenium by dilution. At the Valleyfield operation, the acid bleed only

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 883

accounts for 5-10% of the total effluent flow rate. Dilution by mixing with other plant effluents is no longer accepted under the local environmental regulations, and source treatment of the weak acid bleed must be performed separately.

Alternative Treatment Technology

As part of a larger program to eliminate the toxicity of the plant liquid effluents, the CEZinc Division had the objective of treating the weak acid bleed effluent at the source, subject to the feasibility of developing and implementing a new treatment method capable of achieving a high removal performance for selenium.

The joint Noranda/NTC R&D program on selenium removal really entered into a new area of research. A literature review, contacts within the industry and contacts with treatment equipment suppliers indicated that no treatment technology had ever been applied at full-scale specifically for removing selenium from acidic liquid effluents.

Based on this review, two candidate technologies were first selected as potential avenues for evaluation and testing:

1. Ion exchange. 2. Cementation on a fluidized bed of aluminum pellets.

A third alternative was subsequently added to the R&D program as a result of the work carried out to improve the understanding of the electrochemistry of the cementation of selenium in option 2 above. This third option was:

3. Sodium hydrosulfite reduction and precipitation of selenium.

The feasibility of the three treatment options was evaluated in parallel according to a similar methodology, which progressively led to the establisment of milestones and an identification of the difficulties specific to each of the options. For the three options, the methodology included:

• Bench-scale laboratory tests • On-site pilot-scale test work • Capital and operating cost estimates for full-scale application.

Table I provides a brief description of the three technologies, along with a descriptive summary of the various development steps. Table II provides a summary of the results at each of these steps. As seen from Tables I and II, considerable effort and resources were invested in both bench-scale and pilot testing. In particular for the cementation technology, two different pilot-scale units were built and operated on-site at the plant. The first reactor was a fluidized bed of aluminum pellets; the second cementation reactor was equipped with ultrasonic equipment (Vibrating Bar, 2000 W, 25 kHz) rented from the German supplier Martin Walter as described in SNC-Research (8,9). The introduction of the ultrasonic equipment was a promising method to prevent passivation (inactivation) of the aluminum surfaces.

Tab

le I

- S

umm

ary

of T

echn

olog

y E

valu

atio

ns f

or S

elen

ium

Rem

oval

Cem

enta

tion

Io

n E

xcha

nge

Sod

ium

Hyd

rosu

lfit

e P

reci

pita

tion

Tec

hnol

ogy

Des

crip

tion

Pil

ot-s

cale

T

ests

Eng

inee

ring

S

tudi

es

Ele

ctro

chem

ical

pro

cess

; ce

men

tati

on

on a

lum

inum

sur

face

.

- F

luid

ized

bed

col

umn

wit

h al

umin

um

pell

ets.

2-

mon

th t

esti

ng o

n-si

te.

- F

luid

ized

bed

col

umn

wit

h al

umin

um

pell

ets,

equ

ippe

d w

ith

ultr

ason

ic

vibr

atin

g ba

r.

2 w

eeks

, on

sit

e te

stin

g.

- S

ubje

ct t

o pi

lot

test

res

ults

.

- R

esin

tec

hnol

ogy.

C

hem

ical

red

ucti

on a

nd p

reci

pita

tion

as

elem

enta

ry s

elen

ium

S°.

- A

nion

exc

hang

e co

lum

n te

sted

on-

site

. -

On-

site

pil

ot p

lant

: ch

emic

al f

eed

- A

nion

exc

hang

e co

lum

n te

sted

at

NT

C

sy

ste

m' r

ea

cto

rs'

Pla

te

filte

rs·

Lab

orat

ory.

-

3-m

onth

sit

e te

stin

g.

Pro

cess

Flo

w D

iagr

am.

Equ

ipm

ent

sizi

ng.

Equ

ipm

ent

layo

ut.

Bui

ldin

g re

quir

emen

t.

Cap

ital

cos

t es

tim

ates

.

- P

roce

ss F

low

Dia

gram

.

- E

quip

men

t si

zing

.

- E

quip

men

t la

yout

.

- B

uild

ing

requ

irem

ent.

- C

apit

al c

ost

esti

mat

es.

Tab

le I

I -

Sum

mar

y of

Pil

ot S

cale

Tes

t R

esul

t

Cem

enta

tion

Io

n E

xcha

nge

Sod

ium

Hyd

rosu

lfit

e P

reci

pita

tion

Res

ults

fr

om

Pil

ot-s

cale

T

ests

Res

ults

fr

om

Eng

inee

ring

S

tudi

es

- G

ener

atio

n of

hyd

roge

n se

leni

de g

as.

- In

com

plet

e S

e re

mov

al w

itho

ut a

ddit

ion

of S

O2.

- C

ontr

ol o

f SO

2 ad

diti

on i

mpr

acti

cabl

e.

- S

tric

t co

ntro

l of

F~

and

CF

leve

ls i

mpr

acti

cabl

e.

- S

e re

mov

al <

1 p

pm.

No

engi

neer

ing

stud

ies.

O

ptio

n di

scar

ded

base

d on

pi

lot

wor

k.

- F

ree

acid

lev

el c

ontr

ol

requ

ired

.

- C

o-lo

adin

g of

res

ins.

- SO

2 ad

diti

on r

equi

red.

- E

xces

sive

cap

ital

cos

t.

- C

ompl

ete

rem

oval

of

Se <

0.0

5 pp

m.

- D

osag

e of

rea

gent

s de

pend

ent

on r

eal

tim

e in

form

atio

n on

act

ual

conc

entr

atio

ns o

f S

e.

- B

atch

ope

rati

on r

equi

red.

- F

iltr

atio

n of

sol

ute

key

full

-sca

le v

aria

ble.

- C

apit

al c

ost

one

orde

r of

mag

nitu

de l

ower

th

an i

on e

xcha

nge.

r en

> D

Z n

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 885

Monette (10) reported on the pilot-scale test work using aluminum pellets and showed that high selenium removal (< 1 mg/1) was achieved. The cementation technology is still considered promising since similar work is being pursued by others, namely Kuan et al. (11) for the removal of Se (IV) and Se (VI) using aluminum-oxide-coated sand. In the present application, however, pilot-scale testing brought to light impracticable operating requirements (see Table II). The option was then discarded.

Ion exchange technology was tested at the pilot-scale as well. Various difficulties could not be resolved, in particular the problems of co-loading of the resin with other species, of the stripping of the resin and of the need for pre-conditioning of the free acid levels. Before justifying further pilot-scale work, an engineering study was conducted to obtain preliminary capital cost estimates for a full-scale ion exchange unit for selenium removal. Assuming that an optimal resin could be selected, the results of the capital cost estimates did not make this option attractive, compared to option 3 (sodium hydrosulfite reduction/precipitation). Given the process difficulties to be resolved with the ion exchange technology, the third option was then selected for full scale development and implementation.

Selected Treatment Process - Sodium Hydrosulfite Reduction/Precipitation of Selenium

The basic principle of the selected treatment method is the reduction and precipitation of the selenium present in the weak acid effluent as elemental selenium (Se (IV) —> Se°). Selenite present in the weak acid bleed is a relatively strong oxidizing agent as given in Dean (2):

H2Se03 + 4H+ + 4e" <H> Se + 3H20 E° = + 0.74 V [11]

However, this oxidizing potential is not sufficiently high that it would react with the sulfur dioxide initially present in the weak acid bleed:

S02 (aq) + H20 <-► S04"2 + 4H+ + 2e" E° = -0.158K [12]

The high sulfate content of the weak acid bleed, and the comparatively low selenite and sulfur dioxide concentrations, do not favor the reduction of selenite by the sulfur dioxide.

The use of a stronger reducing agent than sulfur dioxide is required to transform selenite to elemental selenium. Sodium hydrosulfite, Na2S204, also called dithionite, is one such agent. It reacts as follows:

HS204"+ 2H20 <-» 2H2S03 + H+ + 2e" E° = + 0.08 V [13]

This is exactly equivalent to:

S204"2 <-* 2S02(aq) + 2e- [14]

Reaction 14 is particularly interesting because its end product is already present in the weak acid bleed. Removal of the selenite by this reaction would not cause the introduction of another potentially toxic by-product in the treated stream. The overall equation for the reduction of selenite to elemental selenium using sodium hydrosulfite is given by:

LEAD-ZINC 2000

H2Se03 + 4H++ 2S204"2-^ Se 4-+ 4S02(aq) + 3H20 E° = + 0.82 V [15]

Full Scale Process for Selenium Removal

In the design of the full-scale unit, the removal of mercury and the removal of selenium are two treatment operations, which were integrated into a two-stage process:

• First stage: sulfide precipitation of mercury using Na2S. Sulfide precipitation also results in the removal of part of the selenium initially present in the bleed.

• Second stage: complete removal of selenium using sodium hydrosulfite (Na2S2C>4) reduction/precipitation.

Figure 1 presents a conceptual process flow diagram of the new selenium removal unit.

Figure 1 - Block Flow Diagram for the Full Scale Selenium Removal Unit

The treatment operation is a batch process. The dosage of the reagent (Na2S2C>4) for Se removal is very sensitive to the actual concentration of selenium. Since such an operating variable cannot be measured on-line, the use of a batch process makes it possible to obtain a proper estimate of the selenium concentrations in each batch and to adjust the dosage accordingly.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 887

As shown by Figure 1, the treatment scheme includes the following:

• Storage tanks for the weak acid effluent pumped from the acid plant • Filtration of the weak acid to remove calcine particles from the solution • First stage: injection of Na2S and precipitation of mercury • Second stage: injection of sodium hydrosulfite and reduction/precipitation of

selenium • Filtration to remove mercury and selenium precipitates • Pumping of the treated bleed to the central treatment facility for neutralization and

pH adjustment.

As mentioned above, the dosage of sodium hydrosulfite is governed by the concentration of selenium. This operating condition was established as a result of the bench-scale and pilot-scale test work. For each batch, the operator is required to collect a sample at the outlet of the first stage reactor and to obtain a laboratory determination of the selenium concentration. Alternative instrumentation and control techniques are being investigated, which would allow, in the near future, the replacement of the current manual estimation of the selenium level with an automated device.

As it leaves reactors N°3 or N°4, the solution is filtered using a filter press. The filtrate, free of mercury and selenium, is sent to the central treatment facility (high density sludge) for final neutralization (see Figure 2).

According to the SNC-Lavalin (3) design criteria, the full scale unit was designed for an average treatment capacity of 300 m3/d, based on 18 hours operation per day (i.e., 2 batches per day).

Zinc Smelter

Acid Plants

Weak Acid

Bleed

Site Storm Water

Process Effluents

Hg/Se Removal Unit Filter cakes

ψ 0

disposal or recycling

Equal zation >

W

High Density Sludge Central Treatment

Facility

Discharge to river

>

Sludge to storage basin

Figure 2 - General Effluent Treatment Scheme at CEZinc

Operating Results

Commissioning and start-up of the full-scale selenium removal unit took place in December 1998. Once the unit reached close to normal, steady-state operation, an intensive performance evaluation program was undertaken over 18 different treatment batches. Twelve parameters were measured daily at various points along the treatment process chain.

Tab

le I

II -

Ope

ratin

g R

esul

ts,

Tre

atm

ent

of th

e W

eak

Ble

ed f

or S

elen

ium

Rem

oval

A

cid

Bat

ch

Num

ber

4 5 6 7 8 9 10

11

12

13

14

15

16

17

18

19

20

21

Eff

luen

t A

cidi

ty

as H

2S

04

(g/i

)

26

26

26

26

26

26

26

26

26

26

26

26

Se

(mg/

1)

Init

ial

13.6

13.6

13.6

13.6

13.8

13.8

13.8

13.8

13.8

29.7

29.7

29.7

29.3

29.3

34.7

34.7

30.1

30.1

Fin

al

<0

.01

<0

.01

0.13

<0

.01

<0.

01

<0

.01

<0

.01

<0.

01

<0

.01

0.69

0.22

0.17

0.67

0.14

0.05

0.1

<0

.01

<0

.01

Cd

(mg/

1)

Init

ial

4.5

4.5

4.5

4.5

4.5

4.5

4.5

4.5

4.4

3.3

3.3

3.3

3.2

3.2

3.3

3.3

3.3

3.3

Fin

al

—

3.8

3.8

3.9

3.9

3.8

4.0

3.9

3.8

1.9

1.7

1.3

2.2

1.7

2.1

2.6

2.5

2.5

Cu

(mg/

1)

Init

ial

2.3

2.3

2.3

2.3

2.3

2.3

2.3

2.3

2.3

0.32

0.32

0.32

0.33

0.33

0.36

0.36

0.35

0.35

Fin

al

<0.

01

<0

.01

<0

.01

<0.

01

<0

.01

<0

.01

<0

.01

<0

.01

<0

.01

<0.

01

<0.

01

<0.

01

<0

.01

<0

.01

0.1

<0

.01

<0.

01

<0

.01

As

(mg/

1)

Initi

al

8.4

8.4

8.4

8.4

8.2

8.2

8.2

8.2

8.1

6.3

6.3

6.3

6.2

6.2

6.3

6.3

6.1

6.1

Fin

al

<0

.1

<0

.1

<0

.1

1.4

<0

.1

0.15

<0

.1

<0

.1

0.22

0.13

0.07

<0

.1

0.76

<0

.1

<0

.1

0.13

<0

.1

0.1

Pb

(mg/

1)

Init

ial

5.3

5.3

5.3

5.3

5.1

5.1

5.1

5.1

4.9

5.5

5.5

5.5

5.2

5.2

4.6

4.6

4.4

4.4

Fin

al

3.4

2.0

2.0

3.5

2.5

2.2

3.7

2.9

1.6

1.2

0.83

0.36

2.1

0.64

0.41

1.1

0.84

0.95

r m

>

o N

z

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 889

Table III summarizes the results for selected key metals for the initial composition of the weak acid bleed as it enters the treatment unit and for the final composition of the weak acid bleed after treatment.

During this evaluation program, the full-scale plant was run under different combinations of dosages for the two reagents, Na2S and Na2S2C>4. Under optimal combinations of the dosages, a very high removal of selenium (< 0.01 mg/1) is achieved for most of the batches as seen from Table III. Based on the guidelines provided from the pilot work, further variations in dosage were tested to assess the sensitivity of the treatment process under actual full-scale conditions. However, it was found that even under non-optimal dosages of the reagents, low concentrations of selenium were obtained (0.1 - 0.2 mg/1) in the treated bleed.

Aside from selenium, copper is removed below 0.01 mg/1, while some removal of cadmium and lead is also observed. No removal of soluble zinc is achieved (results not presented) which is consistent with the zinc equilibrium diagram for the present conditions. As for mercury, it is essentially removed in the first stage as a result of Na2S precipitation. Final concentrations of mercury in the treated acid bleed are consistently below 0.05 mg/1.

CONCLUSIONS

The treatment method developed by Noranda Inc., CEZinc Division and NTC is the only known full-scale process unit, which is specifically designed for achieving a high removal of selenium from acidic solutions. The treatment of the weak acid bleed is performed at the source, before mixing and dilution by other plant effluents, which are sent to the central HDS treatment facility. Under optimal dosage of the reagents, selenium is consistently removed to below 0.01 ppm in the treated weak acid. The unit has been successfully operating for VA years. More work is currently being pursued in the two areas of automated, on-line estimation of the selenium concentrations in the weak acid bleed before batch treatment (as opposed to the current manual collection of samples followed by laboratory analysis) and of the optimization of the reagent dosage for higher levels of selenium.

ACKNOWLEDGEMENTS

Numerous persons contributed to the success of this R&D program. The authors wish to acknowledge the contributions of the many team members from the staff of Noranda Inc., CEZinc Division and from the Noranda Technology Center. In particular, the authors wish to thank Mr. Mike Agnew and Mrs. Lucy Rosato, who brought to this project the vision and trust which were needed to make it a success.

890 LEAD-ZINC 2000

REFERENCES

1. R. Renner, "EPA Decision to Revise Selenium Standards Stirs Debate", Environmental Science and Technology, Vol. 32, No. 15, 1998, 915-923.

2. J.A. Dean, Ed., Lange Handbook of Chemistry, 14th Edition, , McGraw Hill, New York, U.S.A., 1992.

3. SNC-Lavalin, "Process Design Criteria - Detailed Engineering for the Selenium Project", Report 012097-0000-49EC-0001, 1997.

4. W.J. Kuit, "Mine and Tailings Effluent Treatment at the Kimberley, B.C. Operations of Cominco Ltd.", CIM Bulletin, Vol. 73, No. 824, 1980, 105-112.

5. USEPA, "Development Document for Effluent Limitations, Guidelines and Standards for the Non-ferrous Metals Manufacturing Point Source Category" Volume IV, EPA 440/1-89-019-1. Office of Water Regulations and Standards. Industrial Technology Division, May 1989.

6. SENES Consultants Ltd., "Acid Mine Drainage - Status of Chemical Treatment and Sludge Management Practices", Report prepared for the Canadian Mine Environment Neutral Program (MEND) and CANMET, Canadian Center of Mineral and Energy Technology, Ottawa, Canada, 1994.

7. USEPA, "Treatability Data Base", Version 6-0, May 1998.

8. SNC-Research Inc., "Continuous Cementation of Selenium Using Aluminum Pellets and Ultrasonic Auto-cleaning : Process Feasibility" (in French), Technical Report prepared for Noranda, CEZinc Division by SNC-Research Inc., Montreal, Canada, November 1995.

9. SNC-Research Inc., "Continuous Cementation of Selenium using Aluminum Pellets and Ultrasonic Auto-cleaning: Confirmation of the Application" (in French), Technical Report prepared for Noranda, CEZinc Division by SNC-Research Inc., Montreal, Canada, January 1996.

10. S. Monette, "Results from 3-month Test Work at the Selenium Pilot Plant" (in French), Technical Memorandum, Noranda Technology Center, Pointe Claire, Quebec, Canada, March 1995.

11. W.-H. Kuan, S.L. Lo, M.K. Wang and C.F. Lin, "Removal of Se (IV) and Se (VI) from Water by Aluminum Oxide Coated Sand", Water Research, Vol. 32, N° 3, 1998, 915-923.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 891

RECOVERY OF SULFIDES FROM SULFATE-CONTAINING BLEED STREAMS USING A BIOLOGICAL PROCESS

C.F.M. Copini Budel Zink

P.O. Box 2001,6020 AA Budel The Netherlands

G.H.R. Janssen, C.J.N. Buisman and S. Vellinga PAQUESB.V.

P.O. Box 52, 8560 AB Balk The Netherlands

ABSTRACT

Non-ferrous metallurgical plants are often confronted with the need to neutralize bleed streams or effluents containing sulfates. The treatment of these streams could produce large amounts of gypsum. With a biological sulfate-reduction process, the production of gypsum is avoided and the metals can be recovered as sulfides. In this paper, the biological treatment of an industrial stream is presented. The bleed solution from a roaster gas scrubber was treated by means of a two-step process. First, the acid was neutralized with calcine. The resulting zinc sulfate solution was converted to ZnS with bacteria using PAQUES-THIOPAQ® technology. The precipitated ZnS was de-watered to 70 wt% solids, and was returned to the roasters. The quality of the effluent stream complied with the environmental regulations.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

892 LEAD-ZINC 2000

INTRODUCTION

Budel Zink B.V., a company owned by Pasminco Ltd, has operated a zinc refinery since 1973 at Budel-Dorplein in the Netherlands. The plant produces more than 200,000 t/y of zinc. The conventional roast-leach-electrowinning process used by this plant generates various waste streams containing sulfate and zinc. Until mid 2000, these streams were treated by a conventional process (neutralization with milk of lime) producing large quantities of gypsum. As of July 1, 2000, the Dutch government prohibited further storage of residues at the Budel Zink site. For this reason, alternative water treatment processes were studied to install a process in which the storage of gypsum would not be needed while meeting the effluent quality requirements of the Dutch government. It was soon learnt that the production of high quality, clean, commercial grade gypsum was very difficult. Thus, alternative options, which did not produce gypsum, were studied. This paper provides an overview of the chosen biological process route.

WASTEWATER STREAMS

Within the Budel Zink process various wastewater streams are generated. The main components of these streams are sulfate and zinc. A distinction can be made between streams with low and high sulfate content. Streams with a low sulfate content (< 1 g/L) can be treated in the existing biological wastewater plant, SRB (sulfate reducing bacteria). This plant was commissioned in 1992 and was intended for the purification of groundwater underneath the Budel Zink terrain (1). In the SRB plant, S 0 4 is reduced to S2", after which the contained metal-ions precipitate as metal sulfides. These sulfides are recycled to the main zinc process and no gypsum is produced. An expansion of the plant in 1999 improved the capabilities of the SRB plant to handle both groundwater and low-sulfate wastewater streams.

There are two streams that belong to the high sulfate category:

• Wash tower acid (scrubber discharge from the roaster acid plant). Typically this flow is about 25 m3/h and it contains 10 g/L H2S04, 0.5 g/L HF, 1 g/L HC1 and 0.5 g/L Zn.

• Magnesium bleed. This bleed is necessary to prevent the build-up of magnesium in the electrolyte. Typically 0.5 m /h of purified solution and/or spent electrolyte has to be bled from the circuit.

These two streams cannot be treated in the SRB plant because their sulfate content is too high, and in case of the wash tower acid, the fluoride content is also high. Between 1995 and 1999 various processes were studied to treat these streams. The objective was to develop a process that allowed a direct purge of these streams or to have the capability of bleeding them, after a final treatment, to the SRB plant. Extensive testwork was done at the laboratory and pilot plant scale. Based on these results, the biological process route was chosen.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 893

BASIC FLOWSHEET

General

The chosen biological process route consists of the following steps which are illustrated in Figure 1.

• Neutralization of the wash tower acid, WT A, with calcine • Fluoride removal by precipitation of CaF2 in a Crystalactor • Mixing with magnesium bleed (i.e., with zinc electrolyte) • Biological conversion of ZnSC>4 to ZnS, using H2 as the electron donor. Hydrogen is

produced in a reformer, which converts natural gas and steam to H2 and CO2. • Precipitation and separation of ZnS • De-watering of ZnS • Treatment of the effluent of the bioreactor in the existing SRB plant where the

excess sulfide is converted to elemental sulfur.

Magnesium bleed

calcine </nO)

C O

C'aF, pellets crystalactor

Hj/C'O,

Figure 1 - Principal Scheme for the Full Scale Installation

894 LEAD-ZINC 2000

Neutralization

The wash tower acid (WTA) contains SO4, Zn, Cl, F and traces of other elements. In the first process step, the WTA is neutralized with calcine in two stirred tanks in which the pH is increased from near 0 to 5.1 producing ZnSCV In a third tank, the slurry is aerated, thereby oxidizing and precipitating the iron. Other minor trace elements also co-precipitate with the iron. Finally, in a thickener, the neutralization residue and the precipitated iron are extracted as a slurry in the thickener underflow. This underflow is mixed with sludge from the bioreactor and the SRB plant. After de-watering in a decanter-centrifuge, the cake is returned to the roasters. The effluent is sent to the fluoride removal stage.

To Crystal actor

To Decanter Centrifuge

Figure 2 - Neutralization and Iron Removal Circuit

Fluoride Removal

Fluoride present in the neutralized WTA solution is precipitated as CaF2 in a fluidized bed reactor (Crystalactor) according to the following reaction:

CaCl2 + 2NaF ->■ CaF2 (pellets) + 2NaCl (1)

The Crystalactor® (2) is a fluidized-bed type reactor (Figure 3) in which sand is used as a seeding material to crystallize CaF2. The product consists of pellets (approximately 1 mm) with a sand core, and a covering layer of crystallized CaF2. An important issue in the design of the reactor is to prevent homogeneous nucleation. Thus, the reagent, the WTA, and the recycle stream are distributed in the bottom part of the Crystalactor to achieve optimal mixing, and to prevent a high supersaturation. The pellets are automatically discharged and fresh seed material is added based on the pressure in the column. After atmospheric drying, pellets which are almost free of water are obtained. To obtain pellets of pure CaF2, plans are in place to crush part of the product into smaller particles (approximately 0.2 mm) and to use these as the seeding material.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 895

Figure 3 - Crystalactor®

Sulfate Reduction

In the H2 reactor, sulfate-reducing bacteria convert ZnS04(aq) to ZnS(S) using H2 as an electron donor.

ZnS04(aq) + 4H2 -> ZnS(5) + 4H20 (2)

The H2 is produced on-site in a natural gas reformer. The natural gas and steam are converted to H2 and C02.

CH4 + 2H20 -> 4H2 + C02 (3)

The reformer produces approximately 500 nm3/h of H2 using 200 nm3/h of CH4. It also produces approximately 1 tonne/h of steam.

896 LEAD-ZINC 2000

The gas mixture containing approximately 80vol% H2 is fed directly to the bioreactor. The CO2 gas is used as a carbon source for the bacteria. To achieve good mixing without introducing high shear forces, a gas-lift loop type reactor is used, as illustrated in Figure 4. The fresh H2 from the reformer is mixed with the recycle gas and is distributed over the riser surface of the reactor. A small amount of the gas recirculation flow is bled off to remove inert components (mainly N2). The gas bleed is used as fuel for the burner of the reformer system.

riser ■<-

I Recirculation gas

6rrd t t

t t

t t

t 1

three phase settler

-*· down comer

H2/C02

Figure 4 - Gas-Lift Loop Type Reactor

ZnS Separation and Dewatering

The precipitated ZnS is separated in a thickener and is subsequently sent to a thickener. Flocculant has to be added to obtain a clear decantate and a manageable cake. The dewatering circuit is shown schematically in Figure 5.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 897

ZnS sludge \ Ca. 200 g/1 \

ZnS sludge Ca. 400 g/1

ZnS to roasting Ca. 70 wt% dry solids

Figure 5 - ZnS Precipitation and Dewatering Circuit

DESIGN OF THE FULL SCALE INSTALLATION

General

The design criteria for the full-scale installation were:

• Flow of WTA: 8-40 m3/h • Electrolyte bleed flow: 0.2-1 m3/h • Concentration of SO4 in the WTA: 5-50 g/L • Concentration of SO4 in the electrolyte bleed: 300 g/L • Concentration of fluoride in the WTA: 500 mg/L • Maximum SO4 load: 10 tonnes/day.

The guaranteed effluent concentrations for the complete installation (including the SRB groundwater treatment plant) are:

• S04 < 200 mg/L • Fluoride < 10 mg/L • Zn2+ < 0.3 mg/L • Cd2+ < 0.01 mg/L • Total solids < 5 mg/L.

An overview of the installation is given in Figure 6. A detailed picture of the bioreactor is shown in Figure 7.

898 LEAD-ZINC 2000

Figure 6 - Overview Drawing of the Full Scale Installation

Figure 7 - The Bio Gas-Lift Loop Type Reactor, Volume 500 m3

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 899

The design, engineering and construction of the full-scale installation treating the WTA and the magnesium bleed was completed in July 1999. Mechanical and software tests were completed in August 1999. Commissioning of the installation took about two months. Since October 1999, the installation has been in operation treating the WTA flow and the bleed from the electrolysis circuit.

Safety and Environmental Factors

Special attention has been paid to safety and the environmental aspects. The recirculation gas contains between 1-5% H2S. The gas bleed (containing approximately 30% H2) is first de-sulfurised using a two-stage sodium hydroxide scrubbing system prior to it being used to heat the reformer.

All plant units (except for the bioreactor) operate under a slight vacuum. Exhaust air is scrubbed in a two-stage sodium hydroxide scrubber system. Finally, the air passes a bio-filter to remove all traces of H2S. In this way there are no odor problems.

START-UP

The 500 m bio reactor was commissioned with 20 m of seed sludge. The loading rate has been increased successfully. Within two weeks, the biological capacity was high enough to treat the complete WTA flow (see Figure 8). As of February 2000, a capacity of 350 kg SOVh has been reached. The target is 400 kg SCVh, which will be reached by further optimizations. The effluent concentration of SO4 can be reduced to 50 mg/L, but it is controlled at approximately 500 mg/L.

400

300

j2 200

100

7-2-00 9-2-00 11-2-00 13-2-00 15-2-00 17-2-00 19-2-00

Figure 8 - Sulfate Removal Capacity in kg/h SO4 Removed

900 LEAD-ZINC 2000

The fast biomass growth facilitated the easy start-up of this system. By regulating the amount and type of nutrients, the biomass growth can be controlled. As a result, the start-up was carried out with only 4 m3 of seed sludge per 100 m3 of reactor volume.

Operators were mainly trained on-the-job. During start-up, one experienced technologist assisted each shift.

OPERATIONAL EXPERIENCE

The results achieved during the first 6 months of operation are summarized in Tables I and II

Table I - Operational Data: Crystalactor Target Achieved

Input Stream flow 8-40 m /hour 8-30 m /hour pH 5,2 5-5,5 Fluoride concentration in the input stream max. 500 mg/L 100-250 mg/L Fluoride concentration in the output stream max. 50 mg/L 20-40 mg/L SO4 concentration range 5-50 mg/L 5-15 g/L

Table II - Operational Data: Bioreactor

Input Stream Flow PH SO4 concentration in the input stream SO4 concentration in the output stream ZnS production

Target 8-40 m3/h 7.2 5-50 g/L 300 mg/L 10 tonne/day

Achieved 8-30 m3/h 6.8-8.1 5-15 g/L 50 mg/L 8.5 tonne/day (higher loads were not available yet)

The operating problems encountered were:

Clogging of the calcine mixing tank; the calcine (ZnO) is first suspended in water before it is mixed with the WTA. The suspension vessel had to be adjusted to prevent clogging by the inlet part of the calcine 'dust'. Reformer: The water level-control for the CO2 absorber level, working at 16 bar, was not reliable under all process conditions. Therefore, the existing radar measurement device will be replaced by a differential pressure measurement.

ENVIRONMENTAL IMPACT

Treatment of the WTA by the conventional neutralization process led to the production of large volumes of gypsum. With the successfully-developed bio-process, no gypsum is produced and an improvement of the water quality has been realized. In addition to a clean

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 901

effluent, CaF2 and ZnS are also produced; the ZnS is returned to the roaster. The net result is a reduction of gypsum production from 18 t/d to 0 t/d.

PROJECT SUPPORT

This project was supported by a Dutch E.E.T. grant (Economy, Ecology and Technology), from the Ministry of Economic Affairs, the Ministry of Education, Culture and Sciences and the Ministry of Housing, Spatial Planning and the Environment.

REFERENCES

1. P.J.H. Scheeren, R.O. Koch and C.J.N. Buisman, "Geohydrological Containment System and Microbial Water Treatment Plant for Metal-Contaminated Groundwater at Budelco", 1993.

2. R.J.M. van Lier, C.J.N. Buisman and A.Giesen, "Crystalactor® Technology and Its Applications in the Mining and Metallurgical Industry", Solid/Liquid Separation including Hvdrometallurgy and the Environment, G.B. Harris and S.J. Omelon, Eds., Canadian Institute of Mining, Mettallurgy and Petroleum, Montreal, Canada, 1999, 221-231.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 903

GOETHITE: FROM RESIDUE TO SECONDARY BUILDING MATERIAL UNION MINIERE'S GRAVELIET® PROCESS

J. Winters Union Miniere - UM Research

Kasteelstraat 7 B-2250 Olen, Belgium

L. Vos and C. Canoo Union Miniere - UM Zinc

Zinkstraat 1 B-2490 Balen, Belgium

ABSTRACT

Committed to a sustainable development of zinc applications, Union Miniere is investigating processes for the minimisation of waste and residues originating from the classic hydrometallurgical zinc process. In that context, Union Miniere has developed a method for the cold inertisation and solidification of goethite, the Graveliet® process. The result is a hard and inert gravel that can be used in the building industry. During the past two years, a pilot plant for the production of Graveliet® has been built and operated. In the present paper, the Graveliet® process and the pilot plant are described after which the properties of the produced gravel are discussed, from both a mechanical and an environmental point-of-view. Graveliet fulfils the criteria for use as a secondary building material. The GravelietJ process is a viable technology to avoid stockpiling of goethite residues and it thus assures the long-term development of the hydrometallurgical zinc process.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

904 LEAD-ZINC 2000

INTRODUCTION

Most of the world's zinc production is obtained via the roast/leach/electrowinning (RLE) process. Starting from sulphidic concentrates, this hydrometallurgical process allows the production of special high-grade zinc. During this process, an iron rich residue (goethite or jarosite) is produced. This residue has to be stockpiled, and this is clearly an environmental issue for the future.

Committed to a sustainable development of zinc applications, Union Miniere (UM) is investigating processes for extending its plant feed to recycled 'end-of-life' zinc products and processes for the minimisation of waste and residues. In that context, UM has developed a method for the cold inertisation and solidification of goethite, the Graveliet® process. The result is a hard and inert gravel that can be used in the building industry. During 1998, a pilot plant for the production of Graveliet® was built. It will be operated for about two years. In this paper, first a schematic overview of the Graveliet process is given. Secondly, the pilot plant is described. After that, the material properties of Graveliet® are discussed, from both an environmental and a mechanical point-of-view. It is concluded that Graveliet fulfils the necessary criteria as a secondary building material.

THE GRAVELIET® PROCESS

UM developed a process for the inertisation and solidification of the goethite residues on bench scale during 1995 and 1996 (1). This process consists of mixing goethite with blast furnace slag and converter slag from the steel industry. Converter slag itself is a residue from the steel industry. The idea is based on the technology of concrete making. Analogous processes have already been described in the literature for several kinds of wastes, including the jarosite iron residue (2). During mixing and further curing, pozzolanic reactions take place between the acidic goethite and the basic slag. The product cures into a hard and inert kind of rock. UM has protected this process by a registered patent (3). The brand name of the hardened goethite is Graveliet . The Graveliet® process is a low cost and low energy consuming process, and it is a valuable solution for the iron waste problem in the zinc industry. The process steps are summarised in Figure 1.

The first step consists of the preparation of the raw materials. This includes, on the one hand, the milling of both the blast furnace slag and the converter slag, and on the other hand, the thorough washing and filtering of the goethite. The latter is necessary to remove the remaining soluble sulphates. The milling has an influence on the pozzolanic reaction and thus the final quality of the Graveliet product. In the next step, the three components are intensively mixed, and water is added in order to make a clay-like paste that can be fed to a forming device. After forming, the obtained material is submitted to a curing stage of several days in a conditioned environment.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 905

Figure 1 - Schematic Overview of the Graveliet Process

Depending on the envisaged application, the cured material can be crushed in a final step in such a way that it can be directly incorporated as gravel substitute in concrete.

THE GRAVELIET® PILOT PLANT

The pilot plant was constructed during the first half of 1998 at the UM Overpelt plant in Belgium. The concept is such that it is possible to have two different forming operations performed on the same installation: extrusion of pellets as well as bricks. The first is the basis for the production of gravel that can be used in concrete and will be discussed in detail.

Figure 2 - Stocking of The Goethite; After Washing and Filtering, the Water Content is Around 30%

906 LEAD-ZINC 2000

The goethite is washed and filtered on industrial filter presses. The filtered goethite is stored inside a warehouse as illustrated in Figure 2. The slags arrive at the pilot plant as milled material. The milling of the slags is performed by an external operator. The slags are stored in silos as shown in Figure 3.

Figure 3: The Slags are Stored in Two Different Silos. The Slags are Fed to the Planetary Mixer by Means of Screw Conveyors

The mixing of the components is performed in a batch operation in a kind of planetary mixer, whereas the subsequent operations are all continuous. Batch mixing was chosen in order to have the possibility for adjustments and to assure the correct final mixtures. The mixture is then fed into an extrusion press as illustrated as illustrated in Figure 4. Although the Graveliet® mixture does not behave as a plastic (e.g., as clay), no major problems arise in this operation. The extrudability seems to be largely controlled by the apparent humidity of the mixture.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 907

Figure 4 - Extrusion Press

The extrusion mould determines the form of the product. In the case of pellet production, the mould has circular holes and cylinders are produced. These cylinders are cut to a predetermined length and are rounded in a rotating drum The pellets are subsequently stored in curing rooms where both the humidity and the temperature can be controlled. A curing room is shown in Figure 5.

Figure 5 - Curing Room; Both Temperature and Humidity Can be Controlled

After a given time, the pellets are removed from the curing rooms and stored in an open-air location for further curing. Depending on the application, the pellets can be used as such or can be crushed to a desired size distribution.

908 LEAD-ZINC 2000

DISCUSSION OF THE RESULTS

First, the eco-technical (or environmental) evaluation of Graveliet will be presented. Next, some of the mechanical and other physical properties will be discussed.

Eco-technical Evaluation

According to the Flemish legislation "Vlarea" (4), it is possible to use a waste material or a residue as a secondary building material provided that certain conditions related to the material properties are fulfilled. A distinction is made between shaped materials and non-shaped materials.

The Vlarea legislation stipulates that it is the amount of metals leached under specific conditions that determines the usability of a product as a secondary building material, rather than the total amount of contained metals. To evaluate this parameter, a number of standard leaching procedures have been defined. Depending on the use, different leaching tests must be employed. The metals that are mentioned in the Vlarea as being critical are the following: As, Cd, Cr, Cu, Hg, Ni, Pb and Zn. Based on risk analysis, the certificate granting authority can then decide upon the allowable applications. Only the use as a so-called "shaped" application well be discussed hereafter.

Petrographical Analysis

One of the important questions during the eco-technical evaluation of Graveliet* was whether a chemical reaction really takes place between the added slags and the goethite. This is important since it is assumed to influence the stability of the hardened Graveliet3 mixture. A mere incorporation of goethite in a matrix formed by the reaction product of blast furnace slag and converter slag would, indeed, not be as stable in the long term as a fully reacted Graveliet* in which goethite is chemically bound to the added slags.

In Figure 6, a SEM picture is given which focusses on the reaction rim around a grain of blast furnace slag. An analysis by EDX reveals that, within this zone, goethite must have reacted with the blast furnace slag and the converter slag to form a complex metal oxide/hydroxide compound.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 909

Figure 6 - SEM Picture Focussing on the Reaction Rim Between the Blast Furnace Slag and the Rest of the Matrix (from (5))

Use of Graveliet as a Shaped Material - Leaching Tests

The Vlarea legislation is based on the principle that the concentration rise of any of the critical parameters in the soil, because of leaching from the material ("emission"), must be less than 1% per 100 years (immission; principle of the maximum load of the soil). That definition leads to limit values for the cumulative emission of metals from a secondary building material in a diffusion test.

The standard diffusion protocol that is imposed by Vlarea is the Dutch standard NEN 7345. This test, however, takes 64 days. This is too long to allow a fast evaluation of candidate secondary building materials. Therefore, a two-step leaching test is under development (prEN 12457-3). This test takes only 24 h. It is the latter that has been used to evaluate Graveliet® and its use as a shaped material.

In Table 1, the behaviour of 3 different Graveliet types in the two-step leaching test is compared with the limit values. It is clear that all three fulfil the legal criteria for the use as shaped material. It must be noticed, however, that both Mo and SO4 " leach more extensively than the suggested reference values which are based on Dutch legislation. Since it is expected, that in the near-future, limit values will also be imposed for Mo, the Mo leaching is under further investigation.

910 LEAD-ZINC 2000

Table 1: Leachability (mg/kg) According to the Two-step Leaching Test for Graveliet® Produced in the Pilot Plant (from (5))

Vlarea parameters

As Cd Cr Cu Ni Pb Zn

Non-Vlarea parameters

Mo Sb

so42-

Limit values (mg/kg)

0.8 0.03 0.5 0.5 0.75 1.3 2.8

Reference values

(mg/kg) 0.2 0.1 540

98054

<0.12 < 0.025 < 0.056

0.13 <0.05 <0.1 <0.1

2.72 <0.1 10053

98055

<0.13 < 0.025 < 0.059

0.31 <0.05 <0.1 <0.1

1.39 <0.13 6104

98056

<0.12 < 0.025 < 0.051

0.48 <0.05 <0.1 <0.1

0.78 <0.21 4832

Mechanical Properties

A very convenient and simple way to compare the mechanical strength of two materials is to compare their hardness. The testing device used to measure the hardness is a Zwick type 3117 apparatus. The hardness is measured by pushing a needle into the material with a certain force (fixed value). The depth of the indentation is related to the hardness (hardness scale: Shore D). Materials with a hardness between 70-80 Shore D are considered to be very hard.

The hardness is not an absolute measure of the mechanical properties of the material, but it is a very convenient way to compare materials of different types. The influence of two of the investigated parameters (curing regime and amount of water-soluble zinc in goethite) on the obtained hardness will be discussed below. Also, an indication of the wear behaviour of Graveliet* will be given.

Hardness Evolution in Time as a Function of the Curing Regime

Curing is the hardening of the material under a well-conditioned atmosphere where both temperature and humidity can be controlled. In concrete manufacturing, curing of products is very important.

In Figure 7, the hardness is given as a function of time for one mixture but with three different hardening/curing regimes applied. All three samples were first put into a conditioned chamber for one day. After that, the conditions were changed for two of the samples. The upper curve is for the material that was cured for several days in a conditioned atmosphere (100% relative humidity, 35°C). The other samples were cured under water and in open-air.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 911

40 ·

30 ■ 10 100 1000

tim* (h)

Figure 7 - Hardness (Shore D) as a Function of Time for Three Different Curing Regimes (16/0: 5 Days in Curing Room + Outside; 16/1: 1 Day in Curing Room + Outside; 16/2: 1 Day

in Curing Room + Under water)

It is clear that the hardening rate after the first day is higher for the first material than for the other two samples. The final hardness, however, is almost the same. The material cured in any of the three ways is not brittle. Material that is immediately dried in the open-air, though, does show a pronounced brittleness.

Hardness as a Function of the Content of Soluble Zinc

The presence of zinc in cement causes the hydraulic reaction to be retarded. The cement reaction cannot take place at all when too much zinc is present.

Since the goethite contains both an amount of water insoluble zinc and an amount of water-soluble zinc, it was decided to investigate the influence of the zinc present in the goethite on the hardness. The quantity of soluble zinc depends on the degree of washing of the filter cake that is obtained after the goethite precipitation step in the zinc process.

In Figure 8, the hardness is given for five mixtures in which the amount of water soluble zinc in the goethite was artificially altered from 0.47 wt% up to 3.70 wt%. From the figure, it is clear that the presence of zinc indeed leads to a slower reaction mechanism. However, the final (28 day) hardness of the brick is not altered and is around 80 Shore D units.

912 LEAD-ZINC 2000

Figure 8 - Hardness as a Function of the Amount of Water-soluble Zinc Present in the Goethite

Wear Behaviour of the Pellets (Los Angeles Test)

The reference method to characterise the mechanical properties of gravel (or granulated material), that is used in concrete or road construction, is the Los Angeles test. In this test, a given quantity of material (with a defined size distribution) is put into a ball mill. After a fixed number of rotations, the amount of material that falls through a fine sieve is measured. On the basis ofthat analysis, the material is characterised. Six categories are defined (A = best to F = worst). The Graveliet material that has been tested up to now, falls into the categories E or F. There are strong indications, though, that the wear properties can be improved.

Other Physical Properties

Porosity

The method used to evaluate the porosity qualitatively is very simple. After weighing, aggregates of a known size are put into a vacuum chamber for 1 hour. Then they are immersed in water for 24 hours. After that, the weight is measured again. The water uptake in wt% of the original weight is taken as a measure of the porosity.

The porosity expressed as the wt% water uptake for the pilot material is in the order of only 2%. An explanation for this lies in the forming process. Indeed, during the extrusion of the material, the extrusion chamber is vacuum sucked and, moreover, the material is compressed when it leaves via the mouth of the extruder.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 913

The water uptake of Graveliet has also been compared with the Hg-porosity. The water uptake is sometimes even more than 100% of the measured Hg-porosity, as shown in Figure 9.

as

1 140 120 100 80 60 40 20 -0 —i i

20 30

time (days)

- i —

40

— i —

50 60

Figure 9 - Water Absorption (% of Hg Porosity) as a Function of Time for Brick 25; Taken From (5)

An analysis of the pores and pore structure by an independent research laboratory demonstrated that the size distribution of the pores within the Graveliet product is not optimal and could give rise to problems with the freeze/thaw behaviour. This could be overcome by the addition of pore-formers and is this aspect under further investigation.

Permeability

The water permeability of Graveliet has been compared with that of an ordinary brick. Although the permeability for the baked clay brick decreases only slightly after 5 days (Figure 10), this is not the case for the Graveliet brick. For the latter, the permeability decreases to almost zero after only 2 days, as is shown in Figure 11.

Figure 10 - Water permeability of a reference brick; taken from (5)

914 LEAD-ZINC 2000

Waterpermaebiliteit IIM0809 (monsteropp. 3,14 cm1; dikte 10 mm)

5 mwk

Figure 11 - Water Permeability of a Graveliet Brick; Taken From (5)

From an eco-technical point of view, the low permeability of Graveliet is a good thing. As the leaching of the metals will be strongly hindered.

CONCLUSIONS

The Graveliet process is an important step towards minimization of waste in the main production route of zinc metal. The eco-technical and constructional properties of the Graveliet® are very promising, but have to be proved by semi-industrial applications.

From an environmental point of view, Graveliet® solves a huge stockpiling problem for the zinc industry. The leaching behaviour of the pellets is very good for all metals. However, although the amounts leached are very low, Mo poses a problem since the background levels in soil are almost negligible. For the latter element, a solution to the leaching problem is under investigation.

From the point of view of the building industry, the use of Graveliet® as a concrete filler is very much welcomed, since in the near future, the use of primary gravel will become much more restricted. The market for Graveliet® pellets would become several times bigger, if its mechanical properties could be more improved, especially its Los Angeles test value.

Overall, the Graveliet project was helpful as it allowed Union Miniere to investigate, on a sufficiently large scale, the possibilities of the Graveliet® process. It can be concluded that, to a large extent, the assumptions about the process and the product that were formulated at the beginning of the project have been confirmed.

ACKNOWLEDGEMENT

The present research was partially funded by the European Commission within the framework of the Life/Environment programme.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 915

REFERENCES

1. K.J. Torfs and J. Vliegen, "The Union Miniere Goethite Process: Plant Practice and Future Prospects", Iron Control and Disposal, J.E. Dutrizac and G.B. Harris, Eds., Canadian Institute of Mining, Metallurgy and Petroleum, Montreal, Canada, 1996, 135-146.

2. T.D.W. Robinson, "Improvements in the Treatment of Hazardous Waste", EP031667, 08 July 1981.

3. J. Vliegen and A. Vandenbranden, "Process for the Conversion of Iron Bearing Residues into a Synthetic Rock", EP0931031, 28 July 1999.

4. Vlaams Reglement inzake afvalvoorkoming en -beheer (VLAREA), Openbare Afvalstoffenmaatschappij voor het Vlaamse Gewest (OVAM), 1998 (in Dutch).

5. V. Dutre, R. Dreesen, J. Dresselaers, K. Vrancken and B. Laethem, "Milieuhygienisch onderzoek naar de valorisatiemogelijkheden van Graveliet® als secundaire grondstof -Materiaalkarakterisering en uitloogonderzoek", Vito, Mol (Belgium), Jan. 1999 (in Dutch).

6. V. Dutre, K. Vrancken and B. Laethem, "Milieuhygienisch onderzoek naar de valorisatiemogelijkheden van Graveliet® als secundaire grondstof, Vito, Mol (Belgium), 1998 (in Dutch).

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 917

A MINERALOGICAL STUDY OF JAROFIX PRODUCTS FOR THE STABILIZATION OF JAROSITE RESIDUES FOR DISPOSAL

T.T. Chen and J.E. Dutrizac CANMET

555 Booth Street Ottawa, Ontario, Canada K1A 0G1

ABSTRACT

Noranda Inc., CEZinc presently employs a cement-stabilization (jarofix) process to stabilize its jarosite residues for direct disposal. In this process, Portland cement partly decomposes Na-jarosite to form Ca6Fe2(SC>4)3(OH)i2.nH20, ferric hydroxide and sodium sulphate which further reacts to form gypsum; approximately one-third of the Na-jarosite reacts in the current commercial product. The cured jarofix products also contain minor amounts of various Ca-Al-Fe silicate-sulphate-hydrate phases and traces of calcite. Together with gypsum and a Ca-Fe sulphate phase, these species bond the various particles together, imparting mechanical strength and reducing the permeability. Water-soluble Zn and Mg are stabilized in the Ca-Al-Fe silicate-sulphate-hydrate phases. The freshly cured products usually contain small amounts of moisture and traces of residual Portland cement, which provides additional alkalinity for the neutralization of any residual acid species. Jarofix products stored for more than six years in the laboratory, after curing, are usually dry and exhibit the same micro-textures as those of the freshly cured products. However, the stored products are depleted in the Ca6Fe2(S04)3(OH)i2.nH20 and Portland cement phases, and are enriched in gypsum and calcite which reflects the highly alkaline environment of the jarofix products. The aged products also have increased amounts of the Ca-Al-Fe silicate-sulphate-hydrate phases which have elevated Zn and Mg contents. This implies a further immobilization of residual water-soluble Zn and Mg during storage.

Lead-Zinc 2000 Edited by J.E. Dutrizac, J.A. Gonzalez,

D.M. Henke, S.E. James, and A.H.-J Siegmund TMS (The Minerals, Metals & Materials Society), 2000

918 LEAD-ZINC 2000

INTRODUCTION

Iron is a major impurity in zinc concentrates, and during the processing of the concentrates by the roast-leach-electrolysis option, the iron dissolves and is conveniently precipitated as NH4-jarosite [NH4Fe3(S04)2(OH)6] or Na-jarosite [NaFe3(S04)2(OH)ö]. The jarosite precipitates are filtered, washed and disposed to jarosite ponds (1). A relatively large volume of residue is generated, and hence, a large land space is required for jarosite disposal. Furthermore, the residues contain trace amounts of Zn, Cu, Cd, Pb and As. For these reasons, a reliable low-cost method for the safe disposal of jarosite residues becomes a priority, and many approaches to the problem have been considered. One of the promising options is to stabilize the jarosite residue with Portland cement, and this approach has been developed by Noranda Inc., CEZinc in co-operation with the Noranda Technology Centre. The technology has been implemented by Noranda Inc., CEZinc at its commercial operation in Valleyfield, Quebec since December 1998, and is designated as the jarofix process. The cured cement-jarosite mixtures themselves are defined as jarofix products (2,3).

In the jarofix process, the jarosite residue is mixed with an appropriate quantity of Portland cement; an amount of water is then added to the mixture to give an acceptable workability. For test purposes, the mixtures are blended and cast into cylindrical molds; they are then cured at room temperature for at least 56 days. Jarofix products cure quickly, and have good physical strength and a low hydraulic permeability; they do not undergo a significant volume change from that of the initial jarofix mixture. Both washed and unwashed NFLt-jarosite and Na-jarosite residues from commercial zinc plants have been evaluated, and various proportions of cement to jarosite residue have been tested. As well, the effects of the moisture content on the properties of the jarofix products have been assessed. The influence of temperature on the curing of jarofix products and on the cured jarofix products themselves has also been evaluated. For Canadian climatic conditions, freeze/thaw tests were performed with satisfactory results. Various engineering data, such as bulk density, moisture content, hydraulic conductivity and unconfined compressive strength, have been obtained for the jarofix products. An extensive series of protocol tests to demonstrate the environmental stability of the jarofix products was carried out. These tests include the Environment Canada EPS Proposed Evaluation Protocol for Cement-based Solidified Waste (4) and the United States EPA Toxicity Characteristic Leaching Procedure (TCLP), Method 1312 Leaching Procedure, Equilibrium Extraction, Dynamic Leach Test, and Acid Neutralization Capacity. Jarofix products pass all the relevant environmental criteria for waste disposal. Accordingly, it is concluded that the jarofix process is an economic option capable of generating a non-leachable product from zinc plant jarosite residues that is suitable for disposal.

In a previous study of laboratory-prepared samples, the authors investigated the mineralogy, micro-textures and chemical reactions of freshly cured jarofix products, to determine how the jarosite and the cement particles are bonded together (3). Since the initial mineralogical study, the preparation conditions for commercial operation have been defined and the jarofix process has been incorporated into normal plant operations. In order to evaluate the stability of the commercial jarofix products and to study possible long term transformations of the stabilized residues, the jarofix product generated in the current commercial operation and the products prepared previously in the laboratory, and stored for more than six years after curing, were investigated. The results of these recent mineralogical studies are reported in this paper.

Edited by J.E. Dutrizac, J.A. Gonzalez, D.M. Henke, S.E. James and A.H.-J. Siegmund 919

EXPERIMENTAL

In the commercial operation at Noranda Inc., CEZinc, Na-jarosite residue and gypsum filter cake are mixed with Portland cement, a small amount of lime and a quantity of water to give a workable mixture. The mixture is blended and subsequently trucked to a nearby disposal site. The commercial sample studied was generated in October 1999 during a period of normal operation at Noranda Inc., CEZinc. For study purposes, the mixture was cast into cylindrical molds 4 inches in diameter x 8 inches high, and was cured for more than three months prior to mineralogical investigation. For comparative purposes, jarofix products, prepared in the laboratory and stored for >4 years or >6 years after curing, were also analyzed. These curing conditions simulate, but do not exactly duplicate, those encountered at the commercial disposal site. Polished sections of the jarofix products were prepared, and these were examined using scanning electron microscopy (SEM) coupled with energy dispersive X-ray analysis (EDX) to identify the various phases and their micro-textures. Where the particle sizes were sufficiently large, quantitative electron microprobe analyses were carried out, but with the understanding that many of the analyses were biased by intimately admixed phases. The samples were also analyzed by X-ray powder diffraction to confirm the phases present and to estimate the extent of jarosite decomposition. The details of the mineralogical procedures are available in the literature (5).

RESULTS AND DISCUSSION

Jarofix Product Produced in the Commercial Noranda Inc., CEZinc Operation

It has been determined that NFU-jarosite reacts with Portland cement to form ammonium sulphate which decomposes to ammonia gas in the strongly alkaline jarofix environment (3). The evolution of NH3 gas effectively precludes the use of jarofix technology for NH4-jarosite residues. Consequently, Noranda Inc., CEZinc converted its former NH4-jarosite process to Na-jarosite precpitation. Sodium jarosite residues from the zinc plant consist mainly of Na-jarosite together with trace amounts of zinc ferrite, quartz and hematite; the particle sizes are generally <4 urn in diameter. Portland cement consists mostly of Ca silicate together with a minor quantity of Ca-Al silicate which contains minor to trace amounts of Fe, Na and K. Traces of gypsum and a Ca-Al-Fe oxide phase are also present; their particle sizes are generally 2-25 μτη. In contrast to conventional concrete technology where cement is mixed with much coarser particles of sand and gravel, jarofix technology involves the mixing of cement with much finer particles of jarosite residue to stabilize the jarosite, zinc ferrite, quartz, etc.

The commercial jarofix product was prepared by mixing Na-jarosite residue, gypsum filter cake and a small amount of lime with Portland cement. An appropriate amount of water was then added to provide an acceptable workability. The mixture was thoroughly blended and cured for more than three months prior to study. The product after three months of curing still contains a small amount of moisture; it has good compressive strength and low permeability. X-ray powder diffraction analyses indicate that the cured product consists of major amounts of gypsum and Na-jarosite together with minor amounts of Ca6Fe2(S04)3(OH)i2.nH20 (an ettringite-type compound) and possibly minor amounts of amorphous compounds. Figure 1 shows the unpolished fracture surface of the freshly broken commercial jarofix product. The large crystals are gypsum; the tiny bright grains are ZnFe2Ü4, and the numerous light-grey grains are partly reacted Na-jarosite. As shown, the jarofix product still possesses a moderate degree of porosity. Presumably, the pores were initially occupied by fluid or water during jarofix preparation, and subsequently, the fluid or

920 LEAD-ZINC 2000

water reacted with the cement and jarosite. Figure 2 shows the same product in a polished cross-section mount. The cement-jarosite residue-gypsum filter cake mixture has formed a coherent mass. The smaller (1-6 μπι) particles are partly reacted Na-jarosite; the large lath-like crystals are gypsum, and the large dark grains are various Ca-Al-Fe silicate-sulphate-hydrate cement reaction products. The rarer tiny bright particles are zinc ferrite or hematite. Most of the cement particles have reacted, and nearly all the Na-jarosite particles have partly reacted, despite the fact that only -30% of the original mass of Na-jarosite has been converted to other species. In most instances, as shown in Figure 3, the cement has completely reacted with water and Na-jarosite to form Ca-Al-Fe oxide and various Ca-Al-Fe silicate-sulphate-hydrate phases. Also, the initial large fragments of cement have disintegrated into smaller particles of new compounds which are dispersed among the reacted jarosite particles. Residual cement is rarely present, and the few residual cement particles are extensively rimmed by the cement reaction products. The residual cement particles likely will react gradually with water and Na-jarosite, and eventually will form small particles of new compounds. These fine-grained new compounds, together with gypsum and Ca-Fe sulphate, constitute the interstitial material between the partly reacted jarosite particles. The fine-grained interstitial phases bond the various particles and give the jarofix product its mechanical strength and low permeability.

Sodium jarosite and gypsum are the major species in the commercial Noranda Inc., CEZinc jarofix product. The finer gypsum particles are formed by the reaction of Na-jarosite and cement, whereas the large crystals are from the original gypsum filter cake. Trace amounts of zinc ferrite, hematite and quartz are present in the jarofix product; they originate from the jarosite residue. Trace amounts of calcium hydroxide which is a cement reaction product, of calcite which forms from the air-carbonation of of Ca(OH)2, and of Ca-Al-Fe oxide (~ Ca4AbFe085 or Ca2(Al,Fe)2C>5) which is a common cement decomposition species are also detected in the jarofix product.

Sodium jarosite usually occurs as small (1-4 urn) euhedral crystals in commercial jarosite residues, and the composition of the original unreacted Na-jarosite crystals is relatively uniform. Figure 4 shows the morphology of the reacted Na-jarosite particles in the commercial jarofix product. Although the external morphology of the particles seems to remain unchanged during the cement-jarosite reaction, the jarosite has undergone a significant degree of chemical alteration. The slightly darker phase (matrix) is mainly residual Na-jarosite whereas the brighter phase is an Fe oxide-rich phase mixed with Ca6Fe2(S04)3(OH)i2.nH20 (dark). The reaction has taken place throughout the mass of the jarosite, and the reaction sites appear to be randomly distributed.

Table I presents the average electron microprobe-determined composition of the reacted jarosite particles in the commercial jarofix product, as well as the initial composition of the sodium jarosite in the jarosite residue (i.e., the unreacted Na-jarosite). Because of the intimate mixture of the residual jarosite and its reaction products, the values in Table I provide only the average overall compositions of the reacted particles. Commercial jarosite commonly contains trace amounts of Al and Zn, which are present in the jarosite structure (6); however, Ca and Si are virtually absent. The average compositions given in Table I indicate that the Fe content of the reacted jarosite particles increases slightly and that the sulphate content decreases significantly, relative to the unreacated jarosite. The reaction products are significantly enriched in Ca and Si, and this indicates that soluble silicate species from the cement diffuse into the jarosite particles and react therein. Significant is the fact that the Na present in the original Na-jarosite remains in the reaction product, likely occurring as Na2SÜ4 or substituting for Ca in the Ca6Fe2(S04)3(OH)i2.nH20 phase.