membrane operations · 2018. 6. 8. · 6 fouling in membrane processes 121 anthony g. fane, tzyy h....

Membrane Operations

Edited by

Enrico Drioli and Lidietta Giorno

Related Titles

Seidel-Morgenstern, A. (ed.)

Membrane ReactorsCase Studies to Improve Selectivity

and Yields

2009

ISBN: 978-3-527-32039-4

Koltuniewicz, A., Drioli, E.

Membranes in CleanTechnologiesTheory and Practice

2008

ISBN: 978-3-527-32007-3

Peinemann, K.-V., Pereira Nunes, S. (eds.)

Membrane TechnologyVolume 1: Membranes for Life Sciences

2007

ISBN: 978-3-527-31480-5

Peinemann, K.-V., Pereira Nunes, S. (eds.)

Membrane TechnologyVolume 2: Membranes for Energy

Conversion

2007

ISBN: 978-3-527-31481-2

Li, K.

Ceramic Membranesfor Separation and Reaction

2007

ISBN: 978-0-470-01440-0

Pereira Nunes, S., Peinemann, K.-V. (eds.)

Membrane Technologyin the Chemical Industry

2006

ISBN: 978-3-527-31316-7

Sammells, A. F., Mundschau, M. V. (eds.)

Nonporous InorganicMembranesfor Chemical Processing

2006

ISBN: 978-3-527-31342-6

Freeman, B., Yampolskii, Y.,Pinnau, I. (eds.)

Materials Science ofMembranes for Gas andVapor Separation2006

ISBN: 978-0-470-85345-0

Ohlrogge, K., Ebert, K. (eds.)

MembranenGrundlagen, Verfahren und

industrielle Anwendungen

2006

ISBN: 978-3-527-30979-5

Membrane Operations

Innovative Separations and Transformations

Edited byEnrico Drioli and Lidietta Giorno

The Editors

Prof. Enrico DrioliUniversity of CalabriaInstitute on Membrane TechnologyVia P. Bucci 17 /C87030 Rende (CS)Italy

Prof. Lidietta GiornoUniversity of CalabriaInstitute on Membrane TechnologyVia P. Bucci 17 /C87030 Rende (CS)Italy

All books published by Wiley-VCH are carefullyproduced. Nevertheless, authors, editors, andpublisher do not warrant the information containedin these books, including this book, to be free oferrors. Readers are advised to keep in mind thatstatements, data, illustrations, procedural details orother items may inadvertently be inaccurate.

Library of Congress Card No.: applied for

British Library Cataloguing-in-Publication DataA catalogue record for this book is available from theBritish Library.

Bibliographic information published bythe Deutsche NationalbibliothekThe Deutsche Nationalbibliothek lists thispublication in the Deutsche Nationalbibliografie;detailed bibliographic data are available on theInternet at http://dnb.d-nb.de.

# 2009 WILEY-VCH Verlag GmbH & Co. KGaA,Weinheim

All rights reserved (including those of translation intoother languages). No part of this book may bereproduced in any form – by photoprinting,microfilm, or any other means – nor transmitted ortranslated into a machine language without writtenpermission from the publishers. Registered names,trademarks, etc. used in this book, even when notspecifically marked as such, are not to be consideredunprotected by law.

Composition Thomson Digital, Noida, IndiaPrinting Betz-Druck GmbH, DarmstadtBookbinding Litges & Dopf GmbH, HeppenheimCover Design Formgeber, Eppelheim

Printed in the Federal Republic of GermanyPrinted on acid-free paper

ISBN: 978-3-527-32038-7

Contents

List of Contributors XVIIIntroduction XXIII

Part One Molecular Separation 1

1 Molecular Modeling, A Tool for the Knowledge-Based Designof Polymer-Based Membrane Materials 3Dieter Hofmann and Elena Tocci

1.1 Introduction 31.2 Basics ofMolecularModeling of Polymer-BasedMembraneMaterials 51.3 Selected Applications 71.3.1 Hard- and Software 71.3.2 Simulation/Prediction of Transport Parameters and Model

Validation 81.3.2.1 Prediction of Solubility Parameters 91.3.2.2 Prediction of Diffusion Constants 91.3.3 Permeability of Small Molecules and Free-Volume Distribution 121.3.3.1 Examples of Polymers with Low Permeability of Small Molecules

(e.g., PO2� 50 Barrer) 131.3.3.2 Examples of Polymers with High Permeability of Small Molecules

(e.g., 50 Barrer�PO2� 200 Barrer) 131.3.3.3 Examples of Polymers with Ultrahigh Permeability of Small Molecules

(e.g., PO2� 1000 Barrer) 141.4 Summary 16

References 17

2 Polymeric Membranes for Molecular Separations 19Heru Susanto and Mathias Ulbricht

2.1 Introduction 192.2 Membrane Classification 19

Membrane Operations. Innovative Separations and Transformations. Edited by Enrico Drioli and Lidietta GiornoCopyright � 2009 WILEY-VCH Verlag GmbH & Co. KGaA, WeinheimISBN: 978-3-527-32038-7

V

2.3 Membrane Polymer Characteristics 222.3.1 Polymer Structure and Properties 222.3.2 Membrane Polymer Selection 232.3.2.1 Polymers for Porous Barriers 232.3.2.2 Polymers for Nonporous Barrier 252.3.2.3 Polymers for Charged Barrier 262.4 Membrane Preparation 262.4.1 Track-Etching of Polymer Films 262.4.2 Phase Separation of Polymer Solutions 272.4.3 Composite Membrane Preparation 302.4.4 Mixed-Matrix Membranes 322.5 Membrane Modification 322.6 Established and Novel Polymer Membranes for Molecular

Separations 342.6.1 Ultrafiltration 342.6.2 Reverse Osmosis and Nanofiltration 362.6.3 Pervaporation 372.6.4 Separations Using Ion-Exchange Membranes 382.7 Conclusion and Outlook 40

References 41

3 Fundamentals of Membrane Solvent Separation andPervaporation 45Bart Van der Bruggen

3.1 Introduction: Separation Needs for Organic Solvents 453.2 Pervaporation and Nanofiltration Principles 463.3 Membrane Materials and Properties for Solvent Separation 483.3.1 Solvent-Stable Polymeric Membrane Materials 483.3.2 Ceramic Membrane Materials 493.3.3 Solvent Stability 523.3.4 Structural Properties for Membranes in NF and PV 523.4 Flux and Separation Prediction 533.4.1 Flux Models in NF 533.4.2 Rejection in NF 553.4.3 Models for PV: from Solution-Diffusion to Maxwell–Stefan 563.4.4 Hybrid Simulations 573.5 Conclusions 58

References 58

4 Fundamentals of Membrane Gas Separation 63Tom M. Murphy, Grant T. Offord, and Don R. Paul

4.1 Introduction 634.2 Polymer Structure and Permeation Behavior 644.3 Membranes from Glassy Polymers: Physical Aging 694.4 Membranes from Rubbery Polymers: Enhanced CO2 Selectivity 75

VI Contents

4.5 Summary 79References 79

5 Fundamentals in Electromembrane Separation Processes 83Heinrich Strathmann

5.1 Introduction 835.2 The Structures and Functions of Ion-Exchange Membranes 845.2.1 Ion-Exchange Membrane Materials and Structures 855.2.2 Preparation of Ion-Exchange Membranes 855.2.2.1 Preparation Procedure of Heterogeneous Ion-Exchange

Membranes 865.2.2.2 Preparation of Homogeneous Ion-Exchange Membranes 865.2.2.3 Special Property Membranes 885.3 Transport of Ions in Membranes and Solutions 885.3.1 Electric Current and Ohms Law in Electrolyte Solutions 895.3.2 Mass Transport in Membranes and Solutions 915.3.2.1 The Driving Force and Fluxes in Electromembrane Processes 915.3.2.2 Electrical Current and Fluxes of Ions 915.3.2.3 The Transport Number and the Membrane Permselectivity 925.3.2.4 Membrane Counterion Permselectivity 935.3.2.5 Water Transport in Electrodialysis 945.4 The Principle of Electromembrane Processes 955.4.1 Electrodialysis 955.4.1.1 Electrodialysis System and Process Design 965.4.1.2 Electrodialysis Process Costs 1025.4.2 Electrodialysis with Bipolar Membranes 1075.4.2.1 Electrodialysis with Bipolar Membrane System

and Process Design 1085.4.2.2 Electrodialysis with Bipolar Membrane Process Costs 1105.4.3 Continuous Electrodeionization 1135.4.3.1 System Components and Process Design Aspects 1135.4.3.2 Operational Problems in Practical Application

of Electrodeionization 1155.4.4 Other Electromembrane Separation Processes 115

References 118

6 Fouling in Membrane Processes 121Anthony G. Fane, Tzyy H. Chong, and Pierre Le-Clech

6.1 Introduction 1216.1.1 Characteristics of Fouling 1216.1.2 Causes of Fouling 1236.1.3 Fouling Mechanisms and Theory 1256.1.4 Critical and Sustainable Flux 1256.1.5 Fouling and Operating Mode 1266.2 Low-Pressure Processes 126

Contents VII

6.2.1 Particulate Fouling 1266.2.2 Colloidal and Macrosolute Fouling 1276.2.3 Biofouling and Biofilms 1286.2.4 Case Studies 1286.2.4.1 Water Treatment and Membrane Pretreatment 1286.2.4.2 Membrane Bioreactor (MBR) 1296.3 High-Pressure Processes 1306.3.1 Particulate and Colloidal Fouling 1306.3.2 Biofouling 1326.3.3 Scale Formation 1336.3.4 Cake-Enhanced Osmotic Pressure 1356.4 Conclusions 136

References 136

7 Energy and Environmental Issues and Impacts of Membranesin Industry 139William J. Koros, Adam Kratochvil, Shu Shu, and Shabbir Husain

7.1 Introduction 1397.2 Hydrodynamic Sieving (MF and UF) Separations 1417.3 Fractionation of Low Molecular Weight Mixtures

(NF, D, RO, GS) 1427.4 Reverse Osmosis – The Prototype Large-Scale Success 1447.5 Energy-Efficiency Increases – A Look to the Future 1457.5.1 Success Stories Built on Existing Membrane Materials

and Formation Technology 1467.5.2 Future Opportunities Relying Upon Developmental Membrane

Materials and Formation Technology 1497.5.2.1 High-Performance Olefin–Paraffin Separation Membranes 1497.5.2.2 Coal Gasification with CO2 Capture for Sequestration 1547.6 Key Hurdles to Overcome for Broadly Expanding

the Membrane-Separation Platform 1587.7 Some Concluding Thoughts 160

References 161

8 Membrane Gas-Separation: Applications 167Richard W. Baker

8.1 Industry Background 1678.2 Current Membrane Gas-Separation Technology 1678.2.1 Membrane Types and Module Configurations 1688.2.1.1 Hollow Fine Fiber Membranes and Modules 1698.2.1.2 Capillary Fiber Membranes and Modules 1708.2.1.3 Flat-Sheet Membranes and Spiral-Wound Modules 1708.2.2 Module Size 1708.3 Applications of Gas-Separation Membranes 1718.3.1 Nitrogen from Air 171

VIII Contents

8.3.2 Air Drying 1738.3.3 Hydrogen Separation 1758.3.4 Natural-Gas Treatment 1788.3.4.1 Carbon-Dioxide Separation 1798.3.4.2 Separation of Heavy Hydrocarbons 1828.3.4.3 Nitrogen Separation from High-Nitrogen Gas 1828.3.5 Vapor/Gas Separations in Petrochemical Operations 1838.4 Future Applications 1868.4.1 CO2/N2 Separations 1868.4.2 CO2/H2 Separations 1888.4.3 Water/Ethanol Separations 1898.4.4 Separation of Organic Vapor Mixtures 1918.5 Summary/Conclusion 191

References 192

9 CO2 Capture with Membrane Systems 195Rune Bredesen, Izumi Kumakiri, and Thijs Peters

9.1 Introduction 1959.1.1 CO2 and Greenhouse-Gas Problem 1959.1.2 CO2 Capture Processes and Technologies 1969.2 Membrane Processes in Energy Systems with CO2

Capture 1999.2.1 Processes Including Oxygen-Separation Membranes 1999.2.2 Precombustion Decarbonization Processes Including Hydrogen

and Carbon Dioxide Membrane Separation 2029.2.3 Postcombustion Capture Processes with Membrane

Separation 2059.3 Properties of Membranes for Hydrogen, Oxygen, and Carbon

Dioxide Separation 2069.3.1 Membranes for Oxygen Separation in Precombustion

Decarbonization and Oxy-Fuel Processes 2069.3.1.1 Flux and Separation 2069.3.1.2 Stability Issues 2079.3.2 Membranes for Hydrogen Separation in Precombustion

Decarbonization 2079.3.2.1 Microporous Membranes 2089.3.2.2 Dense Metal Membranes 2099.3.2.3 Stability Issues 2099.3.2.4 Dense Ceramic Membranes 2109.3.3 Membranes for CO2 Separation in Precombustion

Decarbonization 2119.3.4 CO2 Separation in Postcombustion Capture 2119.3.4.1 CO2 Separation Membranes 2119.3.4.2 Membrane Contactors for CO2 Capture 2129.4 Challenges in Membrane Operation 212

Contents IX

9.4.1 Diffusion Limitation in Gas-Phase and Membrane Support 2129.4.2 Membrane Module Design and Catalyst Integration 2149.5 Concluding Remarks 216

References 216

10 Seawater and Brackish-Water Desalination with MembraneOperations 221Raphael Semiat and David Hasson

10.1 Introduction: The Need for Water 22110.2 Membrane Techniques in Water Treatment 22110.3 Reverse-Osmosis Desalination: Process and Costs 22610.3.1 Quality of Desalinated Water 22810.3.2 Environmental Aspects 22910.3.3 Energy Issues 23010.4 Treatment of Sewage and Polluted Water 23210.4.1 Membrane Bioreactors 23410.4.2 Reclaimed Wastewater Product Quality 23410.5 Fouling and Prevention 23510.5.1 How to Prevent 23610.5.2 Membrane Cleaning 23710.6 R&D Directions 23710.6.1 Impending Water Scarcity 23710.6.2 Better Membranes 23710.6.3 New Membranes-Based Desalination Processes 23810.7 Summary 240

References 240

11 Developments in Membrane Science for DownstreamProcessing 245João G. Crespo

11.1 Introduction 24511.1.1 Why Membranes for Downstream Processing? 24511.2 Constraints and Challenges in Downstream Processing 24611.2.1 External Mass-Transport Limitations 24611.2.2 Membrane Fouling 24711.2.3 Membrane Selectivity 24911.3 Concentration and Purification of Small Bioactive Molecules 24911.3.1 Electrodialysis 25011.3.2 Pervaporation 25111.3.3 Nanofiltration 25311.4 Concentration and Purification of Large Bioactive Molecules 25511.4.1 Ultrafiltration 25611.4.2 Membrane Chromatography 26011.5 Future Trends and Challenges 261

References 262

X Contents

12 Integrated Membrane Processes 265Enrico Drioli and Enrica Fontananova

12.1 Introduction 26512.2 Integrated Membrane Processes for Water Desalination 26612.3 Integrated Membrane Process for Wastewater Treatment 27112.4 Integrated Membrane System for Fruit-Juices Industry 27412.5 Integrated Membrane Processes in Chemical Production 27612.6 Conclusions 281

References 281

Part Two Transformation 285

13 Fundamental of Chemical Membrane Reactors 287Giuseppe Barbieri and Francesco Scura

13.1 Introduction 28713.2 Membranes 28913.3 Membrane Reactors 29413.3.1 Mass Balance 29413.3.2 Energy Balance 29613.4 Catalytic Membranes 30113.5 Thermodynamic Equilibrium in Pd-Alloy Membrane Reactor 30113.6 Conclusions 303

References 306

14 Mathematical Modeling of Biochemical Membrane Reactors 309Endre Nagy

14.1 Introduction 30914.2 Membrane Bioreactors with Membrane as Bioreactor 31014.2.1 Enzyme Membrane Reactor 31114.2.2 Whole-Cell Membrane Bioreactor 31214.3 Membrane Bioreactors with Membrane as Separation

Unit 31214.3.1 Moving-Bed Biofilm Membrane reactor 31214.3.2 Wastewater Treatment by Whole-Cell Membrane Reactor 31314.3.3 Membrane Fouling 31314.4 Mathematical Modeling of Membrane Bioreactor 31414.4.1 Modeling of Enzyme Membrane Layer/Biofilm

Reactor 31414.4.2 Concentration Distribution and Mass-Transfer Rates for Real

Systems 31814.4.3 Prediction of the Convective Velocity through Membrane with

Cake and Polarization Layers 32114.4.4 Convective Flow Profile in a Hollow-Fiber Membrane 32314.4.4.1 Without Cake and Polarization Layers 32314.4.4.2 With Cake and Polarization Layer 324

Contents XI

14.4.5 Mass Transport in the Feed Side of the Hollow-Fiber MembraneBioreactor 325

14.5 Modeling of the MBR with Membrane Separation Unit 32714.5.1 Moving-Bed-Biofilm Membrane Reactor 32714.5.2 Submerged or External MBR Process 32714.5.3 Fouling in Submerged Membrane Module 32814.6 Conclusions and Future Prospects 328

References 332

15 Photocatalytic Membrane Reactors in the Conversion or Degradationof Organic Compounds 335Raffaele Molinari, Angela Caruso, and Leonardo Palmisano

15.1 Introduction 33515.2 Fundamentals on Heterogeneous Photocatalysis 33615.2.1 Mechanism 33615.2.2 Photocatalysts: Properties and New Semiconductor Materials

Used for Photocatalytic Processes 33615.2.2.1 Titanium Dioxide 33815.2.2.2 Modified Photocatalysts 33815.3 Photocatalytic Parameters 34015.4 Applications of Photocatalysis 34115.4.1 Total Oxidations 34115.4.2 Selective Oxidations 34315.4.3 Reduction Reactions 34415.4.4 Functionalization 34415.4.5 Hydrogen Production 34515.4.6 Combination of Heterogeneous Photocatalysis with Other

Operations 34615.5 Advantages and Limits of the Photocatalytic Technologies 34615.6 Membrane Photoreactors 34815.6.1 Introduction 34815.6.2 Membrane Photoreactor Configurations 34815.6.2.1 Pressurized Membrane Photoreactors 34915.6.2.2 Sucked (Submerged) Membrane Photoreactors 34915.6.2.3 Membrane Contactor Photoreactors 35015.6.3 Parameters Influencing the Photocatalytic Membrane

Reactors (PMRs) Performance 35215.6.4 Future Perspectives: Solar Energy 35315.7 Case Study: Partial and Total Oxidation Reactions

in PMRs 35415.7.1 Degradation of Pharmaceutical Compounds in a PMR 35415.7.2 Photocatalytic Production of Phenol from Benzene

in a PMR 35715.8 Conclusions 358

References 358

XII Contents

16 Wastewater Treatment by Membrane Bioreactors 363TorOve Leiknes

16.1 Introduction 36316.2 Membranes in Wastewater Treatment 36416.2.1 Background 36416.2.2 Membranes Applied to Wastewater Treatment 36516.3 Membrane Bioreactors (MBR) 36816.3.1 Membrane-Bioreactor Configurations 36816.3.1.1 Membrane Materials and Options 36816.3.1.2 Process Configurations 37116.3.2 Membrane-Bioreactor Basics 37216.3.3 Membrane Fouling 37416.3.3.1 Understanding Fouling 37416.3.3.2 Dealing with Fouling 37616.3.3.3 Cleaning Fouled Membranes 37816.3.4 Defining Operating Conditions and Parameters

in MBR Processes 37916.3.4.1 Biological Operating Conditions 37916.3.4.2 Membrane Filtration Operation 38116.3.4.3 Optimizing MBR Operations 38316.4 Prospects and Predictions of the MBR Process 38416.4.1 Developments and Market Trends 38416.4.2 An Overview of Commercially Available Systems 38616.4.2.1 Flat-Sheet MBR Designs and Options 38816.4.2.2 Tubular/Hollow-Fiber MBR Designs and Options 388

References 391

17 Biochemical Membrane Reactors in Industrial Processes 397Lidietta Giorno, Rosalinda Mazzei, and Enrico Drioli

17.1 Introduction 39717.2 Applications at Industrial Level 39817.2.1 Pharmaceutical Applications 39917.2.2 Food Applications 40217.2.3 Immobilization of Biocatalysts on Membranes 40517.3 Conclusion 407

References 407

18 Biomedical Membrane Extracorporeal Devices 411Michel Y. Jaffrin and Cécile Legallais

18.1 General Introduction 41118.1.1 Use of Membranes in the Medical Field 41118.1.2 Historical Perspective 41118.2 Hemodialyzers 41318.2.1 Introduction 41318.2.2 Physical Principles of Hemodialysis 414

Contents XIII

18.2.3 Dialysis Requirements 41518.2.4 Mass Transfers in a Hemodialyzer 41618.2.4.1 Characterization of Hemodialyzers Performance 41618.2.5 Hemofiltration and Hemodiafiltration 41718.2.6 Various Types of Hemodialyzers 41818.2.6.1 Various Types of Membranes 41918.2.6.2 Optimization of Hemodialyzer Performance 42018.3 Plasma Separation and Purification by Membrane 42118.3.1 Introduction 42118.3.2 The Baxter Autopheresis C System for Plasma Collection

from Donors 42118.3.3 Therapeutic Applications of Plasma Separation 42218.3.3.1 Plasma Exchange 42318.3.3.2 Selective Plasma Purification by Cascade Filtration 42318.4 Artificial Liver 42618.4.1 Introduction 42618.4.2 Physical Principles 42618.4.3 Convection þ Adsorption Systems 42818.4.4 Diffusion þ Adsorption Systems 42818.4.5 Future of Artificial Livers 42918.4.6 Conclusions 430

References 430

19 Membranes in Regenerative Medicine and TissueEngineering 433Sabrina Morelli, Simona Salerno, Antonella Piscioneri, Maria Rende,Carla Campana, Enrico Drioli, and Loredana De Bartolo

19.1 Introduction 43319.2 Membranes for Human Liver Reconstruction 43419.3 Human Lymphocyte Membrane Bioreactor 43919.4 Membranes for Neuronal-Tissue Reconstruction 44019.5 Concluding Remarks 443

References 444

Part Three Membrane Contactors 447

20 Basics in Membrane Contactors 449Alessandra Criscuoli

20.1 Introduction 44920.2 Definition of Membrane Contactors 44920.3 Mass Transport 45220.4 Applications 45520.5 Concluding Remarks 460

References 460

XIV Contents

21 Membrane Emulsification: Principles and Applications 463Lidietta Giorno, Giorgio De Luca, Alberto Figoli, Emma Piacentini,and Enrico Drioli

21.1 Introduction 46321.2 Membrane Emulsification Basic Concepts 46521.3 Experimental Bases of Membrane Emulsification 46821.3.1 Post-Emulsification Steps for Microcapsules Production 47421.3.2 Membrane Emulsification Devices 47621.4 Theoretical Bases of Membrane Emulsification 47921.4.1 Torque and Force Balances 48021.4.2 Surface-Energy Minimization 48521.4.3 Microfluid Dynamics Approaches: The Shape of the Droplets 48621.5 Membrane Emulsification Applications 48821.5.1 Applications in the Food Industry 48821.5.2 Applications in the Pharmaceutical Industry 48921.5.3 Applications in the Electronics Industry 49021.5.4 Other Applications 49121.6 Conclusions 493

References 494

22 Membrane Contactors in Industrial Applications 499Soccorso Gaeta

22.1 Air Dehumidification: Results of Demonstration Tests with RefrigeratedStorage Cells and with Refrigerated Trucks 505

22.2 Refrigerated Storage Cells 50722.3 Refrigerated Trucks 50822.4 Capture of CO2 from Flue Gas 510

References 512

23 Extractive Separations in Contactors with One and Two ImmobilizedL/L Interfaces: Applications and Perspectives 5134Stefan Schlosser

23.1 Introduction 51323.2 Contactors with Immobilized L/L Interfaces 51623.3 Membrane-Based Solvent Extraction (MBSE) and Stripping

(MBSS) 51723.3.1 Case Studies 51923.4 Pertraction through BLME 52523.4.1 Case Studies 52623.5 Pertraction through SLM 52723.5.1 Case Studies 529

Contents XV

23.6 Comparison of Extractive Processes in HF Contactors andPertraction through ELM 529

23.7 Outlook 529References 531

Index 543

XVI Contents

List of Contributors

XVII


Richard W. BakerMembrane Technology andResearch, Inc.1360 Willow RoadMenlo Park, CA 94025USA

Giuseppe BarbieriUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

Loredana De BartoloUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

Rune BredesenSINTEF Materials and ChemistryP.O. Box 124Blindern0314 OsloNorway

Carla CampanaUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

and

University of CalabriaDepartment of Chemical Engineeringand MaterialsVia P. Bucci, cubo 45/A87030 Rende (CS)Italy

Angela CarusoUniversity of CalabriaDepartment of Chemical Engineeringand MaterialsVia P. Bucci, cubo 45/A87030 Rende (CS)Italy

Tzyy H. ChongNanyang Technological UniversitySingapore Membrane TechnologyCentreSchool of Civil and EnvironmentalEngineeringSingapore639798

João G. CrespoUniversidade Nova de LisboaFaculdade de Ciências e TecnologiaRequimte-CQFBDepartamento de Química2829-516 CaparicaPortugal

Enrico DrioliUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

and


Alessandra CriscuoliUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

Anthony G. FaneUniversity of New South WalesUNESCO Centre for Membrane Science& TechnologySchool of Chemical Sciences andEngineeringSydney, NSW 2052Australia

and

Nanyang Technological UniversitySingapore Membrane TechnologyCentreSchool of Civil and EnvironmentalEngineeringSingapore639798

Alberto FigoliUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

Enrica FontananovaUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

and


XVIII List of Contributors

Soccorso GaetaGVS S.P.A.Via Roma 5040069 Zola Predosa (Bo)Italy

Lidietta GiornoUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

David HassonTechnion – Israel Institute ofTechnologyStephen and Nancy Grand WaterResearch InstituteWolfson Chemical EngineeringDepartmentRabin Desalination LaboratoryTechnion CityHaifa, 32000Israel

Tzyy HaurNanyang Technological UniversitySingapore Membrane TechnologyCentreSchool of Civil and EnvironmentalEngineeringSingapore639798

Dieter HofmannGKSS Research CenterCenter for Biomaterial Developmentof the Institute of Polymer ResearchKantstr. 5514513 TeltowGermany

Shabbir HusainGeorgia Institute of TechnologySchool of Chemical & BiomolecularEngineeringAtlanta, GA 30332-0100USA

Michel Y. JaffrinUMR CNRS 6600Technological University of Compiegne60200 CompiegneFrance

William J. KorosGeorgia Institute of TechnologySchool of Chemical & BiomolecularEngineeringAtlanta, GA 30332-0100USA

Adam KratochvilPRISM MembranesAir Products and Chemicals, Inc.St. Louis, Mo 63146USA

Izumi KumakiriSINTEF Materials TechnologyP.O. Box 124Blindern0314 OsloNorway

Pierre Le-ClechUniversity of New South WalesUNESCO Centre for Membrane Science& TechnologySchool of Chemical Sciences andEngineeringSydney, NSW 2052Australia

List of Contributors XIX

Cécile LegallaisUMR CNRS 6600Technological University of Compiegne60200 CompiegneFrance

TorOve LeiknesNTNU - Norwegian University ofScience and TechnologyDepartment of Hydraulic andEnvironmental EngineeringS.P. Andersensvei 57491 TrondheimNorway

Giorgio De LucaUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

Rosalinda MazzeiUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

and

University of CalabriaDepartment of EcologyVia P. Bucci 6/B87036 Rende (CS)Italy

Raffaele MolinariUniversity of CalabriaDepartment of Chemical Engineeringand MaterialsVia P. Bucci87030 Rende (CS)Italy

Sabrina MorelliUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

T.M. MurphyThe University of Texas at AustinDepartment of Chemical EngineeringAustin, TX 78712USA

Endre NagyUniversity of PannoniaResearch Institute of Chemical andProcess EngineeringP.O. Box 1588201, VeszprémHungary

Grant T. OffordThe University of Texas at AustinDepartment of Chemical EngineeringAustin, TX 78712USA

Leonardo PalmisanoUniversity of PalermoDepartment of Chemical EngineeringProcesses and MaterialsSchiavello-Grillone PhotocatalysisGroupviale delle Scienze90128 PalermoItaly

XX List of Contributors

Don R. PaulThe University of Texas at AustinDepartment of Chemical EngineeringAustin, TX 78712USA

Thijs PetersSINTEF Materials TechnologyP.O. Box 124Blindern0314 OsloNorway

Emma PiacentiniUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

Antonella PiscioneriUniversity of CalabriaInstitute of Membrane TechnologyNational Research Council of ItalyITM-CNRVia P. Bucci, cubo 17/C87030 Rende (CS)Italy

and

University of CalabriaDepartment of Cell Biologyvia P. Bucci87030 Rende (CS)Italy

Maria RendeUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

and


Simona SalernoUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

4Stefan SchlosserSlovak University of TechnologyInstitute of Chemical andEnvironmental EngineeringRadlinského 9812 37 BratislavaSlovakia

Raphael SemiatTechnion – Israel Institute ofTechnologyWolfson Chemical EngineeringDepartmentRabin Desalination LaboratoryStephen and Nancy Grand WaterResearch InstituteTechnion CityHaifa, 32000Israel

List of Contributors XXI

Francesco ScuraUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

Shu ShuGeorgia Institute of TechnologySchool of Chemical & BiomolecularEngineeringAtlanta, GA 30332-0100USA

Heinrich StrathmannUniversity of StuttgartInstitute of Chemical TechnologyBöblingerstr. 7270199 StuttgartGermany

Heru SusantoUniversität Duisburg-EssenLehrstuhl für Technische Chemie II45117 EssenGermany

Elena TocciUniversity of CalabriaInstitute on Membrane Technology(ITM-CNR)Via P. Bucci, 17/C87030 Rende (CS)Italy

Mathias UlbrichtUniversität Duisburg-EssenLehrstuhl für Technische Chemie II45117 EssenGermany

Bart Van der BruggenK.U. Leuven, Department of ChemicalEngineeringSection Applied Physical Chemistry andEnvironmental TechnologyW. de Croylaan 463001 Heverlee (Leuven)Belgium

XXII List of Contributors

Introduction

Membrane processes are state of the art technologies in various industrial sectors,including gas separation, wastewater treatment, food processing and medical appli-cations.Modelling methodologies are contributing significantly to the knowledge-based

development of membrane materials and engineering.Micro-ultrafiltration and reverse osmosis are mature technologies for separations

based on molecular exclusion and solution-diffusion mechanisms, respectively.Cleaning andmaintenance procedures able to control fouling to an acceptable extenthave made these processes commercially suitable.Some of the largest plants for seawater desalination, wastewater treatment and gas

separation are already based on membrane engineering. For example, the AshkelonDesalination Plant for seawater reverse osmosis (SWRO), in Israel, has been fullyoperational since December 2005 and produces more than 100 million m3 ofdesalinated water per year. One of the largest submerged membrane bioreactorunit in the world was recently built in Porto Marghera (Italy) to treat tertiary water.The growth in membrane installations for water treatment in the past decade hasresulted in a decreased cost of desalination facilities, with the consequence that thecost of the reclaimed water for membrane plants has also been reduced.Membranes are growing significantly also in gas separation, for example, the

current market size of carbon-dioxide separation from natural gas is more than 70million Euro/year.Medical applications are among the most important in the membrane market,

with hemodialysis, blood oxygenators, plasma separation and fractionation beingthe traditional areas of applications, while artificial and bioartificial organs andregenerative medicine represent emerging areas in the field.Nanofiltration has achieved a good stage of development, gaining attention in

various applications for separations based on both molecular exclusion and chargeinteraction as well as on the solution-diffusion mechanism. In particular, nanofil-tration is considered among the most suitable technologies for solvent separation.More recent processes such as membrane reactors, membrane contactors, andmembranes in life science are also developing very rapidly. The optimal design of

XXIII


chemical transformation processes with control of reagent supply and/or productremoval through catalytic membranes and membrane reactors is one of the mostattractive solutions in process intensification. The catalytic action of biocatalysts isextremely efficient, selective and highly stereospecific when compared to conven-tional chemical catalysts. Membrane bioreactors are particularly attractive in termsof ecocompatibility, because they do not require additives, are able to operate atmoderate temperature and pressure, reduce the formation of by-products, whilepermitting the production of high valuable coproducts. This may allow challenges indeveloping new production lines moving towards zero discharge to be faced. Thedevelopment of catalytic membrane reactors for high-temperature applicationsbecame realistic more recently, with the development of high-temperature-resistantmembranes.The major market for membrane bioreactors is represented by wastewater treat-

ment with the use of submerged modules configuration. These are consideredamong the best available technologies by the European Directives on Environment.Membrane bioreactors are also applied in food, red and white biotechnology. Inthese cases, the external loop configuration is used.Membrane contactors, including membrane crystallizers and membrane emulsi-

fiers, are among the most recent membrane operations with growing interest invarious industrial sectors. For example, membrane emulsification has grown fromthe 1990s, when it was first developed in Japan, to nowadays with applications infood, chemical, pharmaceutical and cosmetic fields. In Europe, the research at theacademic level has achieved a thorough knowledge both from experimental andtheoretical points of view. This is fuelling the industrial interest towards themembrane emulsification technology, especially for those productions that involvelabile bioactive molecules.In general, nowadays the attention towards membrane science and technology is

increasing significantly. Drivers of this interest include the need for technologies toenable sustainable production, directives and regulations about the use of eco-friendly technologies, consumer demand for high-quality and safe products, publicconcern about environment, and stakeholder confidence in and acceptance ofadvanced technologies.Current initiatives recognize that a sustainable solution to the increasing demand

of goods and energy is in the rational integration and implementation of newtechnologies able to achieve concrete benefits for manufacturing and processing,substantially increasing process precision, reducing equipment size, saving energy,reducing costs, and minimizing environmental impact.Membranes and membrane processes are best suited in this context as their basic

aspects well satisfy the requirements of process intensification for a sustainableindustrial production. In fact, they are precise and flexible processing techniques,able to maximize phase contact, integrate conversion and separation processes, withimproved efficiency and with significantly lower energy requirements compared toconventional techniques.This multiauthor book highlights the current state and advances in membranes

and membrane operations referring to three major roles of the membrane: mole-

XXIV Introduction

cular separation, (bio)chemical transformation and phase contactors. Each topicincludes fundamentals and applications of membranes and membrane operations.The largest section is constituted bymembranes inmolecular separation, which is

the most traditional application of membranes. Significant advances of membranescience and technologies are expected in transformation processes and membranecontactors for conventional and innovative applications.

Introduction XXV

Part OneMolecular Separation

This Part will be focused on the fundamentals and applications of membranes andmembrane operations for separation at the molecular level. Both liquid (includingorganic solvents) and gaseous streams will be discussed.The book opens with a chapter on molecular modeling to highlight the powerful

instruments for designing appropriate membrane materials with predictedproperties.This is followed by a chapter on polymeric membranes that discusses the current

achievements and challenges on membranes for molecular separation in liquidphase.Subsequent individual chapters discussmembranes in organic solvent separation,

gas separation and electrochemical separation. A whole chapter is focused on thefundamentals of fouling molecular separation by membranes are completed by achapter focused on fouling. and another on energy and environmental issues.The application part of this section illustrates the membrane-assisted molecular

separation in (i) gases, with a separate chapter dedicated to the CO2 capture usinginorganic membrane; (ii) water desalination; (iii) downstream processing of biologi-cal products. Achapter on integratedmembrane operations illustrates new strategiesin water treatment and chemical production.Membrane separation in the medical field has been included in a chapter focused

on medical extracorporeal devices, which illustrates the use of membranes forseparation of biological fluids and for preparation of bioartificial organs able toaccomplish ex vivo biological transformation (Part headed Transformation).The overall aim of the molecular separation section is to illustrate the current

capability of membranes and membrane operations in assisting and governingmolecular separations and the future perspectives they offer for a more sustainableindustrial growth through innovative process design. Their implementation will leadto concrete benefits in manufacturing and processing, substantially shrinkingequipment size, boosting plant efficiency, saving energy, reducing capital costs,minimizing environmental impact, and using remote control and automation.


Membrane operations have the potential to replace conventional energy-intensiveseparation techniques, such as distillation and evaporation, to accomplish theselective and efficient transport of specific components, to improve the performanceof reactive processes and, ultimately, to provide reliable options for a sustainableindustrial growth.This is in line with the strategy of process intensification and it is expected to bring

substantial improvements in chemical and many other manufacturing and proces-sing industries.Many membrane operations are based on similar materials and structures, while

differing in the method by which they carry out the separation process. Step forwardinnovations can be promoted by appropriate integration of traditional membraneoperations (reverse osmosis, micro-, ultra- and nanofiltration, electrodialysis, perva-poration, etc.) among them and with innovativemembrane operations. In fact, whilebeing already widely used inmany different applications, they can be combined withnew membrane systems such as catalytic membrane reactors and membranecontactors. Nowadays, redesign of industrial production cycles by combining variousmembrane operations suitable for separation, conversion and concentration units isan attractive opportunity because of the synergic effects that highly integratedmembrane processes can promote.

1Molecular Modeling, A Tool for the Knowledge-BasedDesign of Polymer-Based Membrane MaterialsDieter Hofmann and Elena Tocci

1.1Introduction

Most important macroscopic transport properties (i.e., permeabilities, solubilities,constants of diffusion) of polymer-based membranes have their foundation inmicroscopic features (e.g., free-volume distribution, segmental dynamics, distribu-tion of polar groups, etc.) which are not sufficiently accessible to experimentalcharacterization. Here, the simulation of reasonably equilibrated and validatedatomistic models provides great opportunities to gain a deeper insight into thesemicroscopic features that in turn will help to develop more knowledge-basedapproaches in membrane development.The mentioned transport properties for small and medium-sized molecules in

polymers are decisive in many technologically important processes, for example, inbiotechnology and biomedicine, in pharmacological and chemical industries but alsoin integrated environmental protection. The respective penetrants can be anythingfrom rather small hydrogen or oxygen molecules to chemicals like benzene up torelatively large drug molecules.Membrane processes for the separation of gaseous and liquid mixtures are

important examples. In these cases there are already large numbers of applicablematerials and processes. Further improvements (mostly concerning better selectiv-ities at acceptably highpermeabilities), oftenneeding real jumps inperformance, are,however, still needed inmany cases. This applies, although in the opposite sense, alsoto barrier materials where permeations at least of certain types of molecules will beextremely small. Other areas concern biomaterials or material systems for thecontrolled release of drugs.More specific examples for the need to develop new materials with tailored

transport properties are:

. The separation of methane from higher hydrocarbons in natural gas for safer andmore economical transport through pipelines, or for better exploitation;


j3

. The design of packaging materials for conservation of fresh fruits and vegetables,which means good specific permeation and selectivity properties in order tomaintain a modified/controlled atmosphere;

. The control of migration of additives, monomers or oligomers, from packagingmaterials, for example, into food (important for the enforcement of a high level offood quality and safety) or other consumer products;

. The resistance of resins used in composites for aircraft construction to ageingcaused by water absorption;

. Small but continuous fuel loss by permeation through polymeric parts of the fuelsystem;

. Separation of CO2 from flue gases, and separation of NOx from vehicle emissions;

. Efficient and inexpensive proton-conducting membranes for fuel cells;

. components in polymer electronics (such as for light-emitting diodes or displaycomponents) with extremely low permeabilities for oxygen and water;

. Optimum controlled drug release systems, for example, for medical applications,cosmetics or agriculture;

. Transport problems in artificial or bio-hybrid organs;

. Optimum biocompatibility of polymers in contact with cells and blood;

. Optimum chemical degradation behavior (often to a large extent a water-perme-ation problem) for surgical sutures, scaffold materials for tissue engineering,degradable screws in orthopaedic surgery and so on.

In the near future, the use of multifunctional polymer-based materials withseparation/selective transport capabilities is also to be expected in the design ofproductionsystemswithintegratedenvironmentalprotectionor inthecombinationofchemical reactions and separation by attaching a catalytic functionality to the respec-tivematerial [1].Thus, thosemultifunctionalmaterials shouldcontributematerially tothe development of clean energy and/or energy saving and therefore sustainableproduction technologies. In connectionwith these perspectives, there is considerableinterest in new/modified polymer-based materials with tailored transport/catalyticproperties. Also, many sensor applications are based on controlled permeation.Amorphous polymers or respective composites with inorganic components are an

important class of materials to solve many of the above-mentioned problems.However, the design of these multifunctional materials, based on experimentationand correlative thinking alone is unreliable, time consuming, expensive and oftennotsuccessful. Systematicmultiscale computer-aidedmolecular design (CAMD) offers avery attractive alternative, insofar as these techniques allow for the very elaborateinvestigation of complexmaterial behaviorwith regard to the links between structure,dynamics and relevant properties of the discussed multifunctional polymer-basedmaterials on the length and time scales (from Angstroms to micrometers and frompicoseconds to milliseconds, respectively) which are most important for the pene-trant transport and other relevant processes (e.g., selective transport, separation,catalysis, biodegradation, sensor applications) of interest. In the present chapter,molecular modeling tools (i.e., quantum chemistry (QM), atomistic- and mesoscalemodeling)will be in the focus of interest. Consequently, themicroscopic properties to

4j 1 Molecular Modeling, A Tool for the Knowledge-Based Design Design of Polymer-Based

be related with macroscopically determined transport parameters are, for example,chain stiffness parameters, free volume and its distribution, mobility measures forchain segments, energy densities describing interactions of chain segments withpenetrants, microscopic effects of swelling and so on.Over the last 15 years particularly atomistic molecular modeling methods

have found widespread application in the investigation of small-molecule perme-ation [2–15].

1.2Basics of Molecular Modeling of Polymer-Based Membrane Materials

The permeation of small molecules in amorphous polymers is typically following thesolution diffusion model, that is, the permeability Pi of a feed component i can beenvisioned as the product of the respective solubility Si and constant of diffusionDi. Both parameters can be obtained experimentally and in principle also by atomisticsimulations.The molecular modeling of these polymers typically starts with the construction

of normally rectangular packing models. There, the related chain segments of therespective polymer will be arranged in realistic, that is, statistically possible, way.To do this, first the involved atoms are considered to be spheres of the respectiveatomic radius Ri (as obtainable from QM) and atomic weight mi. The bondedinteractions between atoms resulting in bonds, bond angles and conformationangles are then described by mechanic springs or torsion rods with springconstants related to, for example, experimentally known bond strengths. So-callednonbond interactions between atoms that either belong to different molecules orthat in one and the same molecule are further apart from each other than aboutthree bonds are considered via, for example, Lennard-Jones (to describe van derWaals interactions) and Coulomb potentials (to describe electrostatic interactions).The sum of all interatomic interactions written as the potential energy of a packingmodel is then called a forcefield. Forcefields form the core of all atomisticmolecular modeling programs. Equation 1.1 shows the principal structure of atypical forcefield for a system of N atoms with the Cartesian atomic positionvectors~ri.

V ~r1;~r2; . . . ;~rN� � ¼ X

Covalent bonds

Kbðl�l0Þ2 þX

Bond angles

KQðQ�Q0Þ2

þX

Dihedral angles

Kj½1þ cosðnj�dÞ�

þX

nonbonded atom pairs i;j

aijr12ij

!� bij

r6ij

!þ qiqj

e0errij

" #ð1:1Þ

with the following parameters:

1.2 Basics of Molecular Modeling of Polymer-Based Membrane Materials j5

l ¼ actual length of a bondl0 ¼ length of a bond in equilibriumKb¼ force constant for a bond length deformationQ ¼ actual value for a bond angleQ0¼ value for a bond angle in equilibriumKQ¼ force constant for a bond-angle deformationj ¼ actual value for a conformation anglen ¼ periodicity parameter in a conformation potentiald ¼ constant to fix trans-state in a conformation potentialKj¼ force constant for a conformation potentialRij ¼ distance between atoms i and j with (j� i)> 3aij ¼ constant describing repulsive interactions in the Lennard-Jones Potentialbij ¼ constant describing attractive interactions in the Lennard-Jones Potentialqi ¼ partial charge of the ith atome0 ¼ vacuum permittivityer ¼ dielectric constant.

The parameters l0, Kb, Q0, KQ, Kj, n, d, aij, bij, qi, qj and er belong tothe fit parameters, which can be determined by fitting of Equation 1.1 to a sufficientset of data calculated byQMand/or determined experimentally (e.g., X-ray scattering,IR spectroscopy, heats of formation). From a numeric point of view the pairinteraction terms (van der Waals and Coulomb) are most demanding. In thisconnection the typical size of polymer packing models is limited to typically3000–10 000 atoms (leading to lateral sizes of bulk models of a few nm), althoughin other connections now also models with up to 100 000 atoms have been used.Forcefields may be utilized in two directions:Model systems can be, on the one hand, subjected to a static structure optimiza-

tion. There, the fact is considered that the potential energy of a relaxed atomisticsystem (cf. Equation 1.1) should show a minimum value. Static optimization thenmeans that by suited numeric procedures the geometry of the simulated system ischanged as long as the potential energy reaches the next minimum value [16]. In thecontext of amorphous packing models, the main application for this kind ofprocedure is the reduction of unrealistic local tensions in a model as a prerequisitefor later molecular dynamic (MD) simulations.It is, on the other hand, possible to use the potential energy of a model system as

described by Equation 1.1 to calculate the forces~Fi acting on each atom of the modelvia the gradient operation:

~Fi ¼ �qVð~r1;~r2; . . . ;~rNÞ

q~rið1:2Þ

Then, Newtons equations of motion can be solved for every atom of the investigatedsystem:

~Fi ¼ mid2~riðtÞdt2

ð1:3Þ


The necessary starting positions ~rið0Þof the atoms are in the given case usuallyobtained from methods of chain-packing procedures (see below). The startingvelocities~við0Þ of all atoms are assigned via a suited application of the well-knownrelation between the average kinetic Ekin energy of a polyatomic system and itstemperature T:

Ekin ¼XNi¼1

12mi~vi

2 ¼ 3N�62

kbT ð1:4Þ

kBistheBoltzmannconstant.(3N� 6)isthenumberofdegreesoffreedomofanN-atommodel considering the fact that in the given case the center ofmass of thewholemodelwith its 6 translation and rotation degrees of freedom does not move during the MDsimulation. Using Equations 1.2–1.4 it is then possible to follow, for example, themotions of the atoms of a polymer matrix and the diffusive movement of imbeddedsmall penetrant molecules at a given temperature over a certain interval of time.Equation 1.3 represents a system of usually several thousand coupled differential

equations of second order. It can be solved only numerically in small time steps Dt viafinite-difference methods [16]. There always the situation at t þ Dt is calculated fromthe situation at t. Considering the very fast oscillations of covalent bonds, Dtmust notbe longer than about 1 fs to avoidnumerical breakdown connectedwith problemswithenergy conservation. This condition imposes a limit of the typical maximum simula-tion time that for the above-mentioned system sizes is of the order of several ns. Thelimited possible size of atomistic polymer packing models (cf. above) together withthis simulation time limitation also set certain limits for the structures and processesthat can be reasonably simulated. Furthermore, the limited model size demands theapplication of periodic boundary conditions to avoid extreme surface effects.The already mentioned limited lateral dimensions of packing models of just

several nmmakes it impossible to simulate complete membranes or other polymer-based samples. Therefore, on the one hand, bulk models are considered that aretypically cubic volume elements of a few nanometers side length that represent a partcut out of the interior of a polymer membrane (cf. Figure 1.1). On the other handinterface models are utilized, for example, for the interface between a liquid feedmixture and a membrane surface or between a membrane surface and an inorganicfiller (cf. Figure 1.2).

1.3Selected Applications

1.3.1Hard- and Software

The InsightII/MaterialsStudio/Discover software of Accelrys [18, 19] was utilizedfor the amorphous packing model construction, equilibration and the atomistic

1.3 Selected Applications j7

simulations. In most of the following examples the COMPASS forcefield wasapplied [20, 21].For data evaluation also self-programmed software (mostly in BTCL, Fortran, C)

was applied. Data production runs were performed on a 74 processor Opteron LinuxCluster, a SGI Origin 2100 and on SGI Onyx workstation.

1.3.2Simulation/Prediction of Transport Parameters and Model Validation

The quality of atomistic packing models is typically validated via comparisonsbetweenmeasured and simulated properties likewide-angleX-ray scattering (WAXS)

Figure 1.1 Atomic representation of a typical 3-dimensionalpacking model (thickness about 3 Å) starting with a single HyflonAD60X polymer chain. Atom colors: gray¼ carbon, red¼ oxygen,light blue¼ fluorine [15].

Figure 1.2 Atomic representation of a surface model of Pebax/30%KET with water [17].


curves, densities, transport parameters for small and medium sized penetrants. Inthe latter case both validating (if a polymer is already existing and experimentallycharacterized) and predictive (if a polymer has not been synthesized yet or if notransport parameters are available experimentally) applications are possible.

1.3.2.1 Prediction of Solubility ParametersHere, hitherto in most cases the transition-state method of Gusev and Suter [22, 23]was utilized to first determine calculated solubility values Scalc values. There, a fine3D-grid with a grid spacing of about 0.03 nm is layered over a completely refineddetailed-atomistic amorphous polymer bulk packing model (cf. Figure 1.1). Then asmall virtual test molecule of the intended kind (e.g., O2) in a united atomrepresentation is inserted in the polymer matrix at each lattice point of the grid.The resulting nonbonded interaction energy Eins between the inserted molecule andthe whole polymer matrix is calculated for each position of the respective insertedmolecule. Only the van der Waals interactions are considered, that is, the methodwould not work for highly polar penetrants like water. Furthermore, since thepolymer matrix can not locally relax to accommodate larger inserted penetrants itonly works for small molecules (typically just up to O2, N2, etc.). From the insertionenergy data via Equation 1.5 the chemical excess potential mex for infinite dilution canbe calculated and converted in the respective solubility using Equation 1.6.

mex ¼ RT � ln < expð�Eins=kTÞ > ð1:5Þ

Scalc ¼ T0p0T exp �mexkT

� �ð1:6Þ

withRbeing theuniversal gas constant andT0 and p0 being temperature andpressureunder standard conditions (T0¼ 273.15K; p0¼ 1013� 105 Pa).Table 1.1 contains typical solubility prediction data for an ultrahigh free-volume

polymer (PTMSP) and a polymer with more conventional transport properties(PTMSS).As already mentioned the Gusev–Suter method normally only works for small

penetrant molecules like oxygen or nitrogen. For a long time no really generallyapplicable alternativemethod was available to overcome the problem, but a few yearsago Boulougouris, Economou Theodorou et al. [27, 28] suggested a new inverseWidom method based on the particle-deletion algorithm DPD to overcome thisproblem in principle. The related computer code was, however, only applicable tospecial, relatively simplemodel systems. Based on DPD also a generalized version ofthis algorithm was presented in the literature [29] permitting the calculation ofsolubility coefficients formolecules as large as, for example, benzene in polymers forwhich reasonable forcefield parameters exist. Table 1.2 contains solubility data for anumber of penetrants of different size in PDMS obtained in this way.

1.3.2.2 Prediction of Diffusion ConstantsThe following description again follows the already quoted papers ofGusev andSuter.Using theEins valuesmentioned in the foregoing section, thewhole packingmodel in


question is separated into regions of free volume (low interaction energy) and regionsof densely packed polymer (high interaction energy; cf. Figure 1.3). The bordersbetween the energetically attractive regions Eins(x, y, z) around the resulting localinsertion energy minima are given as crest surfaces of locally maximum insertionenergy. In the two-dimensional analogy of a cratered landscape a minimum energyregion would be represented by a crater, while the crest surface of locally maximuminsertion energy would be reduced to the crest line separating one crater from theadjacent ones. From this identification of energetically separated sites where apenetrant would typically sit (approximately the centers of holes) and jump proba-bilities between adjacent sites (which can be calculated by proper integration over thementioned crest lines and craters of the insertion energy function Eins(x, y, z) anefficient Monte Carlo simulation method for the jump-like diffusion of small

Table 1.1 Results of application of the Gusev–Sutermethod to thesolubility of N2 in PTMSP and PTMSS.

Polymer Structure formula

Average simulatedN2 solubilitycoefficient Scalc[cm3(STP)/(cm3 atm)]

Average measured N2solubility coefficientSexp [cm

3(STP)/(cm3 atm)]

PTMSP 1.16 [24] 1.02 [25]

PTMSS 0.19 [24] 0.18 [26]


molecules in a polymer matrix can be developed (cf. Figure 1.4). With this algorithmthe simulation range can almost extend in the ms range. That is, in most cases thenormal diffusive regime can be reached and the respective constant of diffusion Dican be obtained via the Einstein equation from the slope of the mean squareddisplacement si(t):

~siðtÞ ¼ h~riðtÞ�~rið0Þj2i�� ð1:7Þ

!DiðtÞ ¼ hj~riðtÞ�~rið0Þj2i

6tð1:8Þ

Here,~riðtÞis the position vector of penetrant i and is the average over all possibletime origins t¼ 0 and all simulated trajectories of a penetrant of a given kind. Again,as with the solubilities the Gusev–Suter method can only handle small penetrants inthis way, because the respective polymer matrix cannot conformationally adjust tolarger penetrants. Table 1.3 contains a comparison between experimental and

Table 1.2 Results of application of a generalized DPD method to different penetrants in PDMS.

SoluteScalc[cm3(STP) cm�3 bar�1]

Sexp[cm3(STP) cm�3 bar�1]

Oxygen 0.32a 0.224b

Nitrogen 0.13a 0.127b

Acetone 69a 33–66c

Benzene 495a 275–624d

a[29, 30].b[31].c[32].d[33].

Figure 1.3 Free volume for a perfluorinated polymer in redindicating into the densely packed polymer.


calculated values,Dexp andDcalc, respectively for a number of gases in PTMSP.Here,for methane and carbon dioxide it has to be considered that these molecules arenormally already too large to lead to reasonable results with theGusev–Sutermethod.In comparing simulated and experimentally measured transport parameters

one has to be aware that experimental data in the literature depending, for example,on sample preparation conditions and the chosen measurement methodology canshow a considerable scatter, often reaching a factor of two or even more. It is, forexamplewell-known that polyimides often contain residual solventfilling a part of thefree volume and thus leading to systematically lower S and D values from experi-ments than from simulations [34].

1.3.3Permeability of Small Molecules and Free-Volume Distribution

The distribution of free volume in amorphous polymers is of paramount importancefor the respective materials transport behavior towards small and medium-sizedpenetrants.

Figure 1.4 Jump-like diffusion of oxygen molecules in a perfluorinated polymer matrix.

Table 1.3 Results of application of the Gusev–Suter method forthe diffusion constants of different penetrants in PTMSP.

Solute Dcalc [10�5 cm2/s] Dexp [10

�5 cm2/s]

Nitrogen 7.7a 3.50b

Oxygen 7.5a 4.66b

Methane 8.2a 2.64b

Carbon dioxide 9.2a 8.02b

a[24].b[25].


While in rubbery polymers differences in the segmental mobility can be moreimportant than differences in the free-volume distribution for glassy polymers oftencertain basic correlations can be found between the permeability of small moleculesand free-volume distribution. Other important factors are the molecular mobility ofchain segments and the local chemical composition.Experimentally, the free-volume distribution can be best characterized with

positron annihilation lifetime spectroscopy (PALS). There, in organic glasses or-tho-positronium (o-Ps) which has a lifetime of 142 ns in vacuo shows a strongtendency to localize in heterogeneous regions of low electron density (holes). Inpolymeric materials the vacuum lifetime is cut short via the pick-off mechanism,where o-Ps prematurely annihilates with one of the surrounding bound electrons.This lifetime can (under certain assumptions) be converted in an average hole radius[35, 36], while the intensity of the lifetime signal may permit conclusions about theoverall contents of free volume. There are, however, a number of shortcomings withcommon PALSmethodology. Often, the holes forming the free volume are assumedto be just spheres and the shape of calculated hole radius distribution peaks is set toGaussian. Furthermore, positrons in their limited lifetime seem not to be capable ofprobing large holes of complex topology (cf. in particular PTMSP and other ultrahighfree-volume polymers) [24, 37]. Finally the size of the positronium molecule doesonly permit probing of the accessible free volume for molecules about the size ofhydrogen.Atomisticmolecularmodelingutilizing bulkmodels on the other hand canprovide

additional even more detailed information about free-volume distributions inamorphous polymers. In this way, glassy polymers, where individual differencesin chain segmentmobility donot have an as distinct influence on transport propertiesthan in rubbery polymers, can be roughly grouped into three classes regarding theirsmall molecule permeability, as will be outlined in the following for the example ofoxygen.

1.3.3.1 Examples of Polymers with Low Permeability of Small Molecules(e.g., PO2� 50 Barrer)Figure 1.5(a) shows as a typical example a computer-tomography-like atomic mono-layer representation of a bulk model for diisopropyldimethyl PEEK WC (DIDM-PEEK). In this case the oxygen-accessible free volume is obviously organized inrelatively small isolated holes and the respective size distribution (cf. Figure 1.5(b)) ismonomodal and extending only to hole radii of about 5 Å.

1.3.3.2 Examples of Polymers with High Permeability of Small Molecules(e.g., 50 Barrer� PO2� 200 Barrer)Similarly to Figure 1.5(a), Figure 1.6(a) displays an atomic monolayer representationfor a so-called high-performance polymer (here PPrSiDPA with a PO2 of 230 Barrer[38]). Already in this view larger holes are visible than for the case of low-performancepolymers (cf. Figure 1.5(a)) and the hole-size distribution (Figure 1.6(b)) reveals amuchwider range of radii (here extending to 10Å and being bi-modal). This situationis quite typical for polymerswith high gas transport capacity. Amore systematic study


on polyimides [34] did, for example, reveal that the major difference between low-performance and high-performance polyimides with about the same overall contentsof free volume lies in the distribution of the (e.g., oxygen) accessible free volume.Low-performance polyimides show just a monomodal distribution extending up toabout 5 Å, while high-performance polyimides behave more or less similar to theexample illustrated in Figure 1.6.

1.3.3.3 Examples of Polymers with Ultrahigh Permeability of Small Molecules(e.g., PO2� 1000 Barrer)Figure 1.7 then shows respective data for an ultrahigh free-volume and performancepolymer, Teflon AF2400 of DuPont (PO2¼ 1140 Barrer; [39]). One can recognize that

Figure 1.5 (a) Atomicmonolayer representation (thickness about3 Å) of a typical packing model and structure formula forDIDMPEEK. (b) Hole-size distribution for the packing modelshown in Figure 1.5(a).


in this case there is conventional free volume organized in isolated holes in theradius range below 10Å existing in parallel with a partly continuous phase of muchlarger holes that in this case are visible as a peak between 15 and 20Å. The effect isevenmore pronounced for PTMSP, the polymer of this kind with the highest oxygenpermeability so far measured (about 9000 Barrer; [38]). There, the continuity for thelarge-hole phase is more clearly visible already in atomic monolayer representationsof respective packing models [37] and the ratio between the area under theconventional free-volume peak and the continuous hole phase peak in the hole-size distribution is even smaller than for Teflon AF2400.The fact that for the mentioned ultrahigh free-volume polymers the continuous

hole-phase peak appears at rather limited values is related with the limited size of the

Figure 1.6 (a) Atomicmonolayer representation (thickness about3 Å) of a typical packing model and structure formula forPPrSiDPA. (b)Hole-size distribution for the packingmodel shownin Figure 1.6(a).


investigated models (45–50Å) while the thickness of real polymer membranes canextend into the micrometer range.

1.4Summary

Atomisticmolecularmodeling techniques have proven to be a very useful tool for theinvestigation of the structure and dynamics of dense amorphous membrane poly-mers and of transport processes in these materials. By utilizing these methods,information can be obtained that is not accessible by experimental means.

Figure 1.7 (a) Atomic monolayer representation of a typicalpackingmodel and structure formula for Teflon AF2400. (b)Hole-size distribution for the packing model shown in Figure 1.7(a).


Acknowledgments

The work was in part supported by the European projects Growth Program,PERMOD – Molecular modeling for the competitive molecular design of polymermaterials with controlled permeability properties. Contract #G5RD-CT-2000-200,the 6 Framework Programme project MULTIMATDESIGN- Computer aided mo-lecular design of multifunctional materials with controlled permeability properties,Contract no.: 013644, INTAS - RFBR 97–1525 grant.

References

1 Drioli, E., Natoli,M., Koter, I. and Trotta, F.(1995)Biotechnology andBioengineering, 46,415.

2 Boyd, R.H. and Krishna Pant, P.V. (1991)Macromolecules, 24, 6325.

3 M€uller-Plathe, F. (1991) Journal ofChemical Physics, 94, 3192.

4 Sok, R.M., Berendsen, H.J.B. and vanGunsteren,W.F. (1992) Journal of ChemicalPhysics, 96, 4699.

5 Krishna Pant, P.V. and Boyd, R.H. (1993)Macromolecules, 26, 679.

6 Gusev, A.A., M€uller-Plathe, F., vanGunsteren, W.F. and Suter, U.W. (1994)Advances in Polymer Science, 16, 207.

7 Tamai, Y., Tanaka, H. and Nakanishi, K.(1994) Macromolecules, 27, 4498.

8 Chassapis, C.S., Petrou, J.K., Petropoulos,J.H. and Theodorou, D.N. (1996)Macromolecules, 29, 3615.

9 Fried, J.R., Sadad-Akhavi, M. and Mark,J.E. (1998) Journal of Membrane Science,149, 115.

10 van der Vegt, N.F.A., Briels, W.J.,Wessling, M. and Strathmann, H. (1999)Journal of Chemical Physics, 110, 11061.

11 Hofmann, D., Fritz, L., Ulbrich, J.,Schepers,C.andB€ohning,M. (2000)Macro-molecular Theory and Simulations, 9, 293.

12 Tocci, E., Hofmann, D., Paul, D., Russo, N.and Drioli, E. (2001) Polymer, 42, 521.

13 Neyertz, S. and Brown, D. (2001) Journal ofChemical Physics, 115, 708; Neyertz, S.Brown, D. Douanne, A. Bas, C. andAlb�erola, N.D. (2002) The Journal ofPhysical Chemistry. B, 106, 4617.

14 van der Vegt, N.F.A. (2002) Journal ofMembrane Science, 205, 125; vanGunsteren, W.F. and Berendsen, H.J.C.(1990) Angewandte Chemie, 29, 992.

15 Macchione, M., Jansen, J.C., De Luca, G.,Tocci, E., Longeri, M. and Drioli, E. (2007)Polymer, 48, 2619–2635.

16 Haile, J.M. (1992) Molecular DynamicsSimulation Elementary Methods, WileyInterscience, New York.

17 Tocci, E., Gugliuzza,A.,De Lorenzo, L. andDrioli, E. in preparation.

18 (1999) Polymer User Guide, Amorphous CellSection, Version 400pþ , MolecularSimulations Inc., San Diego, CA.

19 (1999) Discover User Guide, Polymer UserGuide, Amorphous Cell Section, Version4.0.0pþ , Molecular Simulations Inc., SanDiego, CA.

20 Sun, H. and Rigby, D. (1997) SpectrochimActa, 53A, 1301.

21 Rigby, D. Sun, H. and Eichinger, B.E.(1997) Polymer International, 44, 311.

22 Gusev, A.A. Arizzi, S. and Suter, U.W.(1993) Journal of Chemical Physics, 99,2221.

23 Gusev, A.A. and Suter, U.W. (1993) Journalof Chemical Physics, 99, 2228.

24 Hofmann, D., Heuchel, M., Yampolskii,Yu., Khotimskii, V. and Shantarovich, V.(2002) Macromolecules, 35, 2129–2140.

25 Ichiraku, Y., Stern, S.A. and Nakagawa, T.(1987) Journal of Membrane Science, 34,5–18; Masuda, T., Iguchi, Yu., Tang, B.-Z.and Higashimura, T. (1988) Polymer, 29,2041–2049; Bondar, V.I. (1995) Ph.D.

References j17

Thesis, Institute of PetrochemicalSynthesis, Moscow.

26 Khotimskii, V.S., Filippova, V.G.,Bryantseva, I.S., Bondar, V.I.,Shantarovich, V.P. and Yampolskii, Y.P.(2000) Journal of Applied Polymer Science,78, 1612–1620.

27 Boulougouris, G.C., Economou, I.G. andTheodorou, D.N. (2001) Journal ofChemical Physics, 115, 8231–8237.

28 Boulougouris, G.C., Voutsas, E.C.,Economou, I.G., Theodorou, D.N. andTassios, D.P. (2001) The Journal of PhysicalChemistry B, 105, 7792–7798.

29 Siegert, M.R., Heuchel, M. and Hofmann,D. (2007) Journal of ComputationalChemistry, 28, 877–889.

30 Simulated data from, Siegert, M.R. (2006)PhD Thesis, Berechnung vonL€oslichkeitskoeffizienten in Polymer-Materialien, FU-Berlin.

31 Kamiya, Y., Naito, Y., Hirose, T. andMizoguchi, K. (1990) Journal of PolymerScience, Part B, Polymer Physics, 28 (8),1297–1308.

32 Singh, A., Freeman, B.D. and Pinnau, I.(1998) Journal of Polymer Science, Part B,Polymer Physics, 36 (2), 289–301; Frahn, J.(GKSS), personal communication.

33 Summers, W.R., Tewari, Y.B. andSchreiber, H.P. (1972) Macromolecules, 5(1), 12–16; Lichtenthaler, R.N., Liu, D.D.and Prausnitz, J.M. (1974) Berichte derBunsen-Gesellschaft — Physical ChemistryChemical Physics, 78 (5), 470–477; Grate,J.W., Kaganove, S.N. and Bhethanabotla,V.R. (1997) Faraday Discussions, 107,259–283; Frahn, J. (GKSS), personalcommunication.

34 Heuchel, M., Hofmann, D. and Pullumbi,P. (2004) Macromolecules, 37, 201.

35 Gregory, R.B. and Yongkang, Zhu (1991)Positron and Positron Chemistry (ed. Y.C.Jean), World Scientific, Singapore, p. 136.

36 Provensher, S.W. (1982) Computer PhysicsCommunications, 27, 229.

37 Hofmann, D., Entrialgo-Castano, M.,Lerbret, A., Heuchel, M. andYampolskii, Yu. (2003) Macromolecules,36, 8528–8538.

38 Yampolskii, Yu.P., Korikov, A.P.,Shantarovich, V.P., Nagai, K., Freeman,B.D., Masuda, T., Teraguchi, M. and Kwak,G. (2001)Macromolecules, 34, 1788–1796.

39 Alentiev, A.Yu., Yampolskii, Yu.P.,Shantarovich, V.P., Nemser, S.M. andPlat�e, N.A. (1997) Journal of MembraneScience, 126, 123–132.


2Polymeric Membranes for Molecular SeparationsHeru Susanto and Mathias Ulbricht

2.1Introduction

In this chapter we describe the state-of-the-art and the challenges in preparation andmanufacturing of polymeric membranes for molecular separations in liquid phase.The processes include separation of aqueous solutions, that is, pressure-drivendesalination using reverse osmosis and nanofiltration, fractionations of small andlarger molecules using ultrafiltration and removal of organic substances by perva-poration (e.g., for shifting equilibria for (bio)chemical reactions). Separations innonaqueous organic systems such as pervaporation and nanofiltration will also becovered. The preparation of charged membranes for electromembrane processes isanother important application area for special polymers. Surface modification ofmembranes has become an important tool to reduce fouling or increase biocompati-bility, but it can also be used to changemembrane selectivity by combining separationmechanisms (e.g., based on size and charge).

2.2Membrane Classification

Synthetic membranes for molecular liquid separation can be classified according totheir selective barrier, their structure and morphology and the membrane material.The selective barrier– porous, nonporous, charged or with special chemical affinity –dictates the mechanism of permeation and separation. In combination with theapplied driving force for transport through the membrane, different types ofmembrane processes can be distinguished (Table 2.1).Selective barrier structure. Transport through porous membranes is possible by

viscous flow or diffusion, and the selectivity is based on size exclusion (sievingmechanism). This means that permeability and selectivity are mainly influenced bymembrane pore size and the (effective) size of the components of the feed:Molecules


j19

with larger size than the largest membrane pore will be completely rejected, andmolecules with smaller size can pass through the barrier; the Ferry–Renkin modelcan be used to describe the effect of hindrance by the pore on rejection inultrafiltration (UF) [1]. Transport through nonporous membranes is based on thesolution-diffusion mechanism [1, 2]. Therefore, the interactions between the per-meand and the membrane material dominate the mass transport and selectivity.Solubility and chemical affinity on the one hand, and the influence of polymerstructure on mobility on the other hand serve as selection criteria. However, thebarrier structure may also change by uptake of substances from the feed (e.g., byplastification), and in those cases real selectivities can bemuch lower than ideal onesobtained from experiments using only one component in the feed or at low feedactivities. Separation using charged membranes, either nonporous (swollen gel) orporous (fixed charged groups on the porewall), is based on charge exclusion (Donnaneffect; ions or molecules having the same charge as the fixed ions in the membranewill be rejected, whereas species with opposite charge will be taken up by andtransported through the membrane). Therefore, the kind of charge and the chargedensity are the most important characteristics of these membranes [1]. Finally,molecules ormoieties with special affinity for substances in the feed are the basis forcarrier-mediated transport through the membrane; very high selectivities can be

Table 2.1 Overview of main polymer membrane characteristicsand membrane-based processes for molecular separationsin liquid phase.

Transmembrane gradient

Selectivebarrier

Typicalstructure

Concentrationdifference

Pressuredifference

Electricalpotential

Nonporous anisotropic,thin-filmcomposite

Pervaporation Reverse OsmosisNanofiltration

Microporousdp� 2 nm

anisotropic,thin-filmcomposite

Dialysis Nanofiltration Electrodialysis

Non- ormicroporous,with fixed charge

isotropic Dialysis Electrodialysis

Mesoporousdp¼ 2 . . . 50 nm

anisotropic,isotropictrack-etched

Dialysis Ultrafiltration Electro-ultrafiltration

Carrier inliquid

immobilizedin isotropicporousmembrane

Carrier-mediatedseparation

Affinity ligand insolid matrix

isotropic,anisotropic

20j 2 Polymeric Membranes for Molecular Separations

achieved; the diffusive fluxes are higher for (immobilized) liquid membranes thanfor polymer-based fixed-carrier membranes [1].Concentration polarization can dominate the transmembrane flux in UF, and

this can be described by boundary-layer models. Because the fluxes throughnonporous barriers are lower than in UF, polarization effects are less importantin reverse osmosis (RO), nanofiltration (NF), pervaporation (PV), electrodialysis (ED)or carrier-mediated separation. Interactions between substances in the feed andthe membrane surface (adsorption, fouling) may also significantly influence theseparation performance; fouling is especially strong with aqueous feeds.Cross-section structure. An anisotropic membrane (also called asymmetric) has a

thin porous or nonporous selective barrier, supported mechanically by a muchthicker porous substructure. This type ofmorphology reduces the effective thicknessof the selective barrier, and the permeate flux can be enhanced without changes inselectivity. Isotropic (symmetric) membrane cross-sections can be found for self-supported nonporous membranes (mainly ion-exchange) and macroporous micro-filtration (MF) membranes (also often used in membrane contactors [1]). The onlyexample for an established isotropic porous membrane for molecular separations isthe case of track-etched polymer films with pore diameters down to about 10 nm.All the above-mentioned membranes can in principle be made from one material.In contrast to such an integrally anisotropic membrane (homogeneous with respectto composition), a thin-film composite (TFC) membrane consists of differentmaterials for the thin selective barrier layer and the support structure. In compositemembranes in general, a combination of two (or more) materials with differentcharacteristics is used with the aim to achieve synergetic properties. Other examplesbesides thin-film are pore-filled or pore surface-coated composite membranes ormixed-matrix membranes [3].Membrane materials. Polymeric membranes are still dominating a very broad

range of industrial applications. This is due to their following advantages: (i) manydifferent types of polymericmaterials are commercially available, (ii) a large variety ofdifferent selective barriers, that is, porous, nonporous, charged and affinity, can beprepared by versatile and robust methods, (iii) production of large membrane areawith consistent quality is possible on the technical scale at reasonable cost based onreliable manufacturing processes, and (iv) various membrane shapes (flat sheet,hollow-fiber, capillary, tubular, capsule; Figure 2.1) and formats includingmembranemodules with high packing density can be produced. However,membrane polymersalso have some limitations. A very well-defined regular pore structure is difficult toachieve, and the mechanical strength, the thermal stability and the chemical resis-tance (e.g., at extreme pH values or in organic solvents) are rather low for manyorganic polymers. In that regard, inorganic materials can offer some advantages,such as high mechanical strength, excellent thermal and chemical stabilities, and insome cases a very uniform pore shape and size (e.g., in zeolites). However, someinorganicmaterials are very brittle, and due to complicated preparationmethods andmanufacturing technology, the prices for many inorganic membranes (especiallythose for molecular separations) are still very high. An overview of inorganicmembranes for separation and reaction processes can be found elsewhere [4, 5].

2.2 Membrane Classification j21

2.3Membrane Polymer Characteristics

2.3.1Polymer Structure and Properties

Polymers formembranepreparation canbe classified intonatural and synthetic ones.Polysaccharides and rubbers are important examples ofnaturalmembranematerials,but only cellulose derivatives are still used in large scale for technicalmembranes. Byfar the majority of current membranes are made from synthetic polymers (which,however, originally had been developed for many other engineering applications).Macromolecular structure is crucial formembranebarrier andother properties;mainfactors include the chemical structure of the chain segments, molar mass (chainlength), chain flexibility as well as intra- and intermolecular interactions.Macromolecule chain flexibility is affected by the chemical structure of the main

chain and the side groups. A macromolecule is flexible when unhindered rotationaround single bonds in the main chain is possible. This flexibility can be reduced byseveral means, for example, by introducing double bonds or aromatic rings in themain chain, by forming ladder structures along themain chain or by incorporation ofbulky side groups. Even larger effects with respect to the possible macroconforma-tions can be imparted by changes of the chain architecture, that is, the transition fromlinear to branched or network structures. Polymer molar mass and its polydispersityhave an influence on chemical and physical properties via the interactions betweenchain segments (of different or even the same molecule), through noncovalentbinding or entanglement. For stability, high molar mass is desirable because thenumber of interaction sites increases with increasing chain length. However, thesolubility will decrease with increasing molar mass.The preceding structural characteristics dictate the state of polymer (rubbery vs.

glassy vs. semicrystalline) which will strongly affect mechanical strength, thermalstability, chemical resistance and transport properties [6]. In most polymeric mem-branes, the polymer is in an amorphous state. However, some polymers, especiallythose with flexible chains of regular chemical structure (e.g., polyethylene/PE/,polypropylene/PP/or poly(vinylidene fluoride)/PVDF/), tend to form crystalline

Figure 2.1 Polymeric membrane shapes and cross-sectionalstructures. Tubular membranes are similar to flat sheetmembranes because they are cast on a macroporous tube assupport. Capillary membranes are hollow fibers with largerdiameter, that is, >0.5mm.


domains. This will lead to higher mechanical stability (high elastic modulus) as wellas higher temperature and chemical resistance than for the same polymer inamorphous state, but the free volume (and hence permeability) will bemuch smaller.For semicrystalline polymers, the melting temperature (Tm) is important, because atthis temperature a transition between crystalline and liquid state will occur. The glasstransition temperature (Tg) is a much more important parameter to characterizeamorphous polymers, because at this temperature a transition between solid (glass)and supercooled melt (rubber) takes place. In the glassy state molecules are frozen,therefore, chainmobility of a polymer is very limited.Heating this polymer over itsTgleads to a muchmoremobile andmore flexible state, with lower elastic modulus andhigher permeability. So-called glassy polymers have aTg beyond room temperature,and rubbery polymers (or elastomers) have a Tg below room temperature. Polymerselection will be more important for membranes with nonporous selective barrier,because flux and selectivity depend on the solution-diffusion mechanism. Formembranes with a porous selective barrier, the mechanical stability will be crucialto preserve the shape and size of the pores.Block- or graft copolymers, which contain two or more different repeating units

within the same polymer chains, are often used instead of homopolymers in order toobtain high-performance polymeric membranes; the overriding aims are synergiesbetween properties of the different components. In addition, blending of polymers orcopolymers is also performed. In these cases, compatibility and miscibility ofboth (co)polymers in one solvent are required in order to get a homogenous solution(cf. Section 2.4.2). The resulting solid membrane can be a homogenous polymerblend, as indicated by one Tg value between those for the two (co)polymers.A heterogeneous (phase separated) polymer blend will be characterized by two(or more) Tg values for the individual phases. Extensive existing knowledge frompolymer blending can also be adapted to membrane preparation [7].Chemical or physical cross-linking of the polymer is applied in order to control

membrane swelling, especially for separations of organic mixtures. In addition, thiscan also enhance mechanical strength as well as chemical stability of a membrane.However, crosslinking decreases polymer solubility, therefore it is often done aftermembrane formation (cf. Sections 2.4–2.6).The hydrophilicity–hydrophobicity balance of the membrane polymer is another

important parameter that ismainly influenced by the functional groups of the polymer.Hydrophilic polymers have high affinity to water, and therefore they are suited as amaterial fornonporousmembranes thatshouldhaveahighpermeability andselectivityfor water (e.g., in RO or hydrophilic PV). In addition, hydrophilic membranes havebeenproven tobe lessprone to fouling inaqueoussystems thanhydrophobicmaterials.

2.3.2Membrane Polymer Selection

2.3.2.1 Polymers for Porous BarriersThe selection of the polymer for a porous membrane is based on the requirementsof the manufacturing process (mainly solubility for controlled phase separation;

2.3 Membrane Polymer Characteristics j23

cf. Section 2.4.2), and the behavior and performance under application conditions.The following material properties are important to be considered:

(i) Film-forming properties indicate the ability of a polymer to form a cohesivefilm, and the macromolecular structure, especially molar mass and attractiveinteractions between chain segments, is crucial in this regard (cf. Section 2.3.1).Poly(ether sulfones) (PES), polysulfones (PSf), polyamides (PA) or polyimides(PI) are examples for excellent film-forming materials [8].

(ii) Mechanical properties involve film strength, film flexibility and compactionstability (especially of a porous structure). The latter is most important forhigh-pressure processes (e.g., for the porous substructure of an integrallyanisotropic RO membrane). Because hollow fiber membranes are self-supporting, the mechanical stability will be especially relevant. Manycommercial flat-sheet membranes are prepared on a nonwoven supportmaterial (Figure 2.2).

(iii) Thermal stability requirements depend verymuch on the application. In order toensure the integrity of a pore structure in the nanometer dimension, theTg of thepolymer should be higher than the process temperature.

(iv) Chemical stability requirements include the resistance of the polymer at extremepH values and other chemical conditions. Cleaning agents such as strong acidsor bases, or oxidation agents are usually used to clean a fouled membrane. Thestability in special solvents is also important in selected cases, that is, whenprocesses with nonaqueous mixtures are considered.

(v) The hydrophilicity–hydrophobicity balance correlates with the wettability ofthe material. This can be important in order to use all the pores in UF, or

Figure 2.2 SEM micrograph of a microtome cross-section of aporous polymer membrane with an anisotropic structure on anonwoven as mechanical support (reprinted from [9], withpermission from Wiley-VCH, 2006).


when a porous membrane is applied as a contactor between a liquid and agas phase, and the phase boundary is stabilized because the liquid will not wetthe dry pores of the m

membrane operations · 2018. 6. 8. · 6 fouling in membrane processes 121 anthony g. fane, tzyy h....

Documents