1:ÜIFFEX

SURFACE CHEMICAL ASPECTS OFMICROBUBBLE FLOTATION

by

Waverly Mitchell Hale

Thesis submitted to the Faculty of theVirginia Polytechnic Institute and State University

in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

in

Mining and Minerals Engineering

APPROVED:

r»*@#6-.: M). :4,igR.H. Yoon, Chairman G. T. AdelV ‘„

J. R. Lucas G. H. Luttrell

dfiä SURFACE CHEMICAL ASPECTS OFij\

MICROBUBBLE FLOTATION\1Q} by

waverly Mitchell Hale

Committee Chairman: Dr. Roe—Hoan YoonMining and Minerals Engineering

(ABSTRACT)

In order to demonstrate the ability of microbubbleflotation to superclean coal to ash levels of less than2%, several Eastern U. S. coals have been tested. Theresults show that the process is capable of producingsuperclean coal with improved recovery as compared to theconventional flotation process.

To further improve and understand the microbubbleflotation process, electrokinetic studies of thehydrocarbon oils used in flotation as collectors have beenconducted. Also, the effect of oil emulsifiers on the zetapotential of oil droplets has been studied. In general,oil droplets are negatively charged and negative zetapotential is reduced with the addition of nonionic andcationic surfactants. On the other hand, the negativecharge is increased with the addition of an anionicreagent. It has also been shown that the negative zeta

potential of oil droplets increases with increasinghydrocarbon chain length.

The effects of different collectors on induction timeand flotation have been determined by conductingmicroflotation and induction time experiments using anElkhorn seam coal sample. The results show that industrialoils combined with the coal have the shortest inductiontimes and, therefore, the highest flotation yields ascompared to pure hydrocarbon oils. It has also been shownthat oil emulsifiers tend to increase flotation yield andreduce particle/bubble induction time.

ACKNOWLEDGMENTS

The author wishes to express his gratitude andappreciation to Dr. Roe—Hoan Yoon for his guidance, supportand useful suggestions throughout the course of this study.Special appreciation is given to Dr. G. T. Adel for hisadvice during this project. He also extends his gratitudeto Dr. J. R. Lucas, the Department of Mining and MineralsEngineering and the U. S. Department of Energy for thefinancial support that made this work possible. Specialthanks are also given to Dr. G. H. Luttrell forcontributions made towards the completion of this work.

students for their help, support and friendship.

Finally, the author wishes to express his deepestthanks to his mom, dad and sisters, for theirencouragement, support and love when it was so oftenneeded.

iv

TABLE OF CONTENTS

PageABSTRACT ....................... ii

ACKNOWLEDGMENTS .................... iv

TABLE OF CONTENTS ................... v

LIST OF FIGURES .................... ix

LIST OF TABLES .................... xii

I. INTRODUCTION .................. 1

1.1 General .................. 11.2 Objectives of the Proposed work ...... 9

II. SUPERCLEANING ULTRAFINE COAL USING MICROBUBBLEFLOTATION .................... ll2.1 General .................. 122.2 Literature Review ............. l42.3 Experimental ................ l9

2.3.1 Coal Samples ............ 192.3.2 Reagents .............. 2l2.3.3 Equipment .............. 212.3.4 Procedure .............. 23

a. Microbubble Flotation ...... 23b. Conventional Flotation ..... 242.4 Results .................. 26

2.4.1 Microbubble Flotation ........ 262.4.2 Microbubble Flotation after

Conventional Flotation ....... 33

v

vi I

2.5 Discussion ................. 432.5.1 Microbubble Elotation ........ 432.5.2 Microbubble Flotation after

Conventional Flotation ....... 462.6 Summary and Conclusions .......... 48

III. ELECTROKINETICS OF HYDROCARBON OIL EMULSIONS . . 493.1 General .................. 493.2 Literature ................. 51

3.2.1 Zeta Potential of Oil Emulsions . . . 5l3.2.2 Zeta Potential of Coal ....... 533.3 Experimental ................ 56

3.3.1 Coal Sample ............. 563.3.2 Reagents .............. 563.3.3 Equipment .............. 593.3.4 Procedure .............. 59

3.4 Results .................. 613.4.1 Effect of Hydrocarbon Chain Length . 6l3.4.2 Effect of Double Bonds in the

Hydrocarbon Chain .......... 633.4.3 Effect of Industrial Oils ...... 63· 3.4.4 Effect of Emulsifiers on Kerosene . . 63a. Effect of Nonionic Emulsifiers . 63b. Effect of Ionic Surfactants . . . 66

3.4.5 Zeta Potential of Coal ....... 693.5 Discussion ................. 7l

3.5.1 Effect of Hydrocarbon Chain Length . 713.5.2 Effect of Double Bonds in the

Hydrocarbon Chain .......... 733.5.3 Effect of Industrial Oils ...... 743.5.4 Effect of Emulsifiers on Kerosene . . 74

a. Effect of Nonionic Emulsifiers . 74b. Effect of Ionic Surfactants . . . 773.5.5 Zeta Potential of Coal ....... 78

vii

3.6 Summary and Conclusions ........... 79

IV. THE EFFECT OF HYDROCARBON OILS ON THE INDUCTIONTIME AND MICROFLOTATION OF COAL ......... 814.1 General .................. 8l4.2 Literature Review ............. 844.3 Experimental ................ 86

4.3.1 Coal Samples ............ 864.3.2 Reagents .............. 864.3.3 Equipment .............. 874.4.4 Procedure .............. 87

4.4 Results .................. 9l4.4.1 Effect of Hydrocarbon Chain Legnth . 91

a. Microflotation Experiments . . . 91b. Induction Time Experiments . . . 93

4.4.2 Effect of Industrial Oils ...... 93a. Effect of Kerosene ....... 93b. Effect of Diesel Fuel ...... 96c. Effect of No. 3, No. 6 Oil

Mixture ............. 964.4.3 Effect of Oil Emulsifiers ...... 99

a. Effect of Kerosene with Span 40 . 99b. Effect of Kerosene with Tween 40. 99

4.4.4 Effect of Surfactants ........ 102a. Effect of Kerosene with a

Cationic Surfactant ....... 102b. Effect of Kerosene with an

Anionic Surfactant ....... 1024.5 Discussion ................. 106

4.5.1 Effect of Hydrocarbon Chain Length . 1064.5.2 Effect of Industrial Oils ...... 108

a. Effect of Kerosene ....... 108b. Effect of Diesel Fuel ...... 108

viii

c. Effect of No. 3, No. 6 OilMixture ............. 109

4.5.3 Effect of Oil Emulsifiers ...... 110a. Effect of Kerosene with Span 40 . 110b. Effect of Kerosene with Tween 40. 110

4.5.4 Effect of Surfactants ........ 111a. Effect of Kerosene with a

Cationic Surfactant ....... 111b. Effect of Kerosene with anAnicnic Surfactant ....... 112

4.6 Summary and Conclusions ........... 113

V. GENERAL CONCLUSIONS AND RECOMMENDATION FOR FUTUREWORK ...................... 1155.1 General Ccnclusions ............. 1155.2 Recommendations for Further Work ...... 117

I I I I I I I I I I I I I I I I I I I I I I l 9

APPENDIX: Tables of Experimental Results Presented inFigures .................. 125VITA ......................... 149

\

LIST OF FIGURES

Page

Figure 2.1 Schematic diagram of the microbubbleflotation system ............. 22

Figure 2.2 Recovery versus ash curve for themicrobubble flotation of coal from theTaggart seam ............... 27

Figure 2.3 Recovery versus ash curve for themicrobubble flotation of coal from theUpper Cedar Grove seam .......... 29

Figure 2.4 Recovery versus ash curve for themicrobubble flotation of coal from theElkhorn seam ............... 30

Figure 2.5 Recovery versus ash curve for themicrobubble flotation of coal from theSplash Dam seam ............. 32

Figure 2.6 Recovery versus ash curve for themicrobubble flotation of coal from theLower Cedar Grove seam .......... 34

Figure 2.7 Recovery versus ash curve comparingconventional flotation to microbubbleflotation using Elkhorn seam coal .... 35

Figure 2.8 Recovery versus ash curve for coal fromthe Upper Freeport coal seam usingconventional and microbubble flotation . . 38

Figure 2.9 Recovery versus ash curve for theUnited Coal Company's Sample No. 1 usingconventional and microbubble flotation . . 39

Figure 2.10 Recovery versus ash curve for the UnitedCoal Company's Sample No. 2 usingconventional and microbubble flotation . . 4l

Figure 2.11 Recovery versus ash curve for coal fromthe Jericho seam using conventional andmicrobubble flotation .......... 42

ix

X1

1

Figure 3.1 Zeta potentials of pure hydrocarbon oils(hexane, octane, decane, undecane) as afunction of pH value ........... 62

Figure 3.2 Effect of C=C double bonds in purehydrocarbon oils on zeta potential asa function of pH value .......... 64

Figure 3.3 Zeta potential of industrial oilsas a function of pH value ........ 65

Figure 3.4 Effect of oil emulsifiers on the zetapotential of kerosene as a function ofpH value ................. 67

Figure 3.5 Effect of ionic surfactants on the zetapotential of kerosene as a function ofpH value ................. 68

Figure 3.6 Zeta potential of the Elkhorn seam coal . 70Figure 3.7 Schematic diagram of the chemical

structure of Span 40 and Span 60 ..... 75Figure 3.8 Schematic diagram of the chemical

structure of Tween 40 and Tween 60 .... 76Figure 4.1 Schematic diagram of the induction time

apparatus ................ 88Figure 4.2 Flotation yield versus pH value for the

Elkhorn coal using pure hydrocarbon oilsas collectors (hexane, octane, decane,undecane) ................ 92

Figure 4.3 Induction time versus pH value for theElkhorn coal using pure hydrocarbon oilsas collectors (hexane, octane, decane,undecane) ................ 94

xi

Figure 4.4 Results of flotation, induction time andzeta potential measurements conducted asa function of kerosene on Elkhorn coal . . 95Figure 4.5 Results of flotation, induction time andzeta potential measurements conducted asa function of No. 3 fuel oil on Elkhorncoal ................... 97Figure 4.6 Results of flotation, induction time andzeta potential measurements conducted as

a function of No. 3, No. 6 fuel oilmixture on Elkhorn coal ......... 98

Figure 4.7 Results of flotation, induction time andzeta potential measurements conducted asa function of kerosene with Span 40 onElkhorn coal ............... 100

Figure 4.8 Results of flotation, induction time andzeta potential measurements conducted asa function of kerosene with Tween 40 onElkhorn coal ............... 101

Figure 4.9 Results of flotation, induction time andzeta potential measurements conducted asa function of kerosene with dodecylaminehydrochloride (DAH) on Elkhorn coal . . . 103

Figure 4.10 Results of flotation, induction time andzeta potential mesurements conducted as afunction of kerosene with sodium dodecylsulphate (SDS) on Elkhorn coal ...... 105

LIST OF TABLES

Page

Table 2.1 Description of Coal SamplesUsed in Present Work ............ 20

Table 3.1 Description of Collectors Used inZeta Potential Experiments ......... 57

Table 3.2 Description of Surfactants Used inZeta Potential Experiments ......... 58

p xTT

CHAPTER IIntroduction

1.1 General

The U.S. is becoming more conscientious about its ownenergy resources due to the 1973-74 Arab oil embargo andthe petroleum price increase of 1979. Of the recoverableenergy resources that are available, approximately 82% iscoal (Hutton and Gould, 1982). Therefore, coal is thelogical source for the United States' power supply.

In 1947, electrical power utilities burned 86 milliontons, or 16%, of the total coal consumption. As of 1980,569 million tons, or 81%, of the coal consumed was burnedby electrical utilities (Hutton and Gould, 1982). The

Electrical Power Research Institute (EPRI) projects thatapproximately 900 million tons of coal will be burned forthe production of electricity by 1990 (Wright, 1985).

However, compared to petroleum products and naturalgas, coal is an extremely "dirty" fuel. Due to itssedimentary origin, coal usually contains a substantialamount of inert material, especially clays, silica and

sulfur. This material causes problems in the furnaces thatare used for combustion and creates environmental problemsbecause of the type of emissions produced by burning. Dueto the introduction of the Clean Air Act in 1967 and with

1

21

increasingly stringent requirements placed on coal burningfacilities to lower sulfur emissions, coal cleaning hasbecome a necessity. Coal cleaning will increase in demandeven more as the higher quality seams are depleted. Also,in order to obtain complete recovery, washing coal willbecome imperative with increasing mining costs.

While much of the nation's coal may be removed fromthe ground in relatively pure form by modest selectivemining techniques, most coal requires preparation. Sincethe 1880's, coal cleaning has been used by themetallurgical industry for the manufacture of coking coal.The principal purpose of coal preparation is to convert rawcoal into an acceptable commercial product by removing thehigh—ash and high-sulfur refuse material.

Until recently, the major emphasis in coalpreparation has been with coarse coal concentration by jigsand specific gravity baths. However, the introduction ofmechanization to coal mining has increased the amount offine material reporting to the preparation facility. Forexample, the increase in the degree of mechanization in theUnited Kingdom from 2% in 1947 to 98% in 1979 has increasedthe percentage of fine coal from 35% to 85% (Konar and

Sarkar, 1982).

Along with the increase of fine material, there hasalso been an increase in the percent reject in the run—of—mine coal from 10% to 45% (Konar and Sarkar, 1982). This

3

has mainly been caused by the nonselectivity of mechanizedmining. Therefore, the fine coal cleaning circuit mustbecome a crucial component of modern preparation facilitiesin order to prevent the costly loss of the fine combustiblematerial.

Even though fine coal cleaning has become an essentialpart of coal preparation, it has proven to be one of themost difficult tasks. Not only are there more technicalproblems with the treatment of fines, but it is also morecostly than cleaning coarse coal. It has been estimatedthat the capital cost for a plant to clean the 20% of finecoal is approximately 40% of the total capital cost.Overall, the total cost of processing fine material is morethan 2.5 times greater than for the remaining 80% of therun—of—mine material (Miles and Brookes, 1984).

The most widely used method of recovering and cleaningcoal fines is the froth flotation process. Flotationutilizes the differences in the surface properties of coaland gangue particles to achieve a separation. However,differences in the physico—chemical characteristics of coaland gangue particle surfaces are almost indefinable

parameters due to the many different ranks of coal andtypes of ash material. To complicate the matter further,these parameters are readily influenced by the amount ofoxidation of the coal, the pH of the slurry, and the

ul

4‘

presence of various ions in the flotation system. Forthese reasons, froth flotation is a technically difficultprocess. Even with the difficulties involved, flotationhas proven to be beneficial to coal preparation.

Until recently, most material less than 28 mesh wasautomatically discarded with the tailings. An example ofhow flotation has changed this practice is The MariettaCoal Company's Ohio facility. The plant was originallydesigned to produce 500 tons per hour of clean coalproduct. It quickly became apparant to Marietta that thislimit had serious economic consequences since 7% of theirmaterial (-28 mesh) was being discarded. This materialcontained 60-70% coal. with the addition of a flotationcircuit, Marietta began collecting 70% of the previously-lost fine coal. This increased the plant's clean productto 650 tons per hour. According to Marietta's President,"There's not a prep plant in the world that can't justifyflotation cells and a fine coal circuit" (Falas and Zick,1982).

Typically, froth flotation has increased theproductivity and recovery of the fine coal in alllpreparation facilities. However, there are many situationswhere the flotation circuit is not working at its optimum.A study completed by WEMCO under EPRI sponsorship showsthat few of the plants sampled are producing coal atoptimum recovery and ash levels. Of the facilities tested,

5

the coal losses in the flotation circuit ranged from 35 to533 tons per eight-hour shift. Assuming a 300-day per yearoperation, with two working shifts, at $30 per ton and a400-ton per shift loss, this represents a $7,200,000 lossper year in saleable coal (Olson, 1983).

Various operating parameters can be modified in orderto improve coal preparation plant productivity. Byoptimizing reagent dosage, pulp level, air rate or plantcircuitry, the flotation system can operate at its fullpotential. In some instances, minor changes in theflotation circuit can make a nonprofitable plantprofitable. For example, testing done by WEMCO (Olson,1983) illustrates that when the pulp level was increased byone inch in a Kentucky plant, the yield increased by 3.2percent and the ash reporting to the tails increased by 5.0percent.

Even though coal flotation is recovering the abundantamount of fine coal previously lost, the major concern inpreparation is removing the inert material, i.e., ash andsulfur. Most ash and sulfur appears in the fine sizeranges, especially in the -28 mesh size fraction.

Therefore, most improvements in coal cleaning can be madein the processing of fines.

As mentioned previously, froth flotation is the mostwidely applied method for cleaning fine coal. However, it

y 6

has not been successful in achieving low—ash products or inreducing sulfur. Existing and pending environmentalregulations mandate that solid fuels now consumed in powerstations contain almost no pyritic sulfur. In order forcoal to replace petroleum products as the major energysource, the ash and sulfur rejection by coal processingneeds to increase. Since conventional coal beneficiationtechniques have not been able to clean coal so that it willcomply with stringent environmental regulations, new trendsare developing in coal preparation.

An interest has been created in response to theenvironmental and coal combustion problems to clean coal tothe ash and sulfur levels comparable to that of petroleumproducts. This ultra—cleaning of coal has been termed"supercleaning". Various physical and chemical cleaningtechniques are being used to reduce coal to these values.Since chemical techniques are costly, more attention isbeing placed on physical methods.

Most of the physical methods in use are based onultrafine grinding or micronizing the coal beforesubjecting it to the separation process. Ultrafinegrinding can be accomplished by several methods: attritiongrinding (stirred ball milling), pressure crushing,mechanical impact and fluid energy milling (Boose, 1983).The purpose of grinding to these small sizes is to liberatethe finely disseminated ash—bearing and sulfur—bearing

7

particles from the coal.

After liberation, several techniques have been testedfor the separation of the gangue from the valuablematerial. Some of the more successful of these processesare the microbubble flotation process, the Otisca processand the Advanced Fuel Technology process. Since thisinvestigation concerns the microbubble flotation process,it will be considered in detail from this point forward.

Microbubble flotation uses the same physico—chemicalsurface characteristics as conventional flotation to makethe separation between coal and ash particles. However,the difference between microbubble and conventionalflotation·is the bubble size. In microbubble flotationbubble sizes range from 50-100 microns (Yoon, 1984; Trigg,1984), while bubble sizes in conventional flotation rangefrom 600-1000 microns (Fuerstenau, 1980). This differencein bubble size allows an increased recovery of themicronized coal particles (Yoon, 1982; Yoon, 1984; Yoonet al., 1984)

As in conventional flotation, microbubble flotationdepends upon hydrocarbon oil collectors to make the coalparticles more hydrophobic so that attachment with the airbubbles will occur. Various oils affect the chemistry ofthe flotation system in different manners. Some oilsimprove the recovery of the coal but reduce the product

grade. Others are more selective but make flotation lesseconomical due to the loss in yield.

Recently, there has been a great deal of interest inthe effect of hydrocarbon oils on the electrokinetics ofcoal flotation. Depending on the hydrocarbon chain lengthof the collector and the grade of the oil, theelectrophoretic mobility of the oil-in—water emulsions maychange with different pH values (wen and Sun, 1981). Also,with the introduction of oil emulsifiers (flotationpromoters) to coal flotation, questions have arisen as totheir effect on the surface chemistry of coal, e.g., zetapotential and induction time. In the present work, some ofthese parameters have been studied in detail to helpunderstand the mechanisms of microbubble flotation.

, 9 l1.2 Objectives of the Proposed Work

The purpose of this investigation has been to studythe supercleaning of coal using the microbubble flotationprocess. Particular attention has been paid to the role ofsome of the physical and chemical variables in thisprocess, especially the effect of hydrocarbon collectors onflotation chemistry.

The investigation consisted of three phases. Thefirst phase was concerned with the cleaning of variousEastern U.S. coals using a batch microbubble flotationapparatus. The results show the versatility of themicrobubble process in producing superclean coal with highcombustible recovery from various feed coals.

In the second phase, studies were conducted on theeffect of pH on the zeta potential of hydrocarbon oilemulsions. Oil emulsions were made with pure hydrocarbonoils, industrial oils and oils with the addition ofsurfactants. On the basis of the results, physico—chemicalmechanisms responsible for the presence of electricalcharge on oil droplets are proposed.

The third and final phase consisted of studies of theeffect of the various oil emulsions on bubble—particleattachment using induction time measurements and micro-flotation experiments.

Since the topics covered in the three phases of the

10

study are to some degree independent of each other, theyare written with self-contained formats. Each chapter hasits own Introduction, Experimental, Results, Discussion andConclusion sections. However, an overall Conclusionssection is included to bring together the effects of thezeta potential of oil emulsions on induction time and coalflotation.

CHAPTER II

Supercleaning Ultrafine Coal UsingMicrobubble Flotaion

2.1 General

Froth flotation is the most widely used beneficiationmethod for fine material. Many of the low—grade orescurrently being mined would have been uneconomical to minewithout flotation to separate the gangue from the valuablematerial. Originally, flotation was developed toconcentrate sulfide ores, but the process has been adaptedto separate nonsulfide ores, nonmetallic ores, oxides andcoal.

The theory of froth flotation is complex and is notcompletely understood. The process exploits the surfacechemistry of particles to make the separation betweengangue and valuable material. Chemical reagents are usedto enhance or alter the surface properties of minerals andto render the particles either hydrophobic or hydrophilic.The separation occurs when air bubbles rise to the surfaceof a pulp carrying hydrophobic particles, while leavinghydrophilic particles behind.

One of the major problems in fines processing is thesize range of the material that is being treated. If aminimum size of 1 micron is assumed, a size ratio of 500:1

ll

112

exists for the material that reports to the fines circuit(Miles and Brookes, 1984). Conventional froth flotation,like any other separation technique, has an optimum sizerange for treatment. Even though coal in the 28 mesh X0 size fraction is usually sent to the flotation circuit,the optimum flotation occurs in the 50 X 140 mesh sizefraction (Zimmerman, 1979). The 140 mesh X 0 fractioncauses problems in the fines circuit due to its reduced 1

flotation rate.

Some of the problems encountered when processingultrafine material using conventional flotation areassociated with the low probability of collision betweenthese particles and large bubbles. It has been shown thatby reducing the size of bubbles used in flotation, therecovery of the ultrafine material improves (Reay andRatcliff, 1975; Collins and Jameson, 1976; Yoon andLuttrell, 1986). Microbubble flotation of fine coal isbased on this principle. In the ultrafine size range,microbubble flotation has made a substantial improvement inrecovery compared to conventional flotation (Yoon 1982;Yoon et al., 1984; Yoon, 1984; Trigg, 1984).

There has been recent interest in supercleaning coalto reduce the ash handling problems, to minimize the wearon turbine blades and to reduce emission pollutants. Inorder to remove the fine gangue within the coal matrix, it

13 ‘

is necessary to grind the coal before cleaning. The smallparticle size produced by grinding requires a cleaningtechnique that can handle the ultrafine sizes. Therefore,microbubble floation is being used for the production ofsuperclean coal.

14

2.2 Literature Review

The use of micron-sized bubbles in mineral processingwas described by Sebba and Yoon (1981). They used bubblesgenerated by a modified glass aspirator (Sebba, 1971) tofloat mineral fines. It was suggested that the micron-sized bubbles floated material in two manners: (1) bybubble-particle attachment, and (2) by the flotation offlocs using numerous bubbles. According to the authors,the second manner of flotation was most likely to occur.Therefore, a quiescent flotation zone was needed to preventthe destruction of the flocs and to inhibit the removal ofthe bubbles from the solid surface. For this reason, thebubbles were generated outside of the flotation cell.

It was suggested that the most likely way to achieveseparation between valuable minerals and gangue materialwas to use a combination of oil agglomeration and micron-sized bubble flotation. A preliminary investigation wasconducted on the separation of fine coal from ash materialusing kerosene as the oily collector. They found that themicron-sized bubbles could be used for mineral flotation.

A study on the cleaning of fine coals using colloidal

bubbles (called microbubbles) prepared with MIBC andDowfroth 250 was completed (Yoon and Miller, 1982). Twotechniques were used in testing the microbubble flotation.The first was to use the small bubbles along with larger

y 15

bubbles produced in a conventional laboratory flotationmachine under turbulent conditions. The second was to usemicrobubbles alone under quiescent flotation conditions.

Using the first technique, a combination of theconventional bubbles and microbubbles, with a -100 meshPittsburgh No. 8 coal, the recovery was generally the sameas when using conventional flotation by itself. However,the ash rejection was improved by 1.8%. On a -200 meshsample of the same coal, the combination of large and smallbubbles resulted in a higher recovery and an improved ashrejection.

In the second technique, the microbubbles wereintroduced without any mechanical agitation duringseparation and without the larger conventional bubbles.The results of the microbubble separation were compared tothat of oil agglomeration. The oil agglomeration resultswere better in terms of separation efficiency, however,there was a difference of only 6.2% between the twoprocesses. This was encouraging for the microbubbletechnique, since in this method, only 4 lb/ton of kerosenewas used as compared to 200 lb/ton used in oilagglomeration.

The work was further continued using the microbubbleflotation technique under quiescent conditions (Yoon,1982). It was shown that variables such as pulp density,frother addition, collector dosage and bubble dilution

116

affect the flotation results. By increasing the frotheraddition, it was found that the ash percentage in the cleancoal product increased. Varying the number of microbubblesper unit volume of suspension produced a decrease in coalrecovery. However, there was a substantial drop in theash in the clean coal product. The effect of pulp densitywas tested over the range of 2-6%. As it increased, thecoal recovery decreased and the ash percent increased inthe clean coal product. when the kerosene addition wasvaried, the coal yield increased up to 8 lb/ton of keroseneand then remained constant with additional dosage. The ashpercentage decreased at first by 3% up to a 12 lb/tonaddition and then increased by 5% as kerosene levels becameexceedingly high (25 lb/ton). It was also found that theseparation efficiency was 22% better for microbubbleflotation than for conventional flotation on a -400 meshcoal sample.

The microbubble generator previously described (Sebba,1971; Sebba and Yoon, 1981) was made from ground glassjoints. This was determined to be impractical on theindustrial scale. Therefore, a new generator was developed

and used in future work (Yoon, 1984). For proprietaryreasons, the bubble generator description was notpresented.

Further work considered the various parameters

17

involved in microbubble flotation (Trigg, 1984). Theobjectives of this work were to improve the flotation ofbituminous coal fines and to study the mechanisms involvedin generating small bubbles and floating fine particleswith microbubbles. Effects of frother addition, collectoraddition, pulp density, slurry pH and particle size werefurther investigated in this study. A comparison ofmicrobubble flotation to conventional flotation and oilagglomeration was completed. It was found that microbubbleflotation is more efficient in floating fine coal thanconventional flotation and is comparable in separationefficiency to oil agglomeration. However, compared to oilagglomeration, microbubble flotation uses less reagents.

Microbubble flotation was used to clean micron—sizedcoal produced by a stirred ball mill (Yoon et al., 1984).The results obtained showed that the microbubble flotationtechnique cleaned coals with improved recovery and ashrejection as compared to the conventional flotation processwhich uses larger bubbles. The improved recovery waspartially explained by the increased flotation rateobtained when using smaller bubbles. The improvedselectivity of microbubble flotation was attributed tofavorable hydrodynamic conditions. Microbubbles have noturbulent wake as compared to larger bubbles; therefore,they do not entrain as many ash particles in the froth.Also, the authors suggested that the increased bubble

18 “1

loading with decreasing bubble size could account for theimproved selectivity.

Further investigation again showed the increasedrecovery and selectivity of microbubbles over largerbubbles (Luttrell et al., 1985). In this report, however,the flotation experiments were performed with a continuousmicrobubble flotation column instead of a batch flotationsystem. The improved recovery was attributed to theincreased flotation rate and the more conducivehydrodynamic conditions for bubble/particle collisionsinvolved with micron—sized bubbles. It was againemphasized that the improved selectivity of microbubbleswas due to absence of a wake with microbubbles whichentrains ash particles during flotation. However, it wasspecifically shown that bubbles of less than 300 microns indiameter have no wake.

The use of microbubbles for fine coal flotation wasshown to result in an increased flotation rate (Yoon andLuttrell, 1986). This improvement in flotation rate wasattributed to the increased probability of collision whichincreases with decreasing bubble size. The values ofprobability of collision were determined for fine coal.

1F

19

2.3 Experimental

2.3.1 Coal Samples

The coal samples used in the present work are listedin Table 2.1. Most of the samples are run—of—mine coal,however, several are clean products from preparationplants. Also, two samples of unknown origin were obtainedfrom United Coal Company. Due to the presence of a reagentaroma, it was assumed that these samples had undergone sometype of pretreatment.

when the coal was recieved, it was reduced to -1/4inch by a laboratory roll crusher, split into 100—gramsamples and placed in air-tight plastic bags for storage ina freezer at -20OC to minimize oxidation. Prior to eachflotation experiment, a 100—gram sample was taken from thefreezer and reduced to -65 mesh by a laboratory hammermill. It was then wet—ground to ultrafine sizes in astirred ball mill, (Brown et al., 1984; Mankosa et al.,

1986) for a specified length of time. For microbubbleflotation, the wet-ground sample was divided into 20—gramlots by a slurry sample splitter.

In some instances, conventional flotation wasperformed on the coal sample before microbubble flotation.In this case, two 100-gram samples were reduced to -100

mesh and used as feed to the conventional cell. The cleanproduct from conventional flotation was then wet-ground in

20 ·

Table 2.1

Coal Samples Used in the Present Work

Feed AshSeam Locatlon Type (% weight)

Elkhorn United Coal Company. R—O—M 15.0V1rg1n1a

Jerlcho Consol, Clean 7.5Pennsylvania

Lower Cedar Grove Amvest Coal Company. R—O—M 1.4Virginia

Pittsburgh No. 8 Consol• Clean 6.5Pennsylvania

Splash Dam Westmoreland Coal Company, R—O—M 6.0V1rg1n1a

' Taggart Hestmoreland Coal Company. R—0-M 28.0V1rg1n1a

UCC Sample 1 United Coal Company. unknown 20.0Virginia

UCC Sample 2 United Coal Company, unknown 26.0V1rg1n1a

Upper Cedar Grove Amvest Coal Company, R-0-M 11.0V1rg1n1a

Upper FreeportU

Department of Energy. R—0-M 22.0Pennsylvania

21

the attrition mill for the specified length of time andsplit into 20—gram lots.

2.3.2 Reagents

Dowfroth M150 (Polypropylene Glycol Methyl Ether, Mw =400) supplied by Dow Chemical Company, was used in all ofthe microbubble and conventional flotation experiments asthe frother. It was previously shown that Dowfroth M150has the needed stability for the generation of microbubbles(Trigg, 1984). The collector used throughout thisinvestigation was kerosene provided by walrond OilDistributor of Roanoke, Virginia.

2.3.3 Eguipment

The microbubble generation system used in all of themicrobubble flotation experiments was previously used byYoon and his coworkers, 1984; Yoon, 1984; Trigg, 1984. In

this system (see Figure 2.1) a frother solution is added tothe reservoir (R) and circulated through the bubblegenerator (B) by a centrifugal pump (P). A small flow is

then bled off of the stream at valve (V) and pumped by aperistaltic pump (S) past the flotation cell (F) and backinto the reservoir. This enables a fresh flow ofmicrobubbles to be circulated by the cell at all times.

22E

WL

F

*1Z

Y lX e

W'WATER RESERVOIR In. - LAUNDER ... -F- FLOTATION CELL V RS "FER|STALTIC FUMPP ‘CENTRIFUGAL FUMP

_ B'BUBBLE GENERATORR -MICROBUBBLE RESERVOIRD,V,X,Y,Z," STDPCOCKS

E- D0

Figure 2.1 Schematic diagram of the microbubbleflctatioh system

·

23

The three—way stopcock (X) allows the microbubbles to flowto the cell and also to return to the reservoir. Once themicrobubble suspension is at equilibrium, the stopcock (Y)is opened so that the microbubbles can be injected into theflotation cell containing the coal slurry.

Conventional flotation work was performed using theDenver laboratory flotation machine, Model D—12. A 4—literflotation cell was used in all experiments.

2.3.4 Procedure

a. Microbubble Flotation

For microbubble flotation, the coal samples wereprepared as previously mentioned. The slurry samples foreach individual flotation experiment contained 20 grams ofcoal with 400 ml of water; therefore, the slurry was at4.7% solids by weight. The sample was then conditioned ina high—speed blender for 3 minutes to insure the dispersionof the fine particles. Kerosene was added to the slurry bymeans of a microliter syringe and emulsified by againagitating the slurry for another 3 minutes in the high-speed blender.

After conditioning, the coal slurry was placed in theflotation cell while microbubbles were being generated.

The frother solution was circulated throughout the bubble

generation system for 1 minute to insure a stable

24 11

microbubble suspension. Once the microbubbles were formed,a desired volume was injected into the flotation cellcontaining the coal slurry. The slurry was allowed tostand for 6 minutes after the microbubbles were injected.This permitted the bubbles to rise and carry coal particlesto the surface of the slurry so that a stable froth layercould form.

By flooding the cell with water from the reservoir (W,Figure 2.1), the froth was forced to flow over the top ofthe column into a launder. The tailings were drained fromthe bottom of the cell through stopcock Y. For multistageexperiments, the froth was repulped and the same procedurewas performed as above.

b. Conventional Flotation

In the conventional flotation experiments, 200 gramsof coal were used in a 4—liter flotation cell. The samplewas conditioned for a 10-minute period to ensure that itwas completely wet. A collector was then added andconditioned for the same period of time. Following this,the frother was added and conditioned for 1 minute.

Flotation commenced by opening the air valve andcontinued until the hydrophobic coal particles weredepleted from the cell. For the second stage, the frothfrom the first stage was repulped and floated with no

A 25

additional reagents. As mentioned previously, the second-stage froth product was then ground in an attrition millfor further cleaning by microbubble flotation.

In all cases, the froth and tailings were filteredusing a Denver vacuum filter and dried in a constanttemperature oven. The samples were then analyzed for ashcontent using a Fisher automatic ash oven and sulfurcontent using a Leco infrared sulfur analyzer.

26

1 2.4 Results

2.4.1 Microbubble Flotation

Numerous Eastern U.S. coals were used in the presentwork to investigate the applicability of the microbubbleflotation technique. It has been shown that microbubbleflotation, as compared to conventional flotation, improvesboth recovery and selectivity of fine coal flotation (Yoon,1982; Yoon et al., 1984). Therefore, microbubble flotationis currently being used for the production of supercleancoal. Superclean coal is defined here as one containingless than 2% ash with a minimum amount of pyritic sulfur.The following experimental results are examples of theproduction of superclean coal by the microbubble flotationtechnique. All data presented in the figures are tabulatedin the Appendix along with the reagent additions used inthe tests.

Figure 2.2 shows a recovery vs. ash curve for aTaggart seam coal obtained from the westmoreland CoalCompany of Virginia. The coal was pulverized in alaboratory stirred ball mill for a 15—minute period,resulting in a mean particle size of 5-6 microns. Threestages of flotation were performed on the coal sample whichassayed 28% feed ash. The first stage of microbubble

flotation reduced the ash level to 6.6% with 91.6%combustible recovery. After two additional stages of

27

I uoo -———- =T: u2-%.. 80nä; {X"’ 60>·EUJ

40ULIJ .CK__] 20

·

C TAGGART SEAMU0

O 5 I0 I5 20 25 30ASH IN CLEAN COAL (% WEIGHT)

Figure 2.2 Recovery versus ash curve for themicrobubble flotatioh of coel from theTeggert seem

28 L

flotation, the ash content was reduced to 1.2% with arecovery of 74.9%.

Similiar results are shown for the flotation of UpperCedar Grove coal in Figure 2.3. In this case, however, agrinding time of 30 minutes was used in an attempt toliberate more of the finely disseminated mineral matter.This resulted in a mean particle size of approximately 3microns. After two stages of microbubble flotation, theUpper Cedar Grove sample was reduced from a feed ash of 11%to a product ash of 1.4% with a 52.2% combustible recovery.Compared to the flotation of the Taggart seam above, thecombustible recovery of the Upper Cedar Grove wassubstantially lower. This may possibly be attributed tothe increased particle surface area, created by the longergrinding time, which causes flotation to require a largerfrother dosage to achieve high recovery.

In an attempt to study the effect of pre—cleaning onthe final product obtained with microbubble flotation, anElkhorn seam coal sample, containing 15% ash, was cleanedin a magnetite bath at a specific gravity of 1.3. Theresulting product, containing 6% ash, was then ground in astirred ball mill for 15 minutes and subjected to threestages of microbubble flotation. The results of thisexperiment are shown in Figure 2.4. In the first stage offlotation, the ash level in the coal was reduced to 2.6%

29

I00 - : _%I-I

80LLJ

IX" 60>-I ILU

40ULUOS_| 20< UPPER CEDAR GROVE SEAM

~0 2 4 6 8 I0 I2ASI-I IN CLEAN COAL (% WEIGHT)

Figure 2.3 Recovery versus ash curve for themicrobubble flotatioh of coal from theUpper Cedar Grove seam

30 _I I

Ioo Z Z -Y: uI nSD. 80 JL1.!EX" 60)·ML1Jg 40UL1JM_] 20g ELKHORN SEAMO

O -0 I 2 3 4 5 6ASH IN CLEAN COAL (% WEIGHTI

Figure 2.4 Recovery versus ash curve for themicrobubble flotation of coal from theElkhorh seam

l 31 }with a recovery of 99.6%. This shows over a 50% rejectionin ash with essentially no loss of combustible material.After the two following stages of flotation, the coalrecovery was 85.9% with an ash level of 1.8%. with anincreased grinding time, the ash level in the concentratemay have been lowered further. However, the reducedparticle size could also have lowered the combustiblerecovery due to the increased surface area which requiresmore surfactant.

A similar test was carried out on the Splash Dam seamcoal containing 6% ash. This coal was cleaned at 1.3specific gravity in a magnetite bath prior to flotation.The clean coal sample assaying 3.5% ash was wet-ground in astirred ball mill for 30 minutes. The pulverized coal wasanalyzed for particle size distribution by the Elzoneparticle size analyzer. The mean particle size wasdetermined to be 3-4 microns. The coal sample prepared assuch was subjected to three stages of microbubbleflotation. The resulting froth product assayed 1.6% ashwith 66.5% combustible recovery as shown in Figure 2.5.Considering the low feed ash of the coal, the ash reductionwas substantial in this test. As the feed ash to a coalcleaning process is reduced, it becomes more difficult toclean the coal further. A major benefit of the microbubble

flotation process is that with proper liberation, low ashcoals can be cleaned to even lower ash levels.

32

I

I00 : If':

I3ESQ I"’ 60)·EUJg 40 I yULLIE_| 20

33 I

Another example of this can be seen with the LowerCedar Grove coal (Figure 2.6). This particular coal has arun—of—mine ash level of 1.4%. A representative sample ofthis coal was attrition ground for 15 minutes before it wassubjected to microbubble flotation. Several stages offlotation were performed on the coal slurry which resultedin an approximate ash reduction of 80%. This produced afroth product with an ash level of 0.3% and a combustiblerecovery of 63.1%. These results demonstrate thatmicrobubble flotation can produce superclean coal from alow ash feed coal.

2.4.2 Microbubble Flotation After Conventional Flotation

In coal preparation, each method of cleaning coal hasits own optimum operating range. The optimum method ofwashing coarse coal is the heavy media circuit. The finecoal circuit uses heavy media cyclones and conventionalflotation. Therefore, it is important to wash coal by themost economical method before cleaning it with a moreelaborate technique.

For this reason, it would be ideal to use the product

from a conventional flotation circuit as the feed tomicrobubble flotation. A series of experiments were

carried out to demonstrate this strategy.

Figure 2.7 shows the advantage of supercleaning coal

34 E

100 -· '-=

F: I2): Ig Ig I‘·’ 60>-GUJ

40UL1JG_,| 20g LOWER CEDAR GROVE SEAMU

0 .0 0.3 0.6 0.9 I.2 I.5ASH IN CLEAN COAL (% WEIGHT)

Figure 2.6 Recovery versus ash curve for themicrobubble flotatioh of coal from theLower cedar Grove seam

35

I

I00 --A MB,-500 ME;S§I .. ·- - · " jlI'- „‘Ü , 'tlg CI,] Ö',.. 80 \CONV, 65 MESH

I cfX /;

60¤ \CONV, -500 MESH

E ILIJ

40OLIJ1_| 20

u1 36

using the microbubble flotation technique. In thisexperiment, two stages of conventional flotation wereperformed on a -65 mesh sample of Elkhorn seam coal, using0.5 kg/ton of kerosene and 0.1 kg/ton of Dowfroth M150.The concentrate from the second stage was split into threerepresentative samples. The first sample was floatedconventionally for two more stages with no additionalreagent addition. As shown in the figure, the four stagesof flotation reduced the ash level from 6% to 3.3% with80.6% combustible recovery. The second and third sampleswere recombined, then attrition ground for 30 minutes with1/16" stainless steel balls. The ground material was thenresplit into two samples.

The first of the two ground samples was cleaned byconventional flotation for three consecutive stages inwhich no additional reagent was used. The second samplewas floated in three stages by the microbubble technique.An additional 1.35 kg/ton of kerosene was added to thissample and 0.7 kg/ton of Dowfroth M150 was added at eachstage. As can be seen in the figure, the microbubbleflotation technique reduces the ash level in the coal to0.7% with 77.2% combustible recovery. The finely groundsample cleaned by conventional flotation reduced the ashlevel to 1.5% with 56.6% recovery. These results show thatmicrobubble flotation cleans ultrafine material with an

1 37

improved recovery and ash rejection over conventionalflotation.

The Upper Freeport coal, supplied by the Department ofEner9Y„ contained finely disseminated ash particles. Thesample was cleaned by two stages of conventional flotationbefore being micronized for 1 hour to obtain a meanparticle size of 1.5 microns. After grinding, the samplewas cleaned by microbubble flotation. Figure 2.8 shows therecovery versus ash curve for the experiment. The curveshows that the coal could be reduced to superclean levelsby a combination of conventional and microbubble flotation.The resulting ash level was 1.8% with 73.2% combustiblerecovery.

Figure 2.9 shows the recovery versus ash curve for acoal sample from United Coal Company. The feed ash of thecoal was 20.3%. After two stages of conventionalflotation, the froth product was ground in a stirred ballmill for 30 minutes to achieve a 3-4 micron mean particlesize. Three stages of microbubble flotation were thenperformed on the sample. The complete cleaning processreduced the ash level to 1.8% with a combustible recoveryof 60.4%. In this experiment, the sulfur content in thefeed and concentrate was determined. The feed sulfur was1.00% and was reduced to 0.93% by the conventional

flotation. Microbubble flotation further reduced thesulfur level to 0.80%. Since the organic sulfur content

‘ II

I00 — „fäI-IQ so .LU uE ' IX‘·’ 60>-CKLIJg 40 IULU I CONVENTIONAL ._| 20

y 39

100éäI-I$9. 80ä

6h.0

>-CEUJ

4013Q: | c0NvE1~1T101~1A1._| 20

Uc

40

for this sample was reported by United Coal Company to beapproximately 0.80%, this result appears to indicate thatpyritic sulfur was completely removed.

The next flotation experiment involved another coalsample from the United Coal Company. This sample assayedat 26.4% ash before cleaning. Two stages of conventionalflotation were performed on the sample before attritiongrinding the coal for 1 hour. This reduced the meanparticle size to approximately 1.5 microns. Aftergrinding, four stages of microbubble flotation werecompleted on the sample. As can be seen in figure 2.10,the ash level was reduced to 2.9% with a 63.5% combustiblerecovery. The sulfur level in this sample was reduced from1.40% to 1.00%.

Figure 2.11 shows the cleaning of the Jericho coalseam provided by Consol Coal Company. As in previoustests, the coal was first subjected to two stages ofconventional flotation, then attrition ground for 30 minutesbefore floating it by the microbubble technique. Threestages of microbubble flotation were performed on the coalsample. The ash in the coal sample was reduced from 7.5%to 1.1% with 52.9% combustible recovery. The sulfur levelin this sample was reduced from 0.78% to 0.63%.

41

I00 ...I

I-IQ aoLIJg IB2 I"’ 60>-CK.UJ

40ULIJ I CONVENTIONAL'_| 20g S 0cc·N0.2U

00 4 8 I2 I6 20 24 28ASH IN CLEAN COAL (% WEIGHT)

Figure 2.10 Recovery versus ash curve for the UnitedCoal Company's Sample No. 2 usingconventional and microbubble flotation

III

42

I"""'t—'°

éäI-I$2 80ä II]Q3* 60 "uUJ

40OUJg I CONVENTIONAL_| 20g JERICH0 SEAMO

00 I 2 3 4 5 6 7 8ASH IN CLEAN COAL (% WEIGI-IT)

Figure 2.11 Recovery versus ash curve for coal fromthe Jericho seam using conventional andmicrobubble flotation

43 l

2.5 Discussion

2.5.1 Microbubble Flotation

The results presented in the foregoing section showthat the microbubble flotation technique is capable ofproducing superclean coal from an ultrafine coal feed. Ingeneral, clean coal can be produced containing less than 2%ash and in some cases, less than 1% ash. By usingconventional flotation followed by microbubble flotation,higher coal recovery with lower ash levels can be achieved.

Researchers have found that the microbubble flotationtechnique produces a higher recovery of fine coal ascompared to conventional flotation methods (Yoon, 1982;Trigg, 1984; Yoon et al., 1984). In the present work, ithas been necessary to grind coal to ultrafine sizes inorder to liberate the fine ash material from the coal.Therefore, the smaller bubbles used in microbubbleflotation are needed to capture the finer particles createdby stirred ball mill grinding.

The high recovery of fine particles by microbubbles iscaused by the increased probability of collection createdby the decreased bubble size (Yoon and Luttrell, 1986).

The larger bubbles used in conventional flotation producestreamlines such that the probability of collision withfine particles is not favored (Luttrell et. al., 1985).

Fine particles having small inertial forces tend to follow

44

these streamlines around the bubble while large particles,having stronger inertial forces, can deviate from thestreamlines and attach themselves to the bubble. Luttrell(1986) has shown that microbubbles produce streamlines thatallow an increased probability of collision with fineparticles.

Another possible reason for the improvement in fineparticle recovery with microbubbles is the reduced dragforce on the smaller bubbles. The drag force on a bubblerising through the pulp decreases as the bubble diameterdecreases. Fine particles may be swept away from thesurface of larger bubbles due to this force. Occasionally,fine particles are not attached to a bubble since they maynot have sufficient inertia to form a contact angle withthe bubble. These fines may be carried into the froth bycontactless flotation forces (Derjaguin and Dukhin, 1981).when viscous drag forces are larger than thebubble/particle attractive forces, the fine particles maybe swept away from the bubble surface.

Even though the microbubble flotation technique hasproven to increase the recovery of fine coal particles,this is not the only reason to use this technique forsupercleaning coal. Compared to conventional flotation,microbubble flotation is also more selective in floatingcoal over ash. (Yoon, 1982; Trigg, 1984; Yoon et al.,

451984). The factor that contributes most towards thisadvantage is that small bubbles create less turbulence asthey pass through the pulp than large bubbles. It has beenshown by Luttrell et al. (1985) that bubbles less than 300microns in size have no wake when rising to the surface ofthe pulp. when a wake forms behind a bubble, a lowpressure zone exists which enables the larger bubbles toentrain fine gangue particles and carry them to thesurface. with the 50-100 micron size bubbles that are usedin microbubble flotation, essentially no wake exists,thereby, reducing the opportunity for fine gangue to beentrained.

The improved selectivity of microbubble flotation mayalso be attributed to the fact that microbubbles create astable froth layer on top of the pulp. The water in thefroth is allowed to drain during flotation, which helps toremove mineral matter entrained in the froth. It is wellknown that the drainage mechanism in the froth is

responsible for the improved grade of the froth productwith increasing froth height.

Another mechanism responsible for the reduction of ashin the froth is the coalescence of bubbles in the frothlayer. If the bubbles in the froth coalesce, the overallsurface area of the bubbles is reduced. Therefore, it ispossible that the less hydrophobic ash particles are

crowded off of the bubble surface leaving only the more

46

)hydrophobic coal particles. Once the ash particles havebeen crowded from the bubble surface, they have theopportunity to drain from the froth during the time periodallowed for froth formation.

2.5.2 Microbubble Flotation After Conventional Flotation

Several microbubble flotation experiments have beenconducted on the froth product from conventional flotation(Figures 2.7-2.11). This would be a logical step inproducing superclean coal since it would be moreeconomical to first take out the easy-to—remove ash formingminerals using the less costly conventional flotationtechnique.

Figure 2.7 compares the results of conventionalflotation and microbubble flotation conducted on the frothproduct obtained from a two stage conventional flotationtest. The conventional froth product was micronized priorto the comparison tests. An optimum amount of reagents wasused in each test. For a fair comparison, no additionalreagents were used in the conventional flotationexperiments. Further addition of frother to the

conventional tests, would have produced an increased

recovery but, would also have increased the ash percentage

in the froth product. On the other hand, additional

reagent dosage is necessary for microbubble flotation since

47

the bubbles are generated outside of the flotation cell andthe surfactant is needed to aid in bubble generation.

It is clearly shown in Figure 2.7 that the recovery ofultrafine particles is greater in microbubble flotation.The results also show that by ultrafine grinding incombination with microbubble flotation, a cleaner productis obtained compared to that obtained with conventionalflotation. Therefore, microbubble flotation is a bettertechnique than conventional flotation for producing super-clean coal.

48‘

2.6 Summary and Conclusions

This investigation has been concerned with theproduction of superclean coal using the microbubbleflotation technique. Several Eastern U.S. coals have beentested to show the versatility of this technique. Inaddition, the effect of pre-cleaning the coal prior tomicrobubble flotation has been studied. As a result of thisinvestigation, the following conclusions can be drawn:

1) The microbubble flotation technique is capable ofproducing a superclean coal that contains less than 2%ash with high recoveries.

2) Various coals from different seams can be super-cleaned with the microbubble flotation technique.

3) Cleaning coal by conventional flotation, prior toultrafine grinding and microbubble flotation, allowscoal to be reduced to lower ash levels with highercombustible recoveries as compared to separation byconventional techniques alone.

uCHAPTER III _

Electrokinetics of Hydrocarbon Oil Emulsions

3.1 General

Although most coals can easily be floated with only afrother, the recovery of these coals can be enhanced by theaddition of hydrocarbon oils which increase the coal'shydrophobicity. However, it is important to use thecorrect amount of collector. An over—dosage of oil causesnon—selective recovery of gangue material, while aninsufficient amount of oil causes a reduction in clean coalrecovery. Not only is it important to use the properamount of oil, but it is necessary to have the correct oilfor the particular system.

Various oils have been used to improve the flotationof coal. The most common oil collectors for coal are:pine oil, creosols, cresylic acids, kerosene and dieselfuel. Recently, pure hydrocarbon oils such as heptane,hexane, octane and decane, have been used (Keller, 1984).

There has also been an interest in mixing industrial oilssuch as No. 6 and No. 2 fuel oil (wen and Sun, 1981).

The most important feature in using oily collectors isto ensure adsorption of the collector on the coal surface.

This requires the proper conditioning of the coal and oil.During conditioning, the oil must be reduced to small

49

450

droplets so that the probability of the coal particlescolliding with the oil is increased. Various surfactantshave been developed which help reduce the droplet size andstabilize the hydrophobic oil in the water. It has alsobeen found that by using these surfactants the flotationrecovery of the coal improves (Scanlon et al., 1983).

Determining the electrokinetic potentials of the oildroplets and the coal particles is essential in studyingthe attachment of the oil to the coal. whether theparticle and droplet attach or repel is dependent on thesurface charge of each. It has been suggested that thesurfactants used to emulsify the hydrocarbon collectorshave an effect on the zeta potential of the oil droplets(Laskowski and Miller, 1984). This change in zetapotential may be responsible for increased oil—particleattachment and therefore, increased flotation recovery.

Little work has been done in determining the zetapotential of oil droplets and studying the effect of oilemulsifiers on the electrokinetics of these droplets.Therefore, the objective of this study has been to measurethe zeta potentials of the various kinds of oil dropletsand to study the effect of various emulsifying agents onthe zeta potentials of these oil droplets.

51 E1 3

3.2 Literature Review

3.2.1 Zeta Potential of Oil Emulsions

The charge of water droplets dispersed in oil and thecharge of oil droplets dispersed in water has been measuredby a cataphoretic method in order to study effects ofelectrolytes on emulsion stabilization (Cheesman and King,1939). These investigators found that water in oil has apositive charge, while oil in water has a negative charge.They considered that charged ions can only originate fromthe aqueous phase. Thus, in a water—in—oil emulsion, thecharge originates within the water droplet. Knowledge ofthe charges of droplets was found to be instrumental inemulsion stabilization.

Eilers and Korff (1939) showed that pure paraffin oildroplets in water exhibited negative zeta potentials. Theyfound that the emulsion stability is directly related tothe oil droplets. As the zeta potentials of the dropletsbecame more negative, the stability of the emulsionincreased. These investigators also studied the effect ofelectrolytes on the zeta potentials of oil droplets.

Further study of the charge of oil droplets in waterwas made by Chattoraj and Bull (1959) using Nujol oil.They employed an electrophoresis technique using a flatcell. It was found that the electrophoretic mobility ofNujol oil changed with varying pH and the concentration of

52

eumulsifying agent (sodium dodecyl sulfate).Burkin and Bramley (1963) studied the electrokinetics

of fuel oil using a capillary tube cell with an opticalwindow. It was shown that the negative zeta potential offuel oil droplets in water increased as the pH increased.with decreasing pH, the zeta potential decreased steadilyuntil a point—of—zero charge was reached beyond which thezeta potential became positive.

Mackenzie (1969) studied the effect of various long-chain reagents on the electokinetics of Nujol oil drops.It was found that sodium dodecyl sulfate, dodecylaminechloride and cetyltrimethylammonium bromide change the zetapotential of Nujol drops substantially when added to theemulsion. It was shown that the anionic surfactantsincreased the negative charge on the oil drops whilecationic reagents caused the oil to become positivelycharged for most pH values.

A microcataphoretic method was used to investigate theelectrophoretic mobility of oil droplets dispersed inaqueous solutions of electrolytes (Kundu and Ghosh, 1970).Toluene, benzene, liquid parafin and heptane oil were usedfor the experiments. It was found that an increase in theoil droplet zeta potential occurred as long—chainelectrolyte concentration increased.

The zeta potential of various pure hydrocarbon oils

53 V

and industrial oils were studied by wen and Sun (1981).The experimental work showed that hydrocarbon oil emulsionshave a negative electrokinetic potential under mostcircumstances. However, by mixing low—grade industrialoils with various contaminants, positive zeta potentialscan be obtained in the lower pH ranges. The effects ofvarious frothers, electrolytes and ions were alsoinvestigated. These additives caused various changes inthe electrophoretic mobility of the oil droplets.

3.2.2 Zeta Potential of Coal

The zeta potential of coal was studied and found tohave significant effects on coal flotation and filtration(Baker and Miller, 1968). It was found that the additionof various reagents changes the charge of pyrite particlesin coal allowing a separation to be possible by flotationbased on electric potential. The use of zeta potentialcontrol as a means of improving coal cleaning techniqueswas discussed.

Anthracite coal electrokinetics was investigated byCampbell and Sun (19706). They found that hydronium andhydroxyl ions are potential—determining ions for anthracitelithotypes. At high pH values, the zeta potential wasobserved to be negative, while in the low pH range,positive charges were observed.

“

54Theeffect of pH on the zeta potential of bituminouscoal was investigated by Campbell and Sun (1970b). It wasshown that the mineral matter in the coal lithotypes playsa major role in determining the zeta potential of theentire coal matrix. As the amount of mineral matterincreased, the point—of—zero charge became more alkaline.The authors also determined that hydronium and hydroxylions behave as potential—determining ions for bituminouscoal lithotypes.

wen and Sun (1977) studied the electrokinetics ofoxidized coal and compared the findings with flotationresults. It was found that as the extent of oxidation ofthe coal increased, the zeta potentials became morenegative. By treating the oxidized coal with dodecylamine,the zeta potential became positive. This was found to havea beneficial effect on the flotation of the oxidized coal.wen and Sun also showed that the inorganic electrolytesreleased from the coal under different pH conditionssubstantially changed the zeta potentials of the coal.

Pretreatment of coal with potassium permanganate andtin chloride was investigated by Celik and Somasundaran(1980). They showed that conditioning the coal with these

chemicals affects the zeta potential and floatability ofparticles. The potassium permanganate and tin chlorideproduced a depression in the flotation recoveries and movedthe point of zero charge to higher pH values.

55

Mori et al. (1983) established a relationship betweenzeta potentials and the elemental composition of coal. Anequation was proposed to predict zeta potential as afunction of oxygen, nitrogen, carbon, hydrogen, sulfur andash content.

56 iiJ

3.3 Experimental

3.3.1 Coal Sample

The coal sample used in this study was provided byUnited Coal Company of Grundy, Virginia, and was taken fromthe Elkhorn seam. The coal was initially reduced to -1/4inch mesh by a laboratory roll crusher. The sample wasthen screened to remove the -20 mesh material. Afterscreening, the coal was cleaned in a magnetite bath with aspecific gravity of 1.3. This reduced the ash level from15 to 6%. The sample was then prepared as describedpreviously in Chapter 2.

For·determining the zeta potential of the coal, arepresentative sample was taken and reduced to -500 mesh.The fine size was needed so that the coal would not settleduring the determination of the electrophoretic mobility.

3.3.2 Reagents

Several oils were used in making emulsions for zetapotential measurements. These oils are listed in Table3.1. The oils were used as received from the distributor,

without further purification. Various oil emulsifiers,listed in Table 3.2, were also used in this study.

All pH adjustments were made by using reagent gradehydrochloric acid (HCL) and sodium hydroxide (Na0H).

57 q1

Table 3.1Descr1pt1on of Collectors Used 1n Zeta Potent1a1 Exper1ments

Q.¤.1.1.a.c.t.¤.r;s Q.h.em.1.s:.a.1£¤.x;m.u.1.a 1-;..11.. $.1:3.1.1.1; EMLI.1:1 §.¤.u.:.c.eHexane CH3(CH2)4CH3 86.18 L1qu1d R.G. A1dr1chOctane CH3(CH2)6CH3 114.23 L1qu1d R.G. A1dr1chDecane CH3(CH2)8CH3 142.29 L1qu1d R.G. A1dr1chUndecane CH3(CH2)9CH3 156.31 L1qu1d R.G. A1dr1chHexene CH3(CH2)2CH=CH2 84.16 L1qu1d R.G. A1dr1chHex1d1ene H2C=CHCH2CH2CH=CH2 82.15 L1qu1d R.G. A1dr1chKerosene unknown ·-- L1qu1d Ind. Walrond 011

Wh1te Kerosene unknown —-— L1qu1d Ind. Walrond 011‘No. 3 Fuel 011 °(D1ese1) unknown --- L1qu1d Ind. Walrond 011

nNo. 6 Fuel 011 unknown --- L1qu1d Ind. Amer1canPharmacel

l R.G. = Research GradeInd. = Industr1a1 Grade

p 58 1

Table 3.2

Oescr1pt1on of Surfactants Used 1n Zeta Potent1al Exper1ments

¤.o.¤.1.¤.11.1.¤Span 40 See F1gure 3.7 501.59 Sol1d R.G. ICI Amer1cas• Inc.

Span 60 See F1gure 3.7 529.65 Sol1d R.G. ICI Amer1cas. Inc.

Tveen 40 See F1gure 3.8 1323.65 L1qu1d R.G. ICI Amer1cas. Inc.

Tveen 60 See F1gure 3.8 1351.71 Sol1d R.G. ICI Amer1cas. Inc.

An.L¤.¤.1..¤Sod1um DodecylSulfate _ Cl2H25OS03Na 288.38 Sol1d R.G. BHO Chem1cal. Ltd.

Oodecylam1neHydrochlor1de CH3(CH2)llNH2HCl 221.82 Sol1d R.G. Eastman

R.G. = Reagent Grade

59|

Sodium chloride was used as the electrolyte and double-distilled water was used in all of the experiments.

3.3.3 Eggipment

All electrokinetic measurements were performed using azeta meter as developed by Zeta Meter, Inc. of New York.The apparatus used a Riddick cell with a molybdenum anodeand a platinum—iridium cathode. The temperature of eachsolution was determined before each measurement and theappropriate correction was made using the table provided inthe instruction manual. Electrophoretic mobility wasconverted to zeta potential by using the Helmholtz-Smoluchowski equation.

3.3.4. Procedure

The oil-in—water emulsions were prepared according tothe procedure given by wen and Sun (1981). For eachelectrokinetic measurement, 0.05 grams of oil was added to250 ml of double—distilled water and shaken by hand for twominutes. when surfactants were added to stabilize theemulsion, the solution was shaken by hand for another twominutes once the chemical was added. The emulsion was thendiluted to 1000 ml before being used for measurements.

A 100-ml sample of the oil emulsion was taken and

60 ”

I

adjusted to obtain the desired pH value. The temperatureof the sample was then measured before being transferred tothe electrophoresis cell. Once the emulsion had beenplaced in the cell, the electrophoretic mobility wasdetermined by recording the time required for each oildroplet to move a distance of 1 micron. Twentymeasurements were taken and averaged to determine themobility, from which the zeta potential value wascalculated by using the Helmholtz—Smoluchowski equation.

For the electrokinetic measurements of coal, 0.1 gramsof -500 mesh coal was added to 250 ml of double-distilledwater containing 0.6 grams of NaCl and shaken by hand fortwo minutes. The solution was then diluted to 1000 mlbefore taking the measurement using the same procedure asdescribed for the oil—in—water emulsions.

6l

ß3.4 Results

3.4.1 Effect of Hydrocarbon Chain Length

Figure 3.1 shows the zeta potentials for various purehydrocarbon oils plotted as a function of pH value. Asshown, the oil droplets are negatively charged throughoutthe entire pH range studied. These results are inagreement with those of other researchers e.g. Cheesman andKing, (1939); Eilers and Korf, (1939); Burkin and

Bramley, (1963); Mackenzie, (1969); wen and Sun, (1983).

Parreira and Schulman (l96la) attributed the negativecharge to the preferential adsorption of OH' ions to thatof H+ ions. Yordan and Yoon (1986) ascribed the

preferential adsorption to the differences in the hydrationenergies of H+ and OH' ions.

It is shown that the zeta potentials of the oilsbecome more negative as the hydrocarbon chain lengthincreases. As the chain length increases from 6 to 11, thenegative zeta potential is shown to increase significantly.These results indicate that the negative zeta potentialincreases with increasing CH2/CH3 ratios, which in turnsuggests that the CH3 group is more electronegative than

the CH2 groups.

62

20 .

O .... -.....------------- ---—----—-—fä>

-200gg 0

" -40 · QI-E-I -¤

·I-‘6°

0G uÜ- ‘~- II·

ÜOE• 0844,8N ·¤¤ ¤ @46**2.2 ' •„.l CHH24 ¤

·|2O4 3 5 7 9 II

pH VALUE

Figure 3.1 Zeta potentials of pure hydrocarbon oils(hexane, octane, decane, undecane) as afunction of pH value

1 63 I3.4.2 Effect of Double Bonds in the Hydrocarbon Chain

As shown in Figure 3.2, an increase in the number ofC=C double bonds has a significant effect on the zetapotential of the oil. The electrokinetic potential becomesless positive in the higher pH range by increasing thenumber of double bonds in the oil structure. This may beattributed to the double bonds being more electronegativethan single bonds, which discourages the adsorption of OH-ions on the surface of the oil droplets.

3.4.3 Effect of Industrial Oils

Figure 3.3 shows the potentials of various industrialoils. It can be seen that kerosene, white kerosene and No.3 fuel oil (diesel) have essentially the same zetapotential. However, the mixture of No. 3 and No. 6 fueloil has a less negative zeta potential. wen and Sun (1986)suggested that the impurities in the No. 6 oil has asignificant effect on its zeta potential behavior. No. 6oil could not be studied by itself due to its inability toemulsify into distilled water.

3.4.4 Effect of Emulsifiers on Kerosene

a. Effect of Nonionic Emulsifiers

The effect of nonionic emulsifiers on the zeta

Y 64

20 E

Q -—-—- —--——-—----——-——-----—---Ä

>E ·2J L ‘

¤F-

°4Q

-

I

-E ¤ « · E'_ *60O0..4 *80 Y ..I-LUN *I00 I HEXENE(I DOUBLE BOND)

E] HEXIDIENE(2 DOUBLE BONDS)*I20

1 3 5 7 9 IIpH VALUE

Figure 3.2 Effect of C-C double bonds in pure —hydrocarbon oils on zeta potential asa function of pH value

p 65 {

6O .Ü KEROSENE

·I WHITE KEROSENEA ZG O NO. 3 FUEL OIL(DlESEL)

> Q NO. 3, NO.6 MIX (4:I)Eäé ••**•"··—··——-———-———·—————-—---—--

..IS -20 QI·-. Z ÜLLI*6 -60 0q_ EI

l 66 Ä

potential of kerosene can be seen in Figure 3.4. Fourdifferent emulsifiers were used at a concentration of 10-5moles/liter. As shown, the addition of nonionicemulsifiers reduces the negative zeta potential of the oil.Also, as ethylene oxide groups are added to the structureof the emulsifier (Tween 40 and Tween 60), the reduction inzeta potential becomes more significant. This suggeststhat the hydrocarbon tail is adsorbed onto the surface ofthe kerosene droplet by hydrophobic bonding, leaving thenonionic—hydrophilic polar head (PEO group) oriented towardthe aqueous phase. The polar head may be stronglyelectronegative due to the presence of oxygen and henceprevent the adsorption of OH- ions.

b. Effect of Ionic Surfactants

Figure 3.5 gives the results obtained usingdodecylamine hydrochloride and sodium dodecyl sulfate asemulsifiers for kerosene. A concentration of 10-5moles/liter of surfactant was used in each experiment. Asshown, the anionic surfactant increased the zeta potentialin the negative direction, while the cationic surfactantincreased the zeta potential in the positive direction.Other investigators Cockbain, 1954; Anderson, 1958;Mackenzie, 1969; Kundu and Ghosh, 1970; reported similiar

results.

67

20

Q ---———————————-———————————--—f'§

> l°2O u

a' · ° GL- -40 n'u.1 ^ Wa}·* _6O A‘ l n

Q; ·@¤ ‘ ‘. . ¤ '|·— AKEROSENE “ ·· ILG E‘EäS8§ER%¢ä‘é% AOKEROSENE/T‘4OQKEROSENE/T"6O A ^

-I20 7 'I 3 5 7 9 IIpH VALUE

Figure 3.4 Effect ofoflw

emulsifiers on the zetapotential of kerosene as a function ofpH value ,"__ _ _

68

40 1

20{Ä

> 0 —-—-—---------- --- -.....----E\?_] -20S|·-· -40Z -60O ¤0- -60< E1 +

69

3.4.5 Zeta Potential of Coal

Figure 3.6 shows the results of the electrokineticmeasurements obtained for the Elkhorn Seam coal. As shown,hydroxyl and hydronium are the potential—determining ionsfor this coal sample and its PZC is found to be pH 3.2.This PZC is slightly lower than that of unoxidized ash—freecoal, which is usually found to be in the vicinity of pH4.5 (wen and Sun, 1981). The fact that the PZC of the

present sample is somewhat lower than that of the ash—freecoal may be due to the presence of mineral matter (6%).

7040

20

SE 0 ---—------—-—--——------.........-al *20I-Z *40UJI-O -60CL

71 I

3.5 Discussion

The results presented in the foregoing section showthat the electrokinetic potential of oil droplets can bedetermined using a microelectrophoresis technique. Ingeneral, oil droplets in water have a negative zetapotential. However, heavier industrial oils tend tocontain contaminants which lessen the negative zetapotential. Nonionic emulsifiers containing electronegativeoxygens tend to reduce the negative zeta potential. On theother hand, the charges of the oil-in—water emulsionsstabilized by ionic surfactants follow the sign of thesurfactant used.

3.5.1 Effect of Hydrocarbon Chain Length

It has been shown in Figure 3.1 that pure hydrocarbonoil droplets in water have a negative charge over most ofthe pH range studied. As mentioned previously, thenegative charge is most likely due to the preferentialadsorption of OH' ions. (Parreira and Schulman, 1961a;Yordan and Yoon, 1986).

The zeta potential vs. pH curves shown in this chaptershow that as the pH is reduced from alkaline to neutral pH,the zeta potential is reduced gradually. As the pH isfurther reduced, the zeta potential increases suddenly

until pH 5-6 before it begins to decrease again. This

72

*

‘discontinuity has also been observed by other investigators(Parreira and Schulman, 1961b; wen and Sun, 1981), andexplained by the adsorption of Cl" ions. The Cl' ions havebeen introduced into the system as HC1 which was used as apH regulating agent. This means that the Cl' ions adsorbon the oil droplets in preference to OH- ions. Athermodynamic explaination for this may be that Cl" ionsare less strongly hydrated then the OH- ions in bulksolution phase. According to the "best" thermodynamic datacomputed by Conway (1978), the enthalpy of hydration of OH-

and Cl“ ions are -446.8 and 363.6 kJ/mole, respectively.It has been shown in the present work that the zeta

potential becomes less negative with decreasing hydrocarbonchain length. Since the ratio of the CH3— to CH2— groupsincreases with decreasing chain length, it may be suggestedthat CH3—groups are more electronegative than CH2—groups.This is the only explaination that can be given for thedecreasing adsorbability of OH- ions on the surface ofshorter chain oil droplets, and hence the decreasing zetapotential. The difference in electronegativities of theCH3— and CH2-groups can be attributed to the inductiveeffect of alkyl chains.

73 2

3.5.2 Effect of Double Bonds in the Hydrocarbon Chain

In Figure 3.2, it is shown that with an increase inthe number of carbon—carbon double bonds in the oilstructure, the negative charge of the oil is reduced.Carbon—carbon double bonds in hydrocarbon structuresacquire more polar character due to the presence of anextra electron pair. This electron pair will impartelectronegativity to the hydrocarbon chain and reduce theadsorbability of OH- ions.

3.5.3 Effect of Industrial Oils

Figure 3.3 shows the results obtained from tests onthe zeta potential measurements of various industrial oils.Kerosene, white kerosene and No. 3 Fuel oil (diesel fuel)have essentially the same composition. The main differencebetween these oils is the temperatures at which they wererefined. white kerosene is the cleanest of the three oils,boiling away from crude oil at the lowest temperature.Kerosene follows next in purity, and last is diesel fuel(Rossini, 1953). As can be seen, these three oils show

approximately the same zeta potential behavior.

In order to emulsify the No. 6 fuel oil, it wasnecessary to mix it with No. 3 fuel oil. In the presentwork, 4 parts of No. 3 fuel oil were mixed with 1 part ofNo. 6 fuel oil. The mixture created an oil with a less

Q 74 Q5

negative zeta potential. This may be attributed to thevarious impurities associated with the No. 6 fuel oil.These impurities are polar compounds of undeterminedpolycyclic structures that contain amine, sulfur, asphalt,nickel and vanadium salts (wen and Sun, 1981).

3.5.4 Effect of Emulsifiers on Kerosene

a. Effect of Nonionic Emulsifiers

As shown in Figure 3.4, the addition of nonionicemulsifiers reduces the negative charge on the kerosenedroplets. Four different emulsifiers were tested in thepresent work. The chemical structures of these emulsifiersare shown in Figures 3.7 and 3.8. Span 40 and Span 60 are

similiar in structure except for differences in hydrocarbonchain length. Tween 40 and Tween 60 are also differentfrom each other by the number of hydrocarbon chains. Themajor difference between the Span and Tween surfactants isthat the latter contains 20 polyoxyethelene groups. Thereduction in negative zeta potential caused by the additionof emulsifiers can be attributed to the increased number ofoxygen atoms in the emulsifying agents. The oxygen atomsexposed on the surface of the oil droplets may reduce theadsorbability of OH- ions due to its high

electronegativity. The fact that Tween 20 and 40 reduce

75 „

SPAN 406066nmN M0~0s¤ALM¤mr6TH2 [H00H[ 0000TH [HT

HT·OOCCHZOOC ' CH3(CH2)[4 COOH

SPAN 60

H00H| 0COO?H [HTHT·OOC0++2000 — 0H3,6000H

Figure 3.7 Schematic diagram of the chemicalstructure of Span 40 amd Spam 60

76TWEEN 40

POLYOXYETHYLENE (20) SORBITAN MONOPALMITATE

CH2

H(I0(CH2 CH20)6HO,.4___)

H·

77 Ä

the negative zeta—potential more significantly than theSpan 20 and 40 may be attributed to the presence of PEGgroups.

Note that Span 60 gives slightly more negative zetapotentials than the Span 40. This difference may beattributed to the longer hydrocarbon chain length of Span60. This result is consistant with the results shown inFigure 3.1, which shows that the negative zeta—potentialincreases with increasing chain length. On the other hand,there are no discernable differences between Tween 40 and60. It seems that the PEO groups have the overwhelminginfluence in the charge mechanism.

b. Effect of Ionic Surfactants

Figure 3.5 shows the results obtained when ionicsurfactants, i.e. sodium dodecyl sulfate (SDS) and dodecylhydrochloride, were added to the kerosene/water emulsions.Using SDS to emulsify kerosene, the zeta potentialincreases in the negative direction. This can be explainedby the adsorption of the SDS at the oil water interface(Cockbain, 1954; Anderson, 1958; Mackenzie, 1969; Kundu

and Ghosh, 1970). The increased amount of negative chargeadsorbed at the interface causes a higher negative chargeon the oil droplets.

As shown in Figure 3.5, dodecylamine hydrochloride has

78

a significant effect on the oil droplet charge. A positivezeta potential is produced by the adsorption of thecationic surfactant on the oil in the lower pH range.However, this positive charge is reduced as the pHapproaches that of of dodecylamine precipitation. As thepH is further increased, the zeta potential becomesnegative because the amine is more active as a non—aminesurfactant.

3.5.5 Zeta Potential of Coal

The zeta potential of the Elkhorn seam coal can beseen in Figure 3.6. The results obtained are consistentwith previous studies (Campbell and Sun, 1970b; wen and ‘

Sun, 1977). As mentioned in section 3.3.1, the coal samplewas cleaned in a magnetite bath with a specific gravity of1.3; therefore, the sample consists mostly of vitrain(Thomas, 1968). The mineral matter in the coal sampleplays a major role in the point of zero charge of the coal.Structural anionic impurities would be expected to cause anacidic shift in the point of zero charge. However, themineral matter contained in the vitrain may also cause theacidic shift of the PZC.

79 HI

3.6 Summary and Conclusions

This investigation has been concerned with the studyof the electrokinetic behavior of hydrocarbon oil emulsionsand coal. The zeta potentials of pure hydrocarbon oils,industrial oils and oils with emulsifiers have beenmeasured. The results of this study may be summerized asfollows:

1) It has been suggested that CH3-groups are moreelectronegative than CH2—groups. This differencein electronegativities can be attributed to theinductive effect of alkyl chains.

2) Carbon—carbon double bonds are more electronegativethan carbon—carbon single bonds.

3) The zeta potential of kerosene and diesel fuel areessentially the same. However, by adding No. 6 fueloil to diesel, the contaminants in the heavy oil,lessen the negative charge of the oil droplets.

4) The reduction in negative zeta potential caused by theaddition of emulsifiers to oil—in—water emulsions can

be attributed to the increased number of oxygen atoms

in the emulsifying agents.

5) Ionic surfactants cause oil—in—water emulsions to take

80 Ä

on the charge of the anionic or cationic surfactant,producing a negatively charged or positively chargedoil droplet.

6) The point of zero charge was determined to be 3.2 forthe Elkhorn coal sample tested in this work.

The results obtained in this work show that the zetapotential of an oil is affected by its chemical structureand by the addition of reagents to the oil. In general,most oil droplets have a negative charge when emulsified inwater.

1

CHAPTER IV

The Effect of Hydrocarbon Oils on the Induction Timeand Microflotation of Coal

4.1 General

Froth flotation occurs when mineral particles attachto bubbles and rise to the pulp surface. This process isconsidered to occur in three basic phases (Derjaguin andDukhin, 1961; Laskowski 1974). They are:

I) the formation of a thin wetting film as the particleand bubble approach each other,

II) the thinning and rupturing of the wetting film, as theparticle and bubble collide, and attachment occurs,and

III) the formation of a stable bubble—particle aggregatecapable of withstanding considerable disruptive forcesoperating in the cell.

Phase I of the flotation process is dictated by thehydrodynamics of the system created by the particle, thebubble, and the flotation cell. Phases II and III are a

function of the surface chemistry of the system, especiallythe hydrophobicity of the mineral. In order for

81

82particle/bubbleattachment to occur, the contact anglebetween the two must be greater than OO and the contacttime during collision must be longer than the inductiontime.

The most important aspect of flotation is theattachment of the particle to the bubble. As the particleand bubble approach one another, a wetting film forms(Phase I). As the distance decreases between the particleand bubble, the wetting film begins to thin, and becomesunstable. The instability of this film causes it torupture after a definite period of time (Phase II). Thetime required for the film to thin and rupture is known asinduction time. As mentioned previously, if the contacttime between the bubble and particle is longer than therequired induction time, particle/bubble attachment willoccur.

Thus, induction time measurements provide kineticinformation for flotation. In this regard, induction timemeasurements are considered more beneficial for the studyof flotation than contact angle measurements which are athermodynamic quantity (Eigeles and Volova, 1960;Laskowski, 1974).

Little work has been done on the effect of oilemulsions on bubble—particle attachment. Therefore, theobjective of this work has been to study how bubble-particle attachment is effected by oil emulsions using

83

induction time and microflotation experiments as areflection of attachment.

y 84 :

4.2 Literature Review

Early researchers suggested that the thinning of theintervening liquid layer was important in the attachment ofparticles to air bubbles in flotation (Frumkin, 1933).However, the existence of induction time and its importancein flotation kinetics was first realized by Sven—Nillson in1934 (Leja, 1982). Since that time, several investigatorshave measured induction time using various techniques(Eigeles, 1950; Glembotsky, 1953; Evans and Ewers, 1953).

Eigeles and Volova (1960) carried out the firstextensive work on induction time. Their findings are asfollows: 1) induction time decreases with increasingcollector concentration; 2) induction time increases with

increasing particle size; and 3) induction time decreaseswith increasing temperature. Further induction timestudi