Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The floatability of coal and other inherently hydrophobic solids in relation to the surface tension of… Hornsby, David Theodore Burt 1981

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1981_A1 H67.pdf [ 17.02MB ]
Metadata
JSON: 831-1.0079033.json
JSON-LD: 831-1.0079033-ld.json
RDF/XML (Pretty): 831-1.0079033-rdf.xml
RDF/JSON: 831-1.0079033-rdf.json
Turtle: 831-1.0079033-turtle.txt
N-Triples: 831-1.0079033-rdf-ntriples.txt
Original Record: 831-1.0079033-source.json
Full Text
831-1.0079033-fulltext.txt
Citation
831-1.0079033.ris

Full Text

THE FLOATABILITY OF COAL AND OTHER INHERENTLY HYDROPHOBIC SOLIDS IN RELATION TO THE SURFACE TENSION OF AQUEOUS METHANOL SOLUTIONS by DAVID THEODORE BURT HORNSBY B.Eng., University of Melbourne, 1968 M.Eng.Sc., University of Melbourne, 1972 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Mining and Mineral Process Engineering) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September, 1981 (c) David Theodore Burt Hornsby, 1981 In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t of the requirements f o r an advanced degree a t the U n i v e r s i t y o f B r i t i s h Columbia, I agree t h a t the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and study. I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e copying of t h i s t h e s i s f o r s c h o l a r l y purposes may be granted by the head o f my department o r by h i s o r her r e p r e s e n t a t i v e s . I t i s understood t h a t c o p y i n g or p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be allowed without my w r i t t e n p e r m i s s i o n . Department of t\Ai The U n i v e r s i t y o f B r i t i s h Columbia 2075 Wesbrook P l a c e Vancouver, Canada V6T 1W5 i i Research Supervisor Jan Leja, Professor Dept. of Mining and Mineral Process Engineering ABSTRACT The r e l a t i v e f l o a t a b i l i t y of coal and other inherently hydrophobic solids i n aqueous solutions of short-chain n-alcohols was studied. An analysis of w e t t a b i l i t y data from the l i t e r a t u r e for such systems indicated that, i n general, a l i n e a r relationship existed between the solution sur-face tension, Y 1 v » and the adhesion tension, Y 1 V C O S 0 (where 0 i s the con-tact angle). Except for very non-polar s o l i d s , the c r i t i c a l surface tension of wetting, Y C (0= 0° intercept), determined with the alcohol solutions, was r e l a t i v e l y constant and independent of the type of s o l i d . The parameter which reflected the r e l a t i v e w e t t a b i l i t y of a s o l i d was the slope of i t s Y 2 V cos 0 vs. Y-^ v l i n e . Since w e t t a b i l i t y and f l o a t a b i l i t y are not necessarily synonymous, a concept of c r i t i c a l surface tension of f l o a t a b i l i t y , Y ^ , was developed to characterize the threshold f l o a t a b i l i t y condition where successful bubble-parti c l e attachment was controlled by either adhesion or aggregate s t a b i l i t y . In both cases i t was concluded that Y ^  would depend on the surface and the physical properties of a p a r t i c l e for a given set of f l o t a t i o n conditions, and that Y c f > Y . Under certain circumstances i t also appeared that selec-tive f l o t a t i o n separation between particles of two inherently hydrophobic s o l i d s , A and B, would be feasible in an aqueous alcohol solution i f Yc.p < Of Yi s Y f > even when the two solids had the same Y value. J. V C T c Small-scale f l o t a t i o n tests in aqueous methanol solutions were performed on narrow size fractions of several inherently hydrophobic s o l i d s . In most cases the wt.% floated decreased from a maximum to a minimum over a discrete range of implying that the particles in a sample had a range of i i i values. The r e l a t i v e f l o a t a b i l i t y of a sample was given by the position of i t s v x band along the Y, axis. For the same size f r a c t i o n , f l o a t a b i l i t y 'cf 3 l v decreased in the order: polytetrafluoroethyl ene (PTFE), sulphur, medium v o l a t i l e bituminous coal, molybdenite. Almost complete selective separation by f l o t a t i o n was achieved between samples whose Y c^ bands did not overlap. Because coal i s a very heterogeneous material, i t was expected that a sample of fine coal particles would display a wide range of w e t t a b i l i t y and f l o a t a b i l i t y c h a r a c t e r i s t i c s , with a corresponding broad range of y ^ values. This was v e r i f i e d in aqueous methanol solutions for narrow p a r t i c l e size fractions of several high rank coal samples from the East Kootenay c o a l f i e l d , B r i t i s h Columbia. Incremental ash c o n t e n t / f l o a t a b i l i t y relationships were developed which could be roughly divided into two main regions'. In one, the coal p a r t i c l e ash content was low and independent of f l o a t a b i l i t y , while in the other they were inversely related. The effects of physical and compo-s i t i o n a l factors other than ash content on the d i s t r i b u t i o n of f l o a t a b i l i t y within a coal sample have been discussed. For a readily floatable coal sample the yield-ash content relationships for the f l o t a t i o n products were independent of the f l o t a t i o n time and con-ditioning time, and were considered to approximate the ideal f l o t a t i o n wash-a b i l i t y characteristics of the sample. Flotation washability was found to be l i n e a r l y related to feed ash content. The shape and position of the incremental ash c o n t e n t / f l o a t a b i l i t y curves f o r different coal samples allowed the r e l a t i v e w e t t a b i l i t i e s of their organic matter to be assessed in terms of degree of oxidation, rank and petro-graphic composition. Coal sample f l o a t a b i l i t y in aqueous methanol solutions was found to be sensitive to small differences in surface properties which were not reflected by bulk compositional parameters. iv TABLE OF CONTENTS Page ABSTRACT i f TABLE OF CONTENTS fv LIST OF TABLES x LIST OF FIGURES x t i i LIST OF SYMBOLS x v i i i ACKNOWLEDGEMENTS xxi CHAPTER 1 INTRODUCTION 1 CHAPTER 2 WETTING IN RELATION TO FLOTATION 4 2.1 Basic Wetting Theory for Liquids on Solids 4 2.1.1 Interpretation of contact angles 10 2.1.2 Estimation of surface and i n t e r f a c i a l 15 tensions of solids 2.1.3 Surface pressure 21 2.2 C r i t i c a l Surface Tension of Wetting 25 2.2.1 Pure liquids on low-energy solids 25 . 2.2.2 Pure liquids on high-energy solids 32 2.2.3 Aqueous surfactant solutions on low-energy 33 sol ids 2.2.3.1 adsorption and w e t t a b i l i t y 42 2.2.4 Wettability and f l o t a t i o n of high-energy 46 solids in aqueous surfactant solutions 2.3 C r i t i c a l Surface Tension of F l o a t a b i l i t y 54 2.3.1 Particle-bubble adhesion 56 2.3.2 Bubble-particle aggregate s t a b i l i t y 67 2.3.3 Selective f l o t a t i o n separation of inherently 73 hydrophobic solids V Page CHAPTER 3 SURFACE PROPERTIES OF SOME INHERENTLY HYDROPHOBIC 77 SOLIDS 3.1 Theories of Inherent Hydrophobicity 77 3.2 Specific Surface Properties 82 3.2.1 Contact angle 82 3.2.2 Zeta potential 83 3.2.3 Flotation response 84 CHAPTER 4 WETTING AND FLOTATION OF COAL 86 4.1 Coal Compositional Properties 86 4.1.1 Rank 87 4.1.2 Type 88 4.1.3 Grade 92 4.2 Oxidation 94 4.3 Surface Properties 95 4.3.1 Rank 96 4.3.2 Mineral matter 98 4.3.3 Petrographic composition 98 4.3.4 Oxidation 99 4.3.5 Electrokinetics 100 4.3.6 Wetting and adsorption 101 4.3.7 Coal w e t t a b i l i t y d i s t r i b u t i o n 105 CHAPTER 5 OBJECTIVES AND SCOPE OF EXPERIMENTATION 108 CHAPTER 6 MATERIALS - PROPERTIES AND PREPARATION 110 6.1 Glassware 110 6.2 Solutions 110 6.2.1 pH 111 6.2.2 Surface tension 111 v i Page 6.3 Coal Samples 112 6.3.1 Description 112 6.3.2 Preparation 114 6.3.3 Proximate and ultimate analyses 116 6.3.3.1 bulk coal 116 6.3.3.2 -210 + 149 ym size fraction 116 6.3.4 Scanning electron microscopy 117 6.3.5 Infrared spectroscopy 120 6.3.6 Petrographic analysis 125 6.4 Non-Coal Materials 129 6.4.1 Description and preparation 129 6.4.1.1 polytetrafluoroethylene (PTFE) 129 6.4.1.2 sulphur 131 6.4.1.3 molybdenite 132 6.4.1.4 quartzite 134 6.4.2 Characterization 135 CHAPTER 7 EXPERIMENTAL DETAILS 138 7.1 Exploratory Test Tube Experiments 138 7.2 Small-scale Flotation Tests 139 7.2.1 Apparatus 139 7.2.2 Procedure 143 7.3 Treatment of Flotation Products 146 7.3.1 Drying 146 7.3.2 Analysis 147 CHAPTER 8 RESULTS AND DISCUSSION 149 8.1 Results of Exploratory Test Tube Experiments on 149 Coal Samples vi i Page 8.2 Results of Small-scale Flotation Tests on Coal 151 Samples 8.2.1 Conditioning time 152 8.2.2 Flotation time 153 8.2.3 Cumulative y i e l d versus f l o a t a b i l i t y 153 8.2.4 Cumulative y i e l d versus cumulative ash 159 content 8.3 Discussion of Results of Coal F l o a t a b i l i t y Tests 165 8.3.1 Flotation washability 165 8.3.2 Distribution of f l o a t a b i l i t y within a coal 171 sample 8.3.2.1 compositional/surface properties 174 8.3.2.2 physical factors and f l o t a t i o n 177 conditions 8.3.3 F l o a t a b i l i t y differences between coal samples 180 8.3.3.1 mineral matter (ash) characteristics 182 8.3.3.2 effect of weathering 189 8.3.3.3 rank and petrographic composition 196 8.3.4 General 197 8.4 The Relative F l o a t a b i l i t y of Some Inherently 199 Hydrophobic Solids 8.4.1 Results of exploratory test tube experiments 199 8.4.2 Results of small-scale f l o t a t i o n tests 202 8.4.3 Discussion of f l o a t a b i l i t y results 208 8.5 Selective Flotation Separation of Inherently 215 Hydrophobic Sol ids 8.5.1 Results of test tube experiments 215 8.5.2 Results of small-scale f l o t a t i o n tests 217 8.5.3 Discussion 221 v i i i Page CHAPTER 9 CONCLUSIONS 226 9.1 General 226 9.2 Coal 226 9.3 Other Inherently Hydrophobic Solids 227 CHAPTER 10 SUGGESTIONS FOR FURTHER WORK 230 REFERENCES 232 APPENDICES A SURFACE TENSION MEASUREMENT 257 B BULK COAL SAMPLE ANALYTICAL DATA 261 C STATISTICAL METHODS OF ANALYSIS OF ANALYTICAL AND 262 EXPERIMENTAL DATA C.l Testing Normality 263 C.2 Homogeneity of Variances 264 C.3 Single Factor Analysis of Variance (ANOV) 264 C.4 SNK Multiple Comparison Test 264 C.5 Confidence Interval 266 C.6 Correlation Coefficient 267 D STATISTICAL ANALYSIS OF C, H, AND N ANALYSIS DATA 268 E STATISTICAL ANALYSIS OF ASH CONTENT DATA 274 F INFRARED SPECTROSCOPY OF COAL SAMPLES 281 F.l Sample Preparation 281 F.2 Pel l e t Preparation 281 F.3 Recording the Spectra 282 F.4 Extracting Absorption Data from Spectra 282 G TREATMENT OF RAW DATA FROM PETROGRAPHIC ANALYSES 286 H METHODS FOR THE ANALYSIS OF COMPOSITION OF FLOTATION 289 PRODUCTS FROM SYNTHETIC MIXTURES ix Page H.l Sulphur/Molybdenite 289 H.2 Sulphur/Coal 289 J SMALL-SCALE COAL FLOTATION TEST CONDITIONS AND RESULTS 291 K STATISTICAL ANALYSIS OF COAL FLOTATION TEST YIELD DATA 299 L FLOTATION WASHABILITY DATA 308 M FLOTATION TEST RESULTS FOR NON-COAL MATERIALS 313 X LIST OF TABLES Tab! e tension, yiv» and p a r t i c l e s i z e , d; . p = 3 g cm-3. Pj = 1 g cm-3, n = 160,, r b = 0.02 cm s Page 2.1 Selected water contact angle, e, and c r i t i c a l surface tension of wetting, y , data for various low-energy solids (200 - 25°C) c 30 2.2 Calculated c r i t i c a l values of contact angle for bubble-p a r t i c l e aggregate s t a b i l i t y , as a function of surface 70 6.1 Analysis data for the -210 + 149 ym size fraction of coal samples used in f l o a t a b i l i t y tests 118 6.2 Petrographic analyses (vol.. %) and mean maximum v i t r i n i t e reflectance measurements, R max %, of particulate coal samples 127 6.3 Properties of non-coal materials tested in f l o a t a b i l i t y experiments 136 7.1 Surface tension of aqueous methanol solutions used in test tube experiments 138 8.1 Comparison of analysis, washability, f l o a t a b i l i t y and ash rejection potential data 186 8.2 Range of c r i t i c a l surface tension of f l o a t a b i l i t y for particles in samples of inherently hydrophobic materials in aqueous methanol solutions (including hydrophilic quartzite for comparison) 206 8.3 Results of f l o t a t i o n tests on mixtures of inherently hydrophobic materials in aqueous methanol solutions; 5 minutes conditioning, 3 minutes f l o t a t i o n , no bubble deflector 218 A. l Aqueous solution surface tension data, dyne/cm, 20°C 260 B. l Proximate and ultimate analysis data for the No. 5 seam bulk sample and the -595 + 0 ym size fraction screened from the adit 23 bulk samples 261 C. l Summary of source and variance calculation for single factor ANOV 265 D. l S t a t i s t i c a l analysis of carbon data (wt. % dmmf) 269 D.2 S t a t i s t i c a l analysis of hydrogen data (wt. % dmmf) 271 D.3 S t a t i s t i c a l analysis of nitrogen data (wt. % dmmf) 273 x i Table E.l Chi-square goodness-of-fit test of normality of calculated coal feed ash contents (wt. % a i r dried) 275 E. 2 Homogeneity of variances, single factor ANOV and SNK multiple comparison of means tests on calculated feed ash content data (wt. % a i r dried), -210 + 149 ym size fraction 279 F. l Absorption data for s p e c i f i c bands from infrared linear absorption spectra of coal samples 284 F. 2 Organic matter structure assigned to absorption bands in infrared spectra of coal (from references 212,229,286) 285 G. l Maceral analysis data (voK %.) and mean maximum v i t r i n i t e reflectance measurements (R max, %) of particulate coal samples 288 J . l No. 5 seam f l o t a t i o n test results 292 J.2 Flotation test r e s u l t s , 0' adit 23 sample, no bubble deflector, 3 minutes f l o t a t i o n 294 J.3 Flotation test r e s u l t s , 38' adit 23 sample, no bubble deflector, 3 minutes f l o t a t i o n 295 J.4 Flotation test r e s u l t s , 75' adit 23 sample, no bubble deflector, 3 minutes f l o t a t i o n , 15 minutes conditioning 296 J.5 Flotation test r e s u l t s , 1111 adit 23 sample, no bubble deflector, 3 minutes f l o t a t i o n , 15 minutes conditioning 297 0.6 Flotation test r e s u l t s , 150' adit 23 sample, no bubble deflector, 3 minutes f l o t a t i o n , 15 minutes conditioning 298 K.l Chi-square goodness-of-fit test of normality of replicate f l o t a t i o n test cumulative % y i e l d data, 75' adit 23 coal sample, 10 vol . % methanol 300 K.2 Homogeneity of variances of replicate cum. % y i e l d data 301 K.3 ANOV and SNK tests on cum. % y i e l d data, 10 vol. % methanol 302 K.4 ANOV and SNK tests on cum. % y i e l d data, 12.5 vo l . % methanol 304 K.5 ANOV and SNK tests on cum. % y i e l d data, 15 vol. % methanol 306 L . l Flotation washability data, No. 5 seam sample 309 L.2 Flotation washability data for the 75' adit 23 sample 310 XI 1 Table L.3 Flotation washabil i t y data for the 11.1'' adit 23 sample L.4 Flotation washability data for the 150' adit 23 sample M.l Cumulative wt. % y i e l d data for f l o t a t i o n tests on non-coal materials in aqueous methanol solutions, no bubble deflector, 5 minutes conditioning, 3 minutes f l o t a t i o n XI 1 1 LIST OF FIGURES Figure 2.1 Three-phase contact equilibrium for a l i q u i d drop on a s o l i d surface; (a) ideal conditions; (b) hysteresis on an inclined surface under gravity, after Wolfram and Faust (34); (c) edge effect at sharp corners, after Oliver et al . (43) 2.2 Zisman w e t t a b i l i t y diagram for a low-energy s o l i d ; 1 -pure l i q u i d s ; 2,3 - different types of behavior found for aqueous solutions of micelle-forming surfactants 2.3 Relationship between Y and surface chemical constitution of low-energy s o l i d s , after Shafrin and Zisman (82) 2.4 Wettability of paraffin by pure n-alcohols and t h e i r aqueous solutions; (a) Zisman diagram; (b) adhesion tension diagram; data from Good (67) and Fowkes and Harkins (105) 2.5 Adhesion tension diagrams for low-energy solids in aqueous surfactant solutions; data from (a) Fowkes (124) and (b) Bernett and Zisman (97) 2.6 Adhesion tension diagrams for t a l c , graphite and s t i b n i t e in aqueous surfactant solutions; • n-butanol, A n-butylamine; data from Fowkes and Harkins (105) 2.7 Adhesion tension diagram for A - PTFE, B - paraffin, C - polypropylene, D - polycarbonate in aqueous solutions of methanol, ethanol, n-propanol and n-butanol, from Wolfram (99); E - cellulose in ethanol solutions, from van Voorst Vader (114) 2.8 Wettability l i n e s for hypothetical solids A (non-polar), and B and D (increasing polarity) in aqueous solutions of a short chain n-alcohol i l l u s t r a t i n g selective wetting region (shaded) between A and B. 2.9 Adhesion tension diagram for a high-energy solid-micelle-forming c o l l e c t o r system (BaSO, - Na dodecyl sulphate) i l l u s t r a t i n g postulated changes in Y ^ (for pure liquids) as Y ] V and e change; 9 and Y - I v data from Schulman and Leja (142) and Elworthy and Rysels (143) 2.10 Selective separation of a mixture of magnetite and quartz made hydrophobic by conditioning in alkali n e dodecyl amine acetate solution, and subsequently floated in aqueous methanol solutions; after Finch and Smith (136) xiv Figure Page 2.11 Suggested relationships between induction time and (a) p a r t i c l e s i z e , adapted from Jowett (170), (b) surface tension of aqueous alcohol solutions and p a r t i c l e size 61 2.12 C r i t i c a l surface tension and contact angle of adhesion, and c r i t i c a l aggregate s t a b i l i t y lines and surface tension of s t a b i l i t y , for two p a r t i c l e sizes d' and d" of an inherently hydrophobic s o l i d with w e t t a b i l i t y l i n e A 65 2.13 Possible differences in c r i t i c a l surface tension of f l o a t -a b i l i t y for three inherently hydrophobic sol ids A, B and D in aqueous solutions of a short chain n-alcohol; (a) the same for each s o l i d ; (b) e A ^ < e B ^ < e D f ; ( 0 e g f < e B f < e » f c f c f 74 2.14 Selective separation region between c r i t i c a l f l o a t a b i l i t y bands for two inherently hydrophobic solids with wetta-b i l i t y bands A and B 76 4.1 Variation of organic matter elemental composition for high rank whole coal, W, e x i n i t e , E, v i t r i n i t e , V, and i n e r t i n i t e , I, as a function of rank; after Mazumdar (217) and Kessler (216) 89 4.2 Broad w e t t a b i l i t y band postulated for a particulate sample of a high rank readily floatable coal 106 6.1 Stati c surface tension of aqueous methanol solutions, 20°C 113 6.2 I l l u s t r a t i o n of sample c o l l e c t i o n points in adit 23, No. 7 seam Greenhills area, East Kootenay C o a l f i e l d , southeast B r i t i s h Columbia, after Teo et al. (284) 115 6.3 SEM images (40X) of the -210 + 149 ym size fraction of the coal samples used in f l o a t a b i l i t y experiments i l l u s t r a t i n g increased f r i a b i l i t y and slimes content of 0' sample 119 6.4 Infrared spectra (% Transmission) from KBr pellets of -210 + 149 ym size fraction of coal samples, including baselines for extracting absorption data, showing changes in 1700, 3430 and 1600 c n r l band i n t e n s i t i e s 121 6.5 Infrared absorption data at s p e c i f i c bands for the six coal samples (duplicate KBr p e l l e t s , -210 + 149 ym size f r a c t i o n , 1.5.mg-coal/300 mg) 122 XV Figure Page 6.6 SEM images of samples tested in f l o a t a b i l i t y experiments, 40X 137 7.1 The a l l - g l a s s "frothless" Partridge-Smith c e l l used for small-scale f l o t a t i o n t e s t s ; (a) assembled; (b) exploded; (c) basic operation, after Stewart (336) 140 7.2 Arrangement of P/S f l o t a t i o n c e l l , gas supply system and solution delivery system 142 8.1 Results of exploratory test tube experiments on f l o a t a -b i l i t y of adit 23 and No. 5 seam coal samples in aqueous methanol solutions 150 8.2 Effect of conditioning time on y i e l d , 3 mins f l o t a t i o n 154 8.3 ' Effect of f l o t a t i o n time on y i e l d , No. 5 seam, 15 mins conditioning, 10 vo l . % methanol 154 8.4 Cum. f l o a t a b i l i t y relationships for the No. 5 seam sample (-210 + 149 ym) 155 8.5 Cum. f l o a t a b i l i t y relationships for the 150' adit 23 sample (-210 + 149 ym), 15 mins conditioning, 3 mins f l o t a t i o n 156 8.6 Cum. f l o a t a b i l i t y relationships for the 111' adit 23 sample (-210 + 149 ym), 15 mins conditioning, 3 mins f l o t a t i o n 156 8.7 Cum. f l o a t a b i l i t y relationships for the 75' adit 23 sample (-210 + 149 ym), 15 mins conditioning, 3 mins f l o t a t i o n 157 8.8 Cum. y i e l d vs. f l o a t a b i l i t y for the 0' and 38' adit 23 samples (-210 + 149 ym), 3 mins f l o t a t i o n 157 8.9 Flotation washability curves for No. 5 seam sample (-210 + 149 ym) 160 8.10 Flotation washability curves for 150' adit 23 sample (-210 + 149 ym), 15 mins conditioning, 3 mins f l o t a t i o n 161 8.11 Flotation washability curves for 1111 adit 23 sample (-210 + 149 ym), 15 mins conditioning, 3 mins f l o t a t i o n 161 8.12 Flotation washability curves for the 75' adit 23 sample (-210 + 149 ym), 15 mins conditioning, 3 mins f l o t a t i o n 162 8.13 Flotation washability data from tests using different bubble deflectors in P/S c e l l , compared with curves obtained with no bubble deflector, No. 5 seam sample 166 xv i Figure ' Page 8.14 Calculated incremental y i e l d and ash content of flo a t s as a function of f l o a t a b i l i t y number, No. 5 seam sample 166 8.15 Calculated incremental y i e l d and ash content of floats as a function of f l o a t a b i l i t y number for 75', 111' and 150' adit 23 and No. 5 seam samples 173 8.16 Comparison of cum. f l o a t a b i l i t y curves for a d i t 23 samples 181 8.17 Comparison of f l o a t product f l o t a t i o n washability curves for the 75', 111' and 150' adit 23 samples and the No. 5 seam sample 183 8.18 Comparison of reject product f l o t a t i o n washability curves for the 75', 111' and 150' adit 23 samples and the No. 5 seam sample 184 8.19 Correlation of measured feed ash content of the 75', 1111 and 150' adit samples and the No. 5 seam sample with t h e i r f l o t a t i o n washability (W), ash rejection potential (R), and f l o a t a b i l i t y (F) at 25, 50 and 75% cum. y i e l d 187 8.20 Flotation ash rejection potential curves for the adit 23 and No. 5 seam samples 188 8.21 Correlation of bulk oxidation parameters with F f l o a t a -b i l i t y for the adit 23 and No. 5 seam samples 191 8.22 Comparison of incremental ash c o n t e n t - f l o a t a b i l i t y number lin e s for the adit 23 and No. 5 seam samples 193 8.23 Results of exploratory test tube f l o a t a b i l i t y experiments on some particulate samples of inherently hydrophobic solids in aqueous methanol solutions 200 8.24 F l o a t a b i l i t y of samples of several inherently hydrophobic solids and hydrophilic quartzite in aqueous methanol solutions in small-scale frothless P/S c e l l , 3 mins f l o t a t i o n 204 8.25 Effect of p a r t i c l e size on the f l o a t a b i l i t y of molybdenite 5 minutes conditioning, 3 minutes f l o t a t i o n 205 8.26 Adhesion tension diagrams i l l u s t r a t i n g two possible explanations for a plateau (discontinuity) in the cumula-t i v e f l o a t a b i l i t y curve for an inherently hydrophobic s o l i d in aqueous methanol solutions, assuming f l o a t a b i l i t y is controlled by bubble-particle aggregate s t a b i l i t y ; (a) a s p l i t w e t t a b i l i t y band; (b) a bimodal p a r t i c l e size d i s t r i b u t i o n 211 x v i i Figure 8.27 I l l u s t r a t i o n of the expected effect of pa r t i a l surface oxidation of a particulate sample of an inherently hydrophobic s o l i d , assuming f l o a t a b i l i t y i s controlled by bubble-particle aggregate s t a b i l i t y ; (a) adhesion tension diagram showing extended w e t t a b i l i t y region in which particles are non-floatable below Yu 0 ; (b) shape of resulting cum. % floated curve 2 8.28 Selective separation of PTFE from No. 5 seam coal in aqueous methanol solutions (1:1 mixture by wt., -210 + 149 ym) xv i i i LIST OF SYMBOLS Symbol a.. a c t i v i t y of component i a slope of li n e a r adhesion tension p l o t ; empirical constant b slope of li n e a r Zisman p l o t ; empirical constant c empirical constant d,d ,d . ,d',d" pa r t i c l e s i z e ; maximum; minimum1,' intermediate max mm ^ d ,d maximum par t i c l e size at r e l a t i v e velocity of impact in max,v max,g turbulent conditions, and under s t a t i c gravitational cond-tions d 1 ,d" intermediate par t i c l e sizes corresponding to c r i t i c a l bubble-c s c s p a r t i c l e aggregate s t a b i l i t y f s p e c i f i c excess i n t e r f a c i a l HelmhoTtz free energy g gravitational acceleration h f i l m thickness h^ thickness of residual hydrated layer 1 perimeter of Wilhelmy sensor plate m empirical exponent n " " ; number of carbon atoms in hydrocarbon chain n. number of molecules of component i n acceleration factor 9 p 3P Q; equilibrium and saturation vapour pressure Ap pressure difference across a curved interface ( c a p i l l a r y pressure) r surface roughness factor r1'r2 principal r a d i i of curvature of curved interface r ,r radius of bubble and vortex b v t , t ' induction time t , t ' contact time c c v r e l a t i v e velocity of impact xix Symbol v. tangential velocity i n vortex A area \'^H fractional area of low-energy and high-energy surface component D Debye thickness of the ionic atmosphere F Helmholtz free energy; f l o a t a b i l i t y ( v o l . % methanol at a given y i e l d ) H Hamaker constant K empirical constant Pf»P »P »P probability of f l o t a t i o n , c o l l i s i o n , adhesion, aggregate stab-t c a s i l i t y R gas constant; ash rejection potential S entropy; spreading c o e f f i c i e n t T absolute temperature U internal energy V vol urne W f l o t a t i o n washability (cum.% ash in floats for a given y i e l d ) d h G ^A'^A'^A'^A W 0 r' < °^ -dhesion; dispersion, hydrogen bond, ionic components Wf,Wf,Wf,Wp,Wr,Wp work of cohesion; dispersion, induction, dipole, hydro-gen bond, pi-bond components a c r i t c r i t i c a l t i l t angle at which l i q u i d drop moves on inclined plane Y , Y A , Y D , Y H , T 1 , T P , Y I R surface ( i n t e r f a c i a l ) tension; anisotropic dispersion, London dispersion, hydrogen bond, induction, dipole, pi-bond components Ya' Yb surface tension of li q u i d s A and B Y A B i n t e r f a c i a l tension between l i q u i d s A and B Y , Y , Y F ' Y R ( . c r i t i c a l surface tension of wetting, adhesion, f l o a t a b i l i t y , c ca C T cs aggregate s t a b i l i t y Y _ solution surface tension at c r i t i c a l micelle concentration cmc XX Symbol Y ^ j Q surface tension of water ^ Y - J v surface tension at liquid/vapour interface Y s ' Y s V Y s v i n t e r f a c i a l tension; solid/vacuum, s o l i d / l i q u i d , solid/vapour 6 edge angle e d i e l e c t r i c permeability t, zeta potential n absolute v i s c o s i t y 9'9A'9C'9M'9max'9R'9W'9Y contact angle; advancing, Cassie, macroscopic, maximum, receding, Wenzel, Young ( i n t r i n s i c ) 6^,»e o f c r i t i c a l contact angle of adhesion and f l o a t a b i l i t y ca C T 6 equilibrium contact angle corresponding to free energy minimum for tot a l system 9YH'9YL high-energy and low-energy components of Young contact angle p. chemical potential of component i IT , T T equilibrium and saturated surface pressure e o n p-j , p s r e l a t i v e density of l i q u i d and s o l i d r. s p e c i f i c excess i n t e r f a c i a l concentration of molecules of component i above the bulk solution concentration r l v ' r s l ' r s v surface excess concentration at liquid/vapour, s o l i d / l i q u i d , and solid/vapour interfaces n n,n F,rt v,ii<, d i s j o i n i n g pressure; e l e c t r i c a l double layer, van der Waals, and s t e r i c components $ interaction parameter xx i ACKNOWLEDGEMENTS The author expresses his sincere gratitude to his research supervisor, Dr. J. Leja, for his patient guidance, support and encouragement. Thanks are also extended to the st a f f members of the Department of Mining and Mineral Process Engineering for th e i r valuable assistance, to various post-doctoral fellows and graduate students for th e i r helpful discussion, and in particular to Dr. M. Bustin of the Department of Geological Sciences for providing and petrographically characterising coal samples. The stimulating comments of Dr. H. Kolny of Vancouver are also acknowledged. Support from grants from Fording Coal Ltd., the National Science and Engineering Research Council of Canada, and the following scholarships, are greatly appreciated: Kit s a u l t Community; Frederick Armand McDiarmid; Henry Dewitt Smith; Bethlehem Copper Corporation. -1 CHAPTER 1 INTRODUCTION Coal production i s steadily increasing in Canada and on a world scale, the total being some 30 m i l l i o n tonnes and 3500 m i l l i o n tonnes respectively in 1978 (1). In many countries over 50% of the total mine output is cleaned (washed) to some extent in coal preparation plants (over 60% in Canada) to reduce the sulphur and mineral matter (ash) content of the feed for coke making and thermal power generation; the majority of the coal which is cleaned is of bituminous class (2). Coal washing had been t r a d i t i o n a l l y confined to the coarse size f r a c t i o n s , but with the wide spread use of mechanized mining methods, and the emphasis placed on maximizing resource u t i l i z a t i o n and reducing the particulate matter content in wash plant effluents, the recovery of fine, coal (nominally <0.5 mm in size) has more recently become a very important aspect of coal preparation. It i s l i k e l y to be even more so in the future i f coal conversion processes are i n i t i a t e d on a large scale. Some of them w i l l require s t r i c t l i m i t s on the quantity of inorganic contaminants in the i r feed material for reasons of process efficiency and pollution control. Premiums w i l l also l i k e l y be paid for a cleaned product which is r i c h in s p e c i f i c microscopic organic components (macerals) found in the coal (3,4). I f coal preparation i s to meet these needs of greater ash rejection and preferential maceral concentration, greater l i b e r a t i o n of the coal components through comminution w i l l become a necessity (5). Froth f l o t a t i o n i s one of the main methods of fine coal cleaning and in several countries i t currently treats an average of up to 12% of the total preparation plant feed (2). Of the 400 to 500 preparation - 2 -plants in the USA, i t has been estimated that about 300 of them use f l o t a t i o n , including a l l plants treating metallurgical coals. The amount cleaned by f l o t a t i o n varies between about 4 and 30% of the total clean coal production at individual plants (347). Due to severe geological disturbance, the Western Canadian Rocky Mountain coals are very f r i a b l e resulting in the feed to preparation plants in the area having t y p i c a l l y 35% <0.5 mm and up to 60% for some seams. The Mclntyre Smoky River plant (Alberta) treats a l l the 0.5 mm x 0 material by f l o t a t i o n . The Kaiser (now B.C. Coal Corp.) Elkview plant and the Fording plant ( B r i t i s h Columbia) use water only cyclones instead of flotation to treat the 0.5 mm x 0.2 mm fraction because of low f l o t a t i o n y i e l d s , problems with f l o a t i n g oxidized coal, and i n i t i a l underdesign of the f l o t a t i o n sections. Flotation of the 0.2 mm x 0 fraction provides about 15% of the total clean coal product (6,189,328,341). The success of selective coal f l o t a t i o n , and similar processes under development l i k e o i l agglomeration, i s the result of the natural or inherent hydrophobicity of the organic matter and the hydrophilic nature of associated mineral matter. Many synthetic organic polymers and some other natural solids are also inherently hydrophobic. Their properties are such that i t i s thermodynamically favourable for a i r to displace water from t h e i r surfaces (or some organic l i q u i d s or aqueous surfactant solutions), the very basis of f l o t a t i o n . The degree to which t h i s displacement occurs i s commonly called w e t t a b i l i t y ; i t r e f l e c t s the surface or i n t e r f a c i a l energies of the s o l i d and wetting l i q u i d and has been related to the surface constitution of homogeneous solids under ideal conditions. - 3 -Coal i s not a homogeneous substance, the surface properties of i t s organic matter depending on the rank, petrographic composition and degree of weathering, as well as the quantity and d i s t r i b u t i o n of associated mineral matter. Although there have been attempts to determine the w e t t a b i l i t y properties of coal as a function of these parameters, there has been l i t t l e success in relat i n g them d i r e c t l y to f l o t a t i o n practice. As a result f l o t a t i o n plant design and operation mainly r e l i e s on empirical testing. The d i f f i c u l t y appears to l i e in the extreme heterogeneity of coal when in the particulate (powdered) state. Most studies have measured only the average properties of massive or particulate samples. But the physical and constitutional properties of fine particles can range from those of v i r t u a l l y pure inorganic mineral species to almost pure organic matter, which i t s e l f can have a considerable range of elemental and structural properties and degrees of surface oxidation. Each p a r t i c l e can therefore be considered to have i t s own unique w e t t a b i l i t y and f l o a t a b i l i t y properties. Consequently there i s a need for a methodology which w i l l allow the d i s t r i b u t i o n of these properties within a particulate coal sample to be characterized. Its formulation and v e r i f i c a t i o n are the basis for t h i s investigation. The f i r s t part consists of an analysis of the l i t e r a t u r e on the wetting of inherently hydrophobic solids by pure li q u i d s and aqueous surfactant solutions, and the theoretical and empirical relationships between w e t t a b i l i t y and f l o a t a b i l i t y . The second part involves the experimental testing of the concepts developed from t h i s study with particulate samples of coal and other inherently hydrophobic soli d s . - 4 -CHAPTER 2 WETTING IN RELATION TO FLOTATION 2.1 Basic Wetting Theory for Liquids on Solids When a l i q u i d phase (including solutions) either p a r t i a l l y or com-pletely wets (spreads on) a s o l i d surface i t must displace another f l u i d phase ( l i q u i d , vapour or gas). A s o l i d surface can be generally c l a s s i f i e d as l y o p h i l i c i f the l i q u i d completely wets i t , or lyophobic i f i t i s only p a r t i a l l y wetted (the terms hydrophilic and hydrophobic apply s p e c i f i c a l l y to water or d i l u t e aqueous solutions). Wetting theory has been comprehen-siv e l y covered in many publications (for example references 7-11) and w i l l only be s e l e c t i v e l y reviewed here. Young (12) q u a l i t a t i v e l y described the mechanical equilibrium of the pa r t i a l wetting configuration i l l u s t r a t e d for a s e s s i l e l i q u i d drop in Fig. 2.1(a) in terms of the angle subtended by the l i q u i d on the s o l i d surface and the surface forces in the three participating interfaces (the t h i r d phase being the vapour of the l i q u i d under equilibrium condi-tions). Assuming a non-deformable, homogeneous, isotropic smooth s o l i d surface the force equilibrium parallel to the s o l i d surface can be expressed mathematically as Y , C O S e v = Y ~ Y i ....2.1 'lv Y 'sv 'si where e Y i s the equilibrium (Young) contact angle and Y - | V > Y s v and Y S - | are the i n t e r f a c i a l tensions of the three interfaces (mNm--'- or dyne cm"-'-)*. For s i m p l i c i t y , the Young equation and following discussion are presented in terms of a s o l i d phase\(s), a l i q u i d (solution) phase ( 1)and a gaseous vapour phase (v). However the principles involved also apply i f either the vapour or the s o l i d i s replaced by a second immiscible l i q u i d , a l l three phases being in mutual equilibrium saturation. - 5 -mg Fig. 2.1 Three-phase contact angle equilibrium for a l i q u i d drop on a s o l i d surface; (a) ideal conditions; (b) hysteresis on an inclined surface under gravity, after Wolfram and Faust (34); (c) edge effect at sharp corners, after Oliver et al. (43) - 6 -The e x p l i c i t thermodynamic d e f i n i t i o n of i n t e r f a c i a l (surface) tension most commonly given i s Y = (|jr) 2.2 9 A T.V.n. p where F i s the Helmholtz free energy of the system , A i s the area of the interface in question, V i s the volume of the system and n^ is the number of molecules of component i in the total system (11). Using thermodynamics Dupre (13) defined work of adhesion, W^ , as the reversible work necessary to separate unit area of the interface between two pure immiscible liquids in equilibrium into unit area of liquid/vapour interface of both l i q u i d s . In terms of a one-component l i q u i d interfaced with a one-component isotropic s o l i d , i t can be expressed as WA = ^ l v + Y s v - Y s l '•••2-3 Dupre also described Young's relationship as WA = Y 1 V (1 + cos 6 y) ....2.4 provided that i s less than the work of cohesion, W^ , of the l i q u i d phase = 2Y-| y ....2.5 Combination of equations 2.3 and 2.4 results in equation 2.1. For the system to be in true thermodynamic equilibrium i t i s necessary that each component of the system should have the same chemical potential, in the three phases ( i . e . the l i q u i d and vapour must be saturated with the F = U-TS where U i s internal energy, S is entropy and T i s absolute temperature. - 7 -s o l i d , and vice versa, and the s o l i d / l i q u i d , solid/vapour and l i q u i d / vapour interfaces must be at adsorption equilibrium). Young's equation has been rigorously derived from f i r s t principles using Gibbs thermodynamics in a system where adsorption and gravity are e x p l i c i t l y considered, provided that the re l a t i o n i s in terms of int e r -f a c i a l tensions, y, and not s p e c i f i c excess i n t e r f a c i a l free energies, f (the Helmholtz free energy change associated with the isothermal -2 -2 reversible formation of unit area of fresh interface, mJm or erg cm ) (11,14). Using the Gibbs model for an interface f = Y + ? yi r.j 2.6 where r. i s the sp e c i f i c excess concentration of molecules (of component i ) in the interface above the bulk concentration. By d e f i n i t i o n the Gibbs interface i s situated such that the s p e c i f i c excess concentration of the bulk (solvent) component ( i = l ) , i s zero. Although y and f are dimensional l y equivalent, they are only numerically equal for one-component li q u i d s where there is no positive or negative adsorption at the interface (r.j =0). In pr i n c i p l e they are also equal for a one-component isotropic s o l i d with an equilibrium surface in contact with a one-component f l u i d or s i m i l a r bulk s o l i d phase. However, in practice, when fresh s o l i d surface i s formed (e.g. by cleavage), the molecules or ions in the surface w i l l normally be in a state of s t r a i n . They w i l l therefore be unable to take up t h e i r equilibrium configuration (or the relaxation time may be very slow) and tension in the surface w i l l not be the y as defined by equation 2.2 (8,11). Aveyard and Haydon (11) give the surface tension of a one-component isotropic s o l i d in vacuum as - 8 -Only at equilibrium ( i . e . no strain) w i l l A ( ^ ) = 0, and Y s = f-The thermodynamic v a l i d i t y of Young's equation under ideal e q u i l i -brium conditions i s now generally accepted (10), although objections have been raised i n the past about both i t s mechanical and thermodynamic basis, p a r t i c u l a r l y by Bikerman (15,16). The equilibrium requirement that the s o l i d and vapour phases be at equilibrium means that the s o l i d surface i s often covered with a f i l m of adsorbed vapour phase. The value of Y s v for the f i l m covered surface i s lower than that for the vacuum/solid interface, Y s > by an amount Tre, the surface pressure of the adsorbed f i l m , i.e. Y S V = Y S - T r e ....2.8 It follows that W^ , equation 2.4, i s the work required to separate unit area of s o l i d and l i q u i d , the f i n a l s o l i d surface having on i t the equilibrium adsorbed f i l m . Combining equations 2.1 and 2.8 gives Y 1 v cos e Y = Y S - Y s 1 - * e ....2.9 It i l l u s t r a t e s that an increase in Tre results in a larger contact angle. In those situations where complete wetting (spreading) of the s o l i d surface takes place no contact i s formed and the Young equation (2.1 and 2.9) does not apply. The necessary condition for spreading was f i r s t put forward by Cooper and Nuttall (17); they defined a spreading c o e f f i c i e n t , S, as 5 = Y s v - ^ s l + V •••• 2' 1 0 or from equations 2.3 and 2.5 S = WA - Wc ....2.11 - 9 -For spreading, S >0 ( i . e . cohesion within the l i q u i d i s less than the adhesion of the l i q u i d to the s o l i d surface), and for non-spreading, S <;0. Harkins (7) differentiated between the " i n i t i a l spreading c o e f f i -cient" which expresses the i n i t i a l conditions causing spreading (often non-equilibrium, and in terms of Y s instead of Y s v ) and a f i n a l spreading c o e f f i c i e n t which describes the equilibrium conditions after mutual saturation and/or adsorption have taken place. For non-spreading a f i n i t e contact angle i s formed, and Young's equation applies. It has often been emphasized that equations 2.1 and 2.10 are of great significance in terms of froth f l o t a t i o n (18). Young's equation describes the thermodynamic condition for f l o a t a b i l i t y of a s o l i d , whereas S >0 applies to the condition of non-f l o a t a b i l i t y or depression. As far as i s known, neither equation 2.1 nor 2.9 have been s a t i s -f a c t o r i l y experimentally v e r i f i e d for l i q u i d s contacting s o l i d s , y^ can be accurately determined and contact angles can be measured r e l a t i v e l y easily and reproducibly; however, the interpretation of the l a t t e r in terms of the physical, chemical and dynamic conditions of the wetting system often presents many problems. Controversy and measurement uncertainty also surround the r e l a t i v e magnitude of ^  , but the most intractable problem has been the direct measurement o f - Y S > Y S V O R Y s i > although considerable progress has been made i n th e i r indirect estimation for low-energy solids based on e and Y-|v data. These aspects w i l l be b r i e f l y expanded on in the following sections. Despite the uncertainties, the Young equation i s s t i l l the thermodynamic backbone of wetting theory in cases where the l i q u i d only - 10 -p a r t i a l l y wets the s o l i d surface, forming a f i n i t e contact angle. 2.1.1 Interpretation of contact angles The equilibrium contact angle, 6y> given by the Young equation (2.1) is a unique function of the i n t e r f a c i a l tensions Y-]v» y s v A N A" Y S T By accepting the thermodynamic v a l i d i t y of the equation, 8y can therefore be considered as a thermodynamic parameter (19). Its magnitude w i l l be affected by any influence which changes one or more of the three in t e r -f a c i a l tensions. Adsorption of surface-active additives i s the most common example (see sections 2.2.3.1 and 2.2.4 for further discussion) and i s the basis of many practical contact angle applications, not least of which i s froth f l o t a t i o n and analogous processes (20-24). The contact angle of li q u i d s on some polymers and paraffins have indicated a small degree of temperature s e n s i t i v i t y , p a r t i c u l a r l y when a l l o t r o p i c phase changes occur (19). Surface charge can also be an important parameter as indicated by the parabolic shape of contact angle - applied potential curves for the a i r / e l e c t r o l y t e solution/mercury system; e reaches a maximum corresponding to zero mercury surface charge (25). Similar 3 behavior has been found for e as a function of zeta potential of s i l v e r iodide, e reaching a maximum in the region of the i s o e l e c t r i c point (26). However, there are other situations (which w i l l be mentioned in chapter 3) where l i t t l e or no correlation with zeta potential has been found. The magnitude of measured contact angles can also depend to a certain extent on the measuring technique [of which there are many, recently Zeta potential i s the difference in potential between the plane of shear in the l i q u i d near the p a r t i c l e surface, and a point in the bulk solution at some distance from i t (8). - 11 -reviewed by Neumann and Good (27)] and on whether the system i s in a tru l y stable state. Dynamic (time dependent) contact angles can result when such factors as adsorption, d i f f u s i o n , charge d i s t r i b u t i o n , molecular orientation, swelling, hydrodynamics etc., have not reached equilibrium (10,23,28-33). In such cases the Young equation does not apply. It i s often found that even when dynamic effects are absent, two or more metastable contact angles can occur, their magnitude depending on whether the l i q u i d has been advancing or retreating across the s o l i d surface prior to the s t a t i c angle being formed. The largest and smallest metastable angles have been termed the advancing and receding contacting angles respectively, e A and e R, while the difference between the two i s generally called contact angle hysteresis. The c l a s s i c example i s a symetrical l i q u i d drop on a horizontal surface becoming asymetrical under the influence of gravity when the surface i s t i l t e d , Fig. 2.1(b). becomes larger and smaller as the t i l t angle increases reaching l i m i t i n g s t a t i c values at c c c r i- t where movement of the drop begins (34). There have been several recent reviews .on the causes and theoretical analysis of contact angle hysteresis (8-10,19,35-37). The two factors affecting hysteresis which have received the most attention are surface roughness and heterogeneity. Roughness describes the deviation of a surface from a smooth, planar configuration, whereas heterogeneity i s due to the presence of areas of dif f e r e n t surface energy; for instance, different chemical phases, impurities in or adsorbed on a surface, and variations in atomic or molecular packing densities and orientation. There i s no sharp separation between roughness and heterogeneity. On a microscopic scale approaching atomic or molecular dimensions high-energy - 12 -crystallographic defects such as dislocations, grain boundaries, steps, ledges etc., can be considered as contributing to both. Roughness i s a d i f f i c u l t parameter to quantify; Wenzel (38) defined a roughness factor, r, as the r a t i o of the actual surface area to the geometric plane area and used i t as a contact angle correction factor such that cos = r cos 0y 2.12 where is the predicted macroscopic angle and 6y i s the equilibrium, i n t r i n s i c (Young) angle for a smooth, homogeneous s o l i d surface. Sub-sequent theoretical analyses of various surface roughness models have determined the change in the total Helmholtz free energy of the system, A F, (which depends on the macroscopic configuration of the system under the influence of gravity) as a function of the macroscopic contact angle, 8^, assuming that the Young equation (2.1) holds on the microscopic scale, and the Laplace equation of c a p i l l a r i t y Ap = Y 1 v O / i ^ + l / r 2 ) ....2.13 holds on the macroscopic scale (where Ap is the pressure d i f f e r e n t i a l across a curved interface and r^ and are the principal r a d i i of curvature). The existence of a number of metastable energy states, with corresponding contact angles, are predicted, the largest and smallest of these being equated with the l i m i t i n g observable advancing and receding contact angles, and 9^  (39-42). The degree of hysteresis has been shown to depend on 6y, and the angle of i n c l i n a t i o n and shape of the asperites or ridges, as well as t h e i r orientation.in r e l a t i o n to the direction of motion of the three-phase-boundary. Their height affects the magnitude of the energy barriers between the metastable states. - 13 -Hysteresis decreases as the asperites become smaller and f l a t t e r , approaching zero for estimated heights of below about 0.1 pM (41). The model system analysed by Neuman (19) confirms a long held be l i e f that macroscopic or phenomenological advancing and receding contact angles measured on rough surfaces w i l l not be Young angles ( i . e . w i l l not s a t i s f y equation 2.1), and therefore cannot be u t i l i z e d in any equations for estimating s o l i d surface energies which have been derived on the basis of the a p p l i c a b i l i t y of Young's equation (e.g. equations 2.21 and 2.26, section 2.2.2). I f there i s an absolute minimum in the AF vs relationship for a particular system (which i s not necessarily the case), then the corresponding theoretical thermodynamic equilibrium contact angle, e , is given by the Wenzel contact angle, although neither of them are Young angles (41). It has long been considered that sharp edges and corners on small angular particles in f l o t a t i o n systems tend to i n h i b i t the spreading of the three-phase boundary produced by bubble attachment, and can have a marked effect on contact angle hysteresis (20-22). A recent investigation by Oliver et al (43) has indicated that spreading of a l i q u i d drop over the sharp edge of a c i r c u l a r d i s c , Fig. 2.1(c), w i l l only occur i f 6 » [(180 - 5) + e y] 2.14 where e Y i s the i n t r i n s i c contact angle for the l i q u i d on the s o l i d , and 6 i s the edge angle. This i s true even i f the l i q u i d completely wets the s o l i d ( i . e . e y = 0) (44). For a model f l a t , patchy (heterogeneous), two-component s o l i d surface, Cassie and Baxter (45,46) proposed that the equilibrium contact angle should be given by the area-weighted average - 14 -cos e c = A L cos e y L + A h cos 0 V H 2.15 where A^ and A^ are the fractional areas of the low-energy and high-energy 4 components respectively. It became commonly accepted that 0^ and 9^ on such a surface would be related or equal to the i n t r i n s i c contact angles of the two components, 0V|_ and 0 V H > respectively. Theoretical analyses of model two-component heterogeneous systems, si m i l a r to those for rough surfaces, have indicated the existence of a number of metastable states in some situations, the maximum and minimum 0^ corresponding to the i n t r i n s i c contact angles 0 V ^ and y^ of the low- and high-energy components (9,47). Therefore,for these model heterogeneous surfaces, 9^ max and 0^ min are Young angles because they s a t i s f y Young's equation l o c a l l y at the three-phase l i n e . But other metastable contact angles can exist between max and 0^ min which are not local Young angles. The theoretical thermodynamic equilibrium contact angle, 0 , corresponding to the absolute minimum in the AF V S 0^ rel a t i o n s h i p ^ i s predicted by 0^ from equation 2.15 but also i s not a Young angle (19). Of course, the dimensions, geometries and orientations of the heterogeneous patches have a bearing on whether metastable states, and therefore hysteresis, exist or not. For instance, the q u a l i t a t i v e d i v i s i o n of s o l i d surfaces into so-called high-energy and low-energy categories is rather a r b i t r a r y ; Fox and Zisman (76) considered that hard solids such as metals, metal oxides, s i l i c a t e s , etc., with s p e c i f i c surface free energies ranging from 500 to 5000 ergs cm-2^  have high-energy surfaces, while r e l a t i v e l y soft solids such as organic polymers and waxes, which generally have s p e c i f i c surface free energies of less than 100 ergs cm-2^have low-energy surfaces. Another common method of c l a s s i f i c a t i o n i s on the basis of whether or not water completely wets the surface ( i . e . hydrophilic or hydrophobic) (10,23). The two c l a s s i f i c a t i o n systems roughly correspond. - 15 -i f the boundaries between high- and low-energy s t r i p s are pa r a l l e l to the direction of motion of the three phase l i n e of the wetting l i q u i d , hysteresis becomes zero; this also appears to be the case when their dimensions become sub-micron in size (47). ' . In some circumstances i t has been observed that the magnitude of the measured contact angle decreased, and hysteresis increased, when the size of a captive bubble or s e s s i l e drop was reduced below a certain value (48,49). Leja and Poling (48) suggested that this effect could be due to the influence of gravity on bubble d i s t o r t i o n ; however, more recent investigations have led to the be l i e f that the three-phase-boundary l i n e i s distorted by surface heterogeneity or roughness such that,for small drops or bubbles, the area of the contorted 1iquid/vapour surface becomes s i g n i f i c a n t l y larger than the uncontorted surface, there-by changing the energy balance and macroscopic contact angle of the system (10,49). 2.1.2 Estimation of the surface and i n t e r f a c i a l tensions of solids Because of the d i f f i c u l t y experienced in obtaining r e l i a b l e experi-mental values of the surface or i n t e r f a c i a l tensions of low- or high-energy s o l i d s , considerable attention has been focussed on thei r indirect estimation from other molecular or bulk phase parameters, and in p a r t i -cular from contact angles via Young's equation (2.1 and 2.9). A quick perusal of these relationships makes i t clear that contact angle and Y - J V only allow the difference Y S V - Y S - J (or Y S - Y s - | i f ^ e i s known) to be calculated. Two similar semi-empirical theories were put forward in the late 50's and early 60's for the estimation of i n t e r f a c i a l tension between two - 16 -immiscible li q u i d s from thei r measured surface tensions. G i r i f a l c o and Good (50) i n i t i a l l y proposed that the free energy of adhesion between two immiscible l i q u i d s , A and B, was given by the geometric mean of the free energies of cohesion of the two liq u i d s "A • *<WC,a "C,b>°'5 •••• 2- 1 6 or from equation 2.5 WA = 2 * ( V Y b ) ° * 5 '••- 2- 1 7 where $, an interaction parameter, was a function of the molecular properties of the two phases. Empirically i t s values were found to range from about 0.5 to 1.15 for water in contact with a variety of organic liquids of various degrees of m i s c i b i l i t y . For l i q u i d s which formed "regular" interfaces i t was found that $ = 1. A system with a regular interface was defined as one in which the dominant cohesive and adhesive forces were of the same type, and the two liquids were mutually en t i r e l y immiscible. It was o r i g i n a l l y proposed that, for non-regular systems 3$ could be estimated from the molar volumes of the two liq u i d s (50), but since then more detailed and sounder theories have been put forward for computing $ from known properties of the two interacting phases (molecular p o l a r i z a b i l i t y , dipole moment, ionization potential) (10,51,52). $ has also been related to Hamaker constants. The i n t e r f a c i a l tension between phases A and B was given by Yab = Y a + Yb " 2 $ ( ^ a Y b } ° ' 5 ....2.18 (from equation 2.17 and = y a + y^ - Y a k . ) - ^ has been shown that when the s o l u b i l i t i e s of the two liq u i d s are appreciable, y a and Y h should be replaced by concentration weighted average values (51). For - 17 -a s o l i d / l i q u i d system equation 2.18 translates into Y S ] = Y s + Y L Y " 2 * ( y s y l v ) 0 - 5 ....2.19 Therefore from equations 2.3 and 2.8 WA + * e = 2 $ ( y s Y 1 v ) 0 " ' 5 ....2.20 which combined with equation 2.4 and rearranged gives cos e y = 2 $ ( Y S / Y 1 v ) ° ' 5 - 1 - * E / Y ] V ....2.21 Fowkes (53-56) put forward a modified version of the Good and G i r i f a l c o theory by proposing that the work of cohesion of a phase (and by implication i t s surface tension) could be broken up into several components on a simple a d d i t i v i t y p r i n c i p l e ; WC = WC + WC + WC + WC + WC + ....2.22 and - d i p h i T , 7 9 . , Y - Y + Y + Y + Y + Y where d,i,p,h and ir represent the London dispersion, induction, dipole, hydrogen bond, and pi-bond contributions. He also considered that inter-actions only take place between l i k e components; t h i s has since been shown to be only an approximation (8,52). Like Good and G i r i f a l c o , Fowkes also assumed that the interaction energy across the interface followed a geometric mean law such that f o r phases which interacted by dispersion forces only (using equation 2.5) a n d Y a b = Y a + Y b " 2 ( l a Y b ' ° ' 5 . . . . 2 . 2 5 - 18 -For a s o l i d / l i q u i d system this translates into cos e Y - 2(y d Y J V ) 0 V 5 / Y 1 v - 1 - v ^ i v •••• 2' 2 6 Equations 2.21 and 2.26 are obviously very s i m i l a r . For liquids where either $ or y^ v are known or can be calculated, then Y s or y^ can be estimated from e, y^ v and -n^ data; because of lack of r e l i a b l e data, i r e i s often considered to be negligible for liquids with y-|V > Y S - For many systems, the v a l i d i t y of this assumption i s open to question (see next section). Plots of cos e vs either ( l / y - | v ) 0 ' 5 or ( Y ^ V ) ° ' 5 / Y - | V for single component l i q u i d s on low-energy polymers have produced straight lines through the cos 6 = -1 o r i g i n ; these plots have been used to prove that •n-g i s n e g l i g i b l e , $ is a constant for a homologous series of liquids on a particular s o l i d surface, and the slope of the l i n e or the cos e = -1 extrapolated intercept are estimates of y (assuming $ = 1) or y^. But s s there i s also evidence that such straight l i n e relationships can be obtained through the cos e = -1 origin even i f s i g n i f i c a n t TT i s included, as long as i t varies inversely with y-|V in a regular manner (57). The analysis of Good and G i r i f a l c o probably has the sounder theore-t i c a l foundation ( p a r t i c u l a r l y after l a t e r modifications); however, Fowkes' concept seems to have been more widely u t i l i z e d because of i t s a t t r a c t i v e s i m p l i c i t y and impressive predictive success for many systems [some of which may only be fortuitous (8)]. Since his i n i t i a l proposal Fowkes has provided a theoretical basis for the calculation of from dispersion interactions and related i t to Hamaker constants (58,59). Theoretical relationships have also been developed between the $ of Good and G i r i f a l c o and the energy of interaction components of Fowkes, so that - 19 -the two theories have become v i r t u a l l y completely interlocked ( 5 2 , 6 0 , 6 1 ) . Equation 2 . 2 5 was used to predict the dispersion component of the surface tension of water, yjj Q (and other polar and metallic l i q u i d s ) , 2 ^ based on the assumption that, for saturated alkane liquids, y-| = Y-J V -Using the best available data for alkanes ( 6 2 ) i t was found that y°j Q was v i r t u a l l y constant for c y c l i c and branched-chain li q u i d s at 2 2 . 0 dyne/cm, but decreased from 2 2 . 8 to 1 9 . 7 dyne/cm as n increased from 6 to 16 for the n-alkanes. This inconsistency has been the basis for some of the c r i t i s m l e v e l l e d at Fowkes1 equation 2 . 2 5 ( 6 3 ) . It led him ( 64 ) to postulate that the surface free energy of alkanes consists of two independent terms, one involving isotropic dispersion forces, y^ y, and the other involving anisotropic dispersion forces y^ , such that Y l v = Y?v + Y?v . . . . 2 . 2 7 The anisotropic component i s supposed to result from the par a l l e l orientation of hydrocarbon molecules and i s a function of chain length. Fowkes argues that because water i s essent i a l l y isotropic i t cannot sense the polarized dispersion forces across the interface and therefore w i l l only interact with the isotropic dispersion forces of the alkane. Based on this reasoning, y£| Q was assumed to be constant and equal to 2 2 . 0 dyne/cm, and was used to calculate the y^ component of the n-alkanes, which increased from 0 to 2 .9 for n = 6 to 16 ( i . e . from about 0 to 10% of y-|V)- Good ( 5 2 ) had also considered that surface free energy should contain a molecular orientation component to take into account entropic factors. On the assumption that each surface free energy component of a substance i s a property which holds at any interface with another phase, - 20 -the Fowkes relationships (equations 2.25 and 2.26) have been extended to allow the estimation of surface and i n t e r f a c i a l tensions between two phases which interact by other surface energy components as well as dispersion (65-67). Extra energy of adhesion terms for each interacting component are included on the right hand side of equations 2.25 and 2.26, given by the geometric mean approach (equation 2.24) or by an harmonic mean approach, or by a combination of the two. Baskin et al (66), using polar and non-polar li q u i d s on polyethylene surfaces with various degrees of oxidation, found that the geometric mean approach gave the best correlation with experimental r e s u l t s . A more empirical approach to estimating s o l i d surface energies has been developed by Neuman et al (19,68). A linear correlation was found between the interaction parameter, $, and Y s-| of the form * = - ay s 1 + b 2.28 based on contact angle data for several fluorinated polymers. It was also assumed that the contact angles were Young angles (section 2.1.1), Y s v was constant for a l l Y-| v» and Y S - J = 0 and Y S V = Y - J v at 0y = 0. a and b were found to be constant for a l l the polymers, with best estimates of -0.0075 and 1.00 respectively. Combining equation 2.28 with the Young equation (2.1) and Good and G i r i f a l c o ' s equation (2.19) gives (2ay - 2b)(y Y - , ) ° " 5 + Yi c o s 9 Y = S V n i — • • • • 2 - 2 9 from which Y s v can be calculated i f 0y and y-|v are known. Y s ] can then be calculated via the Young equation. Equation 2.29 can only be used to estimate Y s v and Y S - J when the s o l i d is as smooth, homogeneous and isotropic as possible, such that T\Q is negligible and 0 = 0y. - 21 -Neumann (19) considers that equation 2.29 can be used to test whether contact angle hysteresis i s due to heterogeneity or roughness. If advancing or receding contact angles measured with several liquids on the same substrate y i e l d y values which are essentially constant and sv independent of y-|V» the contact angles must be Young angles characterizing the low- and high-energy components of the so l i d surface respectively. 2.1.3 Surface pressure There is considerable uncertainty in the l i t e r a t u r e as to the magnitude of the Tre term p a r t i c u l a r l y for low-energy solids being wetted by one-component liq u i d s (8,63,69). The main reason for the controversy appears to be lack of r e l i a b l e experimental data. For the l i q u i d / s o l i d / vapour system T\Q is usually obtained from the expression TT = RT r ^  r dlnp 2.30 e J 0 sv where p is the equilibrium vapour or gas pressure, R i s the gas constant (11). r s v is the Gibbs surface excess concentration at the solid/vapour interface as defined in equation 2.38, section 2.2.3.1. Gibbs' equation is only v a l i d for physical adsorption whereas in many cases involves chemisorption. r s v i s most often determined gravimetrically or volumetrically from powdered solids with large surface areas, whereas contact angle data are measured on smooth macroscopic surfaces. There is no assurance that the surface properties of the s o l i d remain the same when i t i s in the powdered form. Also i n t e r p a r t i c l e c a p i l l a r y condensa-tion can d i s t o r t adsorption isotherms at pressures above about 0.9pQ, where p Q i s saturation pressure (8). In the u t i l i z a t i o n of the Girifalco-Good-Fowkes-Young method of estimating the dispersion and non-dispersion components of low-energy - 22 -s o l i d and l i q u i d surface tensions i t i s often assumed that i r e (or IT , the saturation surface pressure) i s negligible p a r t i c u l a r l y for li q u i d s that form large contact angles on these surfaces (50,53,59,70). Several authors are of the opinion that in many cases this i s not a good assumption (63,71,72). P h i l l i p s and Riddiford opposed i t on rather philosophical grounds and considered that i r e could be assumed to be negligible only for systems in which y s v <Y s]> equivalent to e >90° via Young's equation (2.1), and therefore limited to cases of highly assoc-iated l i q u i d s on the most lyophobic or low-energy s o l i d surfaces. They also pointed out that i f TT - 0, Zisman's empirically determined r e l a t i o n -ship (equation 2.33, next section) becomes incompatible with equations 2.21 and 2.26 of G i r i f a l c o and Good, and Fowkes. Adamson and co-workers (57,72,73) have used ellipsometry to measure the thickness of the adsorbed vapour films at saturation on the same smooth surfaces of pyrolytic carbon, polyethylene and PTFE which were used for contact angle measurement. With a potential-distortion model they calculated -nQ values _2 of approximately 120, 14 and 9 erg cm respectively for water vapour on these three solids (on the assumption that the adsorbed vapour was uniformly distributed over the surface) with corresponding advancing contact angles of about 70°, 88° and 98°, which in a l l cases is an appreciable fraction of y-|v- Data for other li q u i d s and low-energy solids show that ^g/Y-jy i s mostly between 0.1 and 0.3 which means corrections to e calculated by equation 2.26 of about 25° to 45° for low e, and 5° to 10° for values around 90°. They concluded that IT i s not always negligible in high contact angle systems and can vary considerably among systems having the same contact angle. - 23 -Fowkes et al (69), while not disputing the measured -nQ values for water vapour for the above-mentioned systems, believe that water vapour only adsorbs as multilayers on hydrophilic sites known to exist to some degree on almost every low-energy surface unless subjected to the most rigorous p u r i f i c a t i o n (49,71,74). Hu and Adamson (57) also realized that the i r results did not preclude this p o s s i b i l i t y . The hydrophilic sites tend to reduce the contact angle below that found on high purity samples (69,74), but not because of an increase in ir , which acts to increase contact angle (equation 2.9) (49). . An interesting alternative method for d i r e c t l y measuring u of liqu i d s which only interact through dispersion forces on the hydrophobic portion of low-energy solids has been recently devised (69). The advancing contact angles of one l i q u i d are used to measure -nQ generated by the vapour of a second l i q u i d . With water and methylene iodide, CH^^' (which are immiscible and whose vapours do not affect the surface tension of the second l i q u i d ) on polyethylene and PTFE, they found that the advancing contact angle formed by a drop of one l i q u i d was unaffected by the nature of the vapour phase. This led to the conclusion that high-energy liquids (water and CH^^) do not adsorb on low-energy substrates, and was used as proof of the v a l i d i t y of the following equation = 2 ( Y D Y ? ) 0 , 5 - 2 Y l ....2.31 e w s T l v '1 v derived from a combination of equations 2.8, 2.10 and 2.24 for W^» W^  ( f i n i t e contact angle) for the special case that the spreading l i q u i d has only isotropic London dispersion force interactions with the substrate (54,55,69). I f y l v * ( Y D Y D V ) ° ' 5 > then v = 0; for C H 2 I 2 , Y D V = Y 1 v = 50.8 dyne/cm. When methylene iodide was replaced by cyclohexane (also for - 24 -which Y ^ " V = Y - j v - 25.5 dyne/cm), the contact angle for water on polyethy-lene increased in the presence of CgH^ vapour, whereas i t was unaffected on PTFE. The authors concluded that cyclohexane adsorbs on polyethylene but not on P T F E because Y ^ ( P T F E ) < Y 1 W ( C - H . , 0 ) <Y^(polyethylene). However, s iv old s there i s one disturbing aspect about the results and conclusions. Additional evidence i s presented which shows that the PTFE surface contains an appreciable number of hydrophilic sites and that substantial water vapour adsorption occurs on powdered PTFE at these s i t e s . Therefore the following assumptions seem to have been made; (a) the surface pressure can be broken up into two components, one for the hydrophilic areas and one for the low-energy areas; (b) when contact angles are formed with a non-polar l i q u i d (e.g. CH^^) they do not depend on the hydrophilic surface pressure component ( i . e . for non-polar liquids the three phase contact l i n e occurs only on the non-polar portion of the s o l i d surface); (c) when contact angles are formed with polar li q u i d s which interact via both d i s -persion and non-dispersion forces, they are affected by both components of the surface pressure. There have been some recent calculations which have shown that, provided a low-energy s o l i d surface i s smooth and homogeneous, n should be negligible in the case of most liq u i d s for which e>0, except for low-boi l i n g point l i q u i d s . The predicted ^ 's for n-alkanes on PTFE were found to increase with decreasing chain length (52). However, because almost no real surfaces are smooth or homogeneous down to the molecular l e v e l , Tfe can be expected to be appreciable in proportion to the degree of roughness and heterogeneity (8,10). This i s in agreement with the appreciable T T O ' S found for n-alkanes with n <10 on PTFE powders (71). - 25 -A graphical method has been suggested to test whether -rre > 0 (68). This i s to plot y-|V cos e vs y-|V and f i t a quadratic equation of the form Y 1 v cos e - aY-|V2 + bY-,v + c 2.32 to the data. I f points close to the curve's intersection with the 6 = 0 ° boundary f a l l below the quadratic l i n e , i t i s considered to be an indica-tion that iTe i s appreciable. 2.2 C r i t i c a l Surface Tension of Wetting In section 2.1.2 semiempirical relationships were described which have been used to provide estimates of the surface tension of low-energy so l i d interfaces by u t i l i z i n g contact angle and non-polar l i q u i d surface tension data. A purely empirical method for characterizing the surface energy or lyophobicity of a low-energy s o l i d was developed during the 1950's by Zisman and his colleagues [comprehensively reviewed by Zisman (75)].. They found that by reducing Y-|V of a l i q u i d or solution, the contact angle, e, formed by a s e s s i l e drop on a f l a t , homogeneous, low-energy s o l i d surface, also became smaller. To characterize the r e l a t i v e w e t t a b i l i t y of such surfaces they defined the c r i t i c a l surface tension of wetting, Y , as that value of Y-| v below which liquids spread on a particular surface. 2.2.1 Pure liq u i d s on low-energy s o l i d surfaces For a series of homologous liq u i d s [n-alkanes, n-alkylbenzenes, di(n-alkyl) ethers, lin e a r polymethylsiloxanes] on specially prepared smooth, homogeneous, low-energy s o l i d surfaces [ polytetrafluoroethylene (PTFE), tetrafluoroethylene polymers, hexatriacontane, paraffin, - 26 -hydrocarbon and fluorocarbon acid and amine monolayer coatings on high energy platinum substrates] in almost a l l cases a linear relationship was found between the cosine of the equilibrium advancing contact angle and the l i q u i d surface tension, l i n e 1, Fig. 2.2 (76-80). An estimate of Y c i s obtained by extrapolating the l i n e to the cos e= 1 axis. Linear relationships have also been found for a series of miscellaneous liquids on some solids e.g. polyethylene (78). I t can be described by the equation cos 0 = 1 + b(y G - Y 1 v ) ....2.33 The values of both b and Y c have been found to be dependent to some extent on the nature of the liquids used. In some situations deviations from l i n e a r i t y are apparent, especially when high surface tension polar liq u i d s such as H^ O and glycerol produce large contact angles approaching 90° or more (77). In the linear region of the plo t , p a r t i c u l a r l y for n-alkanes on low energy surface such as PTFE, the interaction between the s o l i d and the l i q u i d is presumed to arise primarily from London dispersion forces only. In the higher surface tension region, the more polar li q u i d s are considered to interact with the s o l i d through addi-tional forces such as hydrogen bonding (81). Even though Y c i s an empirical parameter, i t has been correlated with the chemical structure of a s o l i d surface, p a r t i c u l a r l y polymers. The studies of low-energy, condensed organic monolayers on high energy substrates such as platinum showed that the w e t t a b i l i t y ( y c and 0) was always a r e f l e c t i o n of the outermost atoms of the monolayer without regard to the composition of the substrate (79,80). This led Shafrin and Zisman (82) to put forward the i r "constitutive law of w e t t a b i l i t y " - 27 -COMPLETE WETTING in o o PARTIAL WETTING y y c cmc SURFACE TENSION,/ dyne/cm Iv 2.2 Zisman w e t t a b i l i t y diagram for a low-energy s o l i d ; 1 -pure l i q u i d s ; 2,3 - different types of behavior found for aqueous solutions of micelle-forming surfactants - 28 -which states that "in general, the w e t t a b i l i t y of organic surfaces is determined by the nature and packing of the surface atoms or exposed groups of atoms of the s o l i d and i s otherwise independent of the nature and arrangements of the underlying atoms and molecules". Fig. 2.3 i s a we t t a b i l i t y spectrum they prepared relat i n g Y c to the type of chemical group existing in the surface layer. For PTFE, the representative y c has been taken as 18.5 dyne/cm and i s characteristic of a surface con-s i s t i n g of -CFg- groups. Y c values for some common polymers are given in Table 2.1; for a comprehensive l i s t refer to Zisman (75). There has been some discussion in the l i t e r a t u r e about the physical significance of Y c and i t s relationship to the s o l i d surface tension (81,83,84). Good (52) combined equations 2.21 and 2.33, and, by assuming iTe = 0 at cos e = 1 and Y-J V = Y c » arrived at the relationship r t = * 2 Y s ....2.34 A similar analysis for the same conditions (and only for non-polar li q u i d s where Y ^ V = Y - ] V ) > using Fowkes' equation 2.26, yields (70) • • • • 2»3 5 Rhee (85,86) combined the Young equation (2.1) with Zisman's equation (2.33), and obtained a parabolic relationship between Y s 1 and cos e. He proposed that when 9 Y s - | / 9 ( C O S e) = 0, Y S I = 0; assuming i r e = 0, he obtained Y s as a function of Y c and the contact angle at which Y S I = 0. Other methods based on similar assumptions have also been put forward (87). However, these approaches have been c r i t i c i z e d , mainly for the i r assump-tions that TT = 0 and p a r t i c u l a r l y that Y S ] = 0 as e approaches zero yc •= Critical surface tension, dynes/cm at 2Q*C 10 20 30 40 50 " I 1 1 1 1 1 1 1 1 °a i r n i i i - c F H - c H , H — C f > —CF 2H | — C F , A CF2 — C H , — CF,.— CF, — C F , — C F H — Fluonnation i n T X T C H , C H , — C H , | TTT-CH— — C H , — C H " Hydrogenation n — r c, ci., H —CCIH CH, A =CCI, — CCI ,—CH— Chlorination C, 0,1 H CHOH—CHj-— CO,—CH,—, — C H " r . uo ) *u i u Li y — 0'\ I —C(NO,),| 4~C0NHrCH-H | — CH,ONO, _ C H j N H N 0 2 ' — (101) U Crystal ~~ | Monolayer | Polymer Fig. 2.3 Relationship between Y c and surface chemical constitution of low-energy so l i d s , after Shafrin and Zisman (82) - 30 -Table 2.1 Selected water contact angle, e, and c r i t i c a l surface tension of wetting, Y , data for various low-energy solids (20°-25°C) s o l i d JH 20 pure liquids aqueous n-alc. solutions refs. polytetrafluoro- .112: b ethylene 108 " 123 " 18.5 19 18.5 21 C2 C2.C4 c r c 4 61 97 99 paraffin 110 " 108 " 110 " 110 f 22 21 19 21 23 24 23 c 2 c 1 c 2 L3 C1-C4 61 67 99 105 polyethylene 100 95 94 b ii 31 22 22.5 27.5 C 2 »C4 91 61 97 polypropylene 90 108 28 C1~ C4 99 112 polystyrene 82,87 e 84 b 33^ 27 27.0 21.0 27.0 26.9 Cl-C 3 c 2 Cl c 2 C 4 106 61 89 polycarbonate polyvinyl chloride polymethyl-methacrylate 85 " 83 " 74 " II 76^80 " 39 39 26 26.0 26.5 23.0 26.5 26.0 C-i-C l _ u 4 c 2 ^2 99 61 89 11 11 113 polyethylene-terephthalate 71 43 27.0 61 nylon 6,6 65 46 28.0 continued - 31 -Table 2.1 (cont.) sol id "H20 pure 3 aqueous n-alc, liquids solutions refs, eellulose sulphur - ortho. - mono. - amor. ta l c molybdenite graphite s t i b n i t e 30 87 96 81 50-70 88 86 30 f b 70,81 ,89 d 80 c 85 f 61 ,69,80 d 29-39 b 33-82 84 35-77 f d 30 015) 30.5 (115) -31.5 (115)" 35 (117)' 29 45.7 (120) 28 26 114 115 ii 116 105 117 118 119 105 118 121 122 105 118 a - unless otherwise indicated, ^ c values for pure liq u i d s are from Zisman (75) b - advancing s e s s i l e drop; c - equilibrium ses s i l e drop; d - equilibrium captive bubble; e - receding captive bubble; f - advancing t i l t i n g plate C] - methanol; C 2 - ethanol; Q3 - n-propanol; - n-butanol - 32 -(8,63,88,89), as well as the fact that y- depends to some extent on the type of li q u i d s used in i t s estimation (11,61). It has been suggested by Good (52) that i f the Y c obtained by Zisman's method i s to be used as an estimate of the surface energy of a s o l i d , i t should be limited to the data from one-component liquids having the same type of cohesive forces, where y-|v< 2y c (and preferably < 1.25y c), and only when linea r extrapolation i s possible. He recommends that a more consistent estimate of Y c i s obtained by plotting cos 6 vs. 1 / ( Y - | V ) ^ ' ^ However, irrespective of whether or not Y c i s related to Y s or Y D > i t i s s t i l l a useful empirical parameter which can be used to characterize the r e l a t i v e w e t t a b i l i t y of s o l i d surfaces, because, as Zisman (75) has stated, i t s r e l a t i v e values "act as one would expect of Y _ " . One application of y of interest in connection to the study of coal w e t t a b i l i t y has been the assessment of the effect of controlled oxidation on the surface properties of polyethylene. The we t t a b i l i t y of the surface, as characterized by the Y c obtained with methylene iodide/ decalin mixtures, was found to increase regularly as the degree of oxida-tion in sulphuric acid-potassium chlorate mixtures increased (charac-terized by the adsorption of radioactive 45 ^ a++ ions on the polar sites) (66,90). Similar behaviour of Y c for polyethylene exposured to oxidizing flames was observed with ethanol/water mixtures (91). 2.2.2 Pure l i q u i d s on high-energy solids The r e l a t i v e w e t t a b i l i t y of high-energy s o l i d surfaces has also been studied by the Y c technique. Zisman and co-workers (92) found that although the surface tensions of a l l the organic liquids were well below the presumed surface tensions of the high-energy surfaces, some spread - 33 -as predicted, but others didn't. They concluded that a l i q u i d w i l l spread freely as long as i t s adsorbed f i l m on the s o l i d has a Y c which is greater than the l i q u i d surface tension. However, for many polar liquids the adsorbed oriented f i l m presents a surface which has a Y C less than Y-| V and a f i n i t e contact angle is formed. These liquids were termed "autophobic". There is evidence in the l i t e r a t u r e that the lin e a r relationship between cos e and Y l v (equation 2.33) also holds for high-energy surfaces such as glass, A1,,03, graphite and TaC wetted by l i q u i d metals and alloys (85,93-96), and concentrated s a l t solutions (93). 2.2.3 Aqueous surfactant solutions on low-energy solids The w e t t a b i l i t y of low-energy solids has also been studied using aqueous solutions of surfactants (wetting agents) instead of pure l i q u i d s . Bernett and Zisman (97) studied a large number of different types of hydrocarbon surfactants on polyethylene and PTFE including non-ionic, anionic and c a t i o n i c , micelle-forming and non-micelle forming. In a l l cases e decreased as the Y-| v °f the solution decreased, but over the whole concentration range the cos e vs. Y-|V relationship was non-linear. For micelle-forming surfactants (either ionic or non-ionic) the curves could be broken up into roughly lin e a r portions, the break occurring in the v i c i n i t y of the c r i t i c a l micelle concentration (cmc) i l l u s t r a t e d by curve 2, Fig. 2.2. The Y values obtained by extrapolation varied quite s i g n i f i c a n t l y by up to 5 dyne/cm for the different surfactants on the same s o l i d , although they did tend to cluster about the Y c values determined with pure l i q u i d s . This led to the conclusion that Y C - 34 -depended only on the s o l i d surface properties andy] v> and not to any great extent on the type of surfactant (although the position and shape of the wetting lines could be quite d i f f e r e n t ) . None of the solutions of hydrocarbon surfactants was found to be capable of completely wetting the PTFE surface because i t s Y c was lower than the minimum Y-| V of between 26 and 27 dyne/cm which could be obtained with single conventional wetting agents. Results with micelle-forming highly fluorinated a l i p h a t i c acids and sal t s (with Y - | V at cmc, Y C M C > less than 20 dyne/cm) on the same s o l i d surfaces also gave wetting lines broken into two linea r portions, but in this case the break was found to have no relation to thei r cmc values (98). The Y c for PTFE remained v i r t u a l l y unchanged at about 19 dyne/cm, but that for polyethylene was reduced to about 20 dyne/cm. This suggested to Bernett and Zisman that the fluorinated compounds adsorbed at the polyethylene/solution interface in such a way to change i t s surface properties into those resembling a fluorinated surface (which they v e r i f i e d experimentally). The results of l a t e r investigations with similar systems tended to confirm the behaviour observed by Bernett and Zisman, although the shape of the wetting lines differed somewhat, being roughly of the form of curve 3 in Fig. 2.2. In general y-|V became constant and a minimum at the cmc ( Y C M C ) ; i f Y C of the s o l i d surface (determined with pure liquids) was less than Y c m c s then complete wetting did not occur in the surfactant solutions (e>0) even at concentrations above the cmc (99-103). For Y-, > y the wetting l i n e was usually curved rather than li n e a r . - 35 -I f , on the other hand, y c m c was s i g n i f i c a n t l y less than Y c (deter-mined with pure l i q u i d s ) , then the s o l i d surface was not wetted u n t i l Y N = Y » as demonstrated for Na-dodecyl benzene sulphonate and l v cmc Na-dodecyl sulphate on nylon and a polyester (polyethylene terephthalate) (104). This behavior supports Zisman's conclusion that for more polar s o l i d surfaces the long-chain surfactant adsorbs at the s o l i d / l i q u i d interface with i t s hydrocarbon chain oriented towards the solution making i t more lyophobic (in a si m i l a r fashion to autophobic li q u i d s on high-energy s o l i d s , section 2.2.2). Also, i t i s obvious that Y C determined with surfactant solutions w i l l apply only to that one particular surfactant. The wetting behavior observed with short-chain, non-ionic, non-micelle forming surfactants i s generally less complicated than that shown by long-chain, micelle-forming surfactants, p a r t i c u l a r l y for aqueous n-alcohol solutions. In most cases i t can be represented by continuous curves (61,89,91,105,106); as an example, Fig. 2.4(a) is a Zisman plot of the data of Good (67) for paraffin wax. The pure n-alcohols give the typical straight l i n e behavior of pure liquids (with the exception of methanol). Extrapolated values of Y c determined from aqueous n-alcohol solution data in the 1iterature are given in Table 2.1 for some polymers, paraffins and other inherently hydrophobic s o l i d s , as well as contact angles measured in water. A Y c of about 27 dyne/cm for human skin determined with acetone/water and ethanol/ water solutions has also been reported (107,108). As i t has been pointed out by Dann (61), with the exception of the most non-polar solids PTFE and p a r a f f i n , the Y determined with binary liquids is always less than I JO i 0.8 0.6 0.4 -0.2 -0 -0.2 --0.4 --0.6 -Parraf in x pure n-alcohols • methanol in water o ethanol in water • propanol in water 0 10 20 30 40 50 60 70 80 90 Surface tension , v dynes/cm o 40 30 ID a> >» 20 CD in o o 10 x 5 c o £ -IO a> .2 -20 in Qi TJ < -30 -40 1 0 Parrafin • x pure n-alcohols • methanol in water • propanol " " o ethanol •• •• 4 butanol « « O / y  / V Y-^-~- _ \ 2!1 \ 6> \ • \ ° \ \ \ \ (b) \ \ A \ a N \ A > \ » \ \ \ \ J _ A L 10 20 30 40 50 60 70 80 Surface tension,/^ dynes/cm 2 .4 Wettability of paraffin by pure n-alcohols and th e i r aqueous solutions; (a) Zisman diagram; (b) adhesion tension diagram; data from Good (67) and Fowkes and Harkins. (105) - 37 -Y c determined with pure l i q u i d s , and appears to have a l i m i t i n g maximum of around 30 dyne/cm. Another way of plotting 0 and y - | V data i s the adhesion tension diagram, y ^ v cos 0 vs. Y 1 v , which appears to have been f i r s t u t i l i z e d for t h is purpose by Lucassen-Reynders (109). Adhesion tension was f i r s t defined by Freundlich (110) as the difference between the surface tension of a s o l i d and i t s i n t e r f a c i a l tension against a l i q u i d , y ^ - y -j or Y s v ~ Y s l ^ o r Y l v C 0 S 9 v i a Y o u n 9 ' s equation). B a r t e l l and Osterhof (111) considered that i t was a measure of the attraction or "wetting power" of the l i q u i d for the s o l i d . Lucassen-Reynders found that his data for aqueous surfactant solutions on low-energy solids gave straight lines when plotted on an adhesion tension diagram. The same behavior has since been found for a variety of l i t e r a t u r e data for surfactant/low-energy s o l i d systems (123), p a r t i c u l a r l y for aqueous n-alcohol solutions (99,114). Fig. 2.5(a) i s an example of such a diagram for various commercial wetting agents^ on paraffin wax, based on the data of Fowkes (124). Lines of constant contact angle radiate from the o r i g i n , between 0 = 0 ° and 180°. For micelle-forming surfactants the wetting l i n e becomes ve r t i c a l as y ^ y approaches y c m c (as long as y c determined with pure liq u i d s i s less than Y ) (99,103). cmc ' The data in Fig. 2.4(a) are replotted in terms of adhesion tension in Fig. 2.4(b) along with the data from Fowkes and Harkins (105) for 5 Aerosol 0T, Na di-2-ethylhexyl sulphosuccinate; Aerosol MA, Na-dihexyl sulphosuccinate; Nonic 218, condensate of ethylene oxide with dodecyl-mercaptan, the polyether chain containing about 9 units; Tergitol 4, Na sa l t of a highly branched alkanol sulphate; Triton X-100, condensate of ethylene oxide with octylphenol, the polyether chain containing about 9 units (124). T-7 P a r o f f i n : • aerosol OT x nonic 218 o aerosol MA • tergitol 4 A triton X-IOO 10 2 0 30 4 0 50 6 0 70 Solution surface tension,/ dynes/cm 80 4 0 I 3 0 f 20 CD O 10 o c o "S -io . 2 -20 </> a> . c T3 < -30 -40 |-pure ethanol / / / / / / pure 1,4-dioxane, pure butanol-. / A Polyethylene > P T F E • 1,4-dioxane a x butanol-1 & o ethanol I /^e^90^ \°o \ \ \ \ (b) \ \ \ \ _L ± \ \ 10 2 0 3 0 4 0 50 6 0 70 Surface tension,/LV dynes/cm 8 0 Fig. 2.5 Adhesion tension diagrams for low-energy solids i n aqueous surfactant solutions; data from (a) Fowkes (124); (b) Bernett and Zisman (97) - 39 -n-butanol. They a l l f a l l within a f a i r l y narrow linear band. Bernett and Zisman's (97) data for PTFE and polyethylene are shown in Fig. 2.5(b). The large displacement between the curves for the 1,4-dioxane solutions and the alcohol solutions indicates that the adsorption of 1,4-dioxane at the s o l i d surface has a di f f e r e n t effect than that of the alcohols. Fig. 2.6 is based on the data of Fowkes and Harkins (105) ; approximately linea r behavior i s observed for n-butanol on the three naturally hydro-phobic minerals, whereas n-butylamine shows d i s t i n c t l y non-linear behavior, again indicating that different surfactants can have d i s t i n c t l y d i fferent adsorption properties at the s o l i d surface. The lines in Fig. 2.7 are based on data from Wolfram (99) and van Voorst Vader (114) for aqueous n-alcohol solutions on several low-energy solids of various degrees of polari t y (PTFE and paraffin being assumed to be completely non-polar, interacting by London dispersion forces only). The general l i n e a r behavior can be described by the following equation Y L V cos 6 = a y l v + ( l - a ) y c 2.36 where a i s the slope of the straight l i n e , (which can be either positive or negative), and Y c is the extrapolated intercept at the 9 = 0 ° boundary. Using t h e i r own and l i t e r a t u r e data (28,97,99,105,124,125) Bargeman and van Voorst Vader (123) found that a = -1.0 for a variety of surfactants on non-polar surfaces of glyceryl t r i s t e a r a t e , PTFE, paraffin and bee's wax (as indicated by the plots in Figs. 2.4(b), 2.5 and 2.7) and that ( l - a ) Y C was a constant for a given s o l i d which was independent of the surfactant type. By combining equation 2.36 with Fowkes' equation 2.26 (and assuming = 0) they concluded that (1 - a ) Y C = 2(Y^ Y ^ v ) ° ' 5 for - 40 -Fig. 2.6 Adhesion tension diagrams for t a l c , graphite ,and st i b n i t e in aqueous surfactant solutions; • n-butanol; • n-butyl-amine, data from Fowkes and Harkins (105) E u 0) c >> •o 8 0 6 0 4 0 CD 2 0 in O o ^ 0 c o in c 0) •20 o - 4 0 to 0) T3 O -60 - 8 0 p 1 1 y»1 / l 7e = 9 0 ° \<s> |C 1 1 |A 1 1 20 4 0 6 0 8 0 100 surface tension , y dyne/cm Iv Fig. 2.7 Adhesion tension diagram for A - PTFE, B-paraffin, C - polypropylene, D -poly-carbonate in aqueous solutions of meth-anol, ethanol, n-propanol and n-butanol, from Wolfram (99); E - cellulose in ethanol solutions, from van Voorst Vader (114) 2.8 Wettability lines for hypothetical solids A(non-polar), and B and D (increasing polarity) in aqueous solutions of a short-chain n-alcohol i l l u s t r a t i n g selective wetting re-gion (shaded) between A and B - 42 -non-polar s o l i d s . However, they were probably too selective in the data they used; the data in Figs. 2.4(b) and 2.5(b) for PTFE and paraffin indicate that Y c does depend on the type of surfactant. There is also some evidence that slopes more negative than -1 are possible (99,103). The preceding discussion indicates that,for non-polar solids in aqueous surfactant solutions, the parameter which r e f l e c t s any difference in w e t t a b i l i t y properties i s y • However, as a s o l i d becomes more polar (indicated by a larger y c determined with pure.1iquids, Table 2.1), Y c obtained with aqueous solutions of surfactants such as short-chain n-alcohols appears to approach a l i m i t i n g value of about 30 dyne/cm. The slope, a, then becomes the parameter which characterizes the s o l i d w e t t a b i l i t y , becoming less negative, and even positive, as the polarity of the s o l i d increases. If the Y c values of two s o l i d s , A and B, are s i g n i f i c a n t l y different (when determined with solutions of the same alcohol) then a selective A B wetting region w i l l exist i f Y c < Y i v < ' Y c ' indicated by the shaded area in Fig. 2.8. Solid B w i l l be completely wetted, but the solution w i l l s t i l l form a f i n i t e contact angle on s o l i d A. However between solids B and D B D the selective wetting region w i l l be negligible because Y C ~ Y C -2.2.3.1 adsorption and w e t t a b i l i t y The w e t t a b i l i t y of a low-energy s o l i d surface by aqueous surfactant solutions depends on the r e l a t i v e adsorption of the surfactant species at the three participating interfaces. Some investigators consider that a long-chain wetting agent s e l e c t i v e l y adsorbs at the low-energy s o l i d / l i q u i d interface with i t s hydrophilic or polar group oriented towards the solution; the aqueous solution can then spread over the modified surface (105,126). Bernett and Zisman (97) were of the opinion that the - 43 -reduction in y-|V due to surfactant adsorption was the main reason for the decrease in contact angle (via Young's equation), althoughthey did not deny that some adsorption at the other two interfaces could occur [as in the case of highly-fluorinated surfactants on polyethylene (98)j. They f e l t that the orientation of the hydrocarbon chains towards the vapour phase at the liquid/vapour interface in concentrated solutions made an aqueous drop act l i k e a pure hydrocarbon l i q u i d drop. As Adamson (8) has pointed out, i f no adsorption occurred at either the s o l i d / l i q u i d or the solid/vapour interface, then Y S V - YS-J would remain constant as Y-| V decreased, resulting in a horizontal wetting l i n e on an adhesion tension diagram. This has been only rarely observed for non-polar s o l i d s , and then for only a limited intermediate range of y-|V (see for example the data of Padday (102) for Na lauryl sulphate on paraff i n ) . Based on this reasoning, the w e t t a b i l i t y lines and curves in Figs. 2.4 to 2.7 c l e a r l y indicate that surfactant adsorption must be taking place on at least one of the s o l i d interfaces. To explain the substantial differences observed between the Y £ values of polar low-energy solids obtained with pure liq u i d s and those obtained with aqueous n-alcohol solutions (Table 2.1), Dann (61) used Fowkes equation (2.26) to show that i t was due to an interaction between the polar components of the s o l i d and solution surface tensions, Y|? and ; he assumed that 1 T e was zero, i.e. no adsorption. An alternative explanation for the low and r e l a t i v e l y narrow range of Y C values obtained with alcohol solutions was put forward by Murphy et a l . (89). They postulated that preferential adsorption of the alcohol molecules took place at both the solid/vapour and s o l i d / l i q u i d interfaces via the - 44 -hydroxyl group and that the reduction in Y s v due to this adsorption was the cause of the low Y c values. They also concluded that Y c determined with alcohols was r e l a t i v e l y independent of the s o l i d and depended mainly on the alcohol used. However, Brookes et al . (127) did not agree with the above adsorption mechanism, being unable to accept that, in the absence of charge effects, surfactants would adsorb with the i r non-polar portion oriented towards the aqueous phase. From adsorption studies on Graphon (high-temperature graphitized carbon black) and polystyrene latex particles [the l a t t e r containing s i g n i f i c a n t proportions of polar carboxyl groups (COOH) on t h e i r surface] in aqueous short-chain n-alcohol solutions, Ottewill and Vincent (106) concluded that alcohol molecules adsorb on the low-energy parts of the surface in monolayers with t h e i r OH groups directed towards the solution, while adsorption at the hydrophilic (polar) surface sites occurs with the opposite orientation. Contrary to the results of Murphy et al., they found that Y c for polystyrene did not depend on the type of alcohol in the aqueous solutions (methanol, ethanol or n-propanol), see Table 2.1. The change in i n t e r f a c i a l tension due to adsorption i s described by the general form of the Gibbs adsorption equation (11,18), at constant temperature for an equilibrium, completely reversible system, d.Y = - 2 T | d y. 2.37 i 1 for i components (not including the bulk or sol vent component). An important conclusion i s that positive adsorption (r-j >0) at any inter-face causes a decrease in i n t e r f a c i a l tension. In terms of i t s equilibrium a c t i v i t y , a^, the surface excess concentration of adsorbing species i i s given by - 45 -r. = -VRT(dy/dlna.) T s a „ y l j. e x c e p t y . .... 2 - 3 8 Several investigators have related contact angle to the adsorption densities of one surface-active species at the three interfaces by combining Young's equation (2.1) with Gibbs equation (2.38) (109,128-130), based on the underlying assumption of thermodynamic reversible equilibrium. For many systems, this assumption may not be v a l i d , p a r t i c u l a r l y when chemisorption takes place (131). For the purposes of discussion of adhesion tension diagrams, the relationship derived by Lucassen-Reynders (109) i s presented here: d ( Y s v " Y s l ) _ d ( Y l v C O S 6 ) 1 ^ - 1 ^ d Y n Cry.,- T , ' l v Mv lv .. .2.39 where r s v > T ^  and i y v are the excess concentrations of the surface-active species at the three interfaces and d(y-j v cos e)/dy-j v = a, the slope of a we t t a b i l i t y l i n e on an adhesion tension diagram. Despite the r e s t r i c t i o n s on i t s a p p l i c a b i l i t y , this equation allows some insight into the r e l a t i v e adsorption densities at interfaces from contact angle data. A linear w e t t a b i l i t y l i n e ( i . e . constant slope) i s most simply explained by assuming that Y S V » YS-J and Y-| v remain proportional to each other over the entire concentration range ( i . e . t h e i r adsorption isotherms should have the same shape) (109). A negative slope implies that r s v < r s l > r s-| being f i n i t e (104). Bargeman and van Voorst Vader (123) have pointed > out that a slope of -1 means r s l - r s y = r-| y; they also argued that r s v = 0 for non-polar solids (see section 2.1.3). For more polar solids such as nylon and polyester Smolders (104) concluded that r $ v was positive ( i . e . surfactant adsorption occurred l o c a l l y outside the drop). Positive slopes not only s i g n i f y that i s posi t i v e , but also r s v > r s i • Thus, a - 46 -more positive slope means greater adsorption density at the solid/vapour interface, which implies a greater number of polar s i t e s on the s o l i d ( i . e . greater p o l a r i t y ) . It i s worth repeating here the caution emphasized by Good (10,52) about the use and interpretation of Y c values obtained with surfactant solutions or binary polar l i q u i d mixtures (e.g. short-chain n-alcohol solutions). Such y values can in no-way be related to y determined with pure, non-polar l i q u i d s [as has been suggested (132)], and w i l l not, in general, even be proportional to the surface energy of the s o l i d , Y s or Ygji.e. the w e t t a b i l i t y data cannot be used in the equations of Good and G i r i f a l c o (2.21) or Fowkes (2.26). This i s because the components of a surfactant or binary solution w i l l not adsorb equally at the three interfaces, and for solids with any polar s i t e s in t h e i r surface, r s v and 'r ^ w i l l not be zero. However, the w e t t a b i l i t y curves and Y c values obtained with binary or surfactant solutions s t i l l have great p r a c t i c a l value for characterizing the r e l a t i v e w e t t a b i l i t i e s of s o l i d surfaces. 2.2.4 Wettability and f l o t a t i o n of high-energy solids in aqueous  surfactant solutions The concept of c r i t i c a l surface tension of wetting with respect to aqueous surfactant systems is d i r e c t l y relevant to the hydrophobic/ hydrophilic t r a n s i t i o n in conventional froth f l o t a t i o n systems and has begun to be discussed in t h i s connection i n the f l o t a t i o n l i t e r a t u r e in the l a s t few years (23,133-136). The preceding section was concerned with the use of surfactants to increase the w e t t a b i l i t y of low-energy, hydophobic sol i d s in aqueous solutions. In contrast, the basis of most conventional f l o t a t i o n i s to use surfactants c a l l e d c o l l e c t o r s to - 47 -sele c t i v e l y induce .hydrophobicfty (reduce wettability) by adsorbing on the surface of the high-energy, hydrophilic s o l i d to be floated (with or without the help of modifying agents) and turning i t into a low-energy surface. A detailed discussion of adsorption mechanisms and the conditions required for s e l e c t i v i t y can be found in several references, for example (24,137,138,345,346). Parekh and Apian (139,140) appear to have been the f i r s t investigators to use Zisman's technique with drops of pure liq u i d s to determine for collector-coated minerals. Although problems arose for some systems due to possible interaction between the liqu i d s and the adsorbed surfactant, they generally confirmed Zisman's line a r relationship between Y-|V a n d cos e. For an homologous series of linear surfactants (same polar group, different hydrocarbon chain length) they found that Y c reached a minimum above about C-^, indicating a paraffin (-CH^ ) type surface. The slope b (equation 2.33) appeared to depend on the type of adsorption ( i . e . chem4 isorption or physical), being constant and a maximum for chemisorbed co l l e c t o r s . The slope for physically adsorbed co l l e c t o r s was considered to depend on the degree of hydrocarbon chain interaction, reaching a maximum above C.^ - Emphasis was placed on the importance of the slope for characterizing the type or strength of adsorption, and on Y {'for determining the packing and orientation of the adsorbed surfactant species. Finch and Smith (136) attempted to use the time-dependent change in surface tension of freshly formed liquid/vapour interface in dodecyl-amine s a l t solutions at alkalin e pH to estimate y c of minerals e q u i l i b -rated with the same co l l e c t o r solution. The quantity of particles picked up by a captive bubble replaced contact angle as the w e t t a b i l i t y - 48 -parameter. They obtained a good correlation between Y and bubble pick-up for magnetite, allowing a Y c to be estimated for the particular solution conditions, but not for s i l i c a . The same authors appear to have been the f i r s t f l o t a t i o n investigators to examine the relationship between contact angle and collec t o r adsorption in terms of the adhesion tension diagram (23,129,141), their purpose being to emphasize the importance of the liquid/vapour interface in contact angle development. Their papers contain adhesion tension w e t t a b i l i t y curves for a number of mineral-collector systems based on the rather scarce contact angle-surface tension data available in the l i t e r a t u r e . Fig. 2.9 is a further example based on data for the BaSO^ (barite) -Na dodecyl sulphate (SDDS) system (142,143) (SDDS i s a typical anionic micelle-forming surfactant with a cmc of about 8x10 moles/1 and Ycmc ~ ^  dyne/cm). It i l l u s t r a t e s the f a i r l y typical behavior observed as the concentration of the surfactant increases towards i t s cmc. For the purpose of the following discussion the w e t t a b i l i t y curve has been roughly divided into three regions, X, Y and Z. Most f l o t a t i o n is carried out under conditions corresponding to region X, i.e. a large change in contact angle accompanied by only a r e l a t i v e l y small decrease in Y-Jv- In terms of r e l a t i v e adsorption at the three interfaces (equation 2.39) a large positive slope means that r s v »rs-|> a- conclusion o r i g i n a l l y reached by de Bruyn et a l . (144). A bubble transfer mechanism has been postulated to explain t h i s result (145,188) ( i . e . surfactant adsorbed at the liquid/vapour interface of the bubble i s transferred to the solid/vapour interface on attachment, in addition to what was o r i g i n a l l y at the s o l i d / l i q u i d interface). No direct experimental proof - 49 -critical surface tension of wetting, y dyne/cm solution surface tension , y dyne/cm 'lv Fig. 2.9 Adhesion tension diagram for a high-energy s o l i d - micelle-forming coll e c t o r system (BaSO^ - Na dodecyl sulphate) i l l -ustrating postulated changes in y (for pure liquids) as Y-, and e change; e and y, data from Schulman and Leja (142) ana* Elworthy and Mysels (1^3) - 50 -of the v a l i d i t y of the conclusion is available for f l o t a t i o n systems. In region Y the contact angle reaches a maximum (e ) and remains max r e l a t i v e l y constant over a range of Y-|V« This does not mean that further adsorption at the s o l i d surface ceases; r g v - t must continue to increase in direct proportion to r ^ v to maintain the r e l a t i v e l y constant slope. Y l v approaches Y c m c in region Z and a second layer of surfactant species begins to adsorb with the ionized polar groups oriented towards the solution (due to van der Waals association with the hydrocarbon chains of the f i r s t l ayer). Under these conditions contact angle and f l o t a t i o n response rapidly decrease to zero, the conventional explanation being that the s o l i d / l i q u i d and liquid/vapour interfaces become s i m i l a r l y charged and repel each other (24). Because the slope of the w e t t a b i l i t y curve in this region apparently changes from positive to negative, equation 2.39 states that r g v must become less than r - | ; t h i s i s d i f f i c u l t to v i s u a l i z e in terms of a bubble transfer mechanism. The a p p l i c a b i l i t y of equation 2.39 i s also doubtful in the circumstances when a second surfactant layer begins to adsorb in a reverse orientation at one or more of the interfaces involved. The wetting behavior in the three regions in Fig. 2.9 w i l l now be described in terms of the c r i t i c a l surface tension of wetting concept. Point 1 represents the complete wetting by water of the naturally hydro-p h i l i c s o l i d before c o l l e c t o r addition, Q- At point 2 the 2 2 adsorbed c o l l e c t o r on the s o l i d surface has lowered y„ to y such that - 51 -Y c < < Y l v a n c ' A ^ A R 9 E f i n i t e contact angle i s formed. In region Y at higher c o l l e c t o r concentrations the surfactant approaches i t s maximum adsorption density on the s o l i d surface with the hydrocarbon chain 3 oriented to the solution, represented by.point 3 , resulting in Y c being 3 3 the minimum value of j ( Y _ < < Y 1 w ) . This i s b a s i c a l l y the implied argu-ment of Parekh and Apian ( 1 3 9 ) . At higher surfactant concentrations Y-J V continues to decrease while Y c remains constant or starts increasing, causing a decrease in contact angle ( i l l u s t r a t e d by point 4 ) . Ginn ( 1 3 3 ) and Zimmels et al. ( 134 ) have pointed out that Y c w i l l probably st a r t increasing as cmc i s approached due to adsorption of surfactant in the reverse orientation. At point 5, Y c has increased to such an extent 5 5 that Yp ^ Yn,.> and complete wetting occurs (6 = 0). For the system under v |V 5 consideration y , =Y M „ > but Finch and Smith have implied that Y„ = Yn 1 v cmc c 11 v may occur at solution concentrations s i g n i f i c a n t l y below cmc for some systems. Fig. 2 . 9 also i l l u s t r a t e s that i t i s not essential for the second layer of adsorbed surfactant to make the s o l i d surface completely "hydrophilic" (Y = : Y H Q) for 0 = 0 ; the only requirement i s that 5 5 2 5 Y „ > Y i . ( - Therefore i t could s t i l l be a low-energy surface (Y < Y u n ) c i v C n while being completely wetted by the surfactant solution. 'The oblique, dashed connecting lines in Fig. 2 .9 do not represent Zisman line a r wetting l i n e s . Their purpose i s to i l l u s t r a t e the relationship between Y-I v and Y c for a given point on the w e t t a b i l i t y curve. - 52 -The condition for f l o t a t i o n s e l e c t i v i t y between two mineral types, A and B, Y c < Y l v * Y c '....2.40 has been proposed by Finch and Smith (23). In normal f l o t a t i o n practice the aim i s to adjust the solution conditions using collectors and A B modifiers so that Y C <Y^ Q 4y ( i . e . negligible c o l l e c t o r adsorption on mineral B). The preceding discussion of Fig. 2.9 suggests that s e l e c t i v i t y may s t i l l be possible when substantial c o l l e c t o r adsorption occurs on both minerals, as long as t h e i r resulting w e t t a b i l i t y curves are s i g n i -f i c a n t l y d i f f e r e n t ( r e f l e c t i n g different surfactant adsorption densities or mechanisms at the s o l i d surface), and y-|V can be adjusted independently of c o l l e c t o r concentration to meet the s e l e c t i v i t y c r i t e r i o n given above (inequality 2.40). Finch and Smith appear to be the only investigators to have tested t h i s hypothesis (136). They equilibrated mixtures of a narrow p a r t i c l e -4 size range of magnetite and quartz in 4.08 x 10 M dodecylamine acetate (DDA) solution at pH 9.7. After the solids were separated from the amine solution small-scale "frothless" f l o t a t i o n tests were carried out in aqueous methanol solutions of various concentrations (and hence surface tensions). The results are presented in Fig. 2.10. The magnetite became non-floatable at a s i g n i f i c a n t l y higher value of Y-J v than quartz, allowing complete selective separation to be achieved at Y - | V ~- 40 dyne/cm. Flotation recovery decreased from 100% to 0% over a narrow range of for both minerals. Except in the most concentrated solutions, replacing the methanol solutions with water reactivated f l o t a t i o n in those tests where i t had been suppressed, indicating that non - f l o a t a b i l i t y was not - 53 -Methanol cone. (% by volume , not linear) 0 1 0 2 5 5 0 1 0 0 1 0 0 k 8 0 7 0 6 0 5 0 4 0 3 0 2 0 Solution surface tension, y L V dynes/cm Fig. 2.10 Selective separation of mixture of magnetite and quartz made hydrophobic by conditioning in alkaline dodecyl amine acetate solution, and subsequently floated in aqueous methanol solute ions; after Finch and Smith (136) - 54 -due to the ODA being desorbed from the mineral surfaces. Finch and Smith seemed to infer from these results that selective ' separation was possible because the y c values of the c o l l e c t o r coated minerals, determined with aqueous methanol solutions, were s i g n i f i c a n t l y d i f f e r e n t . However, the discussion of the wetting characteristics of low-energy solids in aqueous n-alcohol solutions already presented in section 2.2.3 casts doubt on whether any such differences in methanol solution Y c values would exist between the col l e c t o r coated minerals (even though a difference may exist between their Y c values obtained with B D pure l i q u i d s ) . If in fact Y C - Y c i n aqueous methanol solutions, the question arises as to how the selective separation was achieved? In an attempt to answer this question in the next section, the concept of c r i t i c a l surface tension of f l o a t a b i l i t y , Y cf> is developed. 2.3 C r i t i c a l Surface Tension of F l o a t a b i l i t y Under actual dynamic f l o t a t i o n conditions (laboratory or plant scale), a p a r t i c l e i s c l a s s i f i e d as floatable only when i t i s successfully separated from the f l o t a t i o n s lurry due to i t s attachment to a bubble.^ For this to occur, not only does partia l dewetting of the p a r t i c l e surface have to be thermodynamically favourable (which has been characterized in the previous sections by the requirement that e> 0 or y-|V >Y c) » but many other c r i t e r i a also have to be s a t i s f i e d [see for example the review by Trahar (146) for a detailed breakdown]* Three main mechanistic steps For the purposes of the present discussion, the mechanical or hydro-dynamic entrapment of non-attached particles in the froth product i s not considered to constitute f l o a t a b i l i t y . - 55 -must be f u l f i l l e d during a single bubble-particle encounter for a f r u i t -ful attachment to occur: (a) the bubble must c o l l i d e with the p a r t i c l e (or with a micro-bubble precipitated on i t s surface) (b) the disj o i n i n g f i l m which separates the c o l l i d i n g p a r t i c l e and bubble must t h i n , rupture and recede before they can rebound from each other ( i . e . adhesion must occur) (c) the strength of the particle-bubble aggregate must be s u f f i c i e n t to withstand the disruptive forces operating in the f l o t a t i o n c e l l . These c r i t e r i a have been expressed in terms of dimension!ess pro b a b i l i t i e s (147-149), the overall probability of a p a r t i c l e being floated being given by P f = P c P a P s ....2.41 where P , P g and P g are the pr o b a b i l i t i e s of c o l l i s i o n , adhesion and aggregate s t a b i l i t y respectively. The rate of f l o t a t i o n of particles depends on th e i r P^ value as well as the encounter frequency (the units usually being in terms of number or weight of particles per unit time). The various c o l l i s i o n theories, reviewed in several publications (148,150-153), assume that the probability of c o l l i s i o n , P c ( a l t e r n a t i v e l y referred to as the c o l l i s i o n e f f i c i e n c y ) , i s independent of s o l i d / l i q u i d and vapour/liquid i n t e r f a c i a l properties, and therefore should not depend on whether y, <y . This i s p a r t i c u l a r l y the case for non-- 56 -slimes particles where i n e r t i a ! effects dominate. Although P c for non-slime sized particles is a function of the physical properties of the p a r t i c l e , bubble and solution, i t i s expected to be constant for constant hydrodynamic conditions and s i g n i f i c a n t l y greater than zero (151,152,163); i t w i l l therefore normally not be the controlling factor in determining whether a p a r t i c l e w i l l be floatable or non-floatable. 2.3.1 Particle-bubble adhesion For a p a r t i c l e to be able to adhere successfully to a bubble, a s o l i d / vapour interface and a f i n i t e contact angle must form during the period in which the bubble and p a r t i c l e are in contact before they rebound from each other. The hydrophobicity of a s o l i d surface defined by the contact angle is therefore not enough to describe i t s f l o t a t i o n properties, since the conditions connected with the kinetics of displacement or thinning of the l i q u i d layer between s o l i d p a r t i c l e and bubble must also be f u l f i l l e d . The analysis of thinning and s t a b i l i t y of l i q u i d films on solids given by researchers such as Derjaguin, Frumkin, Scheludko, Kitchener, Schulze and the i r colleagues, has been discussed in many reviews (22,149, 348-351). The basic theory of Derjaguin postulates that, as a l i q u i d f i l m forms between approaching s o l i d / l i q u i d and gaseous/liquid interfaces, the pressure in the f i l m acting normal to the interfaces becomes different The bubble-particle interaction theory of Derjaguin and Dukhin (154) indicates that, for fine slime-size particles (below about 20 ym in diameter) in surface-active electrolyte solutions, diffusiophoretic effects in Zone 2 around a bubble are involved in particle-bubble c o l l i s i o n . In this respect the surface properties of the p a r t i c l e and bubble can be considered to play a role in determining c o l l i s i o n probability. - 57 -from that in the adjacent bulk solution, the pressure d i f f e r e n t i a l being termed the "disjoining pressure", . The di s j o i n i n g pressure was considered to be a function of the f i l m thickness, h, and i t could be either positive (opposing thinning) or negative (assisting thinning). The main components of the n D for hydrophobic solids are believed to be: positive long-range e l e c t r o s t a t i c repulsion due to overlap of e l e c t r i c a l double layers, n^, for charged interfaces of the same sign; negative in t e r -mediate-range van der Waals a t t r a c t i o n , Jiy, between the approaching phases across the intervening l i q u i d medium; positive short-range " s t e r i c " repul-sion, n<., which remains i l l - d e f i n e d but i s thought to be due to the presence of oriented or structured solvation (hydration) layers due to s p e c i f i c i n -teractions between the solvent and the substrate (such as hydrogen bonding). Each component is a complex function of the f i l m thickness and they are assumed to be additive: n Q = n E + n v + n s 2.42 For a completely hydrophilic s o l i d (on which the wetting f i l m i s stable at a l l thicknesses) n D > 0 and anD/3h < 0 for h = h^ to °° [ h r rep-resents a stable residual hydrated layer of molecular dimensions (connected with 1I5) which has been postulated to exist on most s o l i d surfaces to sat-i s f y the overall free energy of the system, even after thermodynamically favourable bubble-solid attachment has taken pliace (22)]. For a very hydro-phobic s o l i d on which spontaneous f i l m rupture and adhesion occur, < 0 and dU^/dh> 0 for h = h r to °°. In between these two extremes the versus h relationship may have a maximum and minimum depending on the r e l a t i v e con-tributions of i t s componentsparts. The maximum acts as the equivalent of a free energy barrier to thinning and may prevent f i l m rupture i f > 0 and i n s u f f i c i e n t activation energy i s supplied to overcome i t . Metastable - 58 -thin films can form i f the positive (repulsive) disjoining pressure i s balanced by the negative external hydrostatic or Laplace c a p i l l a r y pressure (equation 2.13) at a thickness greater than the c r i t i c a l rupture thickness. Collector - and frother-acting surfactants and inorganic electrolytes can change the r e l a t i v e contribution to of the various d i s j o i n i n g pressure components, and as a result can affect the height (depth) of the pressure or free energy maximum and minimum. The existence of a c r i t i c a l rupture thickness on a " p a r t i a l l y " hydrophobic surface has been confirmed experimentally on smooth methylated s i l i c a surfaces (160,162); i t s magnitude decreased with increasing electrolyte concentration. Evans (155) was of the opinion that the c r i t i c a l rupture thickness was the i n t r i n s i c property of a hydrophobic surface. A s l i g h t l y d ifferent way of looking at f i l m thinning in re l a t i o n to f l o t a t i o n is to discuss i t in terms of so-called "induction time". For successful adhesion to occur the induction time, t , must be less than the contact time, t (149). While there appear to be no rigorous de f i n i t i o n s of either parameter, the majority opinion considers that induction time is the time necessary for the thinning of the disjoining f i l m , i t s rupture, and the spreading of the three phase contact over the s o l i d surface (146, 155-157); however others think the f i r s t step i s the most important (152, 158,1 59)'. There have been several experimental studies of induction time but very few quantitative theories which relate i t to the surface and physical properties of p a r t i c l e , bubble and solution. A small induction - 59 -time has generally been found to be associated with a large contact angle. The work of Laskowski and Iskra (163) on methylated s i l i c a has shown that, under conditions where e l e c t r o s t a t i c repulsion between e l e c t r i c a l double layers i s small, induction time (measured with p a r t i c l e s ) , f l o t a t i o n response and contact angle can a l l be correlated as a function of pH. However, under other circumstances, p a r t i c u l a r l y at low electrolyte concentrations when double layer repulsion was s i g n i f i c a n t , f l o t a t i o n response and induction time did not correlate with contact angle (153,163). Eigeles and Volova (164) have shown that induction time increases quite dramatically as p a r t i c l e size increases. An increase in solution temperature results in a decrease in induction time, and i s thought to be due to a reduction in the v i s c o s i t y of the f i l m and bulk solution (164, 165). Compressing the e l e c t r i c a l double layers with large additions of indifferent e lectrolyte greatly reduces induction time (163). It can also be decreased markedly by the addition of small quantities of frother-acting surfactants (shown by direct measurement and from indirect f l o t a t i o n 6 rate studies) (166-169). I t has been proposed that frother molecules can reduce induction time, and thereby enhance f l o t a t i o n k i n e t i c s , by a variety of mechanisms due to t h e i r physical adsorption in diffuse monolayers at approaching interfaces (but without a l t e r i n g hydrophobicity). They are considered to be able to diffuse rapidly across the thin disjoining f i l m between areas of different surface concentration on the bubble and s o l i d surface, thereby breaking up the structure of the hydrated layer on the s o l i d by reducing the number of hydrogen bonds (167,352). Their high mobility and rapid dipole reorientation a b i l i t y in a diffuse monolayer at the gas/water interface may a s s i s t i n the formation of regions which - 60 -are aligned with oppositely charged areas at the solid/water interface, thus reducing e l e c t r o s t a t i c repulsion (24,352). No doubt surface rough-ness and morphology w i l l also a f f e c t induction time by impeding the move-ment of the three phase contact l i n e across the surface of the p a r t i c l e . One of the few quantitative theories for estimating induction time has been put forward by Scheludko et al (157). They considered that, for small p a r t i c l e s , the rate of thinning of the disjoining f i l m to rupture is the c o n t r o l l i n g process, whereas for large particles i t i s the time of expansion of the solution/vapour interface to the size necessary to r e s i s t the detachment forces of the rebounding p a r t i c l e . This expansion (induction) time was d i r e c t l y related to d ( p a r t i c l e diameter), and P s ( s o l i d density), and inversely related to e, y-|V and the mobility of the receding f i l m (which was a complex function of the surface properties). So i t i s at least in q u a l i t a t i v e agreement with the experimental conclusions. In a recent q u a l i t a t i v e analysis of the relationship between induc-tion time and p a r t i c l e size Jowett (170) suggested that t a d n ....2.43 where n i s a function of flow conditions and p a r t i c l e s i z e , being zero for small slime-size particles in laminar flow, and reaching a maximum of 1.5 for medium to large particles under turbulent conditions, Fig. 2.11(a). Experimental evidence has shown that contact times between particles and bubbles are of the order of a few milliseconds (165,171,172), but very l i t t l e correlation has been made between the magnitude of the contact time and the properties of the system. Three of the most popular theories Fig. 2.11 Suggested relationships between induction time and (a) pa r t i c l e s i z e , adapted from Jowett (170), (b) surface tension of aqueous alcohol solutions and parti c l e size - 62 -developed to estimate t assume that the p a r t i c l e c o l l i d e s with a horizontal solution/vapour interface, e l a s t i c a l l y deforms i t , and then rebounds; the resulting equations have the following general form: t c = X[d 3(p s - P i ) / Y 1 V ] ° - 5 ....2.44 The simplest treatment was by Evans (155) in which A was just a numerical constant r e f l e c t i n g p a r t i c l e morphology; in this case .1.5 / x0.5 -0.5 c " d ' ( p s " p l } ' Y l v In P h i l i p p o f f s more complex analysis the solution/vapour interface was treated as a simple harmonic o s c i l l a t o r (173). It resulted in A being a function of d, P - J and Y - J v such that t c * d n (1 <n <1.5), (p s - P l ) 0 - 5 , but v i r t u a l l y independent of Y - J v (a large increase in Y - | v causes only a s l i g h t increase in t ). In both treatments t £ was assumed to be independent of the s o l i d surface properties ( i . e . contact angle, e). Scheludko et al (174) modified Evan's analysis by more rigorously describing the deforma-tion of the solution/vapour interface, and allowed for the situation where rupture and pa r t i a l recession of the disjoining f i l m takes place before the contact period has elapsed. \ became a complex function of d, (p g - P - J ) , Y-| v» s, and the degree of surface deformation, such that t - d n (1 <n <1.5), (p s - P l ) m (0 <m <0.5). t also tended to decrease as Y 1 v decreased at large p a r t i c l e sizes, but the opposite effect occurred for medium to small p a r t i c l e sizes. If a contact angle formed during the contact period such that e > 0, then t increased with e. - 63 -The theoretical analyses of contact time do give values of the right order of magnitude but they have been c r i t i c i z e d for not being in accordance with the experimental photographic evidence of Bogdanov (as quoted in ref. 22) and Kirchberg and Topfer (165). These workers found that only large energetic particles deformed the bubble surface on c o l l i s i o n , before rebounding without adhering, while medium sized particles under f l o t a t i o n conditions s l i d down the surface of the bubble to i t s lower pole, but did not deform the surface. They concluded that contact time was therefore inversely related to p a r t i c l e s i z e , contrary to the theoretical predictions outlined above. A more recent study using large stationary bubbles in a stream of particles has indicated that adhesion may occur mainly in the tr a n s i t i o n region between impact and s l i d i n g (353). Although there i s some disagreement between theory and experimental observation, i t would appear from the preceding discussion that, for a given set of hydrodynamic conditions in a f l o t a t i o n c e l l , the main interrelated factors that determine whether adhesion w i l l be successful during a particle-bubble encounter are: s o l i d surface l.yophobicity ( w e t t a b i l i t y ) ; d isjoining f i l m energy barrier; p a r t i c l e size and density (mass). There w i l l be some combination of these parameters which w i l l determine the threshold adhesion/non-adhesion condition for a p a r t i c l e under the given hydrodynamic conditions. On one side of the threshold condition, P. w i l l be very small, approaching zero. [P_ = 0 corresponds a - a to complete w e t t a b i l i t y ( h y d r o p h i l i c i t y , e = 0) of the s o l i d surface]. On the other side of the threshold condition i t i s expected that'P w i l l increase rapidly as either the w e t t a b i l i t y , p a r t i c l e mass or the thin - 64 -f i l m energy barrier decrease. Using experimental data from the l i t e r a t u r e Jowett (170) showed that a large decrease in P (represented a by f l o t a t i o n response) can resu l t from only a very small increase in induction time above a threshold value (of the order of a few m i l l i -seconds). He concluded that such a change can result from only a small change in either p a r t i c l e size or the state of the s o l i d surface (not necessarily evident as a detectable change i n contact angle). Scheludko (354) (as quoted in ref. 163) has suggested that the contact angle could be d i r e c t l y related to P 3 only i f there were no kinetic resistances to a f i l m thinning, or i f the kinetic resistance depended on the same para-meters as the contact angle. For a homogeneous inherently hydrophobic s o l i d in aqueous short-chain n-alcohol solutions the w e t t a b i l i t y (e) is controlled.by the solution concentration or surface tension ( l i n e A, Fig. 2.12). As Y - ] V decreases towards the thermodynamic l i m i t of y c> the threshold adhesion condition w i l l be reached for a given p a r t i c l e s i z e , d'; i t can be,characterized by the solution surface tension at that point, defined as the c r i t i c a l surface tension of adhesion, y'. e' i s the corresponding c r i t i c a l Ca Ca contact angle of adhesion. y"l and e" are the threshold conditions for ca ca a larger size p a r t i c l e of the same s o l i d , d" (induction time has been shown to increase with p a r t i c l e s i z e , and i t i s therefore expected that a larger degree of lyophobicity w i l l be necessary for the threshold adhesion condition to be met under the same hydrodynamic conditions). This analysis suggests that i f f l o t a t i o n i s carried out in a solution with y ' a <y-|V ^ c a ' the small particles w i l l readily adhere, but the large ones w i l l not. If i t i s accepted that induction time increases rapidly with p a r t i c l e size (based on theoretical and experimental evidence), and the assumption - 65 -Fig. 2.12 C r i t i c a l surface tension and contact angle of adhesion,and c r i t i c a l aggregate s t a b i l i t y lines and,surface tension of s t a b i l i t y , for two p a r t i c l e sizes d 1 and d" of an inherently hydrophobic s o l i d with w e t t a b i l i t y line A. - 6 6 -i s made that t is independent of p a r t i c l e size (due to disagreement between theoretical and experimental predictions) i t is suggested that a l i m i t i n g p a r t i c l e size w i l l be reached where the probability of adhesion w i l l become very small under a given set of f l o t a t i o n conditions, no matter how lyophobic the p a r t i c l e surface i s . Jowett (170) believes that this argument can be used to explain adequately the poor f l o a t a b i l i t y of coarse particles of some col l e c t o r coated sulphide minerals without resorting to particle-bubble aggregate disruption mechanisms (discussed in the next section). Fig. 2.11(b) is presented as a qu a l i t a t i v e i l l u s t r a t i o n of the type of relationship considered to exist between induction time, contact time, pa r t i c l e size and the surface tension of aqueous solutions of a short-chain n-alcohol. The shape of the t vs. Y - j v curves was suggested by the experimental dome-shaped f l o t a t i o n recovery curves found for coal as a function of alcohol concentration (22,175). As Y 1 v i s reduced below Y H Q due to the i n i t i a l addition of small amounts of alcohol, induction time decreases rapidly r e f l e c t i n g the "frother" effect on the kinetics of f i l m thinning. However as the alcohol concentration increases further i t s wetting effect begins to dominate ( i . e . the reduction of contact angle or lyophobicity) so that induction time w i l l begin to increase again. The curve for the smallest p a r t i c l e s , d . , intersects the contact time l i n e r rain at y ; at higher surface tensions they readily adhere. The curve for the ca largest p a r t i c l e s , d , does not intersect the t l i n e , meaning that they 3 r max c w i l l not adhere (and therefore not be floatable) over the whole range of Y-J values. The l i m i t i n g intermediate p a r t i c l e size for adhesion i s represented by the d' curve and Y ^ 3 - In this simple analysis i t i s assumed that t i s also independent of Y 1 v (mainly because the various - 67 -theories previously outlined which connected these parameters gave con-f l i c t i n g predictions). 2.3.2 Bubble-particle aggregate s t a b i l i t y The poor recovery (or f l o t a t i o n response) observed for coarse hydro-phobic particles (inherent or induced by c o l l e c t o r addition) i s usually ascribed to the rupture of particle-bubble aggregates in the turbulent regions of the f l o t a t i o n c e l l (151). There have been a number of attempts to analyse t h e o r e t i c a l l y the s t a b i l i t y of a single p a r t i c l e attached to a solution/vapour interface under the influence of gravitational or turbulent forces (21,22,174,177,178-180). They a l l predict that the maximum floatable p a r t i c l e size i s d i r e c t l y related to tithe contact angle and solution surface tension and inversely related to the p a r t i c l e density and detachment acceleration. I t i s therefore implied that (assuming c o l l i s i o n and adhesion have taken place) there w i l l be threshold values of y-|v and e below which the attachment of a p a r t i c l e of a given size and density w i l l be unstable for a given set of disruption conditions ( i . e . P g -»- 0), even though e s t i l l has-a f i n i t e value. Above these threshold or l i m i t i n g conditions i t i s expected that P s w i l l rapidly increase. Huh and Mason (177) have rigorously analysed the gravitational situation of a spheroidal p a r t i c l e at a horizontal solution/vapour, i n t e r -face for s t a t i c and rotational conditions. Similar predictions were obtained for a more sim p l i f i e d model of the same configuration by Scheludko et a l . (174) who also analysed a dynamic situation where they assumed that s t a b i l i t y was determined by the kinetic energy of the - 68 -pa r t i c l e rebounding from the interface (as a function of r e l a t i v e velocity of impact, v). The resulting equation is dmax,v = * ^ l v ( r - c o s e ) 2 / ( p s - p n)v 2 ....2.45 where f i s a complex function of d, 9, Y-J v and degree of deformation of the interface. Using parameter values considered to.be representative of a practical f l o t a t i o n environment, d m a x v was found to be much less than d„,,„ r, (gravitational situation) for a l l values of contact angle, and that max,g f l o t a t i o n of a l l sizes became impossible fore<20°. Schulze (178) also s p e c i f i c a l l y dealt with the dynamic situation by analysing a spherical p a r t i c l e attached to a bubble subject to disruptive forces in turbulent vortices. Turbulence was modelled by the r e l a t i v e velocity of the aggregate to that of the water, and reasonable agreement was found between experiment and theory. Bubble size was an important parameter in this model. Schulze and others (22) have also emphasized that bubble-particle aggregates w i l l be s t a b i l i z e d by dynamic advancing contact angles and anisometric p a r t i c l e shape (edge eff e c t , section 2.1.1). The precise relationship between Y-| V and e i s d i f f i c u l t to extract from the complex dynamic models of Scheludko et al and Schulze, but on a qua l i t a t i v e basis they are inversely related; i.e. for a given p a r t i c l e s i z e , e increases as Y-| v decreases. A more simple s t a t i c analysis of the maximum floatable size of a c y l i n d r i c a l p a r t i c l e attached by i t s f l a t end to a spherical bubble has been presented by Morris (179) who took into account the two forces acting to disrupt the attachment (gravity and Laplace c a p i l l a r y pressure) and two forces acting to maintain the attachment (surface tension and residual hydrostatic pressure). From his force balance the following - 69 -relationship between e and Y, i s obtained si n e = B / Y 1 v + C ....2.46 where B [d 2 ( p s - ....2.47 and d C ....2.48 (for the case where the length of the cylinder i s equal to i t s diameter, curvature are equal, equation 2.13). Both Morris and more recently Jowett (170) have proposed that this analysis could be used to approximate the s t a b i l i t y situation under dynamic conditions i f g was multiplied by a suitable "g factor", ng, to represent the centrifugal acceleration experienced by the p a r t i c l e r e l a t i v e to the bubble in turbulent vortices. Jowett proposed that where v. i s the tangential velocity at radius r from the vortex centre. L V Schulze (178) considered that the translational v e l o c i t i e s of 20-60 cm/sec, determined by Mika and Fuerstenau (.159) for turbulent conditions in a f l o t a t i o n c e l l , were reasonable estimates of v^. He also claimed that the dimensions of vortices important to aggregate s t a b i l i t y are of the same order of magnitude as the p a r t i c l e s i z e . For r y = 100 ym (0.01 cm) and v t = 40 cm/sec, equation 2.49 gives ng-160. The data in Table 2.2 have been calculated, using equations 2.46 -2.48, to provide a rough i l l u s t r a t i o n of the type of relationship predicted between o and Y-J v for c r i t i c a l aggregate s t a b i l i t y conditions for different p a r t i c l e s i z e s , assuming constant bubble size and g factor. The calculated data indicate that, for the larger p a r t i c l e sizes, e increases as y-] v d ; r b i s the bubble radius, assuming that the principal r a d i i of n ....2.49 - 70 -Table 2.2 Calculated c r i t i c a l values of contact angle for bubble-particle aggregate s t a b i l i t y , as a function of surface tension, Y 1 V » -3 -3 and p a r t i c l e s i z e , d; p = 3 g cm , p, =1 g cm ,n = 160, i> I g r, = .02 cm. Y 1 V dv cm dyne cm"1 0.02 0.015 0.01 0.005 70 46.4 27.3 14.5 5.6 55 51.7 28.8 14.5 5.1 40 63.1 31.5 14.5 4.4 decreases, giving c r i t i c a l s t a b i l i t y lines for sizes d^ and d^ s (150 and 200 ym respectively) in Fig. 2.12 which are approximately line a r and p a r a l l e l to the 8 = 0 ° boundary. Below 100 ym 9 starts to decrease s l i g h t l y as decreases. It can be seen from equations 2.46 - 2.48 that the p a r t i c l e size at which this t r a n s i t i o n occurs becomes smaller as the p a r t i c l e density becomes larger and the bubble size decreases. The intersection of the c r i t i c a l p a r t i c l e size s t a b i l i t y lines with the w e t t a b i l i t y l i n e A gives the c r i t i c a l surface tensions of aggregate s t a b i l i t y , y^ and y'^ for the two p a r t i c l e sizes, below which i t is ex-pected that P s w i l l be very small. Fig. 2.12 has been drawn in such a way as to i l l u s t r a t e another aspect of f l o a t a b i l i t y previously mentioned; depending on p a r t i c l e s i z e , either Y C A or Y c s could be the threshold f l o a t a b i l i t y parameter, whichever i s larger. For the smaller p a r t i c l e s i z e , y' >y' ; therefore, for. y' < Co. CS CS Y , < Y' , f l o a t a b i l i t y i s not possible even though the aggregate would be - 71 -stable i f i t could form. The opposite situation applies for the larger p a r t i c l e s i z e , d". An example of a f l o t a t i o n situation where adhesion was the co n t r o l l i n g f l o a t a b i l i t y condition has been provided by Anfruns and Kitchener (153). The f l o a t a b i l i t y of strongly hydrophobic methylated glass beads was reduced to almost zero in solutions of a wetting surfactant which markedly lowered Y-J v (Na dodecyl sulphate); they could not adhere to the bubbles, even though the s t a t i c receding contact angle measured under the same conditions was about 30°. However, i f attachment did take place the subsequent bubble-particle aggregates were very stable. Admittedly, the experimental conditions were very quiescent in comparison to those experienced in mechanically agitated f l o t a t i o n c e l l s . A greater degree of turbulence tends to improve c o l l i s i o n and adhesion, but also makes detachment more l i k e l y (181). It i s generally considered that the receding contact angle, e R , is the important parameter for adhesion (178), whereas for aggregate s t a b i l i t y the advancing angle, e^, i s applicable (153,174).- Since e R i s usually substantially smaller than for non-ideal systems,it would probably be more appropriate to analyse y and y • in terms of separate receding and C 9 CS advancing w e t t a b i l i t y l i n e s . However, this would not a l t e r the general conclusion that f l o a t a b i l i t y of an inherently hydrophobic s o l i d in aqueous solutions of a non-ionic, non-micelle forming, wetting surfactant depends on YT being greater than either y , or y ; i.e. a p a r t i c l e w i l l only have a I v ca cs si g n i f i c a n t probability of being floated i f Y - j v > Y c f where Y c f i s the c r i t i c a l surface tension of f l o a t a b i l i t y . - 72 -At this point some attention should be given to the minimum contact angles required for f l o a t a b i l i t y , be they s t a t i c or dynamic. Theoretical analyses of minimum contact angles have mostly dealt with conditions necessary for particle-bubble aggregate s t a b i l i t y when subject to multiples of grav-i t a t i o n a l acceleration (21,22). They predict values ranging from less than 1° for small low density particles up to 90° for large heavy ones. Other theoretical analyses have indicated a minimum advancing contact angle of about 10° to 20° for particles above 50 ym diameter for simulated .dynamic conditions approximating those found i n a f l o t a t i o n c e l l (174,177,178). Mackenzie and Matheson (182) used the theory of Bogdanov et a l . (180) to calculate minimum angles in water of 9° and 29° for 100 ym and 500 ym diameter coal p a r t i c l e s respectively. Some experimental data i n the l i t e r a t u r e show f l o t a t i o n response f a l l i n g to zero as the corresponding measured, s t a t i c contact angles ('usually equi-librium or advancing) decrease to between 10° and 20° (153,183-186). In contrast, other studies have shown good f l o t a t i o n response when the corres-ponding s t a t i c contact angles were zero (167,187). In such cases, time dependent (dynamic, non-equilibrium) phenomena involving surfactant diffusion to, and co-adsorption, condensation and re-arrangement at, the three involved interfaces have been proposed (187,188,190). The concept of threshold hydro-phobicity has been used by Leja and Poling (48) to explain the adherence of small particles to bubbles when a f i n i t e contact angle and adhesion on a f l a t surface of the same s o l i d are almost undetectable. They argued that a zero detectable contact angle means only the absence of deformation of the solution/vapour interface, and not necessarily a complete absence of hydro-phobicity. However, at such low levels of hydrophobicity, the induction time may be too long to allow adhesion of medium to large size particles to take place in a f l o t a t i o n c e l l . - 73 -It i s therefore obivous that, at least i n some situations, the c r i t i c a l contact angles predicted by the s t a t i c w e t t a b i l i t y lines (either advancing or receding) in the adhesion tension diagrams w i l l not be the same as the instantaneous ones which actually occur in the f l o t a t i o n system. However, i t should not a l t e r the argument that for a given inherently hydrophobic s o l i d there w i l l be a Y C ^ > Y C which i s a function of the physical and sur-face properties of a pa r t i c l e and the properties of the thin disjoining f i l m ( i f adhesion controls f l o a t a b i l i t y ) . 2.3.3 Selective f l o t a t i o n separation of inherently hydrophobic solids In section 2.2.3 i t wa-s shown that a selective wetting region exists between two inherently hydrophobic solids whose Y values were substantially d i f f e r e n t in aqueous solutions of the same wetting surfactant (Fig. 2.8). I t was suggested that i t could be exploited to achieve a selective separation by f l o t a t i o n between the two s o l i d s . The development of the concept of a c r i v t i c a l surface tension of f l o a t a b i l i t y , Y indicates that a s i m i l a r ex-ploitable selective f l o a t a b i l i t y region may exist between two inherently hydro-phobic solids which have the same Y . The s e l e c t i v i t y c r i t e r i o n i s given by A B o cn Y c f < Y l v « Y c f ....2.50 Some possible situations are i l l u s t r a t e d in Fig. 2.13 for three hypothetical s o l i d s , A, B and D, with d i f f e r e n t wetting characteristics (slope) but similar Y C values. In (a) the c r i t i c a l contact angle of f l o a t a b i l i t y , ecf» i s the same for each s o l i d ( i . e . the physical properties of the p a r t i c l e s , the thin f i l m properties, and the hydrodynamic conditions, are a l l assumed to be constant) so that r e l a t i v e f l o a t a b i l i t y depends only on the surface lyophobicity,e , which i s controlled by Y2V; (b) represents Y ^  increasing from s o l i d A to D (the case, for example i f density or parti c l e size increased from A to D); (c) represents the opposite situation to (b). In some circumstances there may be very l i t t l e difference in Y ^  values, thereby precluding any selective Fig. 2.13 Possible differences in c r i t i c a l surface tension of f l o a t a b i l i t y for three inherently hydro-. phobic s o l i d s , A, B and D in aqueous solutions of a short-chain n-alcohol; (a)e cf the same for each s o l i d ; (b) e{Jf < ejjf < e£ f; (c) e[L < e£f < e§ f - 75 -separation, while in others substantial selective f l o a t a b i l i t y regions may well e x i s t . The arguments presented in the l a s t few sections have assumed that the surface w e t t a b i l i t y properties Of an inherently hydrophobic s o l i d can be represented by a single w e t t a b i l i t y l i n e . However, i f the surface properties are heterogeneous, as they would l i k e l y be in the particulate state, then a wet t a b i l i t y band would be more appropriate, i t s breadth depending on the degree of heterogeneity. A particulate sample would also have a continuous size d i s t r i b u t i o n between an upper and lower l i m i t , and possibly a d i s t r i b u t i o n of p a r t i c l e density (and morphology), both of which can affect Y c f •'• E v e n for closely controlled f l o t a t i o n c e l l and reagent parameters, there w i l l be a di s t r i b u t i o n of hydrodynamic conditions and thin f i l m properties which w i l l also influence Y ^. A l l these factors mean that for a given particulate sample there w i l l be a Y c f band instead of one discrete value. This i s schematically i l l u s t r a t e d in Fig. 2.14 for particulate samples of two different inherently hydrophobic solids with w e t t a b i l i t y bands A and B. I t i s to be expected that, under some circumstances, p a r t i c u l a r l y for samples with wide size d i s t r i b u t i o n s , the bands may overlap to some extent so that com-plete selective separation between the two solids may not be readily achieved. However, even under these conditions, some preferential concentration would l i k e l y occur, and with p r e - c l a s s i f i e d feed samples complete separations may well be possible i f the Y bands, do not overlap, as i l l u s t r a t e d in Fig. 2.14. If the Y ^ bands completely coincide, then no preferential concentration would appear to be feasible. - 76 -Fig. 2.14 Selective separation region between c r i t i c a l f l o a t a b i l i t y bands for two inherently hydrophobic solids with w e t t a b i l i t y bands A and B. - 77 -CHAPTER 3 SURFACE PROPERTIES OF SOME INHERENTLY HYDROPHOBIC SOLIDS 3.1 Theories of Inherent Hydrophobicity Several interpretations of the basis of native f l o a t a b i l i t y or i n -herent hydrophobicity of natural solids have been put forward. S t r e l t s i n (191) considered structural differences in the hydrated layers at hydro-p h i l i c and hydrophobic s o l i d surfaces i n relation to the crystal structure of the s o l i d , equating native f l o a t a b i l i t y to weak surface hydration. Gaudin et al . (192) further emphasized s o l i d crystal structure and proposed that native f l o a t a b i l i t y results when at least some fracture or cleavage surfaces form without rupture of interatomic bonds other than residual (van der Waals) bonds. These surfaces can only interact with the environment via dispersion and other van der Waals forces. Rupture of co-valent or ionic bonds pro-vide surface sites which can react chemically with suitable adsorbate ions or molecules i n an aqueous environment. The energies of the two types of bonds d i f f e r by an order of magnitude (118). Derjaguin and Shukakidse (193) quantitatively introduced the i n t e r -action of e l e c t r i c a l double layers in the wetting f i l m between approaching pa r t i c l e and bubble as a factor in determining whether or not a s o l i d had "bulk" (inherent) hydrophobicity. They analysed the s t a b i l i t y of the wetting f i l m in terms of the c l a s s i c a l DLVO theory (194,195) of c o l l o i d s t a b i l i t y in the absence of adsorption of surface-active ions at the solution/vapour or solid/solution interfaces. Under these conditions the disjoining pressure in the thin f i l m is equal to the sum of the el e c t r o s t a t i c force and the van der Waals force ( a t t r a c t i v e for inherently hydrophobic s o l i d s ) . They determined that for e l e c t r o s t a t i c repulsion the - 78 -c r i t e r i o n of f l o a t a b i l i t y was given by where K i s a constant close to 3, H is a positive constant representing the stronger van der Waals attraction between water molecules themselves than between water molecules and the atoms in the s o l i d surface, 5 i s the zeta potential, e is the d i e l e c t r i c permeability of water, and D i s the Debye thickness of the ionic atmosphere. It predicts that f l o a t a b i l i t y should improve as the zeta potential is reduced. They found quantitative agreement between t h e i r theory and the experimentally determined f l o a t a b i l i t y of antimonite ( s t i b n i t e , Sb,,S3) as a function of pH, and concluded that, "the value of the zeta potential can serve as one of the characteristics in determining the f l o t a t i o n a c t i v i t y of antimonite". The theory of Derjaguin and Shukakidse has since been tested on several other inherently hydrophobic solids as well as s t i b n i t e . Arbiter et al. (118) found that while the f l o a t a b i l i t y and contact angle for s t i b n i t e did increase as the zeta potential was reduced, there was only qu a l i t a t i v e agreement with theory. The use of a frother by Derjaguin and Shukakidse was considered to be one reason for the disparate results. However, for other solids [molybdenite (MoS^), graphite and paraffin] there was no correlation between these parameters even though r e l a t i v e l y large changes in zeta potential were observed. The behavior of s t i b n i t e was considered to be a consequence of i t s crystal structure being based on chain units (instead of sheets as i s the case for most other c r y s t a l l i n e hydrophobic s o l i d s ) . Cleavage of s t i b n i t e breaks some Sb-S bonds and exposes other weak Sb-S interchain bonds which act as sites for s p e c i f i c adsorption of 0H~ in a l k a l i n e solutions. To explain a similar variation - 79 -in zeta potential without accompanying changes in we t t a b i l i t y for the other inherently hydrophobic s o l i d s , Arbiter et al suggested that, for s o l i d surfaces which only interact via van der Waals forces, the electro-kinetic properties are primarily determined by the aqueous phase; i.e. the e l e c t r i c a l double layer i s formed ent i r e l y on the water side of the interface and approximates the solution/vapour interface. The sign and magnitude of the zeta potential therefore depend on the orientation of water molecule dipoles and the surface a c t i v i t y of the solute ions with l i t t l e or no influence from the s o l i d . A similar interpretation was put forward by Jessop and Stretton (196) to explain the negative zeta potential and i t s r e l a t i v e independence of pH which they found for sulphur, para f f i n , charcoal and cetyl alcohol c r y s t a l s . Again no correlation between f l o a t a b i l i t y and zeta potential was observed. In contrast to the results of Arbiter et al (118), Chander and Fuerstenau (1.19) did find some quali t a t i v e correlation between the magnitude of the zeta potential and f l o a t a b i l i t y of molybdenite leached in KOH (to remove surface oxides). They postulated that the dependence of zeta poten-t i a l on pH could be due to the following reaction at oxygenated sites on the fracture edges of molybdenite sheets: HMoO" ( s u r f a c e ) ^ H + (solution) + MoO^  (surface) 3.2 In order to explain the correlation of zeta potential with hydrophobicity (contact angle) and f l o t a t i o n response in some cases but not in others, Chander et al. (197) put forward the concept of degree of surface anisotropy. They defined an anisotropic surface as consisting of two main components, one which i s formed by rupture on van der Waals bonds only, and the other by rupture of ionic and co-valent bonds. Surface anisotropy was considered to be an inherent characteristic of most hydrophobic solids and a function - 80 -of the cleavage characteristics and part i c l e s i z e . It was suggested that t h i s gives r i s e to the p a r t i c l e size and shape dependence of contact angle, zeta potential and f l o a t a b i l i t y and helps explain the lack of correlation often observed between them (198). For hydrophobic surfaces with no hydro-p h i l i c sites they agree with Arbiter et a l . (118) that the zeta potential -pH relationship w i l l be analogous to that expected for the solution/vapour interface. The behavior of completely hydrophobic surfaces (no anisotropy) w i l l be e s s e n t i a l l y independent of pH un t i l i t is chemically attacked by strong acid or a l k a l i solutions. Laskowski and Kitchener (199) have described the hydrophobic-hydrophilic transition in s l i g h t l y different terms. Using Fowkes concept of i n t e r f a c i a l energies consisting of independent contributions (section 2.1.2) they considered that a s o l i d surface interacts with water via three independent short-range forces, such that WA = V + WA h + V ••••3-3 d h G where , and are the works of adhesion due to dispersion, polar (hydrogen bonding) and ionic interactions. The c l a s s i c thermodynamic boundary condition for hydrophilic-hydrophobic transition is S = - = 0 (equation 2.11), where Wc(water) = 2y^ y = 146 erg cm . From Fowkes' data and analysis Wftd * 2 (Y J v Y S ) ° ' 5 (equation 2.24 assuming no vapour adsorption) and has not been found to be larger than Wc for water on low-and high-energy s o l i d s . (This conclusion i s not altered even i f the adsorption of water vapour on the s o l i d surface i s considered). Therefore, in general, the condition for inherent hydrophobicity i s - 81 -WAn + WAe < 146 - WAd ....3.4 i.e. hydrophobicity arises essentially from the weakness of adhesion of water to the sol i d . The facts that methylation of s i l i c a particles had negligible effect on their zeta-potential, and that contact angle on methylated s i l i c a plates was not s i g n i f i c a n t l y affected by changes in solution ionic strength^Ted Laskowski and Kitchener to conclude that the presence of e l e c t r i c a l double layers does not exclude hydrophobicity; however, i t has been proved that they do control the thickness of equilibrium wetting films on hydrohilic s i l i c a , metastable films on hydrophobic methylated s i l i c a , and the S t a b i l i t y of methylated s i l i c a suspension (160-162,199,200). As discussed in section 2.3.1, e l e c t r i c a l double layer forces affect the kinetics of bubble-p a r t i c l e contact in f l o t a t i o n because at long range they may exceed the short range surface forces, even though under s t a t i c conditions < (199,201). The so-called " s a l t f l o t a t i o n effect" with naturally floatable minerals i s thought to be a manifestation of this phenomenon (202,327). More recently the f l o t a t i o n rate constant for polystyrene particles was found to increase by an order of magnitude when the absolute zeta potential was halved (203). The negative disjoining pressure ( i n s t a b i l i t y ) in thin water films on hydrophobic solids i s considered by Laskowski and Kitchener to be due to a deficiency of hydrogen bonding in these films as compared with l i q u i d water. In other words, proximity to any non-polar surface imposes an unfavourable configuration on neighbouring water molecules. In terms of s i l i c a they believe that methylation, or dehydration at high temperatures, removes most of the SiOH hydrogen bonding sites making - 82 -In discussing the controversy surrounding the inherent f l o a t a b i l i t y or no n - f l o a t a b i l i t y of 'clean' galena,Poling and Leja (338) emphasized that before an unequivocal answer can be given, the exact state of the mineral surface has to be defined for the given sample in the particular environment. For sulphides whose properties place them close to the hydrophobic-hydrophilic t r a n s i t i o n [e.g. galena and chalcopyrite (169)], W^  - W^  w i l l be very small and only very minor changes at any of three interfaces would be required to s h i f t the balance one way or the other. According to Finkelstein et al. (204) the i n a b i l i t y of sulphide surfaces to form hydrogen bonds is the principal reason why they, unlike oxygen-containing minerals, are close to the hydrophobic-hydrophilic balance. 3.2 Specific Surface Properties Most experimental studies on native (natural) f l o a t a b i l i t y have characterised the s o l i d in terms of contact angle, f l o t a t i o n response and electrokinetic behavior as a function of pH. The behavior and magnitude of a l l these properties have been found to depend to a greater or lesser degree on the sample o r i g i n , method of surface preparation, p a r t i c l e size and measurement technique. However, some general patterns.have been observed which w i l l be b r i e f l y summarized. 3.2.1 Contact angle Contact angles measured on inherently hydrophobic (floatable) minerals have generally been found to be substantially independent of pH except in strongly acidic or basic solutions (where they decrease). This i s the case for s o l i d surfaces considered to have only a small degree of anisotropy ( i . e . which predominantly interact with the aqueous environment through weak van der Waals forces) such as molybdenite and graphite cleavage surfaces - 83 -and paraffin (118). For s t i b n i t e , contact angle decreased continually with increase in pH (118), in a similar fashion to that found for methylated s i l i c a (199), an indication of considerable anisotropy, the degree of hydration of the reactive hydrophilic sites being a function of pH. Contact angles measured on the fractured ends of molybdenite and ta l c particles were much lower than those measured on the cleavage faces (22), and even on the l a t t e r surfaces the range of measured values can be large (119). In contrast, Olsen et al. (115) found no difference in contact angles measured on different faces of orthorhombic sulphur c r y s t a l s , and 1 i t t l e variation in the values measured on the same surface. 3.2.2 Zeta potential In almost a l l cases the estimated zeta potentials were negative over the entire pH range tested. The smallest zeta potentials were usually obtained in the most acidic conditions, increasing with pH t i l l a plateau was reached in the neutral to mildly alkaline region, and then sometimes increasing again in more alkaline solutions; the extent of each of these regions depends on the type of s o l i d . This behavior has been exhibited by st i b n i t e (193), paraffin (118,196,197), molybdenite (118,119,198), t a l c (197,205) ( M g 3 S i 4 0 1 Q ( 0 H ) 2 ) , graphite (197), cetyl alcohol crystals (196) and sulphur (196,206). There are exceptions; Parreira and Schulman (207) found zeta potential for paraffin to be independent of pH above about pH.5, but in more acidic regions i t decreased rapidly, becoming s i i g h t l y positive below about pH 4.5. They postulated H + and 0H~ as potential determining ions, but their results have been considered to be erroneous (197). On a sample of natural sulphur Chander et al. (197) found zeta potential to increase continually with pH and postulated that i t was caused by the - 84 -formation of S0^Hz surface oxidation species such as $2®3~' I n c o n t r a s t Chibowski and Waksmundzki (206) found that zeta potential on sulphur was constant oyer a wide pH range and concluded that H + and 0H~ were not potential determining for sulphur. The zeta potential of methylated s i l i c a has been found to behave in a fashion similar to the f i r s t sulphur sample (199) (although very time dependent, taking about 24 hours to reach equilibrium). 3.2.3 Flotation response It i s more d i f f i c u l t to generalize about the f l o a t a b i l i t y character-i s t i c s of inherently hydrophobic s o l i d s . For one thing, results as a function of pH are very d i f f i c u l t to find for low-energy solids with very l i t t l e anisotropy such as paraffin or PTFE. Also, comparisons have to be made between data obtained in small-scale f l o t a t i o n tests performed with or without frother addition and with different concentrations of indifferent e l e c t r o l y t e . For example, Derjaguin and Shukakidse (193), using n-hexyl alcohol as frother, found reasonable correlation between zeta potential and f l o t a t i o n recovery of s t i b n i t e as a function of pH (the l a t t e r decreased as the former increased); the alcohol had no effect on zeta potential at the small concentrations used. In contrast, Arbiter et al. (118) found only a s l i g h t reduction in recovery from pH 2 to 11 when no frother was used. They also found contact angle to be independent of n-hexyl alcohol concen-t r a t i o n . Flotation response was r e l a t i v e l y independent of pH from mildly acidic to mildly basic conditions for graphite (118), molybdenite (118), sulphur (197,204), and t a l c (197,205,208) [with and without frother (208)]. Chander et al (197) reported a decrease'in recovery of molybdenite as pH increased in the acidic region, while that of pyrophyllite (hydrous aluminium s i l i c a t e , - 85 -^ 2 ^ 4 ^ 1 0 ^ H ^ decreased continuously from 100% at pH 3,to zero at pH 12. Laskowski and Iskra (163) found that the s a l t f l o t a t i o n response/pH relationship for methylated s i l i c a depended on the degree of methylation of the sample, and in 0.5 N KCl solutions i t correlated with contact angle and induction time as a function of pH. For strongly methylated s i l i c a , f l o t a t i o n response only decreased s l i g h t l y between pH 3-9. It can be concluded from these studies that f l o t a t i o n response of inherently hydrophobic solids w i l l be r e l a t i v e l y independent of solution pH as long as the surface hydrophobicity (contact angle) i s unaffected by acid or a l k a l i , and i f indifferent electrolyte and/or frother i s added to overcome the kinetic barrier to f i l m thinning. - 86 -CHAPTER 4  WETTING AND FLOTATION OF COAL Many coals have been found to have native f l o a t a b i l i t y and inherent hydrophobicity. This fact has been u t i l i z e d for decades 'in the benefica-tion of fine coal by froth f l o t a t i o n and o i l agglomeration. However, coals d i f f e r from most other inherently hydrophobic s o l i d s , l i k e those described in Chapter 3, because they can have extremely heterogeneous structural and compositional properties, which manifest themselves as heterogeneous surface properties. In the geological sense coals are organic sedimentary rocks which have formed from the accumulated remains of carbonaceous plant materials contaminated to various degrees by inorganic substances. The large range of t h e i r Chemical, physical and technological properties i s due to varia-tions in the nature of the deposited materials and the time and conditions of i n i t i a l deposition and biochemical decay (diagenesis), and subsequent burial (metamorphism). A detailed description of the o r i g i n , formation and properties of coal i s given in the books by van Krevelen (209), Francis (210), Stach et al. (211), Berkowitz (212) and Lowry (213). 4.1 Coal Compositional Properties Coal compositional quality i s usually described in terms of three main parameters: (a) rank - represents the degree of chemical and physical change which has occurred in the organic material over geological time, called coal i f i c a t i o n (or maturation). On a qualit a -t i v e basis a rank category refers to the position of a coal in the coal i f ication series peat l i g n i t e - 87 -subbituminous bituminous anthracite [according to North American terminology, ASTM D388-66 (214)]. (b) type - r e f l e c t s the nature and r e l a t i v e proportions of the original plant debris e n t i t i e s . (c) grade- describes the purity of the coal in terms of the amount of inorganic matter incorporated in i t . 4.1.1 Rank The rank of a coal is not a d i r e c t l y measurable quantity. To be defined, a s p e c i f i c physical or chemical property which exhibits regular changes in the course of coal i f i c a t i o n has to be referred to, for example, moisture content, c a l o r i f i c value, volative matter content, v i t r i n i t e reflectance. Each is applicable to a certain region of the coal i f i c a t i o n series and should only be compared for coals of similar type (except the l a s t ) . For this reason the rank indicators should only be determined on one maceral constituent, almost universally v i t r i n i t e . V i t r i n i t e r e f l e c -tance and v o l a t i l e matter content (or fixed carbon) are the accepted rank parameters for high rank coals, the f i r s t increasing with c o a l i f i c a t i o n , while the l a t t e r f a l l s (there is a close relationship between v o l a t i l e matter and hydrogen content). The chemical structure or constitution of the organic material changes with degree of c o a l i f i c a t i o n . No characteristic polymeric units have been found to exist in coal but "average" coal molecules have been postulated (for the v i t r i n i t e component of bituminous coals) which attempt to include a l l the essential molecular features which have been indicated by a variety of chemical and spectroscopic analytical techniques. They generally consist of small clusters of aromatic nuclei (1-3 units in s i z e , and possibly containing N substitution) linked together mainly by a l i p h a t i c - 88 -carbon bridge structures (sometimes containing 0 and S); peripheral groups attached to the nuclei consist mainly of a l i c y c l i c structures, some phenolic hydroxyl (OH) and quinonic carbonyl (C=0), and perhaps a l i t t l e CH^. Very l i t t l e i f any oxygen in the form of carboxyl (COOH), a l i p h a t i c OH, or methoxyl ether (0CH3) i s believed to exist in unoxidized high rank coals (although considerable amounts exist at lower ranks). The reduction in the H/C atomic ra t i o at high ranks (see the whole coal band, Fig. 4.1) has been generally considered to r e f l e c t the increase in the amount of aromatic carbon (aromaticity) at the expense of al i p h a t i c carbon, resulting in ring condensation, and a g r a p h i t i c - l i k e structure for anthracites (211). More recently i t has been shown that a structure based on non-aromatic tetrahedral quaternary C-C bonds, with a much lower degree of aromaticity, can also explain much of the analytical evidence (212). 4.1.2 Tyje On a macroscopic scale two general classes of coal type have been distinguished, those which formed "in s i t u " (autochthonous), and are commonly called humic coals, and those which formed from d r i f t e d accumula-tions of debris (allochthonous), often referred to as sapropelic coals. The humic coals predominate, and depending on th e i r rank, feature four macroscopically d i s t i n c t bands of various degrees of brightness referred to as the lithotypes v i t r a i n , c l a r a i n , durain and fusain, which have f a i r l y c h a r a c t e r i s t i c properties. On the microscopic scale, the smallest recognizable components of the coal structure are called macerals (analogous to minerals in inorganic rocks) while th e i r natural associations with other macerals and with inorganic minerals are termed microlithotypes. Macerals are characterised by th e i r appearance, optical properties and chemical composition and in - 89 -1-21-H/C (atomic) 0-8h 0-6h 0-4 0/C (atomic) x I0" 2 C % daf 4.1 Variation of organic matter elemental composition for high rank whole coal, W, e x i n i t e , E, v i t r i n i t e , V and i n e r t i n i t e . I, as a function of rank; after Mazumdar (217) and Kessler (216) - 90 -most cases can be traced to sp e c i f i c plant debris e n t i t i e s . The type and re l a t i v e proportions of the macerals or microlithotypes in a coal have been shown to have a direct bearing on i t s physical and chemical properties and behavior in coal preparation, coke making and coal conversion processes. Maceral nomenclature and descriptions are given by the International Commission for Coal Petrology (215) and Stach et al. (211). The macerals of hard (bituminous) coals have been c l a s s i f i e d into three groups on the basis of th e i r similar r e f l e c t i v i t y properties, termed v i t r i n i t e , exinite and i n e r t i n i t e . In fresh, unoxidized coals, similar r e f l e c t i v i t y i s equated with similar elemental composition and technological properties. The elemental composition of the organic material ( i n terms of C, H and 0) in the maceral groups, and consequently the whole coal, changes progressively with increase in rank, as i l l u s t r a t e d in Fig. 4.1. The group maceral lines are based on the data of Kessler (216) and the "normal" shaded whole coal band on the selected data of Mazumdar (217); % tfaf (dry ash free) i s i n this case used as the rank parameter. General-l y , as rank increases,the 0/C atomic r a t i o of the three group macerals (and therefore of the whole coal) f a l l s u n t i l i t begins to level off at about the tra n s i t i o n from low-volatile bituminous to semianthracite (^9T%Cj-^). In contrast, the H/C ra t i o of the whole coal is r e l a t i v e l y constant up to about 84%C whereupon i t begins to decrease slowly due to the loss of H in the exinite and then more rapidly due to the loss of H in the v i t r i n i t e and i n e r t i n i t e . In the region of medium v o l a t i l e bituminous rank the properties of exinite approach those of v i t r i n i t e so thatat higher ranks they become v i r t u a l l y indistinguishable. The properties of i n e r t i n i t e remain different u n t i l into the anthracite region. - 91 -For a given carbon content the group maceral curves indicate that i n e r t i n i t e i s richest in oxygen while exinite i s poorest. The situation is reversed i n terms of hydrogen. However, a comparison of th i s sort can be misleading i f one i s interested in the r e l a t i v e chemical composi-tions of group macerals from the same coal (209). In this case they w i l l each have a different carbon content even though they are of the same rank. In general i n e r t i n i t e has the highest carbon content and the lowest H/C r a t i o , exinite has the highest H/C r a t i o and the lowest oxygen content, and v i t r i n i t e has the smallest carbon content, highest oxygen content, with an intermediate H/C r a t i o . V i t r i n i t e i s usually the predominant group maceral ( p a r t i c u l a r l y for northern hemisphere coals) and derives mainly from the woody tissues and bark of trees. I t has intermediate r e f l e c t i v i t y (grey in oil-immersion reflected l i g h t ) which varies regularly with degree of c o a l i f i c a t i o n . V i t r i n i t e r e f l e c t i v i t y has therefore come to be accepted as the most readily determined, unambiguous rank parameter for bituminous coals. In the higher rank bituminous coals v i t r i n i t e i s classed as a reactive coal component because of i t s good swelling and p l a s t i c i t y (softening) properties when subjected to heat (important in coke making). Its r e l a t i v e l y high H/C ra t i o and correspondingly large v o l a t i l e matter content in the lower rank bituminous coals make i t suitable for conversion to l i q u i d fuels via pyrolysis and hydrogenation processes. The macerals in the exinite group are derived from plant spores, pollen, c u t i c l e s , resins and algae and have low r e f l e c t i v i t y (dark grey); they are p a r t i c u l a r l y r i c h in hydrogen, making them a valuable source for producing gaseous and l i q u i d fuels. Unfortunately they usually occur in r e l a t i v e l y small amounts in humic bituminous coals. - 92 -The i n e r t i n i t e group macerals are distinguished by their high r e f l e c t i v i t y ( l i g h t grey to yellow-white) and in some cases the i r characteristic c e l l - l i k e structure. They originate from similar plant material to v i t r i n i t e but have been subjected to greater degrees of oxidation during diagenesis which has reduced thei r hydrogen content. They are mostly inert in coking and liquefaction processes, although semi-f u s i n i t e is considered to have some r e a c t i v i t y (218). The v i t r a i n lithotype essentially consists of v i t r i n i t e while i n e r t i n i t e (and in particular f u s i n i t e ) is the main component of fusain. Clarain and durain are mixtures of the group macerals (or their micro-lithotype associations), cl a r a i n having a f i n e l y layered structure. 4.1.3 Grade Mineral matter i s the general name given to a l l forms of inorganic material in coal, be i t complexed cations and anions or i d e n t i f i a b l e mineral phases. It has been technically c l a s s i f i e d as to the manner and coal i f i c a t i o n period in which i t has become associated with the organic material (211,212,219). Allogenic refers to d e t r i t a l mineral matter transported to the deposition s i t e by wind and water, while authigenic refers to minerals formed "in s i t u " . Minerals formed or deposited during the diagenetic period of coal i f i c a t i o n are called syngenetic, while those formed or deposited during the metamorphic stage are termed epigenetic. As a rule syngenetic minerals are fine grained and intimately intergrown with the organic material. They are d i f f i c u l t to remove with conventional coal preparation methods and are often loosely referred to as inherent ash (ash being the name given to the inorganic residue remaining after the -93 -combustion of the organic matter). By contrast, epigenetic minerals are usually neither fine grained nor intimately associated with the organic matter and are more amenable to removal by coal cleaning (washing). Commonly used synonyms for epigenetic are adventitious or extraneous; they also encompass mineral matter from shale partings and the roof and floor of seams that becomes combined with the coal during mining operations. In unoxidized bituminous coals the amount of organically complexed inorganic ions i s considered to be small. The most common mineral species associated with coals are clays, s i l i c a , carbonates and sulphides (mainly p y r i t e ) , the clays (aluminium s i l i c a t e s ) usually predominating. V i t r i n i t e has generally been found to have the lowest mineral matter content and i n e r t i n i t e the largest (5,21,220,221). Correspondingly mineral association with lithotypes tends to increase in the order: vitrain-clarain-durain-fusain. Excessive mineral matter contents in coal are generally detrimental to coal u t i l i z a t i o n (3,218). In thermal power generation i t can reduce g r i n d a b i l i t y and c a l o r i f i c value, and result in excessive formation of f l y ash, v o l a t i l e trace elements and sulphur oxide gases.v Coke quality can be reduced, and problems created in the d i f f i c u l t s o l i d / l i q u i d sep-aration step i n solvent-refined-coal processes. Certain types of mineral matter can act as catalysts or sites for spontaneous combustion in stock-p i l e s . Increased wear during coal handling and transportation can also result. Inherent ash also includes organically complexed inorganic ions which in low rank or severely weathered coals can account for a s i g n i f i c a n t pro-portion of the total inorganic material. - 94 -On the other hand, some mineral matter can have beneficial effects in s p e c i f i c situations. There are indications that a certain amount i s required for optimum coke strength; i t can act as a catalyst in hydro-genation processes, increasing the y i e l d of l i q u i d product (3,222,223). 4.2 Oxidation The oxygen in "fresh" coal i s the remnant of the oxygen derived from the o r i g i n a l plant material and biochemical environment before burial occurred. When re-exposed to an oxidizing environment, most coals w i l l react with further oxygen, low rank coals being far more sensitive than high rank coals (224). The i n i t i a l stages proceed by the chemisorption of oxygen at aromatic and non-aromatic surface s i t e s , forming peroxide or hydroperoxide complexes i f moisture i s present, and by formation of -C00H, C=0 and phenolic OH functional groups. Prolonged oxidation (under natural conditions called weathering) results in the break-down of the whole coal material into a l k a l i - s o l u b l e substances known as "humic acids" and f i n a l l y into water soluble and gaseous products (212,225,226). Progressive a i r oxidation decreases both the carbon and hydrogen content of a coal while the oxygen content r i s e s . The resulting elemental composition of a severely weathered coal w i l l place i t outside the normal coal band (210). Berkowitz et aL (227) studied the formation of oxygen containing functional groups during a i r oxidation at elevated temperature in Cretaceous and Carboniferous age high rank coals. Only C=0 could be detected in the fresh samples; as oxidation proceeded small amounts of C00H and phenolic OH formed in the Cretaceous coal, but the C=0 increased the most. In contrast, for the Carboniferous coal only very small changes in C00H and C=0 occurred, while phenolic OH increased - 95 -rapidly. The.results pointed out that coal type and history, and the possible c a t a l y t i c effect of mineral matter, could have a marked effect on oxidation paths and kin e t i c s . The analysis of oxygen functional groups in coal by chemical and spectroscopic methods has been reviewed by Ignasiak et al (228) and Speight (229). A multitude of technological tests have also been used as indicators of the degree of oxidation of a coal. They usually measure s p e c i f i c bulk properties with varying degrees of s e n s i t i v i t y . When performed on whole coal samples the interpretation of results i s complicated by the effects of rank, petrographic composition and mineral matter composition. Examples are ig n i t i o n temperature, free swelling index, dilatometry, f l u i d i t y , microhardness, e l e c t r o s t a t i c properties, optical- r e f l e c t i v i t y , composition of pyrolysis gases etc. (209,225,230-234). V i t r i n i t e and exinite are the most readily oxidizable maceral components, becoming less so at higher ranks; i n e r t i n i t e s are not readily oxidized (218). 4.3 Surface Properties The various attempts which have been made to relate hydrophobicity (wettability) and f l o t a t i o n response to the compositional properties of coal have been covered in several reviews (21,221,235-237). Although there is some c o n f l i c t i n g evidence the following general conclusions have been reached:-(a) hydrophobicity and f l o t a t i o n response increase with degree of c o a l i f i c a t i o n up to low v o l a t i l e bituminous and semi-anthracite rank, decreasing somewhat for anthracite., (b) v i t r i n i t e f l o a t s better than i n e r t i n i t e , - 96 -(c) the greater the amount of mineral matter intimately associated with the organic material the lower the hydrophobicity and f l o t a t i o n response, (d) weathering (oxidation) markedly reduces coal hydrophobicity and f l o t a t i o n response. Many years ago Sun (238) made the very reasonable assumption that the surface composition of unoxidized hydrocarbon materials (including coals) reflected t h e i r bulk compositional properties. He developed a semi-empirical f l o a t a b i l i t y index which was a complex function of the moisture and mineral matter content and the organic elemental composition. C, H and S were considered to be hydrophobic components which enhanced f l o a t a b i l i t y , while 0, N,moisture and mineral matter were hydrophilic and had the opposite effect (a very similar d i v i s i o n to the low-energy, high-energy s i t e hypotheses of Zisman, Good, Fowkes and co-workers, sections 2.1 and 2.2). He obtained good correlation between experimental f l o t a t i o n results and the f l o a t a b i l i t y index. However, i t has been pointed out (236) that the index loses i t s u t i l i t y when the hydrocarbon material has been subjected to surface oxidation which can s i g n i f i c a n t l y reduce f l o a t a b i l i t y while i t s effect on bulk composition can barely be detected. 4.3.1 Rank The c l a s s i c a l study of Horsley and Smith (239) showed a regular increase in water advancing contact angles with rank when measured on the v i t r a i n fraction of several coals. A maximum of about 65° was reached at around 88-89%C cj m m^ with a small decrease at higher ranks. Other studies have recorded advancing contact angles up to about 90° for some high rank coals, the magnitude depending on whether a polished or cleavage surface - 97 -was used, and the origin of the coal (238,240-242). Contact angle hysteresis i s generally very large with receding contact angles being very small, normally less than 10°, even for the high rank coals with the largest advancing contact angles. On rough surfaces the receding contact angles have been found to be indistinguishable from zero (240). Brown (235) determined the f l o t a t i o n reagent requirement for optimum recovery as a function of rank and found that i t correlated well with the contact angle relationship ( i . e . minimum reagent addition corresponded to maximum contact angle). The increase in hydrophobicity and f l o t a t i o n response up to low v o l a t i l e bituminous rank correlates with the reduction in oxygen content (and therefore polar surface sites) (243). The decrease in hydrophobicity as anthracite rank is approached is paralleled by the rapid decrease in hydrogen content with a corresponding change in the carbon skeletal structure. Zisman's Y c (measured with pure 1iquids, section 2.2.1) i n -creases as surface groupings change from CH^ (paraffin) to CH,, (polyethylene) to aromatic groupings (polystyrene) to graphite (Table 2.1). Parekh and Apian (120) appear to be the f i r s t investigators to have measured Y c on coal and graphite. They used miscellaneous organic l i q u i d s , water and concentrated s a l t solutions. Y c was found to be r e l a t i v e l y con-stant at about 45 dyne/cm irrespective of coal rank ( l i g n i t e to anthracite). However, the slope of the l i n e a r Zisman plots, b (equation 2.33), gave a similar correlation with rank as did the water contact angles (the slope for graphite was lower than that of anthracite). - 98 -4.3.2 Mineral matter It i s common experience that coal material with the lowest ash (mineral matter) content fl o a t s f i r s t [apart from high ash-content slimes (221)] and that the i n i t i a l ash content of each size fraction o f coal i s important i n determining i t s f l o t a t i o n response. I t i s the basis of the Release Analysis technique (244) which w i l l be b r i e f l y described in section 8.3.1. Contact angle has also been found to increase as the ash content of coal or graphite samples decreases (235,245). Except for rare occurrences a l l the inorganic minerals commonly associated with coal are considered to be naturally hydrophilic (236,238). 4.3.3 Petrographic composition Very l i t t l e research has been done on the r e l a t i v e f l o a t a b i l i t i e s of the maceral constituents of coal mainly because of the d i f f i c u l t y in obtaining s u f f i c i e n t quantities of the individual group macerals, and t h e i r intimate association with varying amounts of hydrophilic mineral matter. In one investigation the f l o t a t i o n response of lithotypes has been found to decrease in the order v i t r a i n , c l a r a i n , durain, fusain (239), but in another i t was c l a r a i n , v i t r a i n , fusain, durain (238). The difference in order i s not surprising as lithotypes, p a r t i c u l a r l y durain and c l a r a i n , can vary substantially in t h e i r maceral and mineral matter composition from one coal to another. There are instances where f l o t a t i o n has been used to concentrate s e l e c t i v e l y v i t r a i n based on i t s greater native f l o a t a b i l i t y , usually with starvation feeding of standard f l o t a t i o n reagents and repeated cleaning of the froth product (5,246,247). The si m i l a r f l o t a t i o n concen-tra t i o n of the r e s i n i t e maceral (one of the exinite group) has also been reported (355). These separations have r e l i e d on differences in f l o t a t i o n - 99 -rate of the coal components, and not on the complete suppression of the unwanted material (356). Brown (235) and Gaudin (21 ) have described various schemes for the selective depression of particular petrographic components by selective oxidization, pH control, and the use of s p e c i f i c depressing reagents such as starch. The estimation of the surface free energy and Y c of low energy hydrocarbon polymers described in sections 2.1.2 and 2.2.1 indicates that the lower the H/C r a t i o of the groups exposed.at the s o l i d surface, the more wettable i t i s . The decrease in the organic matter hydrogen content from exinite to v i t r i n i t e to i n e r t i n i t e for high rank coals, section 4.1.2, would therefore suggest that the w e t t a b i l i t y of these components should increase in the same order. 4.3.4 Oxidation The weathering (oxidation) of coal can occur i n - s i t u , or after ex-posure in open-pit mining, or in stock-piles (224,225,248-250), lower rank coals being more susceptible than those of higher rank. Even i f the degree of oxidation i s not s u f f i c i e n t to a l t e r s i g n i f i c a n t l y the chemical analysis of the coal i t can cause marked deterioration in f l o t a t i o n c h a r a c t e r i s t i c s , in terms of both y i e l d and ash rejection (224,248,251,252). Reagents such as non-polar hydrocarbon o i l s and cationic amine collectors can be used to overcome to some extent the effects of mild oxidation on f l o t a t i o n y i e l d (248,251,253,254), but the excess quantities often required, and the degree of ash rejection achieved, may make the process economically unattractive (255), p a r t i c u l a r l y as weathered coals usually contain a s i g n i f i c a n t proportion of slimes which consume consider-able quantities of reagents (256). However, recently there have been - 100 -reports of substantial improvements in the f l o t a t i o n y i e l d of heavily oxidized coals without s i g n i f i c a n t changes in ash r e j e c t i o n , using newly developed proprietry reagents, for example Dowel! froth conditioner M210 (357). Several techniques have been used to assess the effect of atmospheric oxidation on the surface properties of coal in r e l a t i o n to wetting and f l o t a t i o n . Correlations have been found between degree of oxidation (usually the time and/or temperature of laboratory oxidation) and contact angle (253), f l o t a t i o n y i e l d or rate under standard conditions (224,248, 251,252,257,258), and adsorption densities of anionic and cationic sur-factants from aqueous solution (259). The methods which appear to be the most sensitive to small degrees of surface oxidation are the f l o t a t i o n method of Baranov et a l . (258) in an aqueous solution of hexyl alcohol, and the s a l t f l o t a t i o n technique of Iskra and Laskowski (252). The l a t t e r consists of carrying out the f l o t a t i o n in a 0.5 N potassium chloride solution without any addition of organic f l o t a t i o n reagents. The authors concluded that these two methods were more sensitive to small degrees of oxidation than most other standard indicators. 4.3.5 Electrokinetics Unoxidized coal samples have generally been found to have negative zeta potentials at neutral and basic pH (196,253,260,261,356), with an i s o e l e c t r i c point (iep) in the mildly acidic region ( i f any i s indicated). L i t t l e i f any correlation has been found between the iep and either the rank of whole coals and v i t r a i n s , or the various lithotypes taken from the same coal or different coals. Zeta potential can be either dependent or independent of pH in the neutral pH region. The lack of correlation - 101 -is attributed to the general heterogeneous nature of coal, and in p a r t i -cular the presence of associated mineral matter (196,261,340). Wen and Sun (253) found that laboratory oxidation reduced the iep of a v i t r a i n sample, and made the zeta potential more negative, so that after prolonged oxidation no iep was indicated and zeta potential was v i r t u a l l y independent of pH. In another study the zeta potential of v i t r a i n and durain was reduced in d i l u t e solutions of NaCl and CaCl 2 (261). Klassen et a l . (175) have also reported the reduction in electrokinetic potential in d i l u t e aqueous n-alcohol solutions. It has been postulated that maximum hydrophobicity and f l o a t a b i l i t y of coal should occur at the iep (235). However, to cast doubt on th i s hypothesis there i s evidence showing that f l o t a t i o n response i s either independent of pH over a considerable range, or has an optimum which doesn't coincide with the iep (196,224,262). The maximum f l o t a t i o n y i e l d is usually found near neutral pH, while optimum ash rejection occurs under acidic conditions (235-237). [Iskra and Laskowski (252) found that f l o t a t i o n response for unoxidized and oxidized coal in s a l t f l o t a t i o n was a maximum at low pH, about 2]. 4.3.6 Wetting and adsorption Reagents are used in coal f l o t a t i o n to enhance both the hydrophobicity and f l o t a t i o n response (which are not necessarily synonymous). For d i f f i c u l t - t o - f l o a t coals (low rank or oxidized) additions of non-polar hydrocarbon o i l s such as kerosene or fuel o i l are made to increase hydrophobicity. Various short-chain mono-hydroxyl alcohols are always used to increase f l o t a t i o n rate. Their action has t r a d i t i o n a l l y been - 102 -considered to be the reduction of solution surface tension resulting in smaller bubble s i z e , larger bubble population and the formation of a stable froth layer (339). There has been extensive research into the adsorption of frother-acting reagents on the surface of coal as a function of rank, petrographic composition and mineral matter content (256,263,264), p a r t i c u l a r l y by inves-tigators in the USSR (22,221,257,265-269,358). The general conclusions drawn from the more recent studies are that, for normal alcohols at least, adsorption on unoxidized high rank coal surfaces i s via the hydro-carbon group with the polar group oriented toward the solution, making the surface more wettable. At low degrees of surface coverage by oxygenated polar sites (lower rank coals, or with small degrees of oxidation) adsorp-tion takes place at these sit e s through hydrogen bonding with the frother polar group. For highly oxygenated surfaces (low rank or heavily oxidized) the polar water molecules form a strong hydrated layer which considerably reduces the competing alcohol adsorption. The longer the hydrocarbon-chain, the greater the degree of adsorption. There appears to be some uncertainty in the role that frother-type reagents play in enhancing coal f l o t a t i o n response. Claims have been made that some frothers have "c o l l e c t i n g " properties for coal. Horsley and Smith (239) used the term in reference to pine o i l (a mixture of terpineols) because i t produced a s i g n i f i c a n t increase in contact angle whereas other common frothers had l i t t l e effect [cresylic acid, phenol, MIBC (methyl isobutyl carbinol), 2-butanol]. But, in another study (120) small additions of MIBC and pine o i l did not affect the slope of Zisman plots on coals of different rank, except for a s l i g h t reduction on bituminous coals (equivalent to a decrease in the aqueous contact angle). In contrast, contact angles measured on various unoxidized microlithotypes - 103 -have been found to decrease s i g n i f i c a n t l y after the surfaces were treated with a solution of hexyl alcohol (269). A recent study (359).has also shown that the force required to detach an a i r bubble from a coal p a r t i c l e (either "pure" or coated with n-octane) decreased as the concentration of the frother (diacetone alcohol) increased. The results were explained in terms of frother adsorbing on the coal surface with polar groups towards the solution and free energy changes at the interfaces ( i . e . an increase in w e t t a b i l i t y ) . The preceding findings have to be equated with the generally observed fact there i s an optimum frother addition at which f l o t a t i o n y i e l d i s a maximum (22,236,270). Klassen et a l . (175) found that this optimum also occurred under "frothless" f l o t a t i o n conditions in n-hexanol solutions (floated particles are separated from the suspension without the aid of a stable froth l a y e r ) ; they therefore concluded that "the improvement i n fl o t a t i o n can only be ascribed to the co l l e c t i n g action of the reagent" and postulated that the alcohol molecules only s e l e c t i v e l y adsorbed on the coal surface at polar sit e s with the i r hydrocarbon groups oriented towards the solution, thereby making the surface more hydrophobic. The maximum f l o t a t i o n response was found with normal alcohols having a hydrocarbon chain length of Cg to Cg. Similar conclusions have also been reached with oxidized coals (257,265). However, Lekki and Laskowski (352) are of the opinion that, although "surface-inactive" frothers such as diacetone alcohol, ethyl acetal or ethanol adsorb on mineral (and coal) surfaces, the improved f l o t a t i o n rate or y i e l d i s not caused by a resulting increase in hydrophobicity, but is due to the effect of the very mobile frother molecules on the kinetics of adhesion (discussed in section 2.3.1). - 104 -The results of Klassen et a l . (175) also indicated that the normal alcohols give better ash rejection ( s e l e c t i v i t y ) than other types of frothers used with coal. The complete suppression of coal f l o t a t i o n by excessive additions of frothing agents has been reported (175,236,271). Klassen et al. f e l t that the n-hexanol molecules formed "hydrated assemblies" on the coal surface while Brown (235) thought that a second layer of molecules may adsorb with the i r polar ends directed towards the solution (the same behavior considered to occur for micelle forming surfactants as they approach cmc; n-hexanol does not belong in this category). Szczypa et al. (271) found that the contact angle on both clean and slime-coated coal surfaces decreased to zero as the concentration of a commercial frother increased ( A l i f a l - a mixture of short-chain alcohols); no mechanism was postulated. Coal f l o t a t i o n investigators do not appear to have been con-cerned with the effect of the reduction i n surface tension of aqueous surfactant (frother) solutions on the w e t t a b i l i t y of coal. Investigators studying coal dust suppression are interested in the wetting of coal by aqueous surfactant solutions. A common method for assessing the r e l a t i v e wetting properties of surfactants as a function of concentration has been to determine the time required for the sinking of a powdered sample placed on the surface of the solution in a beaker or test tube (238,272-274). Gl a n v i l l e and Wightman (275,276) have recently shown that in certain cases a minimum threshold concentration of the wetting agent was necessary to produce any discernible wetting of a coal sample. For normal alcohols from methyl to butyl, and isopropanol, the threshold concentration corresponded to the same surface tension for each - 105 -of two different high rank coal samples (about 33 and 40 dyne/cm) (276). They concluded that a c r i t i c a l surface tension was necessary before wetting could begin. The difference between the two values was attributed to differences in sample p a r t i c l e size and surface properties. G l a n v i l l e and Wightman did not attempt to equate these c r i t i c a l surface tension values to Zisman's Y c obtained from contact angle data because thei r system was a dynamic one using a powdered sample, and with binary l i q u i d mixtures instead of pure liquids (see section 2.2.3). In fact c r i t i c a l surface tension determined in this manner is probably more closely related to the c r i t i c a l surface tension of f l o a t a b i l i t y , y ^ (section 2.3), than to the c r i t i c a l surface tension of wetting, Y c> determined on the surfaces of polished coal samples. A direct linear relationship was found between the heat of immersion of the same coal samples in aqueous methanol solutions and solution sur-face tension over the whole range of concentration. G l a n v i l l e and Wightman used this correlation to postulate that methanol preferentially adsorbed over water at the s o l i d / l i q u i d interface, p a r t i c u l a r l y at high concentra-tions. 4.3.7 Coal w e t t a b i l i t y d i s t r i b u t i o n The preceding discussion of coal emphasizes that the particles in a powdered coal sample w i l l have a very wide and continuous range of compositional and surface properties. Therefore, instead of the w e t t a b i l i t y of a coal sample in aqueous alcohol solutions being represented by a d i s -crete l i n e or a narrow band on an adhesion tension diagram (section 2.3.3), a broad band is postulated, as i l l u s t r a t e d in Fig. 4.2. In some cases the band may even extend to the e = 0 boundary, representing completely - 106 -Fig. 4.2 Broad w e t t a b i l i t y band postulated for a particulate sample of a high rank readily floatable coal - 107 -hydrophilic particles (due to weathering, high mineral matter content etc.). The previously quoted work of G l a n v i l l e and Wightman indicates that, at least for unoxidized high rank coals, Y c f° r aqueous alcohol solutions may be in the region of 30 dyne/cm, which is also the case for other low-energy, inherently hydrophobic solids (section 2.2.3). In a particulate coal sample with a range of w e t t a b i l i t y and physical properties ( p a r t i c l e size and density) a range of Y ^ values should also e x i s t , as postulated in section 2.3. By suitable adjustment of the surface tension of an aqueous alcohol f l o t a t i o n solution i t should therefore be possible to determine the d i s t r i b u t i o n of f l o a t a b i l i t y in such a sample, and to divide i t into several fractions of different f l o a t a b i l i t y . - 108 -CHAPTER 5 OBJECTIVES AND SCOPE OF EXPERIMENTATION The purpose of the experimental investigation has been to test whether the proposed concept of c r i t i c a l surface tension of f l o a t a b i l i t y in aqueous surfactant solutions can be used to (a) determine the d i s t r i b u t i o n of f l o a t a b i l i t y within particulate coal samples, (b) assess the r e l a t i v e f l o a t a b i l i t y characteristics of coal samples which differed in t h e i r compositional properties and degree of weathering, (c) characterize the r e l a t i v e f l o a t a b i l i t y of particulate samples of inherently hydrophobic s o l i d s , and (d) see i f selective f l o t a t i o n separation was possible between inherently hydrophobic solids with different y ^ values. Polytetrafluoroethylene, sulphur and molybdenite were chosen as examples of inherently hydrophobic solids because the r e l a t i v e magnitude of th e i r water contact angles available in the l i t e r a t u r e indicated that they differed appreciably in degree of hydrophobicity (PTFE > sulphur > molybdenite); also t h e i r r e l a t i v e densities were substantially greater than that of water. The coal sample properties have been characterized by proximate, ultimate, free swelling index, infrared spectroscopy and petrographic analysis (including v i t r i n i t e reflectance). The surfactant chosen to manipulate aqueous solution surface tension was methanol. It gives a wide range of solution surface tension (about 12 to 72 dyne/cm) but has the least surface a c t i v i t y of any of the alcohols - 109 -(only foams to a small degree, the foam almost immediately breaking down on the cessation of aeration). It i s considered to adsorb only physically on low-energy s o l i d surfaces, including coal (277), (with the possible ex-ception of hydrogen bonding at suitable high-energy sites) but there i s no evidence to suggest that such adsorption enhances surface hydrophobicity ( i . e . i t does not act as a " c o l l e c t o r " ) . Methanol also reaches adsorption equilibrium at freshly created solution/vapour interface faster than longer-chain homologues (<0.005 sec) (278). Qualitative f l o a t a b i l i t y experiments were performed in test tubes to obtain a rapid visual assessment of the Y ^ of a particulate sample. There are several "frothless" small-scale (micro-flotation) tests available for more quantitatively assessing the r e l a t i v e f l o t a t i o n response and threshold f l o a t a b i l i t y of pure minerals under particular solution conditions, some of the more popular ones being the so-called bubble-pick-up (136), the Hallimond Tube (279) and the Partridge-Smith c e l l (280). The l a s t one was selected of i t s better re p r o d u c i b i l i t y (204) and control of parti c l e hydrodynamic carry-over. Samples with narrow nominal size ranges were used to minimize the effects of pa r t i c l e size on f l o a t a b i l i t y ; -210 + 149 ym (-65 + 100 mesh Tyler) was chosen for coal because i t i s in the range generally found to give maximum f l o t a t i o n response (180,182,237). - no -CHAPTER 6 MATERIALS - PROPERTIES AND PREPARATION 6.1 Glassware A l l glassware used to prepare, store and handle single d i s t i l l e d water and aqueous solutions, and carry out test work, was washed in detergent solution (Sparkleen), cleaned with acetone and fresh K^Cr^Oy/ concentrated H 2S0 4 solution (chromic acid), and thoroughly rinsed in d i s t i l l e d water. Glassware was rejected i f beading of d i s t i l l e d water s t i l l occurred after this procedure. The Partridge/Smith (P/S) f l o t a t i o n c e l l parts and test tubes were stored under d i s t i l l e d water when not in use. The medium porosity f r i t t e d glass disc in the bottom section of the P/S c e l l was cleaned at regular intervals in hot aqua regia and concen-trated ammonia to remove fine particles plugging the pores. 6.2 Solutions A l l the aqueous solutions used in the experiments were made up with single d i s t i l l e d water (SDW) obtained from a Pyrex glass d i s t i l l a t i o n unit, Corning Model AG-2, and stored in large Pyrex glass containers. The e l e c t r i c a l conductivity was measured on a regular basis and varied between 1 and 3 ymho/cm at 25°C + 0.2°C. A Radiometer A/S Type CDM2d conductivity meter with a type CDC 104 conductivity c e l l was used. The amount of oxidizable organic matter in the SDW was q u a l i t a t i v e l y tested using the ASTM standard KMnO^  colour retention method (281). The minimum colour retention time was greater than 60 minutes indicating very low levels of oxidizable organic matter. - I l l -The methanol used to make up stock solutions was Fisher C e r t i f i e d ACS grade, Cat. No. A-412 (acetone free, absolute), and used as received. The aqueous methanol solutions were prepared by pipetting the required volume of alcohol into Pyrex volumetric flasks at room temperature (20 + 3°C). 6.2.1 £H A Fisher Accumet Model 230 pH/ion meter and Fisher Standard Combina-tion Glass Electrode were used to measure the pH of the aqueous solutions at room temperature. The pH meter was calibrated with Fisher standard buffer solutions (pH 4.01, 7.41 and 10.4 at 25°C) following the manufacturer's recommended procedures. The pH of the single d i s t i l l e d water was measured at frequent inter-vals and varied between 5.2 and 6.5. For aqueous methanol stock solutions below 40 v o l . % the pH varied between 5.2 and 5.7 and appeared to be independent of methanol concentration; for the more concentrated solutions the pH increased s l i g h t l y with concentration to about 6.0, although i t was d i f f i c u l t to obtain steady readings. 6.2.2 Surface tension The s t a t i c surface tension of the methanol solutions was measured by the Wilhelmy Plate attachment method (8,27) using a Cahn RG electrobalance with a modified dynamic surface tension accessory, and a chart recorder. The measurements were carried out at 20 + 2°C i n a closed compartment in which the atmosphere had been p a r t i a l l y equilibrated with the solution to be measured. Details of the apparatus and procedures are given in Appendix A. The s t a t i c surface tension remained constant over a period of 15 minutes for solutions with methanol concentrations below 20 v o l . % ; at -112 -higher concentrations i t increased s l i g h t l y with time, most l i k e l y due to the evaporation of methanol J The recorded values in Table A . l , Appendix A, are those measured 10 seconds after attachment took place. The mean experimental values and l i t e r a t u r e values (282,283) are plotted in Fig. 6.1. The value for the single d i s t i l l e d water of 72.5 +0.2 dyne/cm (95% confidence i n t e r v a l , tg Q 5 ^ 1 2) compares with a l i t e r -ature value of 72.88 dyne/cm at 20°C (282); for methanol the experimental value of 22.9 +0.1 dyne/cm (tg ^) Q) compares with the l i t e r a t u r e value of 22.50 (282). At low methanol concentrations ( <25 vol.%) the experimental and l i t e r a t u r e values agree reasonably closely; between 25 and 90 v o l . % the former are approximately 1 dyne/cm above the l a t t e r . 6.3 Coal Samples 6.3.1 Description The coal samples used in the experiments came from Cretaceous age seams in the East Kootenay (Crowsnest) C o a l f i e l d , south east B r i t i s h Columbia. The nominally non-weathered No. 5 seam sample was supplied by Fording Coal Ltd., Sparwood B.C., from the Greenhills p i t , sealed in heavy duty polyethylene bags. Based on the dry mineral matter free fixed car-bon content ( F C ^ ^ ) of 76.4% (calculated from the proximate analysis supplied with the sample, Table B . l , Appendix B) this sample was of medium v o l a t i l e bituminous rank according to the ASTM D 388-66(1972) c l a s s i f i c a t i o n system (214). Using the vibrating j e t method i t has been found that the equilibrium surface tension of short chain n-alcohol solutions (methyl to butyl) i s reached in less than 0.005 seconds (278). - 114 -A suite of f i v e other bulk coal samples designated 0', 38', 75', 1111 and 150' (approximately 100 kg each) were provided by Drs. W.H. Mathews and M. Bustin of the Geological Sciences Department, University of B r i t i s h Columbia. These samples were air-picked at various depths (0, 38, 75, 111 and 150 f t . ) in horizontal adit 23, freshly opened in outcropping No. 7 seam on the Kaiser Resources Ltd. property in the Greenhills area. Fig. 6.2 i l l u s t r a t e s the position of the sampling points in the adit. An i n i t i a l layer of coal about 15 cm deep was f i r s t removed from the wall before each sample was collected from a v e r t i c l e channel over the whole seam thickness, and stored in heavy duty bags. Although the analyses in Table B . l , Appendix B, are for the -595 ym (28 mesh Tyler) size fraction only, the R^ dmmf values of about 71% for the deep samples indicate that the coal in adit 23 was of medium v o l a t i l e bituminous rank. 6.3.2 Preparation A 10 kg sample of the No. 5 seam sample was spread out in trays for 24 hours to equilibrate with the laboratory atmosphere and was then crushed to -8 mesh. Half of t h i s material was further size reduced to -28 mesh, and four lots of approximately 1 kg each obtained by r i f f l i n g . One of these lots was dry screened (in approximately 200 g amounts) for 20 minutes on a Ro-Tap machine into several size fractions between 595 ym and 149 ym. A l l sub-samples and size fractions were stored in sealed containers under argon to minimize atmospheric oxidation. The adit 23 bulk samples were sent to the Vancouver laboratory of Commercial Testing and Engineering Co. where they were a i r dried to remove excess surface moisture and the -595 ym size fraction dry screened out from each. Two approximately 500 to 1000 g subsamples were r i f f l e d c o a l Fig. 6.2 I l l u s t r a t i o n of sample-collection points in adit 23, No. 7 seam, Greenhills area, East Kootenay Co a l f i e l d , south-east B r i t i s h Columbia, after Teo et al. (284) - 116 -from the -595 ym material; one was pulverised to -250 ym for analytical purposes while the other was dry screened into several size fractions between 595 and 149 ym in 200 g lots on a Ro-Tap machine for 20 minutes, using standard Tyler stainless steel sieves. The sub-samples and size fractions were stored in sealed containers under argon. 6.3.3 Proximate and ultimate analyses 6.3.3.1 bulk coal The proximate analysis of the No. 5 seam bulk sample was supplied by Fording Coal Ltd., on an a i r dried basis and i s presented in Table B . l , Appendix B on a dry basis. The proximate analyses, total sulphur, free swelling index (FSI) and gross c a l o r i f i c value on the -595 ym size fraction of the adit samples were carried out in the assay laboratory of the Department of Mining and Mineral Process Engineering, while the ultimate analyses were performed by Commercial Testing and Engineering, a l l following ASTM standard procedures (214). The results are l i s t e d in Table B . l , Appendix B, on the-dry and dry mineral matter free basis. The state of oxidation of these samples has also been analysed by several other techniques including chemical functional group analysis, KBr p e l l e t infrared (IR) spectroscopy, solvent extraction in hexane, benzene and pyridine, and analysis of these extracts u t i l i z i n g IR and proton-nuclear magnetic resonance (H-NMR) spectroscopy and gas-liquid chromatography (GLC) (284). 6.3.3.2 -210 + 149 ym size fraction The analytical data for the -210 + 149 ym size fraction of each coal sample are given in Table 6.1, the proximate and total sulphur data being the mean of duplicate analyses. The proximate, total sulphur and FSI - 117 -analyses were performed in the Department of Mining and Mineral Process Engineering following ASTM standard procedures. The ultimate analysis data (C,H and N) are the means of one set of C and H analyses performed in the above laboratory following ASTM standard procedures, and two or three replicate C, H and N instrumental microanalyses done by Canadian Microanalytical Service Ltd., Vancouver. A comparison of the analytical data in Table B . l , Appendix B, for the bulk coal samples with the data in Table 6.1 shows that in almost a l l respects the -210 + 149 ym material has very similar compositional properties to i t s parent bulk material; the most obvious exception appears to be the FSI values for the 38' sample. The replicate raw C, Hand N data were converted to the dry mineral matter free base and single factor analysis of variance (ANOV) and a Student-Newman-Keuls multiple comparison of means test (SNK) (285) performed to determine whether there were s t a t i s t i c a l l y s i g n i f i c a n t differences between the 0', 38', 75', 1111 and 150' samples in terms of their mean, C, H and N contents. Appendix C gives the details of the s t a t i s t i c a l methods used, while the calculations and tests of hypotheses are presented in Appendix D, Tables D.l - D.3. 6.3.4 Scanning electron microscopy The micrographs in Fig. 6.3 are SEM images of the -210 + 149 ym size fraction of the six coal samples (vacuum gold coated, using an ETEC Corporation Autoscan Scanning Electron Microscope). They indicate that the shape and size d i s t r i b u t i o n characteristics of the particles and the small amount of fine "slimes" adhering to the larger particles are similar for the No. 5 seam and 38', 75', 111' and 150' adit samples. The effect of severe weathering on the f r i a b i l i t y and the quantity of slimes can be seen for the 0' (surface sample). - 118 -Table 6.1 Analysis data for -210 + 149 ym size fraction of coal samples used in f l o a t a b i l i t y tests. Sample Depth No. 5 adit 23, No. 7 seam seam 0 38 75 111 150 a i r dried (ad) Moisture (M),• wt.% 0.8 15.9 2.1 0.7 0.9 0.9 Free Swelling Index (FSI) 3 V 3 % 0/0 3/3 7/6% Ihllh 8/7% dry (d) wt. % Ash (A) 19.9 18.4 12.0 11.1 10.1 9.6 Vo l a t i l e Matter (VM) 20.2 29.2 a 25.4 26.9 26.7 26.9 Fixed Carbon (FC)(diff.) 59.9 52.4a 62.6 62.0 63.2 63.5 TOTAL 100.00 100.0 100.0, 100.0 . 100.0 '. 100.0 Total Sulphur (S) 0.33 0.42 0.54 0.65 0.64 0.74 Mineral Matter (MM)b 21 .7 20.1 13.2 12.3 11.3 10.8 dry mineral matter free (dmmf), wt.% Carbon (C) 90.3 80.9 86.7 88.0 89.0 88.0 Hydrogen (H) 4.6 3.2 4.8 5.0 5.1 5.0 Nitrogen (N) 1.5 1.6 1.6 1.6 1 -6 1.7 Oxygen (0) ( d i f f . ) 3.6 14.3 6.9 5.4 4.3 5.3 TOTAL 100.0 100.0 100.0 100.0 100.0 100.0 H/C (atomic) 0.61 0.48 0.66 0.68 0.69 0.68 0/C 0.030 0.133 0.060 0.046 0.036 0.04! 0/H F Cdmmf W t ' % t 0.049 76.4 0.279 65.5 a 0.090 72.1 0.068 70.7 0.053 71 .1 0.060 71.0 aThe r e l i a b i l i t y of these data i s uncertain due to problems associated with determining the v o l a t i l e matter content of severely weathered coals. (ASTM D 3172-73) (214). bMMd = 1.08Ad + 0.55Sd FCdmmf = ( F Cd " °- 5 5 Sd) ' "100/(100 - MMd) Parr formulae - ASTM D 388-66 (1972) (214) - 119 -Fig. 6.3 SEM images (40X) of the -210 + 149 ym size fraction of the coal samples used in the f l o a t a b i l i t y experiments i l l u s t r a t i n g increased f r i a b i l i t y and slimes content of 0' sample - 120 -6.3.5 Infrared spectroscopy The -210 + 149 ym size fraction of each coal sample was examined by infrared spectroscopy using the KBr p e l l e t technique (286) following standardized procedures developed in the Department of Mining and Mineral Process Engineering (see Appendix E). Transmission spectra for the s i x coal samples are shown in Fig. 6.4, including the baselines used to deter-2 mine the absorption for the designated bands. The resulting data are l i s t e d in Appendix E, Table E . l . Most have been obtained d i r e c t l y from linear absorption spectra, but some were derived from transmission spectra. Table E.2 l i s t s the most generally accepted organic structural assignments given to the absorption bands. Fig. 6.5 i l l u s t r a t e s the relationship between depth into the adit and absorption at bands 2920, 2860, 1700 and 1600 cm - 1, and also includes data from the No. 5 seam sample for comparison. The plotted data for the -210 + 149 ym size fractions are the means from duplicate p e l l e t s . They are compared with data taken from spectra discussed in reference 284 for the For a complex substance such as coal, each recorded absorption band (peak or shoulder) i s usually the composite result of the contribution from two or more types of structural vibrations whose individual infrared bands overlap. Therefore data obtained by the simple baseline subtraction method w i l l not represent the absolute absorption of a single component. One way to obtain estimates of the individual contributions i s to decide which components absorb in the v i c i n i t y of a given band that are l i k e l y to be present in the sample. Each of these components is assumed to have a band shape which can be mathematically described by a composite Lorentzian-Gaussian function. The peak height of the component bands is adjusted either graphically or numerically u n t i l the composite band gives a best f i t of the experimental sample band shape. The absorption c o n t r i -bution of each component can then be estimated for a given wave number (343,344). - 121 -1 — 1 r to o — * «o ID io — o I I I I l 3500 2500 2000 1500 1000 w a v e n u m b e r , c m " 1 . 6.4 Infrared spectra (% Transmission) from KBr pellets of -210 + 149 pm size fraction of coal samples, including baselines for extracting absorption data, showing changes in 1700, 3430 and 1600 cm - 1 band in t e n s i t i e s - 122 -38 75 IN 150 No. 5 seam DEPTH IN ADIT , ft Fig. 6.5 Infrared absorption data at spe c i f i c bands for the six coal samples (duplicate KBr p e l l e t s , -210 + 149 ym size f r a c t i o n , 1.5 mg coal/300mg) - 123 --595 ym size fraction of the adit samples. The trends are similar for both the narrow size fraction and the whole coal samples. The absorption data for the 2920 cm - 1 and 2860 cm"1 bands in Fig. 6.5 indicate that there is very l i t t l e difference in the amount of a l i p h a t i c CH3 and CH,,, and CH respectively in the deeper adit 23 coal samples (38', 75', 1111 and 150') for the -210 + 149 ym size f r a c t i o n ; however, the amount of CH^ and CH^ decreases while CH increases for the severely weathered 0' sample. The amount of a l l types of a l i p h a t i c hydrogen i s less in the No. 5 seam sample than in the adit samples, which possibly re f l e c t s i t s higher rank and higher proportion of i n e r t i n i t e group maceral (see Table 6.2, next section), but w i l l also be partly due to i t s higher mineral matter content, Table 6.1 (which e f f e c t i v e l y reduces the concen-trat i o n of organic matter in the KBr p e l l e t ) . Absorption in the region of the 1700 cm"1 band i s generally very weak in unoxidized high rank coals, but develops on oxidation becoming a very characteristic band for weathered coals (212,229,286-288). In a recent study using Fourier Transform Infrared spectroscopy (FTIR), Painter et a l . (289) found that the 1690 cm"1 band absorption intensity was inversely related to the swelling properties of samples of a weathered caking coal also taken from an adit in the Crowsnest c o a l f i e l d . The 1600 cm 1 band, although r e f l e c t i n g the composite effect of several different structures, i s also considered to be an indicator of weathering (229). 'All the absorption data presented are for KBr pellets with density 1.5 mg coal/300 mg. - 124 -In this respect i t does follow a similar trend to the 1700 cm"1 band (Fig. 6.5). However, the 1600 cm"1 band has also been found to be affected to a s i g n i f i c a n t degree by adsorbed water; drying the KBr pellets in a dessicator at room temperature or in an oven at 105 - 110°C for several days reduces the absorption intensity (284,290). This i s thought to be the reason for the lower 1600 cm - 1 absorption for the -210 + 149 ym samples t compared with the -595 ym samples; the former were stored over dessicant in a dessicator for several days prior to making the KBr p e l l e t s , whereas the Tatter were not. The absorption at the 1600 cm"1 and 3430 cm"1 bands parallel one another for the f i v e adit samples. The previously mentioned FTIR study on weathered adit coal samples (289) isolated several characteristic absorption bands in the 1800 to 1500 cm 1 region by subtracting the spectra of a r e l a t i v e l y unoxidized sample from the spectra of the more oxidized samples. The 1765 cm - 1 band was assigned to phenyl esters, the 1700 cm 1 to carboxylic acid groups, the 1690 cm 1 to conjugated carbonyl structures (e.g. quinone) and 1585 cm 1 to carboxylate ion. Their r e l a t i v e i n t e n s i t i e s depended on the degree of oxidation. The absorption at bands 3030, 1440, 1375 and 1260 cm"1 show similar trends to the 1700 and 1600 cm 1 bands for the adit samples, but have lower i n t e n s i t i e s for the No. 5 seam sample compared with the adit samples. The f i r s t three bands have been generally assigned to carbon and carbon/ hydrogen structures, and the l a s t to ether* groups. The strong bands at 1095, 1030 and 1005 cm"1 in a l l the samples are very characteristic of the spectra for k a o l i n i t e , (and also the 910 cm 1 band) and show l i t t l e evidence of s i g n i f i c a n t quantities of other clay minerals which often occur in coal such as i l l i t e and montmorillonite - 125 -(289,291). Kaolinite and quartz are generally found to be the dominant mineral species in coals from the East Kootenay (Crowsnest) c o a l f i e l d (292,293). Quartz has characteristic absorption bands at 1080 cm"1 (with a broad shoulder out to about 1200 cm" 1), 799 and 779 cm"1 (294). The spectra in t h i s region in Fig. 6.4 do not indicate the presence of s i g n i -ficant quantities of quartz in any of the samples analysed. The carbonate minerals c a l c i t e , dolomite and s i d e r i t e have characteristic bands at 880 -870, 710 - 740 and 1430 cm"1; the deeper adit samples and the No. 5 seam sample do show some absorption at 870 cm"1, 740 cm"1 and 1440 cm"1, indicat-ing the possible presence of some carbonate minerals. However, several other minerals including k a o l i n i t e , quartz and gypsum, as well as the organic matter, also absorb to some extent in the 650 - 800 cm 1 region. 6.3.6 Petrographic analysis Maceral analyses and mean maximum v i t r i n i t e reflectance measurements in o i l (R max, 546 nm) were performed on representative samples of the -595 urn size fraction of the f i v e adit samples and the No. 5 seam sample, and on the -210 + 149 ym size fraction of the l a t t e r sample and 75', 111' and 150' adit samples in the Department of Geological Sciences, University of B r i t i s h Columbia, by Dr. M. Bustin. A Leitz Orthoplan r e f l e c t i n g l i g h t microscope was used equipped with a MPV2 photomultip!ier for reflectance measurements. The particulate samples were prepared and polished as grain mounts in a cold setting re s i n . The maceral analyses were performed by counting about 500 points per sample, while the RQ max values were deter-mined by measuring the maximum reflectance of 50 grains per sample of the c o l l i n i t e maceral (the main v i t r i n i t e group maceral) using established techniques (215 ,295). The raw data and t h e i r treatment are given in - 126 -Appendix F, and are summarized in Table 6.2. Duplicate analyses on the -595 ym size fraction have been included for the 75', 1111 and 150' samples, and they agree reasonably closely in terms of total v i t r i n i t e and i n e r t i n i t e and semifusinite contents. A l l the samples contained only very small amounts of recognizable exinite group maceral. The adit samples were v i t r i n i t e r i c h , whereas the No. 5 seam sample contained considerably more i n e r t i n i t e group macerals and semifusinite. The maceral analyses for the two different size fractions are s i m i l a r except that the -210 + 149 ym fraction of the 150' sample has a l i t t l e less v i t r i n i t e than i t s corresponding -595 ym f r a c t i o n . Finely disseminated pyrite was observed in a l l samples except the'O' and 38' adit samples, indicating that weathering may have resulted in the oxidation of pyrite. There was other petrographic evidence that the degree of oxidation •increased from the deepest adit samples to the surface (0 1) sample. Some coal particles had a characteristic microstructure consisting of micropores and microfissures s i m i l a r to those previously reported for laboratory and naturally oxidized coals (211,296). The v i t r i n i t e showed evidence of oxidation in a l l the adit samples; in the three deepest samples the evidence of oxidation was limited to the periphery of some of the v i t r i n i t e grains, whereas in the 0' and 38' samples the oxidation rinds had progressed throughout some of the grains. The i n e r t i n i t e and semifusinite macerals showed l i t t l e evidence of oxidation with the exception of the surface sample in which d i s t i n c t oxidation rinds were established. In an attempt to quantify petrographically the extent of oxidation, the number of pervasively oxidized v i t r i n i t e grains per sample was deter-mined for one set of the -595 ym size fraction samples, and i s given in Table 6.2. Table 6.2 Petrographic analyses (vol.%) and mean maximum v i t r i n i t e reflectance measurements, R max %, of particulate coal samples. Sample mineral matter, ( c a l . ) a inert-initeb semi-f u s i n i t e exinite total vi t r i n i t e heavily oxidized v i t r i n i t e total inerts RQ max +stand.dev. -595 + 0 ym 0' 10 5 4 1 80 11 18 0.95 + .08 38" 7 7 3 T 82 4 15 1.06 + .05 75' Adit 23 111 ' 6 6 6 13 8 12 8 8 3 1 1 2 72 77 77 4 1 24 19 20 1.00 1 .03 + .05 + .05 • 6 9 7 1 77 - 19 150' 6 10 4 t r 80 2 18 1.00 + .06 6 9 3 2 80 - 17 No. 5 seam 12 24 8 1 55 - 42 1.30 + .06 -210 + 149 ym. 75" 6 16 5 <1 73 - 25 1.06 + .04 Adit i]•]« 23 150' 6 9 7 1 77 - 19 1 .06 + .06 6 12 6 1 75 - 22 1.12 + .06 No. 5 seam 12 22 9 <1 57 - 41 1.35 + .05 see Appendix G for method of calculation k i n e r t i n i t e = macrinite + micrinite + f u s i n i t e t r - trace - 128 -The RQ max measurements do not show a consistent trend with depth from the surface for the -595 + 0 ym size fraction adit samples, although the lowest RQ max value corresponds to the surface sample. This agrees with one recent set of findings on sim i l a r adit samples from the same c o a l f i e l d (293), but another recent investigation on another set of si m i l a r adit coals has shown a consistent trend between RQ max and proximity to the surface (296), RQ max decreasing as the extent of oxidation increases. The RQ max values for the -210 + 149 ym size fraction of the three deepest samples are s l i g h t l y higher. According to the rank c l a s s i f i c a t i o n system of McCartney and Teich-muller (297) „ (who set the boundary between medium and high v o l a t i l e at RQ max = 1.12 based on European and United States coals) the RQ max values obtained for the 75', 1111 and 150' adit samples (least affected by weathering) of between 1.00 and 1.12 for the two size fractions indicate that the rank of the adit coal is in the high v o l a t i l e bituminous region, just below medium v o l a t i l e bituminous. This rank c l a s s i f i c a t i o n of the adit coal i s in variance with that based on v o l a t i l e matter y i e l d dis-cussed previously. It has been found in many cases that western Canadian Cretaceous-age coals do give lower v o l a t i l e matter yields than European or eastern U.S. coals of the same v i t r i n i t e RQ max. This has been ascribed to the higher i n e r t i n i t e contents generally found in the western Canadian coals (293)(resulting in lower yields of v o l a t i l e matter)/ It i s because v o l a t i l e matter y i e l d (and hence fixed carbon) i s dependent on both coal rank and type (maceral composition) that the tr a d i t i o n a l rank parameters for coals r i c h in i n e r t i n i t e are not indicative of the true rank (297). The v o l a t i l e matter contents of i n e r t i n i t e r i c h coals w i l l give them misleadingly high rank c l a s s i f i c a t i o n . - 129 -However, the adit samples do not contain abnormally high i n e r t i n i t e con-tents, so i t i s possible that the RQ max values for the 75', 111' and 150' samples have been reduced by the effects of weathering. Another rank c l a s s i f i c a t i o n system has been put forward by Ting (298) based on RQ max and coal type. In t h i s system the adit samples would just be c l a s s i f i e d as medium rank bituminous, type A, meaning that they have RQ max values between 1.0 and 1.5 and v i t r i n i t e contents between 75 and 100 v o l . % (on a mineral matter free ba s i s ) , i.e. v i t r i n i t e r i c h . The RQ max values for the No. 5 seam sample, Table 6.2, place i t roughly in the centre of the medium v o l a t i l e bituminous region according to the McCartney-Teichmuller rank c l a s s i f i c a t i o n system. 6.4 Non-Coal Materials 6.4.1 Description and preparation 6.4.1.1 polytetrafluoreothylene (PTFE) PTFE is a line a r polymer, (~^2~^2~^n' ^ r e e ^ r o m a n ^ s i ^ i f i c a n t branching. The average molecular weights of commercial polymers are in the range 400,000 to 9,000,000, and can have a high percent c r y s t a l l i n i t y depending on the method of manufacture and the molecular weight. It i s a tough, f l e x i b l e material of moderate t e n s i l e strength remaining d u c t i l e in compression at temperatures as low as 4°K. It is almost completely insoluble in a l l organic solvents and i s chemically inert at room temp-erature; the s p e c i f i c gravity of pure PTFE ranges from 2.1 to 2.3 depending on i t s crystal 1 i n i t y and method of manufacture (299,300). Experimental work was performed on a sample of PTFE supplied by Cadillac P l a s t i c (Canada) Ltd., Vancouver, B.C., in the form of 13 mm (0.5 inch) thick commercial sheet X1100, manufactured by TFE Industries, Kalamazoo, Michigan, USA, from Teflon (tradename) resin obtained from the - 130 -Dupont Corporation. The sheet had a quoted density of 2.13 to 2.20 g/cm , with no f i l l e r material. Small cubic blocks of about 5 mm side were cut from the sheet; 20-30 g at a time of these blocks were placed in a "Mini-Blend Container" with about 100 ml of d i s t i l l e d water and ground in an "Osterizer" household blender (stainless steel blades) at the maximum speed setting for about 8 minutes. The ground product was vacuum f i l t e r e d , oven dried at ^ 105°C and dry screened on stainless steel Tyler sieves into several size fractions between 595 ym and 149 ym. The + 595 ym material was recycled to the 5 blender for further grinding. The screened size fractions were washed on the screens with laboratory grade methanol to remove as much undersize material as possible. They were then oven dried and again dry screened repeatedly u n t i l the amount of material passing the 149 ym sieve became negligible. Each size fraction was agitated in methanol and the supernatant decanted u n t i l no v i s i b l e slimes could be observed. The fractions were then rinsed several times in Considerable d i f f i c u l t i e s were encountered in transferring the fine dry size fractions of PTFE (and other hydrophobic materials.such as coal, MqSo and sulphur) from the sieves to containers etc., by tr a d i t i o n a l brushing. A suction system was therefore devised which u t i l i z e d a small vacuum pump. Although slower than brushing (under normal conditions) i t f a c i l i t a t e d complete transferral of the fine p a r t i c l e s , including "dust", from one container to another, and largely eliminated the dispersal of contaminat-ing dust around the laboratory. 'in the dry state very fine PTFE particles adhered tenaciously to larger particles and no amount of screening could remove them. In water large compact p a r t i c l e - a i r bubble aggregates formed which could not be broken down by washing on sieves. The low surface tension of methanol prevented the formation of these aggregates and dispersed the fine p a r t i c l e s , allowing them to be separated from the coarse p a r t i c l e s . - 131 -d i s t i l l e d water, soaked in 10% HC1 solution, rinsed again in d i s t i l l e d water, f i l t e r e d , oven dried and stored in sealed glass containers. A zeta potential of -47 mV in double d i s t i l l e d water for PTFE chips prepared in a similar manner has been reported (using the streaming potential method) (301). 6.4.1.2 sulphur Elemental sulphur (S, AW = 32.06) occurs in several a l l o t r o p i c forms. Orthorhombic is the stable c r y s t a l l i n e form below 95.4°.C at atmospheric pressure. I t consists of ring-shaped molecules containing eight sulphur atoms in a crown configuration which are held together by van der Waals forces ( i . e . molecular c r y s t a l s ) . I t has conchoidal to uneven fracture 3 along several cleavage planes and a density of 2.07 g/cm . Amorphous sulphur results when the ring molecules of sulphur break and successive ones l i n k together to form long chain molecules. It i s not stable below 160°C, but the change to the c r y s t a l l i n e forms of sulphur is extremely slow at room temperature. A l l allotropes are essentially insoluble in alcohols, and i t s c r y s t a l l i n e forms are s l i g h t l y soluble in acetone, benzene and CCI^ at room temperature, and highly soluble in CS^ (302-304). Experiments were performed on a sample of Technical Grade r o l l s u l -phur obtained from Fisher S c i e n t i f i c Co.. The quoted specifications were 99.8% S, of which 97% was orthorhombic and the remaining insoluble in CSg (amorphous). This was checked by dissolving a pulverized sample of the r o l l sulphur in CS^ following the procedure given in Appendix H, section H.2, but omitting the step of heating the sample in the oven, and dis-solving i t at room temperature instead of in b o i l i n g CS^ (to avoid trans-forming the unstable amorphous sulphur into the stable c r y s t a l l i n e form). - 132 -Lumps of the r o l l sulphur were ground by hand in an agate mortar, and the comminuted material dry screened into several size fractions between 595 ym and 149 ym. These size fractions were then deslimed, cleaned, and stored under argon in the same manner as the PTFE, with the exception that,instead of drying in an oven, the sulphur was dried in a dessdcator over dessicant (to minimize oxidation). The ready native f l o a t a b i l i t y of sulphur has been reported on many occasions (192,204,305). However, at least one group of investigators (305) postulated that this observed f l o a t a b i l i t y was due to hydrocarbon contamination because they could not measure f i n i t e contact angles on the polished surfaces of highly purified sulphur samples. Gaudin et al. (192) considered that these contradictory observations were due to contamination of the soft sulphur surface by grinding media, and that pure sulphur i s indeed inherently hydrophobic and floatable. Olsen et al* (115) found no effect of atmospheric surface oxidation on the measured contact angles on cleaved sulphur surfaces, or any differences between the various faces of orthorhombic c r y s t a l s . As already, discussed in section 3.2.2 the reported relationships between zeta potential and pH do d i f f e r somewhat, although they have always been negative over the pH ranges investigated (196,197,206). Flotation response generally appears to be independent of pH (21,197), 100% recovery in d i s t i l l e d water in less than one minute in a Partridge-Smith "frothless" small-scale c e l l being reported by Finkelstein et al. (204) from pH 2 to 12. 6.4.1.3 molybdenite Molybdenite (MoS2, MW 160.07) i s a mineral which has a layered structure with each Mo atom being surrounded by s i x S atoms at the corners - 133 -of a right trigonal prism forming an i n f i n i t e S-Mo-S layer. The layers are held together by van der Waals bonds which can be readily broken and result in cleavage between the sheets along the (0001) plane with irregular fracture across them. MoS^ c r y s t a l l i z e s in the hexagonal system as short tubular prisms, scattered scales and f o l i a t e d masses. It i s opaque, lusterous, lead gray in colour with a measured density between 4.62 and 5.06, and i s soluble only in strongly oxidizing acids such as aqua regia and boil i n g concentrated HgSO^'(302,304). Most sulphides, including molybdenite, are thermodynamically capable of undergoing aqueous oxidation in conditions similar to those encountered in f l o t a t i o n (119,306,307). Because oxides and hydroxides generally exhibit negligible native f l o a t a b i l i t y , the presence of any such oxidation products on the surface of molybdenite would be expected to diminish i t s f l o a t a b i l i t y . However, under normal f l o t a t i o n conditions molybdenum i s considered to enter solution as an thiomolybdate anion (MoS 07 ,x = 0 to 4) X T" ~ X and not to form s o l i d products. On dry oxidation MoSg forms oxides such as M o 0 2 » MoOg and Mo0£ MoOg 5 , 3 i s soluble in water (forming molybdenum blue complexes), MoO^  i s soluble in a l k a l i e s , but M0O2 i s insoluble under most conditions (119,302). Chander and Fuerstenau (119) found that the oxidation product of low temperature roasting ( <400°C) could be removed by dissolution in water (returning contact angles on cleavage surfaces to what they were before oxidation) while the product of high temperature roasting ( >450°C) required leaching in K0H for the restoration of original contact angles. Arbiter et al (197) found the contact angle on molybdenite cleavage surfaces to be r e l a t i v e l y independent of pH between 1 and 9: i t s magnitude i s very dependent on surface - 134 -preparation. The zeta potential oif fresh unoxidized samples and KOH leached samples i s always negative and generally increases with pH in the mildly acidic region, becoming r e l a t i v e l y constant at neutral and basic pH (118,119,198); i t s magnitude is p a r t i c l e size dependent (anisotropy). In one case f l o t a t i o n recovery in a modified Hallimond Tube was found to be independent of pH (118), while in another a qualitative correlation was found with zeta potential (119) ( i . e . the higher the zeta potential the lower the recovery). Experiments were performed on a sample of massive molybdenite of uncertain o r i g i n available in the Department of Mining and Mineral Process Engineering. Small batches of this lump material were dry ground in the Osterizer blender for about 30 seconds at the maximum speed setting without any noticeable heat build-up. The ground material was then dry screened into several size fractions between 297 ym and 74 ym. These were deslimed, cleaned, dried and stored in the same manner as described for sulphur, with the added step of leaching in 0.1M KOH followed by rinsing in dis-t i l l e d water to remove any a l k a l i soluble surface oxide (119). 6.4.1.4 quartzite A l o c a l l y available sample of quartzite was crushed, dry screened i n -to several size fractions between 595 ym and 149 ym, and these size frac-tions washed on the screens with d i s t i l l e d water and oven dried. They were then rescreened dry u n t i l the amount of material passing 149 ym be-came negligible. The -210 + 149 ym fraction was then repeatedly decanted Repeated agitation and decantation in methanol substantially reduced the amount of fine slimes, but a small quantity continued to persist after many cycles. - 135 -with d i s t i l l e d water u n t i l a clear supernatant was obtained, soaked in 10% HC1, thoroughly rinsed with d i s t i l l e d water, oven dried and stored in a glass container. 6.4.2 Characterization The composition of three molybdenite size fractions and the -210 + 149 pm size fraction of quartzite are given in Table 6.3. The Mo analyses by atomic absorption spectroscopy (AAS) were carried out by Canadian Microanalytical Service Ltd., Vancouver. The other elements were deter-mined in the Department of Mining and Mineral Process Engineering using AAS, while SiO^ was analysed by a standard gravimetric technique. The densities of the various size fractions of the PTFE, sulphur, molybdenite and quartzite were determined with an a i r comparison pycnometer (Beckman, model 930) and are also l i s t e d in Table 6.3. The scanning electron microscope images are shown in Fig. 6.6 to i l l u s t r a t e the morphology of the particles of the different s o l i d s . Table 6.3 Properties of non-coal materials tested in f l o a t a b i l i t y experiments. material molybdenite quartzite PTFE sulphur nominal part i c l e size range, ym -210+149 -149+105 -105+74 -210+149 -210+149 -210+149 meas. density g/cm 4.76 4.78 4.79 2.65 2.18 2.04 Mo 55.9 57.5 57.7 MoS2 93.3 95.9 96.2 chemical SiCL analysis F J 4.9 2.1 4.4 •2.4 3.1 2.0 98.7 0.5 wt.% CaO MgO <0.1 A 12°3 - 137 -MoS2, -105 + 74 ym quartzite, -210 + 149 ym 6.6 SEM images of samples tested in f l o a t a b i l i t y experiments, 40X - 138 -CHAPTER 7  EXPERIMENTAL DETAILS 7.1 Exploratory Test Tube Experiments 1 g of sample was placed in each of seven clean, dry test tubes and 10 ml of appropriate solution added. The concentrations and surface tensions of the seven solutions, increasing in surface tension from 100% methanol to 100% single d i s t i l l e d water in steps of about 10 dyne/cm, are given in Table 7.1. The test tubes were sealed with polyethylene stoppers, Table 7.1 Surface tension of aqueous methanol solutions used in test tube experiments. Solution no. 1 2 3 4 5 6 7 vol . % methanol 100 70 40 20 8 2 0 surface tension, dyne/cm 23 31 40 50 60 69 72 handshaken for 10 seconds, placed in a test tube rack and allowed to s e t t l e . A clear supernatant was necessary for photographing the re s u l t s . It occurred rapidly for deslimed samples (PTFE, sulphur, molybdenite, quartzite, but took up to 24 hours for un-deslimed coal samples, p a r t i c u l a r l y in the more concentrated alcohol solutions). No s i g n i f i c a n t changes in f l o a t a b i l -i t y behavior were observed during prolonged s e t t l i n g periods except i f the test tubes were mechanically disturbed. The results were photographed and a note made of the s t a b i l i t y of any particle-laden froth or i n t e r f a c i a l layer, and the degree of p a r t i c l e attachment to gas bubbles trapped in the - 139 -the sediment, by gentle tapping of the test tube. The same procedure was followed for 1:1 mixtures of two s o l i d s . 7.2 Small-scale Flotation Tests 7.2.1 Apparatus The small-scale f l o t a t i o n tests (often referred to as micro-flotation) were carried out in an a l l - g l a s s version of the "frothless" laboratory f l o t a t i o n c e l l designedly Partridge and Smith (280) (referred to as the P/S c e l l ) . Fig. 7.1 shows the c e l l in i t s assembled and exploded form plus a schematic i l l u s t r a t i o n of i t s basic operation. The approximate dimensions of the central f l o t a t i o n column ( i n i t s assembled form) were:-minimum diameter, 20 mm; height 165 mm; measured volume (between the f r i t t e d disc and the top l i p ) , 58 cm . The c o l l e c t i o n chamber was about 50 mm in diameter, 60 mm in height and the central column projected about 20 mm into i t . The top and bottom sections were mated with a KPV 29 Duran 50 clear glass j o i n t with a teflon sleeve, and held together by two small springs attached to lugs on both sections. The f l o t a t i o n gas was dispersed through a 30 mm diameter medium porosity Pyrex f r i t t e d glass disc (nominal maximum pore size 10-15 ym). A plexiglass support bridge with a teflon bushing insert f i t t e d on top of the c o l l e c t i o n chamber and acted as a guide for the shaft of the glass s t i r r i n g rod (20 mm diameter impeller blades). The P/S c e l l is normally used to compare the f l o a t a b i l i t i e s of individual pure minerals under s p e c i f i c solution conditions (e.g. pH, c o l l e c t o r addition) without the use of a frother. The gas i s dispersed through the f r i t t e d d i s c , particles and bubbles are mixed by the impeller, and particle-laden bubbles are buoyed up the central column. Normally a conical bubble deflector is positioned at the top of the column to - 140 -(b) st i rrer c o l l e c t i o n c h a m b e r f l oa ted m a t e r i a l 11 f r i t t e d g l a s s d i s c bubble de f l e c to r so lut ion leve l pa th o f f l o a t e d m a t e r i a l (c) f l o t a t i o n co lumn g a s Fig. 7.1 The a l l - g l a s s "frothless" Partridge-Smith c e l l used for small-scale f l o t a t i o n t e s t s ; (a) assembled; (b) exploded; (c) basic operation, after Stewart (336) - 141 deflect them into the c o l l e c t i o n chamber so that they reach the solution surface outside the perimeter of the column. After the bubbles rupture, their p a r t i c l e load i s dropped into the c o l l e c t i o n chamber and does not return down the f l o t a t i o n column. A stable froth i s therefore not required to hold the particles at the interface (hence the term " f r o t h l e s s " ) . Tests were performed on the No. 5 seam coal sample without a bubble deflector [on a suggestion by Yarar (308)],and:with bubble deflectors made from plexiglass (polymethyl methacrylate), stainless steel and glass (the f i r s t two held in position on the shaft of the glass s t i r r e r by neoprene "0" ring inserts, and the l a t t e r by a teflon c o l l a r at i t s apex). 1 The results (presented in Appendix J , Table J . l and discussed in the next chapter) indicated that better ash rejection and higher f l o t a t i o n rates were achieved without a bubble deflector. Based on this evidence, a l l tests on other coal samples and on PTFE, sulphur, molybdenite and quartzite (and th e i r mixtures) were conducted without a bubble deflector. The general arrangement for introducing the solution into the P/S c e l l and supplying the f l o t a t i o n gas i s i l l u s t r a t e d in Fig. 7.2. The p l e x i -glass s t i r r i n g shaft support bridge and the Fisher Stedi Speed Adjustable S t i r r e r were ca r e f u l l y aligned to minimize stress on, and rotation eccentricity of, the glass s t i r r i n g shaft. The f l o t a t i o n solution was delivered from a 250 ml glass dispensing burette (capped to minimize methanol evaporation). It was fed into the P/S c e l l through a glass tube Stationary deflectors were also t r i e d (not rotating with the s t i r r i n g shaft), a glass one attached to the outer walls of the c o l l e c t i o n chamber, and a stainless steel one suspended from a modified version of the plexiglass support bridge. However, stresses caused by imperfect a l i g n -ment resulted in frequent fracture of the glass s t i r r i n g shafts. so l . burette support br idge P/S cell A D ro tameter 3- way contro l s topcock a tmosphere pressure vent •©- -®- -®- "0-I I operat ing posit ions o f 3-way con t ro l stopcock m\ mist trap water sa fe t y bubbler liquid t rap N 2 supp l y Fig. 7.2 Arrangement of P/S fl o t a t i o n c e l l , gas supply system and solution delivery system - 143 -via a hole in the plexiglass support bridge. The f l o t a t i o n gas was grade K nitrogen (medical), delivered at a pressure of 10 p.s.i.g. by Tygon tubing. It passed from the tank through three Fisher-Milligan gas-washing bottles, the f i r s t empty and in reverse orientation to act as a l i q u i d trap, the second containing d i s t i l l e d water to act as a bubbler, and the t h i r d f i l l e d with glass wool to act as a water droplet (mist) trap. The gas flow was controlled by a three-way stopcock after the t h i r d gas-washing b o t t l e ; 3 i t could be either directed through the Dwyer rotameter (20-250 cm /min capacity) to the f l o t a t i o n c e l l (positionA), closed while maintaining pressure on both sides of the stopcock (position B), vented to the atmos-phere through another gas-washing bottle containing d i s t i l l e d water (posi-tion C) while s t i l l maintaining pressure on the f l o t a t i o n c e l l side of the stopcock, or the pressure in the whole system could be released to atmos-phere (position D). The l a t t e r gas-washing bottle isolated the system from the atmosphere while allowing the pressure to be immediately reduced on the f l o t a t i o n c e l l to stop bubble generation at the end of a specified f l o t a t i o n (aeration) time. 7.2.2 Procedure The following procedure for f l o t a t i o n tests in the P/S c e l l i s essentially based on the outline given by Finkelstein et al (204). Before each series of f l o t a t i o n tests the required number of samples were placed in small polyethylene weigh boats in 2 g lots and allowed to equilibrate with the laboratory atmosphere. The gas supply system was purged with N 2 and the control stopcock l e f t in position D (Fig. 7.2), allowing the gas to flow through the system. - 144 -Each time the f l o t a t i o n solution was changed, the burette and glass delivery tube were rinsed with the appropriate solution before being f i l l e d . After each test the c e l l was rinsed with single d i s t i l l e d water; water was passed under pressure through the f r i t t e d disc from the gas i n -l e t side to keep the pores as free as possible of s o l i d p a r t i c l e s . Polyethylene gloves were used to handle the P/S c e l l components. A l l tests were carried out at room temperature, 20 + 3°C, and at the natural pH of the suspension. The following procedure was followed for each f l o t a t i o n test on the coal and non-coal samples, unless specified otherwise. (a) A l l components of the P/S c e l l were assembled and rinsed with the appropriate f l o t a t i o n solution. (b) The assembled c e l l was f i t t e d into the plexiglass support bridge, the glass s t i r r i n g shaft connected to the f l e x i b l e coupling (tygon tube) on the s t i r r i n g mechanism, the alignment and clearance (1 to 2 mm) of the glass impeller above the f r i t t e d disc adjusted, and the top section clamped firmly in position. (c) The bottom section of the P/S c e l l was removed, any excess solution shaken out, and the pre-weighed sample care f u l l y placed on the f r i t t e d disc (taking care not to leave p a r t i c l e s attached to the glass j o i n t surface). (d) The bottom section was reconnected to the top section with two small springs attached to the lugs provided on the two sections. (e) The gas supply tube was connected to the bottom section of the c e l l and the control stopcock placed in position A for about 30 seconds to purge the a i r from below the f r i t t e d disc. It was then placed in - 145 -position B (closed) which maintained pressure in the system but stopped the flow of through the f r i t t e d disc. (f) Solution was delivered from the burette into the c o l l e c t i o n chamber unti l i t overflowed the l i p of the central f l o t a t i o n column and f i l l e d i t to just above the glass j o i n t . (g) The s t i r r e r was set at 1100 rpm (measured with a tachometer) and the slurry conditioned for the appropriate time. This s t i r r e r speed was necessary to break down bubble-particle aggregates at the solution/ vapour interface which otherwise accumulated at lower speeds. (h) One minute before conditioning was complete the s t i r r e r was reduced to the setting required for f l o t a t i o n , and further solution slowly added from the burette (to prevent carry-over of particles into the c o l l e c t i o n chamber) t i l l the level in the c o l l e c t i o n chamber was about 5 mm above the l i p of the central column (the total volume of the solution was approximately 120 ml). ( i ) At the end of the conditioning period the N 2 was introduced to the P/S c e l l by placing the stopcock in position A. The f l o t a t i o n period "The s t i r r e r speed during f l o t a t i o n was set at the minimum necessary to stop sanding of the suspended particles on the f r i t t e d disc. For a l l coal samples (-210 + 149 ym size fraction) this was standardized at 310 rpm; for the denser PTFE, sulphur and quartzite, the -105 + 74 ym molyb-denite, and a l l mixtures, the speed was increased to 370 rpm, while for the -149 + 105 ym and -210 + 149 ym molybdenite i t was further increased to 430 and 500 rpm respectively. At these speeds the suspended particles (with the exception of any fine "slimes") did not r i s e more than half-way up the central column unless they were attached to the solution/vapour interface or any extraneous gas bubbles (which was frequently the case for the more lyophobic solids such as PTFE and sulphur, especially in the more d i l u t e methanol solutions). 4n those tests where a bubble deflector was used, the level of the solution was maintained at about 5 mm below the top of the deflector. - 146 -began when the f i r s t bubbles emerged from the f r i t t e d disc (a few seconds after the stopcock was opened). The rotameter (always main-tained at roughly the required setting) was adjusted to a gas flow rate of 40 ml/min, standardized for a l l tests. This gas flow rate generated a large number of bubbles but did not cause s u f f i c i e n t turbulence to carry-over hydrodynamically otherwise non-floating, non-slimes particles into the c o l l e c t i o n chamber (with the s t i r r e r operating). (j) About 5 seconds before the f l o t a t i o n period expired the plug in the co l l e c t i o n chamber outlet was removed and the contents drained into a pre-weighed floats beaker. This displaced any bubble-particle aggregates from above the central f l o t a t i o n column before the gas flow was stopped by turning the stopcock to position D (which immediately stopped bubble generation at the f r i t t e d d i s c ) . (k) The s t i r r e r was stopped, the gas supply tube disconnected, the top section of the c e l l undamped, the s t i r r i n g shaft detached from i t s f l e x i b l e coupling, and the whole c e l l assembly removed. (1) The bottom section of the c e l l was detached and the non-floated "rejects" material in i t and the central f l o t a t i o n column washed into a pre-weighed beaker with d i s t i l l e d water. Any remaining material in the c o l l e c t i o n chamber was washed into the " f l o a t s " product beaker. 7.3 Treatment of Flotation Products 7.3.1 Drying The fl o a t s and rejects from tests on coal, PTFE, MoS^ and quartzite samples, and the i r mixtures, were dried in an oven at about 105°C, - 147 -cooled and equilibrated with the laboratory atmosphere, and weighed. The products from tests on sulphur and i t s mixtures were vacuum f i l t e r e d and dried at room temperature in a fume hood and dessicator (to forego the p o s s i b i l i t y of sulphur combustion i n the drying oven) before being equilibrated with the atmosphere and weighed. 7.3.2 Analysis (a) coal Ashcontent ( a i r dried basis) was determined by standard ASTM proce-dures (214). (b) mixtures containing PTFE The PTFE content in f l o t a t i o n products from mixtures was determined by a simple "skin f l o t a t i o n " technique u t i l i z i n g 100% methanol and visual inspection with a stereomicroscope. This technique i s based on the selec-t i v e wetting considerations discussed i n Chapter 2. Pure methanol com-pletely wets sulphur, molybdenite and coal but only p a r t i a l l y wets PTFE. By gentle manual agitation and manipulation of small quantities of meth-anol with the pre-dried product mixtures, a l l the PTFE particles were made to attach to the air/methanol interface and were transferred to another container with a suction nozzle attached to a water aspirator. Almost perfect separation was rapidly achieved with the small quantities being dealt with ( v e r i f i e d by microscopic inspection). This technique was also t r i e d on non-PTFE mixtures with aqueous methanol solutions with solution surface tensions between the estimated c r i t i c a l surface tensions of f l o a t a b i l i t y of the individual s o l i d s . However, i t was not as successful as for PTFE. - 148 -(c) mixtures containing sulphur Sulphur/MoS2 mixtures were analysed by a simple sulphur combustion method, while sulphur/coal mixtures were analysed by dissolving the sulphur in CSg. Both techniques are described in Appendix H. (d) Molybdenite/coal mixtures The Mo content in these mixtures was analysed by Canadian Micro-analytical Service Ltd., Vancouver, using atomic absorption spectroscopy. - 149 -CHAPTER 8  RESULTS AND DISCUSSION 8.1 Results of Exploratory Test Tube Experiments on Coal Samples Photographs of the results of the exploratory test tube experiments on the s i x coal samples described in section 7.1 are presented in Fig. 8.1. (The test tube numbers in the photographs correspond to the solutions in Table 7.1). None of the coal samples exhibited any f l o a t a b i l i t y in the 100, 70 and 40 vol . % methanol solutions (1, 2 and 3 respectively). Foams formed : in the l a s t two solutions quickly broke down and did not appear to have any particles attached. No particles could be observed attached to r i s i n g gas bubbles (released from the sediment by gentle tapping). In solution 4 (20 v o l . % methanol), no stable froths were formed for any of the coal samples, but a layer of slimes and a few weakly adhering particles remained at the solution/vapour interface for the No. 5 seam sample and the 75', 111' and 150' adit samples. Only slimes attached to the interface for the 0' and 38' adit samples. A stable froth laden with a s i g n i f i c a n t proportion of particles re-sulted in solutions 5 and 6 (8 and 2 v o l . % methanol respectively) for the No. 5 seam and 75', 1111 and 150' adit samples (the proportion of particles in the froth can be judged from the amount of sediment remaining in the bottom of the test tube). The solution 6 froth contained:more material than the froth of solution 5, and gas bubbles r i s i n g from i t s sediment appeared to be more heavily laden with p a r t i c l e s . The s t a b i l i z a t i o n of froths by fine hydrophobic coal particles in solutions of common alcohols is well known (309). No stable froth was formed with the 38' adit sample - 150 -No. 5 seam Fig. 8.1 Results of exploratory test tube experiments on f l o a t a b i l i t y of adit 23 and No. 5 seam coal samples in aqueous methanol solutions - 151 -in these two solutions, although a layer of particles covered part of the solution/vapour interface and r i s i n g bubbles had a few particles attached to them. The 0' adit sample showed no f l o a t a b i l i t y in these solutions, with only a small layer of slimes adhering to the interface. None of the coal samples formed a stable, particle-laden froth in d i s t i l l e d water (test tube 7). However, the interface meniscus was com-pletely covered with strongly attached particles for the No. 5 seam and 75', 111' and 150' adit samples, and r i s i n g bubbles released from the sediment by gentle tapping were heavily laden with p a r t i c l e s . Some slimes and a few particles were attached to the interface and r i s i n g bubbles for the 38' adit sample. None of the 0' adit sample particles were floatable in d i s t i l l e d water. The test tube results indicate that the minimum c r i t i c a l surface tension of f l o a t a b i l i t y , Y cf> of the -210 + 149 ym size fraction of the No. 5 seam and 75', 111', and 150' adit samples is between 40 and 50 dyne/cm, the surface tensions of the 40 and 20 v o l . % methanol solutions (3 and 4) respectively. The minimum y ^ for the 28' adit sample appears to be between 50 and 60 dyne/cm (20 and 8 v o l . % methanol, solutions 4 and 5 respectively) while that for the 0' sample i s greater than 72 dyne/cm ( d i s t i l l e d water). The presence of s i g n i f i c a n t amounts of very slowly s e t t l i n g slimes i s indicated for the 0' and 38' adit samples. 8.2 Results of Small-scale Flotation Tests on Coal Samples The cumulative % y i e l d (wt. % floated), the measured ash content of the f l o t a t i o n products (f l o a t s and rejects) and the calculated feed ash content data for each f l o t a t i o n test in the P/S c e l l on the -210 + 149 ym - 152 -size fraction of the s i x coal samples are l i s t e d in the tables in Appendix J. Unless otherwise specified, the f l o t a t i o n tests referred to in the following sections were carried out in aqueous methanol solutions without using a bubble deflector. 8.2.1 Conditioning time Conditioning practice reported in the l i t e r a t u r e for coal f l o t a t i o n varies widely, p a r t i c u l a r l y for laboratory testing. Two types of condition-ing can be distinguished; one is the mixing of the coal sample with water prior to reagent addition, while the other i s the conditioning of the fl o t a t i o n pulp after reagent addition. The U.S. Bureau of Mines appears to follow a standard practice of mixing the coal sample in water for 10 to 15 minutes to wet completely the sample prior to reagent addition (255, 310); however, much shorter "wetting" times have also been reported (311, 312), and in many cases none at a l l . One study (224) showed no effect of mixing times between 2 and 10 minutes on y i e l d and ash content. The length of the conditioning period after reagent addition varies depending on the type of reagents added and the properties of the coal sample being tested. When alcohol-type frothers alone are used, only short periods of generally less than 3 to 5 minutes have been found to be necessary (313), prolonged periods either providing no beneficial effect (224) or causing problems such as increased reagent consumption, reduced y i e l d , higher f l o a t product ash content and d i f f i c u l t - t o - h a n d l e f r o t h s (256,270,314). When non-polar reagents are added as collectors a greater degree of conditioning i s usually necessary for the dispersion of the reagent, and prior emulsification i s generally recommended to minimize this conditioning period (236). - 153 -The effect of conditioning time in aqueous methanol solutions on the floats y i e l d i s i l l u s t r a t e d in Fig. 8.2. The y i e l d tends to decrease up to about 5 minutes and then become r e l a t i v e l y constant for the readily floatable high rank No. 5 seam sample. The severely weathered adit sur-face sample (0 1) showed no f l o a t a b i l i t y even at short conditioning times. The results for the 38' adit sample,which the test tube experiments indicated had only very l i t t l e f l o a t a b i l i t y , i l l u s t r a t e that conditioning time had a pronounced effect on the y i e l d for this p a r t i a l l y weathered coal sample, in both d i s t i l l e d water and 2 v o l . % methanol. 8.2.2 Flotation time Fig. 8.3 i l l u s t r a t e s the fa m i l i a r relationship generally observed between f l o t a t i o n (aeration) time and y i e l d (or recovery). In a l l f l o t a -tion tests performed in the P/S c e l l in aqueous methanol solutions on both coal and other inherently hydrophobic materials, y i e l d became essen-t i a l l y constant before the completion of 3 minutes f l o t a t i o n time. In the few tests carried out in single d i s t i l l e d water the i n i t i a l f l o t a t i o n rate was much less than in the aqueous methanol solutions and in most cases particles continued to f l o a t after 3 minutes. The lower f l o t a t i o n rate in d i s t i l l e d water i s attributed to the smaller bubble population generated for the standard gas flow rate, the larger size of the bubbles, thei r greater degree of deformation and coalescence (165,171), and the absence of frother-acting molecules (such as methanol) to f a c i l i t a t e the rapid thinning of the disjoining f i l m between the c o l l i d i n g particles and bubbles (section 2.3.1). 8.2.3 Cumulative y i e l d versus f l o a t a b i l i t y The cumulative y i e l d data are plotted as a function of f l o a t a b i l i t y ( v b l . % methanol) in Figs. 8.4 to 8.8 (curve A) for the six coal samples Fig. 8.2 Effect of conditioning time on y i e l d , 3 mins f l o t a t i o n Fig. 8.3 . Effect of f l o t a t i o n time on y i e l d , No. 5 seam, 15 mins conditioning, 10 v o l . % methanol. - 155 -FLOATABILITY , VOL. % METHANOL Fig. 8.4 Cum. f l o a t a b i l i t y relationships for the No. 5 seam sample (-210 + 149 ym) 100 01 1 1 1 1 i i_ 30 25 20 15 10 5 floatability, vol. % methanol Fig. 8.5 Cum. f l o a t a b i l i t y relationships for the 150' adit 23 sample (-210 + 149 ym), 15 mins con-dit i o n i n g , 3 mins f l o t a t i o n T 1 1 1 1 1 110 30 25 20 15 10 5 0 floatability , vol. % methanol Fig. 8 6 Cum. f l o a t a b i l i t y relationships for the 111' adit 23 sample (-210 + 149 ym), 15 mins conditioning, 3 mins f l o t a t i o n 8.7 Cum. f l o a t a b i l i t y relationships for the 75' adit 23 sample (-210 + 149 ym), 15 mins conditioning, 3 mins f l o t a t i o n Cum. y i e l d vs f l o a t a b i l i t y for the 0' and 38' adit 23 samples (-210 + 149 ym), 3 mins f l o t a t i o n - 158 -(15 minutes conditioning time, 3 minutes f l o t a t i o n time, no bubble deflector). Fig. 8.4 also contains one minute f l o t a t i o n time data (curve A 1 ) , and Fig. 8.8 contains 3 minute conditioning time data for the 38' adit sample. The S-shaped curves have been f i t t e d by eye through the arithmetic means of the replicate data points at each methanol l e v e l . The lines marked B and C in Figs. 8.4 to 8.7 w i l l be discussed in section 8.3.1. The largest number of replicate tests performed was f i v e for the 75' adit sample at 10 v o l . % methanol, resulting in a mean cumulative y i e l d of 60% and a 95% confidence interval of +9.3%. (The method of calculating confidence intervals i s given i n Appendix C, section C.4, and confidence intervals for the coal samples at 10, 12.5 and 15 v o l . % methanol where there were three or more rep l i c a t e data, are l i s t e d in Appendix K, Tables K.3 to K.5). The cumulative y i e l d - f l o a t a b i l i t y curves in Figs. 8.4 - 8.7 show that the No. 5 seam and the 75', 1111 and 150' adit samples are readily floatable and indicate a minimum c r i t i c a l surface tension of f l o a t a b i l i t y , y ^ min, of between 47 and 60 dyne/cm (25 and 20 v o l . % methanol). There i s a small amount of material (about 3%) which remained non-floatable below 5 v o l . % methanol; microscopic examination and ash analysis showed that most of this consisted of almost f u l l y liberated inorganic p a r t i c l e s . (The lower yields obtained in d i s t i l l e d water, Fig. 8.4, have been previously discussed in section 8.2.2). At methanol concentrations above 25 v o l . % a small, f a i r l y constant amount of very fine slimes always reported to the P/S c e l l c o l l e c t i o n chamber. The majority appeared to be hydrodynamically carried up the central f l o t a t i o n column due to f l u i d motion caused by agitation and the r i s i n g gas bubbles. However, some - 159 -were observed to form films on the bubbles and may have been d i r e c t l y attached to the bubble surface, or trapped in a water layer transported by the gas bubbles (315). The 38' adit sample showed no f l o a t a b i l i t y at 15 minutes condition-ing time (except for the small slimes f r a c t i o n ) , but noticeable f l o a t a b i -l i t y at 3 minutes conditioning time. A minimum of about 58 dyne/cm (10 v o l . % methanol) i s indicated in the l a t t e r case. The 0' (surface) adit sample was non-floatable in a l l solutions implying that y ^ for a l l i t s p a r t icles i s greater than about 72 dyne/cm (water). The small-scale f l o t a t i o n test results just described confirm and quantify the behavior i l l u s t r a t e d by the q u a l i t a t i v e test tube experiments, Fig. 8.1. 8.2.4 Cumulative y i e l d versus cumulative ash content The cumulative ash content data of both the floats and the rejects products are plotted against cumulative % y i e l d of fl o a t s and cumulative wt. % of rejects in Figs. 8.9 to 8.12. The data in Fig. 8.9 represent a l l the f l o t a t i o n tests on the No. 5 seam sample included in Fig. 8.4, as well as other tests also carried out with no bubble deflector but at different f l o t a t i o n and conditioning times (as indicated in the legend in Fig. 8.9). The curves through the data points in Figs. 8.9 to 8.12 have been f i t t e d by eye. It can be seen that for yi e l d s above about 20% a l l the data points l i e very close to the f i t t e d curves for the four coal samples. There is a continuous decrease in cumulative ash content of both the floats and rejects as the y i e l d decreases. In this context i t i s interesting to note the relationship between ash content and y i e l d for replicate tests. For example,the four replicate data points for the - 160 -CUM. % ASH, REJECTS CUM. % ASH , FLOATS Fig. 8.9 Flotation washability curves for No. 5 seam sample (-210+149ym) cum. % ash , rejects 2 4 6 8 cum. % ash , floats Fig. 8.10 Flotation washability curves for the 150' adit 23 sample (-210+149ym), 15 mins conditioning, 3 mins f l o t a t i o n cum. % ash , rejects Fig. 8 11 Flotation washability curves for the 111' adit 23 sample (.-210+149ym), 15 mins conditioning, 3 mins f l o t a t i o n - 162 -Fig. 8.12 Flotation washability curves for the 75' adit 23 sample (-210 + 149 ym), 15 min conditioning, 3 min f l o t a t i o n - 163 -No. 5 seam sample plotted between 45 and 60% y i e l d in Figs. 8.4 and 8.9 (15 v o l . % methanol, 3 minute f l o t a t i o n , 15 minute conditioning) show that the f l o a t s and rejects ash content decreases continually with y i e l d . The same trend occurs for most sets of replicate data for other samples and conditions (Appendix J ) . This implies that the different yields and ash contents obtained from ostensibly identical tests were not solely due to random sampling and analytical errors, but also reflected a basic difference in the f l o t a t i o n conditions between the tests. Conditioning and f l o t a t i o n time were closely controlled which suggests that the values of the solution parameters during each test may have differed s i g n i f i c a n t l y from those measured on the stock solutions, for instance surface tension. Surface tension i s sensitive to temperature fluctuations and changes in solution concentration due to methanol evaporation and adsorption and/or surface-active contaminants ar i s i n g from the coal sample. Fig. 8.4 i l l u s t r a t e s that the y i e l d changed from 80 to 20% between 10.8 and 17.4 vo l . % methanol, equivalent to a nominal decrease in surface tension of approximately 6 dyne/cm, Fig. 6.1. A change of 1 dyne/cm in this f l o a t a -b i l i t y region therefore causes a change in y i e l d of about 10%. Thus, a variation in surface tension of a l i t t l e over 1 dyne/cm between the four replicate tests is s u f f i c i e n t to account for the 13% range in y i e l d observed. As cumulative y i e l d decreases below about 20% the ash content of the floats begins to increase rapidly for the f l o t a t i o n tests carried out at 15 minutes conditioning time (dashed extensions D1 of the floats curves in Figs. 8.9 to 8.12). This apparently anomalous behavior i s attributed to the carry-over of a small quantity of high-ash content - 164 -slimes released during prolonged conditioning in the more concentrated methanol solutions. Others have also reported this strong peptizing effect of aqueous alcohol solutions on clay minerals associated with coal (175). The average ash content of the floats product would increasingly r e f l e c t the influence of a small quantity of high ash con-tent slimes as the amount of floatable low-ash content coarse particles decreases. Extension D" in Fig. 8.9, drawn through two data points for 5 minute conditioning time, indicates that the peptizing effect i s minimized at short conditioning times. The slimes o r i g i n a l l y present in the sample, and those released during the short conditioning period, appear to have lower ash contents than the coarse p a r t i c l e s . This behavior i s a rea-sonable expectation from l i b e r a t i o n considerations. However, i t i s unlikely that f u l l y liberated organic particles (0% ash) are possible because of the small quantities of intimately associated, inorganic material derived from the original plant material (210). The rejects curves in Figs. 8.9 to 8.12 r e f l e c t the material balance in terms of ash content, and approach the feed ash content of the samples as y i e l d approaches 0%. The small differences observed between the measured feed ash contents of the 75', 1111 and 150' adit samples, Table 6.1, have been shown i n d i r e c t l y to be s t a t i s t i c a l l y s i g n i f i c a n t . ^ The The calculated feed ash contents obtained from the measured floats and rejects ash contents for a l l the f l o t a t i o n tests carried out on each coal sample are l i s t e d in Appendix J , Tables J . l to J.6. These data have been s t a t i s t i c a l l y analysed in Appendix E. The differences between the population means of the calculated feed ash contents of the 75', 111' and 150' adit samples are s t a t i s t i c a l l y s i g n i f i c a n t at the 95% confidence l e v e l , and the measured feed ash contents on the a i r dried basis are within the 95% confidence intervals. - 165 -r e l a t i v e position of t h e i r respective rejects curves at yields approaching 0% r e f l e c t these differences. 8.3 Discussion of Results of Coal F l o a t a b i l i t y Tests 8.3.1 Flotation washability The results in Figs. 8.2 and 8.3 i l l u s t r a t e that for the readily floatable coal samples, cumulative y i e l d becomes r e l a t i v e l y constant after prolonged periods of conditioning and f l o t a t i o n i n aqueous methanol solutions. The cumulative y i e l d - f l o a t a b i l i t y curves in Figs. 8.4 to 8.7 for 15 minutes conditioning and 3 minutes f l o t a t i o n therefore represent a close approximation to a l i m i t i n g f l o t a t i o n characteristic for the -210 + 149 ym size fraction of the coal samples in the f l o t a t i o n system under consideration. The close f i t of a l l the floats and rejects data to single curves in Fig. 8.9 (above about 20% cumulative y i e l d for the floats) indicates that these cumulative yield-cumulative ash content relationships are not dependent on the physical f l o t a t i o n conditions within the ranges studied and also closely approximate a 1imiting f l o t a t i o n characteristic of the sample; that i s , for "frothless" f l o t a t i o n in aqueous methanol solutions the ash content of the flo a t s and reject products i s a function of y i e l d only. Additional evidence which supports the conclusion that the curves in Fig. 8.9 represent a l i m i t i n g f l o t a t i o n c h a r a c t e r i s t i c i s given in Fig. 8.13, which contains data from f l o t a t i o n tests in the P/S c e l l per-formed with three different types of bubble deflector on the No. 5 seam sample (see section 7.2.1 for a description of the deflectors, and Appendix J , Table J . l for the test conditions and r e s u l t s ) . The s o l i d l i n e s are the cumulative yield-cumulative ash curves for tests with no bubble deflector taken from Fig. 8.9. In almost a l l cases, when a bubble CUM. % ASH , REJECTS CUM. % ASH , FLOATS Fig. .8.13 Flotation washability data from tests using different bubble deflectors in P/S c e l l , compared with curves obtained with no bubble deflector, No. 5 seam FLOATABILITY NUMBER , FN Fig. 8.14 Calculated incremental y i e l d and ash content of. f l o a t s as a function of f l o a t a b i l i t y number, No. 5 seam sample - 167 -deflector was used, the floats ash content was higher and the rejects ash content lower than those obtained with no bubble deflector at the same cumulative % y i e l d . The tests performed without a bubble deflector therefore gave superior ash rejection. The reason(s) for t h i s difference in f l o t a t i o n e f f i c i e n c y i s obscure. I t may be that some cleaning action takes place at the solution/vapour interface above the central f l o t a t i o n column when no bubble deflector i s used. J o s t l i n g and collapse of bubble-particle aggregates may cause the release of mechanically entrapped or weakly adhering high ash content particles which then drop back down the column. The presence of a bubble deflector would force a l l aggregates away from above the central column before similar cleaning action could take place. The cumulative yield-cumulative ash curves in Figs. 8.9 to 8.12 also appear to represent l i m i t i n g f l o t a t i o n characteristics of the -210 + 149 ym size f r a c t i o n of the coal samples. These curves and those i n Figs. 8.4 to 8.7 are analogous to the well known family of s p e c i f i c gravity washabi-1ity curves developed from sin k - f l o a t tests in liquids of different density on coarse-size fractions of coal (316). They can therefore be considered as f l o t a t i o n washability curves. The cumulative y i e l d - f l o a t a b i l i t y (vol. % methanol) curves (A) correspond to cumulative y i e l d - s p e c i f i c gravity (densimetric) washability curves, while the flo a t s and rejects cumulative yield-cumulative ash curves (D and E.) correspond to "clean coal" and "discard" s p e c i f i c gravity curves. It i s believed that these f l o t a t i o n washability curves represent close approximations to the fundamental f l o t a t i o n characteristics of the coal samples for several reasons: (a) the f l o t a t i o n test conditions were such as to minimize the hydro-dynamic carry-over of particles (except very fine slimes) in the - 168 -col l e c t i o n chamber of the P/S c e l l ( i . e . only particles attached to bubbles were collected). (b) a narrow p a r t i c l e size range was used to minimize the effect of pa r t i c l e size on bubble-particle adhesion and aggregate s t a b i l i t y (see section 2.3.2), so that surface properties could have the dominant effect on f l o a t a b i l i t y . (c) the lyophobicity of the coal particles was not enhanced by the addi-tion of conventional coal c o l l e c t i n g and/or frothing agents. (d) f l o a t a b i l i t y was controlled solely by changing the methanol concen-tr a t i o n (and therefore surface tension) of the f l o t a t i o n solution. (e) the yield-ash content relationship does not depend on f l o t a t i o n kinetics ( i . e . i t i s independent of f l o t a t i o n time). The y i e l d versus ash f l o t a t i o n washability curves for a particular coal sample can be used to estimate the maximum cumulative y i e l d of floats and the maximum cumulative % ash in the rejects that can possibly be achieved by f l o t a t i o n to obtain a desired clean coal ash content. The only other methods known to be available for providing estimates of f l o t a t i o n washability yield-ash characteristics of fi n e coal samples are the standard sp e c i f i c gravity sink-float tests and the Release Analys technique developed by Dell (244). The former method is of only limited value in predicting the coal cleaning potential of froth f l o t a t i o n becaus the d i s t r i b u t i o n of density and surface properties amongst the particles in a sample are not necessarily p a r a l l e l (310); sink-float tests are also d i f f i c u l t to perform on fine coal. Dell's Release Analysis technique, and i t s l a t e r modifications (310, 317), b a s i c a l l y consists of fl o a t i n g a 200 to 300 g sample in a standard mechanical laboratory f l o t a t i o n c e l l , the froth being collected in . - 169 -increments over several time intervals. The increments are recleaned two or more times to enhance the sharpness of separation and to remove clay slimes. As the i n i t i a l froth c o l l e c t i o n proceeds, the c o l l e c t o r and frother reagent additions, aeration rates and c e l l agitation are adjusted so that the most readily floatable material i s collected f i r s t , and the most d i f f i c u l t - t o - f l o a t material i s collected l a s t . The main objections to t h i s technique appear to be: (a) the f l o a t s washability curve applies only to the particular set of test conditions (224) > (b) the levels of reagent additions are adjusted to suit each coal, (c) the basic surface properties of the particles may be masked by the c o l l e c t o r and frother reagents (which are used in s i i g h t excess)» (d) the technique depends on f l o t a t i o n kinetics as well as the.surface properties of the coal p a r t i c l e s , (e) only one combined t a i l i s available for independent ash analysis ( i . e . no cumulative t a i l s curve i s d i r e c t l y produced). However, Dell (244) claims that i f certain precautions are taken, the curve is independent of reagent type and quantity, coal type and quantity, and operator. In comparison, the proposed technique of producing f l o t a t i o n wash-a b i l i t y curves by frothless f l o t a t i o n tests in aqueous methanol solutions appears to have some advantages (in addition to the points emphasized e a r l i e r in t h i s section): (a) the tests are standardized and r e l a t i v e l y simple to perform, (b) the sample size required is small, allowing adequate quantities of narrow size fractions to be readily prepared. The determination of the f l o t a t i o n washability characteristics of several size - 170 -fractions would provide.additional information to coal preparation engineers with which to design, modify and operate coal f l o t a t i o n c i r c u i t s to optimize performance, (c) an independent yield-ash curve i s obtained d i r e c t l y for the reject product,. (d) a fresh sample i s used for each test ( i . e . there i s no r e - f l o t a t i o n ) . There are, however, some foreseeable limitations to the technique. It may not be applicable in i t s present form for d i f f i c u l t - t o - f l o a t coals (e.g. low rank or heavily oxidized) which only give low yi e l d s even at very low methanol concentrations and short conditioning times. Under these circumstances longer chain alcohols may be required to enhance the coal f l o a t a b i l i t y (175) as well as reduce the solution surface tension. It may even be possible to pre-condition the d i f f i c u l t - t o - f l o a t sample with a non-polar c o l l e c t o r before carrying out the f l o t a t i o n tests in aqueous alcohol solutions. Also, the method may not be applicable to very fine coal p a r t i c l e sizes where i t may be d i f f i c u l t to stop s i g n i f i -cant hydrodynamic carry-over of the particles in the f l o t a t i o n system described. The y i e l d - f l o a t a b i l i t y curves in Figs. 8.4 to 8.7 and the yield-ash curves in Figs. 8.9 to 8.12 allow the relationship between cumulative ash content and f l o a t a b i l i t y of the f l o t a t i o n products to be derived for the readily floatable coal samples. The resulting f l o a t s and rejects lines in Figs. 8.4 to 8.7 (B and C respectively) are based on data derived at intervals of 2 vol . % methanol (columns c and d in the tables in Appendix L)'. They make i t clear that the f l o a t a b i l i t y (represented by methanol concentration) of a coal p a r t i c l e increases as ash content decreases. Extension B' emphasizes the adverse effect of high-ash content slimes - 171 -being released during prolonged conditioning periods. However, i f con-ditioning i s short (extension B") the inverse relationship between flo a t s cumulative ash and f l o a t a b i l i t y continues over the whole range. 8.3.2 Distribution of f l o a t a b i l i t y within a coal sample . The cumulative y i e l d curves in Figs. 8.4 to 8.7 imply that each coal sample can be divided into several fractions of different "average" f l o a t a b i l i t y in a manner similar to the d i v i s i o n of coarse coal samples into a number of density fractions. This has been done for the 3 minute f l o t a t i o n time data (curve A) by dividing the abscissa into intervals of 2 vo l . % methanol and determining the fraction (wt. %) of material which floated in each i n t e r v a l . The incremental ash content of each fraction has also been calculated and the resulting data are given in the tables in Appendix L. The "average" f l o a t a b i l i t y of each fraction can be characterized by a " f l o a t a b i l i t y number", FN, which i s defined as the arithmetic mean of the two boundary vo l . % methanol values. For example, in Fig. 8.4, the cumulative y i e l d at 16 vo l . % methanol i s 39.0% (to nearest 0.5%); reducing the methanol concentration to 14 v o l . % increases the y i e l d to 59.5%; therefore, the proportion of the coal sample which floa t s at 14 vo l . % but not at 16.vol. % i s 20.5%, the incremental y i e l d ; i t i s assigned a f l o a t a b i l i t y number of 15 (FN 15). The higher the f l o a t a b i l i t y number of a f r a c t i o n , the more floatable i s the material in i t . Each f l o a t a b i l i t y fraction can be considered to have i t s own c r i t i c a l surface tension of f l o a t a b i l i t y , Y cf> which i s the surface tension of the solution forming the upper boundary of the f l o a t a b i l i t y i n t e r v a l . For example, the y c f of the fraction in FN 15 i s equivalent to the nominal - 172 -surface tension of the 16 vo l . % methanol solution, approximately 52.5 dyne/cm (Fig. 6.1).,, The incremental y i e l d data for the readily floatable coal samples are presented as histograms in Figs. 8.14 and 8.15. The non-floatable fraction (mainly 1iberated mineral matter) and the slimes fraction are not included. These histograms suggest that by careful manipulation of the surface tension of the f l o t a t i o n solution i t should be feasible to separate fine coal into several fractions of different technological properties. The lines in Figs. 8.14 and 8.15 i l l u s t r a t e the relationship between f l o a t a b i l i t y and incremental ash content. ( I f the cumulative y i e l d -f l o a t a b i l i t y and cumulative yield-cumulative ash curves in Figs. 8.4 to 8.7 and Figs. 8.9 - 8.12 were f i t t e d by continuous mathematical functions i t would be possible to produce a continuous relationship between incre-mental ash content and f l o a t a b i l i t y . However, in l i e u of t h i s , the seg-mented relationship provides an adequate approximation). The incremental ash lines can be roughly divided into two main regions; at low FN incre-mental ash content and f l o a t a b i l i t y are inversely related; at high FN they are independent (the incremental ash content reaches a minimum and becomes r e l a t i v e l y constant). Although the tra n s i t i o n between the two regions i s not sharp, the boundary occurs at an incremental ash content of about 5% for the 75', 111' and 150' adit samples (corresponding to fractions FN 11, FN 15 and FN 13 respectively) and about 8% for the No. 5 seam sample at FN 17. The proportion of sample reporting to the high FN If the cumulative y i e l d versus f l o a t a b i l i t y curve for a coal sample was a v e r t i c a l l i n e , a l l the particles within the coal sample would have the same f l o a t a b i l i t y , and therefore the same y f . INCREMENTAL % YIELD , FLOATS INCREMENTAL % ASH , FLOATS Fig. 8.15 - 174 -region cannot be precisely calculated, but the histograms i l l u s t r a t e that i t i s quite substantial, p a r t i c u l a r l y for the a d i t samples (between 30 and 60 wt. % ) . Extensions F' and F" of the incremental ash l i n e in Fig. 8.14 (No. 5 seam sample) indicate that subtracting out the influence of the slimes fraction makes the incremental ash of the non-slimes particles very similar at short and long conditioning times. Consideration w i l l now be given to the factors which may have c o n t r i -buted to the range of f l o a t a b i l i t i e s observed within each readily floatable coal sample, using the results for No. 5 seam for i l l u s t r a t i o n . They w i l l be looked at under two categories; the compositional/surface properties of the p a r t i c l e s , and the physical properties of the particles and condi-tions of the f l o t a t i o n system. 8.3.2.1 compositional/surface properties The incremental ash lines in Figs. 8.14 and 8.15 i l l u s t r a t e that in the low FN region the ash content of the coal particles has a major effect on the i r f l o a t a b i l i t y . It can be concluded that the bulk composi-tion of mineral matter in a p a r t i c l e i s d i r e c t l y related to the "average" w e t t a b i l i t y of the p a r t i c l e surface. In the region where incremental ash content i s low and r e l a t i v e l y independent of f l o a t a b i l i t y , the surface properties of the organic matter become dominant. The main factors which can affect the organic surface are oxidation, maceral composition and rank. The description of the petrographic analyses in section 6.3.6, and the accompanying data in Table 6.2, indicate that there was a variation in the degree of oxidation of v i t r i n i t e p a r t i c l e s within each of the adit 23 samples, as well as between them, a l b e i t small for the deeper - 175 -ones. It i s known that the degree of surface oxidation necessary to adversely affect f l o a t a b i l i t y would be too small to be detected as oxida-tion rinds by petrographic examination. The high FN region for the 75' adit sample covers a wider range of f l o a t a b i l i t y than the two deeper samples, Fig. 8.15. This factor, plus other evidence indicating a greater degree of oxidation of the former sample (FSI, Table 6.1, and amount of heavily oxidized v i t r i n i t e p a r t i c l e s , Table 6.2), suggest that d i f f e r e n t i a l oxidation was one factor which contributed to the range of organic matter lyophobicity amongst the particles with low mineral matter content in the 75' sample, and probably to a lesser extent in the 111' and 150' samples. In comparison to the adit samples, the range of f l o a t a b i l i t y for the No. 5 seam sample in the high FN region i s narrower, and the propor-tion of the sample involved is less. The low FSI value of 3-1/2 (Table 6.1) suggests that i t has undergone considerable weathering; the resulting d i f f e r e n t i a l oxidation of the particles would therefore be expected to contribute to the i r range of f l o a t a b i l i t y . However, none of the sample's other bulk parameters (Table 6.1), infrared spectra (Figs. 6.4 and 6.5), petrographic analysis, v i t r i n i t e reflectance or ready f l o a t a b i l i t y i n d i -cated any s i g n i f i c a n t degree of oxidation ( p a r t i c u l a r l y when compared with the same parameters for the 38' adit sample, which had an FSI of 3). Comparatively low FSI values are a common feature of many unoxidized non-marine coals from the East Kootenay f i e l d (293). The cause of t h e i r reduced swelling properties i s considered to be the large proportion of inert (non-reactive in terms of caking or coking properties) material found in many of them. Pearson (293) has recently produced a contoured plot of FSI as a function of rank (R Q max) and petrographic composition [vol. % inerts = i n e r t i n i t e + 2/3 semifusinite + mineral matter (211, 318, - 176 -319)] based on a large number of unoxidized non-marine samples from Western Canada and A u s t r a l i a , a l l with about the same ash content of 10 +2 wt. % (dry). Based on this c o r r e l a t i o n , the measured RQ max and inerts content of the No. 5 seam sample (1.35% and 41 v o l . % respectively, Table 6.2) predict an FSI between 5 and 6 compared with i t s measured value of 3-1/2 (a coal with the same RQ max, but a more normal inerts content of less than 30 vol. % would have a predicted FSI between 8 and 9). The discrepancy between the predicted and measured FSI values can be explained, at least i n part, by the considerably larger ash content of the No. 5 seam sample, 20 wt. % (dry) (319). However, i t does not rule out the p o s s i b i l i t y of a small degree of otherwise undetected oxidation, although at least one investigation has shown that high rank Fording coals have shown l i t t l e or no reduction in FSI over several months of atmospheric exposure (319,360). No conclusions can be drawn about the effect of maceral composition on p a r t i c l e f l o a t a b i l i t y within a coal sample because i n s u f f i c i e n t material was available in the f l o t a t i o n products to conduct both ash and petrographic analyses, p a r t i c u l a r l y for tests with low y i e l d s . Other investigations have found that in many cases the i n e r t i n i t e rich fractions of coals have higher ash contents than the v i t r i n i t e rich fractions, section 4.3.3. It i s therefore expected that the majority of the i n e r t i n i t e in the coal samples would report to the low FN region, while the low ash content organic matter occurring in the high floatab.ility region would be v i t r i n i t e r i c h . Although rank is the other main compositional factor known to affect coal f l o a t a b i l i t y (section 4.3.1), the normal rank parameters (fixed carbon, c a l o r i f i c value, v i t r i n i t e reflectance) are, by d e f i n i t i o n , - 177 -considered to be bulk properties which only apply to the coal sample as a 3 whole, and not to individual p a r t i c l e s . 8.3.2.2 physical factors and f l o t a t i o n conditions A narrow p a r t i c l e size fraction was used in an attempt to minimize the effect of p a r t i c l e size on the induction time and aggregate s t a b i l i t y , and therefore f l o a t a b i l i t y , as outlined i n section 2.3. In spite of t h i s , a s i g n i f i c a n t range s t i l l existed between the nominal l i m i t i n g sizes of 210 and 149 ym, as well as a small percentage of undersized material con-s i s t i n g mainly of very fine slimes. Owing to the heterogeneous nature of coal the r e l a t i v e densities of the particles would have ranged from about 1.3 g/cirPfor almost f u l l y liberated organic matter to about 2.7 g/cnPfor liberated inorganic s i l i c a t e s (210). For the same surface/wettability properties small, less dense particles would tend to be more floatable and therefore report to the higher FN fractions than would coarse, more dense pa r t i c l e s . Or, conversely, the particles which reported to the same fl o a t a -b i l i t y fraction would not necessarily have the same we t t a b i l i t y lines even though they had approximately the same c r i t i c a l surface tension of f l o a t a -b i l i t y (as i l l u s t r a t e d in Fig. 2.13, section 2.3.3). However, i t is worth noting for coal that both density and we t t a b i l i t y increase with increasing ash content (235); i t follows that their effects on f l o a t a b i l i t y should reinforce each other. ' R 0 max i s derived from a large number of measurements on small v i t r i n i t e segments (the c o l l i n i t e maceral only i s normally specified); the d i s t r i -bution of measured values providesan indication of chemical v a r i a b i l i t y within the v i t r i n i t e of a sample. Whether i t i s s i g n i f i c a n t enough to result in a d i s t r i b u t i o n of v i t r i n i t e surface properties which can be detected by f l o a t a b i l i t y tests remains to be seen. - 178 -Besides surface tension, other solution conditions changed as the methanol concentration was increased from 0 to 25 v o l . %. There was only a small reduction in solution density, p| (about 3%), but the absolute v i s c o s i t y increased by about 61% (320). Greater vi s c o s i t y means lower rates of thinning and receding of the disjoining f i l m between contacting bubbles and p a r t i c l e s , thereby tending to increase induction time and reduce f l o a t a b i l i t y (see section 2.3.1). On the other hand, an increase in v i s c o s i t y , n> acts to reduce turbulence (indicated by the Reynolds P 1ND 2 Number, N D = —'— , where N and D are the impeller speed and diameter Ke n respectively) which would result in greater aggregate s t a b i l i t y and allow larger, more dense particles to f l o a t than in a less viscous solution, assuming the same lyophobicity. The overall effect of vi s c o s i t y on f l o a t a b i l i t y as solution concentration changes is therefore d i f f i c u l t to predict, but i t i s not expected to affect s i g n i f i c a n t l y Yc^r for a given p a r t i c l e . When an alcohol-type frother i s added to a f l o t a t i o n solution the bubble size and coalescence are reduced, and bubble r i g i d i t y and population increase (for a constant gas flow rate) (156,342). Although many d i f f e r -ent theoretical claims have been made about the effect of bubble size on f l o t a t i o n (22,150,152,154,158,159), no general concensus seems to have been reached (321). From the point of view of c o l l i s i o n , most predict that the c o l l i s i o n e f f i c i e n c y w i l l increase with bubble siz e . However, some "frothless" experimental studies have shown just the opposite effect (176,181). In situations where bubble size i s reduced by frother addition i t i s often d i f f i c u l t to extract i t s effect from that of the corresponding increase in bubble population and reduced induction time (see section 2.3.1). The direct effect of bubble size on induction time and hence - 179 -adhesion i s therefore obscure. The bubble-particle aggregate s t a b i l i t y theories predict that a larger bubble enhances s t a b i l i t y (178,179). I t has also been proposed that there i s an optimum particle/bubble size r a t i o for maximum f l o a t a b i l i t y (22,158). In general i t appears that, as long as the bubble diameter i s below about 1 mm, f l o t a t i o n response w i l l depend mainly on the effect of bubble population on c o l l i s i o n frequency and f l o t a t i o n rate, but w i l l not be influenced to any great extent by the effects of bubble size on the pro-b a b i l i t i e s of c o l l i s i o n , adhesion and aggregate s t a b i l i t y (152,170,322). During the small-scale f l o t a t i o n tests i t was observed that the average bubble size decreased considerably from d i s t i l l e d water to 2 vol. % methanol with a corresponding increase in bubble population. A further small change in the same direction occurred from 2 to 5 vol. %, but at higher concentrations between 5 and 25 vol. % (where f l o t a t i o n y i e l d was reduced, Figs. 8.4 - 8.7) the average bubble size remained f a i r l y constant. Similar behaviour has been observed for aqueous solutions of other alcohol frothers (270,323,324). The f l o t a t i o n results have also shown that at 3 minutes f l o t a t i o n time both the y i e l d - f l o a t a b i l i t y and yield-ash washabi-1ity relationships were independent of f l o t a t i o n k i n e t i c s . Therefore, changes in bubble s i z e , r i g i d i t y and population probably had only a small effect on the range of coal p a r t i c l e f l o a t a b i l i t i e s observed between 5 and 25 v o l . % methanol concentration. From the above discussion i t can be concluded that the observed range of f l o a t a b i l i t i e s i l l u s t r a t e d in Figs. 8.14 and 8.15 i s not necessarily due to differences in p a r t i c l e w e t t a b i l i t y alone, but could be caused in part by variations in physical properties of the particles and the f l o t a -tion conditions. - 180 -8.3.3 F l o a t a b i l i t y differences between coal samples The various factors which may have contributed to the range of f l o a t a b i l i t y observed within the coal samples have been discussed in the previous section (8.3.2). The present section w i l l be concerned with the differences in f l o a t a b i l i t y observed between the coal samples. The cumulative y i e l d - f l o a t a b i l i t y curves from Figs. 8.5 to 8.8 for the adit 23 coal samples and the No. 5 seam sample are compared in Fig. 8.16. The plotted points for the 75', 111 1 and 150' samples and the No. 5 seam sample are the mean values from two or more replicate tests. There are obvious large differences in f l o t a t i o n response between the two samples nearest the surface of the adit (0 1 and 38') and the three deepest samples (75', 111', 150'). The difference between the f i r s t two samples becomes apparent at short conditioning times ( l i n e 6, Fig. 8.16 and Fig. 8.2). To determine whether the observed l a t e r a l displacements between the curves for the three deepest samples were s t a t i s t i c a l l y s i g n i f i c a n t , single factor (the coal sample) analysis of variance (ANOV) and Student -Newman-Keuls multiple comparison of means tests (SNK) were conducted on the replicate cumulative y i e l d data at 10, 12.5 and 15 vol. % methanol, at the 95% confidence l e v e l . The data f o r the No. 5 seam sample were also included in the analysis. The s t a t i s t i c a l methods are described i n Appendix C, while the calculations and tests of hypotheses are given in Appendix K. The conclusions from the s t a t i s t i c a l analysis are that the differences between the mean cumulative yields of the three deep adit samples at the three methanol levels are a l l s t a t i s t i c a l l y s i g n i f i c a n t , except between the 111' and 150' samples at 10 vo l . % methanol. There was no difference between the 1111 adit sample and No. 5 seam sample at either 10 vo l . % or 15 vol. % methanol, or between the 150' sample and - 181 -Fig. 8.16 Comparison of cum. f l o a t a b i l i t y curves for adit 23 samples - 182 -the No. 5 seam sample at 10 vo l . % methanol. The discussion in section 8.3.2.2 pointed out that the range of p a r t i c l e f l o a t a b i l i t i e s within a coal sample could be partly due to variations in the physical properties of the particles and the f l o t a t i o n conditions. But in t h i s section the comparisons w i l l be made between tests carried out under ostensibly identical f l o t a t i o n conditions and in the same methanol concentration, as well as on samples with very similar physical properties (in terms of d i s t r i b u t i o n of p a r t i c l e s i z e , density and morphology, Fig. 6.3),, p a r t i c u l a r l y for the three deepest adit samples. The observed f l o a t a b i l i t y differences between the adit samples must therefore be almost en t i r e l y due to differences in t h e i r surface properties, which are in turn a manifestation of the i r compositional characteristics and degree of weathering. 8.3.3.1 mineral matter (ash) characteristics The floats and rejects curves in Figs. 8.17 and 8.18 respectively show that there are some marked differences i n f l o t a t i o n washability between the four samples. Above about 50% y i e l d there i s no discernible difference between the 111' and 150' curves, whereas the 75' curve diverges considerably towards higher ash contents; the f l o t a t i o n washability of the l a t t e r is therefore i n f e r i o r to the f i r s t two samples (and the No. 5 seam sample even more so). For example, i f a clean coal product contain-ing 5% ash is required, the maximum possible y i e l d from the 111' and 150' samples i s about 80%, compared to only about 67% for the 75' sample (and 0% for the No. 5 seam sample). Below 50% y i e l d there i s l i t t l e difference between the three adit samples except that the 75' and 150' appear to have released s l i g h t l y larger quantities of high-ash content slimes during prolonged conditioning, a possible indication of a higher degree of weathering. - 183 -cum. ash floats , wt. % Fig. 8.17 Comparison of f l o a t product flotation, washability curves for 75', 1111 and 150' adit 23 samples and the No. 5 seam sample - 184 -cum. ash rejects , wt. % Fig. 8.18 Comparison of reject product f l o t a t i o n washability curves for the 75', 1111 and 150' adit 23 samples and the No. 5 seam sample. - 185 -A simple method of quantifying the r e l a t i v e washability, W, of the samples i s to determine the ash content of each at which a given cumula-t i v e y i e l d i s achieved. Values of W at 25, 50 and 75% y i e l d are l i s t e d in Table 8.1, as well as the measured feed ash contents of the samples (from Table 6.1). As expected, Fig. 8.19 indicates that there i s a reasonably direct correlation between washability and feed ash content (the calculated li n e a r correlation c o e f f i c i e n t s are shown in brackets; see Appendix C, section C.5 for the method of calc u l a t i o n ) . In terms of y i e l d of clean coal product, the conclusions drawn above about the r e l a t i v e f l o t a t i o n cleaning q u a l i t i e s of the coal samples are v a l i d . But another way of assessing the capacity of a coal to be bene-f i c i a t e d by f l o t a t i o n i s to compare the maximum proportion of the feed ash which can be rejected at a particular y i e l d of clean .coal'.. This w i l l be called the ash rejection potential, R, and i s the r a t i o of the quantity (weight units) of ash in the rejects over the quantity of ash in the feed. Ash rejection potentials have been calculated from the individual f l o t a -tion test data (Appendix J ) . The data are plotted as a function of 4 cumulative y i e l d in Fig. 8.20; the dashed diagonal l i n e represents the l i m i t i n g case for a hypothetical coal sample in which there i s uniform ash d i s t r i b u t i o n and no ash li b e r a t i o n ( i . e . the ash content of a l l particles i s the same irrespective of their f l o a t a b i l i t y ) . The four samples appear to have similar ash rejection potentials over the whole y i e l d range, suggesting that the mineral matter d i s t r i b u t i o n and l i b e r a -tion characteristics of each sample are quite similar ( i . e . at any value of R the maximum difference in y i e l d i s only about 7%). A l l available data are not included for the sake of c l a r i t y , but those that are omitted also l i e on the f i t t e d curves. - 186 -Table 8.1 Comparison of analysis, washability, f l o a t a b i l i t y and ash rejection potential data Adit 23 No. 5 0' 38' 75' 111 ' 150' seam Mineral Meas. Feed Ash 18.4 12.0 11.1 10.1 9.6 19.9 Matter (Dry), wt. % Rank F Cdmmf' w t' % - 70.7 71.1 71.0 76.4 R0 max, 7o - 1.06 1.06 1.12 1.35 Petrographic Inert. + semifus. - _ 21 16 18 31 Composition V i t r i n i t e - 73 77 75 57 (vol . %) Inerts — 25 19 22 41 Weathering Moist.(ad), 15.9 2.1 0.7 0.9 0.9 0.8 Oxidation wt. % FSI 0 3 7 7.5 8 3.5 °dmmf( d i f f>' 1 4 ' 3 6.9 5.4 4.3 5.3 3.6 wt. % 1700 cm"1 abs. .460 .188 .088 .084 .097 .084 H/C (atomic) .475 .664 .682 .688 .682 .611 Washability W75 5.6 4.7 4.7 11.1 (wt. % ash) / 0 W50 - 4.2 3.8 4.2 8.3 W25 - 3.9 3.7 3.7 6.7 Ash R75 • - .610 .640 .610 .580 Rejection R50 - .805 .805 .780 .790 Potential R25 - .910 .910 .900 ,915 F l o a t a b i l i t y F75 - 8.5 12.7 11.2 11.8 (vol. % F50 - 10.9 14.7 13.4 15.1 methanol) F25 - ' 13.4 16.5 15.3 17.0 F0 0 12.0 15.6 18.3 17.2 18.8 - 187 -rr 0.9 S 0.8 *-o Q. .2 0 7 O *£ 0.6 o o 0.5 12 10 8 3 >• 6 .o o XI o 5 R 5 0 { - . I 7 6 5 ) R 7 5 ( - . 8 2 3 7 ) 18 16 14 o c o x: E o > 12 — 10 8 o o o 8 12 16 20 meas. feed ash (dry), wt. % Fig. 8.19 Correlation of measured feed ash content of the 75', 1111 and 150' adit 23 samples and the No. 5 seam sample with thei r f l o t a t i o n washability (W), ash rejection potential (R), and f l o a t a b i l i t y (F) at 25, 50 and 75% cum. y i e l d . - 188 -Fig. 8.20 Flotation ash rejection potential curves for the adit 23 and No. 5 seam samples - 189 -Ash rejection potentials at 25, 50 and 75% y i e l d are given in Table 8.1 for the four samples. Neither the plots in Fig. 8.19 nor the accompanying li n e a r correlation c o e f f i c i e n t s provide any strong evidence of dependence of R on the feed ash content of the samples. The r e l a t i v e positions of the f l o a t a b i l i t y curves in Fig. 8.16 and curve A, Fig. 8.4 have also been quantified as F, the vol>. % methanol at 25, 50 and 75% y i e l d (Table 8.1). . Fig. 8.19 shows that there i s no simple dependence of f l o a t a b i l i t y on feed ash content of the samples. There is also no indication of any relationship between f l o a t a b i l i t y and ash rejec-tion potential (linear correlation c o e f f i c i e n t s of .2032, -.2945 and .3226 at 75, 50 and 25% cumulative y i e l d respectively). It emphasizes that other factors as well as mineral matter content and d i s t r i b u t i o n are important in determining the r e l a t i v e f l o a t a b i l i t y of a coal sample. 8.3.3.2 effect of weathering The marked reduction in f l o a t a b i l i t y of the adit samples from 75' to 38' to 0' corresponds to si m i l a r changes in the values of many of t h e i r bulk parameters which are often used to characterize degree of weathering, such as moisture, FSI, oxygen content, 1700 cm~^  infrared absorption and atomic H/C r a t i o , Table 8.1. The dependency of f l o t a t i o n response on conditioning time for the 38' sample (Fig. 8.2 and 8.16) is no doubt due to a combination of factors. The main time dependent mechanisms which have been put forward to explain s i m i l a r behavior for weathered coals are (239,24:8,250,256,260,270,271 , 325):-(a) the reduction of alcohol frother concentration in the f l o t a t i o n solution due to adsorption at hydrophilic or polar surface s i t e s , resulting in lower f l o t a t i o n rates , - 190 -(b) the formation of hydrated layers on organic matter surfaces due to H^O adsorption at polar sit e s (oxygen containing functional groups) caused by partia l oxidation, (c) the release of hydrophilic slimes (particulate and c o l l o i d a l , organic and inorganic) which adhere to coarse coal p a r t i c l e surfaces rendering them more wettable, (d) the dissolution of water or a l k a l i soluble, multipolar, organic, coal oxidation products which act as f l o t a t i o n depressants. On a quantitative basis the f l o a t a b i l i t y of the surface adit sample (0') i s considered as zero under a l l conditions (Fig. 8.2). Because the f l o a t a b i l i t y of the 38' sample was very dependent on conditioning time, i t is not r e a l l y v a l i d to compare quantitatively i t s f l o a t a b i l i t y at 3 minutes conditioning time (curve 6, Fig. 8.16) with the f l o a t a b i l i t y of the deeper samples at 15 minutes conditioning time. However, i t has been used to provide a rough estimate of the intermediate bulk f l o a t a b i l i t y characteristics of the 38' sample by extrapolating the r e l a t i v e l y l i n e a r portion of the f l o a t a b i l i t y curve to 0% y i e l d . Similar F Q values obtained in the same manner are also l i s t e d in Table 8.1 for the four readily floatable samples. The relationships between the weathering-oxidation bulk parameters and F Q f l o a t a b i l i t y are shown in Fig. 8.21. A reasonable correlation between degree of weathering and f l o a t a b i l i t y i s observed for the 0', 38' and 75' samples. However, for the four readily floatable samples. (75', 111', 150' and No. 5 seam) the only bulk parameter which appears to correlate somewhat with f l o a t a b i l i t y i s the oxygen content. But i t must be pointed out that t h i s correlation may be fortuitous because no Fig. 8.21 Correlation of bulk oxidation parameters with F Q f l o a t a b i l i t y for the adit 23 and No. 5 seam samples - 192 -s t a t i s t i c a l l y s i g n i f i c a n t difference was found between the C, H and N values of the 75', 111' and 150' samples (section 6.3.3.2 and Appendix D); t h i s implies that there is no s i g n i f i c a n t difference between thei r oxygen ( d i f f . ) contents (or t h e i r H/C, 0/C and 0/H r a t i o s , Table 6.1). Also, determining oxygen content by the difference method is not very r e l i a b l e . Moisture content and 1700 cm - 1 absorption for the -210 + 149 ym size fraction samples follow a very similar trend. There is v i r t u a l l y no difference between the four most floatable samples. There is also l i t t l e difference in FSI and atomic H/C r a t i o for the 3 most floatable samples (75', 150' and 111'), but they both decrease s i g n i f i c a n t l y for the No. 5 seam sample (which has the highest f l o a t a b i l i t y ) . At f i r s t glance the low FSI and H/C ra t i o would suggest that the No. 5 seam sample had under-gone a substantial degree of weathering; however, i t has already been pointed out in section 8.3.2.1 that this apparent FSI anomaly can be explained, at least in part, by the large inerts content in the sample, Table 8.1 (the low H/C r a t i o r e f l e c t s the high proportion of i n e r t i n i t e ) . I t emphasizes that FSI can only be used as an indication of degree of weathering (and of consequent f l o t a t i o n response) when coals with similar mineral matter and petrographic composition are compared. The incremental ash lines in Fig. 8.15 (described in section 8.3.2) are compared in Fig. 8.22. The dashed iso-cumulative y i e l d lines show that the t r a n s i t i o n between the two main f l o a t a b i l i t y regions occurs at about 50% cumulative y i e l d and 5% incremental ash content for the three deep adit samples. At low values of FN, the lines for these three samples are displaced r e l a t i v e to one another. Therefore, in a particular f l o a t a b i l i t y f r a c t i o n , particles from the three adit samples have the - 193 -Fig. 8.22 Comparison of incremental ash content - f l o a t a b i l i t y number li n e s for the adit 23 and No. 5 seam samples - 194 -same f l o a t a b i l i t y but different ash contents. I t can therefore be deduced that the lyophobicity of the organic material in the particles must be d i f f e r e n t to compensate for the adverse effect of the different amounts of l y o p h i l i c mineral matter (ash) on f l o a t a b i l i t y ; i.e. the organic material in the 111' sample must be the most lyophobic (least wettable) while that in the 75' sample must be the least lyophobic (most wettable). The small variations observed in maceral composition of the -210 + 149 urn size fractions of the three deep adit samples in terms of i n e r t i n i t e , semifusinite and v i t r i n i t e , Tables 6.2 and 8.1, are not considered to be s u f f i c i e n t to account for the f l o a t a b i l i t y differences observed between the three samples. There is also very l i t t l e difference in the values of the i r rank parameters (FC^ ^ and RQ max, Table 8.1). As discussed e a r l i e r in thi^s section, the small differences in the values of the weathering/oxidation parameters for the 3 deep adit samples are not considered to be s i g n i f i c a n t . However, coal f l o t a t i o n response is well known to be sensitive to small degrees of oxidation which cannot be detected by conventional means (252). Also, degree of weathering is not necessarily d i r e c t l y related to depth into an adit. A recent study has shown that in heavily fractured coal seams in the Rocky Mountain area, local oxidation can occur at any point in a seam due to the v i c i n i t y of f a u l t s (292). I t i s therefore tentatively concluded that the deduced difference in organic matter lyophobicity between the three deepest adit samples i s most l i k e l y caused by small, undetected, differences in degree of surface oxidation. In the region where f l o a t a b i l i t y number is independent of incre-mental ash content for the three deep adit samples (Fig. 8.22), they a l l - 195 -have approximately the same low ash content (which represents the lowest ash content which can be achieved by f l o t a t i o n i f slimes are neglected). It i s therefore deduced that the organic material lyophobicity in a particular f l o a t a b i l i t y fraction in t h i s region must be the same for each sample. But because the incremental y i e l d histograms in Fig. 8.15 are displaced r e l a t i v e to one another along the f l o a t a b i l i t y number ax i s , the amount of material found in a particular fraction varies from sample to sample. For example, the incremental y i e l d in FN 17 i s 4, 7 and 21% respectively for the 75', 150' and 111' samples, a l l with an incremental ash content of about 3%. This i s the effect that would be expected i f the lyophobicity of most of the organic matter i n a sample was reduced by oxidation. As the degree of oxidation progressed, a sample's incremental y i e l d histogram would be shifted to lower f l o a t a b i l i t i e s u n t i l increasing amounts of the particles in the sample became non-floatable even in the most d i l u t e (high surface tension) solutions. Another possible explanation for the different quantities of material reporting to the same f l o a t a b i l i t y f r a c t i o n i s that the mineral matter d i s t r i b u t i o n and l i b e r a t i o n characteristics of the three samples were sub-s t a n t i a l l y d i f f e r e n t . However, the iso-incremental ash lines in Fig. 8.16, and the ash rejection potential curves in Fig. 8.20, indicate that the mineral matter dissemination characteristics are quite similar for the three adit samples. The cumulative amount of material which f l o a t s with less than a particular incremental ash content i s approximately equal for each sample. This i s a reasonable expectation for coal samples from the same adit in a single seam, with a narrow size range, similar feed ash contents, and collected, handled and prepared under the same conditions. - 196 -8.3.3.3 rank and petrographic composition The f l o t a t i o n washability of the No. 5 seam sample is very different from the three deep adit samples (Figs. 8.17 and 8.18), whereas there i s no s t a t i s t i c a l l y s i g n i f i c a n t difference between i t s f l o a t a b i l i t y curve and that of the 111' sample. The comparison of incremental ash lines in Fig. 8.22 shows that for FN > 11 the incremental ash content of the No. 5 seam sample in a particular f l o a t a b i l i t y f r a c tion i s greater than in the adit samples, implying that the lyophobicity of i t s organic material must also be greater to compensate for the deleterious effect of ash content on f l o a t a b i l i t y . Based on the discussion in the previous section, one possible explanation for th i s deduced difference in organic material f l o a t a b i l i t y is that the degree of oxidation of the organic matter in the adit samples was greater than that of the No. 5 seam sample. However, a more l i k e l y alternative i s that the organic matter in the No. 5 seam sample i s of higher rank than that in the adit samples, indicated by the substantially higher values of the FC^ m m^ and R max parameters, Table 8.1. It was pointed out in section 4.3.1 that coal hydrophobicity increases with rank towards a maximum for low v o l a t i l e bituminous coals. At low f l o a t a b i l i t y numbers in Fig. 8.22 the incremental ash lines of the No. 5 seam and 1111 adit samples converge, suggesting that the lyopho-b i c i t y of t h e i r organic material became similar in th i s region. If the maceral composition ( i . e . r e l a t i v e proportions of v i t r i n i t e , i n e r t i n i t e and semifusinite) of the particles from both samples was the same in a particular f l o a t a b i l i t y f r a c t i o n , the incremental ash lines would be expected to remain roughly parallel (as i s the case for the three deep adit samples). But the proportion of i n e r t i n i t e plus semifusinite to v i t r i n i t e in the No. 5 seam sample was about twice as large as in the - 197 -adit samples (Table 8.1). I n e r t i n i t e and semifusinite are usually more intimately associated with mineral matter and t h e i r organic material i s considered to be less lyophobic than in v i t r i n i t e of the same rank (section 4.3.3). Therefore, a concentration of a greater proportion of i n e r t i n i t e in the low f l o a t a b i l i t y number fractions for the No. 5 seam sample is considered to be the reason for the convergence of the incre-mental ash l i n e s . 8.3.4 General The correspondence found between the quantitative results of the small-scale f l o t a t i o n tests and the q u a l i t a t i v e observations obtained with the exploratory test tube experiments indicate that the l a t t e r method may be suitable (with some modifications) as a rapid, simple, inexpensive way of assessing the r e l a t i v e f l o a t a b i l i t y of coal samples in the f i e l d or in the preparation plant. The technique could be made more quantitative for laboratory purposes by adding the additional dimen-sion of measuring wetting rates as in the si m i l a r methods of G l a n v i l l e and Wightman (276) and Guruswamy (274). In the f l o t a t i o n of inherently lyophobic particles in aqueous meth-anol solutions, the methanol is considered to play two main roles; a frother and a wetting agent (section 4.3.6). As a frother, although i t s surface a c t i v i t y i s much lower than that of more conventional frothers, i t decreases the size of bubbles, and, at least at low concentrations, f a c i l i t a t e s the rapid thinning of the di s j o i n i n g f i l m between approaching p a r t i c l e and bubble. The l a t t e r factor results in a decrease in the induction time and therefore enhances p a r t i c l e f l o a t a b i l i t y (section 2.3.1). Reducing bubble size usually means an increase in bubble population ( i f the gas flow rate is constant), resulting in greater f l o t a t i o n rates than - 198 -those obtained in water alone. As a wetting agent, the r e l a t i v e adsorp-tion of methanol at the three interfaces results in the reduction of the coal lyophobicity, manifested by a smaller contact angle and f l o a t a b i l i t y (sections 2.2.3 and 2.2.3.1). Therefore, the two roles of methanol tend to oppose one another. At low concentrations i t s role as a frother i s most important, but at high concentrations i t s wetting a b i l i t y becomes dominant. The results from the small-scale f l o t a t i o n tests in aqueous methanol solutions on the readily floatable coal samples have shown that i t has the potential for being developed into a laboratory technique for providing basic f l o t a t i o n washability data and information about the r e l a t i v e w e t t a b i l i t y and f l o a t a b i l i t y of the organic material in fine coal samples. In addition, the results suggest that i t would be technically feasible to use t h i s method to separate such samples into fractions r i c h in particular group macerals or micro!ithotypes (a process called maceration). The sharpness of the separation would of course depend on the prior l i b e r a t i o n of as much of the mineral matter constituents as possible, and the achieve-ment of a certain degree of l i b e r a t i o n between the main petrographic components. The selective separation of petrographic constituents i s important for both laboratory investigations and industrial applications such as coke making and coal l i q u e f a c t i o n , where feed material r i c h in reactive and hydrogen ri c h components ( v i t r i n i t e and exinite) are an advantage (3,4). Similar schemes have been proposed and developed in the past to provide low ash "ultrapure" coal for the production of electrode carbon (247). They were based on gravity precleaning, grinding, and multistage f l o t a t i o n using starvation quantities of reagents. An example in B r i t i s h Columbia where maceration by f l o t a t i o n may be a t t r a c t i v e is a - 199 -coal deposit in the Bowron River area (Norco Resources Ltd., Vancouver) which i s r i c h in r e s i n i t e . Resinite is one of the exinite group macerals and can have commercial value for the manufacture of p l a s t i c s , varnishes etc. (355). It i s one of the most hydrogen-rich, and therefore floatable, of the coal macerals and could possibly be amenable to selective f l o t a t i o n separation by a method based on c r i t i c a l surface tension of p a r t i c l e f l o a t a b i l i t y . 8.4 The Relative F l o a t a b i l i t y of Some Inherently Hydrophobic Solids 8.4.1 Results of exploratory test tube experiments The results of the exploratory test tube experiments on the PTFE, sulphur, molybdenite and quartzite samples,described in section 6.4,are i l l u s t r a t e d by the photographs in Fig. 8.23 (the test tubes from l e f t to ri g h t correspond to the methanol solutions numbered 1 to 7 in Table 7.1). The -210 + 149 ym PTFE exhibited f l o a t a b i l i t y in a l l the solutions; in pure methanol (far l e f t ) there was only a layer of particles attached to the interface and wetting meniscus; stable froth formed in the second solution from the l e f t (70 vol . % methanol) which contained most of the pa r t i c l e s . In the more d i l u t e solutions, including d i s t i l l e d water (far r i g h t ) , extensive particle/bubble agglomerates were formed which increased in compactness as the solutions became more d i l u t e ; the agglomerates 5 either sank or attached to the interface depending on t h e i r bulk density. PTFE particles also formed compact layers on the interface and wetting meniscus, and small amounts of seemingly dry particles were observed 'Glembotskii et al (221) called these particle/bubble agglomerates or aggregates " a i r - f l o c s " , and postulated that the necessary conditions for t h e i r formation were a high degree of s o l i d lyophobicity, pulp aeration and pulp density, and r e l a t i v e l y l i t t l e pulp agitation. - 200 -PTFE (-210 +149 ym) sulphur (-210 + 149 ym) sulphur (sublimed) molybdenite (-210 + 149 ym) molybdenite (-105 + 74 ym) Fig. 8.23 Results of exploratory test tube f l o a t a b i l i t y experiments on some particulate samples of inherently hydrophobic solids in aqueous methanol solutions - 20T -s i t t i n g on top of the p a r t i c l e coated interface in the three most d i l u t e solutions. Similar behavior was observed for coarser and f i n e r s i z e fractions. These results indicate that the minimum Y f value for a l l the c t PTFE particles in the sample was less than the surface-tension of methanol (about 23 dyne/cm). No f l o a t a b i l i t y of the -210 + 149 urn sulphur sample was observed in solutions 1 and 2; attached particles p a r t i a l l y covered the interface of solution 3 and were attached to r i s i n g bubbles, while stable froths were formed in solutions 4 to 6, with some particle/bubble agglomerates in solution 6. A compact layer of particles formed at the interface in dis-t i l l e d water (solution 7) and trapped large completely particle-coated bubbles beneath i t ; there were many completely coated bubbles also trapped in the sediment. Similar tests were also done on a sample of very fine sublimed sulphur (Fisher laboratory grade, as received) to observe the effect of p a r t i c l e s i z e . Again, no f l o a t a b i l i t y occurred in solutions 1 and 2, although the fine sulphur formed rapidly s e t t l i n g floes which .... accounts for the voluminous sediment. A stable froth was formed in solution 3, while the behavior in solutions 4 to 7 was similar to that for the coarse sulphur except that the froths were more voluminous. A minimum y ^ for sulphur occurring between solutions 2 and 3 i s therefore indicated for both p a r t i c l e s i z e s , that i s , between about 31 and 40 dyne/cm (Table 7.1). For the -210 + 149 ym molybdenite sample there was no p a r t i c l e attach-ment to the interface in solutions 1, 2 or 3, nor any observable bubbles trapped in the dispersed (non-flocculated) sediment. No stable froth formed in solutions 4, 5 or 6, but about a t h i r d of the interface in 4 was - 202 -covered with attached p a r t i c l e s , while i t was completely covered in 5 and 6 with some small particle-laden bubbles trapped underneath and in the sediment. In d i s t i l l e d water about half of the interface was covered with attached p a r t i c l e s . Gentle tapping dislodged most particles from inter-faces of a l l the solutions. A minimum y ^ value between 40 and 50 dyne/cm (solutions 3 and 4) i s indicated for t h i s size f r a c t i o n . A few particles of the f i n e r molybdenite sample (-105 + 74 ym) attached to the interface in solution 3; in solution 4 a few large and small bubbles p a r t i a l l y coated with particles were trapped under the completely p a r t i c l e -coated interface, the large ones being readily broken down by tapping. Most of the particles in solutions 5 and 6 reported to the stable froth which consisted of almost completely coated "armored" bubbles, trapped under a particle-coated interface and wetting meniscus. Small bubbles were trapped in the sediment. In d i s t i l l e d water a few large particle-coated bubbles were trapped under a particle-coated interface and readily broke down on tapping; bubbles r i s i n g from the sediment were completely laden with p a r t i c l e s . A minimum y ^ of about 40 dyne/cm (solution 3) i s indicated for t h i s size f r a c t i o n . The -210 + 149 ym quartzite size fraction was also tested in a similar fashion and showed no detectable f l o a t a b i l i t y in any of the solu-t i o n s , implying that the y ^ of a l l particles in the sample was greater than the surface tension of water, about 72 dyne/cm. 8.4.2 Results of small-scale f l o t a t i o n tests The tests on PTFE, sulphur, molybdenite and quartzite were a l l conducted under the same f l o t a t i o n conditions as the coal samples, with the following exceptions: the conditioning time was set at 5 minutes, and - 203 -the impeller speed was adjusted between 300 and 500 rpm depending on the size and density of the material so as to ju s t keep the particles in suspension without sanding, but ensuring that they were not hydrodynamically transported into the c o l l e c t i o n chamber of the P/S c e l l . The cumulative % floated data are l i s t e d in Appendix M and plotted in Figs. 8.24 and 8.25. The f l o a t a b i l i t y curve for the No. 5 seam coal sample (the data points are mean r e s u l t s , -210 + 149 ym size f r a c t i o n , 15 minutes condition-g ing time, curve A, Fig. 8.4) i s also included for comparison. Substantial differences in f l o a t a b i l i t y behavior are obvious between the various s o l i d samples in Fig. 8.24. The range of solution surface tension over which the % floated drops from near 100% to approximately 0% gives the range of Yc.f values of the particles in each sample, and these are l i s t e d in Table 8.2. Visual observations made during the f l o t a t i o n tests also provided some useful information. When conditioned in solutions of 70 vol . % methanol or l e s s , the PTFE particles formed particle-bubble agglomerates (floes) and particle-coated bubbles which did not break down and contained the majority of the particles present. A l l particles floated as floes as soon as the gas was turned on. In 90 vol. % methanol, most of the particles were dispersed during conditioning but also floated as floes within a few seconds. In 100% methanol the particles were completely dispersed during conditioning, and although the bubbles were much larger than in the binary solutions, a l l particles had floated as p a r t i c l e -coated bubbles within 30 seconds. There was only a small difference in cumulative y i e l d between the tests at 5 and 15 minutes conditioning time for the No. 5 seam coal sample, indicated by data in Fig. 8.2 and Appendix J , Table J . l . floatability, v o l . % methanol (non-linear) 100 90 70 60 50 40 30 20 15 10 5 100 80 60 40 20 0 PTFE •210 + 149 /Am sulphur -210+149 /xm quartzite 210 + 149 /xm 20 30 40 50 60 floatability , solution surface tension dyne/cm 70 8.24 Fl o a t a b i l i t y of samples of inherently hydrophobic solids and hydrophilic quartzite in aqueous methanol solutions in small-scale frothless P/S c e l l , 3 mins f l o t a t i o n - 205 -v o l . % methanol (non-linear) 50 40 30 20 15 10 5 2 floatability, solution surface tension dyne/cm 8.25 Effect of p a r t i c l e size on the f l o a t a b i l i t y of molybdenite, 5 min conditioning, 3 min f l o t a t i o n - 206 -Table 8.2 Range of c r i t i c a l surface tension of f l o a t a b i l i t y for particles in samples of inherently hydrophobic materials -jn aqueous methanol solutions (including hydrophilic quartzite for comparison) material nominal size range y ^ nominal range methanol cone. ym dyne/cm range, v o l . % PTFE -210 + 149 <23 >100 sulphur -210 + 149 33 - 38 63 - 45 molybdenite -105 + 74 40 - 52 4 0 - 1 7 -149 + 105 >43 <33 -210 + 149 >o>55 <^ 13 No. 5 seam coal -210 + 149 48 - 65 23 - 5 quartzite -210 + 149 >72 <0 - 207 -The behavior of the - 210 + 149 ym sulphur was similar to the PTFE, only at much more di l u t e solutions. Up to 25 vo l . % methanol bubble/ p a r t i c l e agglomerates and particle-coated bubbles formed during condition-ing, and a l l particles floated as floes within a few seconds of the gas being turned on. At 40 vo l . % methanol, and above, the sulphur particles remained dispersed but the time taken for a l l the particles to f l o a t increased as the methanol concentration increased. In 50 vo l . % methanol only about half the particles floated, but most of those that did had been collected within the f i r s t minute so that a t the end of the 3 minute aeration period v i r t u a l l y no more particles were being carried into the co l l e c t i o n chamber.7 At 70 vol. % methanol, no detectable sulphur particles were transported to the col l e c t i o n chamber. The behavior of the -105 + 74 ym molybdenite sample was f a i r l y similar to that for the most floatable coal samples. Most of the particles were dispersed during conditioning in a l l solutions. A l l the particles had floated within the f i r s t 30 seconds in the more di l u t e solutions (2 and 5 vo l . % ) , and within the f i r s t minute for the 15 vol. % solution. The small amount of material reporting to the co l l e c t i o n chamber at 40 and 50 vol. % was slimes material, either floated or hydrodynamically transported. In f l o t a t i o n tests on sulphur, coal, or molybdenite, in methanol solutions where some or a l l of the material remained non-floatable, as soon as wash water was added to the solution during product c o l l e c t i o n procedures ( i . e . the surface tension was increased), the non-floatable particles immediately became floatable. This indicates very rapid r e v e r s i b i l i t y of the f l o a t / non-float t r a n s i t i o n . - 208 -These observations indicate q u a l i t a t i v e l y that the f l o t a t i o n rate decreases with solution surface tension, Y ^ , even when a l l the particles are s t i l l floatable (by d e f i n i t i o n , when Y 1 v > Y c f max). The decrease in fl o t a t i o n rate i s caused by the probability of f l o t a t i o n , P f, becoming smaller, due most l i k e l y to a reduction in lyophobicity (contact angle, 0). 8.4.3 Discussion of f l o a t a b i l i t y results F i r s t of a l l , the reasonably good correlation should be noted between the q u a l i t a t i v e results of the test tube experiments and the quantitative results of the small-scale f l o t a t i o n tests. It emphasizes the u t i l i t y of the simple, rapid, inexpensive tests for assessing the r e l a t i v e f l o a t a b i -l i t y and/or w e t t a b i l i t y of inherently hydrophobic solids in the particulate form. No conclusions can be drawn about the possible range of f l o a t a b i l i t i e s in the PTFE sample because a l l particles were floatable in the 100 vol. % methanol solution, Fig. 8.25. The r e l a t i v e l y narrow range of surface tension over which the f l o a t a b i l i t y of the -210 + 149 ym sulphur particles was reduced to zero suggests that both the i r physical properties ( s i z e , shape, density) and surface properties were quite uniform ( i . e . the we t t a b i l i t y band for the sulphur sample was probably quite narrow, section 2.3.3). From the shape of the f l o a t a b i l i t y curve i t cannot be ascertained whether f l o a t a b i l i t y was limited by adhesion or aggregate s t a b i l i t y . The narrow f l o a t a b i l i t y range observed for sulphur suggests that the methodology of froth!ess f l o t a t i o n tests in aqueous solutions of a short-chain n-alcoho! could be the basis for a fundamental laboratory investigation of the influence of contact angle and physical properties on p a r t i c l e f l o a t a b i l i t y . For - 209 -example, the w e t t a b i l i t y l i n e (band) of methylated s i l i c a plates could be determined, and f l o a t a b i l i t y tests carried out on methylated s i l i c a spheres of discrete sizes using techniques similar to those developed by Kitchener and colleagues. The Y c f range of about 12 dyne/cm (from about 40-52 dyne/cm) for the -105 + 74 ym molybdenite sample i s substantially wider than that for the sulphur sample. The small l y o p h i l i c gangue content (about 5%, Table 6.3) is probably too small to account for much of this d i s t r i b u t i o n , p a r t i c u l a r l y when the f l o a t a b i l i t y results for the r e l a t i v e l y high-ash content No. 5 seam coal sample are considered. However, the SEM photograph, Fig. 6.6, i l l u s t r a t e s that there is a s i g n i f i c a n t d i s t r i b u t i o n in both p a r t i c l e morphology and size (compared with the sulphur sample) and this i s one l i k e l y reason for the r e l a t i v e l y large f l o a t a b i l i t y d i s t r i b u t i o n . The p l a t e - l i k e morphology of many of the molybdenite particles makes e f f i c i e n t s i z i n g very d i f f i c u l t on square aperture screens. The cleavage and fracture characteristics of molybdenite (section 6.4.1.3) provide another possible explanation for the r e l a t i v e l y wide range of f l o a t a b i l i t i e s observed for this otherwise ostensibly homogeneous, inherently hydrophobic s o l i d . The SEM photographs of the molybdenite samples show that the " f l a t " cleavage faces of the particles contain the edges of many fractured sheets; i t i s therefore to be expected that the molybdenite particles would have a greater range of w e t t a b i l i t i e s than the sulphur particles (due to both surface roughness and heterogeneity, section 2.1.1), resulting in the w e t t a b i l i t y band for a particulate molybdenite sample probably being much wider than that for a sulphur sample. The water contact angle measured on c r y s t a l l i n e sulphur has been - 210 -found to vary very l i t t l e from one cleavage face to another, whereas a considerable range, was found on cleaved molybdenite surfaces containing edges of broken sheets (section 3.2.1). The range of of approximately 17 dyne/cm for the No. 5 seam coal sample has already been explained in section 8.3.2 in terms of mineral matter content, maceral composition, the physical properties of the p a r t i c l e s , and the f l o t a t i o n conditions. The quartzite sample had no detectable f l o a t a b i l i t y in any of the solutions tested, meaning that i t s minimum y ^ was greater than 72 dyne/cm. The effect of p a r t i c l e size on f l o a t a b i l i t y is further emphasized by the results in Fig. 8.25 for 3 d i f f e r e n t p a r t i c l e size fractions of molybdenite. The data points are. from individual tests. The cumulative % floated curve for the -149 + 105 ym size fraction i s shifted to higher solution surface tensions than the -105 + 74 ym f r a c t i o n , providing direct evidence that f l o a t a b i l i t y decreases as p a r t i c l e size increases. There is a continuous d i s t r i b u t i o n of f l o a t a b i l i t y between about 43 and 58 dyne/ cm, but above 58 dyne/cm very few additional particles became floatable, i.e. about 13% of the particles in the -149 + 105 ym sample are non-float-able in aqueous methanol solutions. The plateau in the f l o a t a b i l i t y curve well below 100% floated indicates a discontinuity in the f l o a t a b i l i t y d i s t r i b u t i o n which was not observed for the f i n e r size f r a c t i o n . Such a discontinuity could be explained in several ways. Assuming that p a r t i c l e f l o a t a b i l i t y is limited by aggregate s t a b i l i t y , such an effect could be caused by a discontinuity in the d i s t r i b u t i o n of surface properties amongst the p a r t i c l e s , i l l u s t r a t e d schematically by the s p l i t w e t t a b i l i t y band in Fig. 8.26(a). The upper part of the w e t t a b i l i t y band represents the non-floatable fraction of p a r t i c l e s ; i t i s always above the - 211 -solution surface tension , y dyne/cm Iv Fig. 8.26 Adhesion tension diagrams i l l u s t r a t i n g two possible ex-planations for a plateau (discontinuity] in the f l o a t -a b i l i t y curve for an inherently hydrophobic s o l i d in aqueous, methanol solutions, assuming f l o a t a b i l i t y i s controlled by bubble-particle aggregate s t a b i l i t y ; (a) a s p l i t w e t t a b i l i t y band; (b) a bimodal p a r t i c l e size d i s t r i b u t i o n - 212 -aggregate s t a b i l i t y l i n e for the smallest particles ( d ^ ) for Y 1 v below YH 20' As Y 1 v increases from y ^ min to y* ^ only the particles in the lower w e t t a b i l i t y band become floatable. However, i t i s hard to imagine what would produce such a s p l i t w e t t a b i l i t y band (except for complete li b e r a t i o n of two phases with very different w e t t a b i l i t i e s ) . The -105 + 74 ym molybdenite size fraction was prepared in the same manner from the same massive sample, but i t does not show such behavior. The gangue content of the two size fractions i s about the same. If the -149 + 105 ym o size fraction had been subjected to an effect l i k e surface oxidation, i t s wett a b i l i t y would be reduced (and the range possibly broadened) in a similar manner to that found for the adit coal sample discussed in section 8.3.3. Some of the larger more oxidized particles may then become non-floatable, i. e . those represented by the region in the extended wettability band above the d m a x aggregate s t a b i l i t y l i n e between YCF max and Y^ i n the schematic i l l u s t r a t i o n in Fig. 8.27(a). Such a situation would result in a continuous f l o a t a b i l i t y curve of the type shown in Fig. 8.27(b); the y i e l d - f l o a t a b i l i t y curve for the p a r t i a l l y weathered 38' adit coal sample at 3 minutes con-ditioning time, Fig. 8.8, i s an example. Another remote p o s s i b i l i t y i s that the pa r t i c l e size d i s t r i b u t i o n between the nominal size l i m i t s of 149 and 105 ym may have been bimodal. Such a situation could result in the^wettability band being to the l e f t of the aggregate s t a b i l i t y lines for the coarse mode of the d i s t r i b u t i o n , Molybdenite i s not known to be very susceptible to atmospheric oxidation; also, low temperature oxidation products are water soluble and do not have a marked effect on hydrophobicity (119) (see section 6.4.1.3). - 213 -Fig. 8.27 I l l u s t r a t i o n of the expected effect, of partial surface. oxidation of a particulate sample of an inherently hyd-rophobic s o l i d , assuming f l o a t a b i l i t y i s controlled-by . bubble-particle aggregate s t a b i l i t y ; (a) adhesion tens-ion diagram showing extended w e t t a b i l i t y region in which particles are non-floatable below Y H Q; (b) shape of resulting cum.% floated curve 2 - 214 -d" t° d m . schematically i l l u s t r a t e d in Fig. 8.26(b) ( i t assumes that max there are no particles in the sample with sizes between d' and d"). y ^ is the surface tension at which a l l the particles in the fine mode, d . r min to d', become floatable, i.e. the start of the f l o a t a b i l i t y plateau. The SEM photograph, Fig. 6.6, gives no indication of such a unique bimodal d i s t r i b u t i o n for the -149 + 105 ym molybdenite sample. There i s one explanation which does not rely on the existence of discontinuities in the surface or physical properties of the particles in the sample. It assumes that f l o a t a b i l i t y was controlled by bubble-p a r t i c l e adhesion and not aggregate s t a b i l i t y , as suggested by Jowett (170) (see section 2.3.1). He considered that the tr a n s i t i o n from floatable to non-floatable can occur over a very narrow size range where induction time, t , becomes greater than contact time, t . Thus, there may have been some large particles in the -149 + 105 urn molybdenite sample which were non-floatable because the i r size was such that t >t over the entire range of Y-| v, as i l l u s t r a t e d in Fig. 2.11. To produce a plateau in the f l o a t a b i l i t y curve between about 58 and 69 dyne/cm (10 and 2 vol. % methanol, Fig. 8.25) i t would also have to be assumed that t c > and the relationship between induction time and p a r t i c l e s i z e , were r e l a t i v e l y independent of methanol concentration in this region. If this reasoning is valid i t would mean that, in some circumstances, the shape of a % floated vs. y-j v curve could be used as an indication of control by either adhesion or aggregate s t a b i l i t y . Only a small proportion of the particles in the -210 + 149 ym moly-bdenite size fraction were floatable in any solution. The overlap of f l o a t a b i l i t y ranges for the three size fractions was probably due,at least i n part, to the overlap of p a r t i c l e size caused by i n e f f i c i e n c i e s in the screening - 215 -process ( i . e . each fraction would contain a proportion of undersize material), as well as to the existence of a range of w e t t a b i l i t i e s (and thin f i l m properties). 8.5 Selective Flotation Separation of Inherently Hydrophobic Solids It is c l e a r l y indicated that a certain degree of selective separa-tion should be possible between the samples of inherently hydrophobic solids for which the f l o a t a b i l i t y curves are presented in Fig. 8.24. Almost complete separation should be possible between the PTFE sample and each of the other 3 samples (sulphur, molybdenite and coal) and also between sulphur and both the molybdenite and coal samples. A clear-cut separation between the -105 + 74 ym molybdenite sample and the coal sample may not be possible because of the small overlap of t h e i r Y .p regions. However, some degree of preferential concentration of the molybdenite in the f l o a t product should be achievable. 8.5.1 Results of test tube experiments An exploratory test tube experiment was conducted on 1:1 mixtures of PTFE and coal (by weight) to see i f this technique could be used as a rapid guide for assessing whether selective separation was possible. Based on the results of the individual test tube experiments on PTFE and the No. 5 seam coal sample (Figs. 8.23 and 8.1) i t was expected that a clear separation would occur in solutions 1, 2 and 3 (the coal being non-f l o a t a b l e ) , whereas both PTFE and coal particles would f l o a t together in solutions 4-7. However, Fig. 8.28 i l l u s t r a t e s that, for a l l the solutions, the particles which collected at the solution interface or in p a r t i c l e / bubble agglomerates were almost a l l PTFE. The small number of black coal particles which can be seen associated with the floated PTFE were easily - 216 -Fig. 8.28 Selective separation of PTFE from No. 5 seam coal in aqueous methanol solutions (1:1 mixture by wt., -210 + 149 urn) - 217 -dislodged by gentle tapping of the test tubes, and appeared to be only mechanically entrapped. The results of similar experiments on sulphur/coal and sulphur/ molybdenite mixtures could not be easily interpreted in terms of f l o a t a b i -l i t y s e l e c t i v i t y . I t was therefore concluded that the simple test tube experiments in the i r present form are not suitable for studying f l o t a t i o n s e l e c t i v i t y between inherently hydrophobic s o l i d s . 8.5.2 Results of small-scale f l o t a t i o n tests Small-scale f l o t a t i o n tests were carried out on 1:1 mixtures (by weight) using the same procedures and conditions that were used for the individual materials (section 7.2), with conditioning and f l o t a t i o n times of 5 and 3 minutes respectively (unless otherwise specified), and no bubble deflector. The impeller speed was adjusted between 300 and 350 r.p.m. to avoid sanding, but without causing non-slimes particles to be hydrodynamically carried into the c o l l e c t i o n chamber of the P/S c e l l . Table 8.3 l i s t s the aqueous methanol solution concentrations used and the g grade and recovery data for the f l o t a t i o n products, (a) PTFE mixtures The tests on PTFE mixtures were carried out at 80 vol. % methanol in an attempt to disperse the PTFE particles as much as possible during conditioning and minimize the formation of large particle/bubble agglom-erates (which may have mechanically entrapped or swept non-floating particles into the c o l l e c t i o n chamber) while maintaining a r e l a t i v e l y small bubble size. The deterioration of f l o a t product quality when The data are given to the nearest percentage.point because of the r e l a t i v e l y low precision of the simple analysis techniques used (section 7.3.2) and d i f f i c u l t i e s associated with obtaining representative analysis samples from the synthetic mixtures due to segregation. - 218 -Table 8.3 Results of f l o t a t i o n tests on mixtures of inherently hydrophobic materials in aqueous methanol solutions; 5 minutes conditioning, 3 minutes f l o t a t i o n , no bubble deflector head1 mixture meth, cone. cum. wt. % meas. grade 0 wt. % c a l . c a l . head b rec. 1 grade wt. % vol. % floated f l oats rejects % wt. PTFE/sulphur 50 80 50 >99 <1 99 50 PTFE/MoS2 50 80 50 >99 <1 99 50 (-105 + 74 um) PTFE/MoS0 (-149 + 105 ym) 50 80 50 >99 <1 99 50 PTFE/coal 50 80 50 >99 <1 99 50 sulphur/MoS2 50 30 80 62 2 99 50 (-105 + 74 ym) 40 45 53 29 88 92 8 33 93 53 50 50 50 15 93 43 27 50 sulphur/MoS,, 50 30 53 92 1 98 50 (-149 + 105 ym) 40 43 99 14 85 50 sulphur/coal 50 25 51 97 1 99 50 30 49 97 3 97 49 30 50 97 2 98 49 30 c 43 98 14 84 50 MoS0 (-105 + 48 d 15 88 53 2 99 47 74 ym)/coal 15 20 86 57 57 82 2 2 99 98 49 48 25 42 93 15 82 48 aUnless otherwise specified the nominal size of both materials in each mixture is -210 + 149 ym. DGrade refers to the amount of the f i r s t component in each mixture. Conditioning time was 15 minutes. dThe measured head grade of the -105 + 74 ym MoS? sample was 96.2% (Table 6.3). - 219 -pronounced aggregation of very hydrophobic particles takes place has been previously noted (22). Despite this precaution, a l l the PTFE particles s t i l l floated as f a i r l y large agglomerates in the f i r s t few seconds; how-ever, as the results in Table 8.3 show, only negligible amounts of sulphur, molybdenite or coal reported to the co l l e c t i o n chamber. Therefore almost perfect separation was achieved between PTFE and the other inherently hydrophobic samples, (b) sulphur/molybdenite The results for the -105 + 74 ym MoS2 mixture i l l u s t r a t e that, as the solution concentration increases (surface tension decreases), the grade of the sulphur in the floats (concentrate) increases while the recovery decreases. This i s the typical inverse grade-recovery relationship. At 40, 45 and 50 vol. % methanol only a r e l a t i v e l y small amount of MoS2 contaminated the sulphur f l o a t product, as predicted by the f l o a t a b i l i t y curves in Fig. 8.24 (most of i t appearing to be hydrodynamically transported into the co l l e c t i o n chamber). At 30 vol. % methanol, where Fig. 8.24 shows a s i g n i f i c a n t proportion of the molybdenite particles being floatable, i t was observed that v i r t u a l l y a l l the sulphur particles floated rapidly in about the f i r s t 15 seconds as particle-coated bubbles and particle/bubble agglomerates with very few MoS2 particles associated with them. It was only after most of the sulphur particles had floated that the MoS2 particles began attach-ing to freshly introduced bubbles. The sulphur recovery for the f l o t a t i o n tests on the.sulphur/moly-bdenite mixtures (for both the -105 + 74 ym and -149 + 105 ym MoS2 size fractions) at 40, 45 and 50 vol, % i s less than the wt. % floated in the tests on sulphur alone (Fig. 8.24). One possible explanation for t h i s - 220 -reduction in sulphur f l o a t a b i l i t y i s that fine molybdenite slimes (of low lyophobicity, or even l y o p h i l i c at these solution concentrations) may have p a r t i a l l y coated some of the sulphur particles and reduced the i r f l o a t a b i l i t y . The best sulphur/molybdenite separation results are 88% grade and 93% recovery (of sulphur) at 40 vol. % methanol for the -105 + 74 ym MoS2 size f r a c t i o n , and 92% grade and 98% recovery at 30 vol. % methanol for the -149 + 105 ym size f r a c t i o n . (c) sulphur/No. 5 seam coal Almost complete separation of sulphur from coal was achieved at 25 and 30 v o l . % methanol and 5 minutes conditioning time ( i . e . a grade of 97% sulphur, and recoveries between 97 and 99%). V i r t u a l l y a l l the sulphur floated within 30 seconds. The reduction in recovery of sulphur to about 84% at 30 v o l . % methanol when the conditioning time was increased to 15 minutes could again be due to the release of l y o p h i l i c slimes from the coal which may have p a r t i a l l y coated some sulphur particles and reduced thei r f l o a t a b i l i t y . (d) -105 + 74 ym molybdenite/No. 5 seam coal The f l o a t a b i l i t y curves in Fig. 8.24 give a reasonable prediction of the results for f l o t a t i o n tests on the molybdenite/coal mixtures, Table 8.3. At 15 v o l . % methanol about 99% of the molybdenite floated, but also a considerable amount of the coal; at 20 vol. % almost a l l the molybdenite was s t i l l recovered (98%) but only a small amount of coal reported to the f l o a t s , resulting in an increase in MoS^ grade to 82%; at 25 vol. % the recovery of MoS2 drops but the grade increases to 93%. So, although com-plete separation between these two samples does not occur, there i s con-siderable preferential concentration. The optimum solution conditions for separating these two samples is between 20 and 25 v o l . % methanol. - 221 -The f l o a t a b i l i t y curves for the coarser size fractions of NoSr,, Fig. 8.25, indicate that the f l o a t a b i l i t y differences between coal and molybdenite would become less as the M0S2 p a r t i c l e size increases. After the gas was introduced at 15 v o l . % methanol the molybdenite was observed to flo a t as small particle/bubble aggregates with very few associated coal p a r t i c l e s . Only when most of the M0S2 had floated did the coal particles s t a r t attaching. 8.5.3 Discussion In the test tube experiments and some of the small-scale f l o t a t i o n te s t s , where both solids in the mixture were s t i l l floatable ( i . e . Y ^ v > A B Y £ a n d Y ~ ) , i t was observed that the more floatable s o l i d (the s o l i d cf cf with the lowest Y .p) floated f i r s t (e.g. the PTFE/coal test tube experiment in dilute solutions, and the 30 v o l . % methanol sulphur/-105. + 74 ym M0S2 and 15 v o l . % methanol -105 + 74 ym MoS2/coal f l o t a t i o n t e s t s ) . It appeared that the particles of the most floatable s o l i d p r e f e r e n t i a l l y coated any available solution/vapour interface and either prevented par-t i c l e s of the second s o l i d from attaching ( i . e . formed a protective "ar-mour" coating), or actually displaced any that may have i n i t i a l l y attached. Perhaps in a situation where the particles of the least floatable s o l i d are much smaller than those of the most floatable s o l i d , i t may be possible for the small particles to attach to the solution/vapour interface in the in t e r s t i c e s between the larger p a r t i c l e s . However, the probability of c o l l i s i o n with such small exposed areas on a crowded interface would l i k e l y be very low. The PTFE/coal test tube experiment i l l u s t r a t e s the situation where the available area of solution/vapour interface i s limited and i s covered by the more floatable s o l i d (PTFE) to the almost complete exclusion of the - 222 -less floatable s o l i d (coal), in the dilut e solutions. The small-scale f l o t a t i o n tests demonstrate the case where fresh solution/vapour in t e r -face (bubbles) i s continuously created. The particles of the more floatable s o l i d p r e f e r e n t i a l l y coat the i n i t i a l bubbles u n t i l the majority of them have been removed; then the particles of the less floatable s o l i d have exposed solution/vapour interface on which to attach, and they begin to f l o a t . That the extent of available free solution/vapour interface can be rate-controlling under hindered f l o t a t i o n conditions i s well known (22,148, 329); heavy p a r t i c l e coatings prevent other floatable particles from attaching, thereby reducing the i r f l o t a t i o n rate. King (329) accounted for i t in his continuous f l o t a t i o n rate expression with a parameter r e f l e c t -ing the average fraction of the bubble surface not covered by particles ( i . e . free surface); he considered i t to be a function of pulp density and average bubble residence time in the pulp; by implication i t was assumed to be independent of mineralogy and surface a c t i v i t y of the adhering particles although i t did contain an adjustable constant. However, with a few excep-tions (221,330), very l i t t l e emphasis has been given to "preferential" or "selective" bubble surface armouring as a rate-controlling parameter. Hemmings (330) found that fine quartz slimes in C-jgTAB solutions (cetyl trimethylammonium bromide) pr e f e r e n t i a l l y adhered to gas bubbles, forming an effective armour coating which prevented f l o t a t i o n of coarser p a r t i c l e s , which, i f on t h e i r own, floated at a faster rate than the fine slimes on t h e i r own. The f l o t a t i o n rate studies of Tomlinson and Fleming (148) on mixed mineral systems, where both minerals had co l l e c t o r induced hydro-phobicity (apatite and hematite.in potassium oleate), also showed this effect but to a lesser degree. While the highest f l o t a t i o n rates were - 223 -obtained under free f l o t a t i o n conditions (bubbles only sparsely coated by particles at a l l times) and some preferential concentration of apatite took place, they found that s e l e c t i v i t y (grade) was improved under hindered f l o t a t i o n conditions where the bubble surfaces were saturated with p a r t i c l e s . Selective separation based on different f l o t a t i o n rates has often been emphasized (148,309,331). However, i t has not been extensively used in plant practice although size c l a s s i f i c a t i o n of plant feed and " s p l i t -conditioning" have been recommended to take advantage of differences in f l o t a t i o n rate of particles of the same size but different w e t t a b i l i t i e s , p a r t i c u l a r l y in the coal industry (312,332). Differences in f l o t a t i o n rate i s also the basis for the Release Analysis technique for assessing coal f l o t a t i o n washability (see section 8.3.1). The present discussion of the f l o t a t i o n of mixtures of two solids with different degrees of f l o a t a b i l i t y indicates that the rate of p a r t i c l e attachment w i l l depend not only on the f l o a t a b i l i t y of the particles trying to attach and the fraction of free solution/vapour interface available, but also on the f l o a t a b i l i t y of the particles already attached and/or competing with them for the same free interface, p a r t i c u l a r l y under hindered f l o t a -tion conditions. The f l o t a t i o n rate of the particles of the more floatable s o l i d in a mixture would probably be much the same as when being floated on t h e i r own; however, there w i l l be a lag time before the particles of the less floatable s o l i d begin f l o a t i n g , resulting in thei r "apparent" f l o t a -tion rate in a mixture being lower than when floated on their own. Thus, for mixtures of two inherently hydrophobic s o l i d s , A and B, in aqueous methanol solutions two means exist for achieving selective separation. The f i r s t is to choose a solution surface tension such that - 224 -Y C F max <  yiy * Y ^ min, as i l l u s t r a t e d in Fig. 2.14 and discussed in section 2.3.3. This c r i t e r i o n ( i . e . the separation of c r i t i c a l f l o a t a b i l i t y bands) could s t i l l be met even i f the w e t t a b i l i t y bands coincide; i t would require substantial differences in the size and/or density of the particles of the two s o l i d s . The second way of achieving at least some selective separation (pre-f e r e n t i a l concentration) under hindered f l o t a t i o n conditions i s to u t i l i z e the differences in f l o t a t i o n rate resulting from the pre-ferential coating of bubbles by the more floatable s o l i d at Y - ^ V anywhere above Y ^ min ( i . e . where both solids are f l o a t a b l e ) . The results and discussion of f l o t a t i o n tests on inherently hydrophobic solids in aqueous methanol solutions presented in the l a s t three sections suggest that t h i s methodology can be used as a laboratory technique in the investigation of f l o t a t i o n funda-mentals, and also have possible industrial application. By studying c a r e f u l l y prepared particulate s o l i d samples which are as homogeneous as possible in terms of surface and physical properties, i t should be possible to assess independently the effect of par t i c l e lyophobicity (contact angle), s i z e , shape and density, and the properties of the d i s j o i n i n g f i l m on f l o a t a b i l i t y (as characterized by the c r i t i c a l surface tension of f l o a t a b i l i t y ) . I t may also be possible to determine under what conditions f l o a t a b i l i t y i s controlled by bubble-particle adhesion or aggregate s t a b i l i t y . - 225 -In some of the more complex mechanistic f l o t a t i o n rate models developed to describe continuous f l o t a t i o n systems, the rate of f l o t a t i o n i s considered to be influenced by three d i s t i n c t p a r t i c l e properties - si z e , mineralogical composition . and surface a c t i v i t y ; the values of each of these properties are distributed over the individual particles in the par t i c l e pop-ulation and result in a distributed set of rate constants (333, 361). The assignment of numerical values to the l a s t property, surface a c t i v i t y , has presented the greatest d i f f i c u l t i e s (329). C r i t i c a l surface tension of f l o a t a b i l i t y i s a parameter which could possibly be used to quantify "surface a c t i v i t y " in this context. It i s quite common in practise for ores to contain trouble-some quantities of an inherently floatable contaminant, such as talc or carbonaceous material, which has to be suppressed with s p e c i f i c reagents.(362). There also are examples where an ore contains two phases which both exhibit native f l o a t a b i l i t y making selective separation d i f f i c u l t to achieve; for example the talcose/ molybdenite ores in the A b i t i b i area of Quebec (334) where talc i s the contaminant, the chalcopyrite/molybdenite deposit at I Island Copper Mine, Vancouver Island, B.C. which contains a carb-onaceous material which often contaminates the molybdenite con-centrate (335), and the separation of bitumen impurities from sulphur concentrate (221). In.the chemical and waste recycling industries there are no doubt many instances where natural or - 226 -synthetic mixtures contain two or more inherently floatable materials which require selective separation [Gaudin has described some examples (21)] . The c r i t i c a l surface tension of f l o a t a b i l i t y concept may provide a new avenue for investigating these practical problems and possibly be the basis for a selective f l o t a t i o n process. - 227 -CHAPTER 9  CONCLUSIONS 9.1 General A concept of c r i t i c a l surface tension of f l o a t a b i l i t y , Y ^,  n a s b e e n developed. It postulates that, for a given set of fl o t a t i o n conditions, (a) an inherently hydrophobic particle in aqueous solutions of a non-i o n i c , non-micelle forming wetting surfactant becomes non-floatable at a c r i t i c a l solution surface tension of f l o a t a b i l i t y , Y ^ , which i s greater than i t s c r i t i c a l surface tension of wetting, Y , (determined with the same solutions), irrespective of whether f l o a t a b i l i t y i s controlled by bubble-particle adhesion or aggregate s t a b i l i t y ; (b) the Y c ^ of a pa r t i c l e depends on i t s w e t t a b i l i t y properties (characterized by the relationship between contact angle and solution surface tension), as well as i t s physical properties such as size and density, and the properties of the thin disjoining f i l m ; (c) selective f l o t a t i o n separation of two inherently hydrophobic p a r t i c l e s , A and B, w i l l be possible in a surfactant solution i f i t s surface A B tension, y^, i s such that Y C F < Y ^ Y ^ Y C f 9.2 Coal (a) Particulate coal samples with a narrow size d i s t r i b u t i o n displayed a considerable range of particile f l o a t a b i l i t y in a sequence of aqueous methanol solutions of varying surface tension. (b) For a readily floatable coal sample the relationship between cumula-tive ash content and y i e l d of the floated material was independent of the conditioning time and f l o t a t i o n time within the ranges studied. - 228 -above about 20% y i e l d . This relationship was considered to rep-resent an approximation of the fundamental or ideal f l o t a t i o n washability characteristics of the coal sample. (c) Flotation washability (defined.as the cumulative ash content f o r a given cumulative yield) wasinversely related to the feed ash content. (d) Sample f l o a t a b i l i t y (defined as the vol. % methanol f o r a given cumulative yi e l d ) did not correlate with feed ash content. For large degrees of weathering f l o a t a b i l i t y was inversely related to bulk parameters used to indicate degree of weathering, but was also sensitive to small surface changes which were not reflected by the same parameters. (e) Free swelling index did not necessarily r e f l e c t the effect of weathering/oxidation on the f l o a t a b i l i t y of the coal samples, p a r t i -c u l a r l y for the sample ri c h in non-reactive petrographic components. (f) The floatability-incremental ash content relationships for the readily-floatable coal samples consisted of two main regions; in one, f l o a t a b i l i t y was inversely related to incremental ash content, while in the other they were independent. The shape and r e l a t i v e position of the curves allowed the e f f e c t of rank, petrographic composition and otherwise, undetected oxidation on organic matter w e t t a b i l i t y to be assessed. 9.3 Other Inherently Hydrophobic Solids (a) The f l o t a t i o n recovery of narrow size fraction samples of several inherently hydrophobic solids in aqueous methanol solutions decreased from a maximum to zero over discrete ranges of Y^, called Y ^ bands. The r e l a t i v e f l o a t a b i l i t y of a sample was biven by the position of i t s - 229 -Y ^ band in the Y-j spectrum. F l o a t a b i l i t y decreased in the order: polytetrafluoroethylene -sulphur - medium v o l a t i l e bituminous coal - molybdenite, for the same size f r a c t i o n . The f l o a t a b i l i t y of molybdenite decreased as the p a r t i c l e size increased. V i r t u a l l y complete selective f l o t a t i o n separation was achieved in aqueous methanol solutions between the samples of inherently hydro-phobic solids whose individual. Y ^ bands did not overlap. Prefer-e n t i a l concentration of the more floatable sample (molybdenite) was obtained from a molybdeni.te/coal mixture whose individual f l o a t a b i l i t y bands overlapped but did not completely coincide. These are believed to be the f i r s t demonstrations of selective separation of inherently hydrophobic solids solely by the man-ipulation of the surface tension of the f l o t a t i o n solution. Under hindered f l o t a t i o n conditions (limited a v a i l a b i l i t y of fresh solution/vapour interface) the most floatable s o l i d in the mixture pr e f e r e n t i a l l y coated the bubbles and almost completely prevented attachment of the less floatable s o l i d (for solution conditions in which both solids were i n d i v i d u a l l y f l o a t a b l e ) . - 230 -CHAPTER 10  SUGGESTIONS FOR FURTHER WORK The current study i s believed to be the f i r s t to have developed the concept of c r i t i c a l surface tension of f l o a t a b i l i t y , y cf» a n d demonstrated that i t can be used to characterize the f l o a t a b i l i t y properties of particulate samples of coal and other inherently hydrophobic s o l i d s , and achieve selective separations between them. The experimental program was essentially exploratory in nature and has opened up several possible avenues for future research and practical application, (a) coal The limitations of the system for determining the f l o t a t i o n wash-a b i l i t y characteristics of very fine coal p a r t i c l e size ranges have to be delineated, as well as i t s a p p l i c a b i l i t y to d i f f i c u l t - t o - f l o a t low-rank and weathered coals. Larger scale experimental systems need to be developed to provide s u f f i c i e n t product material for petrographic, spectroscopic, and chemical as well as ash analyses, so that the relationship between f l o a t a b i l i t y and organic matter properties can be quantitatively assessed. A comparison is required of the various methods of assessing the f l o t a t i o n washability characteristics of fine coal samples. The technique should be investigated as a possible means for obtain-ing group maceral concentrates for laboratory investigations from coal samples which have been ground to li b e r a t e the macerals from each other and the majority of the associated mineral matter. Its a p p l i c a b i l i t y to the production of reactive r i c h concentrates on an industrial scale for coking or liquefaction would be doubtful unless - 231 -an economic reagent recovery system could be devised. In this regard the use of smaller quantities of higher molecular weight, more surface active alcohols to decrease surface tension needs to be investigated. However, they may have decreased s e l e c t i v i t y due to t h e i r a b i l i t y to form more persistent froths and t h e i r possible " c o l l e c t i n g " properties, (b) other inherently hydrophobic solids Uncertainty s t i l l exists as to the fundamental relationships between pa r t i c l e f l o a t a b i l i t y and the surface and physical properties of the p a r t i c l e , and the f l o t a t i o n solution and c e l l conditions. y ^ could provide a convenient parameter for assessing the individual effect of each factor using well characterized homogeneous samples of an inherently hydro-phobic s o l i d . In this context i t may be possible to combine the y ^ and s a l t f l o t a t i o n techniques to investigate the relationship between f l o a t -a b i l i t y , contact angle and e l e c t r i c a l double layer properties. The technique may also provide a means for determining whether p a r t i c l e f l o a t a b i l i t y i s controlled by adhesion or by aggregate s t a b i l i t y . Flotation modelling studies require a method of characterizing the d i s t r i b u t i o n of wetting properties within heterogeneous particulate samples made hydrophobic by c o l l e c t o r adsorption. The.Y f method could be suitable,provided that the wetting surfactant (e.g. methanol) does not interact s i g n i f i c a n t l y with the adsorbed c o l l e c t o r coatings. The technique should be assessed as the basis of a practical process for s e l e c t i v e l y separating natural or synthetic mixtures of inherently floatable materials in the mineral, chemical or waste recycling industries. - 232 -REFERENCES 1. International Coal 1979, National Coal Assoc., Wash., D.C., 1980 2. Coal Preparation S t a t i s t i c s , Appendix 1, 7 t n Intern. Coal Prep. Cong., Au s t r a l i a , 1976,edit. A.C. Partridge 3. Davis, A., Spackman, W. and.Given, P.H., The influence of the properties of. coals on.their conversion into ^clean fuels, Energy Sources, 3(1), 1976, pp. 55-81 4. Brown, T.D., Coal g a s i f i c a t i o n and liqu e f a c t i o n , CIM Bull ., 74(825) ,1981 , pp. 122-133 5. Jowett, A., Preparation of low ash coal for carbonization and conversion processes, Proc. Australas. Inst. Min. Metal!., No. 275, Sept. 1980, pp. 5-16 6. Gagnon, A., Coal processing at Fording Coal Limited, CIM Bui 1., 71(797), 1978, pp. 103-106 7. Harkins, W.D., The Physical Chemistry of Surface Films, Rheinhold Publ. Corp., New York, 1952 8. Adamson, A.W., Physical Chemistry of .Surfaces, ,3rd Edit.V.'John:Wiley.and Sons:, .New York, 1976 9. Johnson, R.E. and Dettre, R.H., Wettability and contact angles, In: Surface and Colloid Science, ed. E. Matijevic, Wiley-Interscience, NewYork, 1969, Vol. 2, pp. 85-153 10. Good, R.J., Contact angles and the surface free energy of s o l i d s , In: Surface and Colloid Science, Vol. 11, Experimental methods, edits . R.J. Good and R.R. Stromberg, Plenum Press, New York, 1979 pp. 1-30 11. Aveyard, R. and Haydon, D.A., An introduction to the principles of surface chemistry, Cambridge Univ. Press, 1973 12. Young, T., Essay on the cohesion of f l u i d s , P h i l . Trans. Roy. Soc., 95, 1805, p. 65 13. Dupre, A., Theorie Mechanique de l a Chaleur, Gautjer-Villars, Paris, 1869, p. 368 (French) 14. Johnson, R.E., Conflicts between Gibbsian thermodynamics and recent treatments of i n t e r f a c i a l energies in solid-liquid-vapour systems, J. Phys. Chem., 63, 1959, pp.1655-1658 15. Bickerman, J.J., Contact angles, spreading and wetting, In: Proc. 2nd Intern. Cong, of Surface A c t i v i t y , London, ed. J.H. Schulman, Butter-worths, London, 1957, Vol. 3, pp.125-130 16. Bickerman, J.J. Physical surfaces, Academic Press, New York, 1970 - 233 -17. Cooper, W.F. and N u t t a l l , W.H., The theory of wetting, and the determi -nation of the wetting power of dipping and spraying f l u i d s contain-ing a soap basis, J. Agric. S c i . , 7, 1915, pp. 219-239 18. de Bruyn, P.L. .and Agar, G.E., Surface chemistry of f l o t a t i o n , In: Froth Flotation,;:5Qth Anniversary Volume, ed. D.W. Fuerstenau, Amer. Inst. Min. Metallurgical and Petrol. Engrs., Inc., New York, 1962, pp.91-138 19. Neumann, A.W., Contact angles and t h e i r temperature dependence: Thermo-dynamic status , measurement, interpretation and application, Advances  in C o l l . and Interface S c i . , 4(2/3), 1974, pp.105-191 20. Sutherland, K.L., and Wark, I.W., Principles of Flotation, Australas. Inst. Min. M e t a l l . , Melbourne, 1955 21. Gaudin, A.M., Flotation, 2nd Edi t . , McGraw H i l l , New York, 1957 22. Klassen, V.I. and Mokrousov, V.A., An Introduction to the Theory of Flo-t a t i o n , translated J. Leja and G.W. Poling, Butterworths, London, 1963 23. Finch, J.A., and Smith, G.W., Contact angles and wetting, Minerals S c i . Engng., 11(1), 1979, pp.36-63 24. Hornsby, D.T., and Leja, J . , Selective f l o t a t i o n and i t s surface chemical c h a r a c t e r i s t i c s , accepted for publication in Surface and Colloid  Science, Vol. 12, ed. E. Matijevic, 1981 25. Smolders, C.A., Contact angles, wetting and dewetting of mercury. Part I I I . Contact angles on Mercury, Recueil des Travaux Chimiques des  Pays-Bas, 80, 1961, pp. 699-720 26. O t t e w i l l , R.H., B i l l e t , D.F., Gonzalez,., G., Hough, D.B. and L o v e l l , V.M., The variation of contact angle at the s i l v e r iodide-liquid-vapour interface with the charge on the s o l i d surface, In: Wetting, '-Spreading  and Adhesion, ed. J.F. Padday, Academic Press, London, 1978, pp. 183-199 27. Neumann, A.W. and Good, R.J. Techniques of measuring contact angles, In: Surface and Colloid Science. Vol. 11 Experimental Methods, e d i t s , R.J. Good and R.R. Stromberg, Plenum Press, New York, 1979, pp. 31-92 28. E l l i o t , T.A. and Leese, L., Dynamic contact angles. Part 1. Changes in a i r - s o l u t i o n - s o l i d contact angles with time, J. Chem. S o c , Part A, 1957, pp. 22-30 29. E l l i o t , T.A. and Ford, D.M., Dynamic Contact angles. Part 7. Impact spreading of water drops i n a i r and aqueous solutions of surface active agents in vapour on smooth paraffin wax surfaces, J. Chem. Soc, Faraday Trans. I, 68, 1972, pp. 1814-1823 30. E l l i o t , T.A. and Morgan, M., Dynamic contact angles. Parts 111-VI. J. Chem. S o c , A, 1966, pp.558-572 - 234 -31. E l l i o t t , G.E.P., and Riddiford, A.C., Dynamic contact angles. I. The e f f -ect of.impressed motion, J. C o l l . Interface S c i . , 23, 1967, pp. 389-398 32. Schulman, J.H. and Leja, J . , S t a t i c and dynamic attachment of a i r bubbles to s o l i d surfaces, In: Surface Phenomena in Chemistry and Biology, edit s . J.F. D a n i e l l i , K.G.A. Pankhurst and A.C. Riddiford, Pergamon Press, New York, 1958, pp. 236-245 33. Schwartz, A.M., The dynamics of contact angle phenomena, Advan. C o l l . In-terface S c i . , 4(4), 1975, pp. 349-374 34. Wolfram, E. and Faust, R., Liquid drops, on a t i l ted plate, contact angle hysteresis and the Young contact angle, In: Wetting, Spreading and Ad- hesion, ed. J.F. Padday, Academic Press, London, 1.978, pp. 213-222 35. Blake, T.D., and Haynes, J.M., Contact angle hysteresis, In: Progress i n Surface and Membrane Science, Vol. 6, eds. J.F. D a n i e l l i , M.D. Rosen-berg and D.A. Cadenhead, Academic Press, New York, 1973, pp. 125-138 36. Neumann, A.W., Contact angles, In: Wetting, Spreading and Adhesion, e d i t . J.F. Padday, Academic Press, Inc., London, 1978, pp. 3-35 37. Leja, J . , Contact angles, wetting and i t s applications, paper presented at 3Qth Canadian Chemical Engineering-Conf., 19-22 Oct. 1980, Edmon*' ton, Alberta 38. Wenzel, R.N., Resistance of s o l i d surfaces to wetting by water, Ind. Eng. Chem., 28(8), 1936, pp. 988-994 39. Johnson, R.E. and Dettre, R.H., Contact angle hysteresis, Part I. Study of an idealized rough surface, In: Contact angle, Wettability and  Adhesion, Adv. Chem. Series No. 43, ACS, Wash., D.C., 1964, pp. 112 -135 40. Dettre, R.H. and Johnson, R.E., Contact angle hysteresis. Part I I . Con-tact angle measurements on rough surfaces, In: Contact Angfte, Wetta- b i l i t y and Adhesion, Adv. Chem. Series No. 43, ACS, Wash., D.C., 1964, pp. 136-144 41. Eick, J.D., Good, R.J. and Neumann, A.W., Thermodynamics of contact angles II:. Rough s o l i d surfaces, J. C o l l . Interface S c i . , 53(2), 1975, pp. 235-248 42. Huh, C , and Mason, S.G., Effects of surface roughness on wetting (theor-e t i c a l ) , J_^jC^lJ_Mnt^rfac^ 60(1), 1977, pp.1.1-18 43. Oliver, J.F., =Huh,C. and Maspn,S.G.., Resistance, to spreading of l i q u i d s by sharp, corners, J. C o l l . Interface S c i . , 59(3), 1977, pp. 568-581 44. Bayramli, E., and Mason, S.G.., Liquid spreading: edge effect for zero contact angle, J. C o l l . Interface S c i . , 66(1 ), 1978, pp. 200-202 45. Cassie, A.B.D., and Baxter, S., Wettability of porous s o l i d s , Trans.  Faraday S o c , 40, 1944, pp. 546-551 - 235 -46. Cassie, A.B.D., Contact angles, Disc. Faraday S o c , 3, 1948, pp.11-16 47. Neumann, A.W. and Good, R.J., Thermodynamics of contact angles. I. Heter-ogeneous s o l i d surfaces, J. C o l l . Interface S c i . , 38(2), 1972, pp. 341-358 48. Leja, J. and Poling, G.W., On the interpretation of contact angle, In: Proc. Intern. Miner. Process. Cong., London, 1960, Instn. Min. Met*•': a l l . , London, 1960, pp. 325-341 49. Good, R.J., and Koo, M.N., The effect of drop size on contact angle, J. C o l l . Interface S c i . , 71(2), 1979, pp. 283-292 50. G i r i f a l c o , L.A., and Good, R.J., A theory for the estimation of surface and i n t e r f a c i a l energies. I. Derivation and application to i n t e r -f a c i a l tension, J. Phys. Chem., 61, 1957, pp. 904-909 Good, R.J., G i r i f a l c o , LM., and Kraus, G., i b i d . I I . Application to surface thermodynamics of teflon and graphite, i b i d , 62, 1958, pp. 1418-1421 Good, R.J. and G i r i f a l c o , L.A., i b i d . I I I . Estimation of surface energies of solids from contact angle data, i b i d , 64, 1960, pp. 561-565 51. Good, R.J., and Elbi ng, E., Generalization of theory for estimation of i n t e r f a c i a l energies, In: Chemistry and Physics of Interfaces - I I , ACS,. Wash., D.C., 1971, pp. 72-96 52. Good, R.J., Surface free energy of solids and l i q u i d s : Thermodynamics, molecular forces, and structure, J. C o l l . Interface S c i . , 59(3), 1977, pp. 398-419 53. Fowkes, F.M., Determination of i n t e r f a c i a l tensions, contact angles, and dispersion forces i n surfaces by assuming a d d i t i v i t y of intermolecu-l a r interactions i n surfaces, J. Phys. Chem., 66, 1962, p. 382 54. Fowkes, F.M., A d d i t i v i t y of intermolecular forces at interfaces. I. Determination of the contribution to surface and i n t e r f a c i a l tensions of dispersion forces i n various l i q u i d s , J. Phys. Chem., 67, 1963, pp. 2538-2541 55. Fowkes, F.M., Attractive forces at interfaces-,• Ind. and Engng. Chem., 56(12), 1964, pp. 40-52. •. -56. Fowkes, F.M., Attractive forces at s o l i d - l i q u i d interfaces, In: Wetting, S.C.I. Monograph No. 25, Soc. Chem. Ind., London, 1967, pp. 3-30 57. Hu, P. and Adamson, A.W., Adsorption and contact angle studies. I I . Water and organic substances on polished polytetra fluoroethylene, J. Coll . Interface S c i . , 59, 1977, pp. 605-614 58. Fowkes, F.M., Intermolecular and interatomic forces at interfaces, In: Surface and Interfaces I. Chemical and Physical Characteristics, ed i t s . J.J. Burke, N.L. Reed and V. Weiss, Syracuse Univ. Press, 1967, New York, pp. 197-224 - 236 -59. Fowkes, F.M., Predicting attractive forces at interfaces. Analogy to s o l u b i l i t y parameter, In: Chemistry.and Physics of Interfaces - I I , Amer. Chem. Soc. Pub!., Wash., D.C, 1971, pp. 154-167 60. Kaelble, D,.H., Physical .Chemistry of Adhesion, Wiley-Interscience, New York, 1971 61. Dann, J.R., Forces involved in the adhesion process. I. C r i t i c a l surface tensions of polymeric solids as determined with polar l i q u i d s , J. C o l l . Interface S c i . , 32(2), 1970, pp. 302-320 62. Aveyard, R. and Haydon, D.A., Thermodynamic properties of a l i p h a t i c hydrocarbon/water interfaces, Faraday Soc. Trans., 61, 1965, pp. 2255-2261 63. P h i l l i p s , M.C. and Riddiford, A.C, Contact angles and the free surface energies of s o l i d s , In: Wetting,: SCI Monograph No. 25, Soc. Chem. Ind., London, 1967, pp.31-42; Discussion, pp.42-56 64. Fowkes, F.M., Surface effects of anisotropic London dispersion forces in n-al k a h e s J . Phys . Chem., 84, 1980, pp. 510-512 65. Owens, D.K. and Wendt, R.C., Estimation of the surface free energy of polymers, J. Appl. Polymer S c i . , 13, 1969, pp. 1741-1747 66. Baskin, A., Nishino, M. and Ter-Minassian-Saraga, L., S o l i d - l i q u i d ad-hesion of oxidized polyethylene films. Effect of temperature on polar forces, J. C o l l . Interface S c i . , 59(3), 1977, pp. 516-524 67. Good, W.R., A comparison of contact angle interpretations, J. Coll Interface S c i . , 44(1),1973, pp. 63-71 68. Neumann, A.W., Good#R.J., Hope, C.J., and Sejpal, M., An equation-of-state approach to determine surface tensions of low-energy solids from contact angles, J. C o l l . Interface Sci ., 49(2), 1974, pp. 291-304 69. Fowkes, F.M., McCarthy, D.C. and Mostafa, M.A., Contact angles and the equilibrium spreading pressures of liquids on hydrophobic s o l i d s , J. C o l l . Interface Sci ., 78(1), 1980, pp. 200-206 70. Fowkes, F.M., Dispersion force contributions to surface and i n t e r f a c i a l tensions, contact angles, and heats of immersion, In: Contact Angle, Wettability, and Adhesion, Advances in Chemistry Series No. 43, ACS Wash., D.C, 1964, pp. 99-111 71. Zettlemoyer, A.C, Hydrophobic surfaces, In: Hydrophobic Surfaces, e d i t . F.M. Fowkes, Academic Press, New York, 1969, pp. 1-27 72. Tadros, M.E., Hu, P. and Adamson, A.W., Adsorption and contact angle studies, J . C o l l . Interface S c i . , 49(2), 1974, pp. 184-195 73. Adamson, A.W., An adsorption model for contact angle and spreading, J. Coll . Interface Sci.., 27(2), 1968, pp. 180-187 74. Dettre, R.H. and Johnson, R.E., Surface properties of polymers. I. The surface tensions of some molten polyethylenes, J. C o l l . Interface - 237 -S c i . , 21 , 1966, pp. .367-377 75. Zisman, W.A., Relation of the equilibrium contact angle to l i q u i d and sol i d . c o n s t i t u t i o n , In: Contact Angle, Wettability and Adhesion, Advances in Chem. Series 43, ACS, Wash., D.C., 1964, pp. 1-51 76. Fox, H.W. and Zisman, W.A., The spreading of liquids on low energy sur-faces. I. Polytetrafluoroethylene, J. C o l l . S c i . , 5(6), 1950, pp. 514-531 77. Fox, H.W. and Zisman, W.A., The spreading of liquids on low-energy sur-faces. I I . Modifiedtetrafluoroethylene polymers, J. C o l l . S c i . , 7(2), 1952, pp. 109-121 78. Fox, H.W. and Zisman, W.A., The spreading of liquids on low energy sur-faces. I I I . Hydrocarbon surfaces, J. C o l l . S c i . , 7(4), 1952, pp. 428-442 79. Shafrin, E.G. and Zisman, W.A., The spreading of liquids on low-energy surfaces. IV. Monolayer coatings on platinum, J. Col 1. S c i . , 7(2), 1952, pp. 166-177 80. Schulman, F. and Zisman, W.A., The spreading of liquids on low-energy surfaces. V. Perfluorodecanoic acid monolayers , J. C o l l . S c i . , 7(4), 1952, pp. 465-481 81. Pittman, A.G., Surface properties of fluorocarbon polymers, In: Fluoro- . polymers, e d i t . L.A. Wall, Wiley-Interscience, New York, 1972, pp. 419-449 82. Shafrin, E.G. and Zisman, W.A., Constitutive relations in the wetting of low energy surfaces and the theory of the retraction method of pre-paring monolayers , J. Phys. Chem., 64, 1960, pp. 519-524 83. Wolfram, E., Adhesion of liquids on surfaces of synthetic resins, K o l l o i d - Z e i t . u. Zeit. Polymere, 182(1-2), 1962, pp. 75-85 84. Schonhorn, H., Surface free energy of polymers, J. Phys. Chem., 69, 1965, pp. 1084-1085 85. Rhee, S.K., C r i t i c a l surface tension for spreading in the systems ( A l -Mg)/Graphite and (Cu-0)/Sapphire, J. Amer. Ceram. S o c , 54(8)), 1971 , pp. 37.6-379 86. Rhee, S.K., Surface tension of low-energy s o l i d s , J. C o l l . Interface Sci ., 44(1), 1973, pp. 173-174 87. Wu, S., Surface tension of s o l i d s : An equation of state analysis, JJ'Cpll Interface S c i . , 71(3), 1979, pp. 605-609 88. Johnson, R.E. and Dettre, R.H., The w e t t a b i l i t y of low energy l i q u i d surfaces, J. C o l l . Interface S c i . , 21 , 1966, pp. 610-622 89. Murphy, W.J., Roberts, M.W. and Ross, J.R.H., Contact angle studies of some low energy polymer surfaces, J. Chem. Soc., Faraday Trans. ]_, 68, 1972, pp. 1190-1199 - 238 -90. Baskin, A. and Ter-Minassian-Saraga, L., Wetting of polyethylene by water, methylene iodide and methylene iodide-decalin mixtures, J. C o l l . Inter- face S c i . , 43(1), 1973, pp. 190-202 91. A l l a n , A.J.G., The spreading of liquids on polyethylene f i l m . The effect of preprinting treatments, J. Polymer. S c i . , 38(134)., 1959, pp. 297-306 92. Fox, H.W., Hare, E.F. and Zisman, W.A., Wetting properties of organic liq u i d s on high energy surfaces, J. Phys. Chem., 59, 1955, pp. 1097-1106 93. Olsen, D.A. and Osteraas, A.J., The c r i t i c a l surface tension of glass, J. Phys. Chem., 68(9), 1964, pp. 2730-2732 94. Eberhart, J.G.,The c r i t i c a l surface tension of sapphire, J. Phys. Chem., 71, 1967, pp. 4125-4128 95. Eberhart, J.G., The c r i t i c a l surface tension of uranium dioxide, J. Nucl . Materials, 25, 1968, pp. 105-107 96. Rhee, S.K., Wetting of TaC by l i q u i d copper and l i q u i d Ag., J. Amer. Ceram. S o c , 55, 1972, pp. 157-159 97. Bernett, M.K. and Zisman, W.A., Relation of we t t a b i l i t y by aqueous so-lutions to the surface constitution of low-energy s o l i d s , J. Phys.  Chem., 63, 1959, pp. 1241-1246 98. o Bernett, M.K. and Zisman, W.A., Wetting of low-energy solids by aqueous solutions of highly fluorinated: acids and s a l t s , J. Phys. Chem., 63, 1959, pp. 1911-1916 99. Wolfram, E., New data on the we t t a b i l i t y of polymers, Ko l l o i d - Z e i t . Zeit. Polymere, 211(1-2), 1966, pp. 84-94 (In German) 100. Ueno, M., Tsubaki, K. and Meguro, K., Wettability studies of methyl group rich surface and benzene ring r i c h surface by the aqueous solutions of nonionic surfactants with homogeneous ethylene oxide chain length, In: Proc. Intern. Conf. on C o l l . and Surface S c i . , Budapest, 1975, ed. E. Wolfram, Elsevier S c i e n t i f i c Publ. Co., Amsterdam, 1975, pp. 133-140 101. Padday, J.F., The effect of surfactants on van der Waals forces at a s o l i d - l i q u i d interface, In: Wetting, Soc. Chem. Ind. Monograph 25, SCI, London, 1967, pp. 234-250 102. Padday, J.F., Spreading c o e f f i c i e n t s , wetting energies, and the ad-sorption of surfactants, In: proc. 4th Intern. Cong, on Surface  Active Substances, Brussels, 1964, Gordon and Breach Science Pub-l i s h e r s , New York, 1967, Vol. I I , pp. 299-308 103. Owens, N.F., Richmond, P., Gregory, D.', Mingins, J. and Chan, D., Con-tact angles of pure liquids and surfactants on low-energy surfaces, In: Wetting, Spreading and Adhesion, ed. J.F. Padday, Academic Press, Inc., London, 1978, pp. 127-146 - 239 -104.. Smolders, C.A. Wetting of polymers - effect of surfactant adsorption, In: Chemistry, Physics and Applications of Surface Active Substances, Vol. I I , Gordon and Breach, London,, 1967, pp. 343-349 105. Fowkes, F.M. and Harkin s, W.D., The state of monolayers adsorbed at the interface solid-aqueous solution, J. Amer. Ceram. S o c , 62, 1940 pp. 3377-3386 106. O t t e w i l l , R.H. and Vincent B., Colloid and.surface chemistry of polymer l a t t i c e s . Part 1. Adsorption and wetting behavior of n-alkanols, J.  Chem. Soc,. Faraday Trans. I, 68, 1972, pp. 1533-1543 107. Ginn, M.E., Noyes, CM. and Jungermann, E., The contact angle of water on viable human skin, J. C o l l . Interface S c i . , 26, 1968, pp. 146-151 108. Adamson, A.W., Kunichika, K., Shirley, F. and Orem, M., Dermatometry for coeds, J. Chem. Educ., 45(11),1968, pp. 702-704 109. Lucassen-Reynders, E.H., Contact angles and adsorption orit s o l i d s , J. Phys. Chem., 67, 1963, 969-972 110. Freundlich, H., Colloid and Capillary Chemistry, Methuen, London, 1926 111. B a r t e l l , F.E. and Osterhof, H.J., Determination of the w e t t a b i l i t y of a s o l i d by a l i q u i d , Ind. Eng. Chem., 19, 1927, pp. 1277-1280 112. Schonhorn, H., Dependence of contact angles on temperature: Polar li q u i d s on polypropylene, J. Phys. Chem., 70, 1966, pp. 4086-4087 113. J a r v i s , N.L., Fox, R.B., and Zisman, W.A., Surface a c t i v i t y at organic l i q u i d - a i r interfaces. V. The effect of p a r t i a l l y fluorinated addit- ; ives on the we t t a b i l i t y of s o l i d polymers, In: Contact Angle, Wettab- i l i t y and Adhesion, Adv. Chem. Series No. 43, ACS, Wash., D.C., 1964 pp. 317-331 114. van Voorst Vader, F., Effects of surfactants on. w e t t a b i l i t y , Chem.-Ing.-Tech., 49(6), 1977, pp. 488-493 (In German) 115. Olsen, D.A., Moravec, R.W. and Osteraas, A.J., The c r i t i c a l surface tension values of group VIA elements, J. Phys. Chem., 71(13), 1967, pp. 4464-4466 116. Waksmundzki, A., Neczaj-Hruzewicz', J. and Planik, M., Slime coatings formation and th e i r properties in the sulphur-calcite systems, Trans. Instn. Min. Metal!. (Sect. C), 80(3), 1971 , pp.C200-C203 117. B a r t e l l , F..E. and Zuidema, H.H., Wetting characteristics of solids of low surface tension such as t a l c , waxes, and resins, Amer. Chem.  Soc. J., 58, 1936, pp. 1449-1454 118. Ar b i t e r , N., F u j i i , Y., Hansen, B. and Raja, A., Surface properties of hydrophobic s o l i d s , In: Advances in i n t e r f a c i a l phenomena of partic- ulate/solution/gas systems; applications to f l o t a t i o n research, edits. P. Somasundaran and R.B. Grieves, AI.ChE Series Vol. 71, No. 150, AIChE, New York, 1975, pp. 176-182 - 240 -119. Chander, S.. and Fuerstenau,. D.W., On. the natural f l o a t a b i l i t y of moly-bdenite, Trans. SME/AI.ME, 252(1), 1972 , pp. 62-68 120. Parekh, B.K. and Apian, F.F., The c r i t i c a l surface tension of wetting of coal, In: Recent Developments in Separation Science,, Vol. IV, e d i t . N.N. L i . , CRC Press, Inc., West Palm Beach, Flo r i d a , 1978, pp. 107-113 121. Schrader, M.E., Ultrahigh vacuum techniques in the measurement of con-tact angles. IV. Water on graphite (0001), J- Phys. Chem., 79, 1975, pp. 2508-2515 122. i b i d , 5. LEED study of the effect of structure on the w e t t a b i l i t y of graphite, i b i d , 84(21), 1980, pp. 2774-2779. 123. Bargeman, D. and van Voorst Vader, F., Effect of surfactants on contact angles at nonpolar s o l i d s , J. C o l l . Interface S c i . , 42(3), 1973, pp. 467-472 124. Fowkes, F.M., Role of surface active agents in wetting, J. Phys. Chem., 57, 1953, pp. 98-103 125. Furmidge, C.G.L., Studies at phase interfaces. I. The s l i d i n g of l i q u i d drops on s o l i d surfaces and a theory for spray retention, J. C d l l .  Sci.., 17, 1962, pp. 309-324 126. Wolfram, E. and Boross-Vargha, E., Effect of.preadsorbed surfactant on the w e t t a b i l i t y of polymers, In: Proc. Intern. Confj. on Colloid and  Surface Science, Budapest, ed. E. Wolfram, Elsevier S c i e n t i f i c Publ. Co., Amsterdam, 1975, pp. 163-167 127. Brooks, J.H., Mingins, J. and Owens, N.F., The effect of the film pressure of an insoluble monolayer at .an air/water interface on the wetting of s o l i d surfaces,"J. C o l l . Interface S c i . , 61(2), 1977, pp. 215-222 128. Smolders, C.A., Contact angles; wetting and dewetting of mercury. Part II Theory of wetting, Recueil des Travaux Chimiques des Pays-Bas, 80, 1961, pp. 650-658 129. Finch, J.A. and Smith, G.W., The Liquid-vapour interface in the study of particle-bubble attachment, Can. Metallurgical Quarterly, 11(4), 1972, pp. 569-572 130. L i n , I . J . and Metzner, A., Co-adsorption of para f f i n i c gases in the system quartz-dodecyl ammonium chloride and i t s effect on froth f l o t a t i o n , Intern. J. Miner. Process., 1, 1974, pp. 319-334 131. Bhargava, A., Francis, A.V. and Biswas, A.K., Inte r f a c i a l studies re-lated to the recovery of mineral slimes in a water-hydrocarbon l i q u i d - c o l l e c t o r system, J. C o l l . Interface S c i . , 64(2), 1978, pp. 214-227 132. ASTM D-2578-67, ASTM Standards, Part 26, 1968, p. 370 - 241 -133. Ginn, M.E., Adsorption of cationic surfactants, on mineral substrates, In: Cationic. Surfactaints, Surfactant Science Series, Vol. 4, e d i t . E. Jungermann, Marcel Dekker,. Inc., New York, 1970, pp." 341-367 134. Zimmels, Y., L i n , I.J. and Friend, J.P., Relation between stepwise bulk association and i n t e r f a c i a l phenomena for some aqueous sur-factant solutions, Coll.Polymer S c i . , 253, 1975, pp. 404-421 135. Finch, J.A. and Smith, G.W., Dynamic surface tension of alkaline dod^ ecylamine acetate solutions in oxide f l o t a t i o n , Trans. Instn. Min. Metal!. (Sect. C), 81 , 1972, pp. C213-C218 136. Finch, J.A. and Smith, G.W., Bubble-solid attachment as a function of bubble surface tension, Can. Metallurgical Quarterly, 14(1), 1975, pp. 45-51 137. Leja, J . , Mechanisms of c o l l e c t o r adsorption and dynamic attachment of particles to a i r bubbles as derived from surface chemical studies, Trans. Instn. Min. Metal!., 66, 1956-1957, pp. 425-437 138. Leja, J . , On the mechanisms of surfactant adsorption, In: VIII Intern. Mineral Process Cong.,Leningrad, 1968, Vol. I I , paper S - l , pp. 215-223 139. Parekh, B.K. and Apian, F.F., Wetting properties of minerals coated with surfactants, In: Proc. Chemical Institute of Canada, Flpe- culation and Dispersion Symp., Chemical Inst. Of Canada, Division of Macromolecular Science, Protective Coatings , Toronto, Nov. 1974 pp. 21-35 140. Parekh, B.K., Contact angle studies of minerals coated with surfactants, M.S. Thesis, Pennsylvania State Univ.,-1972 141. Finch, J.A. and Smith, G.W., Interface adsorption densities i n f l o t a -t i o n , J. C o l l . Interface Sci ., 44(2), 1973, pp. 387-388 142. Schulman, J.H. and Leja, J . , Control of contact angles at the a i r -water-solid interfaces. Emulsions s t a b i l i z e d by s o l i d particles (BaS04), Trans. Faraday S o c , 50, 1954, pp. 598-605 143. Elworthy, P.H. and Mysels, K.J., The surface tension of sodium dod-ecylsulfate solutions and the phase separation model of micelle formation, J. C o l l . Interface S c i . , 21, 1966, pp. 331-347 144. de Bruyn, P.L., Overbeek, J.Th.G. and Schuhman, R.J., Flotation and the Gibb's adsorption equation, Min. Engng., 6, 1954, pp. 519-525 145. Digre, M. and Sandvik, K.L., Adsorption of amine on quartz through bubble i n t e r a c t i o n , Trans. Instn. Min. Metal!. (Sect. C), 77, 1968, pp. C61-C64 146. Trahar, W.J. and Warren, L.J., The f l o t a b i l i t y of very fine p a r t i c l e s -a review, Intern. J. Miner. Process., 3, 1976, pp. 103-131, - 242 -147. Schuhmann, R., Flotation k i n e t i c s . I. Methods for steady state study of f l o t a t i o n problems., J. Phys. Chem., 46, 1942, pp. 891-902 148. Tomlinson, H.S. and Fleming, M.G., Flotation rate studies, In: Proc. 6th Intern. Miner. Process. Cong., Cannes, e d i t . A. Roberts, 1963, pp. 563-579 149. Laskowski, J . , Particle-bubble attachment in f l o t a t i o n , Minerals S c i . Engng., 6(4), 1974, pp. 223-235 150. Sutherland, K.L., Physical chemistry of f l o t a t i o n XI. Kinetics of the f l o t a t i o n process , J. Phys. Chem., 52, 1947, pp. 394-425 151. Jameson, G.J.-, Nam, S. and Moo Young, M., Physical factors affecting recovery rates in f l o t a t i o n , Minerals S c i . Engng., 9(3), 1977, pp. 103-118 152. Reay, D. and R a t c l i f f , G.A., Removal of fine particles from water by dispersed a i r flotation:: Effect of bubble size and pa r t i c l e size on c o l l e c t i o n e f f i c i e n c y , Can. J. Chem. Engng., 51, 1973, pp. 178-185 153. Anfruns, F. and Kitchener, J.A., Rate of capture of small particles in f l o t a t i o n , Trans. Instn. Min. Metal!. (Sect. C) 86, 1977, pp.- C9-C15 154. Derjaguin, B.V. and Dukhin, S.S., Theory of small and medium-size pa r t i c l e s , Trans. Instn. Min. Metal!., 70, 1960-61, pp. 221-246 155. Evans, L.F., Bubble-mineral attachment in f l o t a t i o n , Ind. Engng. Chem. 46(11), 1954, pp. 2420-2429 156. Ar b i t e r , N. and Harris, C C , Flotation Kinetics, In: Froth Fl o t a t i o n . 50th Anniversary Volume,edit. - D.W. Fuerstenau, Amer. Inst. Min. Metallurgical and Petrol. Engrs., Inc., New York, 1962, pp. 215 -246 157. Scheludko, A., Tscha.ljowska, SI. and Fabrikant, A., Contact between a gas bubble and a s o l i d surface and froth f l o t a t i o n , In: Thin Liquid :  l a y e r s " and Boundary Layers, Special Discussions of the Faraday S o c , No. 1, 1970, Academic Press, London, 1971, pp. 112-117 158. F l i n t , L.R. and Howarth, W.J., The c o l l i s i o n frequency of small par-t i c l e s with spherical a i r bubbles, Chem, Engng. S c i . , 26, 1971, pp. 1155-1168 159. Mika, T.S. and Fuerstenau, D.W., A microscopic model of the f l o t a t i o n process, In: 8th Intern. Miner. Process.Cong., Leningrad, Vol. I I , Mechanobr, 1969, pp. 246-249 160. Blake, T.D. and.Kitchener, J.A., S t a b i l i t y of aqueous films on hydro-phobic methylated s i l i c a , J J C h j ^ 68, 1972, pp. 1435-1442 161. .Read, A.D. and Kitchener, J.A., Wetting films on s i l i c a , J. C o l l .  Interface S c i . , 30, 1969, pp. 391-398 - 243 -162. Schulze, H.J., S t a b i l i t y of thin l i q u i d films on hydrophobic quartz surfaces', In: Proc. Intern. Conf. on Colloid and Surface Science, Budapest, ed|t. E. Wolfram, Elsevier S c i e n t i f i c Publ . Co., Amsterdam, 1975, pp. 179-186 163. Laskowski, J. and Iskra, J., Role of c a p i l l a r y effects in bubble-par-t i c l e c o l l i s i o n in f l o t a t i o n , Trans. Instn. Min. Metal!. (Sect. C), 79, 1970, pp. C6-C10 164. Eigeles, M.A. and Vol ova, M.l., Kinetic investigation of effect of contact time, temperature and surface condition,on the adhesion of bubbles to. mineral surfaces, In: 5th Intern. Miner. Process. Cong.,  London, I960, Inst. Min. Metal!., London, 1960, pp. 271-284 165. Kirchberg, H. and. Topfer:, E., The mineralization of a i r bubbles in f l o -t a t i o n , In: 7th Intern. Miner. Process. Cong., Vol.1., ed. N. A r b i t e r Gordon and Breach S c i . Publ., Inc., New York, 1964, pp. 157-168 166. Taggart, A.F. and H a s s i a l i s , M.D., S o l u b i l i t y product and bubble attach-ment in f l o t a t i o n , Trans. ATME, 169, 1946, pp. 259-26.5' 167. Lekki, J. and Laskowski, J.., Dynamic interaction in particle-bubble attachment in f l o t a t i o n , In: Colloid.and Interface Science, Vol. IV, ed i t . M. Kerker, Academic Press, Inc., New York, 1976, pp. 331-345 168. Lekki, J. and Laskowski, J., On the dynamic effect of frother-collector j o i n t action in f l o t a t i o n , Trans. Instn. Min. Metal!. (Sect. C), 80(3), 1971, pp. C174-C180 169. Heyes, G.W. and Trailer, W.J., The natural f l o t a t o i l i ty of chalcopyrite, Int. J. Miner. Process.., 4(4), 1977, pp. 317-344 170. Jowett, A., Formation and disruption of particle-bubble aggregates in f l o t a t i o n , In: Fine Particles Processing, Vol, l , : e d i t . P. Somasun-daran, Amer. Inst. Min. M e t a l l u r g i c a l and Petrol. Engrs.', Inc., New York, 1980, pp. 720-754 . 171. Spedden, H.R. and Hannan, W.S., Attachment of mineral particles to a i r . bubbles in f l o t a t i o n . , Trans. AIME, 183, 1949, pp. 208-213 172. Whelan, P.F. and Brown, D.J., Particle-bubble attachment in froth f l o t a t i o n , Trans. Instn. Min. Metal 1., 65, 1955-56, pp. 181-192 173. P h i l i p p o f f , W., Some.dynamic phenomena in f l o t a t i o n , Trans. AIME, 193, 1952, pp. 386-390 174. Scheludko, A.,Toshev, B.V. and.Bojadjiev, D.T., Attachment, of particles to a l i q u i d surface (Capillary theory of f l o t a t i o n ) , J. Chem. S o c , Faraday Trans. I, 72, 1976, pp. 2815-2828 175. Klassen,. V.I., P-laksine,. I .N. and Vlasova, N.S., Theoretical, bases of the action of reagents in, coal f l o t a t i o n , In:. 3rd Intern. Coal Prep. Cong., Brussels-Liege, I n s t i t u t National de 1'industrie Charbonniere, Liege, 1958, pp. 463-468 - 244 -176. Anfruns, J.P.. and Kitchener, J.A,, The absolute rate of capture of single particles..by single bubbles, In: Flotation.A.M..Gaudi.n.Memorial Volume, Vol. 2, ed i t . M.C. .Fuerstenau, Amer. Inst. Min. Metallurgical and Petrol. Engrs., Inc., New York, 1976, pp. 625-637 177. Huh, C. and Mason, S.G.., The f l o t a t i o n of axisymmetric particles at horizontal l i q u i d interfaces, J. C o l l . Interface S c i . , 47(2), 1974, pp. 271-289 178. Schulze, H.J., New theoretical and experimental investigations on sta-^ b i l i t y of bubble/particle aggregates.in.flotation: a theory on the upper p a r t i c l e size of f l o a t a b i l i t y , Intern. J. Miner. Process., 4, 1977, pp. 241-259 179. Morris, T.M., Measurement of equilibrium forces between an a i r bubble and an attached s o l i d in water, Trans. AIME, 187, 1950, pp. 91-95 180. Bogdanov, O.S., Podnek, A.K., Khainman, V.Ya. and Yanis, N.A., Theory and technology of f l o t a t i o n , Trudy Inst i t u t a Mekhanobr, 124, 1959, pp. 29-47 181. Bogdanov, O.S., Emelyanov, M.F., Maximov, I . I . and Otrqzhdennova, L.A., Influence of some factors on fine p a r t i c l e f l o t a t i o n , In: Fine Par- t i c l e s Processing, Vol . 1, ed i t . P. Somasundarart., Amer; Inst.'Min. Metallurgical and Petrol.Engrs., Inc., New York, 1980, pp. 706-719 182. Mackenzie, J.M.W. and Matheson, G.H., Kinetic and dynamic relationships in coal f l o t a t i o n , Trans. SME/AIME, 226, 1963, pp. 68-75 183. Fuerstenau, D.W., Correlation of contact angles, adsorption density, zeta potentials and f l o t a t i o n rate, Min. Engng., 9, 1957, pp. 1365-1367 184. Fuerstenau, D.W.., HeaTy, T.W. and Somasundaran, P., The role of hydro-carbon chain of alkyl collectors in f l o t a t i o n , Trans. SME/AIME, 229, 1964, pp. 321-325 185. Choi, H.S., Kim, Y.S.and Paik, Y.H., Flotation characteristics of ilmenite, CIM Bull ., 60(658), 1967, pp. 217-220 186. Smith, R...W., Coadsorption of dodecylamine ion and molecule on quartz, Trans. AIME, 226, 1963, pp. 427-433 187. Smith, R.W. and L a i , R.W.M.., On the relationship between contact angle and. f l o t a t i o n behavior., Trans. SME/AIME, 235., 1966, pp. 413-418 188. Leja, J . , Interactions at.interfaces in rela t i o n to froth f l o t a t i o n , In: Proc. 2nd Intern. Cong, of Surface A c t i v i t y , London, .1957, ed. J.H. Schulman, Butterworths, London, 1957, Vol. 3, pp. 273-296 189. P l i t t , L.R. and.Flintoff, B.C. Mineral dressing of coal, paper delivered to 81 st Annual General Meeting; of CIM, Montreal, A p r i l 1979, 55 pp. • - 245 -190. Leja, J . , Discussion and authors reply to ref. 187 Trans. SME/AIME, 238, 1967, pp. 189-191 191. S t r e l t s i n , G.S., Reasons for inherent f l o a t a b i l i t y a n d attachment of hydrophobic particles at the air/water interface, Mekhanobr Inst., 1953; Inherent f l o a t a b i l i t y of minerals from the point of view of their structural c h a r a c t e r i s t i c s , 2nd Conf. Mekhanobr Inst. Metr  a l l u r g i z d a t . 1952; as quoted in refs. 22, 118, 119, 204 192. Gaudin, A.M., Miaw, H.L. and Spedden, H.R., Native f l o a t a b i l i t y and crystal structure, In: Proc. 2nd Intern. Cong, of Surface A c t i v i t y , Vol. I l l , e d i t . J.H. Schulman, Butterworths S c i . Publ., London, 1957 pp. 202-219 193. Derjaguin, B.V. and Shukakidse, N.D., Dependence of the f l o a t a b i l i t y of antimonite on the value of zeta-potential, Trans. Instn. Min. Metal!., 70, 1960-1961, pp. 569-574 194. Derjaguin, B.V. and Landau, L.D., Theory of the s t a b i l i t y of strongly charged lyophobic sols and of the adhesion of strongly charged par-t i c l e s in solutions of el e c t r o l y t e s , Acta Physicochim., USSR, 14, 1941, pp. 633-662 195. Verwey, E.J.W. and Overbeek, J.Th.G., Long-distance forces acting between c o l l o i d a l p a r t i c l e s , Trans. Faraday S o c , 42B, 1946, pp. 117-123 196. Jessop, R.R. and Stretton, J.L., Electrokinetic measurements on coal and a c r i t e r i o n for i t s hydrophobicity, Fuel, 69(1), 1969, pp. 317-320 197. Chander, S., Wie, J.M. and Fuerstenau, D.W., On the f l o a t a b i l i t y and sur-face properties of naturally hydrophobic s o l i d s , In: Advances in int e r -f a c i a l phenomena of particulate/solution/gas systems; applications to  f l o t a t i o n research, edits P. Somasundaran and R.B. Grieves, ATChE Series Vol. 71, No. 150, AIChE, New York, 1975, pp. 183-188 198. Hoover, R.M. and Malhotra, D., Emulsion f l o t a t i o n of molybdenite, In: Flotation - A.M. Gaudin Memorial Volume, Vol. I., edit M.C. Fuers-tenau, Amer. Inst. Min. Metallurgical and Petrol. Engrs., Inc., New York, 1976, pp. 485-505 199. Laskowski, J. and Kitchener, J.A., The hydrophilic-hydrophobic t r a n s i t i o n on s i l i c a , J. C o l l . Interface S c i . , 29(4), 1969, pp. 670-679 200. Schulze, H.J. and Cichos, C.H.R., The equilibrium thickness of very thin l i q u i d films on quartz surfaces, Zeit. Physik. Chemie, 251(3/4), 1972, pp. 145-151 201. Leja, J. and Bowman, C.W., Application of thermodynamics to the Ath-abasca Tar Sands, Can. J. Chem. Engng., 46, 1968, pp. 479-481 202. Laskowski, J . , Coal f l o t a t i o n in solutions with a raised concentration of inorganic s a l t s , C o l l i e r y Guardian, Sept. 17, 1965, pp. 361-366 - 246 -203. C o l l i n s , G.L.. and Jameson,. G.J.:, Experiments on the f l o t a t i o n of fine p a r t i c l e s : the Influence of p a r t i c l e . s i z e and charge,, Chem. Engng. Sci ., 31, 1976, pp. 985.-991 204. Fin k e l s t e i n , N.P., A l l i s o n , S.A.., Lov e l l , V.M. and Stewart, B.V., Natural and induced hydrophobicity in sulfide.mineral systems, In: Advances in. interfacial.phenomena of particulate/solution/gas systems; applications to f l o t a t i o n research, edits.. P. Somasundaran and R.B. Grieves, ArChE Series, V o l . 7 1 , No. 150, ATChE, New York, 1975, pp. 165-175 205. Brien, F.B. and Kar, G., An electrophoresis study of the f l o t a t i o n properties of t a l c , The Trend in Engng., Univ. of Washington, 20, Jan. 1968, pp. 8-12 206. Chibowski, E. and Wakmundzki, A., The use of'streaming potential measurements for sulphur/n-alkane-water interface studies, J. C o l l .  Interface S c i . , 64(2), 1978, pp. 380-382 207 . Parriera, H.C. and Schulman,.J.H., Streaming potential measurements on paraffin wax, In: Solid Surfaces and the Gas-Solid Interface, Advances in Chem. Series 33, ACS, Wash., D.C., 1961, pp. 160-171 208. McHardy, J. and Salman, T., Some aspects of the surface chemistry of ta l c f l o t a t i o n , Trans. Instn. Min. Metall. (Sect. C), 83, 1974, pp. C25-C29 209. Van Krevelen, D.W., Coal, Elsevier Pub. Co., Amsterdam, 1961 210. Francis, W., Coal. Its Formation and Composition, 2nd Edit., Edward Arnold (Publ.) Ltd., London, 1961, 806 pp. 211. Stach's Textbook of Coal Petrology, edi t s . E. Stach et a l , Gebruder Borntraeger, B e r l i n , 1975 212. Berkowitz, N., An Introduction to Coal Technology, Academic Press, New York, 1979, 345 pp. 213. Chemistry of Coal Ut i l i z a t i o n . , Supplementary Volume, ed. H.H. Lowry John Wiley and Sons, Inc., New York, 1963 214. Gaseous Fuels; Coal .and Coke; Atmospheric Analysis, Annual Book of ASTM Standards, 1976, Part 26 , 215. International .Handbook of Coal Petrography, Supplement to the 2nd edition, Intern. Committee for Coal Petrology, Centre National de la Recherche Scientifique, Paris, 1971 216. Kessler, M.F., Interpretation of. the chemical composition of bituminous coal macerals., Fuel, 52(3), 1973, pp. 191-197 217. Mazumdar, B.K., Hydrogen in coal: Part 1. Genetic and structural develop-ment, Fuel , 51(4), 1972, pp. 284-289 - 247 -218. Falcon, R.M.S., Coal, in South A f r i c a , Part I I . The application of petrography to the characterization of coal, Minerals S c i . Engng., 10(1'}, 19.78, pp. 28-52 219. Gluskoter, H . J . M i n e r a l matter and.trace elements in coal, In: Trace Elements in Coal, Adv. Chem. Series No. 141, ACS, Wash., D.C, 1975, pp. 1-22 220. 01 Gorman, J.V. and Walker, P.L., Mineral, matter characteristics of some American coals, Fuel, 50(2), 1971, pp. 135-151 221. Glembotskii, V.A., Klassen, V.I. and Plaksin, I.N., Flotation, trans-lated R.E. Hammond, Primary Sources, New York, 1963 222. Gray, D., Inherent mineral matter i n coal and i t s effect upon hydro-genation, Fuel, 57, 1978, pp. 213-216 223. Strehlow,. R.A.., Harris, L.A. and Yust, C.S., Submicron-sized mineral component of v i t r i n i t e , Fuel, 57, 1978, pp. 185-186 224. Edwards, G.E., Assessment of the froth f l o t a t i o n characteristics of a coal, J. Inst. Fuel, 45, 1972, pp. 514-521 225. Yohe, G.R., Oxidation of Coal, I l l i n o i s State Geological Survey RI 207, Urbana, I l l i n o i s , 1958, 51 pp. 226. Jensen, E.J., Melnyk, N., Wood, J.C and Berkowitz, N., The dry oxid-ation of sub-bituminous coal, In: Coal Science, Adv. Chem. Series 55, ACS, Washington, D.C, 1966, pp. 621-642 227. Berkowitz, N., Fryer, J.F., Boleslaw, S., Ignasiak, B.S. and Szladow, A.J., Behavior differences between carboniferous and cretaceous b i t -uminous coals of.similar rank, Fuel, 53(2), 1974, p. 141 228. Ignasiak, B.S., Ignasiak, T.M. and Berkowitz, N., Advances i n coal analysis, Reviews in Anal. Chem., 11(3), 1975, pp. 278-298 229. Speight, J.G., Assessment o f structures in coal by spectroscopic tech-niques, In: Analytical Methods for Coal and Coal Products, Vol. I I , e d i t . C Karr, Academic Press, New York, 1978, pp. 75-101 230. Nandi, B.N., Ciavaglia, L.A. and Montgomery, D.S., The variation of the microhardness and reflectance, of coal under, cond.itions of oxidation simulating weathering, J. Micros., 109(1), 1977, pp. 93-103 , 231. Prado, J.G., Optical.properties of oxidized v i t r i n i t e and e x i n i t e , J. Micros., 109(1), 1977, pp. 85-92 232. Gray, R.J., Rhoades, A.H. and King, D.T., Detection of oxidized coal and the effect, of oxidation on the technological properties, Trans. SME/  AIME, 260, 1976, pp. 334-341 - 248 -233. Ignasiak, B.S., Nandi, B.N... and Montgomery, D.S.;, Oxidation studies on coking coal related to weathering. 1. Chromatographic analysis of pyrolysis gases as a method for studying coal weathering, Fuel, 49, 1970, pp. 214-221 234. Wachowska, H.M.., Nandi, B.N. and Montgomery, D.S., Oxidation studies on coking coal related to weathering. 4. Oxygen linkages influencing the dilatometric properties and the effect of cleavage of ether linkages, Fuel, 53,:1974, pp. 212-219 235. Brown, D.J., Coal f l o t a t i o n , In: Froth Flotation-50th Anniversary Volume, ed i t . D.W. Fuerstenau, Amer. Inst. Min. Metallurgical S P e t r o l . Engnsi. Inc., New York, 1962, pp. 518-538 236. Apian, F.F., Coal Flotation, In: Flotation - A.M. Gaudin Memorial Volume, Vol. 2, e d i t . M.C. Fuerstenau, Amer. Inst. Min. Metallurgical & Petrol. Engrs., Inc., New York, 1976, pp. 1235-1264 237. Zimmerman, R.E.,/Froth Flotation, In: Coal Preparation, 4th Ed., ed i t . J.W. Leonard, Amer. Inst. Min. Metallurgical &.Petrol. Engrs., Inc., New York, 1979, pp. 10-75 to 10-104 238. ,Sun,.S.C, Hypothesis for different f l o a t a b i l i t i e s of coals, carbons, and hydrocarbon minerals, Trans. AIME, 199, 1954, pp. 67-75 239. Horsley, R.M. and' Smith,H.G., Principles of coal f l o t a t i o n , Fuel, 30(3), 1951, pp. 54-63 240. Bailey, R.. .and Gray, V.R., Contact angle measurements of water on coal, J. Appl. Chem., 8, 1958, pp. 197-202 241. Gayle, J.B. and Smelley, A.G. Effects of temperature variations on con-tact angles for coal and related substances, U.S. Bur. Mines, R.I. 5585, 1960, 16 pp. 242. Brady, G..A. and Gauger, A.W., Properties of coal surfaces, Ind. Engng. Chem., 32, 1940, pp. 1599-1604 243. Banerjee, A.,-Kini, :K-:A'.-,Sarkar, G.G. and L a h i r i , A., Fundamental'studies in coal f l o t a t i o n : relation between contact angle, active oxygen con-tent and. natural f l o a t a b i l i t y , Indian J. Appl. Chem., 22(4), 1959, pp. 169-170 244. D e l l , C C , Determination of f r o t h - f l o t a t i o n washability data, J. Inst. Fuel , 30(200),, 1957, pp. 523-526 245. Majumdar, K.K., Contact angle of natural graphite, J. S c i . Ind., Res. (India), 8B(12), 1949, pp. 232-234 246. Nelson, H., The selective separation of super-low ash coal by f l o t a t i o n , In: Coke Oven Manager's Association Yearbook, Coke Oven Manager's Assoc, of Great B r i t a i n , London, 1948, pp. 288-247. Crawford, A., The preparation of ultra-clean coal in Germany, Trans.  Inst. Min. Engrs., I l l , 1951-52, pp. 204-218 -1249 -248. Sun, S.C., Effects, of oxidation of coals on thei r f l o t a t i o n properties, Min. Engng., 6, 1954, pp. 396-401 249. Berkowitz., N., Deterioration of coal durfng storage and. transportation, In: Minerals. Transportation, eds. N.W.'. Kirshenbaum and G.O. A r g a l l , J r . , World Mining, San Francisco, C a l i f o r n i a . Vol.,1 , 1.972, pp. 154-173 250. Burdon, R.G., Booth, R.W. and Mishra, S.K., Factors influencing the selection of processes for the benefication of fine coal, In: Proc.  7th Intern. Coal Prep- Cong., Sydney, edit. A.C. Partridge, 1976, paper E .1 251. Gayle, J.B., Eddy, W.H. and Shotts, R.Q.,, Laboratory investigation of the effect, of oxidation on.coal, f l o t a t i o n , U.S. Bureau of Mines, RI 6620, 1965 252. Iskra, J. and:Laskowski, J . , New p o s s i b i l i t i e s for investigating a i r -oxidation of coal surfaces at low temperatures., Fuel, 46(1), 1967, pp. 5-12 . 253. Wen, W.W.. and Sun, S.C., E.lectrokinetic study on the amine f l o t a t i o n of oxidized coal, Trans. SME/AIME, 262(2), 1977, pp. 174-180 254. Baranov, L.A., Effect of the weakly acidic polyfunctional nature of the surface, of. oxidiized coal. on. i ts f l o a t a b i l i t y by cationic reagents, Khimiya Tverdogo Topliva, 5, 1971, pp. 71-73; Chem. Abst. 76:16462 g 255. M i l l e r , K.J., Flotation study of some refractory coals, U.S. Bureau of Mines Report of Invest. 8224, 1977, 12 pp. 256. Jowett, A., Influence of adsorption of frothing agent on percentage recovery of coal from f l o t a t i o n pulps, J. Inst. Fuel, 31(210), 1958, pp. 313-318 257. Samylin, V.N. and Stel'makh, N.A., Interaction of f l o t a t i o n reagents with an oxidized coal surface, Koks i Khimiya, 6, 1973, pp. 5-8; CA 79:94561 258. Baranov, L.A., Budrina, O.V., Gubaidulin, Sh.Sh. and Romanovskaya, M.V., Assessing, the oxidation of coal by a f l o t a t i o n procedure, Coke and  Chemistry USSR, 10, 1970, pp. 15-18 259. Laskowski,, J. and Konieczny, E., An attempt to the determination of a i r -oxidation of coal, by adsorption of the surfactants, Archiwum Gornictwa, 15(1), 1970, pp. 29-36 260. Jowett, A., El-sinbawy, H. and Smith H.G., Slime coating of coal in f l o t a t i o n pulps, Fuel, 35, 1956, pp. 303-309 261. Campbell, J.A.L. and Sun, S.C., Bituminous coal e l e c t r o k i n e t i c s , Trans. SME/AIME, 247, 1970, pp. 111-114 262. Celik, M.S. and Somasundaran, P., Effect of pretreatments on f l o t a t i o n and e l e c t r o k i n e tic properties of coal, Colloids and Surfaces, 1(1), 1980, pp. 121-124 - 250 -263. Frangiskos, N.Z., Harris, C C . and Jowett, A., The adsorption of frothing aqents on coals, In: Proc. 3rd Intern: Cong, on Surface A c t i v i t y , , Cologne, Vol.: - IV, Verlag der Universitatsdruckerei Mainz GMBH, 1960, p p ; 404-415 ... . 264. Eveson, G.F., Ward, S.G. and. Worthington, F., The adsorption of phenol from aqueous solutions by coals, J. Inst. Fuel, 29, 1956, pp. 540-544 265. Vlasova, N.S., Kushnere.va, V.G. and Barsukova, A.Y^., Effect of the degree of oxidation of coals on the adsorption of n-aliphatic a l -cohols, Koks i Khimi.ya,. 6, 1973, pp. 3-5 (in .Russian), CA 79: 81447 266. Zubkova, Yu. N. and Przhegorlinskaya, R.V., Effect of the structure of alcohols on their adsorption from aqueous solution by Donetsk coals, Khimiya Tverdogo Topliva, 4, 1974, pp. 100-103; Chem Abst. 82:61630 267. Vlasova, NvS., Kushnereva, V.G. and Chepasova, T.P., Effect of the ash content of coals on adsorption of heteropolar reagents during f l o -t a t i o n , Ugol, 2, 197.3, pp. 45-49; Chem. Abst. .79:94558 -268. Basenkova, V.L. and.Zubkova, Yu. N., The mechanism of the adsorption of surface active agents on coals, Sol i d Fuel Chem., 11(2), 1977, pp. 123-125 269. Butuzova, L.F. and Kucher, R.V., Adsorption of hexyl alcohol and cap-roic acid on the micro components of oxidized fat coal, Solid Fuel  Chem., 11(2), 1977, pp. 107-113 270. Eveson, G.F., Ward, S.G. and Worthington, F., Froth f l o t a t i o n of low-rank coal I I . A s t a t i s t i c a l investigation of six factors influencing froth f l o t a t i o n , J. Inst. Fuel, 30(197), 1957, pp. 298-309 271. Szczypa, J . , Neczaj-Hruzewicz, J. and Sablik, J., Some properties of slime coatings in coal-gangue systems, Trans. Instn. Min. Metal!. (Sect. C), 82, 1973, pp. C167-C169 272. Walker, J r . , P.L., Peterson, E.E. and Wright, C C , Surface active agent phenomena in dust abatement, Ind. Engng. Chem., 44(10), 1952, pp. 2389-2393 273. Papic, M.M. and Mclntyre, A.D., Surface active agents evaluated i n Canadian study, Coal Age, 78, June 1973, pp. 85-87 274. Guruswamy, S., Use. of wetting agents to reduce the consumption of water used in. dust suppression in coal washeries, J. Mines, Metals and Fuels,. 25(7), 1977, pp. 191-193 275. G l a n v i l l e , J.0. and Wightman, J.P., The action of wetting agents on coal dust, Fuel, 58(11), 1979, pp. 819-822 276. G l a n v i l l e , J.0. and Wightman, J.P., Wetting of powdered coals by alk-anol-water solutions and other l i q u i d s , Fuel, 59(8), 1980, pp. 557-562 - 251 -277. Marsh, H., The determination of surface areas of coals - some physico-chemical considerations, Fuel, 44(4), 1965, pp. 253-268 278. Addison, C.C., The properties of freshly formed surfaces. Part IV. The influence of chain length, and structure on the s t a t i c and dy-namic surface tensions of aqueous alcohol solutions, J. Chem. S o c , 1945, pp. 98-106 279. Fuerstenau, D.W., Metzger, P.H. and Seele, G.D., How to. use this modified Hallimond tube for better flotation, t e s t i n g , Engng.. Min. J., 158(3), 1957, pp. 93-95 280. Partridge, A.C. and Smith, G.W., Small-sample f l o t a t i o n testing: a new c e l l , Trans. Instn. Min. Metal!. (Sect. C), 80, 1971, pp. C199-C200 281. Standard spe c i f i c a t i o n for reagent water, ASTM D 1193-77, Annual Book of ASTM Standards, Part 31, 1978, pp. 20-22 282. Jasper, J . J . , The surface tension of. pure l i q u i d compounds, J. Phys.-and Chem. Reference Data, 1(4), 1972, pp. 841-1009 283. Handbook .of Chemistry, 8th Edit., e d i t . N.A. Lange, Handbook Publishers, Inc., Sandusky, Ohio, 1952 284. Teo, K.C., Finora, S. and Leja, J., Ide n t i f i c a t i o n of natural oxidation states in surface coaV from the Fording River Deposit, paper pre-sented at the CIM F i r s t Technical Meeting on Western Canadian Coals, Vancouver, B.C., Feb. 14-15, 1980 285. Zar, J.H., B i o s t a t i s t i c a l Analysis, Prentice Hal!, Inc., Englewood C l i f f s , N.J., 1974 286. F r i e d e l , R.A., Infrared i n coal structure research, In: Applied Infrared Spectroscopy, edit. D.N. Kendall, Reinhold Publishing Corp., New York, 1966, pp. 312-343 287. Albers, T..,,Lenart, L. and Oelert, H.H., Spectroscopic studies on chemically reacted (oxidized) coals, Fuel, 53(1), 1974, pp. 47-53 288. Bouwman, R. and Freriks, I.L.C., Low-temperature oxidation of a b i t -uminous coal. Infrared spectroscopic study of samples from a coal stockpile, Fuel, 59.(5), 1980, pp. 315-322 289. Painter, P.C, Snyder, R.W., Pearson, D.E. and Kwong, J., Fourier trans-form infrared study of the variation in the oxidation of a coking coal, Fuel, 59(5), 1980, pp. 282-286 290. Robin, P rL. and Rouxhet, P.G., Contribution of molecular water in the infrared spectra of kerogens and coals, Fuel, 55, 1976, pp. 177-183 291. Painter, P.C.„ Snyder, R.W.., Youtcheff, i J . , Givert, P;'H. and Suhr, N., Analysis of kaolin i t e in coal by infrared spectroscopy, Fuel, 59(5), 1980, pp. 364-366 292. Pearson, D.E. and Kwong, J., Mineral matter as a measure of oxidation of a coking coal, Fuel, 58(1), 1979, pp. 63-66 - 252 -293. Pearson, D.E., The quality of Western.Canadian coking coal, CIM B u l l . , 73(813},, 1980, pp. ,70-84 294. Painter, P.C, Coleman,.M.M., Jenkins, R.G., Whang, P.W. and Wal ker, P.L., Fourier transform infrared study of mineral matter, in coal. A novel method for quantitative mineralogical.analysis, Fuel, 57(6), 1978, pp. 337-344 295 . Davis, A., The reflectance of coal, In: Analytical Methods for Coal and  Coke, Vol. I, e d i t . C. Karr, Academic Press, New York, 1978, pp. 27-81. 296. Bustin, R.M., Oxidation characteristics of some sheared coal seams of the Mist Mountains. Formation, southeastern Canadian C o r d i l l e r a , Geological Survey of Canada, CurrentiRes'earch-, Taper' 80-1B, pp. 249-254 297. McCartney, J.T. and Teichmuller, M., C l a s s i f i c a t i o n of coals-according to degree of c o a l i f i c a t i o n by reflectance of the v i t r i n i t e component, Fuel, 51(1), 1972, pp. 64-68 298. Ting, F.T..C, Petrographic techniques in coal analysis, In: Analytical Methods for Coal and Coal Products,, Vol. I, e d i t , C. Karr, Academic Press, New York, 1978, pp. 3-26 299. Brydson, J.A., Plastics Materials, Newnes-Butterworths, London, 1975 300. Billmeyer, F.W., Textbook of Polymer Science, 2nd Edit., Wiley-Inter-science, New York, 1971 301. Chibowski, E. and Waksmundzki, A., A relationship between the zeta po-t e n t i a l and surface free energy changes of the sulphur/n-heptane water system, J. C o l l . Interface S c i . , 66(2), 1978, pp. 213-219 302. Collon, F.A. and Wilkinson, GV, Advanced inorganic chemistry, 2nd Edit., Interscience Publ., New York, 1966 303. T u l l e r , W.M., Elemental Sulphur, In: The Analytical Chemistry of Sulphur and Its Compounds, Part I, e d i t . J.H. Karchmer, Wiley-Interscience, New York, pp. 1-88 304. Roberts, W.L., Rapp, G.R. and Weber, J., Encyclopedia of Minerals, Van Nostrand Reinhold Cp., New York, 1974 305. Taggart, A.F., Del Guidice, G.R.M., Sadler, A.M. and H a s s i a l i s , M., O i l -a i r separation of non-sulphide and non-rmetal minerals, Trans. AIME, 134, 1939, pp. 180-206 306. Majima, H. and Peters, E., Oxidation' rates of sulfide minerals by aqueous oxidation at,elevated temperatures, Trans. Met. Soc. AIME, 236, 1966, pp. 1409-1413 307. Peters, E., Direct leaching of s u l f i d e s , Met. Trans. B, 71(4), 1976, pp. 505-517 3Q8. Dr: Yarar,\private„communica.tion - 253 -309. Joy, A.S. and Robinson, A.J.., Flotation, In: Recent Progress in Surface Science, Vol. 2, e d i t s . J.F. D a n i e l l i , K.G.A. Pankhurst and A.C. Riddiford, Academic,Press, New York, 1964, pp. 169-260 310. Cavallaro, J.A. and Deurbrouck, A.W., Froth f l o t a t i o n and washability data of various Appalachian coals using the timed release analysis technique, USBM ,.RI 6652, 1965; 48 pp. 311. Belogu, P, and Gru, G., The treatment of slurry by f l o t a t i o n , In: Third Intern.. Coal Prep. Cong., Brussells-Liege,. June 1958., I n s t i t u t Nation-al de I'industrie Charbonniere, Liege, 1958, pp. 469-480 312. F i r t h , B.A., Swanson, A.R.. and N i c o l , S.K., Flotation c i r c u i t s for poorly fl o a t i n g coals, Intern. J. Miner. :Process,;, 5(4), 1979., pp. 321-334 313. B i t t e r , J.H., Conditioning of the pulp and i t s influence upon the results of f l o t a t i o n , In,: Proc. 5th. Intern. Coal.Preparation Cong., P i t t s -burgh, 1966, paper A-3 314. Allum, J.A.E. and Whelan, P.F., Froth f l o t a t i o n of low-rank coals, J. Inst. Fuel, 27(158), 1954, pp. 142-147 315. Waksmundzki, A., Neczaj-Hruzewicz, J. and Planik, M., Mechanism of carry-over of gangue slimes during f l o t a t i o n of sulphur ore, Tran. Instn.  Min. Met a l l . (Sect. C), 81, 1972, pp. C249-C251 316. Corriveau, M.P. and .Schapiro, N., Projecting data from samples, In: Coal Preparation, 4th Ed., ed i t . J.W. Leonard, Amer. Inst. Min a. Metallurgical & Petrol. Engrs., Inc., New York, 1979, pp. 4-1 to 4-56 317. D e l l , C.C., An improved release analysis procedure for determining coal washability, J. Inst. Fuel, 37, 1964, pp. 149-150 318. Schapiro, N., Gray,, R.J. and Eusner, G.R., Recent developments in coal petrography, Proc. Amer. Inst. Min. Engrs., 20, 1961, pp. 89-112 319. Dr. D. Pearson, private.communication 320. Chemical Engineers Handbook, 5th Edit. , eds. R.H. Perry and CH. Chilton McGraw-Hill Book Co., New. York, 1973 321. Fuerstenau, D.W., Chander, S. and. Abouzeid, A.M., The recovery of fine particles by physical separation methods, In: Benefication of Mineral  Fines. Problems and Research Needs, edits. P. Somasundaran and N. Arbi t e r , Amer. Inst. Min. Metallurgical and Petrol . Engrs., Inc., New York, 1979, pp. 3-59 322. Brown, D.J., A photographic study of froth f l o t a t i o n , Sheffield Univ. Fuel Tech. and Chemical Engng. Soc. J., 16, 1965, pp. 22-34 323. Bennett, ,A.J.R., Chapman, W.R. and D e l l , C.C, Studies in froth f l o -t ation of coal, In: 3rd Intern. Coal Prep. Cong., Inst. National de I'industrie Charbonniere, Liege, 1958, paper E l , pp. 452-462 - 254 -324. Reay, D. and R a t e ! i f f , G.A., Experimental testing of the hydrodynamic c o l l i s i o n model of fine p a r t i c l e f l o t a t i o n , Can. J. Chem. Engng., 53, .1975, pp. 481-486 325. Neczaj-Hruzewi.cz, J.., Szczypa, J. and Czarkowski, H., Influence of f l o t a t i o n reagents on formation of gangue slime coatings on coal, Trans. Instn. Min. Metal!..(Section C), 83(817), 1974, pp. C261-C263 326. Handbook of chemistry and physics,. 51st Edit., edit R.C. Weast, The Chemical Rubber Co., Cleveland, Ohio, 1970-1971 327. Laskowski, J . , The f l o t a t i o n of naturally hydrophobic minerals i n so-lutions with a raised concentration of inorganic s a l t s , Proc. S i l - esian Univ. Techno!., Min., No. 16, 1966, pp. 3-93 (Polish); English summary pp. 84-87 328. Lindsay,^L. J., Kaiser- Resources: Ltd.-El kview P.reparatioh Plant, -In:. Milling-Practice in Canada, CIM, Montreal,.1978, pp. 388-390 329. King, R.P., Model for. the design and control of f l o t a t i o n plants, In: Proc. 10th Intern. Symp. on Application of Computer Methods in  the Mineral Industry, eds. M.D.G. Salamon and F.H. Lancaster, S. Afr. Inst. Min. Metal 1.., Johannesburg, 1973, pp. 341-350 330. Hemmings, C.E., Flotation rate interdependence of different size par-t i c l e s . A lead to the slime problem?, Aust. Min., 66(6), 1974, pp. 62-63, 66-68 331. Trahar, W.J., The selective f l o t a t i o n of galena from sphalerite with special reference to the effects of part i c l e s i z e , Intern. J. Miner. Process., 3, 1976., pp. 151-166 332. Bearce, W.E., Mechanics and control in coal f l o t a t i o n , Colorado Scho51 of Mines Quarterly, .56(3), 1961, pp. 371-381 333. Fuerstenau, D.W. and Healy, T.W., Principles of mineral f l o t a t i o n , In:, Adsorptive Bubble Separation Techniques, e d i t . R. Lemlich, Academic Press, Inc., New York, 1972, pp. 91-131 334. Mathieu, G.I. and Bruce, R.W.-, Treatment of talcose molybdenite ores, Canada Dept. of Energy Mines and Resources, Mines Branch, Mineral Processing Div., Report MPI(0) 73-36, 1973, 32 pp. 335. Shirley, J.F., The recovery of by-product molybdenite, An Operations Manual for Island Copper's Supervisory Personnel, Aug. 1971, 38 pp. 336. Stewart, B.V., A preliminary investigation of the f l o t a t i o n of copper-activated sphalerite without the use of c o l l e c t o r s , Nat. Inst.  M e t a l ! R e p o r t 1587, 1973, 13 pp. 337. F u j i i , S., Osawa, Y. and Sugimura, H., Infra-red spectra of Japanese coal: the absorption bands at 3030, 2920 and 1600 cm"1, Fuel, 49(1), 1970, pp. 68-75 - 255 -338. Poling, G.W. and Leja, J., Discussion of paper by Granville, A., Finkel-s t e i n , N.P. and A l l i s o n , S.A., Review of reactions in the f l o t a t i o n system galena-xanthate-oxygen, Trans. Inst. Min. Metal! (Sect. C), 81, 1972, pp. C1-C30; i b i d , 82, 1973, pp. C125-C126 339. Booth, R.B. and Freyberger, W.L., Froths and frothing agents, In: Froth Flotation, 50th Anniversary Volume, ed. D.W. Fuerstenau, Amer. Inst. Min. Metallurgical and Petrol. Engrs., Inc., New York, 1963, pp. 258-276 340. Campbell, J.A.L. and Sun, S.C., Anthracite coal e l e c t r o k i n e t i c s , Trans. SME/AIME, 247, 1970, pp. 120-122 341. Butcher, S.G., Osborne, D.S. and Walker, A.D., The design and operation of coal preparation plants treating highly f r i a b l e western Canadian coals, In: 8th International Coal Prep. Cong., Donetsk, USSR, 1979, paper G-3. 342. Harris, C.C., Flotation machines, In: Flotation. A.M. Gaudin Memorial Volume, Vol. 2, edit. M.C. Fuerstenau, Amer. Inst. Min. Metallurgical and Petrol Engrs., Inc., New York, 1976, pp. 753-815 343. Jacobs, R.N. and Uytterhoeven, J.B., Assignment of the hydroxyl bands in the infrared spectra of zeolites X and Y, J. Chem. S o c , Faraday Trans. I, 69(2), 1973, pp. 359-386 344. Jones, R.N., Modular computer programs for infrared spectrophotometry, In: Infrared, Correlation and Fourier Transform-aSpectroscopy, e d i t s . J.S. Mattson, H.B. Mark and H.C. MacDonald, Marcel Dekker, Inc., New York, 1977, pp. 1-69 345. Apian, F.F. and Fuerstenau, D.W., Principles of nonmetallic mineral f l o t a t i o n , In: Froth F1otation.50th Anniversary Volume, Amer. Inst. Min. Metallurgical and Petrol. Engrs, Inc., New York, 1962, pp. 170-214 346. Fuerstenau, D.W., Interfacial processes in mineral/water systems, Pure and Appl. Chem., 24, 1970, pp. 135-164 347. Burger, J.R., Froth f l o t a t i o n i s on the r i s e , Coal Age, 85(3), 1980, pp. 99-101, 104-105, 108 348. Sheludko, A., Thin l i q u i d f i l m s , Adv. C o l l . Interface S c i . , 1, 1967, pp. 391-464 349. Blake, T.D., Investigation of equilibrium wetting films of n-alkahes on ^-alumina, J. Chem. S o c , Faraday Trans. I, 71, 1975, pp. 192-208. 350. Clunie, J.S.,, Goodman, J.F. and Ingram, B.T., Thin l i q u i d f i l m s , In: Surface and Colloid Science, Vol. 3, edit. E. Matijevic, Wiley-Inter-science, New York, 1971, pp. 167-239 351. Kitchener, J.A., Foams and free l i q u i d f i l m s , In: Recent Progress in Surface Science, Vol. I, edit s . J.F. D a n i e l l i , K.G.A. Pankhurst and A.C. Riddiford, Academic Press, New York, 1964, pp. 51-93 - 256 -352. Lekki, J. and Laskowski, J . , A new.xoncept of frothing in f l o t a t i o n systems and a general c l a s s i f i c a t i o n of f l o t a t i o n frothers, In: Proc. 11th Intern. Miner. Process. Cong., C a g l i a r i , 1975, Univ. C a l i a r i , 1975, pp. 428-448 353. Schulze, H.J. and Gottschalk, G., Investigations of the hydrodynamic interaction between a gas bubble and mineral particles in f l o t a t i o n , In: 13th Intern. Miner. Process. Cong., Warsaw, Vol. I, edit J. Laskowski, Polish S c i e n t i f i c Publishers, 1979, pp. 79-108 354. Sheludko, A., Elements of the theory of f l o t a t i o n , God. Sof. Univ. Khim. Facult. (Annuaire Univ. Sof i a , Facult. Chim.), 56, 1961/62 (1963), pp. 1-11 (Russian) 355. Klepetko, E., Recovery of resin from Utah coal , Trans. AIME, 177, 1948, pp. 433-448 356. Burdon, R.G. and Mishra, S.K., Improving fines f l o t a t i o n recovery, Coal Age, 85(5), 1980, pp. 220-222, 225-227 357. Nimerick, K.H. and Scott, B.E., New method of oxidized coal f l o t a t i o n , Min. Cong. J., 66(9), 1980, pp. 21-22, 37 358. Vlasova, N.S., Golovnin, I.M., Elyashevich, M.G. and Zhuvtyuk, G.V., Main trends of coal preparation developments by f l o t a t i o n in the USSR, In: 6th Intern. Coal Prep. Cong., Paris, 1973, paper IE, 18 pp. 359. Wojcik, W. and B i l i n s k i , B.,. The enrichment of fine particles of coal by aggregation. Part I. Forces of coal grain adhesion to a i r bubbles, Powder Techno!., 26(2), 1980, pp. 111-114 360. Pearson, D.E. and Creaney, S., Reflectance of carbonized v i t r i n i t e s as a measure of oxidation of a coking coal, Fuel, 60(3), 1981, pp. 273-275. 361. Imaizumi, T. and Inoue, T., Kinetic consideration of froth f l o t a t i o n , In: Proc. 6th Intern. Miner. Process. Cong., Cannes, 1963, edit. A. Roberts, Pergamon Press, Oxford, 1965, pp. 581-593 362. Lyon, G.C. and Fewings, J.H., Control of the f l o t a t i o n rate of carb-onaceous material in Mt. Isa chalcopyrite ore, Proc. Australas. Inst.  Min. Metal!., 237, 1971, pp. 41-46 - 257 -APPENDIX A SURFACE TENSION MEASUREMENT The s t a t i c surface tension of the aqueous stock solution was measured by the Wilhelmy Plate attachment method (8,27) using a Cahn RG electrobalance (Ventron Instruments Corp., Paramount, C a l i f o r n i a , Model 2670 Balance Unit, Model 2054 Control Unit), a Cahn Ventron Dynamic Surface Tension Accessory (Model 2880) and a 10 inch chart recorder, (Beckman Instruments Inc., Fullerton C a l i f o r n i a , Model 1005). The Dynamic Surface Tension Accessory was modified by removing the sweeper arms, guide shaft and solution c i r c u l a t i o n tray from the cabinet and f i t t i n g the platform plus adjustable screw mechanism from a Cenco Interfacial Tensiometer through the bottom of the cabinet. This allowed the solution container to be raised or lowered from outside the cabinet. A l l holes in the cabinet were sealed. The sensor plate was a thin rectangular platinum f o i l , 0.06 mm thick (+ 0.003 mm 95% confidence i n t e r v a l , measured by micrometer, .OOOlinch.vernier) and 10.29 mm wide over the bottom 5 mm length (+ 0.01 mm 95% confidence i n t e r -v a l , measured with a cathetometer, The Precision Tool and Instrument Company Ltd., Surrey, England, .01 mm vernier). This resulted in a perimeter of 20.70 mm. The plate was frequently flattened by pressing i t between two glass micro-scope s l i d e s , cleaned in fresh dichromate/concentrated sulphuric acid solution and thoroughly rinsed and stored in single d i s t i l l e d water. Prior to the measurements a sample of the stock solution to be tested was placed in an open dish in the measuring cabinet to saturate the atmosphere. The sample to be tested was placed i n a freshly cleaned oven dried 60 mm dia-meter Pyrex glass c r y s t a l l i z a t i o n dish and positioned on the adjustable plat-form in the cabinet. Before each measurement the plate was flamed to red heat in the oxidiz-ing part of a propane burner flame. I t was then suspended v e r t i c a l l y above the solution surface from the the balance arm by a thin wire so that i t s - 258 -bottom edge was horizontal. The balance was zeroed at i t s null position as per the manufacturer's instructions. The solution container was gradually raised using the screw mechanism un t i l the solution surface just contacted the plate and a wetting meniscus formed. The control unit automatically maintained the balance at the null position (so that the bottom edge of the plate remained in the same plane as the surface of the solution) and recorded the increases in mass, M, on the chart recorder (operated at 0.5 inch/minute). I f the plate i s immersed below the plane of the solution, the force act-ing on i t due to the wetting meniscus i s given by (27) M9 = -lv I«, cosu:0 - VApg. A.l iv ^ 2 where g = 980.98 cm/sec , Ap i s the density difference between the vapour and the solution, I i s the perimeter of the plate, and V is the volume of displaced solution (V = 0 when the bottom edge of the sensor plate i s in the same plane as the solution surface). The method depends on the solution completely wetting the plate (e = 0°) and does not require any correction factors. Spreading of the solution was checked both v i s u a l l y and by gradually withdrawing the plate from the solution and determining the increase in force to detachment. I f the recorded mass increased by more than 1 mg (^  0.5 dyne/cm) as. the plate was withdrawn from the solution, the reading was discarded. I f e was zero or close to zero, cos e would not change s i g n i f i c a n t l y as the plate was withdrawn, and the only increase in measured mass would be due to the negative solution displacement (-V) u n t i l detachment took place; an increase of 1 mg mass corresponds to r a i s i n g the bottom edge of the plate about 1.5 mm above the surface of the solution. A s i g n i f i c a n t increase would occur i n cos e as the plate was withdrawn i f the equilibrium or advancing contact angle was large because e would decrease to the smaller receding contact angle. This would cause a r e l a t i v e l y large increase in the - 259 -measured mass. The measured mass (and therefore calculated surface tension) of single d i s t i l l e d water remained constant over a 15 minute time i n t e r v a l , but i t increased slowly for pure methanol at a rate equivalent to about 0.03 dyne/ cm/min. This was probably due to either the absorption of water vapour (causing the surface tension to increase), and/or evaporation of methanol (resulting in a lowering of the level of the l i q u i d surface r e l a t i v e to the bottom edge of the sensor plate). The measured mass of the aqueous methanol solutions with concentrations $ 20 v o l . % remained constant over a 15 minute time i n t e r v a l . It increased slowly with time for solutions with concentrations ^ 25 v o l . % , the 70 v o l . % solution having the highest rate equivalent to approx-^ imately 0.25 dyne/cm/min. The l i k e l y explanation is the evaporation of meth-anol causing an increase in surface tension and lowering of the level of the solution surface. For the solutions which displayed time dependent behavior the reading was taken after the chart had moved one main di v i s i o n from the i n i t i a l point of attachment (equivalent to a time lapse of about 12 seconds). Two to three minutes elapsed between placing the sample of stock solution in the measuring container and bringing the solution surface into contact with the sensor plate. For the more concentrated methanol solutions, evaporation and the con-sequent increase in measured mass during this period could explain at least some of the discrepancy between the experimental values and the l i t e r a t u r e values ( F i g . 6.1). Table A.l gives the experimental surface tension data for the methanol aqueous stock solutions and the 95% confidence interval for the replicate data (see Appendix C, section C.5 for method of cal c u l a t i o n ) . Table A.l Aqueous methanol solution surface tension data, dyne/cm, 20°C v o l . % 0 2 5 8 10 15 20 25 30 40 50 70 100 mean 7,2, .5 69 .4 64.7 60.1 58 .8 53 .6 49 .7 46 .8 44.1 39 .7 36 .7 31 .0 22 .9 stand, dev. .36 .20 .24 .16 .33 .21 .22 .17 .08 .11 .11 .22 .12 no.of readings 13 5 4 2 5 5 5 3 2 2 4 7 9 95% confidence interval + .21 .25 .38 .41 .26 .27 .42 .18 .20 .10 - 261 -APPENDIX B BULK COAL SAMPLE ANALYTICAL DATA Table B.l Proximate and ultimate analysis data for the No. 5 seam bulk sample and the -595+0 ym size fraction screened from the adit 23 bul k sampl es. Sam pie No.5 seam adit 23 Depth, f t . 0 38 75 111 150 a i r dried (ad) Moisture(M), wt.% 0.4 15.6 2.5 1.4 1.8 1.4 Free Swelling Index (FSI) 3% 0/0 6/6h 7%/7% 8/8 8/8 dry(d), wt.% Ash(A) 19.5 17.4 12.2 11.2 10.5 9.9 Vo l a t i l e Matter(VM) 20.3 29.8 a 25.8 25.9 26.1 26.1 Fixed Carbon(FC)(diff.) 60.2 52.8a 62.0 62.9 63.4 64.0 Total 100.0 100.0 100.0 100.0 100.0 100.0 Total Sulphur(S) . .23 .38 .55 .66 .55 .68 Mineral Matter(MM) 21.2 19.0 13.5 12.4 11.6 11.1 Gross Cal. Value(CV),Btu/lb 9967 13116 13584 13694 13741 dry mineral matter free (dirmr f ) , wt.% Carbon (C) 77.1 87.4 88.6 Hydrogen (H) 4.0 5.1 5.3 Nitrogen (N) 1.3 1.4 1.4 Oxygen (0) ( d i f f . ) 17.6 6.1 4.7 Total 100.0 100.0 100.0 FC b ^dmmf 76.4 65.2a 71.5 71.8 71.7 71.9 CVdmmf B t u / l b b 12275 15124 15476 15463 15411 these values may not be r e l i a b l e due to problems associated with determining the v o l a t i l e matter content of severely weathered coals (214) b MMd = .1.08 A d + .55 S d FCdmmf = ( F C d " - 1 5 Sd> ' 100/(100-MMd) CVdmmf = ( C V d " 5 0 S d } * :i00/(I00 -MM d ) Parr formulae - ASTM D 388 - 66 (1972) (214) (In the ASTM dmmf basis, a l l the sulphur is assumed to be inorganic sulphur, and therefore the organic sulphur i s zero). - 262 -APPENDIX C STATISTICAL METHODS OF ANALYSIS OF ANALYTICAL AND EXPERIMENTAL DATA The purpose of the following s t a t i s t i c a l analysis i s to determine whether the observed differences between the k sample (group) means, X^  ( i = 1 to k), of the various measured parameters of the -210+149ym size f r a c t i o n of the coal samples r e f l e c t real differences in the population means .(p..), or are only due to experimental/analytical error and would therefore be estimates of the same population mean, y. For each measured parameter i n this investigation the effect of only one factor was tested; coal type. Each coal sample therefore represented a level of the factor. The null hypothesis to be tested for each parameter i s that the population means of the data for each of the k coal samples are equal, i . e . H 0 : v i = y 2 = = vh. The s t a t i s t i c a l methods and tables of c r i t i c a l values used in this pres-entation are from Zar(285), for single factor multiple comparisons when k > 2. The procedure of applying a series of two-sample tests (e.g. Student t test) to a l l possible pairs of samples (groups) to analyse a multisample hy-pothesis i s i n v a l i d . Therefore a single factor analysis of variance (ANOV) has been used to test the null hypothesis. I f the null hypothesis was re-jected ANOV ( i . e . the population means were not a l l equal) at the selected level of significance, «, then a Student-Newman-Keuls multiple comparison of means test (SNK) was used to determine which population means were equal or unequal. The'use of these s t a t i s t i c a l tests i s based on the assumption that (a) the underlying population d i s t r i b u t i o n of replicate data for each sample i s 2 2 normal and (b) the variances of the k samples are equal, i . e . a = a= ..= 2 However, some departure from these, requirements can be tolerated without changing s i g n i f i c a n t l y . Small differences in sample s i z e , n. (the number of replicate data for the i t n coal sample) can also be accommodated. - 263 -C.l Testing Normality The f i r s t assumption (null hypothesis) that each sample is from a nor-mally distributed population can be tested with the chi-square goodness-of-fit test. The data range i s divided into m equal class intervals of size a, the t h q interval having mean X q (the arithmetic average of the interval l i m i t s ) . t h The observed frequency of outcomes in the q i n t e r v a l i s f . The chi-square s t a t i s t i c is used to estimate the deviation between f and the expected f r e -quency, F , assuming a normal population d i s t r i b u t i o n : 9 X 2 V J i i (f„ " F J 2 / F 0 C.l m The calculated sample mean:, X.j, and sample variance, s2., (s. i s the sam-2 pie standard deviation) are used as estimates o f y ^ a n d a. for determining F . _ n-j m _ X. = X. ./n. or £ f X /n. . . . .C.2 l j=l i j ' l q=l q q' i s 2 = SS./n. . . . .C.3 T I T m _. m _ M l f q V " (q^l V q ) 2 / n i ] / ( n i " 1 }  n i 2 n i where SS^ . = X^.y - ( X ^ . . ) 2 / n n . , the sample sum of squares. F q = N P(X q) where P(X q) i s the probability of an outcome X^. occurring t h in the q interval of a normal d i s t r i b u t i o n and i s given by P(X q) = P(Z q) - P ( Z q + 1 ) where Z q = [ ( X q + 1 - X q)/2 - X ^ / s , . P(Z q) is the probability that Z q * ECX"^ - X~q)/2 - X ^ ] ^ and can be obtained from normal probability tables. I f y2 < y.1 (where x 2 i s the c r i t i c a l value of the test s t a t i s t i c w.ith l e v e l of significance «, and v = m - 3 degrees of freedom) then the null hypothesis is accepted that the sample is from a normally distributed population. - 264 -C.2 Homogeneity of Variances The second assumption (null hypothesis) that a l l the samples come from populations with identical variances, can be tested with Bartlett's corrected test s t a t i s t i c B = B/D c k k where B = (In s 2 ) ^ v.) - v. In . . . . . .C.4 k k D = 1. + . [ l / 3 ( k - l ) ] [ i i 1 1/v.' r l / . Z j v.] . . . . ..C.5 k k and v. = n. - 1 degrees of freedom, and s 2 = SS./.^v., the pooled i i ' p i=l i . i=l i error sample variance (an estimate of the common population variance,a 2). C.3 Single Factor. Analysis of Variance (ANOV) ANOV tests the null hypothesis H q: y a = y 2= . . . = y k . Table C.l summarizes the source of variation and the ANOV calculations. The mean squared deviations from the mean (mean square, MS) are estimates of variance. I f the null hypothesis i s true then the groups MS, and the error MS, w i l l both be estimates of a 2 , the variance common to a l l k populations. The error MS (error variance) is the best estimate. I f the k populations are not equal then the groups MS w i l l be greater than the error MS. The test for the equality of means is therefore a one-sided variance ra t i o (F) test:-P_ groups MS error MS C.6 I f ^ * F<*(1) k 1 N k ^ t h e c r l t i c a ^ value), then the null hypothesis, H q, i s rejected, in which case i t i s concluded that a l l the k population means are not equal. C.4 SNK Multiple Comparison Test The SNK test is applied only i f the null hypothesis has been rejected by ANOV1". It examines the null hypothesis H q: y^ = y^ where A and B denote any 1 I t i s possible for the null hypothesis H Q : ^ =^2 = . . . % , t o be rejected by ANOV, but for the subsequent SNK procedure to f a i l to detect differences between any pair of means. This occurrence w i l l not be commonly encountered, but i t re-f l e c t s the fact that the analysis of variance i s a more powerful test, than the multiple comparison test; i.e. Type II errors (not rejecting the null hy-pothesis when i t i s in fact false) are more l i k e l y to occur in multiple com-parison testing than in ANOV. - 265 -Table C.l Summary of source and variance calculation for single factor ANOV Source of Variation Sum of Squares Degrees of Mean Square (SS) Freedom (DF) (MS) Total (X.j-X) k n i ( v a r i a b i l i t y of a l l N z E X 2 -C N-l data about total mean) i = l j=l i j Groups ( X r X ) ( v a r i a b i l i t y of k group means about total mean) z (i X..)2/nrC i=l J'=1 k-1 Groups SS Groups DF Error (X-.-X.) (pooled v a r i a b l i l t y of data within each group) Total SS-Group SS N-k Error SS Error DF N = £ n. x. . i s j datum in group(sample)i i=l 1 J k n-j X = E E X- -/N - total mean i=l j=l J k n i C = ( E E X- •) /N - correction term i=l.j=l 1 J - 266 -possible pair of groups (samples). The number of possible pairs i s k( k - l ) / 2 . The sample means, J.., are ranked in order of increasing magnitude and the difference between the means for a l l pairs, X g - X A, is tabulated. The calculated tfest s t a t i s t i c for each pair i s q = (X B - XA.)/SE . . . • C.7 where SE =/ [ s 2 / 2 ( l / n A + l / n B ) ] • • • • C.8 and s 2 = error MS from ANOV The c r i t i c a l values of the test s t a t i s t i c are "Studentized ranqes", q where v = error DF from ANOV and p = (B-A)+l (the number of means in the range of means being tested). The significance l e v e l , i s the probability of encountering at least one Type I error ( i . e . rejecting the null hypothesis . when i t i s in fact true) during the course of comparing a l l the pairs of means. The conclusions reached by multiple range testing are dependent upon the order in which the pairwise comparisons are considered. The proper pro-cedure i s to compare f i r s t the largest mean against the smallest etc., u n t i l the largest has been compared with the second largest. Then the second l a r -gest is compared with the smallest etc. I f no difference i s found between two means, then i t is concluded that no difference exists between any means enclosed by these two and such d i f f e r -ences are not tested for. C.5 Confidence Interval The probability that the true mean, y, of a population w i l l occur with-in s p e c i f i c l i m i t s , when y and a2 are unknown but estimated by the sample mean and, variance X and s 2 , i s given by P (X - ts//n * y s X + ts//n) = 1 - <* where t i s the c r i t i c a l value of the two-tailed "Student-t" s t a t i s t i c , tx(2) v> v = n-1 degrees of freedom and s is the sample standard deviation. - 267 -The 95% confidence i n t e r v a l , « =0.05, i s therefore given by X + ts//n. C.6 Correlation Coefficient The correlation c o e f f i c i e n t , c i i s a s t a t i s t i c a l parameter which measures the degree of l i n e a r i t y between two discrete random variables X and Y, and i s given by c = s /s s C.9 xy x y where the sample covariance is s x y = l / ( n - l ) • (EXn. Yn. - l z X i Y . ) and s x and $' are the sample standard deviations for the X and Y data (equation C.3). c varies between -1 and 1, having an absolute value of one i f X and Y are completely l i n e a r l y related. I t i s positive i f X and Y are d i r e c t l y related and negative i f they are inversely related. Zero correlation (c=0) does not imply independence of X and Y, but rather the lack of any l i n e a r connection between them. - 268 -APPENDIX D STATISTICAL ANALYSIS OF C, H AND N ANALYSIS DATA The following tables l i s t the replicate sample data for the -210+149jjm size fraction of the 38', 75',,111' and 150' adit coal samples, and the No. 5 seam coal sample, as well as tests for homogeneity of variances and single factor ANOV and SNK multiple comparison of means tests described in Appendix C. A l l sample populations are assumed to be normally distributed. The data are arranged in the tables so that the sample means increase from l e f t to rig h t . 269 Table D.l S t a t i s t i c a l analysis of carbon data (wt.% dmmf) Sample adit 23 No. 5 seam 38' 75' 150' 111' Group 1 2 3^  4 5 Sum(E) 86.64 85.39 87.23 87.50 88.17 86.95 87.96 88.71 87.89 87.46 88.04 88.66 88.54 88.90 89.44 90.39 89.70 . 89.86 91.10 n i X i 4 86.69 4 87.95 4 88.01 3 88.96 4 90.26 19 * x i j x x i j 2 ( E X i j ) 2 / n i 346.76 30063.26 30060.62 .94 351.79 30940.68 30939.05 .74 352.05 30985.54 30984.80 .50 266.88 361.05 23742.06 32590.47 23741.64 .45 32589,. 28 .63 1678.53 148322.01 148315.40 (a) Homogeneity of variances B = 4.60 X (b) ANOV 6.05,4 H 9.49 2 0.05 1.15 Accept H B„ = 4.01 c - - 0.05 C •= 148287.52 Source SS DF MS Total Groups Error 34.49 27.88 6.61 18 4 14 6.97 .47 14.75 F0.05(l),4,14 3 , 1 1 Reject H, i.e. the c ^ r \ m f population means are not a l l equal Continued - 270 -Table D.l (cont.) ( C ) SNK H Q : y B = y A - = 0.05 Comparison B vs A Difference XB " ^A 9 P ' q0.05,14,P Conclusion 5 vs 1 3.57 10.39 5 4.41 Reject H 0 5 vs 2 2.31 6.72 4 4.11 ii 5 vs 3 2.25 6.55 3 3.70 n 5 vs 4 1.30 3.50 2 3.03 n 4 vs 1 2.27 6.12 4 4.11 II 4 vs 2 1.01 2.72 3 3.70 Accept H 0 4 vs 3 .95 II 3 vs 1 1.32 3.84 3 3.70 Reject H 0 3 vs 2 .06 .18 2 3.03 Accept H 0 2 vs 1 1.26 3.67 2 3.03 Reject H 0 The overall conclusion i s . y < y 2 = V3  = Mk < y 5 , i.e. t h e %Cdmmf population means of the 75', 1111 and 150' adit samples are not s t a t i s t i c a l l y d i f f e r e n t , but the No. 5 seam sample i s larger, while the 38' sample i s smaller. - 271 -Table D.2 S t a t i s t i c a l analysis of hydrogen, data (wt.% dmmf) Sample No.5 seam adit 23 38' 150' 75' 111' Group 1 2 3 4 5 Sum(z) 4.51 4.70 4.62 4.70 4.89 4.76 4.78 4.56 4.99 5.00 4.92 5.03 5.14 4.96 5.02 5.02 5.23 4.98 5.01 n i 4 4 4 4 3 19 Ji 4.63 4.75 4.99 5.04 5.07 * X i j 18.53 18.99 19.94 20.14 15.22 92.82 ( E X i j ) V n i  s i 85.84 .09 90.16 .14 99.40 .05 101.40 .08 77.22 .14 454.02 (a) Homogeneity of variances H 0 : a\ = . = 2 - = 0.05 B = 3.39 D = 1 .15 B c = 2.95 Y1 A0.05,4 =9 i.49 Accept H Q (b) ANOV 0 >i : . . . cc = 0. .05 . C = 453.45 Source SS DF MS F Total Groups Error .71 .57 .14 18 4 14 .14 .01 14.04 F0.05(l),4,14 3 , 1 1 Reject H i.e. the %H dmmf population means are not a l l equal Continued. - 272 -Table D.2 (cont.) (c) : SNK = yA - = 0,05 Comparison Di fference q • P q0.05,14,p i Conclusion B vs A h ~h 5 vs 1 .44 8;,i l 5 4.4.4V Reject H Q 5 vs 2 .32 5.90 4 4.11 5 vs 3 .08 1.47 3 3.70 Accept H Q 5 vs 4 .03 ii 4 vs 1 .41 8.16 4 4.11 Reject H Q 4 vs 2 .29 5.78 3 3.70 4 vs 3 .05 1.00 2 3.03 Accept H 0 3 vs 1 .36 7.17 3 3.70 Reject H Q 3 vs 2 .24 4.78 2 3.03 2 vs 1 .12 2.39 2 3.03 Accept H 0 The overall conclusion i s y = y 2 f y 3 = y^ = y^, i.e. there i s no s t a t i s -t i c a l l y s i g n i f i c a n t difference between the % H d m m f population means of the 75', 111' and 150' adit samples, but those of the No. 5 seam and 38' adit samples are smaller. - 273 -Table D.3 S t a t i s t i c a l analysis of nitrogen data (wt.% dmmf) Sample No. 5 seam adi t 23 111' ,75' 38' ,150' Group 1 2 3 4' 5 Sum(E) 1.49 1.73 1.70 1.65 1.46 1.51 1.54 1.69 1.74 1.76 1.61 1.52 1.53 1.83 3 2 3 3 3 14 *, 1.54 1.64 1.64 1.64 1.68 4.61 3.27 4.91 4.92 5.05 22.76 7.09 5.36 8.06 8.09 8.58 37.18 <«,//», 7.08 5.35 8.04 8.07 8.50 37.04 s i .06 .13 .10 .11 .20 (a) Homogeneity of variances hL : cr 0 2 ^ ... - a 2 ' cc 5 = 0.05 B = 2.59 x0.05,4 = 9.49 D = 1.24 B c Accept H Q = 2.09 (b) ANOV Ho : \ = • . . . = y 5 - = 0.05 C = 37.00 Source SS DF MS F Total .18 13 Groups .04 4 .01 0.50 Error .14 9. .02 F0.05(1) 4,9 " 3 ' 6 3 Accept H Q i . e . there i s no s t a t i s t i c a l l y s i g n i f i c a n t difference: between a l l the %Ndmmf P ° P u 1 a t i o n reans - 274 -APPENDIX E STATISTICAL ANALYSIS OF. ASH CONTENT DATA The floats and rejects products from each small-scale f l o t a t i o n test were analysed for ash content. With the corresponding wt.% y i e l d , the ash content of the feed coal sample for each test.was back-calculated. These data are given in Appendix J. The calculated feed ash content data for the 75', 111' and 150' adit samples and the No. 5 seam sample have been analysed for d i s t r i -bution and v a r i a b i l i t y and the mean calculated ash content compared with the measured feed ash content of the -210+149ym size fraction of each coal sample on an a i r dried basis (the measured values in Table 6.1 are on a dry basis). ANOV and SNK tests (Appendix C) were performed to determine whether the ob-served differences between the mean calculated feed ash contents of the 75', 111' and 150' adit samples were s t a t i s t i c a l l y s i g n i f i c a n t at the 95% co n f i -dence l e v e l . Table E.l contains chi-square goodness-of-fit tests to assess whether the calculated feed ash contents come from a normally distributed population. Table E.2 l i s t s the measured feed ash content, and the mean, standard deviation, range and 95% confidence interval for the calculate feed ash content data for . the 75', 111' and 150' adit samples and the No.5 seam sample. I t also contains homogeneity of variance t e s t s , ANOV and SNK multiple comparison of means tests for the 75', 1111 and 150' adit samples. - 275 -TABLE E.l Chi-square goodness-of-fit test of normality of calculated coal feed ash contents (wt. % a i r dried) Ho: sample comes from a normally distributed population «"= 0.05 (a) No. 5 Seam n = 61 m = 10 a = 0.2 X = 19.66 s = 0.52 Interval X q f q P(Z q) P(X q) F q ( fq~ Fq 0) 1.000 .039 2.39) )7 )5.30 .55 18.9 7) .961 .048 2.91) 19.1 1 .913 .082 4.98 3.18 19.3 11 .831 .119 7.27 1.91 19.5 11 .712 .148 9.07 .41 19.7 10 .564 .159 9.66 .01 19.9 6 .405 .144 8.79 .89 20.1 5 .261 .112 6.83 .49 20.3 3 .149 .074 4.53 .52 20.5 3 .075 .042 2.56 .08 20.7 4) .033 .021 1.24) )4 )2.00 2.00 0) .012 .012 .76) Sum (E) 1.000 61.00 10.04 X 2 = 1.0.04 x 2 = 1 4 - 0 7 D o n o t reject H 0 0.05,7 Continued - 276 -Table E.l (cont.) (b) 75' Adit Sample n = 23 m = 6 a = 0.2 X = 11.04 S = 0.29 Interval x q f q P ( Z q ) P(Xq) F q (fq-Fq)?-F q <10.6 10.7 0) )6 6) 1.000 .939 .061 .134 1.39) )4.48 3.09) ^52 10.9 4 .805 .238 5.46 .39 11.1 7 .567 .265 6.12 .13 11.3 3 .302 .186 4.29 .39 11.5 2 .115 .087 1.97 .00 >1T.8 11.7 1) )1 0) .029 .005 .024 .005 .56) )0.67 • 11) .16 Sum (E) 1.000 22.99 1.59 X 2 = 1.59 X 2 Q Q 5 3 = 7.82 Do not reject H 0 Continued - 277 -Table E.I (cont.) (c) 111' Adit Sample n = 17 m = 6 a = 0.2 X = 9.88 s = 0.32 Interval xq fq PCZq) P(xq) <9.2 9.3 0) )2 2 > 1.000 .977 .023 .061 .40) )1.43 1.03) .23 9.5 2 .916 .134 2.28 .03 9.7 3 .782 .211 3.59 .10 9.9 3 .571 .234 3.98 .24 10.1 4 .337 .183 3.12 .25 >10.4 10.3 3) )3 0) .154 .053 .101 .053 1.72) )2.61 .89) .06 Sum (E) 1.000 17.00 .91 X 2 = 0.91. X 2 0 0 5 j 3 = 7.82 Do not reject H 0 Continued - 278 -Table E.l (cont.) (d) 150' Adit Sample n = 16 m = 6 a = 0.2 X = 9.45 s = 0.31 Interval X q P(Z q) "<V (VV! <8.8 8.9 0) )2 2) 1.000 .983 .017 .054 • 27) )1.14 .87) .65 9.1 1 .929 .135 2.16 .62 9.3 3 .794 .230 3.68 .13 9.5 5 .564 .252 4.02 .24 9.7 3 .312 .187 2.99 .00 >10.0 9.9 2) )2 0) .125 .036 .089 .036 1.43) )2.00 .57) .00 Sum (E) 1.000 16.00 1.64 X 2 •= 1.64 x 2 = 7.82 Do not reject H 0'05>3 - 279 -TABLE E.2 Homogeneity of variances, single factor ANOV and SNK multiple comparison of means tests on calculated feed ash content data (wt. % a i r dried), -210 + 149ym size fraction Sample 150' Adit 1111 23 75' No.5 Seam Group 1 2 3 Total Measured feed ash 9.49 10.01 10.98 19.73 Mean c a l . _ feed ash Xi 9.45 9.88 11.04 19.66 n. i 16 17 23 56 61 s. I 0.31 0.32 0.29 0.52 95% confidence interval ±0.17 ±0.16 ±0.13 ±0.13 range 8.91-9.91 9.29-10.37 10.61-11. 72 18.80-20.76 151.17 167.92 253.85 572.94 1429.68 1660.32 2803.64 5893.64 1428.27 1658.65 2801.73 5888.66 (a) Homogeneity of Variances HQ: al - aj = cf <* = 0. 05 B = 0.15 D = 1.03 Be = 0.15 Xo .0 5,2 = : 5.99 Accept H 0 Continued - 280 -Table E.2 (cont.) (b) ANOV (c) SNK H0: y i = y 2 = oc = 0.05 C = 5861.79 Source SS DF MS F Total 31.85 55 Groups 26.87 2 13.43 143.07 Error 4.98 53 0.09 F0.05(1 ), 2,50 = 1 9 , 5 Reject H 0 i.e. the calculated feed % ash populations means are not a l l equal. H0: y B = y A oc = 0.05 Comparison B vs. A Difference X B X q0.05,60,p Conclusion A 3 vs. 1 3 vs. 2 2 vs. 1 1.59 1.16 0.43 22.55 16.74 5.70 3.40 2.83 2.83 Reject H 0 ii ii ii II The overall conclusion i s p i ^ u 2 ^ y 3 . i.e. the calculated feed % ash population means of the 75', 111' and 150' adit samples are a l l s i g n i f i c a n t l y d i f f e r e n t . - 281 -APPENDIX F INFRARED SPECTROSCOPY OF COAL SAMPLES The KBr pellets for obtaining infrared (IR) spectra of the coal samples were prepared following procedures developed in the Department of Mining and Mineral Process Engineering. F.l Sample Preparation 1. Representative 1 g samples of the -210+149ym size fraction of each coal sample were b r i e f l y ground by hand in an agate mortar, placed in polyethylene weigh boats and stored in a dessicator over dessicant for 24 hours. They were then transferred to stainless steel v i a l s with two stainless steel balls (previously cleaned by repeated shaking with small amounts of KBr, and washed and oven dried) and shaken for 60 minutes on a Spex m i l l (Spex Industries, Inc., Scotch Plains, New Jersey). 2. 100 mg + 1 mg of each pulverized coal sample was made up to 1000 mg + 1 mg with KBr (infrared quality, The Hawshaw Chemical Co., stored in an ovennat ^ 105°C) and the 1:10 mixture shaken on the Spex m i l l for 10 minutes. 3. 50 mg + 1 mg of th i s pulverized 1:10 coal -KBr mixture was made up to 1000 mg + 1 mg with KBr and the resulting 1:200 mixture shaken on the Spex m i l l for 10 minutes. The coal -KBr mixtures for a l l the coal samples were prepared concurrently and stored in a dessicator over dessicant u n t i l further use. F.2 Pel l e t Preparation One p e l l e t at a time was prepared and i t s spectra immediately recorded. 1. 300 mg +1 mg of the pulverized 1:200 coal -KBr mixture was placed on a small watchglass and heated in an oven (at ^  105°C) for 5 minutes (to minimize st i c k i n g of the resulting pellet to the d i e ) . 2. The sample was transferred to a polished stainless steel vacuum die (Perkin Elmer, previously cleaned with acetone) and evenly spread and l i g h t l y tamped with a stainless steel spatula before the plunger was inserted. 282 3. The die was placed in the press and vacuum applied. 4. Pressure of about 4800 psi was applied for about 20 seconds and released, this step being performed three times. 5. The resulting p e l l e t (concentration 1.5 mg coal per 300 mg, cross-sectional 2 area 1.33 cm ) was removed from the die and placed in a Wilks Mini Cell Holder. F.3 Recording the Spectra Immediately after preparation, each KBr pe l l e t was placed in the sample beam of a Perkin Elmer double-beam infrared spectrophotometer, Model 283 B, on which the zero and 100% Transmission (T) baseline had been set following normal manufacturer's instructions. The reference beam was attenuated so that the maximum %T spectra baseline at 1800 cm"! was at about 95%. The gain was then adjusted to 20% deflection and the %T and the Linear Absorption (Abs) spectra recorded with the following instrument settings:-s l i t program 7(max. s l i t width) scan time 12 mins. abscissa (wavenumber) expansion IX ordinate (%T or Abs) " IX " " response 1 F.4 Extracting Absorption Data from Spectra Wherever possible the absorption data were obtained from l i n e a r absorption spectra by subtracting baseline absorption from peak (band) absorption. In those instances where the linear absorption peaks or baselines were off-scale at the standard instrument settings then absorption data were derived from the %T specra using the Beer Lambert relationship (in modified form):-Abs. = "log^Q %T(baseline) - log-^Q % T(peak) F.l The baselines are i l l u s t r a t e d in Fig. 6.4 and are similar to those developed by Friedel as outlined by F u j i i et a l l (337). The anchor points for the two baselines used are 3740 to 2350 cm - 1, and 1800 to 840 cm"1. - 283 -The data from duplicate pellets for several absorption bands are l i s t e d in Table F.l for the -210+149 ym size fraction samples. Also included are data derived from spectra discussed by Teo et al (284) obtained from KBr pellets of the -595+Oym size fraction of the adit samples, following the same procedures outlined above. Table F.2 gives the organic structures generally assigned to the absorp-tion bands. TABLE F.l Absorption data for spec i f i c bands from infrared linear absorption spectra of coal samples Size Coal Depth, wavenumber cm"1 Fraction Sample f t . 3430 3050 2920 2860 1700 1600 1440 1375 1260 1095 1030 1005 0 1 .094* . 2 9 6 * . 2 6 8 * . 1 8 3 * . 4 8 3 * 1 .146* . 4 7 1 * . 5 6 1 * . 3 6 8 * . 4 8 4 * . 6 0 5 * . 5 3 8 * .807 .262 .239 . 165 .436 1 .120* .425 .517 .339 .464 .576 .507 38 .430 . 130 .279 .162 .198 .538 .348 . 282 .238 .420 . 653 .579 .427 .129 .285 .168 .179 .532 . 333 .279 .238 .447 . 620 .566 Adit 23 75 .305 .108 .265 .156 .087 .427 .316 .226 .205 .393 . 560 .503 E 2. .307 .109 .278 .159 .088 .427 .321 .233 .203 .399 .575 .509 + 111 .361 .103 .255 .150 .083 .432 .301 .221 . 2 0 0 .312 . 4 4 3 .392 o r — CXI .378 .1.11 .273 .166 .084 . 450 . 320 .234 . 208 .336 .464 .418 150 .462 .122 .284 .166 .104 .488 . 324 . 2 4 5 .215 .355 . 495 . 443 .382 .111 .263 .153 .091 .456 . 318 .231 .209 .336 .489 .441 No. 5 Seam .448 .110 .205 .125 .084 .399 .271 .216 .188 .675 1 .100* 1 .017* .440 .106 .191 .119 .083 .400 .269 .212 .181 .633 1 .029* 0 . 9 6 7 * ** 0 .230 .164 .445 >1 E •p. o + Adit 23 38 75 .257 . 270 .156 .155 .180 .125 . 720 . 570 LO cn LO i 111 150 .268 .297 . 160 .158 .095 .115 . 625 .605 * Absorption derived from % Transmission spectra ** Data taken from spectra discussed in ref. 284 TABLE F.2 Organic matter structure assigned to absorption bands in infrared spectra of coal (from references 286, 212, 229 ) Band cm"1 Assignment 3700-2500 -OH (stretching) from a l l sources; -NH (stretching) 3050 aromatic C-H (stretching) 2920 a l i p h a t i c C-H (stretching), CH3, CH2 2860 al i p h a t i c C-H (stretching), CH 1700 carbonyl C=0 (stretching) 1600 aromatic C=C (stretching); hydrogen bonded or chelated carbonyl, C=0 HQ-electron transfer between aromatic sheets; noncrystalline pseudographitic structures 1440 aromatic C=C (stretching); CH3 (asym. deformation), CH2 (scissor deformation 1375 CH3 (sym. deformation); c y c l i c CH2 1300-1000 phenolic and alcoholic C-0 (stretching); a l l types of ether C-0-C (stretching) 900-700 out-of-plane vibrations of aromatic C-H groups - 286 -APPENDIX G TREATMENT OF RAW DATA FROM PETROGRAPHIC ANALYSES Table G.l l i s t s the raw maceral analysis data obtained by the techniques described in section 6.3,6. In those samples where the micrinite maceral has not been recorded, i t was included in the f u s i n i t e value. The lack of agree-ment sometimes observed between the amounts of semifusinite, macrinite, mic-r i n i t e and f u s i n i t e (the " i n e r t " macerals) from the duplicate sample and from the two different size fractions of the same coal sample, i l l u s t r a t e the d i f -f i c u l t i e s often encountered in distinguishing c l e a r l y between these four mac-erals, particular with fine coal p a r t i c l e sizes. However, in terms of the sum of i n e r t i n i t e group maceral (macrinite, m i c r i n i t e and fu s i n i t e ) and semi-f u s i n i t e , the analyses on the same coal sample are in reasonable agreement. The proportion of mineral matter d i r e c t l y determined during a maceral analysis i s not always very accurate because of d i f f i c u l t i e s in identifying certain mineral matter with incident l i g h t using o i l immersion, and i t s fine dissemination in many coals. Therefore i t has been recommended that the amount of mineral matter on a v o l . % basis be calculated using the formula(298) c a l . mineral matter, v o l . % = 100 MMd/2.8 . . . . G.l (100-MMd)/1.35 + MMd/2.8 MMd is the wt.% mineral matter (dry basis) given in Table 6.1 and Appendix B (which in turn has been calculated from the standard ash and tota l suphur con-tents with the Parr formula). The MM (vol.%) values l i s t e d in Table 6.2 have been determined in this manner and the r e l a t i v e proportions of e x i n i t e , v i t -r i n i t e , i n e r t i n i t e , semifusinite and heavily oxidized v i t r i n i t e correspondingly recalculated. A column of "total inerts" i s also included in Table 6.2. It is an em-p i r i c a l parameter used to quantify the "non-reactive" components in a coal ( i . e . those components which do not melt and/or swell during heating of the coal). Macrinite, micrinite and f u s i n i t e ( i n e r t i n i t e ) are non-reactive; v i t r i n i t e and exinite are reactive; semi-fusinite i s considered to be p a r t i a l l y reactive, - 287 -and established practice divides i t as 2/3 non-reactive or inert and 1/3 re-active(318). Inorganic mineral matter in coal i s also non-reactive; therefore total inerts = i n e r t i n i t e +2/3 semifusinite + mineral matter (293,319) where mineral matter on a v o l . % basis is calculated as described above. Table G.l Maceral analysis data (vol.%) and mean maximum v i t r i n i t e reflectance measurements (R^  max %) of particulate coal samples Sample v i t r . heavily semi - macr. micr. fus. in e r t . exini te mineral max oxid. fus. + semi- matter 0 + c rl v i t r . fus — O • u • -595+Oym 0" 74 12 4 2 0.5 3 9.5 1 3 0.95 ± .08 38' 80 4 3 4.5 0.2 2 9.7 1.5 4 1.06 ± .05 75' 72 4 8 6 3 5 22 1 2 1.00 ± .05 77 8 1 ••- 7 16 1 6 111' 80 ; i 3 5 3 5 16 2 2 1.03 ± .05 82 - 8 •1 8 17 1 2 150' 80 2 4 3 4 3 14 t r 4 1.00 ± .06 83 - 4 1 - 8 13 2 2 whole coal No. 5 seam 58 — 9 6 - 20 35 1 8 1.30 ± .06 -210+149ym 75' 74 - 5 2 - 14 21 <1 3 1.06 ± .04 111' 80 - 8 4 - 5 17 1 3 1.06 ± .06 150' 79 - 6 2 - 11 19 1 2 1.12 ± .06 No. 5 seam 61 - 10 5 - 19 34 <1 5 1.35 ± .05 v i t r . r v i t r i n i t e , semi-fus. = semifusinite, macr. = macrinite, micr. = m i c r i n i t e , i n e r t . = i n e r t i n i t e , t r . = trace a i n e r t i n i t e = macrinite + micrinite + fu s i n i t e - 289 -APPENDIX H METHODS FOR THE ANALYSIS OF THE COMPOSITION OF FLOTATION  PRODUCTS FROM SYNTHETIC MIXTURES H .1 Sulphur/Molybdenite The sulpur content in products from f l o t a t i o n tests on synthetic mixtures of elemental r o l l sulphur and molybdenite were determined by burning o f f the sulphur in a laboratory muffle furnace. The autoignition temperature of sulphur in a i r or oxygen is between about 250 - 270°C (303). The f l o t a t i o n products were ground in an agate mortar (to minimize p a r t i c l e segregation) and up to 1 g weighed into porcelain combustion boats. These were placed in the cold muffle furnace and slowly heated to 300°C to minimize loss of material by s p i t t i n g . The residue was cooled and weighed.. Weight loss on blank tests with 100% molybdenite was less than 1%. The results of analyses on standard sulphur/coal mixtures (25,50 and 75% S by weight) were within 1% of the nominal value. H.2 Sulphur/Coal The sulphur combustion method described in the previous section was not applicable for analysing the products from f l o t a t i o n tests on synthetic elemen-ta l r o l l sulphur/coal mixtures due to the concurrent combustion of coal. The standard Eschka technique for determining sulphur in coal did not provide accurate, reproducible results for samples ri c h in elemental sulphur. There-fore the selective dissolution of c r y s t a l l i n e sulphur in CSr, was u t i l i z e d . The following procedure i s based on that presented in reference 303 for dissolu-tion of sulphur in hot CCl^. The s o l u b i l i t y of c r y s t a l l i n e sulphur in CS2 at room temperature is about 30 g/lOOg of solvent. The CS2 was Anachemia Reagent grade, boil i n g point 46°C. (a) Pulverize sample in agate mortar (to minimize p a r t i c l e segregation) (b) Weigh 0.2 - 0.3 g of sample into tared beaker and dry in oven at <\> 105°C for 1 to 2 hours (to convert any amorphous sulphur into the c r y s t a l l i n e form). Allow to cool and reweigh (c) Add about 10 ml of fresh C S 2 to the beaker plus sample, place on hotplate and allow CS2 to boil for about 30 seconds. - 290 -(d) Decant solution through a small f i l t e r paper into a preweighed beaker; avoid wetting the edges of the f i l t e r paper. (e) Add. 1 0 ml of fresh C S 2 and repeat (c) and (d) with the same f i l t e r paper and preweighed beaker. (f) Remove f i l t e r paper from f i l t e r funnel, tear in half, and place in the beaker containing the original sample on the hotplate. (g) Repeat steps (c) to (f) with a fresh f i l t e r paper, but the same decant beaker. (h) Wash f i n a l f i l t e r paper and funnel with fresh C S 2 and c o l l e c t washings in a weighed decant beaker. ( i ) Slowly evaporate the C S 2 , dry the sulphur residue in an oven at about 100°C for an hour, cool and weigh. Analyses run on standard sulphur/coal mixtures with sulphur contents rang-ing from 10 to 98% by weight gave values within 1 or 2% of the nominal value. Blank runs on pure coal samples showed that less than 1% of the coal was d i s s o l -ved in the CS-,. - 291 -APPENDIX J SMALL-SCALE COAL FLOTATION TEST CONDITIONS AND RESULTS The following tables l i s t the f l o t a t i o n test conditions, the measured cum-ulative y i e l d of f l o a t s , the measured cumulative ash content of the floats and rejects, the back-calculated feed ash content, the mean cumulative y i e l d where there were two or more replicate t e s t s , and the ash rejection potential, R, given by R = (100 - cum, yiel d ) x rejects ash 100 -x. c a l . feed ash A l l tests were carried out on the -210 + 149 ym size fraction of the coal .. sample. - 292 -Table J . l No. 5 Seam Flotation test results bub. cond. defl.. time type :-min: flo t a t i o n , methanol time cone, min, v o l . % cum. wt.% y i e l d mean. . wt.% ,ash ( a i r dried) ,wt.% .floats, rejects c a l . feed ash rejection potential none 2 3 15 71.4 10,46 42.31 ' 19.57 1 .618 5 3 10 89.5 14.17 69.15 19.94 .364 15 56.7 8.70 33.48 19.43 .746 54.2 55.5 8.63 32.35 19.49 .760 20 5.0 5.79 20.04 19.33 .985 25 1.0 3.85 19.67 19.51 .998 10 3 15 49.2 8.26 30.71 19.66 .794 15 0.5 10 63.8 10.08 37.98 20.18 .681 51.0 57.4 8.67 31.50 19.86 .777 15 1 2 95.0 17.56 80.63 20.71 .195 5 93.5 15.02 75.59 18.96 .259 93.6 93.6 15.37 78.16 19.39 .258 10 72.3 10.86 42.61 19.65 .601 74.7 10.93 44.03 19.30 .577 15 1 15 36.1 7.35 25.92 19.22 .862 30.5 33.3 7.11 25.22 19.70 .890 20 2.5 9.78 21.04 20.76 .988 3.4