flotation, theory, reagents and ore testing-ronal d. crozier

flotationTHEORY, REAGENTS AND ORE TESTING

Rala D. Crozier

752rr

flotationTHEORY, REAGENTS AND ORE TESTINGFlotation has become one of the most important techniquesavailable for mineral processing. This work provides a uniqueand authoriative review of sulphide mineral collectorproperties, their manufacture and use with specific ores. Specialemphasis is placed on the different flotation mechanismsinvolved in particle capture of sulphide and non-sulphideminerals and the effect of antagonal mechanisms on reagentselection.

The author provides details, some previously unpublished, ofthe chemical properties, manufacture methods and possiblesurface-active impurities of commercial collectors and frothers.In addition, the chemical composition of a broad number ofNorth American and European commercial reagent designationsare usted.

Ore sampling, sample preparation, testing machines androutines are covered as practica! guides to mine laboratory staff.Suggestions to testing procedures, equipment selection andgraphical data evaluation methods to multivariable problems areprovided.

Arecommended text for practising flotation mili metallurgists,technicians and professional engineers working in the mining.and mineral processing industry. Also of interest to laboratoriesserving geologists, mining exploration and water treatmentreas.

ISBN O 08 041864 3

9780080418643

FlOTA1'~ONTheorv, Reagents and Ore Testing

\

Related Perqamon Titles of ~I"lterest

Books

DOBBY and RAOProcessing of Complex Ores

FINCH and DOBBYColumn Flotation

GILCHRISTExtraction Metallurgy

WILLSMineral Processing Technology

JournalsActa Metallurgica et Materialia

Canadian Metallurgical Quarterly

International Journal of Rack Mechanics and Mining Sciences &Geomechanics Abstracts

Minerals Engineering

The Physics of Metals & Metallography

Scripta Metallurgica et Materialia

Full details of all Pergamon publications/free specimen copy of anyPergamon journal available on request from your nearest Pergamon office.

Fl TITheory, Reagents and Ore Testing

by

RONALD D. CROZIER, C.Eng., F.I.M.M., Ph.D.Consulting Engineer, Santiago, Chile

\

PERGAMON PRESSOXFORD . NEW YORK . SEOUL . TOKYO

1el .e c';~

U,IC

U.5A

I

CONTENTS

lntroducton

Chapter 1. Flotaton FundamentalsDefinition of termsReagents employed in flotation

Chapter 2. The Mechansm oC FlotationCoursing bubble flotationContact angle and Hallimond tube measurementsNascent bubble flotationElectron transfer to mineral surface

Chapter 3. Sulphide Mineral FlotationStereochemistry of xanthates and dithiophosphatesSix coordination complexesSeven coordination complexesDithiolate chemistry and surface compoundsThiol collector bonding with sulphide mineralsThe nature of the bubble-particle bondFrother-collector surface complexes

Chapter 4. Thiol Collector ChemistryProperties of xanthatesManufacture of aIkali metal alkyl xanthatesManufacture of thiophosphatesManufacture of xanthogen formatesManufacture of dialkyl thionocarbamatesMiscellaneous collector structures \

Chapter 5. Cornrnercial Sulphhydryl Collectors

v

1

559

1112131718

2525252629364044

46474955586264

66

Chapter 6. Propertles of Flotaton FrothsFlotation frothers and froth hyclrodynamics

. Dynamic froth stabilityThe effect of the collector on froth propertiesThe influence of the frother on the rate of flotationOver oiling

Chapter 7. Chemcal Properties o FrothersPartially soluble frothersCompletely water miscible frothers

Chapter 8. Flotation ModifiersActivating agentsInorganic depressantsOrganic depressants

Chapter 9. Sulpbides - Depression and ActivationSodium sulphideChemical properties of sulphidesMolybdenum concentrate cleaningCopper depression in the molybdenite separationSodium sulphide as a depressantThe chemistry of the Nokes reagentsSulphidisation of refractory ores

Chapter 10. Mili Tests - Case HistoriesSan Manuel - a non-sulphide moly separation processCODELCO Andina - slime sulphidisationCODELCO El Teniente - molybdenite recovery

Chapter 11. Sulpbide Mili PracticeReagent consumption in the flotation industryCopper mill dataFlotation process designCopper mill flotation practiceNon-sulphide copper oresFlotation of complex ores

vi

858585868892

949498

101102103105

107107107111116117120122

127127136147

174174176200205208209

_""

Chapter u. Noa-metallc !\/1!i!ll!eJrall 1Fliot,il!l1:imnNon-sulphide mineral flotationThe nature of oily collector flotatonStructure of mineral surfacesMechanism of collector adhesionCollectors for industrial mineralsFattyacidsAnionic collector chemistryIndustrially important fatty acid collectorsCationic collector chemistryFactors affecting selective flotationMetallic oxide minerals mill practiceIndustrial minerals mill practice

Chapter 13. Flotaton TestingPlanning a testExperimental designTesting proceduresEquipping the flotation laboratoryComparison of laboratory flotation machinesSampling and sample preparationSampling working concentratorsTesting routinesTest procedures used by a reagent supplierRecommendations on setting test conditionsBatch flotation testsStandard test procedure used by Sherex on coalColumn testingFlotation column testing examples

Bibliography

Subject Index

Author Index

\

vii

212212213216217220221224225225228229230

258258260276278285287291295297303306310313317

324

336

343

lIST OF TABlES

Table 1.- Classification of polar minerals 6Table 2.- Col1ector structures 7Table 3.- Frother structures 8Table 4.- Modifiers 10Table 5.- Xanthate contact angles for selected sulphide mineral 16Table 6.- Properties of dixanthogens and xanthic anhydrides 31Table 7.- Dithiolate reagents 33Table 8.- Reduction potentials for dithiolate /thiol couples 34Table 9.- Products extracted from thiolated surfaces 35Table 10.- Mineral electrochemical properties 38Table 11.- Band-gap energies of sulphide mineral semiconductors 40Table 12.- Decomposition of Na ethyl xanthate on drying at 50C 51Table 13.- Properties of alkyl xanthates 51Table 14.- Xanthate assays, slurry suspension process 52Table 15.- Commercial stability data on xanthate solutions 55Table 16.- Density and viscosity of Aerofloat col1ectors 58Table 17.- Properties of alkyl xanthogen ethyl formates 61Table 18.- Structures, and suppliers of alkyl xanthates 67Table 19.- Structures, and suppliers of xanthic esters 68Table 20.- Structures, and suppliers ofxanthogen formates 69Table 21.- Structures, and suppliers of thionocarbamates 70Table 22.- Structures, and suppliers of thiophosphate col1ectors 70Table 23.- Structures, and suppliers of miscel1aneous col1ectors 72Table 24.- Properties of alcohols used as frothers 94Table 25.- Properties of terpenes and ketones used as frothers 96Table 26.- Properties of water soluble frothers 99Table 27.- Critical pH for flotation 101Table 28.- Effect of col1ector on critical pH foc flotation 102Table 29.- The effect of sodium cyanide on critical pH 102Table 30.- Sodium sulphide saturation solubility 108Table 31.- Effect of pH on the equilibrium values of sulphide ions 110Table 32.- Effect of temperature on the pH of NaSH solutions 111Table 33.- Reagent practice, moly recovery from porphyry ores, Cu circuit 112

ix

Table 34.- Reagent practice, moly recovery'from porphyry ores, moly circuitTable 35.- Regional mine performanceTable 36.- Moly plants using the sulphide process - Shirley (1979)Table 37.- San Manuel copper and moly assaysTable 38.- Molybdenite plant reagent consumptionTable 39.- Microscopic analysis of the moly plant streamsTable 40.- Lab results of sand/slime tests, without added reagentsTable 41.- Lab results of sand/slime tests, with split after 6 minutesTable 42.- Lab results of sand/slime tests, with split after 12 minutesTable 43.- Plant test of the effectiveness of the sand/slime separationTable 44.- Screen analysis and copper distribution - AndinaTable 45.- The effect of NaSH conditioning time on copper recoveryTable 46.- Andina - results of test for addition point of NaSHTable 470- Material balances for El TenienteTable 48.- Floatability of samples from different mine sectorsTable 490- Material balance and screen analysis retreat regrindTable 50.- Effect of Nokes reagent on rougher recoveryTable 51.- Effect of Nokes reagent on first cleaner recoveriesTable 520- Moly/copper separation cleaner tests with columnsTable 53.- Froth flotation in the UoSAo; source: Ll.S. Bureau of MinesTable 540- UoS. consumption of collectors in 1975 and 1985Table 55.- Consumption offrothers in the total U.S. mining industryTable 56.- Industrial copper mili flotation performance dataTable 570- Industrial copper mili flotation reagent dataTable 58.- Average porphyry copper mine performance in 1975Table 590- Collector use frequency comparison for copper milisTable 60.- Frother use frequency comparison for copper millsTable 61.- Collector consumption in copper sulphide ore flotationTable 620- Power and water use in U'S, flotation milisTable 63.- Screen size equivalentsTable 64.- Principal sulphide mineralsTable 650- Secondary minerals in Pb-Zn complex oresTable 66.- Collector reagents for non-metallic mineralsTable 67.- Fatty acid content of oils and fats (weight %)Table 680- The structure of sorne anionic collectorsTable 69.- Critical miscelle concentration for anionic collectorsTable 700- Nitrogenous cationic flotation agentsTable 71.- Cationic amine collector structuresTable 72.- Molecular species solubility, pKa, and polar head diameterTable 73.- Commercially important boron mineralsTable 740- Glass fiber colemanite specificationsTable 75.- Typical composition of spruce pine alaskiteTable 76.- Reagents employed in the flotation of feldspatic ores

x

-

113115119129135138138140142144145146148152157167171171172175176177179/18185/191192194196198199201206210220222222224226227228232233235237

Table 77.- Reagents for fluorspar, sulphide and barite flotation 240Table 78.- Fluorspar grade specifications 242Table 79.- Specification for commercial quartz and feldspar 242Table 80.- Reagents employed in the flotation of glass sand in the USA 243Table 81.- Experimental design and results - glass sands 245Table 82.- Experimental design - glass sands flotation 246Table 83.- Reagents employed in the flotation of phosphate rack in the USA 250Table 84.- Reagents used in potash flotation 252Table 85.- Reagents employed in the flotation of potash in the USA 253Table 86.- Scheelite flotation results 253Table 87.- Mineralogical analysis of Kings Mountain spudomene ore 255Table 88.- Feldspar, quartz, spudomene minerals in N.e. pegmatites 256Table 89.- Flotation testing variables 259Table 90.- Influence of EtX and alpha-terpineol consumption on grade 268Table 91.- Influence of EtX and octanol consumption on concentrate grade 269Table 92.- Three-variable five-Ievel Box-Wilson design layout 273Table 93.- Flotation cell dimensions and standard flotation test conditions 285Table 94.- Reproducibility of time-recovery curves for Leeds cells 287Table 95.- Parameters to determine sample size for different minerals 288Table 96.- Column flotation variables used in reground pulp cleaner study 318Table 97.- Overall experimental plan for the RRC samples in the 2" column 320Table 98.- Optimum flotation conditions for the 2" column 321

\

xi

UST OF FIGURES

Fig. 1.- Cavitation rnechanisrn of bubble forrnation 18Fig. 2.- The electrical double layer rnodel 21Fig. 3.- The structure of ice 22Fig. 4.- Localized bond structure for water 23Fig. 5.- Cation and anion hydration showing water orientation 24Fig. 6.- General stereochernistry for [M(bidentateh]x:!: 26Fig. 7.- [CuII(xanthate)z] or [{Cu(xanth)Z}xCuIIx_z(HzO)z] 27Fig. 8.- Pentagonal bipyramid and capped octahedron geometry 27Fig. 9.- [PbII(ethyl xanthatejjlone pair)] 28Fig. 10.- [PbII(di-isopropylphosphorodithiolate)z(lone pair)] 29Fig. 11.- [M(bidentateh(lone pair)]X:!: 30Fig. 12.- Structure of chalcopyrite and enargite 39Fig. 13.- Wurtzite and sphalerite, from Pauling 39Fig. 14.- Leja and Schulman (1954) interpenetration theory 43Fig. 15.- Frother-collector interaction complex 44Fig. 16.- Frother volume versus alcohol chain length 87Fig. 17.- Effect of frother concentration 88Fig. 18.- Iso-recovery curves, Hallimond tube 88Fig. 19.- Iso-recovery curves for chalcocite 89Fig. 20.- Rate of flotation of copper ores 90Fig. 21.- Performance of frothers on recovery of coal/ash 91Fig. 22.- Illustration of two laboratory time-recovery profiles 93Fig. 23. Ionization of sodium sulphide solutions 109Fig. 24.- Relative effect of the addition of NaSH on pH 111Fig. 25.- Relationship between pH, NaSH depression and floatability 118Fig. 26.- San Manuel molybdenite recovery circuit 128Fig. 27.- Andina rougher circuit in 1976 136Fig. 28.- Andina rougher circuit in 1977 137Fig. 29.- Rougher circuit for Andina's sand/slnne mili test 141Fig. 30.- Overall molybdenum. recovery, El Teniente 149Fig. 31.- Simplified flotation schematic for El Teniente 151Fig. 32.- Sewell ore CuT, CuNS and % oxide copper content 154Fig. 33.- Sewell, Colon and combined ore non-sulphide assay 155

xiii

Fig. 34.- Mineral content of Teniente concentrate - 1984 155Fig. 35.- Monthly moly assays and tonnage produced by sectors 156Fig. 36.- Monthly moly assays for 1989 for Colon and Sewell 156Fig. 37.- Floatability of samples from different sectors of the mine 158Fig. 38.- The alkaline c1eaner flotation circuit at Colon 160Fig. 39.- Performance of the copper c1eaner circuit 161Fig. 41.- The effect of pH on pyrite and molybdenite recovery 163Fig. 42.- Effect of pH on Cu and moly recovery from second c1eaner 164Fig. 43.- Molybdenum content of Teniente tailings water 164Fig. 44.- Potential-pl-l diagram for moly-sulphur-water 165Fig. 45.- First cleaners and copper circuit flow diagram 166Fig. 46.- Effect of grind on copper and moly recovery 168Fig. 47.- Moly content vs particle size - Colon 169Fig. 48.- Moly content vs particle size - Sewell 170Fig. 49.- Proposed new Colon cleaner circuit 170Fig. 50.- Redox measurements of rougher flotation - Teniente 172Fig. 51.- Overall mill performance, world survey 193Fig. 52a.-Collector frequency graphs, world survey 195Fig. 52b.-Frother frequency graphs, world survey 197Fig. 53.- Complex ore differential flotation circuit 204Fig. 54.- Effect of grind on copper recovery from ores 208Fig. 55.- Forces in adsorption and chemisorption 216Fig. 56.- Water clathrates 218Fig. 57.- Effect of brine composition on ratio of oleate chemisorption 228Fig. 58.- Feldspar, mica and quartz flotation, Spruce Pine, NC 236Fig. 59.- Weathered decomposed pegmatite processing, N.e. 236Fig. 60.- Fluorspar milI, Reynolds Mining Corp., Texas 238Fig. 61.- Mill circuit for a fluorspar-zincheavy media concentrate 239Fig. 62.- Pilot plant flowsheet for sand beneficiation 244Fig. 63.- Lock cycle tests and pilot plant - glass sands 244Fig. 64.- Flow diagram for a phosphate rock plant 249Fig. 65.- Potash scrubbing-desliming-reagent conditioning system 251Fig. 66.- Foote spudomene flotation circuit 254Fig. 67.- Spudomene and feldspar-quartz-mica flotation, Kings Mt, NC 255Fig. 68.- Results of 2 kg laboratory flotation tests, single variable search 261Fig. 69.- A two-variable factorial design 262Fig. 70.- Experimental design based on a square pattern 264Fig. 71.- Cubic experiment for a truncated 3 by 3 design 265Fig. 72.- Effect of frother and collector on copper recovery 266Fig. 73.- Collector/frother interaction on recovery and grade 267Fig. 74.- Interaction of octanol and EtX on recovery and grade 270Fig. 75.- Box-Wilson rotatable two-variable five-level design 272Fig. 76.- Box-Wilson rotatable three-variable five-level design 273

xiv

Fig. 77.- False optimum in a one variable at a time search 274Fig. 78.- Steepest ascent experimental procedure result 275Fig. 79.- SSDEVOP: climbing a response surface, from Mular 276Fig. 80.- Standard 5.5 liter (2 kg) flotation cell from ASTM 281Fig. 81.- Flotation cell proposed by ISO committee on coal 282Fig. 82.- Flotation cell impeller and diffuser 283Fig. 83.- Comparison of the performance of flotation machines 286Fig. 84.- Minimum representative sample size 289Fig. 85a.-Standard 2 kg ball milI used in Chile (body) 301Fig. 85b.-Standard 2 kgball mill used in Chile (cover) 302Fig. 86.- Schematic of a flotation column 314Fig. 87.- CIMM's column flotation pilot plant 317Fg. 88.- CIMM recommended flowsheet for cleaner column flotation 322

\

xv

ACKNOWlEDGEMENTS

The author acknowledges with thanks the assistance given by the folIowing companiesand publishers in permitting the reproduction of illustrations from their publications:

John Wiley & Sons Inc., 605 Third Ave, New York, NY 10016Stephen J. Lippard, Editor, Progress in inorganic chemistry, Volume 23, copyright -1977; containing: D.L. Kepert, "Aspects of the stereochemistry of six-coordination",Figures 6.7,8 & 9; and Michael G. B. Drew, "Seven-coordination chemistry", forFigures 10 and 11.

Gordon and Breach Science Publishers, 270 Eight Avenue, New York, NY 10011,Janusz S. Laskowski, Editor, "Frothing in flotation", copyright: 1989; containing RonaldD. Crozier, and Richard R. Klimpel, "Frothers: plant practice", Figures 15, 20, 21, and 22

Plenum Publishing Corporation, 233 Spring Street, New York, NY 10013Jan Leja, "Surface chemistry of flotation", copyright: 1982, Figures 2,55 and 57

Cornell University Press, 124 Robert Place, Ithaca, NY 14850Linus Pauling, "The nature of the chemical bond", copyright: 1960, Figures 3, 12, and 13

Australasian Institute of Mining and MetaIlurgy (Inc), Clunies Ross House, 191 RoyalParade, Parkville, Victoria 3052, AustraliaK.L. Sutherland, and I.W. Wark, "PrincipIes of flotation", copyright: 1955, Figure 25

The Society of Mining Engineers, AIME-SME, Littleton, caLeja, J., & Schulman, J.H. (1954), Flotation theory: molecular interaction betweenfrothers and colIectors at solid-liquid-air interfaces, Trans. AIME., 16, pp.221-8. Figure14, depicting the interpenetration theoryMular, A.L. (1989), Modelling, simulation and" optimization of mineral processingcircuits, Challenges in mineral processing, Eds., K.V.S. Sastry and M.e. Fuerstenau,SME-AIME, Littleton, ca, pp.323-349. Figure 77, 78, and 79Mular, A.L. (1976), Optimization in flotation plants, Flotation, Ed. M.e. Fuerstenau,SME-A1ME, Littleton, ca, pp. 895-936. Figures 75, and 76Redeker, Immo H. and Bentzen, E.H., (1986),Plant and laboratory practice in non-

xvii

metallic flotation, Chemical reagents in the mineral processing industry, Editors, DeepakMalhotra and W.F. Riggs. SME of AIME, Littleton, ca, pp. 3-20 Figures 58, 59, 61,and 67

'rile Institution of Mining and Metallurgy, 44 PortIand Place, London, W1N 4BRGrainger-Allen, T.J.N. (1970), Bubble generation in froth flotation machines, Trans.I.M.M.;]2, pp. C15-C22, figure 1Lekki, J., & Laskowski, J. (1971), On the dynamic effect of frother-collector joint actionin flotation, Trans. I.M.M., Section CM, C174-80. figures 16, 17, and 18Mathieu, G.L and Sirois, L.L. (1984), New processes to float feldspathic and ferrousminerals from quartz, Reagents in the minerals industry, M.J.Jones and ROblatt, Eds,(London: IMM), pp. 57-67; Figure 62, and 63

Mine and Quarry Engineerng, 19 Fairlie Road, London, UKWrobel, SA, (1953), Power and stability of flotation frothers, governing factors, Mineand Quarry Engineering; 19, pp. :?63-7, Figures 16

Pergamon Press, Headington Hill Hall, Oxford OX3 OBWBeas-Bustos, E., and Crozier, RD. (1991) Molyfcopper separation from concentrate ofthe combined acid and basic circuits at El Teniente, Reagents in Minerals Engineering,Camborne School of Mines, September 18-20. Figures 31-51Crozier, RD., (1991) Sulphide collector mineral bonding and the mechanism offlotation, MineralsEngineering, vol 4, Nos 7-11, pp.839-858, Figure 5

The author is also grateful to Mr. Farlow Davis for permission to use unpublishedmaterial on molybdenum depressants; to CODELCO-Chile division management forproviding the data used in the case histories on the Andina and El Teniente mines; tothe Director of Chile's Centro de Investigation Minera y Metalurgica for providing thecolumn testing data; and to Sherex Chemical Company, Inc, which provided flotationtesting procedures and details of the test equipment. Sabine Slotta is thanked for themany tedious hours spent copy editing the manuscript.

xviii

INTRODUCTION

This text is primarily addressed to the mill metallurgist and to technicalservice and development engineers working for reagent suppliers, rather thanto flotation researchers or mill process designers. With this audience inmind, flotation theory is covered so as to emphasise the existence of distinctlydifferent mechanisms in the flotation of simple metallic sulphides, complexsulphides, metallic non-sulphides (loosely known as oxides), industrialminerals, and coals.

As a primary target of flotation testing is correction of a chemical problem ina working mill, be it the flotation media (water) or the selection of acollector, frother or modifier to improve the recovery efficiency of a mill, aconsiderable amount of space has been devoted to attaching the properchemical composition to commercial trade names, as well as providingrelatively detailed information on the properties of the industrial collectors,frothers and modifiers in normal use. Reagents produced by former majorcollector suppliers such as Minerec Corporation (out of business), and theDow Chemical Company, no longer in this niche business, are also identified,to aid in the interpretation of older published mill data and because manylaboratories still have samples of Dow and Minerec reagents which can beused to broaden a particular flotation study. A brief description of thesynthesis of collectors is given in case special samples need to be prepared,and their physico-chemical properties are summarised, as these data aregenerally not available in standard handbooks. Water quality, per se, is toovast a subject to try and summarise, but as the interaction between collectorsand cations in solution is a common problem in mineral pulps, sorne of thecoordination chemistry of collectors is touched on, as is the semi-conductorproperties of the principal sulphide minerals. .

\The technology to treat low grade ores has evolved by trial and error in theoperating mills, and very little is based on fundamental research. Most of thepublished fundamental research has been of relatively little value in resolvingpractical problems, primarily because it has been carried out using syntheticmixtures of pure minerals. In addition, test equipment normally used in

- 1-

2 FLOTATION - Theory, Reagents and Ore Testing

micro-flotation laboratories usually mimics the obsolete pneumatic flotationmachines of the turn of the century. Thus, the results of much of thepublished fundamental research cannot be applied to the design of sulphidemineral flotation circuits. Because of mechanistic incompatibility of agitationflotation and pneumatic flotation, some of the conclusions and mathematicalrelations proposed in the literature on the rate of flotation of pure minerals isparticularly dangerous if it is used in the development of scale-up algorithmsand in choosing the key process variables in the design of experiments. Thiscriticism of simplified flotation testing equipment does not mean that there isnot a crying need for an accurate micro-flotation apparatus capable ofobtaining scale-up data and accurate processability responses from gram-sized samples obtained from core drilling programmes. BUT, just as thegrade and mineralogical data must be inserted into a representative 3-D gridof an ore body to generate reliable geologic modelling of a mining project'sreserves, the same broad analysis must be applied to the floatability responseto provide useful recoveries and operating costs. The statistical confidencerange of the floatability model based on bench flotation data is poor, and ifbased on micro-flotation experiments it is considerably poorer than thereserve estimate because chemical analysis of exploration cores is moreaccurate. Thus, designing a mill based on current theoretical knowledgeusing micro-flotation data could well be disastrously more expensive thaninvesting in adits, tonnage samples, and pilot plant tests.

The deve!opment of an optimum metallurgical sequence to process acomplete!y new ore is always expensive, therefore it must be carefullybudgeted in all new projects. An improper geologic evaluation of theprobable ore grade and reserves can have serious financial consequences fora mining company if the ore runs out before all the capital invested has beenrecovered, so valuation of an ore deposit justifies a large expenditure indrilling and exploration. Similarly, unexpected processing difficu1ties, whichresult in a recovery reduction of 3 or 4% on predicted efficiencies, willaffect the cash flow values used to calculate the debt service, and can alsohave a crucial effect on corporate existence, considering the mammothinvestments required for today's low grade ores. So the allocation ofresources, human and financial, to testing budgets during the pre-investmentdecision period must be carefully weighed.

Much has been written on how to design and manage the proper explorationof an ore body and the design of an optimum drilling programme.Computer programmes are capable of processing drill, core and tunnellingdata, and presenting these in terms of three dimensional topologies of themineral species, and spewing out the proven, probable and inferred reservesof the future mine, including an indication of the confidence leve! of the

Introduction 3

quantity of ore that will be available. However, no similar computercapability has been standardised in the more mundane area of testing oresfor processability, particularly if flotation is involved.

The chemistry of mineral surfaces is dealt with systematically, and is readilyavailable in standard publications, though not necessarily for conditionsapplicable to flotation. The review of the mechanisms of flotationsummarised in this text is slanted to emphasise factors that affect theselection or design of flotation reagents, not equipment and process circuits.The greater emphasis and breadth of information on sulphide reagentsreflects the writer's expertise and not the relative size of the industrialminerals reagent market. The testing procedures designed for sulphide ores,whose behaviour is more complex, can all be easily adapted to non-sulphideores and also to non-metallic flotation, cleaning or separation processes, sosulphide testing is generally covered more fully.

As will be apparent from the text, the selection of appropriate reagents forthe optimum flotation of a particular ore is more an art than a science; butthe use of a systematic approach to obtain laboratory flotation data canreduce the hit or miss aspects of this particular art. Information on theperformance of flotation reagents is frequently contradictory because non-identical flotation circuits may be conforming to a different flotation regime,and therefore the reagents in the full-scale plants act different1y from thecase histories described in the literature. Even if we limit our review tocurrent1y fashionable flotation equipment, we must qua1ify any reagentrecommendations, as frothers and collectors will not react in the samemanner if the flotation is carried out in small subaeration machines, forcedair units, modern giant cells or flotation columns. Quite obviously, there arealso operational differences between the generically similar machine designsof different manufacturers, which can affect the choice of the reagents.

This mechanistic factor, i.e. that a particular reagent's flotation response isdifferent, depending on whether the flotation occurs because of bubblecapture, or mineral surface catalysed micro-bubble formation, is glossed overin most texts on flotation. In these notes it will be over-emphasised, becauseit is so important in diagnosing prablems in cleaner or scavenger circuits, andin extrapolating applicability of a particular reagent from one ore to anotheror from one type of flotation equipment to ~other (i.e., from cell to columnflotation, where only bubble capture occurs).

To add complexity to our information matrix, there are at least four differenttypes of minerals that are upgraded by flotation: (1) metallic sulphideminerals, (2) metallic oxides or tarnished sulphide minerals, (3) non-metallic


insoluble minerals, and (4) non-metallic soluble, or semi-soluble minera!s. Ineach case the mechanism of flotation can be bubble capture or nascentbubble formation, or a combination of the two.

Sub-groups of these four minerals are ores whose mineral disseminationrequires that they be ground to colloidal sizes to liberate the valuable species;minerals where differential flotation is required to separate complex ores thatinclude mixtures of hydrophyllic minerals; and those that are naturallyhydrophobic and amenable to flotation without the use of collectors. It is alsoobvious that the milI metallurgist, the flotation equipment designer, and thereagent designer have quite different goals, and therefore possibly conflictingpoints of views, on how to apply the technology used in flotation. Hardwaredesigners, for example, should be aware that designations such as a "strong"or a "weak" collector do not describe the strength of adhesion between amineral and a bubble; rather, it is an inverse description of selectivity, or theeffect of a particular reagent on the rate of flotation of individual minerals, soa change to a 'stronger collector' in a formu1ation is unlikely to compensatefor poor over-all recovery because of an overly deep cell. When industriallyproven flotation machines are involved in a side-by-side mill test and one isnot performing competitively, its inadequate performance cannot becompensated by a reagent suite change, as the effect of the reagent changecould affect the operation of both machines identically, in which case thecause of the difference is more likely to be due to a physical operatingvariable, such as pulp density, percent solids, air feed rate, impeller rpm,mineral particle size distribution or sorne other operating detail that has notbeen fully controlled during the test periodo

Froth flotation, the surface dependent process par excellence, consists of adeceptively simple procedure - the separation of valuable constituents o atwo or three phase mixture by levitation - but, in fact, flotation embodies avery broad gamut of non-obvious sequential micro-processes, thecombination of which have as a final result the levitation of a desired mineralspecies, leaving other minerals and gangue in the tailings. The same reagentand micro-process sequence does not act the same when applied to differentminerals, i.e. non-metallics, metal oxides and metal sulphides can bedepressed or promoted by a particular reagent. This behaviour results inwhat appear to be completely contradictory results reported in researchpublished by different reputable metallurgists. The reason for thisinconsistency is the great diversity in mineral morphology of commerciallyimportant raw materials. Thus, there is no single valid theory of themechanism of flotation which can be used to design universally applicabletesting procedures, and therefore no 'cookbook' mineral processability testrecipes.

\Chapter 1

FLOTATION FUNDAMENTAlS

Definition of terms

Frotlt flotation is a process used to separate minerals, suspended in liquids, byattaching them to gas bubbles to provide selective levitation of the solidpartic1es. It is the cheapest and most extensively used process for theseparation of chemically similar minerals, and to concentrate ores foreconomical smelting.

Floatable minerals can be c1assified into non-polar and polar types, accordingto Wills (1988). The segregation into these two types of minerals is based ontheir surface bonding. The surfaces of non-polar minerals have relativelyweak molecular bonds, difficult to hydrate, and in consequence such mineralsare hydrophobic. Non-polar minerals inc1ude graphite, sulphur, molybdenite,diamond, coal, and tale, all naturally floatable in the pure state. The orescontaining these minerals usually require the addition of non-specificcollectors to their pulp to aid the natural hydrophobicity of the floatablefraction, i.e. oily collectors, such as fuel oil, kerosene, coal distillates, etc.

Polar minerals have strong covalent or ionic surfacebonding, and exhibithigh free energies at these surfaces. Therefore, surface hydration is rapid dueto the strong reaction with water molecules which form multilayers on themineral. Thus these species are hydrophyllic. The listing in Table 1 isadapted from Wills (1988). The minerals are listed in groups of increasingpolarity, divided into c1asses dependent on the magnitude of the polarity.Minerals in group 3 have similar degrees of polarity, but this polarity can bechanged easily in the case of the minerals grouped under 3 (a) because theyare susceptible to sulphidisation.

The minerals in group 1 are all sulphides, with the exception of the puremetals. Their flotation characteristics are not uniform, and, based onFinkelstein and Poling (1977), must be further divided into three more sub-

-5-


groups: those rninerals that form a xanthate, dithiophosphate, mercaptan,thionocarbamate, or xanthogen frmate dimer (dithiolate) on their surfaceon contact with the collector (covellite, chalcopyrite, stibnite, arsenopyrite,pyrite, sphalerite, orpiment, realgar, and the native Au, Pt, Ag and Cu); thosethat form metal xanthates (galena, bornite, chalcocite); and the others, where

1the analytical data is ambiguous on whether the collector bond is with themetal cation or with the surface sulphur atoms in the mineral.

TabLe 1.-

Group 1

CLASSIFICATION OF POLAR MINERALS

Group 1 continued

OrpimentPent LanditeReaLgarCinnabarALabanditeNative Au,

GaLenaChaLcopyriteCoveLL iteBorniteChaLcociteEnergiteArgentiteMiLLeriteCobaLtiteArsenopyritePyritePyrrhotiteSphaLeriteStibnite

Group 3a

PbSCuFeS2CuSCUS FeS4CU2SCU3AS,Sb)S4A92SNiSCoAsSFeAsSFeS2FE7S8ZnSSb2S3

Group 2

BariteAnhydriteGypsumAngLesite

Group 3

As2S3CFe,Ni)9S8AsSHgSMnS

Ag, Pt, Cu

BaS04CaS04CaS042H20PbS04

MaLachiteAzuriteChrysocoLLaWuLfeniteCerrusite

Group 4

CU2C03(OH)22CuC03CuCOH)2CUSi03 2H20PbMo04PbC03

FLouriteWitheriteMagnesiteDoLomiteApatiteScheeLiteSmithsoniteRhodochrositeSideriteMonazite

Group S

CaF2BaC03Mg C03CaMg(C03)2CaSCCF,CL)CP04)3CaW04Zn si Li cate cLayMnC03FeC03CCe,La,Di)P04

HematiteGoethiteChromitePyroLusiteBoraxWoLframiteCoLumbiteTantaLiteRutiLeCassiterite

Fe203FeO(OH)FeCr204Mn02Na2B407CFe,Mn) W04CFe,Mn)CNb,Ta)206FeTa206Ti02Sn02

Zircon ZrSi04Hemimorphite Zn4Si207COH)2.2H20BeryL Be3AL2Si6018Garnet Ca3AL2CSi04)3

source: WiLLs (1988)

Flotation Fundamentals

Table 2.- COLLECTOR STRUCTURES

7

Type

Soluble Collectors

Structure Water Solubility

Alkyl xanthates orAlkyl dithiocarbonatesSodium 0,0 Dialkyldithiophosphates

Insoluble OilsMercaptans

2-Mercaptobenzothiazol

O,O-Aryl,O-aryldithiophosphoricacids

Dialkyl xanthogenformates

Dialkyl thiono-carbamates

Xanthic Esters

S

R-O-C-SNa

R-O SR-oIP(s-Na+

R-O-8H

~J -+8 e-s NaPh-O S

Ph-oIP(s-Na+

s OR-O-C-S-C-O-R'

S

R-O-C-NHR'

S

R-O-C-S-R'

Soluble

Soluble

very sltlysoluble

Soluble

Insoluble

Insoluble

Insoluble

Insoluble

Metal oxide and non-sulphide mineral collectors

OsolubleSlightly

Fatty acids R-C-OH in water

OSulphonates R-S-O-OH Soluble

O

Alkyl amines R-HNH2 SolubleQuaternary ammonium \compounds RR'R"R"'NCl Soluble

The reason to make this differentiation is that the mechanism of flotation isunique when the collector reacts with the mineral surface to form adithiolate. Nascent bubble flotation is the rate dominant mechanism, when

s FLOTAnON - Theory, Reagents and Ore Testing

dimers are Iorrned, while coursing bubbles dominare when there is true oilycollector flotation, With the exception of pentlandite, ambiguous mineralsare less often encountered in the flotation laboratory,

TabLe 3.- FROTHER STRUCTURES

Type

ALiphatic ALcohoLsR = ALkyL gp with 5 to 8 CMISC

Structure

R-OH

CH3CHCH2CHCH3CH3 OH

Water SoLubiLity

s Light

Pine OiLs ell

TerpineoLs /~~ /CH" CIIHC ""-:::CH HC I C( 3 sLightI I ZI CII I %HC, )H HZC IZ CHOII

CH""-ClJ/I I

"3 C-1-CH 3 CIJaOH~ - te~pifll!(Jl fl!flchyl alcoool

CresyLic acid

ALkoxyparaffins1,1,3-triethoxybutaneor TES

/OCH2CH3CH3CHCH2C-H

I \CH2CH3OCH2CH3

sLight

sLight

PoLygLycoLethersDowfroths

Aerofroths

R(OR' )xOHR(OR' )xOH

miscibLe at Low 'x'to partiaLLysoLubLe at high 'x'

The largest volume flotation operations are those that involve the upgradingof commodities: iron ores, coal, phosphates, limestone, potash, copper... butthe technically most interesting are the separation of similar metals fromcomplex ore, such as mixed lead, zinc, silver and copper ores, orcopper /molybdenum concentrates.

Flotation Fundamentals

Bu/k flotation is a sornewhat imprecise term that covers nearly a1l the normalrougher or scavenger flotations, where a single mineral or a group of relatedminerals are separated from gangue and other low value minerals in a singleflotation step. An exarnple would be the recovery of a mixed coppersulphide concntrate from an ore containing pyrite and gangueo Othersinelude the separation of KCI from sodium chloride, upgrading phosphaterock, desulphurising iron ores, etc.

Differential flotation is the terrn norma11y used to describe the separation ofcomplex ores, and is generally restricted to describing the separation ofsimilar minerals from each other (e.g. copper, Iead, zinc, silver and gold froma single ore, or molybdenum from copper in concntrate, etc.) where thesuccessful and economic recovery of each component involves thesophisticated use of collectors, depressants and flotation activators.

Reagents employed in flotation

The reagents employed in flotation are generally interfacial surface tensionmodifiers, surface chemistry modifiers, and/or flocculants. Usua11y they areelassified under five headings: co11ectors (sometimes known as flotationpromoters), frothers, modifiers, activators and depressants. The designationco11ector or promoter, interchangeably, reflects the opinion of the individualmetallurgist, who may consider the primary function of the collector to be aflotation rate accelerator, or a fine particle agglomerator, in a particularflotation system.

Collectors are reagents that coat and/or react with mineral surfaces andmake them water repellent or attachable to air bubbles. The chemicalstructures of most commonly employed collectors are shown in Table 2.Sulphide ore collectors a11 contain sulphur and are thiols or can hydrolyse toa thiol. Non-sulphide and non-metallic minerals are normally floatedemploying collectors such as fatty acids, amines, quaternary ammoniumcompounds, sulphonates or petroleum oils.

Frothers are surface active reagents that aid in the formation andstabilisation of air-induced flotation froths. The commonly employedfrothing agents are aleohols which are only slightly soluble in water, or themore modern frothers, which are generally varieties of polyethers orpolyglycol ethers that are completely miscible with water. Examples are Iistedin Table 3.

Modifiers, activators and depressants The boundaries between the functionsof a specific in

la FLOTATION - Theory, Reagents and Ore Testing

control, an environment modifier such as lime may be used; but limecontains caleium, and the caleium cation is a known depressant for pyrites incopper flotation or for quartz and tale in the flotation of silver ores. So limewould need to be c1assified under both headings. The silicates and thephosphates act as modifiers when they are used to control the effect ofcationic and anionic impurities in flotation water, but can also act asactivators or depressants in different mineral systems. Another example of adual function is that of the sulphhydryl anion which can be a depressant forsulphide copper, especially in the copper/ moly separation, but, with verycareful concentration control, this anion can be used as a sulphidiser toactivate non-sulphide or oxide copper.

TabLe 4.- MODIFIERS

pH ModifiersLime:Soda ash:

Caustic soda:

Acids:

Resurfacing agentsCations:Anions:

Organic CoLLoids:

Ba++, Ca++, Cu++, Pb++, Zn++, Ag+Si03, P04-- CN-, C03' S=

dextrin, starch, gLue, etc.

Chapter 2

THE MECHANI5M OF FlOTAT~ON

The morphology of froth flotation is not well defined because full-scalesystems are normally opaque. In macro terms, froth concentrates form byselective capture of mineral particles in the bubble generation zone, followedby a second zone where there is coagulation of the discrete bubble streaminto a loose froth. These first two zones are norma1ly identified as the "pulp"in a flotation cell, and can be fixed in height by a level control mechanism.The third region appears when the gas phase predominates over the liquidphase; here, the froth starts to condense and returns water to the pulpoFinal1y a structurally stable mineral-laden froth forms, and is removed overthe lip of the flotation machine. The height (or thickness) of this stable frothcan be control1ed by the designer and operator by fixing the total froth heightand the pulp height. The operating variables, that can be fixed by theforeman within the flotation bank, are: pulp mineral content (pulp density)and flow rate; aeration gas rate; in some cases agitation intensity (inconventional cells); pulp/froth interface height; and total overflow height,which fixes the froth drainage height and the froth removal rateo

A material balance algorithm for a single cell can quantify both the pulp massflow rate in and out of the cell, and also the concentrate mass rate, if assaysare available for the different streams. At a lower confidence level, the rateof return of water to the pulp, and the rate of flow of the differentcomponents of the concentrate, can also be calculated. Under steadyconditions, the height of the interface between the concentrate froth and thedraining froth can be inferred from froth physical property measurements,such as density or metal concentration profiles,

The key mechanistic variable that is undefined is exactly how a bubblese1ectively captures a mineral containing ore particles. Observation confirmsthat for selective capture the desired mineral must be natural1y floatable orbe preferential1y coated with a collector. Once the desired particle is tagged,there appear to be two main, non-exclusive,bubble loading mechanisms inoperation: a) multiple particle bubble collisions in the pulp which eventuallyresult in particle adhesion, and b) micro-bubble formation directly on themineral surface within the bubble formation zone of the pulpo There isconsensus that the forces that join the mineral particle to the bubble, and

-11-


that define the rate of capture of the mineral by the bubble, involve electrontransfers when the mineral is conductive, and surface tension forces whenthe mineral is an insulator.

Coursng bubble Ilotation

The froth flotation literature labels these two particle capture processes asthe coursing bubb1e process, and the air precipitation 01' nascent bubbleprocess (the latter wiIl be used in this text). Bubb1e capture, 01' coursingflotation, predominares when oily collectors are used to concentrate non-metallic minerals or non-sulphide metallic ores, 01' when collectorscontaining double bonded sulphur atoms are used with metallic sulphide oreswhich do not form dithiolates on the mineral surface, but ionise and formmetal salts. As we see from the data in Table 53 (page 175), the bubblecapture mechanism occurs mainly with minerals where the concentrationratio is less than 5:1, and selectivity is usually obtained by employing specificdepressants rather than making collector adjustments.

In these cases, the flotation efficiency of different reagents can be measuredby simple tests such as Hallimond tube experiments. though even with thesesimple ores one would hesitate to extrapolate the se1ectivity results ofHallimond tube flotation tests (which normally are carried out using only acollector) to a specific industrial flotation, as the concentrate obtained isprimarilydependent on the rate of flotation. As is weIl known, the flotationrate depends more on the frother employed, rather than the collector.

With oily collectors, contact angle measurements, and the correspondingcalculated adhesion force between a bubble and a mineral particle, may be aguide to the mineral carrying capacity of a bubble, at least when very dilutepulps are involved; but under normal industrial flotation conditions, wherepulps contain at least 30% solids, the rheology of the inter-bubble fluid is theproperty that fixes the mineral-carrying capacity of the froth, not the strengthof the particle to bubble bond, though the differential value of this force fortwo or more minerals and the gangue may help predict the se1ectivity of areagent.

KIassen and Mokrousov (1963) describe the fundamental difference betweenthe collectors for these two types of flotation by pointing out that: "In physicaladsorption, the reagent maintains its chemical individuality, whereas, inchemisorption, [when e1ectron transfers occur] it forms new compounds".They confirm other researchers, such as Gaudin (1957), that when thecollector acts as a micro-bubble nucleator, the dosage must be such that onlya small proportion of the mineral surface is covered by the collector, which

Mechanism o' Flotation

means that the surface of a Iloatable sulphide mineral is j


Ilotation systems in general, be they sulphide or non-sulphide. For example,Laskowski (1974) cites Klassen and Mokrousov (1963) saying:

"These authors point out that flotation is typically a non-equilibrium process. The phenomena that occur on thesurfaee of particles and bubbles in agitated, aerated mineralpulps, when these particles strike bubbles, are dynamic-non-equilibrium processes which eannot be analysed based oncontact angle values obtained at equilibrium".

Leja (1982, p.7) says:

"in the opimon of the author, the contaet angle is anindieator, but is not a measure of the hydrophilic character[of an ore]".

The seminal work on the signifieance of contact angles in fiotation has beendone by Gardner and Woods (1973, 1974, and 1977), and summarised byWoods (1976, 1977, 1981). The key to their work was the design of a cellwhich allowed the simuitaneous measurement of contact angles and a surfacepotential on an eleetrode, and a miero-fiotation device equipped with anelectrode capable of transferring an electrieal potential to a bed of mineral ormetal particles whieh permitted dynamic measurement of electrochemicalpotentials and flotation reeoveries. The initial work was with gold andplatinum, which was then extended to pyrite and galena surfaces.

Based on contact angle measurements, their major eonclusions were:

"The presence of dixanthogen on the [platinum or gold]surface clearly renders the surfaee strongly hydrophobie....when xanthate ions are adsorbed on platinum at potentialscathodic to dixanthogen formation, the contact angle in thisregion remained close to zero and henee specifical1yadsorbed xanthate ions do not make the surfacehydrophobic" .

In addition they found:

"It is interesting to compare the results of investigations ofthe interaction of xanthate with a galena surfaee and withlead. Lead methyl, ethyl and butyl xanthates on the metalsurface were found to be hydrophilic even at low eoverage.Only at high potentials, where lead xanthate is oxidised to

Mechanisrn of Flotation

dixanthogen, were significant angles observed. H thepotential was decreased after the dixanthogen was Iormed,the immediate contact angle rernained that characteristic ofdixanthogen. However, dixanthogen slowly reacts with thesurface to reform lead xanthate, the contact angle decreasedand the surface became hydrophilic".

And finally very pertinent to the thesis in this text:

"It has been generally considered that chemisorbed xanthatewill be attached to a metal atom in the sulfide surface.However, it has been pointed out by Winter (1975) thatxanthates can form bonds with sulfur, compounds of thetype ROCSS-S-SSCOR being quite stable. Also, suchcompounds cannot readily be distinguished fromdixanthogens by spectroscopic analysis and could weIl beformed at sulfide surfaces. The possibility of involvement ofsuch compounds, and the instability of surface species whichgive rise to the release of monothiocarbamates intosolution, add complications to the general picture of theinteraction of sulfides with coIlectors...u

15

Data on contact angles fOI xanthates from which many of the generalisationson the properties of coIlectors have been based are reported by KIassen andMokrousov (1963), and surnmarised in Table 5. Extending the range of thedata in Table 5, Taggart (1947) cites a series of contact angle measurementswhere the nurnber of carbon atoms in the xanthate alkyl group went from 1to 16, and the contact angle increased from 50 for methyl xanthate, steadilyto 80 fOI amyl, 87 fOI hexyl, 90 fOI 7 carbons, 94 for 8, and 96 for 16carbons (cetyl).

Statements in text books that longer chain alkyl group xanthates are "strongercollectors'' are based on these contact angle values. In fact, beyond sorneminirnum bond strength between the mineral and the bubble, persistence ofadhesion is onIy a function of the froth rheology and turbulence. What doescorrelate with contact angles for xanthates is that longer chain alkyl groupscoincide with lower selectivity, so amyl xanthates wiIl result in the productionof a larger volume, lower grade, rougher concentrate.

\

Xanthates with alkyl chains longer than hexyl provide lower flotationrecoveries. Decyl and longer chain xanthates are good depressants. A weIlknown, but generaIly unpublished, property of xanthates is that they may beused as emergency copper depressants in moly/ copper circuits. Operators


habitually correct for a momentary ineffectveness of the inorganic Cudepressant by the addition of more primary copper collector, be it a xanthateor xanthate derivative. Curiously, the amount of additional colIector that willact as a depressant with a particular moly circuit pulp corresponds to thatamount which will generare a measurable bubble contact angle in the lab.An nteresting experiment (which 1 believe has not been performed) wouldbe to see if a colIector concentration which generates any contact anglecorresponds to an over oiling dose industrially,

We have commented on "over oiling", a term which was first used around thetime of World War 1 to describe the phenomena that excessive amounts ofcolIector in a flotation systern will result in a decrease in recovery. Thisdecrease in recovery with increased dosage tends to be ignored in publishedpapers on flotation research.

The reason for this omission seems to be that Hallimond tube and othermicro-flotation tests do not detect the drop in recovery with excess collectorand/or frother, and for many of these papers the only flotationmeasurements carried out have been on a micro-scale.

TabLe 5. XANTHATE CONTACT ANGLES FOR SELECTED SULPHIDE MINERALS

N of Carbon GaLena ChaLcopyrite Bornite Pyriteatoms in xanthate

1 O O O O2 59 60 60 603 68 69 68 674 74 78 73 745 88 90 86 826 100 94 95 95

Effect of Branching

ButyL XanthateNormaL 74 78 73 74Branched 77 78 78 78

The Hallimond tube has been very popular in research laboratories because iteliminates the cost of assaying head, tailings and concentrate. The need toassay is avoided by studying artificial ores made by mixing pure minerals.The most common system has been a mixture of galena and sand (or purequartz). To perform a flotation experiment in a Hallimond tube, a verydilute pulp (usually about 1% solids) is placed in the bubble generating unit,which is the size of a test tube. The tube has a side arm welded to the upper

Mechanism of Flotation 17

regin; to operare, the unit is tilted until the liquid is level with the side arm,and air is turned on to the bubble generating device (usually a glass frit).The concentrate overflows into the side arm, is collected for a fixed time, andweighed. As a frother normally sequesters collectors and interferes with theoperation of Hallimond tubes, they are usually omitted. Thus we have theanachronism that a device without a controllable froth is used to simulatefroth flotation.

Because Hallimond tube flotation does not require the use of a frother, theeffect of frothers on flotation performance also tends to be ignored in artic1esreporting fundamental research, In addition, many of these articles onlyreport the collector concentration in the water phase, and do not inc1ude thepulp density. Therefore, it is important to translate the published reagentdose shown in recovery curves into mill terms (grams or pounds per ton ofdry mineral), using a best estimate of pulp density (1 - 30%?) if unspecified,when evaluating whether the information reported has any possibility ofbeing pertinent to mill problems. Typical sulphide ore flotation dosage ofcollectors used industrialIy are about 0.02 to 0.2 pounds per short ton of ore,or 10 to 100 gm per metric tonne of ore. These ranges translate toapproximately 5 mgjl to 50 mgjl or 0.029 to 0.29 mol/litre of the waterphase, for the molecular weights of the typical xanthates studied.

Nascent bubble flotatioo

Klassen and Mokrousov, in their "Introduction to the Theory of Flotation",postulate that the firstbubble to adhere to a mineral particle is a micro-bubble (formed by desorption of dissolved gases) which serves as a bridge forthe adherence of a bubble large enough to be capable of levitating theparticle. Klassen proposed this mechanism based on a thermodynamicanalysis showing that the bridging micro-bubble accelerates macro-bubbleattachment rates. Taggart, already in 1930, had observed that it was probablethat micro-bubbles were involved in flotation, and that they were formed onthe mineral surface by precipitation from dissolved gases in the cavitationregion of the pulpo Summarising work published over the past 20 or 30 years,flotation is thought to occur because bubbles form on the surface of avaluable mineral partic1e at points selectively covered by frother-collectormolecules complexes. Simultaneously, secondary bubbles are formed atmiscellaneous electrical discontinuities that do not provide strong bondingbetween the partic1e and its bubble; i.e. in the case of gangue partic1es.Possibly, if Klassen is right, the mineral particles first nuc1eate micro bubblesgenerated from dissolved gases, and these micro-bubbles then act as bridgesfor the formation of larger bubbles or the capture of larger bubbles thatcould be recirculating in the impeller zone. The important point is that, in

18 FLOTAnON - Theory, Reagents and Ore Testing

the flotation process, there is arate limiting step, which is the velocity of theformation, or attachment, of a micro- or macro-bubble to the mineralsurface. Once the bubbles have attached themselves to ore partic1es, thebubbles rise in the pulpo First they form a lattice which appears like a loosesponge, then later the bubbles tighten into a froth. This tightening processresults in the squeezing out of a great part of the occ1uded water, whichreturns to the pulpo This returning water mechanically c1eans the froth anddetaches the gangueo During this c1eaning stage there may be sornesecondary capture of valuable mineral that has fallen from bubbles in thestable froth region, but this secondary capture is not a controlling step in theoverall mineral recovery process.

In an agitation flotation machine, air is normally drawn into the impellerthrough a hollow shaft by mechanical suction induced by the rotatingimpeller, or with a positive air supply. After the air reaches the impellerregion, the flotation bubbles form by ,a two-stage process. First, cavitation"sausages'' form at the trailing edges of the impeller blades. These break upto form bubbles on the surface discontinuities of the mineral partic1es. Thesediscontinuities can be chemisorbed frother-collector complexes adhered tothe mineral surface, or naturally hydrophobic areas.

Grainger-Allen (1970) studied the formation of froths in a flotation machineequipped with flat rotor blades. He employed a cryolite pulp where therefractive index would allow photography of the bubble formation process.His high speed movies confirmed that the bubbles are formed from cavitationinduced air sausages which break up into bubbles onto the mineral partic1es(shown schematically in Fig. 1).

Fig.1.- Cavitation mechanism ofbubble formation

Electron transfer to mineral surface

Sulphide collector adsorption on to a sulphide mineral requires the transferof two electrons to the surface of the mineral partic1e. This electron transfer


mechanism is very similar to the electrochernical process of metal corrosion,During the corrosion process, metal goes into solution at a point on itssurface in contad with water, where the electrolyte is lacking in dissolvedoxygen; to go into solution the metal atom becomes a cation by losing thenumber of electrons corresponding to its normal valance. To maintain aconstant surface potential, these electrons must travel through the conductivemetal to an o),.)'gen rich region where they take part in an oxidation reaction,the result of which is that metal cations in solution, that have diffused fromthe point of dissolution, are converted into insoluble metal oxide; so, in ironcorrosion, the rust is deposited far from the point where a pit is beinggenerated. In flotation, the electrochemical process is very similar. Thecollector molecule approaches the metal surface and is adsorbed only after itdonates or has donated electrons to the mineral partic1e. So as not togenerate a charge on the partic1e, the excess electrons migrate to an oxygenrich point on the partic1e surface where they are consumed by reduction ofadsorbed oxygen to form OH- ions, or H20, or H202, raising the pH, aphenomena very easily measured during the adsorption of collectors onminerals. Under most practical flotation environments, deaerated pulps willDOt float, and it is thought that sorne depressants (sodium sulpbide, forexample) may block flotation by scavenging oxygen. The reaction involved inxanthate adsorption on a mineral, see Woods (1976), are therefore:

ELECTROH TRAHSFER ro A "IHERAL SURFACE

step 1 s..

s..

s..

s su n

2 (RO C-S >ads---+ (RO e-s>2 + 2 e-ol'

step 2

s s" _ u

(RO e-s ~ds + RO C-Ss~l;;-+ (RO e-s>2

These reactions are ba Ian ced electrochenically by:

~02 + 2H+ + 2e- ~ H20~02 + H20 + 2e ~ 20H-

02 2H+ Ze- \ HZOZ+ + ~

Indirect evidence of this last reaction has been provided by Jones andWoodcock's (1978) discovery of perxanthates in tailings pulp from flotation

20 FLOTATIN - Theory, Reagents and Ore Testing

mills in Australia. They postulare the reaction to form perxanthates in aflotation eell to be as follows:

s s

2ROC-S + H202 --~ ROC-S-O- + H20

and point out that the only source of hydrogen peroxide in a pulp would bethe last reaetion aboye.

The nature of the dixanthogen forrned will depend on the metal cationinvolved. In the case of galena the infrared speetra show that the anodiereaction involved is

s s

PbS + 2ROC-S + 1/202 + 2H+ ----~ PbCROC-S)2 + So + 2e

The electrical double layer theory, originally proposed by Helmholtz in 1879,evolved from a postulate that when a erystalline solid surface is generated bya fracture, the eleavage process severs the ionic bonds and leaves anelectrieal imbalance on the surface, and, if the solid is in contact with water,this imbalanee is compensated by a eharge gradient. This primitiveHelmholtz model was refined by Perrin in 1904, and is shown in Fig. 2 as case(a), labelled as the compact layer theory; illustration (b) is based on Gouy(1910) and Chapman (1913) who refined the model to inelude dilutesolutions, based on the assumption of the existence of a diffuse continuum ofions in a non-structured die1eetrie layer around the mineral partiele; (e) addsion interaction as proposed by Debye and Huekel (1923) and Stern (1924),which ineludes the premise that the individual ions can approach the mineralsurface up to a minimum distance "d". This theory in fact is a triple layertheory, as can be seen from the illustration. While (d) corresponds to a casewhere a specific ion in solution has entered the double layer and resulted in acharge reversal.


IHP OHP

-1-+.L~'" sx Water dipole in

the downposition

'"..

".Sz::

Water d ipo Ie inthe up position

IHP= Inner- He Isoho ltz P IemeOIHP= oueee He Iraho ltz p Inne

Solvated cations

edsorbed anion in IHP

So Ioa't.ed cat ion:

Soluated anian

ELECTRICAL DOUBLE LAYER110DEL

DlP OHPb'" Ila4 II _I L '1'

,1 I~ I---r--- 'I'dI II I .,I I '1'5

Helmholtz-Perrinc:onpact layer

Gouy-Chapl'l1an d iffuselayer

Stern doub le la!Jl'r

(a) (b) (e)

Charge reuersa 1because ofadsorp-t ion af~-:--::;; ~ counterions

(a)

\

Fig. 2.- The electrical double layer model

22 FLOTATIN - Theory, Reagents and Ore Testing

The structure 01 water is also a factor. Most theories 01' the rnechanism offlotation stick to the electrical double layer theory as an adequate descriptionof the influence of the water phase on collector adhesion to a mineralsurface. Possibly it is an over-simplification, as the structure of water iscontroversial. Linus Pauling suggests that for pure water, at ambient andlower temperatures there is considerable regularity to the structure; betweenO and 4 he suggests that the crystalline regularity 01' water moleculescorresponds to a mixture of ice 1 crystals (Fig. 3), which has a density of 0.92,mixed with ice Il, a structure similar to quartz, which has a density of 1.17. Asthe temperature approaches the maximum density of water, at 4C, therelative proportion 01' ice II rises. Above 4C, water becomes moreamorphous, the crystalline regularity becoming more and more discontinuousas the temperature increases.

Fig. 3.- The structure of ice, showing the orientation of the hydrogenbonding, adapted from Pauling (1967).

Mechanism of Flotation

Lone e lect.ronpairs

23

Fig. 4.- Localised-bond structure for H20 with sp3 orbitals for oxygen

Fig. 4 shows the spatial arrangement of the orbitals in water, and theexplanation for the abnormally high dipole that this molecule exhibits. Inmineral pulps, the aqueous phase is alkaline and contains common salinesalts in solution, usually including calcium, magnesium, silicates, sodiumchloride and excess hydroxyl groups. Schematically, the orientation of thewater molecules in the hydrated anions and cations will be as shown in Fig. 5.Laskowski proposes that the hydration of a mineral surface in an impureaqueous solution will consist of multilayers of water molecules ordered intoice type crystals with stray water molecules and small hydrated ions captivewithin the clathrate structure. Presumably the value of the measuredelectrical double layer charges reflects the energy necessary to penetrate thishydration layer.

The topology of the layer may well be even more complicated thanLaskowski's model, as concentrated transition metal salt solutions are knownto have peculiar solubility characteristics . which reflect a crystal-likeregularity. For example, a 60.2 weight per cent zinc chloride solution, aconcentration which corresponds to 5 molec~les of water per molecule ofzinc chloride, becomes a perfect solvent for polyacrylonitrile polymers, whilea solution with less than a 45% zinc chloride (a nine to one mole ratio) is anon-solvent (see R.D. Crozier, et al, U.S. Patent 3,346,685, 1967). X-raydiffraction data on the concentrated solutions gives diffuse crystalline

24 FLTATIN - Theory, Reagents and Ore Testing

scattering, though the liquid ordering involved has not been determined,

The electrical behaviour of metal sulphide minerals is more complicated;rnost are semi-conductors, so their electro-potential properties can varyconsiderably from point to point on the particle surface as a function of thecation, the crystal structure, and the homogeneity of the ore. For theseminerals the flotation process is complex and very difficult to predict fromsimple laboratory experiments.

sodium cation chlorine anion

Fig. 5.- Cation and anion hydration showing water orientation

Jan Leja (1982) provides an extensive review of the mechanism of adsorptionof fatty acids, sulphonates and amines on inorganic and metallic ores, andprobably the most extensive, easily available reference on the adsorptionmechanisms. In addition, there is considerable recent literature on zetapotential and other electrochemical sweep experiments, with interpretationof the results in terms of flotation conditions. Though, as virtually all of thedata is based on pure, or at least selected mineral samples, the results oflaboratory measurements are at best indicative rather than quantitative, ifextrapolated to full scale equipment.

Chapter :3

SUlPHIDE MINERAL FlOTATION

Stereochemistry of metal xanthates and dthiophosphates

An essential factor in the evaluation of spectroscopic data of the compoundsformed by collectors on mineral surfaces is the stereo-chemistry of thecomplexes that the transition metals can form with thiolates in solution.These complexes are particularly important in the interpretation of theexistence of dithiolates on the mineral surface, as the complexes themselveshave sulphur-to-sulphur bonding, and many have active lone pairs.

Six coordination complexes

Iron, nickel, cobalt, chromium, Mn IlI, and CuIl can form six-coordinationcomplexes with tris(bidentate ligands), with geometries as illustrated in Pig. 6and 7. Structural parameters for xanthates and dithiophosphates are reportedby Kepert (1977). Por the ferric compound, the constants are b = 1.22 or1.23 and G = 20.5 or 21.8, according to three different investigators.

The reaction of the formation of the aboye bidentate complex is:

But the reaction is known to continue, through self reduction, to produceanother, more massive, six-coordination complex and dixanthogen:

8[FeI I I(EtX)3] -------~ 2[FeII(EtX)3]3FeIII"+ 3EtX-XEt\ dixanthogen

A pseudo four-coordination dithiolate complex is formed by CuIl. In the caseof copper xanthate, it forms a puckered square planar equatorial girdle oftwo xanthate molecules and either water or a bridging ligand in the axialposition. It is considered pseudo four-coordination because the axial bond is

- 25-


CompLex

Xanthateplanar

b

1.341.321.241.221.211.191.211.19

Edge length S andheight h

"Zfor [lHb identate)3 1

24.723.819.921.118.716.820.320.1

Fig. 6.- General stereochemistry for [M(bidentateh]X

much longer (3.5 A) and weaker than the sulphur bonds. The cupricreactions are:

which react further, in solution:

[CUI(EtX)2J2cuII + EtX-XEtdixanthogen

This complex will hydrate to form a bridging six-coordination cuprouscomplexo The water molecules, though, can be substituted by copper ions,For both the cupric and the cuprous bidentate complex, the geometry isshown in Fig. 7.

Seven coordination complexes

Seven coordination complexes, in all cases of interest in flotation, involvelone electron pairs in one of the coordination axis. These electrons can betransferred to the mineral surface or can be used to form polymericcomplexes in the pulpo The most common geometry of the sevencoordination complexes are the pentagonal bipyramid and the cappedoctahedron shown in Fig. 8.

Sulphide mineral flotation 27

)('antlha-t:e

/:.:. ;"'::::"':::" ~_-l/

s

IIr ce (xanthate ) 2]

e ..

POLYMER Cu

According to Mumme and Winter (1971), as reported by Drew (1977),bivalent lead forms a seven coordination complex with a lone electron pairwhich can polymerise. The opinion that polymer formation can occur isbased on examination of samples where "close to the axial positionpresumably occupied by a lone pair an equatorial S atom of anothermolecule" was found. Drew describes the complex as a pentagonalbipyramid. The five positions in the equatorial plane surrounding the metalatom are at a shorter bond distance from the central metallic atom than thetwo axial locations. Certainly the existence of a complex will be sufficient toconfuse surface infra-red spectra, and definitely it will show up as the lead

Pentagonal bipYl'a~id Capped octahedl'onFig. 8.- Pentagonal bipyramid and capped octahedron geometry


xanthate salt if leached from the concentrate by an organic solvent. Similarstructures, with lone pair bridging, are produced for [Sb(dedtchOone pair)],[Bi(dedtc>J(lone pair], and [Te(ethyl xanthatej-flone pair)], where dedtc isdiethyl dithiocarbamate.

Dithiophosphates are more likely to produce a polymeric structure with lead,where the bridging bonds between the sulphur atoms in two complexes areconsiderably longer than the equatorial plane Pb ones (3.232 A, 3.175 A,compared to 2.761 A, 2.772 A, see Fig. 9). The sketch of a trimer (Fig. 10)shows that this complex wiIl have more available electron pairs to form bondswith the mineral surface, which may explain the preference fordithiophosphates as coIlectors for galena.

Note that for the trivalent metals the geometry of the complexes are cappedoctahedrons. The evidence on polymer formation is tenuous, which probablymeans that xanthates reacting with lead cations in solution will usually onlyform dimers and trimers at the mineral surface. The complex wiIl haveavailable Ione electron pairs to transfer to the mineral particle.

s~~~;---- XAHTHATE~ L-.~~~~~) _~ep'~.,.

Sulphur ai:ol"ls .------...

[PbIICci:h!:>ll xAntnato)3 Clono pair)]

Fig. 9.-[PbII(ethyl xanthate)3(lone pair)]

DltlER

Trivalent arsenic, antimony, telurium, bismuth, and vanadium form a verysimilar complex with xanthates and dithiophosphates, except that they haveseven coordination because there is a lone electron pair protruding out ofone of the faces in the axial position. Fig. 11 shows the capped octahedronstructure in a diagram where dirnensions are easier to visualise, and sornetypical key dirnensions for arsenic and antirnony xanthates and a bisrnuthdithiophosphate reported by Kepert.

Sulphide mineral flotation

lone pail:'

Seven Coordination COMPlexunits fOPMing polYMep

Fig. 10.- [PbIl( di-isopropylphosphorodithiolateh(lone pair)]

Dithiolate chemistryand surface compounds

29

Dixanthogens: before going on to analyse the Finkelstein and Poling (1977)review of the nature of the compounds formed on the surface of mineralswhen they react with collectors, it is important to c1arify the probable identityof the dixanthogens mentioned in the older flotation literature. Sixty yearsago Cambron and Whitby (1930) were pointing out that reacting xanthateswith iodine or chlorine produced impure liquid dixanthogens, while a moreselective oxidiser, such as sodiurn tetrathionate, will produce soliddixanthogens without the need to recrystallise. More recent reviews, amongthem Reid (1962), have stressed that the melting point is variable, and anindicator of purity. Table 6 contains representative melting point and boilingpoint data on common dixanthogens, mostly from Reid, but recheckedagainst unpublished data taken by Professor J.C. Vega de Kuyper at theCatholic University of Santiago, Chile, in 197~/81.


5StructuraL Parameters of Tris(bidentate) (Lone pair) CompLexesCompLex M- a M- b aMa bMb aa ab

[AS(S2CNEt~)1] 2.35 2.85 90.1 106.0 54.8 67.2[As(S2COEt 3 2.28 2.94 92.0 107.6 56.2 68.7[Sb(S2COEt)3] 2.52 3.00 87.5 113.1 53.0 74.5[Bi{CS2PCOPr)2}3] 2.70 2.87 91.2 99.9 55.6 62.1

Fig.11.- [M(bidentateh(lone pair)jX:!:

All the flotation literature surveyed, including recent compendia such as Leja(1982), always refer to dixanthogens as oily liquids. This suggests that impureproducts have been used in flotation experiments, because methyl, ethyl andisopropyl dixanthogens are solids at room temperature, as is themercaptobenzothiazole dithiolate. The ethyl and mercaptobenzothiazoledithiolates, which are readily available commercially, are yellow solids, in allgrades.

The 1983 Merck Index, under dixanthogen (compound 3402) lists as alternatenames thioperoxydicarbonic acid diethyl ester, or dithiobis[thioformic acid]O,O-diethyl ester, and describes the product as yellow needles, mp 28-320 ,onion like odour, solubility in alcohol 2g/100 mI, freely soluble in benzene,ether, petr. ether, oils. Almost insoluble in water. Use: insecticideformulations, herbicide, topical parasiticide. Under dithiobis[benzothiazole](monograph 3389) it also lists the dithiolate of mercaptobenzothiazole aspale yellow needles, density 1.5, mp. 1800 , (minimum commercial


specification, mp 168) insoluble in water, use: rubber accelerator. Underdixanthogen, the latest edition of Hawley's Condensed Chernical Dictionarylists ethyl dixanthogen (CAS: 502-55-6) uneler the trade name "Sulfasan", asyellow crystals, mp. 30, insoluble in water, and used as a herbicide,insecticide, parasiticide, It also lists the mercaptobenzothiazole dithiolatewith CAS: 120-78-5, yellow salid, properties, d= 1.54, mp. 168, use: rubberaccelerator.

TabLe 6.- PROPERTIES OF DIXANTHOGENS AND XANTHIC ANHYDRIDES

s S

Xanthic anhydrides = R-O-C-S-C-O-R

MethyLEthyLIsopropyLPropyLter-ButyLIsobutyLAmyLHexyL(MBT)2

ye LLow so LidyeLLow soLidyeL Low soLidbrown oiLbrown oiLyeLLow oiLyeL Low oi LyeLLow oi Lunknown

54-5552-5554-55557 1367

77 d=1.12677 ND=1.541

Dixanthogen

S S

R-O-C-S-S-C-O-R

MethyLEthyLIsopropyLPropyLter-ButyLIsobutyLAmyLHexyL(MBT)2

R=CHrOR=CH3CH2-0R=(CH3)2CH2-0R=CH3(CH2)2-0R= (CH3)3C-OR=(CH3)2CH2CH2-0

yeLLow soLid 23-23.5 907yeLLow soLid 32 1077 d=1.2604yeLLow soLid 57.5-58.5yeLLow oi L 117yeLLow soLid 58yeLLow oi L 165 d=1.0S

yeLLow oiL ND=1.5569yeLLow soLid 180

A caveat in reviewing very old work on elittiolates is that in pre-WW 11Iiterature there is a confusion between eliethyl elixanthogen and Minerec A.Por example, Taggart, in his Hanelbook of Mineral Dressing (1945), says onpage 12-11: "Minerec is principally dixanthogen, formed from treatingxanthate with ethyl chlorocarbonate, which is a strong oxieliser"; but Minerec


A is not a dixanthogen, it is an impure xanthogen formateo Taggart's errorarises from reading a rather deceptive U.S. patent of AH.Fischer's (1928),that was a consequence of a patent priority problem between GuggenheimBrothers and du Pont, Guggenheim Brothers had hired Dr. AH. Fischer todevelop an improved collector for their El Teniente mine. Dr. W. ADouglass (1927), working for du Pont, was simultaneously carrying out ageneral programme on testing xanthate derivatives as improved col1ectors.He obtained a clear priority on a xanthogen formate patent in the U.S.A, sothe only U.S. patent issued to Guggenheim Bros. did not cover xanthogenformates, ergo the misinformation of Taggart and others. Minerec thenlicensed the du Pont patent, and later was the only commercial producer ofxanthogen formates.

Physical or chemical adsorption of a col1ector to a solid surface involves aninteraction of e1ectrical forces; each type of bonding involves a reduction infree energy of the system, and the production of a definite amount of heal.The difference is that in physical adsorption the bond is amorphous, while inchemical adsorption there is a directional bonding involved. Metal1urgical1y,the differences between the two bonds is that, in physical adsorption the bondis weak and easily reversible, so a reduction in the concentration of theabsorbent in the liquid phase will desorb the species from the surface, whilechemical adsorption is irreversible and normally localised on active sites ofthe mineral surface.

Experiments performed over half a century ago demonstrated that, whenthiol collectors are employed to float sulphide minerals, the collectionmechanism involves the transfer of electrons to the mineral particle and theformation of dithiolates. What has remained controversial has been whetherdithiolates are always involved in the collection mechanism or whether insorne cases the metal xanthate is formed and deposited on the mineralsurface. Based on the premise that the active collector is a metal xanthate, anumber of metallurgists have suggested that the effectiveness of a particularxanthate for the collection of a specific mineral could be deduced from thesolubility of the metal xanthate.

The role of alkyl dithiolates in flotation systems, and the type of surfacebonding involved, was extensively reviewed by Finke1stein and Poling (1977).The relevant dithiolates they studied are shown in Table 7.

They devoted much effort to evaluating the considerable volume of publisheclinformation on the flotation of galena with xanthates. Among the morebizarre theories put forward that they evaluated was that the interaction of axanthate with galena resultecl in the formation of elemental sulphur on the


Table 7. - DITHIOLATE REAGENTS: N0t1ENCLATURE AND ABBREVIATIOI~S

33

DithioLate Corresponding Thiol

Name FormuLa Abrv Name Abrv

DiaLkyL dixanthogen ROCCS)SSCS)COR )(2 ALkyL xanthate )(

CRO)2PCS)SSCS)PCOR)2 CDTP)2 DiaLkyL dithio- DTPphosphate

Bis CdiaLkoxyphosphono-thioyL) disuLphide2,2' DithiobisbenzothiazoLe s S

\ /CSSC

/ \N N

CMBT)2 Mercaptobenzo- MBTthiazoLe

mineral surface, and that it was the natural floatability of the sulphur thatcaused the flotation of galena. More important was the controversy that, inall cases, for a xanthate to act as a collector, it must have reacted to formdixanthogen. The analysis is best summarised in their own words:

"As Granville, Finkelstein and Allison (1972) noted, the casefor dixanthogen as the essential hydrophobic entity restedprimarily on the electrochemical evidence, in essence, that abubble only adheres strongly to galena when the potential ofthe surface is hi;h enough to allow dixanthogen to beformed. The same is true at the present time. .... Woods(1977) identified two stages in the reaction between thexanthate and galena. At potentials between -0.2 V and + 0.2V on the hydrogen scale, a reaction that he identified aschemisorption took place. At potentials more anodic than0.2 V, a reaction took place that increased in rate steadily asthe potential was raised, until it was limited by theaccumulation of producto This reaction is the formation ofdixanthogen....."

After the evaluation of considerable contradictory data on the collectorcomposition found on galena surfaces, including infra-red analysis of galenatreated with dixanthogens, they concluded thatxhe active species on galena isthe lead xanthate. The work they reviewed on dixanthogen adsorptionshowed that galena reacts as follows:

PbS + )(2 ~~ PbX2 + So

34 FLOTAnON - Theory, Reagents and Ore Testing

This explains why sorne 01' the early research suggested that elementalsulphur could be the active collector on galena al'ter exposure to xanthates,

Finkelstein and Poling also reviewed the available literature on surfacereactions 01' dithiophosphates and mercaptobenzothiazole with differenrminerals. They noted that dithiophosphates are weaker collectors than thecorresponding xanthates, and that they are less readily oxidisable, as shownby their dithiolate./thiol reduction potentials (Table 8), and that their leadcomplexes are more soluble than the corresponding lead xanthates. Theyalso concluded that the active coUector on the galena surface is the leadthiolate, as is the case with the data they could examine formercaptobenzothiazole. The reduction potential data for the xanthates anddithiophosphates is 01' interest because, at least theoreticalIy, it can becompared with the rest potential 01' the mineral in water to see if the reactionwil1 occur. The general conclusions 01' Finkelstein and Poling on the coUectorspecies present on the mineral surface at flotation is summarised in Table 9.

TabLe 8.- REDUCTION POTENTIALS FOR DITHIOLATE/THIOL COUPLES

HomoLogue Xanthate DTP HomoLogue Xanthate DTP

methyL -0.004 0.316 isobutyL -0.127 0.158ethyL -0.060 0.255 amyL -0.159 0.050propyL -0.091 0.187 isoamyL n/d 0.086isopropyL -0.096 0.196 hexyL n/d -0.015butyL -0.127 0.122

Finkelstein and Poling's most interesting conclusion in their overal1 survey 01'published data on infra-red surface analysis (which predates Woods review)is that the active species on galena is alead xanthate. As there is considerableevidence that the lead atom in galena is labile, there should be Pb++ presentin solution in flotation pulps, so the lead xanthate identified on the mineralsurface is likely to be the seven coordination complexes depicted in Fig. 9.These complexes have lone pairs which can be donated to the mineralsurface.

Woods' 1977 analysis, based on voltamograms for galena, and the conceptthat significant contact angles are essential to flotation, is as foUows:

"The more important question is whether the presence 01'dixanthogen is essential for flotation to occur. It is apparentthat flotation takes place with a number 01' mineral-coUectorsystems without disulfide formation. A problem to be


considered for the interaction of xanthates with galena lesin the faet that bulk lead xanthate is hydrophilic andmultilayers of this compound on galena itself give rise to alow contact angle.

The electrochemical measurements indicate that the initialchemisorbed xanthate layer is hydrophobic and thatflotation commences at potentials where this species isformed with only small quantities of dixanthogens beingpresent. The chemisorbed layer on galena, therefore, has adifferent effect on the superficial tension from the leadxanthate and could be the important species in flotation ofgalena; this does not exclude the possibility thatdixanthogen also plays an important supporting role."

35

TabLe 9.- PRODUCTS EXTRACTED FROM THIOLATED SURFACES(FinkeLstein and PoLing (1977) )

MineraL Product formed on mineraL surface by coLLector

Name FormuLa EthyL PropyL ButyL AmyL HexyL Dithiophos MBT

Orpiment AS2S3 + Stibnite AS2S3 @ @ @ @ @ eReaLgar AsS @ e @ @ + @ @Cinnabar HgS @ @ @ @ +SphaLerite ZnS @ @ @ @ + @ @Antimonite AS2S2 @ + + + + @ @ALabandite MnS @ + + + +Bornite CuSFeS4 @ + + + + Cu(DTP)2 Cu(MBP)2ChaLcocite cU2S @ + + + + Cu(DTP)2 Cu(MBP)2Galena PbS + + + + Pb(DTP)2 @

Pyrrhotite F~S8 I!l I!l I!l I!l I!l @ @Arsenopyrite FeAsS I!l I!l I!l I!l I!l @ ePyrite FeS2 I!l I!l I!l I!l I!l (DTP)2 (MBT)2

+Fe(MBT)ChaLcopyrite CuFeS2 I!l I!l I!l I!l I!l Cu(DTP)2 CU(MBP)2CoveLL ite CuS I!l I!l + + I!l + + I!l + + I!l++ Cu(OTP)2 Cu(MBP)2Mo l i bdeni te MoS2 I!l +? I!l + ? I!l + ? I!l + ? I!l Mo(DTP)x Mo(MBP)x

\

+ = metaL xanthate I!l = dixanthogen @ = no positive identification

36 FLOTA'flON - Theory, Reagents and Ore Testing

Thol colector bondng with metallic sulphde minerals

To evaluate the validity of f1otation theories, an important factor is to definethe type of bonding that exists between the surface of a mineral and the thiol;i.e. is this a sulphur bond with the metal? The evidence is that the collector-mineral bond is a sulphur-to-sulphur bond and not a sulphur-to-metal bond,as is generally assumed. The following quote from Ronald Woods (1977)provides a corroborating opinion on this very controversial subject:

"It has generally been considered that chemisorbed xanthatewill be attached to a metal atom in the sulphide surface.However, it has been pointed out by Winter (G. Winter,(1975)"Xanthates of Sulfur: Their Possible Role inFlotation", Inorganic and Nuclear Chemistry Letters, Vol.11, pp. 113-118) that xanthates can form bonds withsulphur, compounds of the type ROCSZ-S-S2COR beingquite stable. Also, such compounds cannot readily bedistinguished from dixanthogens by spectroscopic analysisand could well be formed at sulphide surfaces."

Strictly speaking, from the point of view of differentiating between flotationmechanisms, it is more important to demonstrate that the chemisorptionbond between the collector and the mineral surface is not a metal-sulphurbond, rather than that it is a sulphur-sulphur bond. For example, Clark(1974) points out that chemisorption on the surface of either an n or p typesemiconductor results in a bond with hybrid covalent and ionic characteristicswhich makes it non-specific to a particular atom. The electron availabilityfrom the double bonded sulphur atom of the thiono group depends on thefollowing equilibrium:

e 00S.... h S:

'e vI t-~---I>

:0:

'RAs the species is the anion of xanthic acid, it is obvious that the single bondedsulphur can provide an electron for an ionic bond. A very similar picture canbe given for the dithiophosphates and the > mercaptans, while thethionocarbamates, xanthogen formates and xanthic esters exhibit resonancein their sulphur-to-carbon bonding, as well as a similar resonance in thenitrogen or oxygen atom.


The authority for conc1uding that electrons donated by a collector to asulphide mineral form a sulphur to sulphur bond is Linus Pauling, In his "TheNature of the Chemical Bond", pages 442/448, he says:

"Many sulfide minerals have structures closely related tothose of sphalerite and wurtzite. Chalcopyrite, CuFeSb isan example (Fig. 12). Its structure is a tetragonalsuperstructure of sphalerite, with the copper and iron atomsin the zinc positions of sphalerite. Energite, CU3AsS4' has astructure that is a superstructure of the wurtzitearrangement, The sulfur atoms are in the same positions asin wurtzite, and the atoms of copper and arsenic replacethose of zinc in an ordered way, so as to give discrete AsS4groups (Fig. 13). The observed As-S bond length, 2.22 A,agrees exactly with the calculated value for a single bond,2.22 A (from the covalent radii, with the correction forelectronegativity difference). The Cu-S bond length, 2.32 A,corresponds to a bond number about 0.7 (the appropriatesingle-bond radius of copper is 1.23 A). Approximately thesame Cu-S bond length is found in other copper sulfideminerals. The copper-sulfur bonds have only a smallamount of ionic character, and the conclusion may be drawnthat the electric charge of the copper atom is negative,probably close to -1."

Pauling makes a similar analysis for a number of other sulphide minerals. Inall cases he reaches the same conclusions, that in transition metal sulphidecrystals the metal is negatively charged, and the sulphur atom is positive andan el

flotation, theory, reagents and ore testing-ronal d. crozier

Documents

ore sampling

ore testingrala

ore testingbyronald

ore testingflotation

mineral surfacechapter

mineral processing industry

particle capture of

manufacture methods