《高等选矿学》课程教学资源（文献资料）Anisotropic minerals in flotation circuits

CANADIAN MINERAL PROCESSING 203 07). Anisotropic minerals in flotation circuits ough J.S. Laskowski East Norman B. Keevil Institute of Mining. University of British Columbia, Vancouver, British Columbia, Canada n,G. sses: New ABSTRACT Anisotropic minerals are important constituents of many ores. This group includes both valuable minerals (e.g., molybdenite in Cu-Mo ores) as well as gangue minerals (e.g., talc in platinum- bearing sulphide ores in South Africa, graphite in Cu-Ni sulphide ores in Canada, chrysotile in Ni sul- phide ores in Australia, and clay minerals in all types of ores). Aqueous suspensions of anisotropic minerals exhibit different properties than suspensions of isotropic minerals. The presence of anisotropic minerals in flotation circuits affects the flotation process. KEYWORDS Anisotropic minerals, Clays, Talc, Molybdenite, Flotation, Pulp properties, Rheology RESUME Les mineraux anisotropes sont des constituants importants de nombreux minerais.Ce groupe comprend a la fois des mineraux de valeur(p. ex. la molybdenite dans les minerais Cu-Mo) et des mineraux de la gangue (p. ex. le talc dans les minerais sulfures contenant du platine en Afrique du Sud le graphite dans les minerais Cu-Ni du Canada, le chrysotile dans les minerais sulfures de nickel en Australie et les mineraux argileux dans toutes sortes de minerais). Des suspensions aqueuses de mineraux anisotropes possedent des proprietes differentes de celles des suspensions de mineraux isotropes. La presence de mineraux anisotropes dans les circuits de flottation affecte le processus de flottation. MOTS CLES Mineraux anisotropes, argiles, talc, molybdenite, flottation, proprietes de la pulpe, rheologie INTRODUCTION Mineral crystallochemistry, as shown by Gaudin, Miaw, of zero charge) varies from one oxide to another and corre- and Spedden(1957), is responsible for the properties of the sponds to their relative strength as acid or base. At the pzc, solid-liquid interface, which are determined by the chemi- the concentrations of negative and positive charges are cal composition of the solid and the electrical charge of the identical: solid surface. [-MOH2+]=[-MO] (3) Isotropic minerals In the case of isotropic minerals, all sides of the crystal As these reactions reveal, the concentration of H+ and are created by breaking the same bonds, resulting in homo- OH ions determines the charge (and potential at the inter- geneous mineral surfaces that have identical electrical face)of oxides; these ions are referred to as potential deter- charges; an example of this is quartz. The new surfaces are mining for these minerals. created by breaking identical Si-O bonds when larger In all studies of particle-particle interactions in which pieces of quartz are crushed. As a result, all the new sur- the effect of attractive (dispersion) forces and repulsive faces have the same composition. The surface hydroxylates (electrostatic) forces are discussed (DLVO theory), form surface silanol groups (SiOH), which ionize as in sili- isotropic solid particles are used, which retain a uniform cic acid. In general, such surface hydroxyls are amphoteric charge independent of the distance between the two inter- and become positively charged in acids and negatively acting particles. This is shown in Figure 1， for two fine charged in alkalis. The charge results from the following quartz particles suspended in water. reactions, where M is a metal: The electrical repulsive forces are calculated based on zeta-potential measurements carried out in most cases with -MOH+H+=-MOH2+ (1) the use of electrophoresis. In this experiment, the elec- trophoretic mobility of the solid particles(Figure 2) is -MOH+OH=-MO+H2O (2) measured in an electrical field and this is used to calculate the zeta potential via Smoluchowski's equation. As a result, the surface acquires either a positive or neg- The driving electrical force, E·q, where E is the electri- ative net electrical charge. The switch-over point(pzc, point cal field and q is the electrical charge of the particles, is CIM Journal Vol. 3， No.4

5204J.S.Laskowskiformed by rupture of ionic or covalentbonds and is hydrophilic"The defini-tion used here is more general and doesnot necessarily require that one of thecrystal sides behydrophobic.Examplesof this are clay minerals, which are typ-ically anisotropic and are usuallyhydrophilic.Clays are the best known examples ofFigure I.Electrical repulsion between two ncgatively charged identical particles suspended inanisotropic minerals. However, molyb-water.denite, graphite, and talc are alsoanisotropic minerals and are inherentlyhydrophobic.While clays are very dif-ferent from molybdenite, graphite, andtalc,what theyhave in common is asheet structure (alsoreferred toas lami-nar crystal structure).The sheet structures of clay mineralsdiscussed in this paper are made up oflayers of silica tetrahedra condensedFigure2.Definition sketchfor electrophoresis showing a negativcly charged spherical particlewithgibbsite [Al(OH)J (inkaolinite) ormoving in electricalfield.brucite[Mg(OH),] (in talc).Kaolinite,aI:1 layer silicate,has two types of basalplanes, the tetrahedral Si-O (at the bot-tom of Figure 4) plane and the octahe-dral Al-OH (upper) plane (Figure 4;Carty, 1999).The 1:1 layers are heldtogether in the crystal by hydrogenbonds.The bottomtetrahedral planecar-ries negative electrical charge at all pHvalues as a result of isomorphous substi-tution of some Si4+ by A13+.The octahe-Figure3.Schematicofananisotropicparticle in an electricalfield.dral basal plane of kaolinite, as well asits edges, carries a charge that dependsopposed by a hydrodynamic friction given by a viscouson solution pHand arisesfrom thepresence of amphotericdrag (Stokes'law),and electrophoretic friction caused byAIOH groups on these surfaces. Thus, the“topochem-the oppositely charged ions moving in the directionistry"of the exposed planes diffcrs quitc significantlyopposed to that of the particle. The resulting Smolu-In montmorillonite, a 2:1 layer silicate, the hydrated alu-chowski's equation allows the calculation of the zeta poten-mina layer is sandwiched between a pair of silica layers,tialfrom themeasured electrophoretic mobilityof particlesand alumina is exposed only at the edges. Thus, for thisof a few microns in size.mineral, while the planes always carry negative electricalSpherical isotropic particles are considered in this typeof derivation and the question then arises of what would bethe result of the electrophoretic experiment if the particlesOCTAHEDRAL PLANEwere anisotropic and had different electrical charges on var-?OHious sides as shown in Figure 3. This question has not yetO+.been satisfactorilyanswered020SrAnisotropic mineralsoIn anisotropic particles,the surface charge on differentAl*sesides of the crystal is different (as in the case of clays).Thedefinition adopted by Chander,Wie, and Fuerstenau (1975)TETRAHEDRALPLANEis as follows:"An anisotropic surface consists oftwo broadtypes-one,which isformed by therupture of van derFigure 4. Structural basic unit cell ofkaolinite (l:1 layer aluminosilicate)Waals bonds and is hydrophobic, and the other, which isshowing tetrahedral and octahedral basal planes.CIM JournalVol.3, No.4

205Anisotropicmineralsinflotation circuits(1957),nativehydropho-walentbicity results when at leastdefini-some fracture or cleavaged doessurfaces form by ruptureof theof weak secondary bonds.imples-8OHowever, the edges arere typ-created by rupture of cova-isuallylent bonds and such sitesspontaneously react withples ofnolyb-watertoformhydrophilicM-OH sites.:alsoThecurled tubularerentlyHstructure of chrysotile isry dif-much more complex thane, andFigure 5.The edges of montmorillonite platelet.the:simpleedge/platenisastructure of clays. Ins lami-chrysotile[Mg,Si,O,(OH], the dimensions of the silicatetetrahedral layer are about 9% smaller than the correspon-ineralsding ones in the octahedral brucite layer (Figure 7). The:up of?OHimperfectfit of octahedral and tetrahedral layers causes thedensed020crystal structure tobend, curl,and formconcentrichollownite)orSit4cylinders. The bending of the sheets is continuous andinite,aOresults in tubes that give the mineral its fibrous naturefbasalMg*2(asbestos mineral).lebot-As chrysotile has a spiral shape,a tetrahedral-octahedraloctahe-TETRAHEDRALBASALPLANEedge is likely to occur at the end of each tube as well asure 4;along the length of it (Figure 7b).The tubes curl such thate heldFigure6.Basicstructuralunitoftalc(2:1layermagnesiumsilicate)the magnesium-rich octahedral layer is exposed on the out-irogenside. Therefore, the surface charge of this site is likely to benecar-similar to that of brucite,which is positively charged over aall pHcharge, only the edges may carry eithera positive or negabroad pH range. Due to this separation of charge betweensubsti-tive charge, depending on pH (Figure 5).edges and faces, as well as the flexibility of the chrysotileoctahe-The basic structural unit of talc is shown in Figure 6.Asfibres, they are ableto align themselves in a number of con-wellasshown by Burdukova, Becker, Bradshaw, and Laskowskifigurations.Thisresults inaverycomplextangledstructureepends(2007), because of the substitution of some Si4+ ions withand has adverse effects on chrysotile slurry rheologyhotericA13+ and Ti3+ in talc tetrahedral layers, these basal planesThe crystallochemical structures of graphite and molyb-ichem-exhibit negative electrical charge. However, the charge atdenite, two inherently hydrophobic minerals, are shown in Figly.the edges depends on pH. The most important differenceure8.In molybdenite,sheetsofmolybdenumatomsareed alu-between talc and the clays is that it is inherently hydropho-sandwiched between two sheets of sulphur atoms.The sulphurlayers,bic.In talc, the layersorthisof silica tetrahedractricalare held together bybrucite (O)outer layervan der Waals bondsandthebreakingTetrahedral Planeprocess proceeds byOHrupturing these weak厦bonds. This surface24Octahedral Planeof talc can then inter-Sryact with water onlysilica (T) inner layerthroughdispersion4/*3T-O Edgeforces,makingithydrophobic(B)(A)(Laskowski &Kitch-ener, 1969).Accord-ilicate)Figure7. (a)The curved morphology of chrysotile (Klein & Hurlbut, 1993); (b)A simplified structure of chrysotilefibreing to Gaudin et al.(Yada, 1971).CIM Journal Vol. 3, No. 4

206J.S.Laskowskiand molybdenum atoms within the layers are strongly cova-PROPERTIESOFAQUEOUSSUSPENSIONSOFlently bonded, but the successive layers of sulphur atoms areISOTROPICANDANISOTROPICMINERALSheld together by weak van der Waals bonds.These bondsprovide excellent cleavage characteristics parallel to theIsotropic mineralsbase of the hexagonal crystals,producing a hydrophobicAs Figure 1 shows, the electrical repulsion forcessurface (sulphurdoes notform hydrogenbonds with water).between interacting solid particles in water will entirely dis-Asimilar situation exists in graphite.appear when these particles do not carry electrical charge.Practically, the charge is indirectly characterized by thezeta-potential measurements, which providethe iso-electricpoint (iep),thepH (strictly speaking,the concentration ofpotential-determining ions) at which the zeta potential ofthe mineral is equal to zero. Around this pH, the suspensionis very unstable; the particles aggregate (coagulation) andsettle quickly.Because aggregation results in the formationof a network between aggregating particles, the rheologicalmeasurements give high shear yield values for such a case.Comparison of the experimentally determined yieldstress versus pH curves with the zeta-potential-pH curves(since formost systems,potential-determining ions are H+oMoosand OH) can yield very important information on theocnature of the particle surface charge. Figure 9 (after John-Figure 8.Crystallochemical structure of graphite and molybdeniteson, Franks, Scales,Boger, & Healy[2000]) shows such acomparison. As this figure indicates, the maximum yieldstress occurs exactly at the pH of the iep for this mineral. At2000this point, the van der Waals attraction is not opposed bySolids (wt%)any electrical repulsion; the suspension is unstable and itcoagulates.Becauseofthestructurethatdevelopsbetween(Bd) 65.3athe coagulating particles, the rheological measurements-provide high yield stress values. As the pH moves away1from the iep, the yield stress values decrease both in highero61.4and lower pH ranges; in these pH ranges the particles arestabilized against aggregation by electrical double layers.57This behaviour is typical for isotropic minerals.42.3Anisotropicminerals08456791011(a)DHClays Because of the importance of clays, this group hasbeen studied extensively.Figure 10 shows the yield stressvalues plotted against pH for kaolinite suspensions at dif-Rank3ferent volumetric solids content (Johnson et al., 2000).2(os a sto) 0.001MThe rheological measurements in this case do not corre-KNO3late with the electrokinetic measurements at all. The iso-1electric point forkaolinitedetermined from the0zeta-potential measurements is approximately 3.5, whereasthepointof maximum coagulation ofkaolinite suspensions-1lies at approximatelypH5.5.The lack of correlation raises serious questions about the-2applicabilityof Smoluchowski'sequationtothecalculation ofzeta potentialfromthe measured electrophoretic mobilityforplate-like anisotropic particles.The case is depicted in Fig57891046ure3.Thebehaviour of plate-likeanisotropicparticles in anpH(b)electrical field is unknown and there is no mathematicalmodel thatallowscalculation ofthezeta-potentialvaluesfromFigure 9.Electrophoretic measurements showing that the maximumthe electrophoretic mobility of suchparticles.All such meas-coagulation occurs at the iso-electric point of zirconia suspensions(Johnson et al., 2000).urements must therefore be treated as estimates only.CIM Journal 1 Vol. 3, No. 4

Anisotropic minerals in flotation circuits207SOFThis discussion, based ona numberof piecesofexperi-charges of the two different sites of kaolinite platelets are theLSmental evidence (Rand & Melton, 1977; Tombacz & Szek-largest (Figure 12).ers,2006;Williams&Williams,1978),leadstotheconclusion that the zeta-potential values for kaolinite alu-Hydrophobic anisotropic minerals This group'orcesmina edges and the zeta potential for silicate faces wouldincludes talc, molybdenite, and graphite.The most charac-y dis-have been different if it had been possible to measure themteristic feature of the anisotropy of these minerals is inher-independently.Such an estimate is shown in Figure11.Fromiarge.ent hydrophobicity of the basal surfaces of these particles.ythethis plot, it can be inferred that maximum coagulation inFigure 13 shows the zeta-potential values of talc particlesectrickaolinite suspensions takes place at approximately pH 5.5calculated from the electrophoretic measurements carried outlonofthe pH level at which differences between the electricalas a function of pH (Fuerstenau & Huang, 2003).Theseial ofnsion) and450ationKaolin-ogical0.26360case.(ed):yield"urves0.24/270reH+Aan theJohn-0.22180iuchayield0.20ral.At90H0.18ed byand ittween-ments4n36891011?awayPHnigheres areFigure 10. The yield stress-pH curves for kaolinite suspensions atFigure 12.Coagulation of clay particles over a pH range of 4-6.ayers.different volumetric solids content (Johnson et al.,2000).2030iphas20BALMATTALCstress0.002MKNOat dif-10-AW).corre-Dedge(Au)eiso--10-20thenereas-20isionsMIEN-30-401faceut the-40ionofityforA-501Fig--60inan-6034567891011c2468101214naticalpHpHsfrommeas-Figure1l.The likely zeta-potential valuesforfaces and edges ofkaoliniteFigure 13. Zeta potential of talc as a function of pH (Fuerstenau&(Johnson et al., 2000).Huang,2003).CIM JournalIVol.3, No.4

208J.S.Laskowski0.42.010MKNO30.3(Ba)0.21.5PointofZeroIsoelectricPointChargeE0.1A1.0A0-0.1HSS0.5Maximm-0.2Aggregation102→10"MKCI-0.30.0239678101145923456191011Final pHpHFigure 14.Potentiometric titration curve for talc that identifies the point-Figure 15.Casson yield stress curve for talc aqueous suspensions asfunction of pH (Burdukova et al., 2007).of-zero chargefor talc to be a pH ofapproximately7.7 (Burdukova et al.,2007).BALMAT TALC CRYSTAL60pH4-4.5EDGE2x10MKNDPointofsaomaximumPoint of netcoagulation/zerochargeu400pH786200FACEBASALPLANEOEDGEn0.20.40.60.81.01.20.0CONCENTRATIONOFDODECYLAMMONIUMACETATE,MMFigure 16.Proposed charge distribution on the surface of talc particles(Burdukova et al., 2007).Figure 17.Effect of dodecyiammonium acetate, measured at a pH of 44.5,onrecedingcontactangleson thefaceandedgeofa talccrystal(Fuerstenau&Huang,2003).values agree very well with many other reported dataobtained with the use of the same experimental technique.They allgive the iso-electric point for talc at approximatelyThe differences between the surface properties of thepI12.5basal planes and edges of talcparticles, as directly meas-The results obtained using the titration technique, asured by Fuerstenau and Huang (2003), clearly show theshown in Figure 14, indicate that the point-of-zero chargeanisotropy of talc particles. The receding contact anglesfor talc occurs at approximately pH 7.7 (Burdukova et al.,measured on the cleaved face of a talc crystal exceeded 60o2007).This does not correlate at all with the position of theover a very broad pH range (4-12), whereas the contactiepfor talc.For comparison, this iep value is also given inangles measured on the edgesprepared by cutting the talcFigure 14.crystal were less than 200.This is even more convincing inFurther evidence for the anisotropic properties of talcFigure 17, which shows the effect of a cationic collectorparticles comes from the rheological measurements. Such(dodecylammonium acetate)on thewettabilityof talcfacesexperiments indicate that the maximum coagulation of talcandedges (atpH4-4.5).aqueous suspensions occurs at approximately pH 5.5 (Fig-Figure 17 reveals that dodecylammonium cations orienture 15).From these results, the model describing electricaldifferently at the face-solution and edge-solution inter-charge distribution on talc faces and edges was derivedfaces. While adsorption of this collector at the edges(Figure 16).makes them hydrophobic, the adsorption on the facesCIM Journal1Vol.3, No.4

209Anisotropic minerals inflotation circuitsresults from the hydrophobic chain interaction with thecleaved graphite and molybdenite surfaces by Arbiter, Fujii,hydrophobic surface making this surface hydrophilic atHansen, and Raja (1975) gave values of approximately 800highercollectorconcentrations.for graphite, and, depending on sample preparation,approx-The hydrophobic-to-hydrophilic ratio, that is, the face-imately70-90for molybdenite.Chander and Fuerstenauto-edge ratio, changes with particle size for anisotropic(1972)confirmed largedifferences in contactangles meas-minerals.As Figure 18 demonstrates, the adsorption of twoured on hydrophobicfaces and hydrophilic edges ofmolyb-polysaccharides (dextrin and guar gum) onto talc stronglydenite crystals. Lopez Valdivieso, Madrid Ortega, Reyesdepends on the size of talc particles that is on the face-to-Bahena, Sanchez Lopez, and Song (2006) reported thatedge ratio.For example, adsorption of dextrin onto awhilecontact angles measured onmolybdenitefaceswere in-150+75 μum size fraction of talc is more than 2mg/m2the range of 60o, the edges were completely hydrophilic.whereasforthe sizefraction of-63+53um it is onlyUsing atomic force microscopy (AFM),Lopez Valdivieso etapproximately1 mg/m?, and for the-37 μm fraction it is aal. (2006) detected microcrystals and micro-edges on thelittle higher than 0.5 mg/m? (Rath, Subramanian, &faces of molybdenite crystals. This demonstrates that11Laskowski, 1997). For both dextrin and guar gum themolybdenite faces are highly heterogeneous.Thesefindingsadsorption expressedpersquaremetreofthemineral sur-are in agreement with those of Komiyama et al. (2004), whoface clearly decreases with decreasing particle size.found terraces and rims of nanometric size on molybdeniteBecause the face-to-edge ratio decreases with decreasingfaces and explained the catalytic activity of molybdeniteparticle size, these results suggest that the adsorption takeswith the occurrence of such high-energy crater structures onplace mostly on thefaces of the talc particles. In a flotationbasal planes.process, this should lead to a depressed flotation oftalc.It is important to note here that the hydrophobic-to-hydrophilic ratio is different for particles used in contactangle (larger specimens),flotation (fine particles),andzeta-potential (very fine particles)measurements.The par-5.0ticles used in flotation and zeta-potential measurements10MKNO3have a higher ratio of hydrophilic-to-hydrophobic surfaces4.5than the specimens utilized in contact angle measurements;therefore, these measurements may provide different corre4.0-lations (Chanderetal.,1975).-150+75)1-HBotn/A(-63+53)umMolybdenite flotation Therate of flotation of very fine-37μm10particles decreases with the size of such particles because ofo（-150+75)Guargum-63+53)umthe low probability of particle-to-bubble collision. For100ppm37ummolybdenite, decreasing particle size also increases theRon1.2edge-to-face ratio, making theseparticles less hydrophobic.Therefore, the floatability of fine molybdenite decreasesCnetsharply with the particle size (Castro, Jara, Munoz, &Laskowski,2011;Castro &Laskowski,1997).It was alsouf4ashown that the floatability of these fine particles is very sen-1.0sitive to the presence of hydrophilic macromolecules (e.g.,polysaccharides and flocculants; Castro & Laskowski,of the0.52004).In the Cu-Mo industry,there are two situations wheremeas-the use of flocculants may affect Mo recovery: (i) increasedowthe-useofrecvcledwaterinwhichtheconcentrationofresidual142F81012anglespolymer builds up and (i) the need for thickening pulps ofpHed600Cu-Mo bulk concentratebefore pumping over a very longontactdistances (e.g., between a grinding plant and flotation plantFigure18.Effect of pH on the adsorption densities of dextrin and guarhetalcgum for three particle sizefractions of talc (Rath et al.,1997)or between the Cu concentrator and the Mo plant).cinginInherently hydrophobic minerals can be floated in con-llectorBecause molybdenite, graphite, and talc are naturallycentrated solutions of inorganic salts, even without othercfaceshydrophobic,it is also possibleto measurewettability of theirflotation agents (Klassen & Mokrousoy,1963).This issurfaces; there are quite a few papers on the flotation proper-called the salt flotation process.Klassen(1966)showed that:orientties ofthese minerals.The problem,however,is thatonly thewhile bituminous coals-which are very hydrophobic-inter-faces of these minerals are hydrophobic, and therefore onlyfloatverywell inconcentratedelectrolytes,thesaltflotationedgestheseparts ofthe surfaceof suchparticles areresponsibleforof lesshydrophobicsub-bituminouscoalsandanthracitesfacestheir natural floatability. Wettability measurements onwas not satisfactory.The relationship between coal surfaceCIM Journal [ Vol. 3, No. 4

210J.S.Laskowskihydrophobicity and its salt fotation response was con-graphite is rejected from the copperconcentrate bydepress-firmed by Fuerstenau, Rosenbaum, & Laskowski (1983).ing chalcopyrite and floating off graphite. Several optionsBecause simple inorganic ions cannot render solid surfaceswere considered: the use of sodium sulphide (which turnedhydrophobic, only highly hydrophobic minerals respondout not to be economical due to its high consumption),thewell to the salt flotation process. Flotation of molybdeniteuse of cyanides, and the use of thioglycolate (TGA) orin 0.5 M NaCl solutions was shown tobe very successfultrithiocarbonate.Thioglycolate (HSCH,COOH)was(Castro&Laskowski,2011;Castro&Laskowski,2012).patented in 1948 but it was not used until recently.TheseHowever,in theflotationof Cu-Mo sulphide ores,pyriteagents are much less toxic than cyanides. Initially,TGA wasmust be depressed and this is commonly achieved with theapplied to depress copper sulphides and float graphite;use of lime.Any change in the pH of such a well-equili-trithiocarbonate has been used for the last couple of years.brated system as sea water leads to hydrolysis of divalentcations.Because magnesium hydroxide is much less solu-Clays Clays appear in almost all types of ores. They formble than calcium hydroxide, increasing the pH with limevery fine suspensions in water that maybe very stable,leads totheformation of magnesium hydroxy-complexesresulting in tailings disposal problems. Because of theirand hydroxides. It was shown that molybdenite flotation inanisotropic character, clay particles,depending on pH,mayalkaline pH is extremely sensitive to the presence ofcarry both positive and negative electric charges, and thusMgOH*and Mg(OH) species in the pulp (Castro, Rioseco,can easily interact with all other minerals.This often leads&Laskowski,2012;Laskowski&Castro,2012).to a slime coating and inhibited flotation of valuable miner-als. Desliming is the most radical remedy and also requiresTale flotation In plants treating platinum-bearing ores inthe use ofdispersing agents.In thecase ofclays,polyphos-South Africa, polysaccharides (e.g., guar gum, carbo-phates (e.g:,hexametaphosphate) or water glass are com-xymethyl cellulose [CMCD) areused to depress talc.Formonlyutilized.some depressants,ionic strength of the pulp is very imporInpotash oreflotation plants inCanada,differential flota-tant.For instance, while in the case of carbo-xymethyl cel-tion of sylvite (KCI)from halite (NaCI) is carried out in sat-lulose(anionicpolymer)inKCl solutions,flotationtestsaturated brine with the use of long-chain primary amines.pH 9 showed that depression is negligible at 10-3 M, but isDesliming involves selective flocculation of slimes with thesatisfactory at 10-2 M; in the presence of Ca2+ and Mg2+use ofpolyacrylamideflocculants and flotation ofthefloccu-ions,depression is satisfactory even at 10-3M concentra-lated slimes with some secondarybranched amines (Cana-tions (Shortridge, Harris, & Bradshaw,1999).ElectrolytedianPatentNo.1,211,235,1982;Perucca&Cormode,concentration (ionic strength)is not that important when1999).Because such a desliming is not complete,so-calledguar gum is used in the flotation tests. As it was later con-blinders are still used (e.g., guar gum, CMC, starch). Byfirmed,macromolecules of carboxymethyl cellulose coil inadsorbing onto remaining slimes, the macromolecules of thesolution at higher ionic strength and this seems to increaseblinders prevent thecollector from adsorbing onto slimes.adsorption of CMC(Pawlik,Laskowski,&Ansari,2003).Successful flotation ofkaolinite after flocculating withWhereas guar gums with larger molecular weight turned outmodified polyacrylamide was reported by Yuehua, Wei,to be stronger talc depressants, this was not observed withHaipu, and Xu (2004).Ma, Bruckard, and Holmes (2009)CMC(Shortridge,Harris,Bradshaw,&Koopal,2000)showed that kaolinite can be floated with ether monoamincTalc is a significant constituent of the gangue in plat-and etherdiamineinNaCI solutionsinum-bearing sulphide ores.Talc depression increases theKaolin clayhasfoundmany important applications,forgrade of the sulphide concentrates.This picture was, how-instance, in themanufacturingofpottery,or as an inertfiller.ever,recently found tobemore complex (Robertson,Brad-This requires prior beneficiation and removal of impuritiesshaw, & Harris, 2003).Because of the platc-likc shape andsuch as anatase, rutile, and iron oxides.The size of such par-hydrophobicityoftalcfaces,thepresenceof talcparticlesticles is below 10 μm.The best-known technique is ultra-leads to dry,large-bubbled froths.Depression of talc andflotation, in which coarse calcite with the addition of fattybetter pulp dispersion leads to more unstable froths withacids (e.g.,tall oil) is used as a carrier to piggy-back finelower water recoveries and reduced entrainment.anatase in the conventional flotation process. Several carri-erless processes were also developed.In all of theseGraphite flotation Graphite also frequently appears as aprocesses,fatty acids areused along with calcium salts (acti-difficult-to-handle gangue in sulphide ores. Inco has threevator)under alkalineconditions (Willis,Mathur &Yoon,sulphide-producing regions, the most important of which are1999).Theproblem is incontrolling theconcentration ofthein the Sudbury area (Ontario)and in Thompson (Manitoba).activator,which may induce clay coagulation when in excess,The Thompson circuit also includes a graphite separationormay even cause flotation oftheclayparticlesratherthan thestage following the copper-nickel separation. The presencecoloured impurities.Perhaps the most significant improve-of graphite in the nickel concentrate helps to maintain thement was the introduction ofhydroxamate collectors patentedreducing potential of the electric furnace bath.However,in1986 (suchas the S6493reagent manufacturedbyCytec)CIM Journal I Vol. 3, No. 4

Anisotropicminerals inflotationcircuits211press-Slime coating Perhaps the most instructive example,This is thena clear example of electrostatic stabilization;theptionswhich illustrates industrial application of dispersants inspecific adsorption of phosphateanions increases thenegativeurnedflotationprocesses,wasprovidedbyJowettand his cowork-charge of the coal and clay particles, bringing about strong1), theers. Jowett, El-Sinbawy, and Smith (1956) demonstratedcoulombic repulsion,whichkeeps clay particles apart fromA) orthat hydrophobic coal particles become completelythe coal surface,thereby restoring natural coal floatabilitywashydrophilic in the presence of clays,which form a ‘slimeThe problem of slime coating in flotation can befurtherThesecoating' on coal surfaces (Figure 19).illustratedbytheexampletakenfromtheflotation of nickelAwassulphide ores in Western Australia; thephite;gangue in these ores contains serpen-ears.tine minerals (antigorite, chrysotile,+SODIUM+CLAYSlizardite).The effect of lizardite (fine-CLAYSORTHOPHOSPHATEformgrained silicate) and chrysotile (fibrousitable,silicate)on the flotation of pentlanditetheirwas studied byEdwards, Kipke,and, mayAgar(1980).InprocessingofCanadianithusultramafic nickel ores with serpentineleadsgangue,Dai et al.(2009)recommendedniner-COALdesliming and flotation at pH 10.1,COALCOALquiresadjusted using soda ash.phos-com-Figure 19.Effectofclays anddispersant on wettabilityofcoal surface inaqueous solution (JowettetFibrous gangue particles in flota-al., 1956).tion circuits As Figure 7 shows,flota-fibrous chrysotile tubes curl in a wayn sat-Coal hydrophobicity can,however,be restored by simplythat exposes themagnesium-rich octahedral layer on thenines.adding a dispersant. In their paper, Jowett et al. (1956) usedoutside; therefore,the surface charge ofthis site is likelytoth thesodium hydrogen phosphate (today we would rather usebe similar to that of brucite.To avoid problems associatedoccu-hexamethaphosphate).Both coal and clay particles werewith electrophoretic measurements of anisotropic miner-Cana-found to develop more negative zeta-potential values in50-als, the electrical charge of brucite will beused to characnode,100mg/L sodiumhydrogenphosphate solutions (Figure20).terize the electrical charge of the chrysotile surfacecalledFigure21 shows zeta potential of brucite and silica plotted). Byagainst pH.ofthe-20Figure 22 shows the zeta potential-pH curves forles.lizardite and pentlandite (Bremmel et al., 2005).As:withL6.9Wei,2009)NUMBERSINDICATEPHimine-30Ocoalparticlesau202s, forbruciteeodeafiller.+shel particlesirities01par-7.7-40ultra-ee'fatty207.6&7.9finecarri-Q8.1these40silica8.0B4-50(acti-8.1SYoon,8.4-60ofthe<cess,/1I24681012an the0100200300400PHrove-Concentration,mg/LentedFigure20.EffectofNa,HPO,onthezetapotentialofcoalandclayFigure 21.Zeta potential of brucite and silica as a function of pHytec).particles (Jowett et al., 1956).(adapted after Miller et al., 2007).CIM Journal Vol. 3, No. 4

J.S.Laskowski212Figures 21 and 22 indicate, the surface of chrysotile fibresnegative, it was possible to eliminatethedepressing effectand the surface of lizardite are both positively charged overof the slimes.Closed water circuits inmineral processingplants area very broad pH range. Over this pH range, pentlanditeparticles are charged negatively. It could then be expectedalmost universal today. In such circuits, electrolyte build upthat fine particles of the serpentine minerals will formin recycle streams over time results in a high electrolyteslime coatings on the surface of pentlandite.With the useconcentration of theprocess water.The discussedphenom-of a scanning electron microscope, it was possible toena and the effect of CMC on slime coating may look dif-demonstrate that such coatings indeed appeared on the sur-ferently at the high ionic strength of the process waterface of pentlandite conditioned in thepulpin the presence(Wellham, Elber, & Yan, 1992). Peng and Seaman (2011)of these minerals.Using small-scale flotation experimentshave recently shown that the effect of CMC stronglywith pure pentlandite,Edwards et al.(1980)demonstrateddepends on the ionization of thesemacromolecules; this isthat at chrysotile-to-pentlandite levels above approxi-expressed as degree of substitution (DS), with the chargemately 1:200 there was almost complete depression of thedensity of CMCmacromolecules increasing with the valuepentlandite flotation (these ores may contain more thanof DS.IntheirflotationteststheyusedCMC samples with80% serpentine minerals).The effect of lizardite was lessDS varying from 0.4 to 0.9 and found that the effect ofsevere. With the use of carboxymethyl cellulose,theCMC strongly depends on the degree of substitution.Whileadsorption of which makes the surface of silicate mineralsthe flotation selectivity in de-ionized water increased withincreasing DS, the opposite effect was observed when theflotation tests were carried out on the Mt. Keith process40water.In the process waterthat is,water at a higher ionicstrength-theless anionic CMC (with smallerDSvalues)aoaprovided more selectiveflotation results.20lizarditeThemajoreffect brought about bythepresenceoffibrous gangue particles in theflotationfeed results from theshape of such particles.Rheology of fine particle systems0depends on many factors; particle size, shape, and solidsconcentration influence the rheological behaviour to a greatextent.In general,the rheological behaviour becomesmorepentlandite20non-Newtonian as the particle size decreases. Also, asdemonstrated by Barnes, Hutton,and Walters (1989),therheology of the fine suspensions strongly depends on theshape of the particles (Figure 23). The effect of the solids46810concentration on the suspension rheology ismuchmorepro-PHnounced if the particles are not sphericalAs Figure 23 shows,at a particular volume solid concen-Figure 22.Zeta potential of lizardite and pentlandite particles in 10-3MKNO, as a function of pH (Bremmel et al.,2005).tration,thesuspensionjamsup',makingflowimpossible;Sespheresgrains+nickel-sulfde ore6chrysotileoplatesorodsaposaAuaieyea ianepa30a2010110203040o1Percentage Volume ConcentrationSolids content (v/v%)Figure 23.Dependence of the apparent viscosity (at a shear rate of 300 s)Figure 24. Rheological properties of the aqueous suspensions ofof differently shaped particles in water on concentration (Barnes et al.,chrysotile and the nickel sulphide ore that contains chrysotile (Kilicka-1989),plan etal.,2010).CIM Journal ↓ Vol. 3, No. 4