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