J. CHEM. SOC. FARADAY TRANS., 1994, 90(21), 3301-3307 Thermodynamics and Kinetics of the Reaction of Copper(l1) and lron(ll1) with Ultra-small Colloidal Chalcopyrite (CuFeS,) Ewen Silvester, Franz Grieser* and Thomas W. Healy School of Chemistry, University of Melbourne, Parkville, Victoria, 3052 Australia Dan Meisel" and James C. Sullivan Chemistry Division, Argonne National Laboratory, Argonne, lL 60439, USA The kinetics of oxidation of colloidal chalcopyrite (CuFeS,) have been examined in the presence of aqueous Fe"' and Cu" at pH 2.3 and in the presence of radiolytically generated Fe(OH),+ at neutral pH. The reaction of Cu" with CuFeS, has been ascribed to an ion-exchange reaction of copper for iron in the CuFeS, lattice. The initial copper sulfide phase that forms, as a result of the exhange, is rich in Cu' and thermally converts to the known covellite (CuS) phase.Oxidation of CuFeS, by Fe"' at low pH leads to the dissolution of the chalcopyrite phase and the formation of Fe2+(aq), Cu"(aq) and So(s). The formation of an unstable intermediate CuS, phase is suggested from thermodynamic calculations. Under neutral pH conditions, CuFeS, is oxidized by Fe(OH),+ adsorbed on the particle surface. Oxidation is restricted to the first monolayer of the particles due to the forma- tion of an Fe'" hydroxide layer at the particle surface. The oxidation of CuFeS, by Fe(OH)'+ is slower than by Fe3+(aq), presumably due to the lower redox potential of the former species. The oxidation of chalcopyrite in acidic Fe"' solutions has received a great deal of attention'-' owing to the applica- tions in commerical hydrometallurgy of this mineral.2 Under acidic conditions the reaction between CuFeS, and Fe"' gen- erally follows CuFeS2(s)+ 4Fe3+(aq)+Cu2+(aq)+ 5Fe2+(aq)+ 2S0(s) (1) although it is not clear whether this equation accurately describes the process relating to the initial reaction kinetics.X-Ray photoelectron spectroscopy (XPS) of air oxidized CuFeS, surfaces under acidic conditions6 has shown that, in the early stages of the reaction, iron migrates into solution while copper remains in what is essentially an oxidized chal- copyrite crystal lattice as Cu', with a composition approach- ing CuS,. Similar findings have been made with electrochemically oxidized CuFeS, electrode^.^ The use of an 'ultra-small '-particle-sized chalcopyrite sol allows a more detailed investigation of the initial reaction kinetics due to the higher proportion of material resident at the solid/aqueous solution interface.This paper deals with the spectroscopic, thermodynamic and kinetic aspects of the reaction of 'ultra-small' chalcopyrite with both Cu" and Fe"' at low pH. In addition the pH dependence of the reaction of Fen' with this mineral has been investigated using pulse radiolysis techniques to generate a transient aqueous Fe"' species [Fe(OH)'+] at neutral pH. Experimental The preparation and dialysis of chalcopyrite sols have been described previously.8 These sols were saturated with argon in gas syringes and stored until required; usually less than 5 days.In all experiments the CuFeS, sol concentration was determined from the sol absorbance at 480 nm (&480 = 7100 dm3 mol- cm-'). Reaction of Chalcopyrite with CU" and Fe"' at pH 2.3 Cu" and Fe"' solutions were prepared from perchlorate salts in lop2 mol dmp3 perchloric acid. Chalcopyrite sols pre- pared at (3.0 k0.3) x mol dm-3 were reacted with Cu" or Fe"' solutions in equal volumes such that, upon mixing, the chalcopyrite concentration was (1.5 & 0.2) x mol dm-3 and the pH ca. 2.3. All the solutions were saturated with argon gas prior to reaction and all experiments were conducted at a temperature of 25 "C. In a separate set of experiments the long-term reaction products were determined by analysis of the metal ion concentrations/speciation in solution. Separation of the aqueous phase from the particles was by filtration of the sol through an Amicon PM30 membrane (pore diameter ca.20 A)in an Amicon 8010 stirred cell assembly. Total copper and iron concentrations in the filtrate were determined by a stan- dard atomic absorption (AAS) technique. Fe" was determined spectrophotometrically using the o-phenanthroline method.' Stopped-flow Measurements Reaction kinetics were studied by the stopped-flow technique using a Hi-tech SF-51 stopped-flow unit. In this system, time- resolved absorbances were recorded at single wavelengths over a pathlength of 0.3 cm. Chalcopyrite sols were reacted with iron(1Ir) solutions in the range 5 x lo-' to mol dm-3 and the sol bleaching monitored at 480 nm.Under moderately acidic conditions Fen' speciation is dominated by Fe3+, Fe(OH)2+ and Fe2(OH)24+, with Fe3+ being the prin- ciple form (> 70%). Both Fe(OH)*+ and Fe,(OH),4+ absorb light at 480 nm;" however, the concentrations of these species were sufficiently low to allow this contribution to the measured absorbance to be ignored. The solubility of crystalline haematite (Fe20,) is exceeded at the higher concentrations of Fe"' reacted ; however, the amorphous Fe(OH), phase, with a precipitation edge at a higher pH, is more likely to form initially." The reduction of Fe"' by the sol results in the formation of Fe", of which the dominant species under acidic conditions is Fe(H,0),2+(aq).'o This species absorbs in the UV region below 300 nm and very weakly in the near IR (NIR)12 and so does not contribute to the measured absorbance at 480 nm.At low pH direct acid dissolution of the sol is possible. The contribution of this pathway to the overall dissolution is dis- cussed in the subsequent sections. Pulse Radiolysis of CuFeS,-Fe" Solutions at Neutral pH A series of sol diluents containing Fe" (prepared from FeS04.7H,0) were mixed with the CuFeS, sol in varying ratios to give a series of mixtures, all containing 5 x Table 1 Initial solution concentrations used in the pulse radiolysis experiments CuFeSJmol dm - Fe"/mol dm - PH 1.0 x 10-4 5 x 10-4 6.63 5.0 x 10-5 5 x 10-4 6.70 2.6 x 10-5 5 x 10-4 6.57 1.3 x 10-5 5 x 10-4 6.15 moI dm-3 Fe", but with varying sol concentrations.The CuFeS, sol and Fe" concentrations studied and the corre- sponding pH of the mixed solutions are given in Table 1. All the sols and diluents were saturated with N,O prior to mixing and the mixing was carried out shortly before irradia- tion. Irradiation of Solutions The irradiation of aqueous solutions with ionizing radiation has been described previously. l3 The Argonne National Laboratory EINAC facility and spectrophotometric detection system used have been described el~ewhere.'~*'~ Solutions and sols were irradiated in quartz cells with optical paths between 0.5 and 2.0 em. Radiation pulses were multiples of 40 ns pulses (up to a maximum of 10 pulses) with a repetition rate of 60 Hz.In this way radiation doses between 2 and 17 krad were applied to the sols. Dosimetry was achieved by irradiation of mol dm-3 KSCN solutions saturated with N2O.I3 In N,O saturated Fe" solutions the following series of reactions occur after primary irradiation H,O-+e-(aq), OH', H,O,, H, ,H30+ [predominantly e(aq)- and OH'] (2) e-(as) + N,O + H,O --+ N, + OH' + OH-(3) Fez+ + OH'-+ Fe(OH),+ (4) yielding an Fe'" species which can then react with the sol. In the absence of the chalcopyrite sol the absorbance spectrum immediately after irradiation (100 ms) was found to corre- spond closely to a previously measured spectrum for Fe(OH)'+.'* This first hydrolysis product of iron" was observed to be kinetically stable in the second before further hydrolysis and the eventual formation of Feu' hydroxide.After ca. 10 s, the absorption spectrum resembled that of iron'" hydroxide particles in the <10 nm size range.16 Changes in the sol absorbance were monitored at 480 nm. At longer times and at higher radiation doses, the formation of Fe"' hydroxide contributed significantly to the measured absorption at this wavelength. Results and Discussion Reaction of Chalcopyrite with Cu" at Low pH Fig. 1 shows the UV-VIS absorption spectrum of a 1.5 x mol ~im-~ CuFeS, sol (a) before and (b) after reaction with mol dmm3 Cu'. Reaction of the chalcopy- rite sol with aqueous Cu" results in significant bleaching of the sol absorption at 480 nm indicating loss of this mineral phase.Accompanying the observed bleaching is a pro-nounced broadening of the absorption band. This was attrib- uted to an increased scattering component in the sol extinction. When heated to >70°C for ca. 1 min the reacted sol developed a strong absorption band in the NIR region [Fig. l(c)] which, for the reasons outlined below, can be assigned to the presence of crystalline covellite" (nominally CuS, although more accurately described as Cu,S * CuS,"). J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 1.o 0.8 0.6 0 m -e8 % 0.4 0.2 I I I I I I I 10.0 300 400 500 600 700 800 900 wavelength/nm Fig. 1 UV-VIS absorption spectrum of a 1.5 x rnol dm-j CuFeS, sol (a) before and (b) after reaction with Cu" (pH 2.3) at an initial concentration of rnol dm-'.Heating the reaction product yielded a sol with spectrum (c), corresponding to that of crystalline covellite (CuS). Copper sulfide sols when first precipitated from solution are 'golden-brown' in colour. Electron diffraction studies indicate that the particles are a poorly crystalline form of covellite while XPS analysis shows that copper is exclusively in the form of Cu'.'' With time, and at an enhanced rate at higher temperature, these sols change to a green colour and concomitantly become more crystalline. The NIR band that forms is attributed to a transformation of a third of the copper in the particle to Cu", consistent with the known structure of crystalline covellite.The intensity of the NIR absorption band gives a good quantitative measure of the crystalline covellite concentration. In the case shown in Fig. l(c) it is readily calculated that greater than 90% of the copper in the system [both from Cu"(aq) and CuFeS,] has been converted into covellite, indicating a nearly quantitative replacement of chalcopyrite by covellite under these condi- tions. The absence of the NIR band before heating suggests that the initial copper sulfide product is exclusively in the Cu' phase. Fig. 2 shows the aqueous concentrations of iron and copper as well as the concentration of adsorbed copper after reaction of CuFeS, sols with Cu" over a range of initial con- centrations.Virtually all (>95%) the aqueous iron was found to be in the divalent form. Feu is significantly leached from the particle even in the absence of Cu', indicating that acid dissolution is a major reaction pathway at low pH. The absence of any measurable aqueous copper under these con- ditions suggests that acid dissolution takes the form CuFeS,(s) + 2H+(aq)+CuS(s) + Fe2'(aq) + H,S(aq) (5) J. CHEM.SOC. FARADAY TRANS., 1994, VOL. 90 IL E0 n h Y + Na k n "0 2 I I 1 2[C~~+(aq)]/lO-~rnol dm-3 Fig. 2 Concentrations of aqueous Fe" (a),aqueous Cu" (W) and adsorbed Cu" (O),10 min after reaction of CuFeS, with Cu" at initial concentrations of up to 2 x rnol dm-3. Initial [CuFeS,] = 1.5 x rnol dm-3; pH 2.3.Insert: Aqueous Cu" concentration as a function of initial [Cu2+(aq)], on a less sensitive scale (x axis, Cu2+(aq)/10-3 rnol dm-j; y axis, CuAaq)/lO-' mol dm-3). For systems in which Cu" was added, the more than quan- titative replacement of copper for iron indicates that very little sulfide departs the system in the form of H2S(g), suggest- ing a direct ion-exchange reaction, i.e. CuFeS,(s) + Cu2+(aq)+2CuS(s) + Fe2+(aq) (6) Thermodynamic calculations show that, in the presence of Cu" significant transformation of CuFeS,(s) to CuS(s) should occur, the extent of which depends on the H2S(aq) concentra- tion in the solution. While this concentration is unknown, it is useful to consider an extreme case. It has been noted pre- viously that CuFeS, sols contain a slight excess of adsorbed S2-(ca.3%)." This excess sulfide would correspond to an H,S(aq) concentration of cu. mol dm-3, and, in turn, to solution concentrations of Cu2 '(as) and Fe2+(aq) in equi- librium with CuFeS,(s) and CuS(s) of the order of 10-l4 and mol dm-3, respectively. This concentration of Fe2'(aq) is very close to the value observed experimentally. At higher concentrations of added Cu" the level of H2S(aq) would be considerably lower, due in part to direct precipitation of CuS(s). Accordingly, the equilibrium level of Fe2 '(aq) should be higher. It is surprising that, even at the highest concentra- tions of Cu", the release of Fe" is limited to mol dm-j. Given that this concentration corresponds to two thirds of the total iron in the particle it seems likely that the observed level is a result of an inhibited reaction with the particle core rather than thermodynamic constraints.The adsorption of Cu" above the level at which Fe" is liberated into solution indicates that a further mechanism of copper retention must be operating; this is presumably the adsorption of Cu" onto the modified particle surface. An attempt was made to study the kinetics of the reaction of Cu" with chalcopyrite by the stopped-flow technique; however, the bulk of the reaction occurred within the mixing time of the apparatus (ca. 20 ms). Acid dissolution was observed to be a much slower process, confirming that the reaction of Cu" with the particle is not preceded by acid dis- solution.While the reaction of Cu" with CuFeS, does not outwardly appear to be a redox process, electron transfer must be involved. Copper is added as Cu" although no Cu" is evident in the UV-VIS absorption spectrum of the reacted colloid, i.e. the NIR band of 'green' copper sulfide is initially absent. Similarly, while iron is believed to be present in chalcopyrite as Fe"', it enters the solution as Fe". These redox changes can be reconciled in terms of the reaction Cu'Fe"'S:'-(s) + Cu"(aq) -+ Cu$"-S'(s) + Fe2+(aq) (7) where Cu\S"S' represents a Cu' sulfide mineral of overall stoichiometry CuS. Reaction of Chalcopyrite with Fe"' at Low pH The reaction of CuFeS, sol with Fe"' resulted in bleaching of the sol absorbance at 480 nm and the formation of Fe" in solution, indicating that oxidation of the sol had occurred. In Fig.3 is shown the concentration of copper that appears in solution with increasing initial Fe"' concentration. For initial Fe"' concentrations of < rnol drn-j, copper is not detected in solution, even though oxidation of the sol occurs. Above this concentration, copper is detected in the aqueous phase, increasing to a level close to that expected for the complete oxidation of the chalcopyrite phase. Since neither Cu' nor Cu" hydrolyse at this pH, the copper retention that is observed can only be due to association with sulfide. The observed behaviour can be accounted for by the sequential reactions, CuFeS,(s) + 2Fe3+(aq) -,CuS,,(s)+ 3Fe2+(aq)+ (2 -y)S'(s) (8) CuS,(s) + 2Fe3+(aq)+Cu2&(aq)+ 2Fe2+(aq)+ ySo(s) (9) where the parameter (2 -y) represents the degree to which the partially oxidized lattice has nucleated to form S'(s).This interpretation is consistent with previous studies on chal-copyrite electrodes2' and chalcopyrite slurries' which have found preferential release of iron in the initial stages of reac- tion. XPS studies indicate that the copper remains as a copper(r) sulfide phase most accurately described as an iron deficient chalcopyrite. 6,7 If no adsorption of Fe" occurs the number of electrons transferred per CuFeSz unit N,, can be obtained from the equation, At low initial concentrations of Fe"' the sol bleaching is a reliable measure of the concentration of CuFeS, sol reacted, with little contribution from light scattering to the sol absorption spectrum.At higher initial Concentrations of Fe"' (>5 x mol drn-j) light scattering is evident, indicative 41 t? Ne 3-2-I 1 ' . '.--C 0I ' ' """I ' ' """I -5 -4 -3 -2 log aFeo+ Fig. 3 Concentration of Cu" (0)produced in solution as a function of initial Fe"' concentration, and the number of electrons transferred per CuFeS, unit (0)calculated as described in the text. Initial [CuFeS,] = 1.5 x mol d~n-~;pH 2.3. 3304 of either elemental sulfur formation or sol aggregation. For these systems the Cu" concentration has been used as an approximate measure of chalcopyrite oxidation.In Fig. 3 is shown the reaction stoichiometry, calculated as described above, over the range of initial Fern' concentrations studied. At high excess of oxidant the stoichiometry has the expected upper limit of four electrons per CuFeS, ,consistent with that found by Linge' using electrochemical techniques and consis- tent with reaction (1). The data do not exhibit the lower limit of two electrons per CuFeS, expected from reaction (8) which is presumably due to the concurrent acid dissolution reac- tion, the effect of which is greater at lower initial Fe"' concen- trations and the stoichiometry of which is zero is terms of eqn. (I). In the thermodynamic interpretation of this system, it is instructive to construct a predominance area diagram in terms of the solution species involved in the reaction dynamics. Such a diagram is shown in Fig.4,where the sta- bility regions have been calculated using the log absorption coefficients, log aFe3+ and log aFez+,as variables. This figure has been constructed for constant Cu2'(aq) activities of and lo-, mol drn-,, total dissolved sulfur concentra- tion of mol dm-, and constant pH (2.3). The assump- tion of constant H,S(aq) activity is inaccurate but it allows a qualitative description of the system. As an initial approx- imation, the Gibbs energy of formation (AfGe) of the CuS, phase has been taken to be the same as that for covellite. The oxidation of CuFeS, can be considered in terms of its reaction path, as is shown in Fig. 4,starting at aFe3+= and uFe2+= lo-'' mol dm-,.Oxidation of CuFeS, under these conditions results in the formation of Fe2+(aq), both from CuFeS, and as a reduction product of Fe" and Cu2'(aq). The increasing concentration of Cu'+(aq) leads to a broadening of the CuS(s) stability field, as shown by the successive CuS(s)/Cu2 +(aq) boundary lines on this figure. An equilibrium point is not marked on this diagram; under the conditions considered it would be the intersection point between the three predominance regions: CuFeS, ; CuS, Fe2+, H,S; and CuS, Fe2+, So, the position of which is con- trolled by the activity of H,S(aq) which will change during the reaction. A copper sulfide product with a Af G" close to that of covellite should not form to a significant extent since the bulk of the reaction occurs outside the stability field of log aFeZ+ -1 5 -1 0 -5 0 0 -5 0)Y (D m --1 0 // '/ / , -1 5 Fig.4 Phase diagram for the Cu-Fe-S system in terms of the master variables log uFe3+and log uFez+showing the reaction path for the oxidation of CuFeS, starting at initial conditions of aFe3+= rnol dm-, and uFez+= lo-" mol dm-'. Diagram prepared for conditions of constant activity of Cu2+ lod4 or lo-' mol drn-,), constant pH (2.3) and constant total dissolved sulfur of mol dm-3. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 CuS(s). The retention of copper that is observed in these systems must therefore be due to either the formation of a CuS,(s) phase of significantly lower Af G * than that of covel- lite, or to kinetic restrictions on the release of copper from the oxidized chalcopyrite lattice.In Fig. 4 is shown where the CuS,(s)-Cu2'(aq) boundary would need to be in order for a copper sulfide to be the dominant oxidation product. The position of this line corresponds to a Gibbs energy of formation of the order of -130 kJ mol-', compared to that for crys- talline covellite of -48.9 kJ mol- '. This value would seem to be unreasonably low and kinetic restrictions on the release of copper from the oxidized chalcopyrite lattice must be favoured as the reason for the observed retention of copper. Kinetics of Chalcopyrite Oxidation of Fe"' at Low pH Bleaching of the sol absorbance occurred on the millisecond to second timescale.From the preceding discussion it can be noted that a considerable excess of Fe"' is necessary for the reaction to be described by reaction (1). Even with a high excess of Fe"' it appears unlikely that this reaction would adequately describe the sol bleaching on short timescales. Acid Dissolution of Chalcopyrite The effects of perchloric, nitric and hydrochloric acids upon the absorbance of chalcopyrite at 480 nm were investigated to determine the extent of direct acid dissolution in these systems. In Table 2 are shown the initial sol absorbances at 480 nm [A(O)],the apparent first-order rate constant fit to the bleaching of the sol absorbance at this wavelength, and the sol absorbance at longer times (20 s) after mixing [A(oo)] for sols reacted with these acids.The measured A(0) values compare well with the expected value (A = 0.31; [CuFeS,] = 1.5 x mol drn-,) indicating that very little bleaching of the sol occurs within the mixing time of the stopped-flow system. After mixing, some bleaching of the sol absorbance is observed, although as will be shown in the fol- lowing section, both the extent of bleaching and the fitted rate constants are small compared to that observed in the presence of Fe"'. Also shown in Table 2 is the effect of 5 x lo-, mol dm-, Al"' (in 5 x lo-, mol dm-, HNO,) upon the sol absorbance. Sol bleaching under these conditions is less than that observed for an equilvalent concentration of nitric acid alone. It appears that aluminium has the effect of protecting the sol from dissolution, presumably via specific binding to the sulfide surface thus preventing interaction of protons with the surface ligands.Chalcopyrite Oxidation by Fe"' at pH 2.3 Fig. 5(a) shows the measured initial sol absorbances for the range of Fe"' concentrations studied. At all concentrations partial oxidation of the chalcopyrite sol occurred within the mixing time of the stopped-flow apparatus (20 ms). At low initial Fe"' concentrations, the measured A(0)values decrease Table 2 Initial absorbance [A(O)],apparent first-order rate constant (k),and final absorbance [A(oo)],for the acid dissolution of chalcopy-rite" HCI 0.306 0.32 0.277 HC10, 0.294 0.978 0.277 HNO, 0.303 0.7 12 0.288 HNO,b + Al(NO,), 0.311 0.07 0.294 ~~ " Concentration of all acids was 0.005 mol drn-,.Optical path length was 0.3 an. 5 x lo-, mol dm-, Al(NO,), in 0.005 mol dm-, HNO,. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 -0.28 0.26 0 0.24 1 0 0 I.,.,.,.,,0.22 01 23456 5 4 -3 k.I s2 1 B e 0 0' ' ' ' ' I " ' ' 1 0 1 2 3 4 5 6 0.10' ' ' . ' ' ' . ' . J' I 01 23456 [Fe"']/10-3 mol dm-3 Fig. 5 (a) Initial absorbances at 480 nm for a CuFeS, sol oxidized by Fe"' at pH 2.3, using the stopped-flow technique. The dashed line shows the expected A(0)value. (b) Fitted first-order rate constants for CuFeS, sol bleaching at 480 nm. The decays were recorded over a 2 s duration.(c) A(m) (sol absorbance after 20 s) values for kinetic measurements made over a 2 s duration. with increasing Fe"' concentration suggesting that the rate of this initial process is dependent on oxidant concentration. At higher Fe"' concentrations the measured A(0)values increase, presumably due to the formation of light absorbing or scat- tering products. Previous studies have shown that CQ. 35% of the CuFeS, in colloidal chalcopyrite resides in the first mono- layer." At all but the lowest concentrations studied, ca. 30% of the sol absorbance is lost during the mixing time. It appears therefore that very little of the oxidation of the first monolayer of the CuFeS, particles is captured by the stopped-flow technique. Bleaching of the sol absorption was observed over several orders of magnitude in time without any distinct separation between different reaction steps. Single-exponential fits were performed on the measured sol bleaching over a several periods of time (2,20 and 200 s) in order to obtain an indica- tion of the rates involved.At all Fe'" concentrations the fitted rate constant value decreased with increasing measurement time, indicating that the reaction is not a single-exponential process. Fig. 5(b)shows the fitted rate constants for measure- ments made over a duration of 2 s. The initial increase in the fitted rate constant with increasing Fe"' concentration further confirms the dependence of the rate of oxidation of the first monolayer on oxidant concentration.In terms of the rate- determining step, this dependence could be interpreted as either a collisional encounter in solution or adsorption of Fe"' onto the particle surface. At the higher Fe"' concentra-tions, the CuFeS, sol bleaching rate is independent of Fe"' concentration. In this domain, which corresponds to particle oxidation beyond the first monolayer, the rate-determining step is possibly either the release of ions into solution from the particle surface or electron transfer from a species in the solid particle to an adsorbed Fe"' species. Fig. 5(c) shows A(co) values corresponding to the data shown in Fig. 5(a) and (b).Again, the presence of an absorb- ing or light scattering product is evident at higher Fe"' con-centrations. Similar observations were made on all the timescales studied.Pulse Radiolysis of Fe"-CuFeS, Solutions at Neutral pH To overcome the mixing time restrictions of the stopped-flow technique, pulse radiolysis experiments were undertaken with the oxidant created in situ by the radiation pulse. Initial con- centrations of Fe(0H)' +,generated according to reactions (2)-(4), were varied by altering the radiation dose applied to CuFeS, sol-Fe" mixtures. In this way initial concentrations of Fe(OH)'+ could be generated in the range 1.2 x to 1.0 x lop4mol dm-3. Unlike the low pH systems, CuFeS, sol bleaching was adequately described by first-order kinetics at all doses and it is these rate constants which are con- sidered in the following discussion. Fig.5 shows the first-order rate constant (@-I) against dose [or initial Fe(OH), concentration], for several sol con- + centrations. Only at the lowest sol concentration does the measured first-order rate constant vary linearly with the oxidant concentration, as might be expected from the relative concentrations of sol particles and the oxidant present. Using the results from the lowest sol concentration, the bimolecular rate constant can be calculated to be 8.5 x lo4 dm3 mol-' s-' in terms of CuFeS, molecules, or 1 x lo8 dm3 mo1-l s-' in terms of CuFeS, particles. It is clear that the reaction observed at near-neutral pH is substantially slower than the reaction at low pH (see previous section). This is probably due to the lower redox potential of the Fe(OH),+/Fe2+ couple compared to that for Fe3+/FeZ+.Interestingly, the fitted rate constant value decreases with increasing sol con- centration. It seems that the rate of oxidation is determined by the number of oxidant species per particle and not by the total concentration of surface sites. Hence, the oxidation occurs by adsorbed Fe"' species and not solution species. Fig. 7 shows the total sol bleaching [derived from A(w) values] against initial Fe(0H)' concentration for several sol+ concentrations. Also shown is the expected bleaching for the 61 1 0 10 000 20 000 dose/rad I 0 4 8 12 [Fe(0H) *+I ,,,,/10 -mol dm -' Fig. 6 First-order rate constants for CuFeS, sol bleaching at 480 nm as a function of dose, or initial Fe(OH)'+ concentration (lower scale), for sol concentrations of (0)5 x lo5and (m) mol dm-3 1.3 x lo-', (0)2.6 x (0) J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 0.00 T,, 1 dose/rad I I 0 4 8 12 [Fe(OH)2+]iniJ10-5 mol dm-3 Fig. 7 Total CuFeS, sol bleaching at 480 nm [as derived from mea- sured A(m) values] us. applied dose, or initial Fe(OH)'+ concentra- tion (lower scale), for sol concentrations of (0)1.3 x lo-', (0) 2.6 x (0) lo-' and (I)mol dm-3. The expected sol5 x bleaching for the first monolayer of reaction for the three lowest sol concentrations is shown by the dashed lines. The monolayer absorb- ance of the highest sol concentration is off scale, as indicated. first monolayer of the particles (dashed lines).For all sol con- centrations, sol bleaching is limited to the first monolayer and is apparently diminished at higher doses. As was noted in the experimental section, the generation of Fe"' at pH 6.5 will result in the formation of solid iron(@ hydroxide after several seconds, the absorption spectrum of which will overlap with that of the sol at 480 nm. At higher doses the contribution of iron(@ oxide to the measured absorbance is clearly significant. It is likely that the limitation of the reac- tion to the first monolayer is also a result of the formation of iron(@ hydroxide given that it would probably form as a surface coating on the CuFeS, particles. It can be readily shown that layers of greater than 10 A in thickness could form on the particles at the higher doses employed.For elec- tron transfer to occur across such a layer, charge carriers must be injected into one of the bands of the iron(u1) hydrox- ide film. The rate of reaction then becomes a function of both the relative positions of the conduction and valence bands of the two solids and the conductivity of the film.,' Fig. 8 shows schematic diagram of the relevant energy levels for the CuFeS,/Fe(OH),/water interface at pH 7. It is clear that the concentration of valence band holes in iron(rI1) hydroxide, as controlled by the Fe(OH),+/Fe2+ couple, would be low under these conditions. Given that the mechanism of electron transfer in this system is hole transfer from adsorbed Fe(OH),+ to iron(@ hydroxide and then to CuFeS, ,the rate of electron transfer would be significantly diminished by the presence of the iron(m) hydroxide layer.In addition to the restrictions on electron transfer, there would also be restrictions on the accompanying charge-balancing steps. A thin product layer on the particles would be sufficient to provide a physical barrier to ion movement into solution and limit reaction to the first monolayer of the CuFeS, particles. Conclusions In the low-pH reactions of both Cu" and Fe"' with CuFeS, particles, Fe" is preferentially released from the colloid during the initial stages of the reaction. While the reaction of Cu" with CuFeS, appears to be an ion-exchange process, of I 0.6 I -Fe(OH)2'/Fe2' v) W I 0.8 I __t r5 1.0 ' Lu 1.2 CuFeS2 particle core aqueous 1.4 1.6 1.8 2.22.0 1 2.4 1 Fig.8 Redox levels for a chalcopyrite/iron(III) hydroxide interface, in contact with an aqueous solution at pH 7 (us. SHE). Chalcopyrite band positions are estimates from ref. 19, while iron(iI1) hydroxide bands are estimates based on the positions of these bands for haema- tite (Fe,O,). Also shown are the positions of the H+/H, and Fe(0H)' +/Fez-+ redox couples in solution and the relevant corrosion potentials (Edccomp)for CuFeS, . copper for iron in the CuFeS, structure, an intra-lattice elec- tron-transfer step must be involved. The initial copper sulfide phase which forms is rich in Cu' and is thermally converted into the known covellite phase.For the reaction of Fe"' with chalcopyrite the release of Cu" appears to be kinetically rather than thermodynamically controlled. The low-pH reac- tion of Fe"' [dominantly Fe3 '(as)] with colloidal chalcopy- rite is much faster than that at near-neutral pH [Fe(OH)2+(aq)], presumably due to the lower redox poten- tial of the latter. Further, the reaction at neutral pH is restricted to the first monolayer of the chalcopyrite particles due the formation of an iron(rr1) hydroxide layer at the parti- cle surface. The dynamics of the oxidation reaction by Fe(OH), + are determined by intra-particle transfer from the lattice to Fe(OH)2 + adsorbed at the particle surface. This work was supported in part by the Australian Research Council (ARC) in the form of a Special Research Centre grant to the Advanced Mineral Products Centre.Work at the Argonne National Laboratory is performed under the auspices of the ofice of Basic Energy Sciences, Division of Chemical Sciences, US-DOE under contract no. W-31-109- ENG-38. E.S. acknowledges the receipt of a Commonwealth Post-Graduate Resarch Award. References 1 H.G. Linge, Hydrometallurgy, 1976,2,51. 2 J. E. Dutrizac, Metall. Trans. B, 1978,9B, 431. 3 P. B. Munoz, J. D. Miller and M. E. Wadsworth, Metall. Trans. B,1979, lOB, 149. 4 D. L. Jones and E. Peters, Extractiue Metallurgy of Copper ZI, The Metallurgical Society of AIME, New York, 1976. 5 J. E. Dutizac, Metall. Trans. B, 1981, 12B, 371. 6 A. N. Buckley and R. Woods, Aust.J. Chem., 1984,37,2403. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 3307 7 A. N. Buckley, I. C. Hamilton and R. Woods, in Flotation of 15 S. Gordon, K. H. Schmidt and E. J. Hart, J. Phys. Chem., 1977, Suljide Minerals, Developments in Mineral Processing, ed. K. S. 81,104. E. Forssberg, Elsevier, Amsterdam, 1985, vol. 6. 16 P. Mulvaney, R. Cooper, F. Grieser and D. Meisel, Langmuir, 8 E. J. Silvester, T. W. Healy, F. Grieser and B. Sexton, Langmuir, 1988,4, 1206. 1991, 7, 19. 17 E. J. Silvester, F. Grieser, B. Sexton and T. W. Healy, Langmuir, 9 A. I. Vogel, Quantitative Inorganic Analysis, Longmans, London, 1991,7,2917. 1953. 18 D. J. Vaughan and J. R. Craig, Mineral Chemistry of Metal Sul-10 C. F. Baes and R. E. Mesmer, The Hydrolysis of Cations, Wiley, fides, Cambridge University Press, Cambridge, 1978, p. 180. New York, 1976. 19 E. J. Silvester, F. Grieser, D. Meisel, T. W. Healy and J. C. Sulli-11 R. M. Garrels and C. L. Christ, Solutions, Minerals and Equi- van, J. Phys. Chem., 1992,%, 4386. libria, Freeman, Cooper and Company, San Francisco, 1965. 20 T. Biegler and M. D. Home, J. Electrochem. Soc., 1985, 132, 12 M. Ehrenfreund and J-L. Leibenguth, Bull. SOC. Chim. Fr., 1970, 1363. 7,2494. 21 S. R. Morrison, Electrochemistry at Semiconductor and Oxidized 13 I. G. Draganic and Z. D. Draganic, The Radiation Chemistry of Metal Electrodes, Plenum Press, New York, 1980. Water, Academic Press, New York, 1971. 14 D. Meisel, W. A. Mulac and M. S. Matheson, J. Phys. Chem., 1981,85,179. Paper 4/04335I; Received 15th July, 1994