- Email: [email protected]
The interaction of xanthates and amines with pyroxene activated by copper and nickel
Minerals Engineering 19 (2006) 799–806 This article is also available online at: www.elsevier.com/locate/mineng
The interaction of xanthates and amines with pyroxene activated by copper and nickel q C.T. OÕConnor a
a,*
, V. Malysiak b, N.J. Shackleton
b
Mineral Processing Research Unit, Department of Chemical Engineering, University of Cape Town, Private Bag, Rondebosch, Cape Town 7701, South Africa b Anglo Platinum Research Centre, P.O. Box 6540, Homestead 1412, South Africa Received 25 July 2005; accepted 9 September 2005 Available online 2 November 2005
Abstract In the flotation of platinum group minerals pyroxene is a major gangue mineral. This study shows using zeta potential and ToF-SIMS analyses that when pyroxene is exposed to copper, nickel and/or calcium ions in solution at alkaline pHs a surface reaction occurs between the hydroxy species, in the case of copper and nickel, or divalent calcium ions and the pyroxene surface which influences the floatability of the pyroxene. Following exposure to copper and nickel ions at pH 9 addition of SIBX resulted in the pyroxene floating readily. It is also shown, in the case of copper, that the activating effect of the adsorbed copper hydroxy species can be reduced by the addition of complexing amines prior to SIBX. Proposals are made to explain these observations. 2005 Elsevier Ltd. All rights reserved. Keywords: Flotation activators; Precious metal ores
1. Introduction Two gangue minerals which are commonly found associated with platinum group minerals are pyroxene and feldspar. Pyroxene is the most abundant mineral in the Merensky Reef, constituting about 60% of the reef by volume. In its most common form it has the general formula of [(Mg, Fe, Ca) Si2O6] consisting of single channels of linked tetrahedra. In a previous study by Phillips et al. (1993) it was found that pyroxene reports to the concentrate due to true flotation. Since it is assumed that sulphide collectors generally do not exhibit any affinity for siliceous minerals it is thus often inferred that the pyroxene surface can be activated by dissolved metal ions, such as copper and nickel ions, or their hydroxy forms. Being uncharged, pyroxene may be expected to behave in a manner similar to quartz. There are numerous q
This is a slightly modified version of a paper that was presented at the centenary of Flotation Conference in Brisbane. * Corresponding author. Tel.: +27 21 6502701; fax: +27 21 6503782. E-mail address: [email protected] (C.T. OÕConnor). 0892-6875/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.mineng.2005.09.040
reports of the activation of quartz by ions such as copper (James and Healy, 1972). Nagaraj and Brinen (1995) have studied copper ion adsorption on pyrite and pyroxene and reported that copper adsorbed onto both minerals but they did not study the ultimate effect on flotation. In the case of nickel, Mackenzie and OÕBrien (1969) suggested that the adsorption of NiOH+ ions might involve hydrogen bonding between the OH groups of the metal complex and oxygen atoms of the quartz surface. The present paper presents results of an investigation into the adsorption of copper and nickel ions onto pyroxene. The effect of adsorption of these ions on the flotation behaviour of pyroxene was also investigated as well as methods to remove the adsorbed metal species from the gangue surface. The effect of calcium ions present in the process water on the flotation of pyroxene was also investigated. 2. Experimental The experimental procedures used were similar to those described previously (Malysiak et al., 2002).
800
C.T. OÕConnor et al. / Minerals Engineering 19 (2006) 799–806
2.1. Minerals Natural pyroxene from the Merensky Reef in the Northern Province, South Africa, was crushed to 2 mm and selected by handpicking. Chemical analysis indicated that the pyroxene contained 13.2% Mg, 11.0% Fe, 3.7% Ca, 2.7% Al, and 25.1% Si. Copper activation was attained by dosing with copper sulphate. Nickel activation was attained by exposing samples to pentlandite which was synthesised at Anglo Platinum Research Centre (Malysiak et al., 2002). Calcium activation was attained by exposure to synthetic process water using different concentrations of calcium. All mineral samples were stored under argon in a freezer and freshly ground in an agate mortar just prior to each experiment. The products were screened to obtain size fractions of 25 lm for zeta potential determinations and +38—106 lm for microflotation tests. 2.2. Reagents Purified sodium isobutyl xanthate (SIBX) was provided by SENMIN. Other chemicals were of analytical grade quality. Sodium carbonate (0.1 M) and hydrochloric acid (0.1 M) were used for pH adjustment. As already mentioned copper sulphate was used as activator. The amines used were ethylenediamine (EDA), diethylenetriamine (DETA), tetraethylenediamine (TETA) and ethylenediamine-tetraacetic acid (EDTA).
2.4. Time of flight secondary ion mass spectrometry (ToF-SIMS) Surface analysis of minerals was carried out using a PHI TRIFT II ToF-SIMS instrument operating in the static SIMS regime. A pulsed primary ion beam bombards the sample surface, causing the emission of atomic and molecular secondary ions. A small percentage of the secondary ions are charged and can therefore be extracted by an electric field into a mass spectrometer. The mass spectra are recorded by measuring the time difference between pulsing the primary ion gun and the arrival of secondary ions on a fast dual microchannelplate detector at the spectrometer, by means of a multi-stop time-to-digital converter. When mixtures were used they were conditioned in synthetic water in the presence of the desired reagents for 20 min. The samples were filtered and washed with water (conductivity 0.7 lmS cm1), adjusted to the desired pH, to remove any physically attached ions. All samples were dried in an argon atmosphere at ambient temperature. The 15 kV, 600 pA gallium beam was used throughout the investigation. About 30 grains for each mineral were imaged and analysed for Ca, Mg, Al, Si, Fe, Ni and Cu during positive ion analysis and O, OH, S and xanthate during negative ion analysis. The intensities obtained are normalised for the elements of interest and presented as a relative percent surface coverage. 2.5. Microflotation tests
2.2.1. Synthetic water composition Water, with a specific conductance of 0.7 lS cm1 and a surface tension of 72.8 mN m1 at 20 C, produced by a MILLI-RO PLUS apparatus, was used during the study. This water was modified by the addition of various chemical salts to produce synthetic process water (I = 3.5 · 102). The synthetic water contained amounts of key ions similar to those typically found in circuit water, viz. Ca2+ 80 ppm, Mg2+ 80 ppm, Na+ 135 ppm, Cl 2 270 ppm, SO2 250 ppm, NO 40 ppm, 4 3 135 ppm, CO3 and a TDS value of 1030. 2.3. Zeta potential determinations Zeta potential determinations were carried out on dilute dispersions of the individual minerals studied using a Malvern Zetasizer 4. During the experiments the interaction between sodium isobutyl xanthate (SIBX: 5 · 105 M), copper sulphate (CuSO4: 5 · 105 M) and various amines (5 · 105 M) and the pyroxene surface was investigated. The instrument gives the electrophoretic mobility from which the zeta potential was calculated using the Smoluchowski equation, since ja 1, ja being the ratio of particle radius to double layer thickness (Hunter, 1993). The zeta potential experiments were carried out in synthetic process water at various pH values at 25 C. Conditioning of the mineral was carried out for 20 min. The Eh was allowed to vary naturally but was generally in the region of 100 mV.
A microflotation cell (volume 250 cm3) was used to determine the flotation response/hydrophobicity of pyroxene and pentlandite–pyroxene mixtures. The cell consists of a conical tapered cylindrical tube with air introduced through a needle at the base of the cell (Wesseldijk et al., 1999). Mineral loaded bubbles rise through the cell and are deflected off the cone at the top of the cell, after which they burst, resulting in the minerals dropping into the concentrate launder. Each test was conducted at pH 9 using a 2 g sample of pure pyroxene. During the tests, various combinations of reagents were investigated. These were generally SIBX, [CuSO4 + SIBX] and [CuSO4 + AMINE + SIBX]. The sequence of reagent addition was as indicated by the sequence of reference in brackets. Conditioning periods for the reagents tested were 2 min for SIBX and 5 min for CuSO4 and the amines. 3. Results and discussion 3.1. Surface analyses 3.1.1. ToF-SIMS analyses ToF-SIMS analysis is a well-established technique for determining the occurrence of atomic/molecular species on the surface of mineral samples. It must be noted that ToF-SIMS results present relative and not absolute values. Table 1 shows the relative % abundances of copper on the
C.T. OÕConnor et al. / Minerals Engineering 19 (2006) 799–806 Table 1 Relative % abundance of copper on the surface of pyroxene subsequent to treatment by copper sulphate with or without further addition of diethylenetriamine (DETA) Condition of ore
pH
Relative % abundance of Cu
Untreated +CuSO4 +CuSO4 +CuSO4 + DETA +CuSO4 + DETA
4 and 9 4 9 4 9
0 0.2 2.0 0.1 0.6
Table 2 Dominant species present in solution at pH 6 and 9 Element of interest
pH 6 2+
pH 9 +
Ni , NiOH Cu2+ Ca2+
Ni Cu Ca
surface of pyroxene subsequent to treatment by copper sulphate and then treatment with diethylenetriamine (DETA). In the first instance these results show that copper hardly adsorbs on the surface at pH 4. This is consistent with results presented by other researchers. James and Healy (1972) have shown that Co2+ did not adsorb onto a quartz surface at pH 2, the pzc value. They proposed that although the non-hydrolyzed cobalt cation is strongly attracted to the negative surface indicated by the zeta potential measurements it is so strongly solvated that it cannot approach the surface. Reduction of the charge through hydrolysis to either the MOH+ or M(OH)2 form reduces the attraction to the solid but also reduces the solvation energy which falls off more rapidly with hydrolysis and thus bonding with the solid surface becomes feasible. Similar results were obtained for a copper, zinc, and lead–quartz system. As will be shown later (Fig. 4) zeta potential studies showed that at pH 4 the surface of pyroxene was also negatively charged and so a similar explanation could apply in this case as well. Table 1 shows that
801
Ni(OH)2, NiOH+ Cu(OH)+, Cu(OH)2 Ca2+
at pH = 9 copper is significantly adsorbed. Speciation diagrams (Malysiak, 2003) show that at pH 9 the dominant form of the copper is the monovalent hydroxy species Cu(OH)+ or the hydroxide Cu(OH)2 (Table 2) and the hydroxy species appears to be able to bind with the pyroxene surface through a hydrogen bonding mechanism. Table 1 also shows the effect of the amine, diethylenetriamine (DETA), on the copper concentration on the surface. As can be seen the amine causes significant decreases in the surface concentration of copper. Fig. 1, which shows the surface coverage of pyroxene after exposure to copper sulphate, SIBX and various amines at pH 9, illustrates clearly that, at that pH, copper adsorbs onto the pyroxene surface but that its surface concentration is reduced when exposed to all the amines tested except for EDTA, prior to SIBX addition. It is not possible to draw any conclusions on the differences between the EDA, DETA and TETA since these are too small to be significant. Kelebek et al. (1996) have shown that DETA complexes readily with nickel and copper hydroxide species. However the inability of EDTA to remove copper from the surface is interesting. EDTA complexes readily with ions in solution whereas EDA is less active. It is thus possible that, given the concurrent presence of calcium and
Pyroxene (95% Confidence Interval)
10 8 6 4 2
Cu+EDTA+SIBX
Cu+TETA+SIBX
Cu+EDA+SIBX
Cu+DETA+SIBX
Cu+SIBX
SIBX
-2
No Reagents
0
Pyrox only
Cu Surface Coverage (Relative %)
12
±1.96*Std. Err. ±1.00*Std. Err. Mean
Fig. 1. Relative copper surface coverage of pyroxene exposed to copper sulphate, SIBX and various amines at pH 9 using ToF-SIMS.
802
C.T. OÕConnor et al. / Minerals Engineering 19 (2006) 799–806 Pyroxene 95% Confidence Interval
Pyroxene 95% Confidence Interval
10
1.1
(b) Nickel Surface Coverage (Relative %)
1.0
0.9
0.8
0.7
0.6 Nicke l Surfa ce Co
Nickel Surface Coverage (Relative %)
(a)
0.5
0.4
8 7 6 5 4 3 2
0.3 pH 6
9
pH 9
±1.96*Std. Err. ±1.00*Std. Err. Mean
1
±1.96*Std. Err. ±1.00*Std. Err. Mean
Fig. 2. Relative nickel surface coverage at pH 6 and pH 9 using ToF-SIMS of (a) pyroxene only; and (b) pyroxene exposed to nickel ions.
magnesium ions in the water, the EDTA, from a stoichiometric aspect, is totally consumed by complexing with the solution ions and that there is little left, at the dosage used in the present study, to complex with the adsorbed copper hydroxy species. The EDA on the other hand will hardly complex with the solution species and thus be readily available to complex with and remove back into solution the adsorbed copper species. The authors are presently quantifying such interactions with a view to testing this hypothesis. As a matter of interest, in a separate set of experiments (Malysiak et al., 2004) it was found that the amines could not remove copper from the surface of pentlandite. Clearly the copper hydroxy species bind very strongly to the pentlandite and that this cannot be overcome by the amine– copper hydroxy interaction. Fig. 2 shows the results of a ToF-SIMS study of the results obtained for pure pyroxene and after it had been exposed to nickel ions by contacting it with a solution into which nickel has leached from the pentlandite present in the pulp. The relative abundance of nickel on the pyroxene surface was found to be almost four times greater at pH 9 than pH 6. The use of pH 6 as opposed to the value of 4 used in the case of copper was because of the frequent use of pH 6 conditions on flotation plants in which pyroxene and pentlandite occur. In these experiments, but not shown in this figure, the pyroxene was then exposed to SIBX and amines as had been done in the case of copper (Fig. 1). Similar results were observed, viz. that the amines EDA, DETA and TETA caused the surface concentration of the nickel to decrease significantly but this did not happen in the case of EDTA. Fig. 3 shows the surface concentration of copper as a function of the sequence of addition of copper sulphate, SIBX and EDA. It is clear that it is necessary to add the
EDA before exposing the copper activated pyroxene to the collector SIBX in order to reduce the surface concentration of copper. The bonding of the SIBX to the Cu(OH)+ or Cu(OH)2 is presumably sufficiently strong to ensure that the amine is not able to compete successfully to react with the copper species present on the surface. ToF-SIMS images of pyroxene exposed to pentlandite, copper sulphate and SIBX showed clearly the presence of copper and nickel on the surface of the pyroxene and also the presence of xanthate species on the surface on which the copper and nickel respectively appear. These images are not presented here since they need to be in colour to be properly appreciated. Images obtained after treatment by EDA showed that this amine removed much of the copper from the pyroxene surface. Images of pyroxene exposed to synthetic water containing calcium at concentrations of 80 and 500 ppm respectively showed that at pH 9 and at the higher concentration the relative abundance of calcium on the surface was about twice that of concentration shown at pH 6. Calcium exists as the divalent ion at both pHs and thus the concentration effect is presumed to be the main reason for this observation. 3.1.2. Zeta potential determinations Zeta potential studies were carried out on pyroxene in the presence of SIBX, CuSO4, [CuSO4 + SIBX], and [CuSO4 + SIBX + various amines] respectively at pH values of 6, 8 and 10. Fig. 4 shows these results where the amine used was DETA. Similar measurements, not presented here, were made in the case of EDA and TETA. A number of observations are made in these experiments which are consistent with the ToF-SIMS results. In the first instance little copper is seen to adsorb at pH 6 which has been explained above as being a result of the strongly
C.T. OÕConnor et al. / Minerals Engineering 19 (2006) 799–806
803
Pyroxene (95% Confidence Interval) 14
Cu Surface Coverage (Relative %)
12 10 8 6 4 2 0 ±1.96*Std. Err. ±1.00*Std. Err. Mean
-2 Pn only SIBX No Reagents
Cu+EDA+SIBX Cu+SIBX+EDA Cu+SIBX EDA+Cu+SIBX
Fig. 3. Relative copper surface coverage of pyroxene exposed to copper sulphate, SIBX and EDA in different sequences at pH 9 using ToF-SIMS.
5
Zeta Potential [mV]
0 -5 -10 -15 -20 -25 5
6
7
8
9
10
11
pH No Reagents
CuSO4
SIBX
CuSO4+SIBX
CuSO4+DETA
CuSO4+DETA+SIBX
Fig. 4. Zeta potential of pyroxene exposed to various reagents. Order of addition follows sequence in legend.
solvated Cu2+ ion. Addition of SIBX hardly affects the zeta potential values. Addition of copper sulphate markedly increased the zeta potential at pHs of 8 and higher. It has been mentioned above that at alkaline pHs the copper exists as either the positively charged Cu(OH)+ species or as Cu(OH)2. This increase in zeta potential may indicate that it is the former species that is dominant. Consistent with the ToF-SIMS results the zeta potential becomes more negative after the addition of the amine due to the removal of the positively charged copper species. SIBX addition thereafter has little effect on the zeta potential. When SIBX is added after copper sulphate, the zeta potential value becomes more negative due to the charge balanc-
ing role of the SIBX in its reaction with the positively charged copper hydroxy species. Fig. 5 shows the results of an experiment in which pyroxene was exposed to calcium ions illustrating that this resulted in a significant increase in the positive charge of the surface, especially at pH of 10, indicating that a surface reaction is occurring between the pyroxene and the calcium ions. 3.2. Microflotation tests The above results clearly indicate that the surface of pyroxene is modified when exposed to copper, nickel and calcium in solution by virtue of the fact that these ions
804
C.T. OÕConnor et al. / Minerals Engineering 19 (2006) 799–806 Zeta Potential 20 15 10 5 0 -5 -10 -15 -20 -25
5
6
7
8
9
10
11
pH
80ppm of Ca ions
500ppm of Ca ions
Fig. 5. Zeta potential of pyroxene exposed to calcium ions at two different concentrations.
adsorb onto the surface of the mineral. This may inadvertently modify the flotation behaviour of these minerals. Consequently a set of experiments was carried out to ascertain the effect of these ions on the flotation of pyroxene. A microflotation technique was used which essentially measures the relative hydrophobicity of the mineral. Fig. 6 shows these results at pH 4 and 9 in the presence of SIBX alone and SIBX and copper sulphate. At pH 9 when pyroxene is exposed to a solution containing copper sulphate followed by the addition of sodium isobutyl xanthate (SIBX) it becomes strongly hydrophobic and the recovery is greater than 70%. At pH 4 there is an increase in recovery from 45% to 59%. At pH 9 this increased from 45% to 80%. All these microflotation results are consistent with the observations made in both the ToF-SIMS and zeta potential experiments. In the former case as was shown in Table 1 the copper surface coverage (the copper existing at that pH as a copper hydroxy species) was greater at pH 9 than at pH 4. Since copper is still in its divalent form at pH 6 it is assumed that the coverage would still be low. Similarly the zeta potential results at pH 9 in the presence of copper
100 90 80 70 60 50 40 12 30 . 20 10 0 0
sulphate was about zero whereas at pH 6 it was <20 mV. Fig. 7 shows the recoveries obtained at pH 9 when the pyroxene was exposed to EDA and EDTA respectively after copper activation. As expected the EDA resulted in a decrease in recovery due to its ability to decrease the surface concentration of the activating copper species. The EDTA, as expected from the surface analysis results, did not have any noticeable effect of the recovery of the pyroxene. There are two issues of particular fundamental interest arising out of these results. Firstly the mechanism whereby the copper interacts with the pyroxene surface and the subsequent reaction of the SIBX with the copper activated pyroxene is of special interest since it is that interaction which renders the pyroxene hydrophobic. Secondly the mechanism whereby the amines are able to reduce the copper-activation effect is also of particular interest. Nagaraj and Brinen (1996) in a study of a copper– pyroxene system and using XPS data, found that copper was present as Cu+ rather than Cu2+ after collector treatment of copper activated pyroxene. This suggests that a chemical reaction is occurring during the adsorption of
Recovery (%)
0.01 0.89
5
10
15
20
25
Time (min) pH 4, No Reagents
pH 9, No Reagents
pH 4, SIBX
pH 9, SIBX
pH 4, CuSO4 + SIBX
pH 9, CuSO4 + SIBX
Fig. 6. Recovery of pyroxene following various reagent treatments at pH 4 and 9.
C.T. OÕConnor et al. / Minerals Engineering 19 (2006) 799–806
805
Recovery (%)
100 80 60 40 20 0 3
10
20
Time (min) No reagents
CuSO4+SIBX
CuSO4+EDA(4.00E-05M)+SIBX
CuSO4+EDTA(4.93E-05M)+SIBX
Fig. 7. Recovery of pyroxene following addition of EDA and EDTA respectively before the addition of SIBX after copper activation (pH 9).
xanthate onto the copper activated pyroxene, which involves the formation of either a CuX or a Cu(OH)X species. The following mechanism for the adsorption of copper at pH 9 is proposed. Cu(OH)2 + 2X $ Cu(OH)X [+OH + X ] $ CuX2 + 2OH In this mechanism two different pathways to the formation of a surface xanthate complex are proposed. One possibility is that adsorbed copper hydroxide reacts with a xanthate ion to form the Cu(OH)X species which renders the pyroxene hydrophobic. Alternatively the Cu(OH)X species could react further to form copper xanthate, viz.: CuX2 + 2OH ! CuX + 1/2X2 + 2OH This mechanism allows for the presence of monovalent copper but requires the formation of dixanthogen which is unlikely to be forming at the Eh values at which these experiments were carried out, viz. 100 mV. In both of the above cases it is important to note that these proposals assume that the rate of surface adsorption is less than rate of copper complex formation. Both the microflotation and zeta potential results showed that on its own SIBX had little effect on the behaviour of pyroxene at either pH. In either case the presence of SIBX showed little difference relative to the measurement for the pure mineral. In preliminary experiments it was observed that a lower recovery was obtained in the presence of calcium ions indicating that the presence of these ions in some way inhibits the possibility of inadvertent activation of pyroxene by the pentlandite dissolution products, viz. Ni2+. It is not clear why the calcium should have this effect. Calcium exists as the divalent ion at pH 9 whereas nickel exists as the hydroxy species (cf. Table 2). Since the heat of formation of a Ca–O bond is 36 kJ/mol and of Ni–O is 13 kJ/mol it may be that in some way the effect is driven by thermodynamic considerations. The role of the amines appears to be related to their effect on the equilibrium concentration of the various species in solution at the pH of interest. A series of calcula-
tions were carried out using the following molar concentrations of ions in solution: Cu2+ = 105; Ca2+ = 103; Ni2+ = 105,6,7; Fe3+ = 105 and [amines] = 106,5,4. The calculations were carried out at pH 9 using the programme JESS (Joint Expert Speciation Simulation) which calculates the equilibrium concentration of all the species in solution taking into account precipitation. These calculations, which represent preliminary findings in view of the fact that the authors are presently carrying out much more extensive analysis of the solution chemistry of this and other similar systems, revealed that, relative to EDA, DETA and TETA, EDTA complexes under the above conditions much more significantly with the other ions in solution than with copper even though EDTA would generally be expected to complex readily with copper. Thus those amines which were able to reduce the effect of copper activation of pyroxene, viz. EDA, DETA and TETA, may achieve this as a result of their not complexing as readily as EDTA with the ions such as calcium, iron and nickel. Thus the solution is essentially starved of free EDTA ions relative to EDA, DETA and TETA. These latter amines then complex with the copper ions in solution and in doing so affect the equilibrium between the copper hydroxide species which has precipitated out on the surface of the pyroxene and the solution ions of Cu2+ and 2OH. The end result is that the copper hydroxide species is resolubilized and thus the inadvertent activation of the pyroxene is reduced. Investigations are presently being carried out to test this hypothesis on similar systems. 4. Conclusion This study has shown clearly using zeta potential and TOF-SIMS analyses that when pyroxene is exposed to copper, nickel and/or calcium ions in solution at alkaline pHs a surface reaction occurs between the hydroxy species (in the case of copper and nickel) and the divalent calcium ions and the pyroxene surface which alters the surface of the pyroxene and thus influences its floatability. At pH 4 no adsorption of SIBX occurs possibly due to the
806
C.T. OÕConnor et al. / Minerals Engineering 19 (2006) 799–806
inhibiting effect of the solvation of the Cu2+. Following exposure to copper and nickel ions at pH 9, at which pH these species exist as hydroxy species, it was shown that addition of SIBX resulted in the pyroxene floating readily due to an interaction between the xanthate and the adsorbed metal hydroxy species. It is also shown that the activating effect of the adsorbed copper hydroxy species can be reduced by the use of complexing amines such as EDA, DETA and TETA which appear to be able to influence the equilibrium concentrations of the activating ions in a way which results in the equilibrium between precipitated copper hydroxide and the solution ions being disturbed in favour of dissolution of the precipitated species. Acknowledgements The authors wish to thank Anglo Platinum for permission to publish this paper. Anglo Platinum and the University of Cape Town are thanked for their financial support. Professor G.E. Jackson is thanked for carrying out the JESS calculations. References Hunter, R.J., 1993. Introduction to Modern Colloid Science. Oxford University Press.
James, R.O., Healy, T.W., 1972. Adsorption of hydrolyzable metal ions at the oxide water interface. Journal of Colloid and Interface Science 40, 53–64. Kelebek, S., Wells, P.F., Fekete, S.O., 1996. Differential flotation of chalcopyrite, pentlandite and pyrrhotite in Ni–Cu sulphide ores. Canadian Metallurgical Quarterly 35 (4), 329–336. Mackenzie, J.M.W., OÕBrien, N.T., 1969. Zeta potential of quartz in the presence of nickel (II) and cobalt (II). Society of Mining Engineers AIME 244, 168–172. Malysiak, V., 2003. Pentlandite–pyroxene and pentlandite–feldspar interactions and their effect on separation by flotation, Ph.D. thesis (unpublished), University of Cape Town. Malysiak, V., OÕConnor, C.T., Ralston, J., Gerson, A., Coetzer, L.P., Bradshaw, D.J., 2002. Pentlandite–feldspar interaction and its effect on separation by flotation. International Journal of Mineral Processing 66 (1–4), 89–106. Malysiak, V., Shackleton, N.J., OÕConnor, C.T., 2004. An investigation into the floatability of a pentlandite–pyroxene system. International Journal of Mineral Processing 74 (1–4), 251–262. Nagaraj, D.R., Brinen, J., 1995. SIMS study of metal ion activation in gangue flotation. In: Proceedings XIX International Mineral Processing Congress, SME, pp. 253–257. Nagaraj, D.R., Brinen, J., 1996. SIMS and XPS study of the adsorption of sulphide collectors on pyroxene: a case for inadvertent metal activation. Colloids and Surfaces A: Physicochemical and Engineering Aspects 116, 241–249. Phillips, D.M.A., Martin, C.J., Malysiak, V., 1993. Current technology in gangue flotation and concentrate. AngloPlatinum Report CJM12/ DML. Wesseldijk, Q.I., Reuter, M.A., Bradshaw, D.J., Harris, P.J., 1999. The flotation behaviour of chromite with respect to the beneficiation of UG2 ore. Minerals Engineering 12 (10), 1177–1184.
The interaction of xanthates and amines with pyroxene activated by copper and nickel q C.T. OÕConnor a
a,*
, V. Malysiak b, N.J. Shackleton
b
Mineral Processing Research Unit, Department of Chemical Engineering, University of Cape Town, Private Bag, Rondebosch, Cape Town 7701, South Africa b Anglo Platinum Research Centre, P.O. Box 6540, Homestead 1412, South Africa Received 25 July 2005; accepted 9 September 2005 Available online 2 November 2005
Abstract In the flotation of platinum group minerals pyroxene is a major gangue mineral. This study shows using zeta potential and ToF-SIMS analyses that when pyroxene is exposed to copper, nickel and/or calcium ions in solution at alkaline pHs a surface reaction occurs between the hydroxy species, in the case of copper and nickel, or divalent calcium ions and the pyroxene surface which influences the floatability of the pyroxene. Following exposure to copper and nickel ions at pH 9 addition of SIBX resulted in the pyroxene floating readily. It is also shown, in the case of copper, that the activating effect of the adsorbed copper hydroxy species can be reduced by the addition of complexing amines prior to SIBX. Proposals are made to explain these observations. 2005 Elsevier Ltd. All rights reserved. Keywords: Flotation activators; Precious metal ores
1. Introduction Two gangue minerals which are commonly found associated with platinum group minerals are pyroxene and feldspar. Pyroxene is the most abundant mineral in the Merensky Reef, constituting about 60% of the reef by volume. In its most common form it has the general formula of [(Mg, Fe, Ca) Si2O6] consisting of single channels of linked tetrahedra. In a previous study by Phillips et al. (1993) it was found that pyroxene reports to the concentrate due to true flotation. Since it is assumed that sulphide collectors generally do not exhibit any affinity for siliceous minerals it is thus often inferred that the pyroxene surface can be activated by dissolved metal ions, such as copper and nickel ions, or their hydroxy forms. Being uncharged, pyroxene may be expected to behave in a manner similar to quartz. There are numerous q
This is a slightly modified version of a paper that was presented at the centenary of Flotation Conference in Brisbane. * Corresponding author. Tel.: +27 21 6502701; fax: +27 21 6503782. E-mail address: [email protected] (C.T. OÕConnor). 0892-6875/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.mineng.2005.09.040
reports of the activation of quartz by ions such as copper (James and Healy, 1972). Nagaraj and Brinen (1995) have studied copper ion adsorption on pyrite and pyroxene and reported that copper adsorbed onto both minerals but they did not study the ultimate effect on flotation. In the case of nickel, Mackenzie and OÕBrien (1969) suggested that the adsorption of NiOH+ ions might involve hydrogen bonding between the OH groups of the metal complex and oxygen atoms of the quartz surface. The present paper presents results of an investigation into the adsorption of copper and nickel ions onto pyroxene. The effect of adsorption of these ions on the flotation behaviour of pyroxene was also investigated as well as methods to remove the adsorbed metal species from the gangue surface. The effect of calcium ions present in the process water on the flotation of pyroxene was also investigated. 2. Experimental The experimental procedures used were similar to those described previously (Malysiak et al., 2002).
800
C.T. OÕConnor et al. / Minerals Engineering 19 (2006) 799–806
2.1. Minerals Natural pyroxene from the Merensky Reef in the Northern Province, South Africa, was crushed to 2 mm and selected by handpicking. Chemical analysis indicated that the pyroxene contained 13.2% Mg, 11.0% Fe, 3.7% Ca, 2.7% Al, and 25.1% Si. Copper activation was attained by dosing with copper sulphate. Nickel activation was attained by exposing samples to pentlandite which was synthesised at Anglo Platinum Research Centre (Malysiak et al., 2002). Calcium activation was attained by exposure to synthetic process water using different concentrations of calcium. All mineral samples were stored under argon in a freezer and freshly ground in an agate mortar just prior to each experiment. The products were screened to obtain size fractions of 25 lm for zeta potential determinations and +38—106 lm for microflotation tests. 2.2. Reagents Purified sodium isobutyl xanthate (SIBX) was provided by SENMIN. Other chemicals were of analytical grade quality. Sodium carbonate (0.1 M) and hydrochloric acid (0.1 M) were used for pH adjustment. As already mentioned copper sulphate was used as activator. The amines used were ethylenediamine (EDA), diethylenetriamine (DETA), tetraethylenediamine (TETA) and ethylenediamine-tetraacetic acid (EDTA).
2.4. Time of flight secondary ion mass spectrometry (ToF-SIMS) Surface analysis of minerals was carried out using a PHI TRIFT II ToF-SIMS instrument operating in the static SIMS regime. A pulsed primary ion beam bombards the sample surface, causing the emission of atomic and molecular secondary ions. A small percentage of the secondary ions are charged and can therefore be extracted by an electric field into a mass spectrometer. The mass spectra are recorded by measuring the time difference between pulsing the primary ion gun and the arrival of secondary ions on a fast dual microchannelplate detector at the spectrometer, by means of a multi-stop time-to-digital converter. When mixtures were used they were conditioned in synthetic water in the presence of the desired reagents for 20 min. The samples were filtered and washed with water (conductivity 0.7 lmS cm1), adjusted to the desired pH, to remove any physically attached ions. All samples were dried in an argon atmosphere at ambient temperature. The 15 kV, 600 pA gallium beam was used throughout the investigation. About 30 grains for each mineral were imaged and analysed for Ca, Mg, Al, Si, Fe, Ni and Cu during positive ion analysis and O, OH, S and xanthate during negative ion analysis. The intensities obtained are normalised for the elements of interest and presented as a relative percent surface coverage. 2.5. Microflotation tests
2.2.1. Synthetic water composition Water, with a specific conductance of 0.7 lS cm1 and a surface tension of 72.8 mN m1 at 20 C, produced by a MILLI-RO PLUS apparatus, was used during the study. This water was modified by the addition of various chemical salts to produce synthetic process water (I = 3.5 · 102). The synthetic water contained amounts of key ions similar to those typically found in circuit water, viz. Ca2+ 80 ppm, Mg2+ 80 ppm, Na+ 135 ppm, Cl 2 270 ppm, SO2 250 ppm, NO 40 ppm, 4 3 135 ppm, CO3 and a TDS value of 1030. 2.3. Zeta potential determinations Zeta potential determinations were carried out on dilute dispersions of the individual minerals studied using a Malvern Zetasizer 4. During the experiments the interaction between sodium isobutyl xanthate (SIBX: 5 · 105 M), copper sulphate (CuSO4: 5 · 105 M) and various amines (5 · 105 M) and the pyroxene surface was investigated. The instrument gives the electrophoretic mobility from which the zeta potential was calculated using the Smoluchowski equation, since ja 1, ja being the ratio of particle radius to double layer thickness (Hunter, 1993). The zeta potential experiments were carried out in synthetic process water at various pH values at 25 C. Conditioning of the mineral was carried out for 20 min. The Eh was allowed to vary naturally but was generally in the region of 100 mV.
A microflotation cell (volume 250 cm3) was used to determine the flotation response/hydrophobicity of pyroxene and pentlandite–pyroxene mixtures. The cell consists of a conical tapered cylindrical tube with air introduced through a needle at the base of the cell (Wesseldijk et al., 1999). Mineral loaded bubbles rise through the cell and are deflected off the cone at the top of the cell, after which they burst, resulting in the minerals dropping into the concentrate launder. Each test was conducted at pH 9 using a 2 g sample of pure pyroxene. During the tests, various combinations of reagents were investigated. These were generally SIBX, [CuSO4 + SIBX] and [CuSO4 + AMINE + SIBX]. The sequence of reagent addition was as indicated by the sequence of reference in brackets. Conditioning periods for the reagents tested were 2 min for SIBX and 5 min for CuSO4 and the amines. 3. Results and discussion 3.1. Surface analyses 3.1.1. ToF-SIMS analyses ToF-SIMS analysis is a well-established technique for determining the occurrence of atomic/molecular species on the surface of mineral samples. It must be noted that ToF-SIMS results present relative and not absolute values. Table 1 shows the relative % abundances of copper on the
C.T. OÕConnor et al. / Minerals Engineering 19 (2006) 799–806 Table 1 Relative % abundance of copper on the surface of pyroxene subsequent to treatment by copper sulphate with or without further addition of diethylenetriamine (DETA) Condition of ore
pH
Relative % abundance of Cu
Untreated +CuSO4 +CuSO4 +CuSO4 + DETA +CuSO4 + DETA
4 and 9 4 9 4 9
0 0.2 2.0 0.1 0.6
Table 2 Dominant species present in solution at pH 6 and 9 Element of interest
pH 6 2+
pH 9 +
Ni , NiOH Cu2+ Ca2+
Ni Cu Ca
surface of pyroxene subsequent to treatment by copper sulphate and then treatment with diethylenetriamine (DETA). In the first instance these results show that copper hardly adsorbs on the surface at pH 4. This is consistent with results presented by other researchers. James and Healy (1972) have shown that Co2+ did not adsorb onto a quartz surface at pH 2, the pzc value. They proposed that although the non-hydrolyzed cobalt cation is strongly attracted to the negative surface indicated by the zeta potential measurements it is so strongly solvated that it cannot approach the surface. Reduction of the charge through hydrolysis to either the MOH+ or M(OH)2 form reduces the attraction to the solid but also reduces the solvation energy which falls off more rapidly with hydrolysis and thus bonding with the solid surface becomes feasible. Similar results were obtained for a copper, zinc, and lead–quartz system. As will be shown later (Fig. 4) zeta potential studies showed that at pH 4 the surface of pyroxene was also negatively charged and so a similar explanation could apply in this case as well. Table 1 shows that
801
Ni(OH)2, NiOH+ Cu(OH)+, Cu(OH)2 Ca2+
at pH = 9 copper is significantly adsorbed. Speciation diagrams (Malysiak, 2003) show that at pH 9 the dominant form of the copper is the monovalent hydroxy species Cu(OH)+ or the hydroxide Cu(OH)2 (Table 2) and the hydroxy species appears to be able to bind with the pyroxene surface through a hydrogen bonding mechanism. Table 1 also shows the effect of the amine, diethylenetriamine (DETA), on the copper concentration on the surface. As can be seen the amine causes significant decreases in the surface concentration of copper. Fig. 1, which shows the surface coverage of pyroxene after exposure to copper sulphate, SIBX and various amines at pH 9, illustrates clearly that, at that pH, copper adsorbs onto the pyroxene surface but that its surface concentration is reduced when exposed to all the amines tested except for EDTA, prior to SIBX addition. It is not possible to draw any conclusions on the differences between the EDA, DETA and TETA since these are too small to be significant. Kelebek et al. (1996) have shown that DETA complexes readily with nickel and copper hydroxide species. However the inability of EDTA to remove copper from the surface is interesting. EDTA complexes readily with ions in solution whereas EDA is less active. It is thus possible that, given the concurrent presence of calcium and
Pyroxene (95% Confidence Interval)
10 8 6 4 2
Cu+EDTA+SIBX
Cu+TETA+SIBX
Cu+EDA+SIBX
Cu+DETA+SIBX
Cu+SIBX
SIBX
-2
No Reagents
0
Pyrox only
Cu Surface Coverage (Relative %)
12
±1.96*Std. Err. ±1.00*Std. Err. Mean
Fig. 1. Relative copper surface coverage of pyroxene exposed to copper sulphate, SIBX and various amines at pH 9 using ToF-SIMS.
802
C.T. OÕConnor et al. / Minerals Engineering 19 (2006) 799–806 Pyroxene 95% Confidence Interval
Pyroxene 95% Confidence Interval
10
1.1
(b) Nickel Surface Coverage (Relative %)
1.0
0.9
0.8
0.7
0.6 Nicke l Surfa ce Co
Nickel Surface Coverage (Relative %)
(a)
0.5
0.4
8 7 6 5 4 3 2
0.3 pH 6
9
pH 9
±1.96*Std. Err. ±1.00*Std. Err. Mean
1
±1.96*Std. Err. ±1.00*Std. Err. Mean
Fig. 2. Relative nickel surface coverage at pH 6 and pH 9 using ToF-SIMS of (a) pyroxene only; and (b) pyroxene exposed to nickel ions.
magnesium ions in the water, the EDTA, from a stoichiometric aspect, is totally consumed by complexing with the solution ions and that there is little left, at the dosage used in the present study, to complex with the adsorbed copper hydroxy species. The EDA on the other hand will hardly complex with the solution species and thus be readily available to complex with and remove back into solution the adsorbed copper species. The authors are presently quantifying such interactions with a view to testing this hypothesis. As a matter of interest, in a separate set of experiments (Malysiak et al., 2004) it was found that the amines could not remove copper from the surface of pentlandite. Clearly the copper hydroxy species bind very strongly to the pentlandite and that this cannot be overcome by the amine– copper hydroxy interaction. Fig. 2 shows the results of a ToF-SIMS study of the results obtained for pure pyroxene and after it had been exposed to nickel ions by contacting it with a solution into which nickel has leached from the pentlandite present in the pulp. The relative abundance of nickel on the pyroxene surface was found to be almost four times greater at pH 9 than pH 6. The use of pH 6 as opposed to the value of 4 used in the case of copper was because of the frequent use of pH 6 conditions on flotation plants in which pyroxene and pentlandite occur. In these experiments, but not shown in this figure, the pyroxene was then exposed to SIBX and amines as had been done in the case of copper (Fig. 1). Similar results were observed, viz. that the amines EDA, DETA and TETA caused the surface concentration of the nickel to decrease significantly but this did not happen in the case of EDTA. Fig. 3 shows the surface concentration of copper as a function of the sequence of addition of copper sulphate, SIBX and EDA. It is clear that it is necessary to add the
EDA before exposing the copper activated pyroxene to the collector SIBX in order to reduce the surface concentration of copper. The bonding of the SIBX to the Cu(OH)+ or Cu(OH)2 is presumably sufficiently strong to ensure that the amine is not able to compete successfully to react with the copper species present on the surface. ToF-SIMS images of pyroxene exposed to pentlandite, copper sulphate and SIBX showed clearly the presence of copper and nickel on the surface of the pyroxene and also the presence of xanthate species on the surface on which the copper and nickel respectively appear. These images are not presented here since they need to be in colour to be properly appreciated. Images obtained after treatment by EDA showed that this amine removed much of the copper from the pyroxene surface. Images of pyroxene exposed to synthetic water containing calcium at concentrations of 80 and 500 ppm respectively showed that at pH 9 and at the higher concentration the relative abundance of calcium on the surface was about twice that of concentration shown at pH 6. Calcium exists as the divalent ion at both pHs and thus the concentration effect is presumed to be the main reason for this observation. 3.1.2. Zeta potential determinations Zeta potential studies were carried out on pyroxene in the presence of SIBX, CuSO4, [CuSO4 + SIBX], and [CuSO4 + SIBX + various amines] respectively at pH values of 6, 8 and 10. Fig. 4 shows these results where the amine used was DETA. Similar measurements, not presented here, were made in the case of EDA and TETA. A number of observations are made in these experiments which are consistent with the ToF-SIMS results. In the first instance little copper is seen to adsorb at pH 6 which has been explained above as being a result of the strongly
C.T. OÕConnor et al. / Minerals Engineering 19 (2006) 799–806
803
Pyroxene (95% Confidence Interval) 14
Cu Surface Coverage (Relative %)
12 10 8 6 4 2 0 ±1.96*Std. Err. ±1.00*Std. Err. Mean
-2 Pn only SIBX No Reagents
Cu+EDA+SIBX Cu+SIBX+EDA Cu+SIBX EDA+Cu+SIBX
Fig. 3. Relative copper surface coverage of pyroxene exposed to copper sulphate, SIBX and EDA in different sequences at pH 9 using ToF-SIMS.
5
Zeta Potential [mV]
0 -5 -10 -15 -20 -25 5
6
7
8
9
10
11
pH No Reagents
CuSO4
SIBX
CuSO4+SIBX
CuSO4+DETA
CuSO4+DETA+SIBX
Fig. 4. Zeta potential of pyroxene exposed to various reagents. Order of addition follows sequence in legend.
solvated Cu2+ ion. Addition of SIBX hardly affects the zeta potential values. Addition of copper sulphate markedly increased the zeta potential at pHs of 8 and higher. It has been mentioned above that at alkaline pHs the copper exists as either the positively charged Cu(OH)+ species or as Cu(OH)2. This increase in zeta potential may indicate that it is the former species that is dominant. Consistent with the ToF-SIMS results the zeta potential becomes more negative after the addition of the amine due to the removal of the positively charged copper species. SIBX addition thereafter has little effect on the zeta potential. When SIBX is added after copper sulphate, the zeta potential value becomes more negative due to the charge balanc-
ing role of the SIBX in its reaction with the positively charged copper hydroxy species. Fig. 5 shows the results of an experiment in which pyroxene was exposed to calcium ions illustrating that this resulted in a significant increase in the positive charge of the surface, especially at pH of 10, indicating that a surface reaction is occurring between the pyroxene and the calcium ions. 3.2. Microflotation tests The above results clearly indicate that the surface of pyroxene is modified when exposed to copper, nickel and calcium in solution by virtue of the fact that these ions
804
C.T. OÕConnor et al. / Minerals Engineering 19 (2006) 799–806 Zeta Potential 20 15 10 5 0 -5 -10 -15 -20 -25
5
6
7
8
9
10
11
pH
80ppm of Ca ions
500ppm of Ca ions
Fig. 5. Zeta potential of pyroxene exposed to calcium ions at two different concentrations.
adsorb onto the surface of the mineral. This may inadvertently modify the flotation behaviour of these minerals. Consequently a set of experiments was carried out to ascertain the effect of these ions on the flotation of pyroxene. A microflotation technique was used which essentially measures the relative hydrophobicity of the mineral. Fig. 6 shows these results at pH 4 and 9 in the presence of SIBX alone and SIBX and copper sulphate. At pH 9 when pyroxene is exposed to a solution containing copper sulphate followed by the addition of sodium isobutyl xanthate (SIBX) it becomes strongly hydrophobic and the recovery is greater than 70%. At pH 4 there is an increase in recovery from 45% to 59%. At pH 9 this increased from 45% to 80%. All these microflotation results are consistent with the observations made in both the ToF-SIMS and zeta potential experiments. In the former case as was shown in Table 1 the copper surface coverage (the copper existing at that pH as a copper hydroxy species) was greater at pH 9 than at pH 4. Since copper is still in its divalent form at pH 6 it is assumed that the coverage would still be low. Similarly the zeta potential results at pH 9 in the presence of copper
100 90 80 70 60 50 40 12 30 . 20 10 0 0
sulphate was about zero whereas at pH 6 it was <20 mV. Fig. 7 shows the recoveries obtained at pH 9 when the pyroxene was exposed to EDA and EDTA respectively after copper activation. As expected the EDA resulted in a decrease in recovery due to its ability to decrease the surface concentration of the activating copper species. The EDTA, as expected from the surface analysis results, did not have any noticeable effect of the recovery of the pyroxene. There are two issues of particular fundamental interest arising out of these results. Firstly the mechanism whereby the copper interacts with the pyroxene surface and the subsequent reaction of the SIBX with the copper activated pyroxene is of special interest since it is that interaction which renders the pyroxene hydrophobic. Secondly the mechanism whereby the amines are able to reduce the copper-activation effect is also of particular interest. Nagaraj and Brinen (1996) in a study of a copper– pyroxene system and using XPS data, found that copper was present as Cu+ rather than Cu2+ after collector treatment of copper activated pyroxene. This suggests that a chemical reaction is occurring during the adsorption of
Recovery (%)
0.01 0.89
5
10
15
20
25
Time (min) pH 4, No Reagents
pH 9, No Reagents
pH 4, SIBX
pH 9, SIBX
pH 4, CuSO4 + SIBX
pH 9, CuSO4 + SIBX
Fig. 6. Recovery of pyroxene following various reagent treatments at pH 4 and 9.
C.T. OÕConnor et al. / Minerals Engineering 19 (2006) 799–806
805
Recovery (%)
100 80 60 40 20 0 3
10
20
Time (min) No reagents
CuSO4+SIBX
CuSO4+EDA(4.00E-05M)+SIBX
CuSO4+EDTA(4.93E-05M)+SIBX
Fig. 7. Recovery of pyroxene following addition of EDA and EDTA respectively before the addition of SIBX after copper activation (pH 9).
xanthate onto the copper activated pyroxene, which involves the formation of either a CuX or a Cu(OH)X species. The following mechanism for the adsorption of copper at pH 9 is proposed. Cu(OH)2 + 2X $ Cu(OH)X [+OH + X ] $ CuX2 + 2OH In this mechanism two different pathways to the formation of a surface xanthate complex are proposed. One possibility is that adsorbed copper hydroxide reacts with a xanthate ion to form the Cu(OH)X species which renders the pyroxene hydrophobic. Alternatively the Cu(OH)X species could react further to form copper xanthate, viz.: CuX2 + 2OH ! CuX + 1/2X2 + 2OH This mechanism allows for the presence of monovalent copper but requires the formation of dixanthogen which is unlikely to be forming at the Eh values at which these experiments were carried out, viz. 100 mV. In both of the above cases it is important to note that these proposals assume that the rate of surface adsorption is less than rate of copper complex formation. Both the microflotation and zeta potential results showed that on its own SIBX had little effect on the behaviour of pyroxene at either pH. In either case the presence of SIBX showed little difference relative to the measurement for the pure mineral. In preliminary experiments it was observed that a lower recovery was obtained in the presence of calcium ions indicating that the presence of these ions in some way inhibits the possibility of inadvertent activation of pyroxene by the pentlandite dissolution products, viz. Ni2+. It is not clear why the calcium should have this effect. Calcium exists as the divalent ion at pH 9 whereas nickel exists as the hydroxy species (cf. Table 2). Since the heat of formation of a Ca–O bond is 36 kJ/mol and of Ni–O is 13 kJ/mol it may be that in some way the effect is driven by thermodynamic considerations. The role of the amines appears to be related to their effect on the equilibrium concentration of the various species in solution at the pH of interest. A series of calcula-
tions were carried out using the following molar concentrations of ions in solution: Cu2+ = 105; Ca2+ = 103; Ni2+ = 105,6,7; Fe3+ = 105 and [amines] = 106,5,4. The calculations were carried out at pH 9 using the programme JESS (Joint Expert Speciation Simulation) which calculates the equilibrium concentration of all the species in solution taking into account precipitation. These calculations, which represent preliminary findings in view of the fact that the authors are presently carrying out much more extensive analysis of the solution chemistry of this and other similar systems, revealed that, relative to EDA, DETA and TETA, EDTA complexes under the above conditions much more significantly with the other ions in solution than with copper even though EDTA would generally be expected to complex readily with copper. Thus those amines which were able to reduce the effect of copper activation of pyroxene, viz. EDA, DETA and TETA, may achieve this as a result of their not complexing as readily as EDTA with the ions such as calcium, iron and nickel. Thus the solution is essentially starved of free EDTA ions relative to EDA, DETA and TETA. These latter amines then complex with the copper ions in solution and in doing so affect the equilibrium between the copper hydroxide species which has precipitated out on the surface of the pyroxene and the solution ions of Cu2+ and 2OH. The end result is that the copper hydroxide species is resolubilized and thus the inadvertent activation of the pyroxene is reduced. Investigations are presently being carried out to test this hypothesis on similar systems. 4. Conclusion This study has shown clearly using zeta potential and TOF-SIMS analyses that when pyroxene is exposed to copper, nickel and/or calcium ions in solution at alkaline pHs a surface reaction occurs between the hydroxy species (in the case of copper and nickel) and the divalent calcium ions and the pyroxene surface which alters the surface of the pyroxene and thus influences its floatability. At pH 4 no adsorption of SIBX occurs possibly due to the
806
C.T. OÕConnor et al. / Minerals Engineering 19 (2006) 799–806
inhibiting effect of the solvation of the Cu2+. Following exposure to copper and nickel ions at pH 9, at which pH these species exist as hydroxy species, it was shown that addition of SIBX resulted in the pyroxene floating readily due to an interaction between the xanthate and the adsorbed metal hydroxy species. It is also shown that the activating effect of the adsorbed copper hydroxy species can be reduced by the use of complexing amines such as EDA, DETA and TETA which appear to be able to influence the equilibrium concentrations of the activating ions in a way which results in the equilibrium between precipitated copper hydroxide and the solution ions being disturbed in favour of dissolution of the precipitated species. Acknowledgements The authors wish to thank Anglo Platinum for permission to publish this paper. Anglo Platinum and the University of Cape Town are thanked for their financial support. Professor G.E. Jackson is thanked for carrying out the JESS calculations. References Hunter, R.J., 1993. Introduction to Modern Colloid Science. Oxford University Press.
James, R.O., Healy, T.W., 1972. Adsorption of hydrolyzable metal ions at the oxide water interface. Journal of Colloid and Interface Science 40, 53–64. Kelebek, S., Wells, P.F., Fekete, S.O., 1996. Differential flotation of chalcopyrite, pentlandite and pyrrhotite in Ni–Cu sulphide ores. Canadian Metallurgical Quarterly 35 (4), 329–336. Mackenzie, J.M.W., OÕBrien, N.T., 1969. Zeta potential of quartz in the presence of nickel (II) and cobalt (II). Society of Mining Engineers AIME 244, 168–172. Malysiak, V., 2003. Pentlandite–pyroxene and pentlandite–feldspar interactions and their effect on separation by flotation, Ph.D. thesis (unpublished), University of Cape Town. Malysiak, V., OÕConnor, C.T., Ralston, J., Gerson, A., Coetzer, L.P., Bradshaw, D.J., 2002. Pentlandite–feldspar interaction and its effect on separation by flotation. International Journal of Mineral Processing 66 (1–4), 89–106. Malysiak, V., Shackleton, N.J., OÕConnor, C.T., 2004. An investigation into the floatability of a pentlandite–pyroxene system. International Journal of Mineral Processing 74 (1–4), 251–262. Nagaraj, D.R., Brinen, J., 1995. SIMS study of metal ion activation in gangue flotation. In: Proceedings XIX International Mineral Processing Congress, SME, pp. 253–257. Nagaraj, D.R., Brinen, J., 1996. SIMS and XPS study of the adsorption of sulphide collectors on pyroxene: a case for inadvertent metal activation. Colloids and Surfaces A: Physicochemical and Engineering Aspects 116, 241–249. Phillips, D.M.A., Martin, C.J., Malysiak, V., 1993. Current technology in gangue flotation and concentrate. AngloPlatinum Report CJM12/ DML. Wesseldijk, Q.I., Reuter, M.A., Bradshaw, D.J., Harris, P.J., 1999. The flotation behaviour of chromite with respect to the beneficiation of UG2 ore. Minerals Engineering 12 (10), 1177–1184.