Pentlandite–feldspar interaction and its effect on separation by flotation

Int. J. Miner. Process. 66 (2002) 89 – 106 www.elsevier.com/locate/ijminpro

Pentlandite–feldspar interaction and its effect on separation by flotation V. Malysiak a,*, C.T. O’Connor b, J. Ralston c, A.R. Gerson c, L.P. Coetzer a, D.J. Bradshaw b a Anglo Platinum Research Centre, P.O. Box 6540, Homestead, 1412, South Africa Mineral Processing Research Unit, Department of Chemical Engineering, University of Cape Town, Rondebosch, Cape Town 7700, South Africa c Ian Wark Research Institute, University of South Australia, Mawson Lakes, Adelaide, South Australia, SA 5095, Australia b

Received 22 May 2001; received in revised form 5 October 2001; accepted 11 February 2002

Abstract The major loss of PGM and base metals in beneficiation of the Merensky Reef ore occurs during the separation of the siliceous gangue from the base metal sulphides and PGM minerals by selective flotation. The predominant gangue minerals in Merensky ores fed to the PGM concentrators are pyroxene and feldspar. Other important gangue minerals are talc and chlorite. A significant percentage of these minerals reports to the concentrate, thus diluting the concentrate grade and ultimately increasing the transport and smelting costs. Feldspar is naturally hydrophilic and requires activation for flotation. The present work investigates how collector adsorption (sodium isobutyl xanthate), copper sulphate activation, and the distribution of ions on mineral surfaces may influence selectivity in the flotation of these ores. The possible chemical reactions taking place on the surface of synthetic pentlandite, natural feldspar, and a 1:1 mixture were investigated. In the pH range studied, zeta potential and time of flight secondary ion mass spectrometry (ToF-SIMS) analyses indicated the presence of xanthate and copper ions on pentlandite surfaces. Xanthate ions did not adsorb onto feldspar surfaces. However, at pH 9, feldspar surfaces became coated with copper species and the subsequent adsorption of xanthate ions caused the feldspar to float. The results are explained in terms of the interaction between various copper and xanthate species present and the zeta potentials of the minerals at the pH values of interest. D 2002 Elsevier Science B.V. All rights reserved. Keywords: feldspar; pentlandite; flotation; zeta potential; ToF-SIMS

*

Corresponding author. Fax: +27-11-828-8990. E-mail address: [email protected] (V. Malysiak).

0301-7516/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 1 - 7 5 1 6 ( 0 2 ) 0 0 0 0 7 - 8

90

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

1. Introduction The Merensky ore is found in a platinum group metal (PGM) bearing reef in the Bushveld Igneous Complex in South Africa. Flotation is used to separate the valuable minerals from the siliceous gangue, with selective separation achieved by the adsorption of a collector onto the surface of the PGM and sulphide minerals. The aim is to maximise recovery of PGM and sulphide minerals and minimise the amount of gangue minerals in concentrates, since the latter have an adverse effect on smelting. Feldspars are by far the most abundant minerals in the earth’s crust, constituting about 65% of the latter by volume. They are characterised by a wide variation in chemical composition, crystal structure, and content of isomorphous admixtures, influencing their behaviour in technological processes. Feldspars are naturally hydrophilic and thus liberated feldspars should report to the concentrate only by entrainment and entrapment. However, a study carried out by Phillips et al. (1993) at a PGM concentrator indicated that 61% of the total gangue in the concentrate was due to flotation and 34% due to entrainment. It is highly unlikely that xanthate collectors exhibit any affinity for feldspar minerals to cause flotation. Therefore, it is likely that activation of the mineral surface by metal ions is probably occurring prior to xanthate adsorption. Reports published to date deal with conventional flotation of feldspar and feldspar flotation in a fluoride-free environment (e.g. Joy et al., 1966; Malghan, 1976). It is assumed that under normal flotation conditions, xanthate collectors should adsorb onto the surface of PGM and sulphide minerals only. Hodgson and Agar (1989) have shown that xanthate chemisorbs directly onto the pentlandite nickel sites at pH 9 and, subsequently, dixanthogen forms from the chemisorbed xanthate. They also found that calcium ions competed with xanthate ions for adsorption onto the surface sites of pentlandite. Rao and Finch (1991) found that the presence of cationic species in water appears to enhance the pyrrhotite xanthate-dixanthogen uptake at pH 8.4. They noted that this was especially apparent with calcium ions due to the formation of Ca(OH)+ species at the mineral surface, thus providing a greater number of positive surface sites. Activation of minerals with metal ions is a well-known phenomenon (Somasundaran, 1986). Isobutyl xanthate interactions with pentlandite were also studied by Bozkurt et al. (1998). They concluded that dixanthogen was the main adsorption product on pentlandite. It is well established that floatability of sulphide minerals is depressed by excessive oxidation. The alteration of pentlandite surfaces due to oxidation was studied by, among others, Richardson and Vaughan (1989), Buckley and Woods (1991), Kelebek (1993), and Legrand et al. (1997). The iron containing sulphide minerals are known to react in solution through the loss of iron ions from the sulphide lattice to form hydroxide overlayers. The aim of this study was to investigate the interactions between collector adsorption, ionic activation, and the distribution of ions on mineral surfaces in the flotation of a pentlandite – feldspar system. Specifically, the possible effect of interactions between sodium isobutyl xanthate, copper sulphate, and calcium ions on feldspar and pentlandite surfaces was investigated at pH 4 and 9 by means of zeta potential determinations, microflotation tests, and time of flight secondary ion mass spectrometry (ToF-SIMS) analysis.

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

91

2. Experimental methods 2.1. Minerals Natural feldspar from the Merensky Reef in the Northern Province, South Africa, was crushed to 2 mm and selected by hand-picking. Chemical analysis indicated that the feldspar contained 15.7% Al, 9.4% Ca, 2.7% Na, 1.3% K, 0.5% Fe, 0.3% Mg, and 22.1% Si. Pentlandite was synthesised at Anglo Platinum Research Centre. After the hydrogen desorption, reduced iron (11.0 g) and reduced nickel (11.86 g) were mixed with sulphur flake (11.41 g) and transferred to a quartz ampoule. The evacuated, sealed ampoule was heated in a furnace to 1150 jC. All mineral samples were stored under argon in a freezer and freshly ground in an agate mortar under argon in a glove box just prior to each experiment. The products were screened to obtain size fractions of 25 Am for zeta potential determinations and +38 – 106 Am for microflotation tests. The surface area of the feldspar and pentlandite were found to be 0.85 and 0.30 m2/g, respectively, using the BET method. 2.2. Reagents Purified sodium isobutyl xanthate (SIBX) was obtained from SENMIN. Other chemicals were of analytical grade quality. High purity argon was used throughout the study. Water with a specific conductance of 0.7 AS cm 1 with a surface tension of 72.8 mN m 1 at 20 jC, produced by a MILLI-RO PLUS apparatus, was used during the study. All experiments were carried out using aerated disodium tetraborate solution at 10 3 M (I=310 3 M) as a background electrolyte. Borate was selected for its buffer capacity, which simulates the buffering effect of pulp solution at PGM concentrators. Sodium hydroxide (0.1 M) and hydrochloric acid (0.1 M) were used for pH adjustment. The concentration of calcium ions was maintained at 510 4 M by the addition of Ca(NO3)2. Sodium isobutyl xanthate (510 5 M) and copper sulphate (510 5 M) were used as collector and activator, respectively. 2.3. Zeta potential determinations Zeta potential determinations were carried out on dilute dispersions of the individual minerals studied using a Malvern Zetasizer 4. The instrument gives the electrophoretic mobility from which the zeta potential was calculated using the Smoluchowski equation, since ja1, ja is the ratio of particle radius to double layer thickness (Hunter, 1993). The zeta potential determinations were carried out at pH 4, 6, 8 and 10 at 25 jC. A mineral sample weighing 0.08 g was dispersed in 80 cm3 of electrolyte solution and the pH was adjusted to the desired value. Conditioning of the mineral for zeta potential determinations was carried out for 30 min. The pH was checked prior to taking the reading. The Eh was allowed to vary naturally.

92

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

2.4. Microflotation tests A microflotation cell was used to determine the flotation response of selected minerals. 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. The tests on single minerals were carried out at pH 9 and mineral mixtures at pH 4 and 9. Blank experiments without any reagent additions were also performed to determine the contribution from entrainment and natural floatability. A 2 g sample (1 g of each mineral for mineral mixtures) was added to 250 cm3 of electrolyte, adjusted to the desired pH and conditioned with copper sulphate for 5 min and xanthate for 2 min. Flotation was carried out by introducing air at a flow rate of 5 cm3/min. Concentrates were collected at time intervals of 2, 5, 10 and 20 min for a single mineral and 3 and 20 min for mixtures of minerals. The floated and non-floated fractions were dried and weighed. In case of mineral mixtures, microflotation products were analysed for sulphur using a LECO analyser, thus enabling the recovery of each individual mineral to be determined. 2.5. Time of flight secondary ion mass spectrometry (ToF-SIMS) Surface analysis of samples 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 microchannel plate detector at the spectrometer, by means of a multistop time-todigital converter. The procedures for sample preparation were identical to those used for the microflotation tests. The samples were filtered and washed with water (conductivity 0.7 AS cm 1), adjusted to the desired pH, to remove any physically attached ions. All samples were dried in an argon atmosphere at ambient temperature. The 25-kV, 2-nA gallium beam was used throughout the investigation. About 15 grains for each mineral were imaged and analysed for Ca, Mg, Al, Si, Fe, Na, B, Cu, Ni, and K during positive ion analysis and O, OH, S, and xanthate during negative ion analysis. 2.6. SOLGAS Water program Speciation diagrams were calculated using the program SOLGAS Water (Eriksson, 1979) to determine the thermodynamically predicted species, both in solution and precipitated from the bulk solution onto the mineral surfaces. The aim of these calculations was to predict the solution conditions at which homogeneous (i.e. not mineral surface mediated) precipitation occurs, which may be reflected by the presence of colloidal particles on the mineral surfaces. Thermodynamic data have been taken from the literature. Equilibrium constants for Cu+, (EtX)2, HetX, CaOH+ liquid and the solid (EtX)2, CuEtX,

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

Fig. 1. Zeta potential of feldspar in Na2B4O7 at 10

3

93

M, I=310 3.

Cu(EtX)2, and Ca(OH)2 species were taken from Forssberg et al. (1984). Equilibrium constants for Cu(OH)+, Cu(OH)2, Cu(OH)3 , Cu(OH)42 , Cu2(OH)2+, and the solid Cu(OH)2 species were taken from Lindsay (1979).

3. Results 3.1. Zeta potential determinations Fig. 1 shows the zeta potentials at various pH values for feldspar, indicating that feldspar has a pHiep (isoelectric point pH) of about 4. A pH of 2 is usually reported in

Fig. 2. Zeta potential of pentlandite in Na2B4O7 at 10

3

M, I=310 3.

94

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

the literature (Fuerstenau and Fuerstenau, 1982). The higher iep obtained on the present sample is due to adsorption of a cation, probably iron, on the surface (Hudiburgh and Clifford, 1976). In the presence of SIBX (510 5 M), the zeta potential values are not significantly altered compared with those obtained without SIBX, and thus, it is concluded that xanthate ions do not adsorb onto the surface of feldspar. It is evident from Fig. 1 that, in the presence of copper (II) ions (510 5 M), the feldspar surfaces became more positively charged above pH 4. This indicates an adsorption of Cu2+ ions as well as various positively charged copper hydroxide species (CuOH+, Cu2(OH)22+) which are predominant below pH 9.5 (Acar and Somasundaran, 1992). The concentrations of these species are pH-dependent and they are known to specifically adsorb onto mineral surfaces. The iep value of copper oxide/hydroxide occurs at pHf9.5 (Fullston et al., 1999). In the presence of copper ions, subsequent introduction of xanthate ions (510 5 M) shifted the zeta potential versus pH curve to more negative

Fig. 3. Feldspar recovery – time curves in Na2B4O7 at 10

3

M, I=310 3, pH 9.

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

95

values, indicating that adsorption of xanthate ions onto the copper-activated feldspar surfaces occurred. The zeta potential versus pH data for synthetic pentlandite (Fig. 2) indicate that the mineral surfaces are highly oxidised, as would be expected after conditioning in an aerated solution for 30 min. Unoxidised sulphides, in general, have a pHiep of about 2 (Acar and Somasundaran, 1992). As the mineral surfaces oxidise and become coated with impurities, a higher pHiep is displayed. The results show that the synthetic pentlandite has a pHiep of about 8.6, which is close to that of iron oxide (Acar and Somasundaran, 1992), and that adsorption of xanthate ions occurs onto the pentlandite surfaces. The zeta potential experiments revealed that unlike xanthate, copper (II) activation is nonselective. Silicate and sulphide minerals became more positively charged above pH 4 in the presence of copper (II) ions and, subsequently, xanthate ions adsorbed onto the copper (II)-activated mineral surfaces.

Fig. 4. Pentlandite recovery – time curves in Na2B4O7 at 10

3

M, I=310 3, pH 9.

96

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

3.2. Microflotation tests Microflotation tests were carried out on single minerals at pH 9 as well as on mineral mixtures at pH 4 and 9 in order to determine possible interactions between the minerals and the reagent distribution on their surfaces. The possible effect of subsequent adsorption of xanthate ions on copper (II)-activated feldspar surfaces was also investigated. To evaluate the contribution from entrainment, flotation tests were conducted without any reagent addition. The data obtained on single minerals are given in Figs. 3a,b and 4a,b. Figs. 5 and 6 summarise the results obtained on the mixture of minerals. The feldspar results presented in Fig. 3a and b (3a gives the recovery– time curves and 3b gives the fraction remaining versus time on log-linear coordinates) showed that xanthate on its own did not affect the floatability. However, a high recovery of feldspar was observed in the presence of copper (510 5 M) and xanthate ions (510 5 M). In the presence of calcium ions (510 4 M), calcium (II) and copper (II) ions appeared to compete for feldspar surfaces. This led to inhibition of feldspar flotation, thus the ultimate recovery obtained was similar to that without any reagent addition. Rao and Finch (1991) showed that the presence of cationic species in solution enhanced the xanthatedixanthogen uptake by pyrrhotite at pH 8.4. In the presence of Cu2+ ions, xanthate uptake was, however, significantly higher compared with that observed in the presence of Ca2+ ions, which do not cause precipitation of xanthate. When the copper sulphate was added after xanthate addition, the recoveries were essentially the same as in the case of xanthate only, indicating that the copper sulphate was unable to induce desorption of the xanthate. The pentlandite microflotation data summarised in Fig. 4a and b (4a gives the recovery– time curves and 4b gives the fraction remaining versus time on log-linear coordinates) revealed that collectorless flotation, i.e. flotation in the absence of collector as a result of mild oxidation of the mineral itself, was low at pH 9 and can mostly be

Fig. 5. Feldspar (in mixture with pentlandite) total recovery in Na2B4O7 at 10

3

M, I=310 3.

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

Fig. 6. Pentlandite (in mixture with feldspar) total recovery in Na2B4O7 at 10

97

3

M, I=310 3.

attributed to entrainment. In the presence of xanthate ions, the recovery increased as expected. The addition of copper sulphate and subsequent introduction of xanthate collector increased the recovery to 97.0%. Calcium (II) ions added at the concentration levels tested did not affect either the recovery and/or flotation rate of pentlandite. When the copper (II) sulphate was added after SIBX addition, the recoveries were virtually identical to those obtained using only SIBX, thus indicating that the copper (II) sulphate was not able to induce desorption of the xanthate. Microflotation tests carried out on mineral mixtures (Fig. 5) revealed lower feldspar recoveries at pH 4 compared with pH 9. It is important to note that feldspar flotation at pH 4 was reduced in the presence of copper (II) and calcium (II) ions. Pentlandite recoveries in the mixtures (Fig. 6) showed that the mineral floated well at pH 4 without any reagent added, probably due to the dissolution of oxide and hydroxide overlayers at acidic pH (Buckley and Woods, 1991). Zeta potential determinations suggested that copper (II) ions do not adsorb onto the mineral surfaces at pH 4 to the same extent as observed at pH 9. Nevertheless, the copper (II) ions surface coverage was sufficient to enhance pentlandite floatability at pH 4. As with the single mineral, in the presence of calcium (II) ions (510 4 M), the flotation response of pentlandite in the mixture with feldspar was not significantly affected compared with the case of copper (II) ions. 3.3. ToF-SIMS analysis ToF-SIMS analyses were carried out on the microflotation products of the tests conducted on mineral mixtures at pH 9. It should be noted that the ToF-SIMS technique is not quantitative, hence only the trends obtained are compared rather then the absolute values. Fig. 7 shows that surface coverage of feldspar by copper (II) species is lower in the presence of calcium (II) ions compared with that observed in the presence of xanthate ions.

98

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

Fig. 7. Relative percentage abundance of copper (II) ions on feldspar surface.

This lower copper (II) ion concentration probably contributed to the lower feldspar recovery observed during the microflotation tests. Fig. 8 reveals that the surface coverage of pentlandite by copper (II) species is hardly affected by the addition of calcium (II) ions. There is a higher concentration of copper (II) ions observed on the pentlandite surfaces compared with the feldspar surfaces. The results presented in Fig. 9 clearly showed that the presence of copper (II) species increased the xanthate ion concentration on the feldspar surfaces. The microflotation data

Fig. 8. Relative percentage abundance of copper (II) and (I) ions on pentlandite surface.

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

99

Fig. 9. Relative percentage abundance of xanthate ions on feldspar surface.

revealed that this xanthate ion surface coverage is sufficient to induce positive flotation of feldspar.

4. Discussion A calculation of the speciation of ethyl xanthate at pH 9 was performed for a range of Eh conditions (Fig. 10). The equilibrium constants used for this SOLGAS Water calculation are given in Table 1 as well as the equilibrium constants for the other solution speciation calculations described in this section. The sodium isobutyl xanthate species were not used due to the lack of reliable equilibrium constant values available in the literature. It was assumed that the ethyl and isobutyl would behave similarly, due to their

Fig. 10. Speciation diagram for ethyl xanthate (EtX) at pH 9.

100

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

Table 1 Species and equilibrium constants used for solution speciation calculations Species

Equilibrium constant

Stoichiometric contributions Cu2+

H+

e

EtX

Ca2+

Solution species Cu2+ e H+ EtX Ca2+ Cu1+ Cu(OH)1+ Cu(OH)2 Cu(OH)3 Cu(OH)24 Cu2(OH)2+ (EtX)2 HetX CaOH+

0 0 0 0 0 2.59 7.70 13.80 26.80 39.60 10.70 2.54 1.52 12.70

1 0 0 0 0 1 1 1 1 1 2 0 0 0

0 1 0 0 0 1 0 0 0 0 0 2 0 0

0 0 1 0 0 0 1 2 3 4 2 0 1 1

0 0 0 1 0 0 0 0 0 0 0 2 1 0

0 0 0 0 1 0 0 0 0 0 0 0 0 1

Solids (EtX)2 (liquid) CuEtX Cu(EtX)2 Cu(OH)2 Ca(OH)2

2.37 23.01 26.75 7.70 22.81

0 1 1 1 0

2 1 0 0 0

0 0 0 2 2

2 1 2 0 0

0 0 0 0 1

closely related molecular properties. However, due to the higher insolubility of sodium isobutyl xanthate compared with ethyl xanthate, the formation of the dimer may occur at a lower Eh (i.e. under less oxidising conditions) than for ethyl xanthate. As there is little dependence of the major species shown in Fig. 10 on proton concentration, the analogous diagram for pH 4 is almost identical and it is not given. The similarity of the zeta potential data for feldspar (Fig. 1) in the presence or absence of xanthate ions indicates no change in surface speciation on addition of SIBX across the entire range of pH values examined (from pH 4 to 10). As expected from the zeta potential data (Fig. 1), the addition of SIBX had essentially no influence on the flotation response of feldspar, either in the single mineral system at pH 9 (Fig. 3a and b) or in the mixed mineral system at pH 4 or 9 (Fig. 5). This suggests that the dimer [SIBXs (l)] is not present as this would deposit on the mineral surfaces where it may alter the zeta potential determinations and increase the flotation response. SIBX is therefore probably present in solution in the anionic form (SIBX ), which would not be expected to adsorb electrostatically onto the negative feldspar surfaces. This is consistent with the speciation data of Fig. 10. The lack of change in either the zeta potential curve or the flotation response indicates that the xanthate ions do not adsorb chemically either. Pentlandite does not show the same lack of response on addition of SIBX. In this case, the zeta potential data are more negative across the entire pH range compared with pentlandite in the absence of SIBX (Fig. 2). It would therefore appear that adsorption of xanthate ions onto the pentlandite surfaces is occurring. The degree of adsorption

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

101

Fig. 11. Copper speciation diagram for pH 9 at a range of Eh conditions.

decreases with increasing pH until pH 10, where the zeta potentials for pentlandite with and without the addition of SIBX are approximately equal. It has already been shown in the feldspar study that SIBX is probably not in the SIBXs (l) form. It is most probable that the negatively charged SIBX species is adsorbed onto the positively charged pentlandite surface at pH values below 8.5. Above pH 8.5, the zeta potential of the pentlandite surfaces becomes negative and there is negligible adsorption of anionic isobutyl xanthate. As implied by the zeta potential determinations, an increase in the flotation response of pentlandite in the presence of xanthate ions is observed both in the single mineral system at pH 9 (Fig. 4a and b) and in the mixed mineral system at pH 4 and 9 (Fig. 6). For reasons described above regarding the relative zeta potentials and speciation at pH 4, the floatability of pentlandite is higher at pH 4 compared with pH 9. Fig. 11 shows the speciation predicted at pH 9 for a solution containing 510 5 mol dm 3 copper (II) ions. At oxidising Eh values (i.e. above 0 mV) at pH 9, the solubility of copper (II) ions is approximately equal to the concentration of Cu(OH)2 species in solution of 7.910 7 mol dm 3. The remainder of the copper (II) ions added, 4.9210 5 mol dm 3 from an initial solution concentration of 510 5 mol dm 3, would then be present as Cu(OH)2 precipitant, assuming no adsorption onto the mineral surfaces. Cu(OH)2 precipitants would first appear at about pH 7 (Fig. 12). Although the data in Fig. 12 were

Fig. 12. Copper speciation as a function of pH at a constant Eh of 300 mV.

102

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

determined at an Eh of 300 mV, the onset of Cu(OH)2 is not strongly dependent on Eh. At low pH, the positive copper (II) ion species predominate. The adsorption of the positive copper ions at acidic pH onto the negatively charged surface of feldspar would cause the zeta potential to increase relative to untreated feldspar. This is indeed what is observed (cf. Fig. 1). The precipitation of deprotonated negative Cu(OH)2 colloids at high pH would lead to the marked decrease in zeta potential observed above pH 8 (cf. Fig. 1). The relationship between the zeta potential curve of the CuSO4-treated pentlandite and the untreated pentlandite is not the same as for feldspar. In the case of pentlandite, as for feldspar, the zeta potential curve for the CuSO4-treated mineral sample is more positive above pH 4 than for the untreated sample. However, the increase in relative zeta potential occurs between pH 4 and 6, above which the difference between the two curves remains approximately constant. The positive zeta potential of the untreated pentlandite between pH 4 and 8.5 rules out the possibility of an electrostatic attraction with the positive copper (II) ions, which dominate the copper (II) solution species up to pH 7 (Fig. 12). At high pH, in the case of the CuSO4-treated pentlandite, the shape of the zeta potential curve mirrors that of the untreated pentlandite sample. This suggests that the deposition of Cu(OH)2 does not play a significant role in the pentlandite surface alteration. It can therefore be concluded that the copper (II) adsorption processes occurring on the pentlandite surfaces are not the same as those on feldspar surfaces. Moreover, the form of the interaction between the copper (II) ions in solution and the pentlandite surfaces is most likely to be chemical (charge transfer mechanism) rather than electrostatic (attraction between opposite charges). The degree of this chemical interaction is sufficient at high pH to remove enough copper (II) ions from solution and thus inhibit Cu(OH)2 precipitation. The ToF-SIMS analyses (Figs. 7 and 8) indicate that the copper ion surface concentration on the pentlandite surfaces is higher compared with that on the feldspar surfaces when the minerals are present in a mixture. The zeta potential curve for feldspar treated with CuSO4 and SIBX is significantly more negative compared with feldspar treated with CuSO4 only. As SIBX on its own did not alter the zeta potential of feldspar, it would appear that xanthate ions complex with the copper (II) ions on the mineral surfaces. In the pentlandite system treated with both CuSO4 and SIBX, the shape of the zeta potential curve is almost identical to that of the CuSO4treated pentlandite sample but at a zeta potential approximately 20 mV lower. This may indicate that the surface concentration of SIBX remains fairly constant across the pH range studied. The speciation diagrams for a solution containing 510 5 mol dm 3 SIBX at pH 4 and 9 as a function of Eh are shown in Fig. 13a and b, respectively. The predominant copper (II) xanthate species at pH 9 is likely to be a mixture of Cu(OH)2 and either Cu(I) and Cu(II) xanthate precipitants. Between approximately 250 and 400 mV, almost half of the precipitated Cu(OH)2 colloids are likely to be converted to hydrophobic Cu(II)-X colloids, assuming that the system is not kinetically controlled. This is reflected in the single mineral feldspar flotation study carried out at pH 9 where the addition of xanthate ions to the copper sulphate-treated sample resulted in a marked increase in flotation response (Fig. 3). For pentlandite, the increase in flotation response upon the addition of xanthate ions to the copper (II) sulphate-treated sample is even more profound (Fig. 4), resulting in a recovery of almost 100%. It is possible that the adsorbed copper (II) ions (not precipitated

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

Fig. 13. Speciation diagrams at (a) pH 4 and (b) pH 9 for 510

5

mol dm

3

103

isobutyl xanthate and copper.

Cu(OH)2) complex with xanthate ions in a more evenly distributed in form and thus greater hydrophobicity is achieved. In the mixed mineral system at pH 4, feldspar flotation diminishes, while at pH 9, flotation increases compared with the untreated mixed mineral flotation experiment and the experiment with only isobutyl xanthate added. At both pH 4 and 9, the flotation response of pentlandite increases compared with the mixed mineral cases with either no reagents or just SIBX added. It would thus appear that at pH 4 pentlandite preferentially adsorbs copper (II) ions compared with feldspar. Moreover, Cu(OH)2 precipitation is unlikely at pH 4. Therefore, on the feldspar surfaces, there is no conversion of the hydrophilic colloids to the hydrophobic Cu(II)-X colloids. Introduction of isobutyl xanthate thus leads to an increased flotation response for pentlandite but not for feldspar. At pH 9, two competing processes are likely to be occurring, viz. adsorption of copper (II) ions from solution onto the minerals surfaces (via a chemical adsorption process in the case of pentlandite and via electrostatic attraction for feldspar) and Cu(OH)2 precipitation.

104

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

Only half of the pentlandite mass was present during the tests with the mineral mixture as compared to the single mineral flotation case. Therefore, even if Cu(OH)2 precipitation is inhibited in the single mineral system due to the rapid adsorption of copper (II) ions from solution onto the pentlandite surfaces, this is not necessarily true for the mixed mineral case. Some xanthate ions will adsorb onto the copper-activated pentlandite surfaces, thus increasing the hydrophobicity and flotation response. Residual xanthate ions will react with the precipitated Cu(OH)2, thus rendering these previously hydrophilic colloids hydrophobic and increasing the flotation response of feldspar. There is evidence from ToF-SIMS analyses (Fig. 9) that the presence of copper (II) species increased the xanthate concentration on the feldspar surfaces. If less copper ions were to be added to solution, or if a greater concentration of pentlandite was present, Cu(OH)2 precipitation could be inhibited and thus only the flotation response of pentlandite would be increased rather than the flotation response of both pentlandite and feldspar. On addition of 510 4 mol dm 3 calcium ions to the solution prior to addition of copper and xanthate ions in the single mineral system (pH 9), the feldspar recovery was similar to that obtained with untreated feldspar and feldspar treated with SIBX. Ca(OH)2 is neither predicted to precipitate at the calcium concentration studied nor any specific interactions between xanthate and calcium ions are likely to occur (Table 1). Therefore, if the zeta potential determinations were to be carried out on the feldspar treated with calcium (II) ions, similar behaviour as for the feldspar treated with CuSO4 could be predicted at low pH. At higher pH, however, the decrease in surface charge should be less pronounced. The nature of the adsorption of the calcium (II) and copper (II) ions onto the feldspar surfaces is likely to be electrostatic. Therefore, after preadsorption of calcium (II) ions onto the mineral surfaces, there is no further driving force for copper (II) ions adsorption. Indeed, the ToF-SIMS analyses revealed lower copper (II) ion concentration in the presence of calcium (II) ions on the feldspar surfaces (Fig. 7). As xanthate ions do not interact with calcium (II) ions, the flotation response of the feldspar was similar to that obtained with no treatment. For pentlandite, the zeta potential is close to neutral at pH 8.5, and hence, electrostatic interaction of copper (II) ions and the pentlandite surfaces will not take place. The interaction appears to be chemical in nature. It is apparent from the single mineral flotation response that calcium ions do not chemically interact with the pentlandite surfaces and hence do not interfere with the subsequent adsorption of copper ions. This is also indicated in the ToF-SIMS analyses (Fig. 8). This interpretation of the nature of the interactions of the minerals with calcium ions is derived from the mixed mineral flotation responses. At pH 4, the flotation recovery of feldspar was low and even more reduced with the addition of calcium ions while the flotation response of pentlandite is not significantly affected. At pH 9, the flotation response of feldspar decreases due to the lack of adsorption of both copper ions and Cu(OH)2 colloids.

5. Conclusions A significant percentage of gangue minerals occurring in Merensky ore reports to the concentrate, thus diluting the concentrate grade. The objective of this study was to

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

105

investigate how sodium isobutyl xanthate adsorption, copper sulphate activation, and the distribution of ions on mineral surfaces may influence selectivity in the flotation of these ores. The possible chemical reactions taking place during a beneficiation process were studied on synthetic pentlandite, natural feldspar, and a 1:1 mixture of these minerals using zeta potential determinations, microflotation tests, and ToF-SIMS analyses. The results obtained indicate that the pentlandite and feldspar used during the study have a pHiep of about 8.6 and 4.2, respectively. The zeta potential versus pH curves and ToF-SIMS analyses indicate the presence of xanthate and copper (II) ions on pentlandite surfaces in the pH range studied. Pentlandite recovery was significantly enhanced in the presence of xanthate ions and even more in the presence of copper and xanthate ions at pH 4 and 9. Xanthate ions do not adsorb onto feldspar surfaces. However, it is evident from the data obtained that at pH 9 feldspar surfaces become coated with copper (II) species and subsequent adsorption of xanthate ions caused the feldspar to float. Feldspar floatability was higher at pH 9 compared to pH 4 at the experimental conditions tested. When calcium (II) ions are present at pH 9, they inhibit copper (II) ions adsorption, thus causing the flotation recovery to decrease to values observed in the absence of copper. The pentlandite recovery was not significantly affected by calcium ions at the concentration tested. These observations have been shown to be generally consistent with the predictions of the forms of species present at the pH values of interest.

Acknowledgements The authors would like to thank Anglo Platinum management for support during the study and the permission to publish the paper. They also record their thanks to Mrs. N. Shackleton and Dr. N.D. Plint for carrying out ToF-SIMS analysis and for valuable comments.

References Acar, S., Somasundaran, P., 1992. Effect of dissolved mineral species on the electrokinetic behaviour of sulphides. Minerals Engineering 5 (1), 27 – 40. Bozkurt, V., Xu, Z., Finch, J.A., 1998. Pentlandite/pyrrhotite interaction and xanthate adsorption. International Journal of Mineral Processing 52, 203 – 214. Buckley, A.N., Woods, R., 1991. Surface composition of pentlandite under flotation-related conditions. Surface and Interface Analysis 17, 675 – 680. Eriksson, G., 1979. An algorithm for the computation of aqueous multi-component multiphase equilibria. Analytica Chimica Acta 112, 375 – 383. Forssberg, K.S.E., Antti, B.M., Palsson, B.I., 1984. Computer-assisted calculations of thermodynamic equilibria in the chalcopyrite – ethyl xanthate system. In: Jones, M.J., Oblatt, R. (Eds.), Reagents in the Minerals Industry. The Institution of Mining and Metallurgy, London, pp. 251 – 264. Fuerstenau, D.W., Fuerstenau, M.C., 1982. The flotation of oxide and sulphide minerals. In: King, R.P. (Ed.), Principles of Flotation. South African Institute of Mining and Metallurgy, Johannesburg, pp. 109 – 158. Fullston, D., Fornasiero, D., Ralston, J., 1999. Zeta potential study of the oxidation of copper sulphide minerals. Colloids and Surfaces, A: Physicochemical and Engineering Aspects 146, 113 – 121. Hodgson, M., Agar, G.E., 1989. Electrochemical investigations into the flotation chemistry of pentlandite and pyrrhotite: process water and xanthate interactions. Canadian Metallurgical Quarterly 28 (3), 189 – 198.

106

V. Malysiak et al. / Int. J. Miner. Process. 66 (2002) 89–106

Hudiburgh, G.W., Clifford, K.R., 1976. Amine flotation of feldspar from a magnetite concentrate. Society of Mining Engineers of AIME 260, 161 – 165. Hunter, R.J., 1993. Introduction to Modern Colloid Science. Oxford Univ. Press, Oxford, New York. Joy, A.S., Manser, R.M., Lloyd, K., Watson, D., 1966. Flotation of silicates: 2. Adsorption of ions on feldspar in relation to its flotation response. Institution of Mining and Metallurgy 75, C81 – C86. Kelebek, S., 1993. The effect of oxidation on the flotation behaviour of nickel – copper ores. XVIII International Mineral Processing Congress, Sydney, 23 – 28 May 1993. Legrand, D.L., Bancroft, G.M., Nesbitt, H.W., 1997. Surface characterization of pentlandite, (Fe,Ni)9S8, by X-ray photoelectron spectroscopy. International Journal of Mineral Processing 51, 217 – 228. Lindsay, W.L., 1979. Chemical Equilibria in Solids. Wiley, New York. Malghan, S.G., 1976. Selective flotation of feldspar in non-fluoride medium. Society of Mining Engineers of AIME, Prepr. No. 76-H-331. Phillips, D.M.A., Martin, C.J., Malysiak, V., 1993. Current technology in gangue flotation and concentrate cleaning and its application to Amandelbult during 1991 – 1993. Anglo Platinum report CJM12/DML. Rao, S.R., Finch, J.A., 1991. Adsorption of amyl xanthate at pyrrhotite in the presence of nitrogen and implications in flotation. Canadian Metallurgical Quarterly 30 (1), 1 – 6. Richardson, S., Vaughan, D.J., 1989. Surface alteration of pentlandite and spectroscopic evidence for secondary violarite formation. Mineralogical Magazine 53, 213 – 222. Somasundaran, P., 1986. In: Somasundaran, P. (Ed.), Advances in Minerals Processing. Society of Mining Engineers of AIME (SME), Littleton, Colorado, pp. 137 – 153. 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.