Effects of free cyanide and cuprous cyanide on the flotation of gold and silver bearing pyrite

Minerals Engineering 71 (2015) 194–204

Contents lists available at ScienceDirect

Minerals Engineering journal homepage: www.elsevier.com/locate/mineng

Effects of free cyanide and cuprous cyanide on the flotation of gold and silver bearing pyrite Bao Guo a, Yongjun Peng a,⇑, Rodolfo Espinosa-Gomez b a b

School of Chemical Engineering, University of Queensland, St. Lucia, Brisbane, QLD 4072, Australia Minseg Pty Ltd, Carindale, Brisbane, QLD 4152, Australia

a r t i c l e

i n f o

Article history: Received 17 August 2014 Accepted 15 November 2014

Keywords: Cuprous cyanide Pyrite Flotation Cyclic voltammetry Electrochemical impedance spectroscopy

a b s t r a c t At gold and silver mineral processing plants, cyanide species are always present in the process water recycled to flotation circuits, despite the cyanide destruction process. The effect of cyanide, in particular, cuprous cyanide on gold and silver flotation has not been well understood. In the present study, free cyanide and cuprous cyanide species were isolated and their effects on the flotation of a pyritic ore were evaluated. It was found that free cyanide depressed gold and silver flotation through their carrier, pyrite. Cuprous cyanide mainly in the form of Cu(CN)2 depressed pyrite flotation similarly as free cyanide. 3 Electrochemical studies including open circuit potential (OCP), cyclic voltammetry and electrochemical impedance spectroscopy (EIS) techniques were carried out to understand the underpinning depression mechanism of cyanide species on pyrite flotation using xanthate as collector. It was found that all surface electrochemical reactions were inhibited by either free cyanide or cuprous cyanide. The surface layer as a result of xanthate adsorption on pyrite was completely removed in the presence of these cyanide species, which was suggested to contribute to the hydrophilic pyrite surface. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction When gold and silver are associated with sulphide minerals, they are often recovered with these sulphide minerals by flotation with the flotation concentrate leached to recover the precious metals or smelted to recover the precious and base metals. At Hidden Valley (HV) Mine, gravity separation is used to recover the coarse gold, followed by a bulk flotation step to recover most of the gold and silver minerals. The flotation concentrate is then leached and the gold and silver in solution are removed by a Counter Current Decantation (CCD) circuit and then recovered in a Merrill Crowe circuit. The precipitate from the Merrill Crowe circuit is calcined and smelted to produce a dore. The flotation and CCD tailings are sent to a Carbon in Leach (CIL) circuit to recover these precious metals and the carbon elution solution is sent to a separate Merrill–Crowe circuit and the precipitate is also calcined and smelted to produce a dore. The remaining cyanide in the final tailing is sent to an Inco Detox and a Caro’s Acid detox circuits, producing the main source of process water which is recycled to the upstream circuits such as grinding and flotation. At the time that this research was initiated, the Au and Ag recoveries were below the

⇑ Corresponding author. Tel.: +61 7 3365 7156; fax: +61 7 3365 3888. E-mail address: [email protected] (Y. Peng). http://dx.doi.org/10.1016/j.mineng.2014.11.016 0892-6875/Ó 2014 Elsevier Ltd. All rights reserved.

values achieved in the feasibility study. A contributing factor was the process water containing cyanide species which may have a deleterious effect on gold and silver flotation despite cyanide destruction processes. This is a common problem encountered in precious mineral processing plants (Adams, 2013). Cyanide is among the most commonly used flotation reagents to depress iron sulphides and enhance the separation efficiency of base metal sulphide flotation. At early stages, (Sutherland and Wark, 1955; Wark, 1938) reviewed the depression of various sulphide minerals by NaCN based on bubble contact tests. Then, several attempts have been made to explain the depression behaviour and the underpinning mechanism for pyrite (Grano et al., 1990; Hodgkinson et al., 1994; Janetski et al. 1977; Prestidge et al., 1993; Wet et al., 1997), pyrrhotite (Prestidge et al., 1993), chalcocite (Castro and Larrondo, 1981), and sphalerite (Buckley et al., 1989; Prestidge et al. 1997; Seke, 2005; Seke and Pistorius, 2006). It has been proposed based on thermodynamic considerations (Elgillani and Fuerstenau, 1968; Wang and Forssberg, 1996) that cyanide preferentially adsorbs on pyrite surface as iron cyanide compounds, inhibiting the chemisorption and oxidation of xanthate. However, the existence of insoluble iron cyanide compounds on pyrite surface has not been well confirmed by experimental studies (Prestidge et al., 1993). No continuous surface layers are formed as a result of the interaction between pyrite and cyanide according to electrochemical impedance spectroscopy studies

B. Guo et al. / Minerals Engineering 71 (2015) 194–204

(Wet et al., 1997). It was also proposed that cyanide as a strong reducing agent significantly decreased the flotation pulp potential, thus inhibiting the oxidation of xanthate which is essential for the hydrophobic surface of pyrite (Kocabag and Guler, 2007; Ralston, 1991; Miller et al., 2006). Janetski et al. (1977) and Wet et al. (1997) showed that cyanide reduced the surface electrochemical activities, limiting both anodic and cathodic reactions on pyrite surface. The above cyanide species are discussed on a basis of free cyanide including hydrocyanic acid (HCN) and the cyanide anion (CN). On the other hand, weak acid dissociable (WAD) cyanide species (cyanide complexes with Cu, Zn, Ag, Cd, and Ni, which dissociates under mildly acidic conditions to free cyanide) are also commonly seen in the process water and copper cyanide species are of the most importance. Despite the large volume of research on cyanide–pyrite interactions, there has been little research on the interactions of copper cyanide species with pyrite. It was first suggested by Wark (1938) that the contact of pyrite with air bubbles was still possible at a CuSO4/NaCN ratio of approximately 3/1 under which the formation of cuprous dicyanide complex Cu(CN) 2 was favoured in addition to cyanate (OCN), an oxidation product of cyanide. Copper cyanide speciation largely depends on cyanide concentration, cyanide to copper molar ratio, pH and salinity of the solution (Dai et al., 2012; Lu et al., 2002; Lukey et al., 1999). The interaction of various copper cyanide species with sulphide minerals needs to be examined in detail. In this study, the Hidden Valley Mine was taken as a case study to understand how free cyanide and cuprous cyanide affect gold and silver flotation. 2. Thermodynamic consideration The cyanide anion in aqueous solution forms hydrocyanic acid according to the following dissociation reaction (Marsden and House, 2006).

HCN $ Hþ þ CN

ð1Þ

The logarithmic exponent of the dissociation constant pKa of HCN at 25 °C is 9.21 (Lu et al., 2002). A pH value higher than 9.21 increases CN concentration while a pH value lower than 9.21 promotes the presence of aqueous HCN in the solution. HCN is a volatile substance and the formation of aqueous HCN would promote its volatilisation, which is responsible for cyanide loss in many gold processing plants (Smith and Mudder, 1991). A significant portion of the cyanide in the tailing dam is volatilised to HCN as the pH deceases and decanted solution has more contact with air (Lotter, 2006). It is estimated that volatilisation rates from ore processing operations such as leaching are generally low, with only 1% of total cyanide being lost through HCN volatilisation at an operation pH of around 10 (Heath et al., 1998 and Adams, 1990). The processing circuits where cyanide loss should be considered in the present paper include grinding and flotation. It is generally observed that in flotation circuits, the pulp pH is typically adjusted in the mildly alkaline range with little cyanide loss. However, the pH of the milling pulp drops significantly as a result of certain interactions including the hydrolysis of metal species to generate acid and carbon dioxide uptake from the air forming carbonic acid and oxidation of sulphur species in the sulphide minerals. Low pH may cause cyanide loss via HCN volatilisation during milling and probably change the cyanide speciation. In the present study, the HCN volatilisation rate from the flotation pulp with modified pH 10 is generally low as long as the cyanide is dosed after grinding circuit, while the majority of cyanide exists as CN in the solution. In copper cyanide system, the loss of cyanide is determined by both the dissociation of the copper cyanide complex and

195

subsequent volatilisation of the formed aqueous HCN. The rate of the first process was found to be dependent on the dissociation rates of the different cuprous cyanide complexes as shown from Eqs. (2)–(5).

CuCN $ Cuþ þ CN K sp

ð2Þ

Cuþ þ 2CN $ CuðCNÞ2 K 2

ð3Þ

Cuþ þ 3CN $ CuðCNÞ2 3 K3

ð4Þ

Cuþ þ 4CN $ CuðCNÞ3 4 K4

ð5Þ

The copper cyanide speciation under the present experimental conditions was modelled using a computer program Visual MINTEQ (version 3.0) (Gustafsson, 2012). In these modellings, copper(I) cyanide (CuCN) was specified as finite solid phase and was dissolved by CN at 25 °C. The solubility product (Ksp) of CuCN was defined at a value of 1019.5 with its reaction enthalpy DH° = 19 kJ/mol. In this program, the corresponding equilibrium constants for Eqs. (3)–(5) were defined as 1023.9, 1029.2, 1030.7, respectively and DH° values for those reactions were defined as 121, 167.4, 214.2 kJ/ mol, respectively. Parameters such as pulp potential Eh, pH and the molality of CuCN, CN and the balanced Na+ were considered to formulate the input data for the calculation. Cu+/Cu2+ redox couple was specified with log K = 2.69 and DH° = 6.9 kJ/mol. Oversaturated solids were allowed to precipitated. From this thermodynamic analysis it is possible to know the concentration of each cuprous cyanide species of interest. The distribution of cyanide species as a function of pH at CN/Cu mole ratio of 3/1 is presented in Fig. 1. The initial cyanide concentration was fixed at 10 ppm and the calculation suggests that there is no influence on cyanide speciation when Eh ranges from 0 mV to 550 mV (SHE) which is consistent with the expected pulp potential in flotation. At pH 10, the dominant cyanide species is cuprous tricyanide Cu(CN)2 3 (accounting for 81% of the initial cyanide) while Cu(CN) 2 accounting for 13%. This agrees with the previous study on the copper cyanide system (Lu et al., 2002). It is noted from Fig. 1 that both HCN and Cu(CN) 2 become dominant in the solution when pH drops down to 7, which is commonly observed during the grinding process. This suggests that H+ ions at low pH compete  with Cu(CN) 2 to complex with CN following Eq. (6).  þ CuðCNÞ2 3 þ H ! CuðCNÞ2 þ HCNðaqÞ

ð6Þ

Thermodynamic calculation for cyanide speciation at CN/Cu mole ratio of 2/1 is presented in Fig. 2. The initial cyanide concentration was fixed at 10 ppm and Eh at 300 mV (SHE). The only dominant

Fig. 1. Copper cyanide speciation at [CN]/[Cu] = 3/1, [CN] = 10 ppm, Eh = 300 mV (SHE), 25 °C.

196

B. Guo et al. / Minerals Engineering 71 (2015) 194–204

Fig. 2. Copper cyanide speciation at [CN]/[Cu] = 2/1, [CN] = 10 ppm, Eh = 300 mV (SHE), 25 °C.

species dissolved in the solution from pH 7 to pH 9 is cuprous dicyanide Cu(CN) 2 . When pH is increased to 10, a small amount of Cu(CN)2 species appears, while Cu(CN) 3 2 is still dominant. There is no free cyanide species either HCN or CN over the whole range of pH calculated. However, it has been suggested that Cu(CN) 2 is not stable in contact with air, oxidising to cupric species according to Eq. (7) (Vukcevic, 1996).

CuðCNÞ2 þ 0:25O2 þ 0:5H2 O þ OH ! CuðOHÞ2 þ 2CN

ð7Þ

It is seen from Fig. 3 that at CN/Cu mole ratio of 2/1 and pH 10, 3% of the initial cuprous species are oxidised to cupric species and precipitate as Cu(OH)2 in alkaline solution at Eh of 300 mV (SHE); while the oxidised copper is increased to 26% with Eh rising up to 500 mV (SHE). The oxidation of cuprous cyanide species and precipitation as copper hydroxide were also measured experimentally by Dai and Breuer (2009). The precipitation of Cu(OH)2 adversely affects the subsequent experiments due to copper being lost before reaching the expected cyanide to copper ratio of 2/1 and the cupric species formed may activate pyrite flotation accidently (Chandra and Gerson, 2009; Finkelstein, 1997; Weisener and Gerson, 2000). It is worth noting that practically higher cyanide concentrations are used to prepare the stock solutions and then added to flotation pulp. Furthermore, during the preparation of stock solution, pH is usually modified to above 10 to prevent any HCN volatilisation and dissolve CuCN solids as sufficient as possible because the stability constant of the equilibrium products is very close to the Ksp of CuCN; otherwise, CuCN would easily precipitate, especially at lower

Fig. 3. Copper(II) precipitation as a result of copper(I) oxidation at [CN]/[Cu] = 2/1, pH = 10, 25 °C.

pH values (Lu et al., 2002). However, it is found that the oxidation of Cu(CN) 2 is accelerated at a higher concentration (shown in Fig. 3) and a higher pH value (not shown here). In contrast, Cu(CN)2 3 is relatively stable with much higher tolerance to a solution chemistry change. Therefore, CN/Cu = 3/1 and a cyanide concentration of 10 ppm (equals to 0.3846 mM) were used for the present flotation study. To avoid the interference of free cyanide and its loss by HCN volatilisation resulting in the change of cyanide speciation, pH value was modified to 10 in flotation pulp. At this pH, the CN- concentration was lower than 0.5 ppm which was proved previously to have a negligible effect on the flotation performance of the same ore (Guo et al., 2013). The effect of cyanide concentration on copper cyanide speciation was presented in Fig. 4, showing Cu(CN)2 3 as the dominant species over a wide range of cyanide concentration. This implies that higher signal to noise ratio could be obtained using higher cyanide concentrations in electrochemical studies without changing cyanide speciation. It was also calculated from Visual MINTEQ that the copper cyanide speciation at CN/Cu = 3/1 showed no change in the presence of sodium borate (Na2B4O7) which was used as pH buffer for the electrochemical studies. 3. Materials and methods 3.1. Ore properties Samples from the primary sulphide deposit were collected at Hidden Valley Mine, Papua New Guinea. In laboratory, these samples were naturally dried, and crushed to a size of 2.36 mm and then homogenised. The chemical compositions of these samples were analysed at Newcrest Analytical Laboratory, Orange, Australia by Fire Assay for gold, X-ray fluorescence (XRF) for Fe and S and Inductively Coupled Plasma Mass Spectroscopy (ICP-MS) for the other elements. The Au and Ag grade of the ore were 2.13 ppm and 35.1 ppm, respectively. It also contained 3.87% Fe, 1.61% S, 86 ppm Cu, 147 ppm As and 244 ppm Zn. 3.2. Cyanide preparation The chemicals, copper(I) cyanide (CuCN, 99.99%) and sodium cyanide (NaCN, 99.9%), were obtained from Aldrich. NaCN was used to study the effect of free cyanide on flotation with a cyanide concentration of 10 ppm (equals to 0.3846 mM) in flotation pulp and 100 ppm (equals to 3.846 mM) in the electrolyte solution for electrochemical studies. A stock solution of 0.1 M NaCN was prepared with 18 MX cm deionized water. The pH of the stock

Fig. 4. Concentration dependence of copper cyanide speciation at [CN]/[Cu] = 3/1, pH = 10, Eh = 300 mV (SHE), 25 °C.

B. Guo et al. / Minerals Engineering 71 (2015) 194–204

solution was stabilized at 11.5 using standardised 1 M sodium hydroxide (NaOH) to avoid any HCN loss by volatilisation. Cuprous cyanide solution was prepared by dissolving CuCN powders in excess of NaCN solution to give a final solution with CN/Cu = 3. The NaCN solution was purged with nitrogen gas of high purity to remove O2 before and during the addition of CuCN powders because relatively unstable Cu(CN) 2 may form when the cyanide to copper ratio locally turns to be less than 3 during the addition of CuCN powders. Free cyanide or cuprous cyanide solution was added in flotation pulp with the same cyanide concentration. 3.3. Grinding and flotation The grinding and flotation were conducted at room temperature using Brisbane tap water. 1 kg crushed ore sample at 60% solid density with 10 Kg laboratory stainless steel rods was ground in a tumbling mill to obtain a 80 wt.% passing size (P80) of 150 lm at determined grinding time. Stainless steel was used to avoid iron contamination and the formation of ferricyanide after the addition of cyanide, which may bring additional interference for studying the impact of other cyanide species in flotation. Before grinding, the grinding mill and rods were cleaned by grinding with sands. The mill discharge was transferred to a 2.5 L mechanical flotation cell with an agitation speed of 800 rpm, and the pH of the flotation pulp was modified to 10 before conditioning with any reagents. Then 20 g/t promoter Aerophine 3418A (sodium diisobutyldithiophosphinate) was added at the beginning of flotation and 30 g/t collector PAX (potassium amyl xanthate) was added before the third stage of flotation (after cumulative flotation time of 3 min). 30 g/t Nasfroth 250 as frother was split between the first and third stage of flotation at a ratio of 80/20. Four concentrates were collected at cumulative times of 1, 3, 5, and 10 min with an airflow rate of 6 L/ min. The flotation froth was scraped at intervals of 10 s. 3.4. Mineralogy analysis Samples for Mineralogy Liberation Analyser (MLA) analysis were mounted in epoxy resin and polished to give a high quality finish and then carbon coated. In the MLA measurement, extended BSE liberation analysis (XBSE) mode was used to obtain bulk modal mineralogy information. Mineral discrimination is based on BSE grey level contrast. XBSE implements area X-ray analysis to efficiently analyse ore samples containing phases with sufficient BSE contrast to ensure effective segmentation. Sparse phase liberation analysis (SPL) mode was used to obtain gold and silver minerals association information. SPL searches BSE images for particles containing phases of interest using a BSE grey scale range and then performs an XBSE analysis on them. It does not provide bulk mineralogy information, as only selected particles in the sample are analysed. 3.5. Electrochemical measurements A hand-picked natural massive cubic pyrite specimen of a high purity mineral originated from Spain was used to make a working electrode. The electrode was connected with a copper wire using silver loaded conducting epoxy, and then mounted into a non-conducting epoxy resin, exposing only one side with a geometric surface of approximately 0.25 cm2 and this value was used to calculate the current density. A fresh electrode surface was prepared between the experimental runs by abrading with silicon carbide abrasive paper (1200 grits), rinsed with deionized water prior to each experiment. The electrode was then immediately transferred into the electrochemical cell as a stationary working electrode.

197

Conventional three-electrode system was employed for the electrochemical measurements. A platinum plate with surface area of 1 cm2 was utilised as an auxiliary electrode (counter electrode). Potentials were measured against an Ag/AgCl reference electrode filled with 3 M KCl which has a potential of +0.204 V against a standard hydrogen electrode (SHE). A Radiometer PGZ100 potentiostat was used in combination with a Frequency Response Analyser (FRA). The background electrolyte performed in all electrochemistry experiments was pH 10 buffered solutions (prepared by mixing 100 mL 0.025 M Na2B4O710H2O and 36.6 mL 0.1 M NaOH). Open circuit potential (OCP), cyclic voltammetry and electrochemical impedance spectroscopy (EIS) were performed in the above buffered solutions with a volume of 200 mL at room temperature in the absence and presence of 1 mM PAX and 3.846 mM cyanide species individually and in combination. Deionized water (18 MX cm) was used in all electrochemical experiments. The potential scan rate was 20 mV/s for all cyclic voltammetry experiments. The variations in the current were recorded as a function of scan potential and reported as current density. Open circuit potential was taken as the direct current (DC) voltage (starting potential) of EIS measurements. The alternating current (AC) voltage (amplitude) was 10 mV. The initial and final frequencies were 102 Hz and 104 Hz, respectively, with a frequency per decade as 5. Typically, the working electrode surface was allowed to react with the addition of chemicals for 5 min to enhance stabilisation at the open circuit potential after which the EIS was obtained. All measurements were performed in triplicate.

4. Results and discussion 4.1. Flotation Hidden Valley primary ore was floated during a flotation time of 10 min with the mass recovery of flotation concentrate reaching 6.8%. Fig. 5 shows the grade-recovery plots of flotation concentrate for the four elements Au, Ag, Fe, and S. At the baseline flotation (in the absence of any cyanide species), Au displayed good floatability reaching a recovery of 93.7% and a concentrate grade of 20 ppm at the completion of 10 min flotation; while at the same time Ag, Fe and S recoveries reached 91.3%, 32.0%, and 95.8%, respectively. The plant operations normally float pyrite at natural pH. According to the above thermodynamic analysis, pH 10 has to be used for the purpose of isolating individual cyanide species in this study. Good flotation was achieved at pH 10 as indicated in Fig. 5 providing an ideal flotation baseline. It is worth noting that after PAX was added at the third stage of flotation, iron sulphide minerals were almost effectively floated with S recovery increased from 40.7% to 93.7%. Obviously, PAX displayed strong collecting power for iron sulphide flotation in addition to the flotation by 3418A in the first and second stages. The flotation performance of Hidden Valley primary ore in the presence of 10 ppm free cyanide is also included in Fig. 5. Au recovery of 80.7% was obtained at the end of flotation, which is obviously lower than its recovery in baseline flotation. Ag recovery had a similar dependence on free cyanide with 82.0% at the end of flotation. On the other hand, the addition of cyanide had little influence on the grade of Au and Ag in flotation concentrate, partially related to the decrease of mass recovery to 4.2% in the presence of free cyanide. The Fe and S recoveries were decreased to 5.7% and 4.2%, respectively, indicating a strong depressive effect of free cyanide on the flotation of iron sulphide minerals. It seems that the depression of Au and Ag was associated with the depression of iron sulphide minerals. Around 13% of the Au and 9.3% of the Ag may be depressed by free cyanide as a result of their associations with iron sulphide minerals. It is well known that cyanide

198

B. Guo et al. / Minerals Engineering 71 (2015) 194–204

Fig. 5. Grade-recovery plots of flotation concentrate for Au, Ag, Fe, and S in the absence and presence of 10 ppm cyanide in the form of free cyanide (NaCN) and cuprous cyanide (CN/Cu = 3/1).

has a strong depressive effect on iron sulphide minerals including pyrite, pyhrrotite, and arsenopyrite. The iron sulphide minerals carrying gold and silver were examined by mineralogy analysis in the following section. Furthermore, free cyanide concentrations at 30 ppm and 50 ppm were also tested but the recoveries showed no further decrease. Fig. 5 also shows the grade-recovery plots of flotation concentrate for Hidden Valley primary ore in the presence of 10 ppm cyanide in the form of cuprous cyanide. It is seen that cuprous cyanide species had a similar depression effect on Au, Ag and their bearing iron sulphide minerals as free cyanide, resulting in recoveries of 5.4% Fe, 4.7% S, 81.2% Au and 81.8% Ag at the end of flotation. Meanwhile, similar mass recoveries and concentrate grades of these elements were obtained with free cyanide and cuprous cyanide. 4.2. Mineralogy analysis Mineralogical liberation analysis was conducted on concentrates from both baseline flotation and flotation in the presence of free cyanide. In baseline flotation, 50.2% of the concentrate mass originated from sulphide minerals, mainly pyrite (FeS2, 48.6%) with a very small amount of chalcopyrite (CuFeS2, 0.3%), sphalerite ((Zn,Fe)S, 0.6%), and galena (PbS, 0.7%) (Fig. 6(a)). The rest of the concentrate was non-sulphide minerals such as quartz, feldspar, clay minerals and other silicates. The S content 1.61% given in the feed represented total sulphur. However, it is indicated from Fig. 6 that all of the sulphur in the flotation concentrate were sulphide sulphur, while most were with pyrite. Therefore, it is deduced from the total S recovery of 93.7% that pyrite recovery of more than 90% was achieved in the baseline flotation.

When free cyanide was added in flotation, pyrite content in the flotation concentrate was decreased significantly to only 5.7%, while more than 90% of the concentrate consisted of non-sulphides (Fig. 6(b)). This is consistent with Fe and S grades of the concentrate assayed by the XRF. In the baseline flotation, Fe and S grades were 24.3% and 25.2%, respectively, whilst Fe and S grades in the flotation concentrate were decreased to 5.2% and 1.5%, respectively in the presence of free cyanide (Fig. 5). Similarly, it is also deduced from a total S recovery of 4.7% that pyrite recovery of less than 5% was obtained in the presence of free cyanide. These mineralogical analyses together with the flotation performance demonstrate that free cyanide has a strong depressive effect on pyrite flotation. The association of gold and silver with other minerals in flotation concentrates are shown in Figs. 7 and 8, respectively. From Fig. 7, 68% of the floated gold was liberated with free surface in baseline flotation concentrate, but it was increased to 85.3% in the flotation concentrate when free cyanide was present. Although the accuracy of detecting gold associations was not reliable in this study due to a very small amount of gold particles detected, it is suggested that some of the unliberated gold particles were depressed by free cyanide in flotation. On the other hand, silver association results were more reliable because a large amount of silver mineral particles were detected. Fig. 8 clearly shows that 63.1% of the silver minerals were liberated with free surface, and 16.9% were associated with pyrite in baseline flotation concentrate, whilst the silver minerals associated with pyrite were decreased to 9.3% in the flotation concentrate when free cyanide was added, indicating that silver-bearing pyrite flotation was depressed by free cyanide. Since free cyanide and cuprous cyanide showed the similar flotation behaviour in terms of Fe and S recoveries, it is con-

B. Guo et al. / Minerals Engineering 71 (2015) 194–204

199

Fig. 6. Mineralogical composition of flotation concentrates: (a) baseline flotation without any cyanide and (b) flotation in the presence of 10 ppm free cyanide.

Fig. 7. Gold associations in flotation concentrate: (a) baseline flotation without cyanide and (b) flotation in the presence of 10 ppm free cyanide.

Fig. 8. Silver associations in flotation concentrate: (a) baseline flotation without cyanide and (b) flotation in the presence of 10 ppm free cyanide.

cluded that cuprous cyanide species, mainly in the form of Cu(CN)2 3 , also depress the gold and silver carrying pyrite. 4.3. Electrochemical analysis The above flotation tests indicated the depressive impact of free cyanide and cuprous cyanide species upon the floatability of minerals, with xanthate as collector. As a significant portion of the precious metals is associated with pyrite, the flotation strategy is aimed at maximising the recovery of pyrite. Electrochemical

processes are of particular importance in sulphide systems as a result of their semi-conductive properties. In the following sections, open circuit potential (rest potential), cyclic voltammetry, and impedance spectroscopy were performed to reveal the depressive mechanism of cyanide species via the investigation on their electrochemical responses to pyrite. 4.3.1. Open circuit potential Open circuit potential (OCP) measurements were initiated to characterise the state of pyrite surface oxidation under open circuit

200

B. Guo et al. / Minerals Engineering 71 (2015) 194–204

conditions. The measured potential is a mixed potential which shows the equilibrium between the cathodic reduction and anodic oxidation processes in pyrite–xanthate–cyanide systems. The OCP of pyrite electrode in xanthate and/or cyanide solution with borate background at pH 10 is presented in Fig. 9 with error bars given based on triplicated measurements. No attempt was made to deoxygenate the solution in these measurements. In the background solution, the OCP of pyrite was 25 mV, but it was decreased to 100 mV in the presence of 3.846 mM free cyanide and 94 mV in the presence of 3.846 mM cuprous cyanide, respectively. The inhibition of anodic reactions tends to increase the mixed potential, while the inhibition of cathodic reactions tends to decrease the mixed potential. The observed decrease of pyrite potential is therefore indicative of strong inhibiting effect which cyanide species have on the cathodic oxygen reduction reaction. Xanthate is also a reducing reagent which donates electrons to the mineral surface on which it chemisorbs and ends up as metal xanthate complex or as dixanthogen (X2). Therefore, free cyanide or cuprous cyanide in combination with PAX has a stronger inhibiting effect on the surface reactions. As shown in Fig. 9, the addition of 3.846 mM free cyanide to 1 mM PAX solution decreased the pyrite potential to 125 mV and the addition of the same amount of cuprous cyanide to PAX solution decreased the pyrite potential to 121 mV. The pulp potential Eh was also measured against Ag/AgCl (3 M KCl) reference electrode in flotation. The reduction of Eh was noticeable, from 157 mV in the baseline flotation to 51 mV with the addition of 10 ppm free cyanide. The addition of 10 ppm cuprous cyanide also led to a decrease of Eh to 67 mV. Therefore, the reducing condition caused by cyanide species was identified. In fact, it is always observed that the decrease of potential co-exists with the depression of pyrite flotation (Kocabag and Guler, 2007; Wet et al., 1997). In the study by Kocabag and Guler (2007) using singe mineral flotation, there was a critical pyrite potential of about 100 mV (SHE) below which pyrite flotation was strongly depressed. It was also observed for a gold ore that the pyrite potential was decreased from 195 mV to 125 mV (SHE) with pyrite flotation being almost completely depressed (Wet et al., 1997). However, the critical potential above which pyrite floats can only apply to oxygenated atmosphere that used in the present study. Hydrophobic pyrite surface could be established at low potentials and low pH in a nitrogen atmosphere (Miller et al., 2002). 4.3.2. Cyclic voltammetry The surface reactions of pyrite exposed to borate background solution were investigated using cyclic voltammetry during which the solutions were purged with high purity nitrogen for 30 min to

Fig. 9. Open circuit potentials of pyrite electrode exposed to various solutions based on pH 10 borate buffer solution.

remove dissolved oxygen prior to each experiment, and nitrogen purging was maintained above the solution throughout the experiment to prevent the possible return of oxygen to the system. Measurements were performed at different anodic switching potential in order to identify the anodic peaks and their corresponding cathodic peaks from which the underpinning electrochemical reactions could be determined. The measurements were performed in triplicate, and the CV curves among these measurements were highly consistent. The results of the first curve are illustrated in Fig. 10. Electrochemical reactions corresponding to anodic and cathodic peaks in Fig. 10 are shown below.

A1 : Fe þ 2H2 O ¼ FeðOHÞ2 þ 2Hþ þ 2e

ð8Þ

A2 : FeðOHÞ2 þ H2 O ¼ FeðOHÞ3 þ Hþ þ e

ð9Þ

A3 : FeS2 þ 3H2 O ¼ FeðOHÞ3 þ 2S0 þ 3Hþ þ 3e

ð10Þ

þ  A4 : FeS2 þ 11H2 O ¼ FeðOHÞ3 þ 2SO2 4 þ 19H þ 15e

ð11Þ

C1 : S0 þ Hþ þ 2e ¼ HS

ð12Þ

The correlation of pyrite voltammograms with surface reactions in alkaline solution has been frequently reported (Ahlberg et al., 1990; Miller et al., 2002). In the first scan from 1200 to 600 mV, there were no visible anodic or cathodic peaks. However, an anodic peak (A3) commenced in the second scan when the upper limit was increased to 200 mV. This peak was attributed to the oxidation of pyrite resulting in the formation of ferric hydroxide and a sulphur-rich sub-layer (elemental sulphur or iron deficient sulphide layer) (Buckley et al. 1985). On the return scan, a cathodic peak C1 appeared with the starting potential of about 400 mV. This peak corresponds to the reduction of elemental sulphur to hydrosulphide ions (HS). A second cathodic peak C2 corresponds to the reduction of ferric hydroxide to ferrous hydroxide commencing at a more negative potential when the anodic switching potential of higher than 0 mV was applied. However, the peak C1 and C2 were not well separated when more positive anodic switching potential was applied. Anodic peak (A4) appeared with the starting potential of about 200 mV due to the further oxidation of pyrite to ferric hydroxide and sulphate as shown in Eq. (11). This reaction was identified to be much more rapid and electrochemically irreversible (Wang and Forssberg, 1996). Consequently, the current density of peak C2 resulting from the reduction of ferric hydroxide to ferrous hydroxide became more negative due to this extensive oxidation of pyrite. At a more negative potential ferrous hydroxide can be reduced to elemental iron at a third cathodic peak (C3). A

Fig. 10. Cyclic voltammetry of pyrite electrode in pH 10 borate buffer solution with various anodic switching potentials.

B. Guo et al. / Minerals Engineering 71 (2015) 194–204

rather large and sharp anodic peak (A1) appeared due to the oxidation of elemental iron as shown in Eq. (8). Iron(II) sulphide may form in the presence of hydrosulphide ions. Consequently peaks A3 and A4 also correspond to the oxidation of iron(II) sulphide forming at A1 together with pyrite itself. There was a general increase of current density for C3 and A1 with the more extensive oxidation of pyrite. When a more positive potential (e.g., 400 mV) was applied, not all of the ferrous species were involved in the formation of iron(II) sulphide. Excess ferrous species in the form of ferrous hydroxide remained at the surface and oxidised to ferric hydroxide at peak A2 on the scan towards anodic potential. The sweeps in Fig. 10 were initiated in the negative direction at potential of 300 mV which was a little more negative than the open circuit potential. The first cycle reflects a freshly polished surface. The appearance of peak C1 in the first cycle indicates the formation of sulphur-rich sub-layer during electrode handling and preparation. A potential window from 800 mV to 300 mV was selected to investigate the electrochemical behaviour of pyrite in xanthate and/or cyanide solution. The results using a second scan are illustrated in Fig. 11. For pyrite electrode in borate background solution as shown in Fig. 11(a), the peak C1 on the cathodic scan corresponding to the reduction of elemental sulphur to hydrosulphide ions (HS) was also evident. However, peak A1 on the anodic scan also contained a contribution from the oxidation of hydrosulphide to elemental sulphur and its deposition on pyrite surface, represented by the reversal of Eq. (12). When pyrite was treated with PAX, a reaction occurred and resulted in a peak A2 on the anodic scan corresponding to the oxidation of xanthate to dixanthogen in Eq. (13).

2X ¼ X2 þ 2e

ð13Þ

The reversible potential for xanthate oxidation to dixanthogen can be computed from the Nernst equation, as follows:

E ¼ E0  0:059log½X 

ð14Þ

E0 = 0.158 V (SHE) for the amyl xanthate/dixanthogen couple (Winter and Woods, 1973). The reduction of dixanthogen to xanthate in the solution was also evident due to the increased cathodic current density at peak C1 where the reduction of dixanthogen and elemental sulphur overlapped. After the addition of 3.846 mM free cyanide to PAX solution as shown in Fig. 11(b), the peaks corresponding to the formation of elemental sulphur and dixanthogen disappeared. Thus free cyanide inhibited all electrochemical reactions on freshly abraded pyrite. It was also revealed by Janetski et al. (1977) that the oxidation of xanthate shifted to more anodic position and the anodic curve

201

for pyrite oxidation showed little change with the increase of free cyanide concentration. It seems that the cyanide concentration in the present study is sufficiently high, resulting in the merging of xanthate oxidation peak with the background current due to the anodic reactions of the mineral itself. In the presence of 3.846 mM cuprous cyanide, the anodic and cathodic reactions were also largely inhibited and peak A2 for the oxidation of xanthate disappeared as shown in Fig. 11(b). However, the anodic oxidation to elemental sulphur still occurred, but less noticeably compared to that in background solution. The difference in the electrochemical reactivity of pyrite in the absence and presence of cyanide species contributed to different pyrite flotation behaviour. The elemental sulphur on mineral surface provides hydrophobicity and therefore collector-less floatability for pyrite (Chander and Briceno, 1987). Thus the inhibition of elemental sulphur formation by free cyanide prevents the collector-less flotation of pyrite. On the other hand, the traditional theory on xanthate-induced flotation of pyrite considers xanthate adsorption as an electrochemical process which involves the formation of hydrophobic dixanthogen (Allison et al., 1972; Haung and Miller, 1978; Majima and Takeda, 1968; Ralston, 1991; Valdivieso et al., 2005; Wood, 1984). One exception is that pyrite floatation occurs at low potentials and low pH values in the nitrogen atmosphere due to formation of ferric xanthate species instead of dixanthogen on pyrite surface (Miller et al., 2002). As indicated above in the oxygenated atmosphere, the formation of dixanthogen was completely inhibited by either free cyanide or cuprous cyanide resulting in poor flotation performance. Considering the difference between free cyanide and cuprous cyanide on the formation of elemental sulphur but with similar flotation recoveries obtained, collector-less flotation might not play a significant role in this study. Wang and Forssberg (1996) showed an enhanced oxidation peak at 150 mV (SCE) on the voltammograms of pyrite electrode in 6 mM NaCN solution at pH 10.5. This peak was attributed to the adsorption of cyanide onto pyrite and the formation of soluble cyanoferric species. However this was not observed in the present study, probably due to the use of borate reacting with pyrite and overlapping the interaction between cyanide and pyrite (Wang, 1996). 4.3.3. EIS measurements Impedance measurements were conducted to give an in-situ detection of possible modification of pyrite surface exposed to cyanide solutions. The impedance (Z) consists of real (Zr) and imaginary (Zi) parts that can be calculated by the means of phase angle. As shown in Fig. 12, the impedance spectra are commonly

Fig. 11. Cyclic voltammetry of pyrite electrode exposed to various solutions at pH 10 made by borate buffer solution: (a) effect of PAX; (b) effect of cyanide species.

202

B. Guo et al. / Minerals Engineering 71 (2015) 194–204

represented as either a Bode plot (with log frequency on the X-axis and both the absolute value of impedance and the phase-shift on the Y-axis) or a Nyquist plot which consists of the imaginary part of impedance (Y-axis) and the real part of impedance (X-axis). At higher frequency region (100–10,000 Hz), the Z values are low and relatively constant, while the phase angle values decrease towards zero. This is a typical response of a resistor to an AC with high frequency, corresponding to solution resistance. In the intermediate frequency region (0.1–100 Hz), the relationship between Z and frequency becomes linear with a slope of 1 and the phase angle reaches maximum within this region. This part corresponds to capacitive behaviour caused by the electrical double layer at the mineral/solution interface and/or the possible surface layer. The impedance spectrum can be modelled by an electrical circuit with physical elements. A number of different equivalent circuits may be fitted to impedance data for pyrite electrode depending on electrode preparation and the solution chemistry (Lin and Say, 1999; Pang et al., 1990; Venter and Vermaak, 2008). In this study, based on the criteria of simplicity and electrochemical interpretation, the electrical circuit shown in Fig. 13 was used to model a simple charge-transfer process at the mineral/solution interface. In this model, a resistor Rs represents the solution resistance and another resistor Rct is resistive component of the mass transport impedance at the interface. The capacitance due to double layer charging and/or possible surface layers is represented by a constant phase element (CPE) in place of an ideal capacitor, since a CPE can provide a useful modelling element containing various disturbances due to the physical nature of electrode surface and reactions (surface roughness, ‘‘leaky’’ capacitor, non-uniform current distribution, etc.).

Fig. 13. Electrical circuit that models impedance in Fig. 12.

The main disturbance, surface roughness of the electrode surface is indicated by a factor n, usually varying between 0.5 and 1. When n = 1, a CPE is equivalent to an ideal capacitor. The formation and growth of a surface layer such as dixanthogen causes a decrease in the electrode capacitance. Since the impedance of a capacitor is inversely proportional to the capacitance, a decrease in the electrode capacitance can be conveniently observed in Bode plots with an increase in the impedance values in the intermediate frequency. Therefore, the kinetics of surface electrochemical reactions and the formation of surface layers on pyrite can be detected. The Bode plots and Nyquist plots for pyrite in the presence of PAX and cyanide species individually or in combination were compared with that in the absence of these reagents in Fig. 12. Good reproducibility of the electrode pre-treatment and electrolyte preparation was confirmed by carrying out measurements in separate solutions and with freshly abraded electrode. The experimental data were fitted well with the proposed equivalent electrical circuit using computer program Zview. The extracted model parameters from the equivalent circuit are given in Table 1. The solution resistance, Rs, was similar in all of the experiments, since the additions of small amounts of xanthate and cyanide species did not further increase the already high solution conductivity.

Fig. 12. EIS of pyrite electrode exposed to various solutions at pH 10 made by borate buffer solution: (a) and (b) Bode plots; (c) and (d) Nyquist plots; (a) and (c) effect of cyanide species; (b) and (d) effect of PAX in the absence and presence of cyanide.

203

B. Guo et al. / Minerals Engineering 71 (2015) 194–204 Table 1 Elements of the equivalent circuit simulated from the EIS in Fig. 12.

Rs, X CPE, 105 S ⁄ sn n Rct, kX

Borate

PAX

NaCN + PAX

NaCN

[Cu(CN3)]2 + PAX

[Cu(CN3)]2

46 48.8 0.90 3.9

45.1 40.0 0.88 5.1

45 49.6 0.91 28.5

44.3 48.9 0.89 29.8

45.2 51.1 0.88 6.9

46.1 50.2 0.89 17.2

The addition of xanthate into the background solution resulted in lowering the capacitance from 48.8  105 to 40.0  105 S ⁄ sn. The higher impedance values in the intermediate frequency region of Bode plots in Fig. 12(b) also reflected a smaller electrode capacitance, presumably due to the formation of a continuous surface species on pyrite. The formation of a surface layer consisting mostly of dixanthogen as a result of xanthate adsorption was proposed previously using impedance techniques (Venter and Vermaak, 2008). Dixanthogen is formed by anodic oxidation of xanthate at pyrite surface coupled with cathodic reduction of adsorbed oxygen or surface ferric hydroxide (Haung and Miller, 1978; Janetski et al., 1977; Valdivieso et al., 2005). In 1 mM amyl xanthate solution, a hydrophobic pyrite surface was observed at a potential more than 0.019 V (SHE) corresponding to the reversible potential for the amyl xanthate/diamyl dixanthogen couple (Miller et al., 2006). This is in agreement with the cyclic voltammetry in the present study. The impedance in the present study was measured at an open circuit potential (65 mV vs. Ag/AgCl, 3 M KCl) which is well above the reversible potential of xanthate–dixanthogen couple. Thus the formation of dixanthogen was expected to be responsible for the decrease in capacitance. It was also presented by Leppinen (1990) using surface analyses that dixanthogen as multilayers covered on monolayer ferric ethyl xanthate (Fe(EX)3), depending on the collector concentration. The conclusion drawn by Miller et al. (2002) and Du Plessis (2004) indicated that dixanthogen formation was found above a certain potential, while ferric xanthate species was responsible for creating the hydrophobic pyrite surface state at a low potential in the absence of oxygen. Therefore, the surface xanthate adsorption layer indicated in the present study might also be attributed to the formation of ferric xanthate compounds. The adsorption of xanthate species is in line with its high recoveries of pyrite in the flotation test work. With the addition of free cyanide, the impedance values of pyrite electrode were significantly higher in the low frequency range of Bode plots as shown in Fig. 12(a) and the diameter of the semicircle in Nyquist plots was also significantly increased as shown in Fig. 12(c). These changes were reflected by the increase of Rct from 3.9 kX to 29.8 kX as shown in Table 1. The higher value of charge transfer resistance indicated that free cyanide appeared to act as an inhibitor for the electrochemical reactions on the pyrite surface, which is in line with the cyclic voltammetry observations. No change in the impedance values in the intermediate frequency range of Bode plots and of course no significant change in the capacitance indicated no continuous surface layers forming resulting from the interaction of pyrite with free cyanide. With the addition of cuprous cyanide, the impedance increase in the low frequency range of Bode plots and a larger semicircle in Nyquist plots were also observed in Fig. 12(a) and (c), respectively. One would still expect a clean pyrite electrode surface interacting with cuprous cyanide species with the surface electrochemical reactions inhibited on the basis of almost constant capacitance values and increased charge transfer resistance values compared to the background solution. It is worth noting that the action of cuprous cyanide as a barrier for surface electrochemical reactions was less pronounced compared to free cyanide, which can be deduced from smaller Rct value (17.2 kX) in the presence of cuprous cyanide

(Table 1). This is also in line with the observation in cyclic voltammetry study. The simultaneous presence of both PAX and cyanide species and their resulting impedance spectra in Fig. 12(b) and (d) suggest that both free cyanide and cuprous cyanide at the concentration studied completely inhibited the interaction between PAX and pyrite, but did not adsorb onto the mineral surface, since the capacitance went back to a similar value in background solution with the addition of cyanide species to PAX containing solution. It is apparent that in the presence of cyanide, the hydrophobic xanthate species are unstable and decompose from pyrite surface by forming much more stable cyanoferrous or cyanoferric complexes which might be soluble. The above electrochemical measurements indicate that both the inhibition of surface reactions (surface cleaning effect) and the detachment of surface xanthate species contribute to the depressive effect on pyrite flotation. Free cyanide and cuprous cyanide play a similar role in terms of inhibiting xanthate oxidation and adsorption, while the surface cleaning effect of cuprous cyanide is relatively weaker than free cyanide. However, cuprous cyanide species depress pyrite flotation at a similar efficiency as free cyanide. 5. Conclusion Gold and silver flotation at alkaline pH was depressed through their association with pyrite by either free cyanide or cuprous cyanide. The predominant cuprous cyanide species depressing pyrite flotation was Cu(CN)2 3 on a basis of thermodynamic calculation. The surface electrochemical reactions, especially the xanthate oxidation on pyrite surface were inhibited by either free cyanide or cuprous cyanide based on electrochemical studies on a pyrite electrode. The changes in the impedance characteristics of the pyrite electrode were consistent with the detachment of adsorbed xanthate species in the presence of these cyanide species, which was suggested to contribute to the hydrophilic pyrite surface and lower recovery in flotation. Acknowledgements The authors gratefully acknowledge the financial support of this work from Morobe Mining Joint Venture. The first author also thanks the scholarship provided by the University of Queensland. References Adams, M.D., 1990. The chemical behavior of cyanide in the extraction of gold Part 1: kinetics of cyanide loss in the presence of absence of activated carbon. J. South African Inst. Min. Mettall. 90 (2), 37–44. Adams, M.D., 2013. Impact of recycling cyanide and its reaction products on upstream unit operations. Miner. Eng. 53, 241–255. Ahlberg, E., Forssberg, K.S.E., Wang, X., 1990. The surface oxidation of pyrite in alkaline solution. J. Appl. Electrochem. 20, 1033–1039. Allison, S.A., Gold, L.A., Nicol, N.G., Granville, A., 1972. A determination of the product of reaction between various sulfide minerals and aqueous xanthate solutions, and a correlation of the products with electrode rest potentials. Metall. Trans. B, Process Metall. 3 (10), 2613–2618. Buckley, A.N., Hamilton, I.C., Woods, R., 1985. Investigation of the surface oxidation of sulphide minerals by linear potential voltammetry and X-ray photoelectron

204

B. Guo et al. / Minerals Engineering 71 (2015) 194–204

spectroscopy. In: Forssberg, K.S.E. (Ed.), Flotation of Sulphide Minerals, vol. 4. Elsevier, Amsterdam, pp. 41–60. Buckley, A.N., Wood, R., Wouterlood, H.J., 1989. An XPS investigation of the surface of natural sphalerite under flotation related conditions. Int. J. Miner. Process. 26 (1–2), 29–49. Castro, S., Larrondo, J., 1981. An electrochemical study of depression of flotation of chalcocite by cyanide and iron-cyanide complex. J. Electroanal. Chem. 118, 317– 326. Chander, S., Briceno, A., 1987. Kinetics of pyrite oxidation. Miner. Metall. Process. 4, 171–176. Chandra, A.P., Gerson, A.R., 2009. A review of the fundamental studies of the copper activation mechanisms for selective flotation of the sulfide minerals, sphalerite and pyrite. Adv. Colloid Interface Sci. 145 (1–4), 97–110. Dai, X., Simons, A., Breuer, P., 2012. A review of copper cyanide recovery technologies for the cyanidation of copper containing gold ores. Miner. Eng. 25 (1), 1–13. Dai, X., Breuer, P., 2009. Cyanide and copper cyanide recovery by activated carbon. Miner. Eng. 22, 469–476. Du Plessis, R., 2004. The thiocarbonate flotation chemistry of auriferous pyrite. PhD Dissertation, The University of Utah. Elgillani, D.A., Fuerstenau, M.C., 1968. Mechanisms involved in cyanide depression of pyrite. Trans. Soc. Min. Eng. 241, 437–445. Finkelstein, N.P., 1997. The activation of sulphide minerals for flotation: a review. Int. J. Miner. Process. 52 (2–3), 81–120. Grano, S., Ralston, J., Smart, R.St.C., 1990. Influence of electrochemical environment on the flotation behaviour of Mt. Isa copper and lead-zinc ore. Int. J. Miner. Process. 30, 69–97. Guo, B., Peng, Y., Espinosa-Gomez, R., 2013. The effect of weak acid dissociable cyanide on gold and silver flotation in Hidden Valley Mines. In: World Gold 2013 Conference, Brisbane. Gustafsson, J.P., 2012. Visual MINTEQ, version 3.0. Compiled in Visual Basic NET2005, KTH, Department of Land and Water Resources Engineering, Stockholm, Sweden. Haung, H.H., Miller, J.D., 1978. Kinetics and thermochemistry of amyl xanthate adsorption by pyrite and marcasite. Int. J. Miner. Process. 5 (3), 241–266. Heath, A.R., Rumball, J.A., Browner, R.E., 1998. A method for measuring HCN(g) emission from CIP/CIL tanks. Miner. Eng. 11 (8), 749–761. Hodgkinson, G., Sandenberghs, R.F., Hunter, C.J., 1994. Pyrite flotation from gold leach residues. Miner. Eng. 7 (5–6), 691–698. Janetski, N.D., Woodburn, S.I., Woods, R., 1977. An electrochemical investigation of pyrite flotation and depression. Int. J. Miner. Process. 4, 227–239. Kocabag, D., Guler, T., 2007. Two-liquid flotation of sulphides: an electrochemical approach. Miner. Eng. 20, 1246–1254. Leppinen, J.O., 1990. FTIR and flotation investigation of the adsorption of ethyl xanthate on activated and non-activated sulfide minerals. Int. J. Miner. Process. 30 (3–4), 245–263. Lin, H.K., Say, W.C., 1999. Study of pyrite oxidation by cyclic voltammetric, impedance spectroscopic and potential step techniques. J. Appl. Electrochem. 29, 987–994. Lotter, N., 2006. Cyanide voltalisation from gold leaching operations and tailing storage facilities. Master Thesis, The University of Pretoria. Lu, J.M., Dreisinger, D.B., Cooper, W.C., 2002. Thermodynamics of the aqueous copper–cyanide system. Hydrometallurgy 66 (1–3), 23–36. Lukey, G.C., Van Deventer, J.S.J., Huntington, S.T., Chowdhury, R.L., Shallcross, D.C., 1999. Raman study on the speciation of copper cyanide complexes in highly saline solutions. Hydrometallurgy 53 (3), 233–244.

Majima, H., Takeda, M., 1968. Electrochemical studies of the xanthate–dixanthogen system on pyrite. Trans. Am. Inst. Min., Metall. Petrol. Eng. 241, 431–436. Marsden, J.O., House, C.L., 2006. The chemistry of gold extraction, second ed. Society for Mining, Metallurgy, and Exploration, Inc., Littleton, pp. 233–270 and 486– 497. Miller, J.D., Kappes, R., Simmons, G.L., LeVier, K.M., 2006. Pyrite activation in amyl xanthate flotation with nitrogen. Miner. Eng. 19 (6–8), 659–665. Miller, J.D., Du Plessis, R., Kotylar, D.G., Zhu, X., Simmons, G.L., 2002. The lowpotential hydrophobic state of pyrite in amyl xanthate flotation with nitrogen. Int. J. Miner. Process. 67, 1–15. Pang, J., Briceno, A., Chander, S., 1990. A study of pyrite/solution interface by impedance spectroscopy. J. Electrochem. Soc. 137 (11), 3447–3455. Prestidge, C.A., Ralston, J., Smart, R.S.C., 1993. The competitive adsorption of cyanide and ethyl xanthate on pyrite and pyrrhotite surfaces. Int. J. Miner. Process. 38 (3–4), 205–233. Prestidge, C.A., Skinner, W.M., Ralston, J., Smart, R.S.C., 1997. Copper(II) activation and cyanide deactivation of zinc sulphide under mildly alkaline conditions. Appl. Surf. Sci. 108, 333–344. Ralston, J., 1991. Eh and its consequences in sulphide minerals flotation. Miner. Eng. 4 (7–11), 859–878. Seke, M.D., 2005. Selective flotation of galena and sphalerite at Rosh Pinah Mine. PhD dissertation, University of Pretoria, Pretoria. Seke, M.D., Pistorius, P.C., 2006. Effect of cuprous cyanide, dry and wet milling on the selective flotation of galena and sphalerite. Miner. Eng. 19 (1), 1–11. Smith, A., Mudder, T., 1991. The Chemistry and Treatment of Cyandation Wastes. Mining Journal Books Ltd., London. Sutherland, K.L., Wark, I.W., 1955. Principles of flotation. Australian Institute of Mining and Metallurgy (Inc.), Melbourne, pp. 113–149. Valdivieso, A.L., Sanchez, A.A.L., Song, S., 2005. On the cathodic reaction coupled with the oxidation of xanthates at the pyrite/aqueous solution interface. Int. J. Miner. Process. 77, 154–164. Venter, J.A., Vermaak, M.K.G., 2008. EIS measurements of dithiocarbonate and trithiocarbonate interactions with pyrite and copper. Miner. Eng. 21, 559–567. Vukcevic, S., 1996. A comparison of alkali and acid methods for the extraction of gold from low grade ores. Miner. Eng. 9 (10), 1033–1047. Wang, X.H., 1996. Interfacial electrochemistry of pyrite oxidation and flotation I: effect of borate on pyrite surface oxidation. J. Colloid Interface Sci. 178, 628– 637. Wang, X.H., Forssberg, K.S.E., 1996. The solution electrochemistry of sulfide– xanthate–cyanide systems in sulfide mineral flotation. Miner. Eng. 9 (5), 527– 546. Wark, I.W., 1938. Principles of flotation. Australian Institute of Mining and Metallurgy (Inc.), Melbourne, pp. 182–233. Weisener, C., Gerson, A., 2000. An investigation of the Cu(II) adsorption mechanism on pyrite by ARXPS and SIMS. Miner. Eng. 13 (13), 1329–1340. Wet, J.R.D., Pistorius, P.C., Sandenbergh, R.F., 1997. The influence of cyanide on pyrite flotation from gold leach residues with sodium isobutyl xanthate. Int. J. Miner. Process. 49 (3–4), 149–169. Winter, G., Woods, R., 1973. The relation of collector redox potential to flotation efficiency: monothiocarbonates. Separat. Sci. 8 (2), 261–267. Woods, R., 1984. Electrochemistry of sulphide flotation. In: Jones, M.H., Woodcock, J.T. (Eds.), Principles of Mineral Flotation, The Wark Symposium. The Australasian IMM, Victoria, Australia, pp. 91–115.