Table of Contents
Cyanide has been used in the mining industry for gold and silver recovery for almost a century. It is cost-effective and achieves excellent extractions from a wide range of ores. However, considerable interest has been expressed in non-cyanide lixiviants since the early 1980’s. In addition to environmental considerations, several factors have influenced the search for alternative lixiviants for gold.
A major drawback of the cyanidation process lies in its inherently slow kinetics. The treatment of refractory ores and presence of cyanicides (e.g. pyrrhotite, stibnite, etc.) present major challenges in processing these ore-types. Oxidized copper minerals, as well as chalcocite and bornite, dissolve in cyanide solutions, and in many cases, recovery of gold from fouled solutions becomes very difficult, if not impossible.
Thiosulfate leaching of gold and silver has been known since 1858. In the Patera process, which was used for silver recovery for many years (Lidell 1945), silver was extracted from silver chloride by sodium thiosulfate leaching of chloridizing-roasted ores and concentrates. It was also recognized that unreacted silver sulfide could be leached when copper sulfate was added to thiosulfate solutions (Russell, 1885). The leaching chemistry of this process was recently investigated in great detail (Flett et al. 1983).
Studies on the ammonium thiosulfate extraction of gold and silver from sulfide concentrates and various (ammoniacal) pressure leach residues (Berezowsky and Sefton, 1979), and dissolution of pure gold under pressure oxidation conditions in ammonium thiosulfate solutions (Tozawa et al., 1981), have shown that the chemistry of these systems is quite complex. Gold dissolution was strongly affected by temperature, copper and thiosulfate concentrations, and oxygen partial pressures or on the degree of copper oxidation.
Thiosulfate leaching of manganiferrous (Zipperian et al., 1988, Kerley, 1981), carbonaceous (Hemmati et al., 1989), and low grade ores (Langhans et al., 1992), and high grade copper concentrates (Changling et al., 1992) were also reported recently.
Gold from Copper-Bearing Ores
Gold is commonly associated with chalcopyrite; association with tetrahedrite, tennantite, as well as bornite and chalcocite also occurs. Unlike gold associated with pyrite and arsenopyrite in refractory ores, the gold particles in many chalcopyrite ores are often discrete, allowing a significant portion of gold to be liberated by fine grinding (Gasparrini, 1983). This association often allows a high grade copper-gold concentrate to be produced which is usually shipped to a smelter. Gold remaining in copper tailings can be recovered by conventional cyanidation.
Major processing difficulties arise with oxidized or transition type ores where copper sulfides are partially or wholly oxidized. These types of ores are not readily amenable to flotation concentration and are highly soluble in ammonia and alkaline cyanide solutions as well as acidic lixiviants (Hedley and Tabachnik, 1968, Shantz and Reich, 1978). Treatment options for such ores have recently been reviewed and the chemistry of leaching with CN-/NH3 and CN-/CO3²- mixtures have also been discussed in detail (Muir et al. 1989). It has been concluded that cyanide consumption can be reduced by ammonia addition; however, copper dissolution increases and gold recovery decreases as the ammonia to cyanide ratio increases.
Another option is the removal of oxidized copper minerals by acid leaching, and gold recovery by subsequent cyanidation. This option, however, is likely to be limited because of acid consuming gangue and neutralization costs preceding cyanidation (Dorr and Bosqui, 1950). Thiourea leaching from an oxidized gold-copper ore with 90-95% gold extraction was also reported (McInnes et al., 1989). However, acid pre-leaching of soluble copper was required in order to reduce thiourea consumption to acceptable levels.
Ammoniacal Thiosulfate Leaching
Ammonium thiosulfate has been used in photography for many decades. It is also used as fertilizer and from an environmental standpoint, has a definite advantage over cyanide. Dissolution rates of gold and silver in thiosulfate are comparable to those for thiourea and faster than cyanide. Even though the consumption of thiosulfate may be much higher than cyanide or thiourea, ammonium thiosulfate costs are a fraction of those reagents. The ability of thiosulfate to form stable complexes with copper ions, in addition to gold and silver, raises the possibility of using this reagent for leaching copper-bearing gold ores. Furthermore, when compared to thiourea leaching where large amounts of acid may be consumed by copper oxides, carbonates, and other gangue minerals, ammoniacal thiosulfate can be highly selective and potentially applicable to heap leaching or even in-situ leaching operations. However, due to the lack of established methods for the recovery of copper, gold, and silver from the leach solutions, thiosulfate has not received as much attention, for instance, as thiourea, as an alternative to cyanide.
Dissolution of gold and silver in ammoniacal thiosulfate solutions can be summarized as shown in Equations (1) and (2), respectively (Berezowski and Sefton, 1979).
2 Au + ½ O2 + 4 S2O3²- + H2O = 2 Au (S2O3) 2³- + 2 OH-
2 Ag + ½ O2 + 4 S2O3²- + H2O = 2 Ag(S2O3) 2³- + 2 OH-
Silver chloride is readily soluble in thiosulfate solutions, and silver sulfide can be solubilized in the presence of copper ions. In the presence of ammonia, respective ammonia complexes of gold and silver are stable at high pH. These equilibria were illustrated in the form Eh-pH diagrams for copper, gold and silver by Zipperian et al. (1989) for the ammonia-thiosulfate-water system. These diagrams show that gold and silver complexes are stable in solutions over the whole range of pH and potentials, whereas the stability domains of solid copper species expand as the copper concentration increases. Nevertheless, under most leaching conditions, i.e. Eh values of 200-400 mV and pH from 4 to 10, copper complexes are stable and should remain in the aqueous phase.
In the present study, rotating disk experiments were carried out to gain a better understanding of the interactive effects of thiosulfate, ammonia, and copper concentrations on gold dissolution.
Rotating disk experiments were run at constant temperature (25°C) and constant rotation speed (600 RPM). The effect of ammonia, thiosulfate and copper ion concentrations on gold dissolution were studied by analyzing aliquots of leach solution periodically. Since the ammonia and ammonium ions form a buffer solution, the solution pH could not be adjusted independently.
The recovery of gold and silver from a copper-bearing gold ore sample obtained from southern New Mexico was investigated using agitation leaching, bottle-roll and column leach techniques.
These experiments were carried out with a low grade ore analyzing 2.1 g/t Au (0.061 ± 0.008 oz/ton) , 56.7 g/t Ag (1.65 ± 0.12 oz/ton), and 0.86% acid soluble (oxide) copper (1.04% total). The ore sample was ground to 100% minus 100 mesh with repeated screening and grinding, and thoroughly homogenized. One hundred gram samples of this ore were leached with 300 ml leach solution in a 1 liter resin kettle at 25°C without additional copper. Initially 2, 4, 6, and 8 hour leach tests were performed and the residues were analyzed by fire assay. The recoveries were calculated from the difference in assay values. The maximum gold extraction (54%) was obtained at low concentrations of thiosulfate (0.1 M) within 4 to 6 hours. Therefore, subsequent ore leaching tests were run for 6 hours using the bottle-roll technique.
These experiments were conducted utilizing a three level, two parameter factorial design, varying free ammonia and thiosulfate concentrations concurrently. The results of these experiments were analyzed using a computer software for statistical analysis to ascertain the interactive effects among ammonia, thiosulfate and cupric ion concentrations.
Results and Discussion
The effect of varying the ammonia concentrations on gold dissolution from a rotating gold disk is shown in Figure 1 (a). Without free ammonia, no gold dissolution was detected. The gold dissolution rates increased as the ammonia concentrations were increased, and reached a maximum at 1.0 M free ammonia. Therefore, subsequent rotating disk experiments were carried out at 1.0 M free NH3 and varying the thiosulfate (0.13-0.66 M) and cupric ion (0.01-0 .1 M) concentrations.
Figure 1(b) shows that in the presence of 0.1 M cupric ions (6.35 g/l) and 1.0 M free ammonia concentrations, gold dissolution rates increased as the thiosulfate ion concentrations increased up to 0.5 M. The deep blue color of the solutions persisted, indicating the stability of the cupric-ammonia complex, and the solution pH remained fairly constant at about 10.1. Neither copper sulfide precipitation nor passivation at the surface of the gold disk was observed under these conditions.The dissolution rates decreased when the thiosulfate ion concentrations were further increased and the solutions became unstable.
Figure 1(c) shows the gold dissolution from a rotating disk as a function of copper concentration. The dissolution rate increased as the copper concentration was increased up to 0.04 M when the rate started decreasing. At or above 0.1 M copper ion concentrations the solution became very unstable. The azure blue color of the solution disappeared, and (black) copper sulfide precipitates formed. The chemical reactions involved in this phenomena can be summarized as follows:
Cupric ions are reduced by thiosulfate to the cuprous state while thiosulfate is oxidized to tetrathionate (Tykodi, 1990) , as shown in Equation (3).
2 S2O3²- + 2 Cu²+ = 2 Cu+ + S4O6²-
Cuprous ions are unstable in solutions and will precipitate as oxide (Cu2O), or thiosulfate (Cu2S2O3), but they can be stabilized by either ammonia or thiosulfate complexes depending on the thiosulfate/ammonia ratio. The overall equations for precipitation of Cu2S2O3 and subsequent complexation can be written (Flett et al., 1983) as shown in Equations (4) and (5).
3 (NH4)2S2O3 + 2CuSO4 = Cu2S2O3 + 2 (NH4)2SO4 + (NH4)2S4O6
5 (NH4)2S2O3 + 2CuSO4 = Cu2S2O3 · 2 (NH4) 2S2O3 + 2 (NH4)2SO4 + (NH4)2S2O4
In the presence of excess ammonia, cuprous ammonium complexes, which are colorless, may also form. Therefore, complete conversion of cupric ions to the cuprous-ammonium-thiosulfate complex (Equation 5) will require a molar ratio of thiosulfate to cupric ion equal to 2.5 to 1. It is also observed that the highest dissolution rate in Figure 1(c) at 0.1 M thiosulfate takes place at a cupric ion concentration of 0.04 M, and above these concentration levels the dissolution rate decreased.
The kinetics of cupric ion reduction by thiosulfate has been studied by Byerly et al. (1973). They have shown that, in purely aqueous solution, reduction of cupric ions (0.005 M) with thiosulfate (0.1 M) is very fast, and follows a pseudo first order kinetics. In the presence of ammonia (0.2 M NH3) however, the reaction is much slower and the rate is inversely dependent on the ammonia concentration.
Flett et al. (1983) in their study on the mechanism of the modified Patera process have concluded that in the absence of air, cuprous-ammonium-thiosulfate was the active species involved in the dissolution of Ag2S by an exchange mechanism, and the end product was cuprous sulfide (Cu2S), whereas in the presence of oxygen, cupric sulfide (CuS) was the end product.
Hiskey and Alturi (1988), suggested that the presence of excess ammonia stabilizes the cupric ions in the form of cupric ammonium complexes, as illustrated by Equation (6), and the gold dissolution reaction may proceed according to Equation (7).
2Cu (S2O3)2³- + 8NH3 + O2 + H2O = 2Cu (NH3)4²+ + 4S2O3²- + 2OH-
Au + 4S2O3²- + Cu(NH3)4²+ = Au (S2O3)2³- + 4NH3 + Cu(S2O3)2³-
These equations suggest that cupric ammine complex ion is an oxidant for gold dissolution, and the thiosulfate leaching of gold can proceed under oxygen deficit conditions which may have significant implications for heap leaching for copper bearing gold ores. However, if an excess of thiosulfate and ammonia, with respect to soluble copper species, can not be maintained in the leach solutions, insoluble copper and silver thiosulfate or sulfide compounds may precipitate. This phenomena will most likely cause gold losses due to gold occlusions in these precipitates or due to insufficient thiosulfate concentrations for gold dissolution.
The effect of temperature was also studied with a limited number of experiments. These experiments showed that, initially, gold dissolution increases with increasing temperature. A further increase in temperature causes rapid oxidation of thiosulfate, hence the dissolution rate decreases.
Due to the difficulties experienced in matching the standard matrices for atomic absorption analysis to the leach solutions aged at higher temperatures, these results were not quite reproducible. Nevertheless, large increases in the leach rates, as reported in the literature for leaching ore samples, were not observed with the rotating disk system. The calculated activation energies were in the range of 5-10 kCal/mole (21-42 kJ/mole).
The results of the factorial design experiments with 2 midpoint replicates are given in Table 1. The maximum recoveries obtained by 6 hour bottle-roll tests were obtained with high thiosulfate (0.5 M) and mid ammonia (0.25 M) levels. A regression analysis of the results indicated that thiosulfate terms shows the strongest effects on both gold and silver extractions, whereas ammonia-thiosulfate interactions are the least significant.
The regression equations for gold and silver extraction, given in equations (8) and (9) respectively, were utilized to plot the contour maps, shown in Figure 2.
% Au = – 2.315 + 290.411 (X1) + 118.389 (X2) – 286.222 (X1)² – 175. 111 (X2)² – 92.222 (X1) (X2)
r² = 0.839
% Ag = – 4.932 + 296.426(X1) + 61.241(X2) – 314.815(X1)² – 117.037(X2)² – 56.667(X1)(X2)
r² = 0.770
These figures show that at low thiosulfate concentrations, extraction of metals is independent of ammonia concentrations, whereas at high thiosulfate levels, both ammonia and thiosulfate have a strong influence. The maximum gold and silver extractions predicted by this model (i.e. 78.0% and 67.0%) are in close agreement with the experimental values (Table 1).
Copper dissolution was complete within 6 hours of leaching and it was expected that longer leaching times would improve the gold and silver recoveries. However, longer leaching test results showed that both gold and silver recoveries decreased drastically with time (Table 2).
A preliminary column leach test was conducted in a 5 cm diameter column with a 5 kg ore sample crushed to minus 6mm. Mid- levels of thiosulfate (0.35 M) and ammonia (0.25 M) concentrations were selected for this test and the leach solution was percolated through the column for 9 days. At the end of the test, the column was drained, washed and the tailings dried. Head and tailings samples were assayed, and the average recoveries for gold and silver were determined to be 40.0% and 55.3%, respectively. Due to the lack of reliable analytical techniques for the solutions, gold and silver dissolution during the column leach test could not be monitored. Hence, these results are inconclusive.
Copper dissolution was almost complete after two days and was not monitored thereafter. Even though the leach solutions were clear and the dark-blue color persisted, it is possible that degradation of thiosulfate and subsequent precipitation of copper compounds might have hindered the gold recovery. It appears that higher thiosulfate concentrations (i.e. 0.5 M) and a larger solution volume to keep the total dissolved copper concentration around 0.05 M would have been more appropriate for the column leach tests. Larger-scale experiments would also allow solution analysis by fire assaying thus enabling the progress of the leach tests to be monitored.
Summary and Conclusions
Although thiosulfate leaching of gold and silver has been known for over a century, recent studies have demonstrated the complexity of this system and the lack of understanding of the fundamental dissolution mechanisms. Preliminary results obtained in this study utilizing the rotating disk technique, have also confirmed the earlier conclusions: the catalytic action of copper ions, the necessity of maintaining thiosulfate and soluble copper above a certain molar ratio, and the lack of gold dissolution in the absence of free ammonia. The decrease in gold dissolution rate at high thiosulfate concentrations can be attributed to the rapid reduction of cupric ions to cuprous, similarly, reduction in dissolution rates at high ammonia concentrations (i.e. 4.0 M) may be due to reduced cupric ion activity through complexation.
Ore leaching experiments have also demonstrated the need for closely monitoring the reagent concentrations and the leaching times. Although the initial tests conducted with an oxidized copper ore yielded encouraging recoveries for gold and silver, longer leaching tests showed that solubilized values may reprecipitate in-situ and be lost to the tailings. Furthermore, routine analytical techniques were not readily applicable because of interferences and the degradation of the thiosulfate solutions. These aspects of thiosulfate leaching may hinder the potential for large-scale applications of this reagent.
This investigation has been supported by the Department of The Interior’s mineral institute program administrated by the Bureau of Mines under allotment grant number G1114135. We would also like to thank the Lordsburg Mining Company for providing the ore samples for this study.