The DeLamar Silver Mine is a Joint Venture between Earth Resources Company, the Canadian Superior Mining Company and the Superior Oil Company. The property is a consolidation of several old operations in the district. The more widely known are the DeLamar and Sommercamp Mines.

The location is 20 miles northeast of Jordan Valley, Oregon and 55 miles southwest of Boise, Idaho. The mine is on the south-slope of DeLamar Mountain in the Owyhee-Mountains of southwest-Idaho.

Initial interest started in 1969 with serious-exploration beginning in 1970. The feasibility study was completed in 1974, and in 1975 Mountain States Engineers were retained to design and build the plant. Processing of ore began in March of 1977 with the first bullion poured in April of that year.

The plant is currently processing some 2200 TPD of ore with a grade of four ounces of silver and 0.03 oz. of gold per ton of ore. Cyanide consumption is 2.5 lb/ton, zinc consumption of 0.3 lb/ton and flocculent consumption of 0.5 lb/ton. A flowsheet is shown as Figure I.

The grinding circuit is of two stage design. The primary mill is an 18 ft. dia. by 9 ft. long semi-autogenous unit. The secondary mill is a 9½ ft. dia. by 15 ft. long overflow ball mill. The mill discharges are combined and sent to 15 in. cyclones for classification. Cyclone overflow reports to leaching with the underflows reporting to the ball mill. The cyclone feed pump is an 8×6 metal case pump.

The primary mill has 1000 installed horsepower and rotates at 68% of critical speed. The ball mill has 700 installed horsepower and rotates at 72% of critical speed.

Because the primary mill is horsepower limited the ball charge is varied to maintain the desired throughput. The mill power is held at 980-990 HP by modifying the feed rate. Four inch forged steel balls are added any time the ore feed fails to reach 100 tons per hour for four hours. During periods of extremely hard ore, the ball charge has been as high as 14 v/o and at other times the mill has operated fully autogenous for weeks.

The main point being that due to extreme variability of DSM ore the ball ration changes daily. If the ore suddenly goes from very hard to very soft, the cyclone underflow is recycled to the primary mill to maintain a liner protecting load until the ball charge is ground down or the mine can supply some competent rock. This also reduces the load on the ball mill and prevents overloading the circuit. The other extreme case is a sudden change from hard to clay rich, soft ore. In this instance, all of the cyclone underflow is put to the ball mill and several tons of balls are charged to the primary mill. Obviously, during normal operations some intermediate combination of the above is used.

The ball mill ball ration is varied to maintain the static charge level two inches above the discharge trunnion line. This results in a ball mill power draw of 550 HP.

Ideally, one would feed an ore mixture such that fully autogenous grinding was possible at the desired feed rate. This luxury is not available at DSM. Between the extreme variability of the ores and the operating constraints of the weather, the mine has not been able to supply a blended ore feed. During the very dry and very cold seasons the mill feed is predominately high clay ores. During the other seasons, the feed stream is predominately rocky to facilitate materials handling problems due to mud. Even with these feed pertubations, it is necessary to shut down the plant for an hour each shift to clear the belt lines and transfer chutes of mud during the fall and spring seasons.

As a result of the high (+30%) clay content of DSM ores; pulps are quite viscous. This unusually high viscosity results in decreased throughput and impaired classification. A viscosity reducer is added to the primary mill and carries through the entire circuit. The most dramatic effect is to allow a higher cyclone overflow density with no change in classification, i.e., 38% solids vs. 28% solids. The increased density results in a significant increase in leach tank residence time. This is beneficial in that increased residence time helps compensate for high throughput surges from the grinding circuit. Some typical grinding circuit data are shown in Table I.

Over the past three years since starting the primary mill liners have gone through a steady evolution. The high wear areas have been identified and the liner sections modified to extend life. Presently, the grates are of pearlitic chrome-moly with a hardness of 321/375 BHN. The balance of the liners are of martensitic chrome-moly with hardness of 363/444 BHN.

The changes in liner design are shown in the following slides.

Typical liner life is estimated as follows:

Feed End Plates……………………400,000 Tons
Feed End Clamp Bars……………400,000 Tons
Cylinder Plates……………………..300,000 Tons
Cylinder Clamp Bars……………..150,000 Tons
Grates………………………………….500,000 Tons
Grate Clamp Bars…………………300,000 Tons

Due to the continuing changes in liner sections we do not have exact numbers for liner life and scrap loss. Liner life varies widely with ore characteristics. When milling “hard” ore with a “high” ball charge liner wear is much faster than when milling “soft” ore with a “low” ball charge.

In conclusion the use of semi-autogenous grinding at DeLamar Silver-Mine has been most successful. A conventional two or three-stage crushing plant followed by rod mill-ball mill grinding would be unable to cope with the wet clays during the spring and fall seasons.

The unexpected presence of a high percentage of clay has had one beneficial effect. The average power for grinding has been 12.5 KWH/T rather than the design value of 15.5 which results in the throughput being some 300-500 tons per day over design.





Although the barrens at four hours were affected by all the variables, the digestion pH and sulfate concentration were of the greatest statistical significance. This means that there is a high degree of correlation between the barrens and the sulfate and pH values. The effect of sulfate and digestion pH on the barrens is pictured in Figure 3. Several significant observations can be made. First, the digestion pH has an optimum range that does not shift as the sulfate increases (Figure 3A). The minimum value of the barrens obtained in this range is, however, dependent on the sulfate concentration. It is not surprising that the barrens increase as the pH is lowered since uranyl peroxide dissolves in acid. It is surprising, however, that as the pH increases above the optimum range, the barrens begin to increase dramatically. Apparently, uranyl peroxide will dissolve in base or large excesses of hydrogen peroxide plus base:


A second observation is that the barrens increase with increasing sulfate level though the increase in barrens is more dramatic at high sulfate levels. As the uranium concentration is increased, the effect of a given level of sulfate on the barrens is decreased (Figure 33). These effects are consistent with the formation of a di-sulfate-uranyl complex.

There is a much greater interaction between the chloride concentration and the digestion pH than was seen with the sulfate effect. Figure 4 shows that at a given pH, the barrens increase significantly with increasing chloride (Fig. 4A). For example, at pH 4.75, the barrens increase from .001g/L U3O8 to 1.44g/L U3O8 when the chloride increases from 0 to 100g/L. The corresponding change for increasing sulfate was only .09g/L to 1.44g/L U3O8. A second observation on the effect of chloride is that the pH corresponding to lowest barrens decreases with increasing chloride. Thus, at 100g/L chloride, the digestion pH should be lower than at 25g/L chloride.

Finally, the effect of a given level of chloride is lessened by increasing the uranium concentration (Fig. 4B). The similarity of the chloride effect to the sulfate effect at high digestion pH and the strong interaction between digestion pH and chloride level indicates that the simple mono-chloro-uranyl complex does not account for the total chloride effect. Apparently, higher chloro complexes are involved.

The interaction between the chloride and sulfate concentrations is pictured in Fig. 5. Several observations can be made. First, the increase in barrens is approximately equal along either axis which means that the effect of chloride or sulfate,independent of the other ion, is similar. Second, the levels of chloride and sulfate are not additive, that is, the barrens at 100g/L chloride alone are much greater than the barrens calculated for 50g/L chloride and 50g/L sulfate. And third, the pattern that is observed by increasing either ion concentration is independent of the concentration of the other ion; the barrens increase slowly from. 0-50g/L of chloride or sulfate, but increase dramatically from 50-100g/L. The magnitude of the increase is affected by the total ion concentration. These observations lead to an important principle. If the chloride strength is high and sulfate low, it is much better to add sulfate than it is to add chloride. This has a bearing on pH control, i.e., in a chloride eluant, sulfuric acid should be used for pH control, not hydrochloric.

Figure 6 shows how the uranium concentration and initial pH interact with the digestion pH. As can be expected from the reaction equation for hydrogen peroxide and uranium, increasing the uranium concentration promotes the formation of uranyl peroxide (Fig. 6A), the barrens being lowest at high uranium levels. There is a slight decrease in the optimum digestion pH as the uranium level increases, supporting the conclusion of an increase in reaction rate at higher uranium values as high pH is not needed to attain low barrens. Outside the optimum digestion pH range, the uranium concentration has little effect on the barrens. The effect of the initial pH is slight. Generally, the higher the initial pH, the lower the barrens. This effect is overshadowed by the dependence of the barrens on the digestion pH.

When the precipitation reaction is allowed to continue for a full 24 hours, the barrens at any set of conditions are improved over the four hour measurements. In Fig. 7, the barrens at 24 hours are portrayed as a function of sulfate, chloride and digestion pH. The barrens follow the same behavior pattern as was observed at four hours (compare Figs. 3, 4). The digestion pH has the largest effect and has an optimum range for low barrens. The barrens increase with increasing sulfate and chloride. However, the effect of high sulfate and chloride is more pronounced at the 24 hour reading. The barrens show a greater rise with increasing sulfate and chloride than was true at the short digestion time indicating that at high concentrations of chloride and/or sulfate, the uranyl complexes that are formed are quite stable. Consequently, it is impossible to precipitate uranyl peroxide at concentrations of chloride and/or sulfate in excess of 100g/L; a longer digestion time does lower the barrens but not sufficiently to make the process usable.

It has been demonstrated that the solution composition and the precipitation conditions have a profound impact on the precipitation of uranyl peroxide. The reaction occurs easily in solutions having less than 50 g/L of sulfate or chloride but is inhibited in higher concentrations. Above 100g/L sulfate or chloride, it is extremely difficult to reach low barrens. The initial pH should be high to help promote the formation of uranyl peroxide. The digestion pH should be kept in fairly strict limits to prevent the redissolution of uranyl peroxide. The reaction works best with high uranium concentrations. By looking at the interplay of these factors, it has been possible to gain a greater understanding of the process chemistry involved. This understanding should make, it easier to effectively apply uranyl peroxide precipitation to specific plant needs. As long as low barrens can be obtained, the uranyl peroxide itself is a sufficiently pure, easily handled product to be considered as a routine processing step.

It should be emphasized that greater purity is an inherent advantage of uranyl peroxide precipitation. The reaction of hydrogen peroxide with uranium is a specific chemical reaction under acidic conditions producing an insoluble, well characterized and crystalline product. ADU, on the other hand, is an amorphous, un-characterized uranyl oxide/hydroxide produced by raising the pH of the solution. Under these conditions it is easy to precipitate other basic metal oxides/hydroxides and thus impurities are incorporated directly into the ADU. With peroxide precipitation, most of the impurities are simply absorbed on the surface of the precipitate and can be removed with washing.

Zinc Refining

The Effect of Air Sparging

The effect of variations in the rate of air sparging (0 to 5000 cm³/min) on the zinc deposit is shown in Figure 7. It can be seen that with no air sparging the surface of the zinc deposit was covered with a black powder (Fig. 7(a) which suggests that the limiting current for zinc deposition was exceeded under these conditions. The substantial decrease in current efficiency (89.9%) and the higher than average decrease in HCl concentration in the spent electrolyte suggests that increased hydrogen evolution occurred. The deposit orientation was [114], [105] which seems to be a characteristic of powdery type zinc deposits.

The maximum rate of air sparging, 4750 cm³/min, yielded the smooth, compact deposit shown in Figure 7(a), which was discussed in detail in the previous section, A one-half reduction in the rate of air sparging to 2375 cm³/min had no significant effect on the current efficiency (99.7%), but as indicated in Figure 7(a), the deposit surface was rough and nodular. Although this aspect of the study was not investigated further, it appears from these results that the rate of air sparging can be reduced to <4750 cm³/min but should be >2370 cm³/min to achieve smooth 24-h zinc deposits.

The deposit obtained with no air sparging was not suitable for SEM analysis but the cross section shown in Figure 7(b) indicates it to be nodular and to contain voids. Although less nodular, the cross section of the deposit obtained with an air sparging rate of 2370 cm³/min, Figure 7(c), also reveals the presence of voids in the deposit. The morphology of this deposit, Figure 7(d), is similar to that obtained at the higher air sparging rate, Figure 5(c).

The Effect of Additives

A series of 24-h zinc deposits obtained from zinc chloride electrolytes containing various additives is shown in Figure 8. The deposit obtained from the addition-free electrolyte was rough and nodular over part of its surface and dendrite formation occurred along the edges (Fig. 8(a)). The additives, TBACl, Percol 140 and Separan NP10, were effective in producing a deposit having a smooth surface and dendrite free edges, Figures 8(b), (c) and (e). Pearl glue was less effective in producing a smooth zinc surface and in eliminating dendritic growth at the edges of the deposit, Figure 8(d).

The structural details of the zinc deposits obtained in the presence of various additives (excepting TBACl) are shown in the series of SEM and OM photomicrographs in Figure 9; the deposit structure obtained in the presence of TBACl was described earlier and is shown in Figure 5.

The deposit obtained from the addition-free electrolyte, Figure 9(a), consists of large, poorly defined hexagonal zinc platelets very similar to the 1-h deposit described in a previous paper, (cf. Fig. 1(a)). The 24-h deposit also contains pores or voids as indicated by the OM photomicrograph, Figure 9(a).

Percol 140 was effective in reducing the deposit grain size, and this resulted in a smooth, compact deposit, Figure 9(b). The deposit consists of fairly distinct hexagonal platelets, aligned at intermediate angles to the Al substrate, This deposit morphology bears a strong resemblance to the characteristic zinc deposit morphology obtained from acid sulphate electrolyte, (cf. Fig. 1(b)).

The presence of Separan NP10 in the zinc chloride electrolyte resulted in poorly defined zinc platelets which were vertically aligned to Al substrate, Figure 9(c). Although the surface of the deposit appears smooth (Fig. 8(c)), the cross section reveals it to be uneven and to contain large voids, Figure 9(c).

Unlike the 1-h deposit obtained in the presence of pearl glue, the 24-h deposit consisted of nodules having a fine grain structure which formed on the surface of the smooth initial deposit layer, Figure 9(d).

In general, TBACl was the most effective addition agent in terms of smoothing the deposit, refining the grain size and eliminating dendritic edge growth (see Fig. 5). It had the added advantage of being the least complex of the additives studied and as such would be least likely to form degradation products which would have harmful effects in other parts of a process circuit.

The current efficiency, energy requirement and orientation results obtained for zinc deposits electrowon from zinc chloride electrolytes containing various additives are summarized in Table 4. In all cases the CE was >90% and the additives generally produced deposits having a predominantly 110 orientation; i.e., the zinc platelets are aligned vertically to the aluminum substrate.


Effect of NaCl Concentration

Situations may arise when the zinc electrolyte will contain large concentrations of NaCl; e.g, from a brine leaching process to remove associated PbCl2- For this reason, the effect of varying concentrations of NaCl on the zinc deposit structure and on zinc deposition current efficiency was studied. The photographs in Figure 12 indicate that dendritic growth and powder formation increase as the NaCl concentration is increased to 3 M.

The effect of increasing NaCl concentration on the structural characteristics of the zinc deposits is shown in the SEM and OM photomicrographs of Figure 13. For NaCl concentrations <2M, the deposit morphology (Fig. 13(a), (b)) is similar to that obtained under standard conditions (Fig. 5(c)) but the deposit orientation becomes more basal; i. e., [002], [103], (Table 6) as compared to intermediate [101] for standard conditions and vertical [110] for the increased Zn concentrations. For 3 M NaCl, the deposit is nodular and contains many deep voids (Fig. 13(c)). A substantial decrease in CE, from 96.3% at 0 M NaCl to 77.3% at 3 M NaCl, occurred also, Table 6.

The effect of deposition time or total ampere- hours on the quality of the zinc deposits electrowon from an electrolyte containing 15 g/L Zn and 3 M NaCl under otherwise standard conditions is shown in Figure 14. The deterioration in the deposit quality with time can be seen from the series of photographs, Figure 14(a) to (d).

After 6-h (12-Ah) depositon time, the deposit (Fig. 14(a) is fairly smooth near the center, but powder formation is prevalent on the sides of the deposit; further, the deposit edges indicate the on-set of dendrite formation. The CE for this deposit was 96.7%.

The 12-h (24 Ah) deposit (Fig. 14(b) is also smooth at the center but large dendrites have formed at the deposit edges and at the bottom of the deposit. The CE has decreased to 94.9%. After 18-h (36-Ah), the deposit (Fig. 14(c)) is rougher and more dendrites have formed on the surface of the deposit. The CE for this deposit was 89.9%.

Finally, after 24-h (48-Ah), the deposit (Fig. 14(d)) consists of a rough, dark-grey, powdery surface with large dendrites at the deposit edges. The CE was only 76.0%.

The fact that the Zn deposits deteriorate with increasing Cl concentration may be attributed to an increase in the formation of Zinc-chloro complex ions (e.g. ZnCl4=) at the expense of Zn2+ cations. An increase in total Cl concentration would favor the formation of zinc-chloro complex ions which could result in a corresponding decrease in Zn2+ concentration and hence give rise to a limiting current condition for Zn deposition which favors powder formation and dendritic growth.

The effect of increasing deposition time on the structural characteristics of the zinc deposits obtained from an electrolyte containing 3 M NaCl is shown in the SEM and OM photomicrographs, Figure 15. The deposit obtained after 12 Ah is even and compact, Figure 15(a), and its morphology is similar to that obtained after 48 Ah under standard conditions; i.e., in the absence of NaCl (cf. Fig. 5(c)).

Increasing the total Ah to 24 and 36 results in thicker deposits which remain even and compact; however, as revealed by the SEM photomicrographs (Fig. 15(b) and (c)), the tendency toward powder formation increases with increasing deposition time.

After 48 Ah (Fig. 15(d)), a high degree of powder formation occurs; the cross section is thinner than that after 36 Ah (cf. fig. 15(c) and (d)) because much of the growth has occurred on the deposit edges as large dendrites (see Fig. 14(d).

Effect of Current Density

The effect of varying the current density on zinc deposition from chloride electrolyte was studied over the range 161 A/m² (15 ASF) to 646 A/m² (60 ASF). The deposits obtained at the various current densities for a total deposition tine of 48-Ah are shown in Figure 16. The deposits were all very similar; the 646 A/m² deposit (Fig. 16(a)) showed more edge growth than was observed at the lower current densities. Variations in the current density had no significant effect on the deposit morphology (Fig. 16(b) which remained similar to that obtained under standard electrolysis conditions (Fig. 5(c). The cross section (Fig. 16(c)) for the deposit obtained at 484 A/m² was smooth and compact.

The current efficiency was >96% for all current density values, except 161 A/m² (15 ASF) where it decreased to 82% (Table 7). As expected, the energy requirement increased with increasing current density (Table 7) because of the higher voltages required. The deposit orientation was not affected by these variations in the current density.

Effect of Impurities

The effect of the impurities: Co, Cu, Cd, Ni, Fe(II), Fe(III), Sb and Pb, both individually and in various combinations, on the zinc deposit quality and on the current efficiency of zinc deposition was also studied. Zinc deposits obtained from chloride electrolyte containing various concentrations of Co, Cu, Fe and Sb under standard electrolysis conditions are shown in Figure 17.

The addition of 0.08 mg/L Sb to the electrolyte had no significant effect on the physical appearance of the zinc deposit (Fig. 17(a)) but the CE was reduced to 90% (Table 8). At an Sb concentration of 0.2 mg/L (Fig. 17(b)), the deposit edges became rougher but the CE increased to 94.12. At 0.5 mg/L Sb, the deposit edges consisted of a mixture of dendrites, powder and large modules; the CE was 88.6%.

The addition of 5 and 10 mg/L Cu to the electrolyte, Figures 17(c) and (d), respectively, resulted in zinc deposits characterized by black powder formation along the top and bottom edges. The deposit surface was noticeably rougher at 5 mg/L Cu; the CE decreased to 91% in both cases (Table 8).

The deposit obtained from an electrolyte containing 30 mg/L Co (Fig. 17(e)) was similar to that obtained from an “impurity-free” electrolyte; the CE was 94%. The presence of cadmium (10 mg/L) in the electrolyte resulted in a rough nodular zinc deposit; the CE was 91%.

The deposit obtained from an electrolyte containing nickel (10 or 30 mg/L) was rough and nodular. Black powder formed along the deposit edges, particularly at the higher Ni concentration. Deposit re-solution occurred and the CE decreased to 816% for 10 mg/L Ni and to 51.7% for 30 mg/L Ni.

The addition of 200 mg/L Fe(II) or Fe(III) to the electrolyte had no significant effect on the physical appearance of the zinc deposit but the CE was reduced to 89%.

Although reasonably high levels of certain metallic impurities (e.g. Co, Fe) could be tolerated, combinations of two or more of the impurities were detrimental to zinc electrowinning from chloride electrolyte. 30 mg/L Co and 0.2 mg/L Sb, Figure 18(a)), resulted in a rough, nodular deposit; some re—solution occurred and the CE was 88%.

A similar deposit (Fig. 18(b)) was obtained when the electrolyte contained 5 mg/L Cu and 0.2 mg/L Sb; black powder formation occurred along the deposit edges and the CE was 86.2%. The entire surface of the deposit obtained from an electrolyte containing 30 mg/L Co, 1 mg/L Cu and 0.2 mg/L Sb was covered with a black powder, Figure 18(c); the CE was 84.4%.

The deposit obtained from an electrolyte containing 1 mg/L Cu and 10 mg/L Pb was relatively smooth but showed some signs of re-solution near the bottom (Fig. 18(d)); the CE was 89.9%. The SEM photomicrograph (Fig. 18(e)) reveals a poorly crystalline deposit, typical of that obtained from sulphate electrolytes containing similar levels of Pb.


A diagram of the laboratory mineral processing equipment used in the precipitation of copper and the subsequent regeneration of cyanide. The general technique employed in these operations was as follows:

1) To pregnant solution in a 1-liter stainless steel precipitator (A) add the sulfide (Na2S; NaHS; CaS) required for precipitation.

2) Bolt the head of the precipitator in place over a Teflon gasket; add the required volume of sulfuric acid to the acid storage container (B).

3) Add acid through valve (C) into the pregnant-sulfide mixture in (A) with valves (H) and (D) closed and with magnetic stirrer (E) operating.

4) Keep system closed and stir for 5 min. to complete precipitation.

5) Open outlet valve (D) to allow the evolved cyanide to pass into traps (F) and (G) containing alkaline absorbing solution.

6) For gas stripping, connect inlet to a source of air or nitrogen, open inlet valve H, and pass gas into and upward through the contents of the precipitator (A) and then out through exit valve (D) into the traps (F) and (G).

7) For steam stripping, a water-cooled condenser and receiver (not shown) are inserted between (A) and (G), and (A) heated and stirred to remove a cyanide-water mixture which is collected (F) and cooled in an ice-water mixture. (F) is vented into trap (G) to prevent loss of cyanide.

8) The cyanide contents of (F) and (G) were determined by direct titration with silver nitrate. The copper bearing precipitate was filtered from the liquors in (A) after the stripping operation, was dried, and analyzed for copper content.

The types of receivers (F) and (G) were varied to facilitate cyanide recovery and manipulation (cooling, volume of distillate handled, quantity of cyanide recovered etc.).


In a fast mineral processing reaction, copper minerals dissolve in cyanide and act as cyanides in the recovery of precious metals by cyanidation. However, the application of cyanide leaching to the recovery of copper from copper ores requires considerably different conditions than used in precious metal cyanidation (higher cyanide concentrations; shorter leaching times; oxidizing conditions unnecessary).

Cyanide Complexes of Copper: The predominant species formed when copper dissolves in cyanide is one containing 3 moles of cyanide to 1 mole of copper (Cu(CN)3). This indicates that at least 2.32 lb NaCN equivalent are required to dissolve one pound of copper, if no side reactions occur.

Cuprous Copper Minerals: The reaction of cyanide with typical cuprous sulfide and oxide copper minimum. Fig. 1:—Simplified precipitation-regeneration system. Not shown in the diagram are a condenser and receiver for steam stripping.

The reaction is generally written as follows:


CU2S + 6 NaCN = 2 Na2Cu(CN)3 + Na2S [1]

CuaS + 3 Ca(CN)a = 2 CaCu(CN)a + CaS [21


Cu20 + 6 NaCN + HaO = 2 Na2Cu(CN)3 + 2 NaOH [SI Cu20 + 3 Ca(CN)a + HaO = 2 CaCu(CN)3 + CalOH). [41

Cupric Copper Minerals: Cupric copper undergoes reduction in cyanide solution. Therefore, when cupric minerals are dissolved in cyanide, the following reactions occur:

2 CuC03 + 8 NaCN = 2 Na2Cu(CN)3 + 2 NaaCOa + (CN).

Cyanogen, (CN)2, in alkaline solution, undergoes a further reaction as follows:

(CN|! + 2 NaOH = NaCNO + NaCN ■ H-.0

The over-all reaction therefore is:
2 CuCOa + 7 NaCN + 2 NaOH = 2 Na2Cu(CN)3 + 2 NaaCOa + NaCNO + HaO
In this reaction, the reduction of the cupric copper and the subsequent oxidation of cyanogen to form cyanate results in the loss of 0.5 mole of NaCN for each mole of copper dissolved (0.39 lb NaCN equiv./lb Cu).

Thiocyanate Formation: A second important source of cyanide loss occurs through the formation of thiocyanates. The exact mechanism of this thiocyanate formation is not known. The reaction of cyanide with sulfides and thiosulfates does not readily produce thiocyanate. However, higher poly-thionates such as the tetrathionates and pentathio-nates react readily with cyanide to give thiocyanate as follows:

NaaSiOo + 3 NaCN + HaO =

NaaSOi + 2 HCN + Na2S2Oj + NaSCN [81 Na_S:,0.; + 4 NaCN + HaO =

NaaSOi + 2 HCN + NaaSaOa + 2 NaSCN [91

Ferrocyanide Formation: In general, ferric iron is not soluble in alkaline cyanide solutions. Ferrous iron, however, does dissolve to form ferrocyanide, Fe(CN),f‘. Since ferrocyanides do not decompose readily in cold, dilute sulfuric acid (used in subsequent copper recovery), such compounds may constitute a source of cyanide loss in the copper cyanidation process.

Leaching Tests on Copper Minerals: Five-gm samples (minus 100 mesh) of several copper min­erals (Ward’s National Science Establishment, Roch­ester, New York) were leached for six hours in 1000 ml. of NaCN solution at various cyanide to copper ratios. In the tests on bornite and chalco- pyrite, 10-gm samples of mineral were leached for four hours using Ca(CN)a solutions. Cyanide con­sumptions were calculated by analysis of the preg­nant solutions for non-regenerative cyanide loss (by distillation) and for ferrocyanide (by iron analysis). Gold Mining Equipment.

Rate of solubility: The rapid rate of extraction of copper from several copper minerals was demon­strated by leaching 50-gm samples of synthetic ores (2% Cu, —200 +325 mesh) prepared from various copper minerals and quartz. These ores were leached in open beakers at 40% solids and a cyanide ratio of 3.0 gm NaCN equivalent per gm of con­tained copper. Results are shown in Fig. 2. As in­dicated, the rate of extraction of copper from the minerals tested increased in the order of bornite, covellite, chalcocite and malachite. Copper extrac­tions of 81.6% (chalcocite) and 94.2% (malachite) were obtained in 15 min., thus demonstrating that samples of a copper ore (minus 28 mesh and minus 10 mesh) and samples of flotation tailings and cleaner flotation tailings were leached with cyanide. The predominant copper mineral in these samples was chalcocite. Samples of the ore suspen­sion were withdrawn periodically, immediately fil­tered and washed and the filter cakes assayed for copper.

As indicated in Fig. 3, the rate of solubility of copper was rapid in the case of the flotation tailing (1.08% Cu), although the cyanide to copper ratio (2.34:1) was lower than that required to give op­timum copper extraction (i.e., usually 3.0-3.5:1). Leaching of the copper ore at coarser sizes also was rapid, since from the minus 10 mesh sample, 75.1% of the total copper dissolved in the first 30 min and 86.7% in 4 hr. From the minus 28 mesh sample of the same ore, 83.9% of the copper was extracted in 30 min and in 2 hr essentially all the extractable copper was in solution.

Leaching was conducted in a laboratory Fagergren flotation machine and in the first two min. 84% of the copper was dissolved. Thus the rapid solubiliz­ing action of cyanide on copper was demonstrated clearly and was subsequently utilized in flotation operations to eliminate the activation of pyrite by contaminating surface films of copper and to permit the production of final copper concentrates of high grade.

Gold Leaching Plant Flowsheets

The Kalgoorlie ore deposits occur in Pre-Cambrian rocks and essentially consist of calc schist and quartz-dolerite-greenstone. The ores contain free gold and tellurides. Auriferrous pyrite is the most abundant sulphide and is occasionally associated with chalcopyrite, tetrahedrite and arseno-pyrite. The gold associated with the pyrite is in a very fine state and is not liberated even after very fine grinding.

The crushing section reduces run-of-mine ore to minus 1/2 inch which is further reduced by rod mills and ball mills to approximately 75% minus 200 mesh. Rod mill discharge is pumped over strake tables ahead of the ball mills with the concentrate being amalgamated, retorted and finally smelted. Final ground product goes direct to flotation before gold leaching by cyanide.

The pyritic flotation concentrate and tailing are treated in separate thickeners with the overflow from both thickeners going to flotation circuit water storage. Underflows from each thickener are cyanided in separate agitators.

The cyanided concentrate is filtered and the filter cake treated in Edwards roasters with the calcine again being cyanided and filtered. Post-cyanidation is the term used to describe this treatment method. The calcine filter cake is repulped and pumped to the flotation tailings cyanide agitators. Filtrates from both stages of filtering are combined and flow to the high grade precipitation section, clarified, and the gold precipitated with zinc dust. This precipitate is then filtered, roasted, fluxed and smelted.

The tailings cyanide agitators discharge to a washing thickener. The underflow is filtered and the filter cake repulped with barren solution and pumped to the residue dam. Filtrate is circulated back to the washing thickener. Overflow from the washing thickener passes to the low grade precipitation section and is treated similarly to the high grade solution. Barren solution is circulated back to both cyanide sections.

Until 1956, a second 500 ton per day mill operated using a pre-cyanidation flowsheet. In this case, the ground ore was cyanided with the cyanide tailings filter cake) being gassed with sulfur dioxide and then floated for sulfide recovery. Flotation concentrate was then roasted, recyanided, and returned to the grinding circuit. In 1956 this plant was converted to the post cyanide flowsheet.Gold and Silver Leaching

The Emperor flowsheet is so different that it is presented in detail on two pages. It is unique in that tellurium metal is produced along with gold bullion and copper cement both unusual products from a gold ore. The flowsheet is distinguished in that there are three separate but allied circuits; the main cyanidation circuit supported by roaster calcine cyanidation, a tellurium metal recovery circuit, and a copper cement recovery operation.

The mine ore is washed in a blade type mill after which the minus 4 plus 1/2 inch product is hand picked for removal of waste. The slime portions go to flotation for recovery of sulfides with the slime tailings going to waste.

After grinding, the tellurium and copper minerals are recovered by flotation. The concentrate containing some of the gold goes through an oxidation step with NaOH, Na2C03 and Ca (OCl which renders the gold soluble for cyanidation and the tellurium leachable. The concentrate residue containing copper is roasted and leached for production of cement copper.

The flotation tailings are cyanided for gold extraction following usual practice. The cyanide residue is conditioned with SO£ gas from the copper concentrate Edwards roaster and then floated with copper sulphate and xanthate. After filtration, the concentrate is fed to the roasters along with the above main stream copper concentrate. Flotation tailings are used for mine fill.

Roaster calcines, following leaching for copper, are washed and cyanided separately and then the calcine-residue is fed back through the gassing tower circuit.Gold and Silver Plant flowsheet


Alternative reagents for leaching gold

With the growing environmental pressures on the use of cyanide in certain regions and the unsuitability of cyanide for certain complex ores, various reagents have been evaluated. Some of these alternative reagents are summarized in Table 4.

It appears that complexation of copper with ammonia lowers the consumption of CK Similarly, present work on leaching copper-molybdenum sulfides indicates that carbonate also complexes with copper, which depresses the amount of copper leached with CI2. This may have applications to copper-gold ores. Then there are various other processes including the K-process using undisclosed reagents However chlorine and thiourea have received the most attention and have specific applications.

As shown, the advantage of chlorine and thiourea is that they dissolve gold much faster than CN- and sometimes give higher recoveries; but unfortunately greater reagent quantities are often required (Table 5).

The problem with chlorine is that it reacts with sulfide minerals leading to high CI2 consumption and acid production (equation 3).

With thiourea, it is the dimerisation and degradation of thiourea that leads to high consumption (equation 4). With other reagents there is the problem of cost, or stability, or how gold is recovered.


Of all the reagents examined so far, thiourea offers the best prospects for the difficult ores which cannot use cyanide. It is particularly suited to the acidic residues from bacterial leaching or pressure leaching of refractory sulfides because it complexes gold in acid media. A recent examination of a number of Australian oxidised gold ores identified some other promising applications. Furthermore, over the last 3 years there have been developments which could make thiourea even more attractive. Recent work has snown that SO2 inhibits dimerisat’on and degradation of thiourea by controlling the Eh and that a modified thiourea is even more stable. Thus certain types of ore which consume base or CN- could be economically treated with modified thiourea reagents. Furtner research in this area is currently being undertaken.


In the past few years much progress has been made in understanding and developing the C.I.P. process. However, there is still much fundamental worK to be aone ana mprovements to be made with the treatment of refractory ores. No two ores are exactly alike and each needs to be evaluated with a range of options at our disposal.

By understanding the characteristics and properties of carbon, its activity and performance has been improved. Whilst no immediate threat to current C.I.P. practice exists, alternatives to carbon and alternatives to CN- may develop in a few years time.


Solvent extraction and ion exchange

Solvent extraction is established technology for copper and uranium, but the extraction of gold has been hampered by the lack of a selective reagent. In 1984 however, it was reported that mixtures of secondary amnes and TBP were selective for aurocyanide at pH 9. These reagents are widely used in the uranium industry and ODen up the prospect of solvent extracting heap leach eluates or developing solvent-in pulp shows that the solvent mixture is quite selective for gold over copper. Gold is extracted at pH 9 and stripped at pH 11.

Resins, however, have attracted most attent on because they offer high loadings of gold without fouling by organics and are relatively easy to regenerate .Some Russian plants have used resins for years but the main disadvantages are related to their selectivity, stripping and screening. As recently discussed, strong base resins are difficult to strip and require reagents like thiocyanate or thiourea. Weak base resins on the other hand are simply stripped by pH adjustment, but most such resins are not pure and have strong base impurities. Hence they do not strip efficiently when the pH is raised. These resins also load copper zinc and nickel cyanide complexes just as readily as gold and show less discrimination than carbon. Thus the ideal resin has yet to be developed. It is therefore interesting to note that an entirely new approachs being developed which uses ion-exchange fibres woven into cloth. These are polyacrylonitrile fibre with imidazole groups or polybenzimidazole functional groups which represent a new concept in substrate and functional group design.