Table of Contents
Numerous low-grade gold-silver ore deposits are being mined and milled throughout the Western United States. In addition to gold and silver, many deposits contain as much as 15 ppm of mercury. During cyanidation, 10 to 30 pct of the mercury and 85 to 90 pct of the gold are typically solubilized. The reactions for gold and mercury extraction with cyanide are —
2 Au + 4 CN- + O2 + 2 H2O → 2 Au(CN)2- + 2 OH- + H2O2………………………………………….(1)
2 Au + 4 CN- + H2O2 → 2 Au(CN)2- + 2OH-………………………………………………………………..(2)
Hg²+ + 4 CN- → Hg(CN)²4-…………………………………………………………………………………………(3)
2 Hg + 8CN- + O2 + 2H2O → 2Hg(CN)²4- + 4 OH-……………………………………………………….(4)
Mercury builds up in the recycled leach solutions because only part of the mercury is adsorbed on carbon in the loading circuit. Mercury and gold are stripped from the carbon with caustic cyanide solution and electrowon on steel wool cathodes. Mercury must either be recovered or precipitated because of the health hazard during smelting of cathodes and regenerating of activated carbon. One gold mill recovers mercury by retorting the cathodes prior to smelting. Another mill autoclaves the ore to extract minimal mercury. Soluble mercury can be precipitated with sulfides, as shown by the chemical equation
Hg(CN)²4- + MS → HgS + M²+ + 4CN-………………………………………………………(5)
Silver, iron, and zinc sulfides have been used to precipitate mercury from gold cyanide solutions but require slurry filtration before sulfide addition to reclaim excess reagents. Silver sulfide addition would tie up considerable silver, and silver sulfide excess would have to be recovered. Iron sulfide is undesirable because of cyanide loss due to ferrocyanide. Calcium and sodium sulfides, which were used in this investigation, are not harmful to the gold recovery process. The objective of this Bureau of Mines investigation was to study the effects of operating parameters on mercury extraction and examine methods for mercury removal from gold cyanide leach slurries.
Material Equipment and Operating Procedure
Ore used in this study was obtained from a gold mining operation in northern Nevada. The ore body is a well-oxidized hydrothermal deposit with gold finely disseminated as micrometer-size particles. The limestone-siltstone host rock has been silicified to a jasperoid which has partially weathered to a clay. A screen analysis of the ore (table 1) shows gold and mercury disseminated throughout all screen fractions. Chemical analysis is given in table 2.
Ore was dry-ground to minus 10-mesh in a laboratory roll crusher. Charges weighing 1 kg were wet-ground in distilled water at 50 pct solids in a ball mill using a 10-kg ball charge. The total grinding time was between 30 min and 1 h, depending on the mesh size required. The ore was generally ground to 90 pct minus 200 mesh, except as noted. The slurry was filtered, and the filter cake was dried at ambient temperature for several days.
The sodium cyanide, lime, and sodium and calcium sulfides used in this study were all reagent grade. A 16- by 20- mesh coconut shell activated carbon was used in the sulfide precipitation investigation.
Leaching tests to determine the effect of NaCN, pH, temperature, and retention time were conducted in an air-agitated Pachuca-type vessel, shown in figure 1. This vessel was a 1-L separatory funnel with attachments at the bottom for introducing air to aerate and agitate the solution. The vessel was heated with electrical heating tape, covered with insulation and aluminum foil, and regulated by a powerstat. An NaOH scrubbing column was placed in the top of the vessel to prevent HCN losses. Pachuca leaching tests were initiated by adding 50 g of dried, ground ore to 200 mL of leaching solution at a predetermined temperature in the leaching vessel and leaching for varying lengths of time. Airflow to the vessel was 2 L/min. Upon test completion, the leach slurry was filtered on a Buchner funnel with No. 2 Whatman filter paper. The filtrate volume was measured, and aliquots were collected for analysis. Residues were washed, dried at room temperature, weighed, and analyzed for gold and mercury by a fire assay- atomic adsorption technique. Following a 24-h cyanide leach, mercury precipitation was accomplished by adding either sodium or calcium, sulfide to the leach slurry. The slurry was mixed by rolling in a 250- to 4,000-mL closed plastic bottle for varying lengths of time at 20° to 25° C. The solids content was about 40 pct.
Additional sulfide precipitation tests were conducted with carbon added to the leach slurry. Five hundred grams of dry-ground ore was leached with 870 mL leach solution containing 0.34 lb NaCN per ton of solution and 2.5 lb lime per ton of ore. The slurry was rolled in a plastic bottle for 24 h at 20° to 25° C. Following cyanide leaching, varying amounts of sulfide were added to the slurry and mixed by rolling for 1 h. The slurry was contacted with about 18 g/L activated carbon for an additional hour. The carbon was screened from the slurry, and the slurry was contacted with fresh carbon a second and third time to simulate a three-stage carbon-in-pulp adsorption. Gold, silver, and mercury analyses were accomplished using radioisotopes. The isotopes were added to the leach slurry after a 24-h cyanide leach and counted on a Tracor Northern TN-1710 gamma counter. The carbon was counted after removal from the slurry. Because all of the gold was adsorbed on the carbon, the ratio of gold counts in the slurry before carbon adsorption to the counts on the carbon after carbon adsorption is equivalent to 100 pet adsorption. The amount of mercury or silver adsorbed on the carbon was determined from the mathematical equation:
Pct Hg/100 = H counts on total carbon/Hg counts on 5 mL starting slurry/Au counts on total carbon/Au counts on 5 mL starting slurry………………………………………………….(6)
(Pct Ag is determined by substituting for Hg.)
Radioisotope analysis compared well with atomic adsorption analysis; therefore, self-determination correction factors for different sample compositions and gamma ray energies were not used.
In addition to the carbon-in-pulp tests, work was done by adding CaS to a laboratory ball mill, along with NaCN and lime. Calcium sulfide was varied from 0.012 lb to 0.09 lb per ton of ore. One thousand grams of minus 10-mesh ore, 1.25 g CaO (2.5 lb/ton ore), 0.8 g NaCN (0.53 lb/ton solution), 1,000 mL water, and the CaS were added to the ball mill and ground for 45 min. The resulting slurry was washed into a 9-L bottle with 2,000 mL water and mixed. Gold and silver were removed from the slurry by adsorbing on 6- by 16-mesh activated carbon at ambient temperatures in three stages. One gram of carbon was used in each stage. The first-stage carbon was in contact with the slurry for 4 h, the second stage for 2 h, and the third stage for 18 h. Gold, silver, and mercury on the carbon were analyzed by neutron activation and gamma counting. The residue and solution were determined by atomic adsorption.
Results and Discussion
Effect of Process Variables on Mercury Extraction
Cyanide Concentration
The effect of NaCN concentration on mercury extraction was determined by varying the NaCN concentration between 0.34 and 20 lb NaCN per ton of solution. Minus 325-mesh ore was slurried and air-agitated for 8 h at pH 11.5 and 20° C. The results of these tests are shown in figure 2. Mercury extraction increased from 10 pct at 0.34 lb NaCN per ton of solution to 40 pct at 20 lb NaCN per ton of solution.
Currently, cyanide mill operators add 0.02 to 2 lb NaCN per ton of solution. Some gold producers are interested in increasing mercury recovery; to others, mercury is a nuisance. The cost of increased cyanide would have to be weighted against the value of mercury. Gold extraction was unaffected over the range studied.
Leaching Time
The leaching time was varied from 8 to 48 h. Ore ground to minus 200-mesh was leached at pH 11.5 and 20° C with 10 lb NaCN per ton of solution and air agitation. The effect of leaching time on mercury extraction is shown in figure 3. The data show that 24 pct of the mercury was extracted during the first 8 h; additional leaching for 40 h extracted another 4 pct. A leaching time of 1 h or less was required to recover 82 pct of the gold, and 15 h was required to recover 90 pct. In some mill operations the ore is in contact with cyanide solution during grinding, solid-liquid separation, leaching of thickener underflow, and carbon-in-pulp treatment so that the total retention time is more than 24 h. Controlling mercury extraction by reducing contact time would be difficult.
Solution pH
Sodium cyanide leach solution containing 10 lb NaCN per ton of solution was adjusted to pH values between 10.6 and 11.9 with lime and mixed with minus 325- mesh ore for 8 h. The results of these tests are shown in figure 4. Mercury extraction increased from 21 pct at pH 10.6 to a maximum of 32 pct at pH 11.5. Gold extraction increased from 91 pct at pH 10.6 to a maximum of 97 pct at pH 11.5.
Gold operations generally leach between pH 10.5 to 11. A mill operation interested in increased mercury extraction from a similar type of ore could increase both gold and mercury extraction by operating near pH 11.5.
Temperature
A test series to determine temperature effects on gold and mercury extraction was conducted using minus 325-mesh ore. Leach solution of pH 11.5 containing 10 lb NaCN per ton of solution was air- agitated with the ore at 20° to 50° C for 8 h. The test results are shown in figure 5. Mercury extraction increased only slightly, but gold extraction decreased from 98 pct at 20° C to 85 pct at 50° C.
The decreased gold extraction may be caused by the partial destruction of sodium cyanide or reduction of dissolved oxygen at elevated temperature.
Particle Size
Different particle size fractions were leached at pH 11.5 and 20° C for 8 h with a cyanide leach solution containing 1 to 10 lb NaCN per ton of solution. Test results for mercury are shown in figure 6. Generally, mercury extraction increased as particle size decreased and sodium cyanide concentration increased. Mercury extraction from minus 10- plus 48-mesh and minus 48- plus 270-mesh fractions increased by 5 and 10 pct, respectively, as the NaCN increased. Mercury extraction from minus 270-mesh and minus 325- mesh particle size fractions increased from about 9 pct to 30 pct as the NaCN concentration was increased from 1 to 10 lb NaCN per ton of solution. Grinding to less than minus 270-mesh does not improve mercury extraction.
Additional tests showed that gold extraction increased from 73 pct to 91 pct as the particle size decreased from minus 10 plus 48 mesh to minus 270 mesh.
Mercury Removal by Sulfide Precipitation
Figures 2-6 show that >10 pct of the mercury was extracted from the ore under all conditions. Mercury must therefore be recovered as a byproduct or precipitated as an insoluble product and disposed of to the tailings pond. Sulfide precipitation offers an alternative method for gold producers who are faced with a mercury problem and are not interested in mercury recovery.
Mercury Precipitation from Leach Slurry
Gold ore having 17 ppm mercury was leached with solutions containing 0.34 and 8 lb NaCN per ton of solution and lime, and then contacted with Na2S. The results of these tests are shown in figure 7. Increasing Na2S lowered mercury extraction when low concentrations of cyanide (0.34 lb NaCN per ton of solution) were used. However, increasing the sulfide in solution increased mercury extraction when high concentrations of cyanide were present in the leach solution. This is probably due to the formation of a soluble HgS²2- complex. High concentrations of Na2S have been used to extract mercury from concentrates; therefore, care must be taken to control the amount of sulfide added.
To determine the effect of Na2S concentration and time on mercury sulfide precipitation, additional tests were conducted. Cyanide solutions were contacted with ore for 24 h to leach 12 pct of the mercury; then Na2S was added. The results are listed in table 3. Sodium sulfide precipitated >99.9 pct of the solubilized mercury with 0.02 lb Na2S per ton ore, but within 0.5 h, the precipitated mercury started to redissolve. After 4 h, about 30 pct of the precipitated mercury sulfide had redissolved.
Because of redissolution of mercury sulfide with Na2S, tests were conducted with CaS. Cyanide solutions at 0.34 and 1.0 lb NaCN per ton of solution were used to extract 12 and 27 pct, respectively, of the mercury in 24 h. After leaching,
CaS was added and samples were taken periodically. The results of these tests are listed in table 4. Calcium sulfide precipitated about 96 pct of the solubilized mercury in 0.25 h, and 99.8 pct in 0.5 h, using 0.02 lb CaS per ton of ore. After 7 h, about 7 pct of the precipitated mercury sulfide had redissolved. Increasing the CaS concentration to 0.09 lb per ton of ore resulted in nearly 100 pct of the soluble mercury precipitating, and after 24 h the mercury sulfide did not redissolve
A comparison between mercury sulfide redissolution with Na2S and CaS is shown in figure 8 (test 1, table 3; and test 2, table 4). About 10 times more mercury sulfide redissolved with Na2S than with CaS after 4 h.
Mercury Precipitation Prior to the Carbon-In-Pulp Circuit
Tests were conducted to determine the effect of sulfide addition on carbon adsorption of gold and mercury in a laboratory carbon-in-pulp (CIP) circuit. The results of these tests are given in table 5.
Gold and silver adsorption on carbon without prior sulfide addition was nearly 100 pct after the first stage, while only 61 pct of the soluble mercury adsorbed. Additional adsorption stages recovered all the gold and silver and 9 pct more of the mercury. Addition of 0.023 lb CaS per ton of ore did not decrease gold and silver recovery, but reduced mercury adsorption to about 5 pct. Increasing the CaS concentration to 0.047 lb CaS per ton of ore caused a slight decrease in silver adsorption in the first stage and slightly less overall recovery, but mercury adsorption dropped to 0.9 pct in the three stages.
Sulfide precipitation could be used in the present mill operation with only minor modification. Sulfide addition should be made prior to adsorption and
enough time provided to precipitate the mercury. A schematic of a northern Nevada mill operation with required changes for sulfide precipitation is proposed in figure 9. Sulfide addition could be made to the thickener underflow and the clear thickener overflow. In this flowsheet, the underflow goes through a series of leach tanks. Sulfide would be added to the last tank prior to entering the CIP circuit. Mercury in the thickener overflow would be recovered by sulfide precipitation and filtration. The slurry from the CIP circuit would contain precipitated mercury and go to the tailing pond. The remaining part of the flowsheet would be stripping, electrowinning, and refining. Additional tests are required to determine if the strip solution is low enough in mercury so that the gold cathode is not contaminated during the electrolysis.
Mercury Precipitation in a Laboratory Grinding Circuit
Several gold operations add NaCN and lime to the grinding circuit to extract gold and silver during grinding. Some of the mercury is also extracted at this time. Tests were conducted to determine if CaS added with NaCN and lime to the ball mill would prevent mercury extraction and affect gold, silver, and mercury loading on carbon. The results of these tests are given in table 6 and figure 10.
Mercury extraction and adsorption on carbon were decreased from 15 to 0.4 pct and 5.5 to 0.17 pct, respectively, by increasing the amount of CaS added to the simulated ball mill circuit from 0.012 to 0.096 lb per ton of ore. Gold extraction may decrease slightly, but additional research is required to verify this. Gold loading on carbon was unaffected. Silver extraction was unaffected, but carbon loading may have decreased slightly. Both gold and silver recovery on carbon was lower than in previous sulfide precipitation tests. This was due to using much less carbon (0.33 g/L rather than the normal 20 g/L). Decreasing the carbon concentration made it possible to analyze for the small amount of mercury adsorbed on the carbon.
Adding CaS to the ball mill would simplify the flowsheet in figure 9. The extra equipment would be unnecessary since the mercury would remain with the residue and be disposed of in the tailings pond. Essentially no change in the mill operation would be required.
Conclusions
Mercury extraction was decreased from 40 to 10 pct by (1) decreasing NaCN concentration from 20 to 0.34 lb NaCN per ton solution, (2) decreasing the pH from 11.5 to 11, and (3) increasing particle size from minus 270 mesh to minus 10 plus 48 mesh.
Calcium sulfide addition to cyanide leach slurry or to a laboratory grinding circuit containing NaCN and lime reduced mercury dissolution to <0.5 pct. Mercury loading on activated carbon was reduced to <0.2 pct. Gold loading on activated carbon was affected very little by the sulfide addition, but silver may be slightly affected. Additional tests are required to verify this.