Bioleaching of cobalt and zinc from pyrite ore in relation to calcitic gangue content

dloleeching of a pyrite ore containing high concentrations of cobalt (0.1%) and zinc (O.oS5%) was affected by small amounts of calcltlc gangue (from 0.01 to 1 .Ol%). Results from an air-lift percolator and from Erlenmeyer flasfc experlments show that a small percentage of calcite raises the pH and arrests the growth of the acidophilic bacterium Thlobaciiius ferrooxidans. In percolator experiments, when calcite is completely removed by the continuous addition of small quantities of acid, and the pH of the liquor becomes acid, the micro-organism begins to grow and to bio-oxidize the pyrite ore. The growth of T. ferrooxidans shows different lag phase spans (from 13 to 190 days) depending on carbonate dissolution. The metals Fe, Zn and Co are released into the leaching solution together at dlfferent rates after a lag-time whkh depends on calcite concentratlons in pyrite gangue. Metal ratios in the mlneral bulk are dlfferent from those In the liquor, Zn dissolving 5 times more readlly than Co. Bloleaching rates for metal removal from pyrite ere higher In percolator (for Fe, from 5 to 15 mglllh) than In flask experlments (from 0.5 to 2 mgilih), but the lag phases are shorter (from 2 to 65 days). The differences between the two systems are related to calcite dissolution and gypsum precipitation.

Bioleaching of cobalt and zinc from pyrite ore in relation to calcitic gangue content F. Baldi, A. Bralia, F. Riccobono and G. Sabatini The continuing world-wide depletion of metal ore deposits and the accumulation of tailings, which cannot be further processed by conventional means, is encouraging the development of bioleaching technology (bio-hydrometallurgy) for recovering metals such as copper and uranium (Brierley 1978;Kelly et al. 1979) and'also 'strategic' metals such as gold, silver and cobalt (Ehrlich 1987;Pinches et al. 1987;Wichlacz & Thompson 1987). The bioleaching of sulphide ores may be performed by means of the acidophilic and chemolithotrophic bacterium Tbiobacih ferrooxidans or co-cultures of this micro-organism with other thiobacilli or acidophilic bacteria (Dugan et al. 1970;Groudev et al. 1978;Schafer 1983). However, this biotechnological process is not always successful. High concentrations of toxic metals (Hg, As, Sb, etc.) in the sulphide ore lattice inhibit bacterial catalysis (Imai et al. 1975;Barbie 1977;Ingledew 1982;Mahapatra & Mishra 1984;Paknikar & Agate 1987) as these metals are released in the leaching solution. Metal-resistant strains of T. ferrooxidans can cope with unfavourable conditions of this kind (Tuovinen & Kelly 1974;Sugio et al. 1981;Baldi & Olson 1987), and grow well in the presence of high concentrations of these toxic cations. The metabolic activity of acidophilic T. ferrooxidans can also be influenced by carbonate compounds (Ishikawa et al. 1983;Khalid et al. 1987). It is obvious that carbonates buffer the acidic bioleaching solution and inactivate acidophilic bacteria (Wallace et al. 1976;Boseker et al. 1978). However, it is of industrial interest to investigate the concentrations of carbonates that affect the bioleaching process, since base metal ore deposits often contain carbonate gangue.
In industrial situations, in which tonnes of mineral ores are bioprocessed, a small percentage of carbonates might cause the failure or delay of bioleaching. Consequently, the carbonate content of ores might be an economic limiting factor in an industrial bioleaching plant. Sulphide ores from different mines in Italy: Funtana Raminosa, Monteneve and Campiano, contain valuable metals including Zn, Cu and Co which may possibly be recovered by biological methods. Preliminary bioleaching experiments with these ores were conducted in our laboratory for several months. The sulphides were inoculated with different strains of T.ferrooxidans isolated from the respective

Pyrite
The pyrite from the Campiano mine (southern Tuscany) was chosen because of its high Co and Zn content. Twenty kilograms of pyrite was ground to obtain 500 g of 250 to 500 pm grain size containing 1 .Ol% of total carbonates. The sample was split into two subsamples: one was treated with 20% (v/v) acetic acid solution to remove carbonates (0.01% carbonate residues). The purity of pyrite and type of carbonate gangue was determined by X-ray diffractometry (Siemens D500) using internal standards. The megacrystal from the Campiano mine was 98.1% pyrite, the remainder being 1.01% CaCO, as calcite, 0.1% Co, 0.065% Zn, and other trace metals (Table I).
Metal Ana&is Metal concentrations of Fe, Zn, Co, Ca and others (Table 1) were determined by holding 0.5 g of sample treated with 50 ml 6 M HCl both as original pyrite and as carbonate-free pyrite, at 90°C overnight. The sample was cooled to room temperature and made up to 100 ml. Metals were determined by flame atomic absorption spectrophotometry (AAS, Perkin Elmer model 5000). The standard deviation was 5.1% for Fe, 6.3% for Ca, 6.1% for Zn and 7.5% for Co. Metal analysis in the leaching solution of percolators and Erlenmeyer flasks was carried out by flame AAS after adding 2 ml cont. HNO, and 7 ml of double distilled water to 1 ml of sample.
Organism and Culture Conditions T. ferrooxidans strain 13661 was kindly supplied by G. J. Olson and grown in 9K medium (Silverman & Lundgren 1959). The bacteria were cultivated routinely in 50 ml of 9K medium in 250 ml Erlenmeyer flasks and agitated at 150 rev/min at 30°C or in 2.0 litres of 9K medium in a 2.5 litre Erlenmeyer flask equipped with a sterile glass tube plugged with glass wool, through which compressed air was passed to vigorously aerate the inoculated medium. When all FeSO, was oxidized, 2.0 litres of culture was harvested by centrifuging at 3600 x g for 20 min. The cells were washed twice with 5 mM HaSO,, and then inoculated into air-lift percolators or the Erlenmeyer flasks.

Standards
T. ferrooxidans ATCC 13661 used here is the same as that used in an interlaboratory bioleaching study (Bioleaching working group, in preparation) of a reference pyrite (NIST-8455; National Institute of Standards & Technology Gaithersburg, MD, USA). The average bioleaching rate of the reference pyrite using this strain, according to eight international laboratories, was 12.4 mg/l/h. The average value of our laboratory for three replicates was 7.8 mg/l/h f 1.1 (standard deviation).

Microbial Protein Analysis
The cell density in solution was determined by measuring protein content with a protein assay kit (BioRad). Bovine serum albumin (Sigma) solution was used as standard. Pyrite oxidation by cells was determined by measuring soluble iron.
Carbonate-free pyrite (20 g), original pyrite (1.01%) and mixed samples with intermediate carbonate content (0.25, 0.50 and 0.75%) were placed in percolators and immersed in 25 ml of 9K medium (minus FeSO,). Pulp density was 80%. All five percolators were inoculated with 0.25 ml of a suspension of a T. ferrooxidans culture containing 3.7 mg protein/ml. Two percolators with sterile pyrite were not inoculated for use as controls. All percolators were continuously aerated, with humidified sterile air from an air-compressor, for one year at room temperature.
The liquor (bioleaching solution) was sampled periodically (at intervals of 3 to 10 days) and analysed for Ca, Co, Fe, Zn and pII. At the outset, other metals were also measured: Ag, Cd, Cu, Mn, Ni and Pb. Losses from the removal of sample aliquots (4% of total volume) and evaporation were 20 to 25% per week. The volume of the liquid in the percolators was maintained by refilling with 5 mM HaSO,. This led to the progressive dissolution of carbonates.

Colloid Phase
The colloid phase was analysed by draining all the liquid from the percolators at the end of the experiment. The liquor was centrifuged at 3600 x g for 30 min to remove cells and inorganic particles. The supernatant was spun at 242,500 x g for 4 h with an ultracentrifuge.
An aliquot of the residue was analysed for metal content and another aliquot was analysed by X-ray diffractometry.

Salt Precipitate in Percolators
When the experiments had been underway for several weeks, a white precipitate was observed on the glass walls of percolators at the air-liquid interface. This compound was dried and analysed by X-ray diffractometry.
Erlenmeyer F/ask Experiments The bioleaching experiments were repeated in Erlenmeyer flasks (125 ml) for a period of 3 months at the optimal temperature (30°C) (Kelly & Harrison 1989). Pyrite (3 g) was immersed in 30 ml of 9K medium (minus FeSO,), with a pulp density of 10% (w/v). The carbonate content of the pyrite samples was: 0.01, 0.05, 0.10,0.20,0.30,0.40,0.60, and 1 .Ol% (original pyrite). The intermediate carbonate contents were obtained by mixing the original pyrite with different quantities of carbonate-free pyrite. The pyrite samples were inoculated with 0.3 ml of a dense T. ferrooxidans culture containing 4.4 mg protein/ml. Two uninoculated samples were used as experiment controls. One ml of solution was collected each week and centrifuged at 12,000 x g to remove particles. Then the supernatant was analysed for metals as described above. In this case evaporation losses were negligible, but sampling took 3.3% of the total volume. The initial addition of 30 ml of 9K medium to 3 g of pyrite was sufficient to dissolve up to 0.30% of the carbonates.

Percolator Experiments
The continuous additions of acid to the percolators was not sufficient to completely dissolve the carbonate residue except in the pyrite sample with 0.01% of calcite Pyrite bioleaching affected by carbonate gangtre content x a  (pulp density 80%). The bioleaching process was first followed by determining metal concentrations in solution and by measuring pH values (Figure 1). Later only the concentration of Fe released into solution was used as an index of bioleaching activity. The pyrite containing 0.01, 0.25 and 0.50% of calcite was leached by T.ferrooxidam strain ATCC 13661, but not the samples with 0.75 and 1.01% of carbonates. For pyrite with a given calcite content, the soluble metals, Fe, Zn and Co, were first detected in solution at the same time, but for different calcite contents. The start of the bio-oxidation was directly related to carbonate concentration.
The lag phases detected were: 13 days for 0.01% of calcite content, 38 days for 0.25% and 190 days for 0.50%. The maximum concentrations of Fe, Zn and Co in solution were reached after 80, 120 and 315 days, respectively, after inoculation and ranged from 4600 to 7500 mg Fe/l, 45 to 62 mg Zn/l and 14 to 18 mg Co/l (Figures 24). In uninoculated samples and in those with 0.75 and 1.01% of calcite, no metals were detected in solution. In the latter, pH was still neutral after one year.
For soluble Fe (Table 2), the bioleaching rate was closer (5.05 to 14.7 mg/l/h) to the rates reported in the interlaboratory calibration (12 mg/l/h). The bioleaching rate for Fe, Co and Zn (Table 2) increased significantly with the percentage of dissolved calcite. There is probably an increase in CO, concentrations from calcite dissolution in the interstitial water of the pyrite bulk. This excess CO, would favour the metabolism of the chemolithotroph T. ferrooxidans, which uses this gas as a carbon source (Ingledew 1982). The concentration ratio of soluble metals was different from the metal content of pyrite ore (Table 3). Zinc is more soluble than Co and Fe, because it is probably  va 1985). Cobalt on the other hand, replaces Fe in the pyrite lattice.
The dissolution pattern of Ca was completely different from that of other metals ( Figure 5). First, the 9K medium and then the continuous refilling with acid water (pH 2) dissolved calcite. In the first weeks, dissolved Ca concentrations did not depend on the calcite content of the pyrite, and reached 350 to 420 mg/l in all percolator liquors. This range depended on continuous sampling, evaporation and refilling with acid solution, Conversely, the time required to entirely dissolve the calcite by acid additions depended on the content of calcite in the pyrite ore. T. ~~FFOOX&Z~S began to grow in the mineral bulk after the calcite had completely disappeared and the optimal growth pH was attained. However, the Ca2+ in solution remained relatively constant, until the soluble metal concentrations  Determination of total metals in the colloid phase (1.08 mg/ml), in the presence of carbonate-free pyrite, showed high concentrations of Fe (157 mg/g dry wt), Zn (1.25 mg/g) and Co (0.37 mg/g). Thus, 5.6% of Fe was in colloidal form and the remaining 94.4% was soluble; 4.2% of Zn and 4.4% of Co occurred as colloids. Colloidal co-precipitation of metals may take place, resulting in the trapping of small particles in the ore bulk, but accounts only for about 5% of the decline in metal concentrations in the later stage of the leaching bioprocess (Figures 2 to 4), with respect to a major metal loss due to the continuous liquor sampling. This decline was mostly accounted for by liquor sampling and was observed to depend on the number of samples taken.

Flask Experiments
This experiment was performed to confirm the effect of calcite on pyrite bioleaching. Only Fe, Ca and titratable acidity were determined in the liquor as indexes for bioleaching and calcite dissolution. Calcium dissolution data is not reported because it followed the same pattern as in percolators. In the flask experiments, the lag phase spans were significantly different from those in the percolator study and were characterized by Fe  concentrations in the liquor. In fact, the bioleaching of pyrite occurred more rapidly. The lag phases determined from Fe solubilization curves ( Figure 6), were 2 days for 0.01% calcite content, 15 days for 0.05 to O.lO%, 20 days for 0.20 to 0.30%, 32 days for 0.40%, and 65 days for 0.60%. The original pyrite (1.01% calcite) was not oxidized by T. ferrooxidans strain ATCC 13661 after 90 days of incubation.
The Fe concentrations were inversely correlated to calcite content in pyrite. The maximum concentration of Fe (1700 mg/l) in solution occurred in the sample without calcite, whereas in the pyrite with more calcite, Fe concentration dropped to 250 mg/l, as, for example, with 0.60% of calcite ( Figure 5).
The Fe bioleaching rates were lower than in the percolator study, ranging from 2 mg/l/h in the sample with carbonate-free pyrite down to 0.36 mg/l/h in the sample with 0.60% of carbonate ( Figure 6). The acid production rate from pyrite bio-oxidation also diminished in the bioleaching liquor depending on the calcite content of the pyrite ores (Figure 7).

Discussion
Thiobacillrrs ferrooxidans strain ATCC 13661 can oxidize pyrites provided they contain only traces of calcite. However, this acidophilic micro-organism can survive at neutral pH for a long time, and as soon as conditions are favourable, it starts oxidizing the mineral bulk. The survival of the organism at neutral pH is not surprising because of its isolation from high pH natural environments. Olson et al. (1979Olson et al. ( , 1981 regularly f ound an abundance of acidophilic iron-and sulphur-oxidizing autotrophs around alkaline pyrite-containing coal mines in southern Montana (USA), despite the high pH of drainage waters. In addition, strains of T.ferrooxidans have been isolated by us from calcareous sulphide deposits in Sardinia (Italy). T. ferrooxidam probably survives in a state of dormancy or is localized in acidic microzones (Van Voast & Hedges 1975). The results of the two studies were contradictory for the following reasons: (1) in the percolator system, the bioleaching rates of Fe were higher and increased with the amount of dissolved calcite, the same was true of Zn and Co; (2) in the flask experiments, the Fe bioleaching rate was about half that in the percolator system for minimum calcite content (up to 0.15%), and about 30 times less when the calcite content was above 0.25%; (3) on the other hand, the lag phase spans for T. ferrooxidans in the flask study were significantly shorter. These differences in metal extraction between the percolator and flask experiments can be explained by the different characteristics of the two systems and depend on if the acid solution is added, on different sites of gypsum precipitation (outside or in the mineral bulk), and on the final pH of the bioleaching solution. In the percolators, a few ml of acidified water per week were added to maintain the constant volume of the leaching liquor, and the larger lag time is linked to the more gradual dissolution of calcite, which keeps the micro-organism in a dormant state. In the flask study, the larger amount of acidified medium added at the beginning dissolved more calcite (0.40%) and the lag phase spans were consequently shorter.
The bioleaching rates of metals tended to be higher in percolators with pyrite ore containing more dissolved calcite, probably because the CO, generated by dissolving carbonates in the interstitial water of the pyrite bulk is utilized by the chemolithotroph T. ferrooxidans as a carbon source. In addition calcite is transformed to gypsum which precipitated on the percolator wall at the water-air interface, outside the mineral bulk. The optimal pH to bio-oxidize the mineral ore is also reached in the liquor.
In the flasks, on the other hand, the bioleaching rates decreased with calcite content because the newly formed gypsum precipitated in the mineral bulk. The pyrite assumed a milky-brown colour and the bio-available surface of the mineral was diminished.
The acidity in the liquor is not optimal, especially in those with high calcite content, and consequently affects the Fe concentrations, which tend to precipitate from the liquor as jarosite.
Regarding the biological extraction of Co and Zn, Zn was more readily extracted than Co.
The removal of carbonates could provide a solution in cases in which pyrite contains high concentrations of more soluble strategic metals. Chemical or biological pretreatment to remove the carbonic gangue could be considered in cases of promising good metal recovery from sulphide ores.