Environ. Sci. Technol. 2003, 37, 158-164
The Effect of Natrojarosite Addition to Mine Tailings J A S N A J U R J O V E C , * ,† C A R O L J . P T A C E K , †,‡ DAVID W. BLOWES,† AND JOHN L. JAMBOR† Department of Earth Sciences, University of Waterloo, Waterloo, Ontario N2L 3G1, Canada, and National Water Research Institute, Environment Canada, Burlington, Ontario L7R 4A6, Canada
An increasingly common practice for metallurgical plants is to discard their wastes by combining them with mine tailings prior to disposing the blended material to a containment facility. This practice has occurred since 1985 at the Kidd Creek tailings impoundment, where natrojarosite, a waste produced from the adjacent Zn refinery, is combined with mine tailings and is deposited in a single impoundment. To assess the environmental impact of the co-disposal, a laboratory column experiment was conducted. The column material was flotation tailings from the Kidd Creek site containing 3 wt % natrojarosite residue. Dilute sulfuric acid was passed through the column to simulate the acid generated in the unsaturated zone of the tailings impoundment. The results of this experiment were compared to the results of a previous experiment conducted on unamended flotation tailings. The results showed that the effluent from the column containing the natrojarosite-bearing mixture had a faster decrease in pH, earlier increases in the concentrations of dissolved metals such as Pb and Cd, and a greater persistence in effluent metal concentrations such as Pb, Zn and Ni. To prevent the observed enhanced release of dissolved metals from mine waste disposal areas, natrojarosite should not be co-disposed with tailings.
Introduction
potential use because of its commercially low Fe content (3). The waste also typically contains impurities, mostly heavy metals, which may be leachable over time (5). Thus, disposal to a dedicated containment facility, such as a lined pond, is common practice. An alternative is co-disposal with concentrator tailings, as has been done since 1985 at the Kidd Creek metallurgical plant near Timmins, Ontario (6). At Kidd Creek, 10000 tones of ore are processed daily, and the tailings are combined with the natrojarosite-dominant wastes from the adjacent metallurgical plant prior to discharge to a 1200ha impoundment. The potential environmental consequences related to the co-disposal were not investigated prior to implementing the practice. However, thermodynamic calculations (7) and laboratory studies (8, 9) suggest that natrojarosite is unstable above pH 3.5 and under the moderate to low Eh conditions that prevail in many tailings impoundments. A field study initiated in 1991 to assess the effect that co-disposal has had on the pore-water geochemistry of the Kidd Creek tailings indicated the presence of elevated concentrations of Fe, Na, K, Zn, Pb, As and SO4 at shallow to intermediate depths in the impoundment. Because natrojarosite was co-disposed with tailings only since 1985 these higher concentrations suggest that natrojarosite has dissolved (6). To evaluate further the effects of the co-disposal at Kidd Creek, a column experiment was conducted over a period of eight months. The column contained the natrojarosite refinery waste and flotation tailings in a proportion similar to that co-disposed to the main impoundment. The objectives of the current study were (1) to determine if natrojarosite dissolves under the geochemical conditions prevalent in tailings impoundments, (2) to determine if the heavy metals that are associated with the natrojarosite residue are mobile under such conditions, and (3) to determine the effects of natrojarosite on the buffering capacity of the tailings. A second column experiment with the flotation tailings, specifically directed toward determining acid neutralization and metal-attenuation mechanisms occurring in the Kidd Creek flotation tailings, was described previously (10). Comparison between the observations of the current experiment and the results of the experiment presented previously (10) aids in the identification of the changes in water chemistry that can be attributed to the effect of the jarosite waste.
Iron is an abundant impurity component of ores processed in most hydrometallurgical operations. The elimination of this Fe is an important consideration in nearly all hydrometallurgical circuits (1). To remove the Fe, the most widely used technique is the Jarosite Process, whereby Fe in the leach circuit is precipitated as a jarosite, most commonly natrojarosite [NaFe3(SO4)2(OH)6]. The popularity of the process, which has been reported to be involved in more that 80% of world zinc production, is attributable to its simplicity and low cost (2, 3). The occurrence, crystallography and thermodynamics of the jarosite group of minerals have been reviewed by Stoffregen et al. (4). Most processing circuits use natrojarosite, which most commonly occurs as a solid solution with jarosite (K) and hydronium jarosite (H3O) members of the jarosite family of minerals. The chief environmental concern associated with the Jarosite Process is the disposal of the jarosite residue, which is produced in large volumes and has little
Prior to the 1970s, electrolytic processing of Zn concentrates containing significant proportions of Fe was hampered because of concurrent electro-deposition of Fe with the Zn (5, 11). In the Jarosite Process, Zn sulfide concentrates are roasted in air to produce a calcine that consists mostly of Zn oxide, Zn ferrite, and a small amount of Zn silicates (3, 12, 13). The calcine is then leached with H2SO4 to dissolve ZnO and produce a ZnSO4 solution that is almost Fe-free (3). The potentially significant loss of the Zn present in residual ferrite and silicates is avoided by subsequently dissolving this residue in hot concentrated H2SO4. To remove Fe by natrojarosite precipitation, pH is raised and NaOH is added to the leach solution. The success of the Jarosite Process in the zinc industry has led to applications to the hydrometallurgy of Cu, Ni, Co and other metals (14).
* Corresponding author e-mail:
[email protected]. † University of Waterloo. ‡ National Water Research Institute, Environment Canada.
Fresh unoxidized tailings were collected from the Kidd Creek flotation concentrator. The mineralogy of the tailings sample that was used in the experiment was determined by Jambor
158
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 1, 2003
The Jarosite Process
Materials and Methods
10.1021/es025589b CCC: $25.00
2003 American Chemical Society Published on Web 11/22/2002
et al. (15). The experimental sample consisted of 15 wt % sulfides and 8 wt % carbonates and Fe oxides and hydroxides; the remainder was mainly quartz (49 wt %) and chlorite (24 wt %). Electron microprobe analyses and X-ray studies indicated that calcite [CaCO3], dolomite [CaMg(CO3)2], ankerite [Ca(Fe,Mg)(CO3)2] and siderite [FeCO3] are present in the tailings but are rarely of end-member composition. The second component, which is the natrojarosite residue, was collected at the Kidd Creek zinc refinery. Mineralogical investigation (16) of the Kidd Creek jarosite impoundment, which was the dedicated repository prior to the initiation of co-disposal, showed that the main components are Znbearing natrojarosite [Na0.75(H3O)0.24K0.01Fe3(SO4)2(OH)6], in which Na is partly replaced by H3O, and a spongy-textured Zn ferrite. The residue also contains minor to trace amounts of quartz [SiO2], cassiterite [SnO2], sphalerite [(Zn,Fe)S], hematite [Fe2O3], goethite [FeOOH], cryptomelane [KMn8O16], native sulfur, celestine [SrSO4], barite [BaSO4], anglesite [PbSO4], an unidentified MnO2 phase, and traces of other phases. A portion of the unamended tailings sample was first thoroughly mixed and coned and quartered. One-quarter was used in the experiment described previously (10), a second quarter was used in the experiment described here and remaining quarters were retained for further study. The column, referred to as the natrojarosite column, was packed with the blend of unamended tailings and 3 wt % natrojarosite residue. This blend is representative of the material that is co-disposed in the Kidd Creek impoundment (17). Thus, the only difference between the natrojarosite column and the column discussed by Jurjovec et al. (10) was the presence of the 3 wt % natrojarosite residue in the natrojarosite column. The laboratory apparatus consisted of an input-solution reservoir, pump, and a sampling cell (10). The natrojarositetailings mixture was loaded into a clear acrylic column, 9.0 cm in internal diameter and 10 cm long. The flow direction was upward, from bottom to top, to ensure removal of the gas initially present amidst the solids. The column was saturated with a background solution, which was equilibrated with respect to calcite and gypsum to prevent removal of these minerals during the preparatory stage. After saturation, a conservative tracer test was conducted to determine the column flow parameters. To simulate acidic pore-water generated in the unsaturated zone of tailings impoundments, 0.1 M H2SO4 was used as the input solution (pH ) 0.99) so that the experiments could be completed in a reasonable length of time. Samples of column effluent were collected versus time. Other details of the experimental procedure were reported previously (10). The pH and Eh were measured in sealed cells. The pH was measured using a combination glass electrode (Orion Sure-Flow Ross 8165BN) calibrated with pH 4 and 7 standard buffer solutions, and checked with pH 1.00 buffer solution (Fisher). The Eh was determined using a platinum combination electrode (Orion Model 9678BN), checked against ZoBell (Nordstrom, 1977) and Light solutions (Light, 1972). Alkalinity was determined by titration of filtered samples using HACH digital titrator, indicator, and standardized sulfuric acid solutions. Concentrations of Fe were determined colorimetrically using the Ferrozine method (Gibbs, 1979). Concentrations of SO4 were determined by ion chromatography, and concentrations of Na, K and Si by flame atomic absorption spectroscopy. The concentrations of Al, Ba, Be, Ca, Cd, Co, Cr, Cu, Li, Mg, Mn, Mo, Pb, Sr, V, and Zn were determined by inductively coupled plasma optical emission spectroscopy. Upon completion of the experiment, two samples were collected from the upper third of the column, which is where the largest mass of residual minerals was expected because the flow direction was upward. A sample of the starting
TABLE 1. Summary of Major Reactions Occurring in the Column Experiment mineral
Log K
+ + 0.4Mg Fe + CO3 ankerite-dolomite Fe2+ + CO32 T siderite Al3+ + 3H2O - 3H+ T gibbsite Ca2+ + SO4 T gypsum Fe3+ + 3H2O - 3H+ ‚ ferrihydrite Fe3+ + 3H2O - 3H+ ‚ goethite 0.75Na + 0.24H3O + 0.01K +2SO4 + 6OH T natrojarosite Ca2+
a
0.6Fe2+
Reference 23.
b
2+
2-T
18.16a 10.45b -8.11c 4.58c -4.891c -1.0c -10.5a
Reference 44. c Reference 19.
material also was retained for analyses. Two polished thin sections of the starting material and two of the leached material were examined by optical microscopy and by scanning electron microscopy (SEM) with a linked energydispersion system. The pore-water geochemistry was interpreted with the assistance of the equilibrium geochemical speciation/mass transfer model MINTEQA2 (18). The thermodynamic database was adopted from the speciation model WATEQ4F (19). Solubility data for siderite (20, 21), FeH3SiO4(aq) (22), ankeritedolomite (23), natrojarosite (23) and Co species (24) were added. The degree of saturation is expressed as the saturation index (SI), where SI is equal to the difference of logarithms of ion activity product and solubility constant (SI ) log IAP - log Ksp). A SI value of zero indicates equilibrium, a negative value undersaturation, and a positive value supersaturation.
Results and Discussion At many field sites, acidic pore water eventually reaches the saturated zone, which was simulated in the experiment. Pore waters with very low pH have not been observed at the Kidd Creek site, at which tailings are currently being deposited; thus, exposure to weathering has been minimal. Although, pore waters at pH
