Oxidation characteristics, acid neutralization, secondary minerals, and trace elements associated with pyrrhotite oxidation in historical waste rock

Abstract

The Detour Lake Mine is an open pit, greenstone-hosted gold mine in Ontario, Canada. Mining produced waste-rock piles that were constructed from 1983 – 1999. Redevelopment and expansion of the open pit required the excavation and relocation of the waste-rock piles, providing an opportunity to collect samples of waste rock that had been weathering in situ for 30 years. Samples of this weathered waste rock and extracted porewater were analyzed for solid-phase and aqueous geochemistry, characteristics of sulfide-mineral oxidation, stable calcium isotopes (40/44Ca), and secondary minerals.

Characterization of weathered waste rock suggest the waste rock was potentially acid generating or of uncertain acid-generating potential and had depleted neutralization potential compared to unweathered samples. Porewater was circumneutral with measurable alkalinity. Mineralogical observations identified pyrrhotite as the sulfide mineral with greatest abundance and oxidation characteristics. Pyrite was present in abundances of approximately half that of pyrrhotite; chalcopyrite was present in lesser amounts and trace amounts of sphalerite and pentlandite were identified. Pyrrhotite was typically strongly altered, with intensity ranging from thick rims surrounding intact-cores to complete replacement by iron-(oxyhydr)oxides. Pyrrhotite, pyrite, chalcopyrite and pentlandite contained trace elements; no sphalerite grains were encountered during the measurements. Of the trace elements measured, Ni occurred in the highest concentrations in pyrrhotite, pyrite, and alteration rims associated with these sulfide minerals. Nickel concentrations in the associated porewater were elevated and sorption sites were calculated to be saturated, suggesting the oxidation of sulfide minerals released Ni that accumulated in porewater, with some attenuation by sorption to iron-(oxyhydr)oxide phases. Synchrotron studies identified Ni in alteration rims associated with an oxidizing pyrrhotite grain, in addition to phases consisting of mixed oxidation states of Fe and S.

An additional synchrotron study focused on iron speciation at 1 µm intervals across transects of three partially oxidized pyrrhotite grains to better understand how pyrrhotite oxidation proceeds. Grains were selected from an unweathered sample, a sample of intermediate weathering characteristics, and a sample with more advanced weathering characteristics. Analysis of the spectral features suggests that at the 1 µm scale oxidation products were mixtures of Fe-(oxyhydr)oxides, Fe-(hydroxy)sulfates, and Fe-depleted sulfides. Measurements on visually unoxidized pyrrhotite grains in the samples suggested variable bonding arrangements of Fe among the grains studied, either due to the mineral structure or early-stage oxidation that were not apparent in other features of the spectra. The pyrrhotite standard and some ferrous iron [Fe2+] standards had discordant spectral features; ratios of ferric iron [Fe3+] to [Fe2+ + Fe3+] calculated for the transect spots exhibited a clustering at a Fe3+ component of 0.2. Together, these results suggested oxidation from Fe2+ to Fe3+ may retain characteristics of the Fe2+ pre-edge spectral feature until a threshold component of Fe3+ is exceeded.

Existing datasets of porewater and solids chemistry were evaluated from Detour Lake Mine and Diavik Diamond Mine to attempt to discriminate the contributions of carbonate minerals and non-carbonate minerals to acid neutralization based on Ca concentrations. Diavik waste rock had low concentrations of calcite and Ca-bearing minerals, and periods with no measurable alkalinity in the drainage water, in contrast to Ca-rich Detour waste rock and porewater samples with measurable alkalinity. The porewater from both Diavik and Detour had higher molar proportions of Ca to major cations than the solids, indicating preferential release during weathering and/or preferential retention in the aqueous phase. The formation of the secondary Ca-bearing phase gypsum, but not Mg- or K-bearing phases, affected the molar proportions of Ca in porewater. Calculations based on measured alkalinity and concentrations of calculated dissolved inorganic carbon provided a lower bound for concentrations of dissolved Ca derived from carbonate-mineral dissolution. Results were consistent with expectations based on Diavik and Detour lithology and measured porewater alkalinity trends.

A small number of samples were analyzed for stable Ca isotopes. This is the first study believed to apply stable Ca-isotopes to mine-waste systems. No trends were discernable between the Ca-isotope values and alkalinity or ratios of major cations, but the dataset was small. Two-component mixing using assumed carbonate and non-carbonate endmembers appeared to over-ascribe Ca to carbonate minerals in the Diavik samples, indicating endmembers may not have been representative, and/or another mechanism was affecting porewater values. Mixing calculations could not be completed on the Detour samples because Ca-isotope values of the endmembers were lower than the porewater values, suggesting a confounding mechanism, likely gypsum precipitation/dissolution, affected the Ca-isotope ratios. Calculated fractionation factors were consistent with previously reported fractionation factors for gypsum.

Secondary minerals associated with sulfide-mineral oxidation were characterized by automated quantitative mineralogy (mineral liberation analysis, MLA). Grain sizes, modal abundances and mineral habits of gypsum produced by MLA suggested geochemical conditions were spatially and temporally variable in the historical waste-rock pile. The secondary mineral jarosite was inferred to be present by MLA and by X-ray diffraction analysis, though porewater was calculated to be undersaturated with respect to this phase. Calculated mineral associations and grain habits produced by MLA suggested that jarosite formed in microenvironments associated with oxidizing sulfide minerals. MLA calculates elemental concentrations based on idealized mineral formulae in the database, and comparisons to results from four-acid digestions of solid samples revealed that the MLA results under-represented concentrations of some trace elements by an order of magnitude. Modal abundances provided by MLA and concentrations of trace elements measured by electron-probe microanalysis were used to calculate concentrations of trace elements in sulfide minerals and their alteration rims. Combining these techniques resulted in an increase of up to 170% (relative percent difference) of some trace elements compared to MLA, but concentrations remained lower than the four-acid digestion results. The discrepancy may be (i) an artefact of the MLA resolution, which may not capture mineral grains < 1.5 µm; (ii) the MLA sample size, which considers a much small number of particles compared to four-acid digestion; and/or (iii) the incorporation of trace elements as impurities in silicate minerals, which are digested by the four-acid method.

This research provided new insights into the complex processes of pyrrhotite oxidation that affect solid phase and aqueous phase geochemistry of mine-waste systems. Results of this research illustrate that integrating standard and novel bulk and microanalytical techniques will contribute to more robust predictive models of mine-drainage chemistry and, therefore, reduced environmental risk from mine drainage.