Fate and transport of arsenic in uranium mine tailings: Rabbit Lake Mine, Saskatchewan, Canada

The mineralogical controls on the long-term stability of secondary arsenic precipitates formed from iron-rich hydrometallurgical solutions were studied in arsenic-rich uranium mine tailings deposited in the Rabbit Lake in-pit tailings management facility (RLITMF), northern Saskatchewan, Canada. Core samples (n = 42) of tailings were collected to a depth of 70.6 m below the surface of the tailings near the centre of the pit. Total arsenic and iron concentrations in iron-rich samples of the mine tailings ranged from 56 to 17,300 mug/g and 10,000 to 63,200 mug/g, respectively (Fe/As molar ratios of 5.3 to 303). Synchrotron-based X-ray absorption spectroscopic studies of tailings samples, fresh mill precipitates, and reference compounds showed that the arsenic in iron-rich areas of the tailings existed as As5+ and showed that it was adsorbed to 2-line ferrihydrite through inner sphere bidentate linkages. Furthermore, under the conditions in the RLITMF, the 2-line ferrihydrite did not undergo any measurable conversion to more crystalline goethite or hematite, even in tailings discharged to the RLITMF 10 years prior to sampling. Stability field diagrams generated from pH, Eh, and temperature measurements on tailings samples (mean values of 9.36, +235 mV, and 4.4°C, respectively) supported arsenic and iron in the tailings as existing in the As5+ and Fe3+ states. The concentration of arsenic in the tailings pore fluids ranged from 0.24 to 140 mg/l. The variability in the concentration of dissolved arsenic was inversely proportional to the Fe/As molar ratio in the tailings solids. Furthermore, there was no correlation between the age of the tailings and arsenic concentrations in the tailings pore fluids. A three-month long mill-scale hydrometallurgical experiment (2,700 m3 of effluent treated/day) was conducted over a pH range 1-11 to develop a thermodynamic database for the dominant mineralogical controls on arsenic in the mill and the resulting mill tailings. The arsenic concentrations in the uranium barren leach process waste solutions (raffinate) ranged from an average of 526 mg/l at pH = 1.0 (initial) to an average of 1.34 mg/l at pH = 10.8 (discharge as tailings). Geochemical modeling of chemistry data (PHREEQC) showed arsenic solubility is controlled by the formation of scorodite from pH 2.4-3.1, that scorodite is unstable (releasing arsenic back in to solution) above pH 3.1, and arsenic adsorption to the surface of 2-line ferrihydrite was the dominant control on the solubility of arsenic from pH 3.2-11.0. Minor alterations to the thermodynamic properties of arsenite and arsenate adsorption to 2-line ferrihydrite improved the fit between measured mill-scale and modeled concentrations for pH 3.2-11.0. Three-dimensional, reactive multi-component transport modeling (MODFLOW and MT3D99) of the RLITMF and the surrounding groundwater regime was conducted using geochemical data collected during a field monitoring program, laboratory-based diffusion cell experiments, and data from the literature to quantify the fate and long-term (10,000 year) transport of arsenic. Results showed that adsorption of arsenic to the tailings (or filter sand located adjacent to the tailings) and diffusive transport of dissolved arsenic in the tailings should reduce the source term concentration of arsenic to between 39 and 70% of the initial concentrations over the 10,000 year simulation period. Based on these simulations, the arsenic concentrations in the regional groundwater, 50 m down gradient of the tailings facility, should be maintained at background concentrations of 0.001 mg/l over the 10,000 year period. These findings indicated that the engineered in-pit disposal of U mine tailings should provide long-term protection for the local groundwater regime from arsenic contamination, provided there are sufficient adsorption sites in the tailings management facility.