Résumé
Acid Mine Drainage (AMD) is an undesired product of the weathering of sulfide minerals present in ores deposits and mining wastes. These leachates are responsible for extreme pollution of the freshwater ecosystems. When produced from arsenic-rich tailings, AMD are concentrated in highly toxic arsenic (As) worsening their adverse effects on the environment. It is then essential to develop cost-effective methods to treat this pollution. Biological treatment based on bacterially mediated Fe-oxidation followed by Fe-As co-precipitation proved to be a promising strategy. The present study aimed at the optimization of a field-scale system for the treatment of arsenic-rich AMD from the ancient Pb-Zn Carnoulès mine in Southern France. Twenty years of monitoring of this AMD provided a fine comprehension of the biogeochemical processes leading to As attenuation at this site. Two field units were designed with the objective to maximize the surface available for the Fe- and As-oxidizing biofilm growth. Two filling materials were compared: plastic support (PS, specific surface 160 m2/m3) and wood/pozzolana mixture (WP, 80%-20% of the mass respectively, specific surface 400 m2/m3). Forced aeration was provided so that dissolved oxygen was not a limiting factor. The field pilots were installed on July 2019 and fed by As-rich AMD water ([Fe] = 500-900 mg/L, [As] = 50-100 mg/L, pH = 4.3 ± 0.8) during five months at controlled flow rate (hence controlled Hydraulic Retention Time, HRT). Water samples were collected periodically at the inlet and outlet of the systems for chemical characterization, including the determination of dissolved Fe(II) concentrations, total dissolved Fe, As and S concentrations and redox arsenic speciation. Our results showed that a steady state regarding iron oxidation was reached more rapidly in the WP pilot compared to the PS pilot where a latency phase was observed. The two field pilots showed an average Fe oxidation efficiency of 92±3% (PS) and 97±1% (WP) for a theoretical HRT of 17.5 h (PS) and 19.3 h (WP). Fe precipitation efficiency reached 38±5 % (PS) and 42±5 % (WP). As precipitation efficiency reached 60±2 % (PS) and 73±5 % (WP). The removal rates (in mol. L-1.s-1) showed no significant differences between the two pilots over the studied period (p >0.05). The abatements obtained in this study were improved compared to those obtained with a previous field pilot tested on Carnoulès site that was fully passive (without forced aeration) and filled with river sand material. Further work is needed to characterize the diversity and the activity of the bacterial community colonizing the pilots, and to assess the possible effect of the filling material on these communities. A particular attention will be paid to arsenic oxidation activity since it will contribute to the formation of more stable and less toxic As(V)-rich precipitates. Long term monitoring of the two pilots is necessary to confirm the robustness of the systems. We can hypothesize that the pozzolana-wood system would show a better robustness towards possible environmental and operational fluctuations.