Résumé
Acid Mine Drainages (AMD) containing arsenic (As) are ubiquitous and represent a threat for aquatic ecosystem and human health. In the perspective of dealing with legacy mine sites, the development of passive or semi-passive affordable, effective and sustainable process for the treatment of these effluents is essential. However, most of these processes are still running at laboratory conditions. The present study is focused on the evaluation in field conditions of two semi-passive treatment devices for the arsenic and iron removal from high-As AMD. These devices exploited natural attenuation process, which involves bacterial iron and arsenic oxidation followed by As co-precipitation with ferric(hydroxy)sulfates. Iron oxidation rate was optimized by forced aeration. Two support media maximizing the surface available for the Fe and As-oxidizing biofilm were compared: plastic support Biofill® (device “PS”) and wood/pozzolana mixture (device “WP”). The performance of the two devices of 1 m3 each were evaluated on the AMD from the abandoned Carnoulès mine in Southern France. This AMD exhibited pH 3.2-5.1, 42-102 mg/L As in the As(III) form and 430-1000 mg/L Fe(II). The treatment was carried out under controlled flow rates in discontinuous operation during seven periods from August 2019 to September 2020. The physico-chemistry of inlet and outlet waters, the concentration of total dissolved Fe and As and their redox speciation were monitored and interpreted by geochemical modeling. The performances of the devices relative to iron and arsenic oxidation and precipitation rates were evaluated. With both “PS” and “WP” devices, we reached 92 ± 6 % iron oxidation, 43 ± 11 % iron removal and 67 ± 10 % arsenic removal. Forced aeration and the use of support media improved performances compared to our previous passive field-scale experience, where the clogging of the systems was an important limitation. No significant differences were evidenced between the two support media in stationary phase (Mann-Whitney p >0.05). However, the WP device showed a better robustness towards operational fluctuations. The sludge of both PS and WP devices contained jarosite besides amorphous Fe-As phases since the earlier period of the monitoring. It contained up to 90 g/kg (dry wt.) of arsenic mainly under As(V) form (60-90%). This is advantageous for sludge management since As(V) is more stable than As(III). These results demonstrated the feasibility of our semi-passive treatment under real field conditions. Our study provided fundamental and technical advances for a full-scale treatment facility adapted for an As-rich AMD.