Algal/bacterial membrane bioreactor for bioremediation of hazardous landfill leachate containing 1,4-dioxane: An...

This study developed an artificial neural network (ANN) model to predict the 1,4-dioxane removal efficiency from hazardous landfill leachate treated by an algal/bacterial symbiosis system. Three algal/bacterial membrane bioreactors were operated in parallel at a hydraulic retention time (HRT) range of 6-24 h, addressing the contribution of enzymatic activities and microbial communities in 1,4-dioxane detoxification. The achieved 1,4-dioxane removal efficiency of 82.8 % ± 3.2 % was supported by secreting extracellular polymeric substances (55.4 ± 2.9 mg/g), alcohol dehydrogenase (1.3 ± 0.03 U/mg), aldehyde dehydrogenase (0.9 ± 0.01 U/mg), and dehydrogenase (2.2 ± 0.04 U/mg) at 12 h-HRT. The genera Thioclava, Afipia, and Mycobacterium showed relative abundances of 7.4 %-7.9 %, 2.9 %-6.5 %, and 4.5 %-9.8 %, respectively, producing specific enzymes required for the aerobic metabolic and co-metabolic degradation of 1,4-dioxane into 2‑hydroxy‑1,4-dioxane and other intermediates. Additional batch experiments and kinetic studies were conducted to reveal the removal mechanisms of 1,4-dioxane in algal/bacterial biosystems and the uptake, biosorption, and assimilation of the generated metabolites in algal biosystems. An ANN model explained > 50 % of the 1,4-dioxane data variability and justified the effect of organic load variation on algal/bacterial activities. Future studies should improve the ANN prediction accuracy by including more inputs (e.g., pH, temperature, and membrane fouling) and optimizing the hidden layer number.