Removal of micropollutants in Moving Bed Biofilm reactors (MBBRs): Microbial‐diversity‐and‐functional‐relationships

Numerous pollutants such as pharmaceuticals and personal care products are continuously released into municipal wastewater treatment plants (WWTP). Present at concentration of nano- to milligram per liter, they are defined as micropollutants. Micropollutants are only partially removed, possibly due to design and operational limitation of conventional WWTP. Eventually, micropollutant parent compounds and transformation products are discharged into receiving water bodies, possibly causing acute and chronic toxic effects on aquatic organisms even at very low concentrations. Therefore, research currently focuses on the enhancement of conventional WWTPs via physical-chemical and biological treatment processes. Biofilm-based treatment processes, such as the Moving Bed Biofilm Reactor (MBBR), were shown to harbour bio-catalytic potential that can enhance the biotransformation of a number of micropollutants compared to conventional activated sludge. In MBBRs, biofilm grow on plastic carriers kept in suspension in the reactor basin via mechanical mixing or aeration, offering a suit of benefits, amongst all comparably small footprint. Despite few existing evidences in aerobic MBBR, an in-depth understanding of the fate of micropollutants in such systems under different operational conditions is still required. In this context, this PhD thesis investigated different optimization strategies using MBBRs towards the removal of 23 commonly detected micropollutants (i.e., pharmaceuticals) in municipal wastewater. Specifically, I studied the impact of (i) biofilm thickness on the diffusion, sorption and biotransformation of the selected pharmaceuticals in nitrifying MBBR; and (ii) of organic carbon quality and availability on micropollutant biotransformation in anoxic pre- and post-denitrifying MBBRs. In both case, the influence of (i) and (ii) on the microbial activity (nitrification and denitrification) and microbial community composition and diversity were investigated. The existence of possible relationships between microbial diversity (analyzed via 16S rRNA amplicon sequencing) and biotransformation of micropollutants was evaluated to investigate which microbial processes and factors underlay the removal of micropollutants. The PhD objectives were evaluated in long- and short-term experiments in three laboratory- scale MBBR systems for pre-denitrification (MBBR1), nitrification (MBBR2) and post-denitrification (MBBR3). Biokinetics of nitrification, denitrification and micropollutant biotransformation rate constants (kBio, L g-1 d-1) were estimated through batch experiments using Activated Sludge Models (ASMs) and ASM for Xenobiotics (ASM-X), respectively. In the pre-denitrifying MBBR1 study, denitrification, biotransformation of micropollutants and microbial community were evaluated in three-stage (S) and single-stage (U) MBBR configurations. The three-stage configuration produced a prolonged exposure of the biofilm to a gradient of organic carbon loading and complexity, leading to a significant differentiation of denitrification and biotransformation kinetics in the three MBBR sub-reactors. The highest and lowest biotransformation kinetics were found in the first and the last stage, respectively (up to 4-fold decrease for selected compounds), suggesting a possible a correlation of micropollutant biotransformation with denitrification rates. The long term-operation with carbon availability and complexity gradient led to higher (p<0.05) biodiversity in the three-stage system, with a more diverse and even microbial community in the last stage. Specific taxa such as Candidate division WS6 and Deinococcales were selected in S, possibly due to oligotrophic conditions occurring in the last reactor stage. The influence of biofilm thickness was studied in nitrifying MBBR2 using newly developed Z-carriers that allow the control of defined biofilm thickness. The use of thinner biofilms (~ 50 µm), rather than thicker biofilms (>200 µm), had a positive effect on nitrification rates and on the biotransformation kinetics of a number of compound such as diclofenac (kBio up to 6 L g-1 d-1) and the three sulfonamide antibiotics. However, the biotransformation of more than 60% of targeted compounds was enhanced in thicker biofilms, that exhibited higher (p<0.05) microbial diversity and were more even. Additionally, a biofilm model was developed and calibrated to evaluate sorption and diffusion of micropollutants in nitrifying biofilms. Sorption was significant only for eight out of the targeted compounds. All compounds removed by sorption were predicted to carry a net positive charge at the experimental pH, suggesting the importance of electrostatic interactions on sorption in biofilms. Sorption coefficients Kd (L g-1) and effective diffusivity coefficients f increased with increasing biofilm thickness, suggesting reduced diffusion limitation and higher surface area accessibility in the thickest, least dense biofilm (~500 µm). Two types of commonly dosed degradable carbon sources (methanol and ethanol) were investigated in two parallel post-denitrifying systems (MBBR3). The methanol-dosed MBBR exhibited in the enhancement of kBio (up to 2.5-fold) for a number of micropollutants (nine out 23) compared to the ethanol-dosed MBBR, while for 10 compounds biokinetics were similar between the two reactors. The higher denitrification rates exhibited by the ethanol-dosed MBBR during batch experiments likely influenced the biotransformation of the sulfonamides antibiotics, in analogy with what observed in MBBR2. A strong cometabolic effect (i.e., an enhancement of micropollutant biotransformation in the presence of organic carbon) was observed for venlafaxine, carbamazepine, sulfamethoxazole and sulfamethizole. However, an increase in methanol or ethanol loading to the MBBRs during continuous-flow experiment did not influence the removal of the targeted micropollutants, most likely due to the short hydraulic residence time (2 hours) used in the study as well as in full-scale reactors. Diversity-function relationships (assessed through Pearson correlation analyses) were tested by comparing diversity estimators against biomass-normalized biotransformation rates. A positive influence of biodiversity for most of the targeted compounds (~60%) was shown in MBBR2 study, while biotransformation of few compounds (diclofenac and sulfonamides) was positively associated to microbial activity (i.e., nitrification). Similarly, a positive association (p<0.05) with the specific denitrification rate was shown in MBBR1, while biotransformation of most of the detected pharmaceuticals in wastewater did not associate or negatively associated with biodiversity. The relationship between biodiversity and micropollutant biotransformation may depend on whether its biotransformation is catalysed by a narrow (i.e., performed by few species) or broad processes. It is likely that for highly redundant microbial processes (such as denitrification), micropollutant biotransformation may be catalysed by broadly distributed enzymes and pathways, and microbial diversity provides no benefit. Conversely, increasing biodiversity under nitrifying conditions may be necessary to increase the inclusion of microorganisms with specific functionality towards micropollutant biotransformation. Overall, the biotransformation rates were significantly enhanced in MBBR3 compared to MBBR1 and MBBR2 for the majority of micropollutants (~60%) suggesting the positive impact of easily degradable carbon sources (such as methanol or ethanol) on micropollutant removal. Finally, the removal of compounds such propranolol atenolol, citalopram, venlafaxine (under post-denitrifying conditions) and diclofenac (under aerobic conditions) was improved compared to conventional activated sludge. It can be thus concluded that MBBRs can offer a suitable technology that can be optimized for the removal of micropollutants in municipal wastewaters under a range of operating conditions (nitrifying, pre- and post-denitrifying).