In biofilm systems for wastewater treatment (e.g., moving bed biofilms reactors-MBBRs) biofilm thickness is typically not under direct control. Nevertheless, biofilm thickness is likely to have a profound effect on the microbial diversity and activity, as a result of diffusion limitation and thus substrate penetration in the biofilm. In this study, we investigated the impact of biofilm thickness on nitrification and on the removal of more than 20 organic micropollutants in laboratory-scale nitrifying MBBRs. We used novel carriers (Z-carriers, AnoxKaldnes) that allowed controlling biofilm thickness at 50, 200, 300, 400, and 500 μm. The impact of biofilm thickness on microbial community was assessed via 16S rRNA gene amplicon sequencing and ammonia monooxygenase (amoA) abundance quantification through quantitative PCR (qPCR). Results from batch experiments and microbial analysis showed that (i) the thickest biofilm (500 μm) presented the highest specific biotransformation rate constants (kbio, L g(-1) d(-1)) for 14 out of 22 micropollutants; (ii) biofilm thickness positively associated with biodiversity, which was suggested as the main factor for the observed enhancement of kbio; (iii) the thinnest biofilm (50 μm) exhibited the highest nitrification rate (gN d(-1) g(-1)), amoA gene abundance and kbio values for some of the most recalcitrant micropollutants (i.e., diclofenac and targeted sulfonamides). Although thin biofilms favored nitrification activity and the removal of some micropollutants, treatment systems based on thicker biofilms should be considered to enhance the elimination of a broad spectrum of micropollutants.

Addition of external carbon sources to post-denitrification systems is frequently used in wastewater treatment plants to enhance nitrate removal. However, little is known about the fate of micropollutants in post-denitrification systems and the influence of external carbon dosing on their removal. In this study, we assessed the effects of two different types and availability of commonly used carbon sources -methanol and ethanol- on the removal of micropollutants in biofilm systems. Two laboratory-scale moving bed biofilm reactors (MBBRs), containing AnoxKaldnes K1 carriers with acclimated biofilm from full-scale systems, were operated in continuous-flow using wastewater dosed with methanol and ethanol, respectively. Batch experiments with 22 spiked pharmaceuticals were performed to assess removal kinetics. Acetyl-sulfadiazine, atenolol, citalopram, propranolol and trimethoprim were easily biotransformed in both MBBRs (biotransformations rate constants k between 1.2 and 12.9 L g d), 13 compounds were moderately biotransformed (rate constants between 0.2 and 2 L g d) and 4 compounds were recalcitrant. The methanol-dosed MBBR showed higher k (e.g., 1.5-2.5-fold) than in the ethanol-dosed MBBR for 9 out of the 22 studied compounds, equal k for 10 compounds, while 3 compounds (i.e., targeted sulfonamides) were biotransformed faster in the ethanol-dosed MBBR. While biotransformation of most of the targeted compounds followed first-order kinetics, removal of venlafaxine, carbamazepine, sulfamethoxazole and sulfamethizole could be described with a cometabolic model. Analyses of the microbial composition in the biofilms using 16S rRNA amplicon sequencing revealed that the methanol-dosed MBBR contained higher microbial richness than the one dosed with ethanol, suggesting that improved biotransformation of targeted compounds could be associated with higher microbial richness. During continuous-flow operation, at conditions representative of full-scale denitrification systems (hydraulic residence time = 2 h), the removal efficiencies of micropollutants were below 35% in both MBBRs, with the exception of atenolol and trimethoprim (>80%). Overall, this study demonstrated that MBBRs used for post-denitrification could be optimized to enhance the biotransformation of a number of micropollutants by accounting for optimal carbon sources and extended residence time.

Solid-liquid partitioning is one of the main fate processes determining the removal of micropollutants in wastewater. Little is known on the sorption of micropollutants in biofilms, where molecular diffusion may significantly influence partitioning kinetics. In this study, the diffusion and the sorption of 23 micropollutants were investigated in novel moving bed biofilm reactor (MBBR) carriers with controlled biofilm thickness (50, 200 and 500 μm) using targeted batch experiments (initial concentration = 1 μg L, for X-ray contrast media 15 μg L) and mathematical modelling. We assessed the influence of biofilm thickness and density on the dimensionless effective diffusivity coefficient f (equal to the biofilm-to-aqueous diffusivity ratio) and the distribution coefficient K (L g). Sorption was significant only for eight positively charged micropollutants (atenolol, metoprolol, propranolol, citalopram, venlafaxine, erythromycin, clarithromycin and roxithromycin), revealing the importance of electrostatic interactions with solids. Sorption equilibria were likely not reached within the duration of batch experiments (4 h), particularly for the thickest biofilm, requiring the calculation of the distribution coefficient K based on the approximation of the asymptotic equilibrium concentration (t > 4 h). K values increased with increasing biofilm thickness for all sorptive micropollutants (except atenolol), possibly due to higher porosity and accessible surface area in the thickest biofilm. Positive correlations between K and micropollutant properties (polarity and molecular size descriptors) were identified but not for all biofilm thicknesses, thus confirming the challenge of improving predictive sorption models for positively charged compounds. A diffusion-sorption model was developed and calibrated against experimental data, and estimated f values also increased with increasing biofilm thickness. This indicates that diffusion in thin biofilms may be strongly limited (f ≪ 0.1) by the high biomass density (reduced porosity).