Liang, Chuanzhou; de Jonge, Nadieh; Carvalho, Pedro N; Nielsen, Jeppe Lund; Bester, Kai

Chemical Engineering Journal 415, 128963, 2021

doi.org/10.1016/j.cej.2021.128963

Feast-famine moving bed biofilm reactors (MBBRs) have shown high potential for removing organic micropollutants from wastewater. However, the relationship between biofilm community during feast-famine adaptation and micropollutant removal is yet unclear. In this study, we determined the biotransformation kinetics of 36 micropollutants and characterized the microbial communities in an MBBR during a 71-day adaptation period of feast-famine regime (raw/effluent wastewater). The feast-famine regime significantly changed the biodegradation rate constants (k) of 24 micropollutants in different ways: 66 times enhanced degradation for propranolol, while more than 10 times for atenolol, metoprolol, tramadol and venlafaxine, less than 2.8 times for losartan, iomeprol and iohexol were detected. 25–60 days of adaptation time was needed to reach the maximum k. Biofilm accumulated during the adaptation, but the kDNA (k relative to the biofilm with DNA concentration as a proxy for kbiomass) of most micropollutants (except propranolol, metoprolol and venlafaxine) declined. This might indicate that the proliferation of potential degraders for micropollutants was slower than other microorganisms under the feast-famine regime. The microbial community changed significantly during the first 8 days of operation, followed by a relatively steady evolution towards the enrichment of nitrifiers until day 71. A multivariate statistical correlation analysis revealed that the development of occurrence of 88 individual taxonomic groups were found to exhibit a significant positive correlation to the kDNA of micropollutants (p < 0.05, r > 0.5), which represent potential biomarkers linking to biotransformation of micropollutants. These results fill the knowledge gaps between dynamics of biofilm communities and micropollutant removal in the feast-famine regime, which is essential for designing highly efficient MBBR.

Figure 2 (a) Reaction rate constant k (h-1) and (b) kDNA (L mg-1 h-1) where both k values are normalized by the each values at day 1. kDNA is the reaction rate constant k divided by the DNA concentration (mg L-1) in the respective batch test	Figure 4 (a) PCA of biofilm samples and (b) heatmap of relative abundance of 25 domain genera or OTUs during adaptation of the feast-famine condition. The trajectory number in the PCA and the X axis of the heatmap means the adaptation day of samples