Main content area

Denitrification and microbial community in MBBR using A. donax as carbon source and biofilm carriers for reverse osmosis concentrate treatment

Li, Li, Yan, Guokai, Wang, Haiyan, Chu, Zhaosheng, Li, Zewen, Ling, Yu, Wu, Tong
Journal of environmental sciences (China) 2019 v.84 pp. 133-143
Arundo donax, activated sludge, anaerobic ammonium oxidation, biofilm, carbon, denitrification, gamma-Proteobacteria, genes, high-throughput nucleotide sequencing, microbial communities, moving bed biofilm reactors, nitrate nitrogen, prediction, quantitative polymerase chain reaction, reverse osmosis, water reuse
In this study, raw Arundo donax (A. donax) pieces were applied as carbon source and biofilm carriers for denitrification in a lab-scale moving bed biofilm reactor (MBBR) for the treatment of reverse osmosis concentrate gathered from local wastewater reuse plant. At stable phase (about 60 days), efficient denitrification performance was obtained with 73.2% ± 19.5% NO3−-N average removal and 8.10 ± 3.45 g N/(m3·day) NO3−-N average volumetric removal rate. Mass balance analysis showed that 4.84 g A. donax was required to remove 1 g TN. Quantitative real-time PCR analysis results showed that the copy numbers of 16S r-RNA, narG, nirS, nosZ and anammox gene of carrier biofilm and suspended activated sludge in the declination phase (BF2 and AS2) were lower than those of samples in the stable phase (BF1 and AS1), and relatively higher copy numbers of nirS and nirK genes with lower abundance of narG and nosZ genes were observed. High-throughput sequencing analysis was conducted for BF2 and AS2, and similar dominant phyla and classes with different abundance were obtained. The class Gammaproteobacteria affiliated with the phylum Proteobacteria was the most dominant microbial community in both BF2 (52.6%) and AS2 (41.7%). The PICRUSt prediction results indicated that 33 predictive specific genes were related to denitrification process, and the relative abundance of 18 predictive specific genes in BF2 were higher than those in AS2.