INVESTIGATION OF THE REMOVAL OF CONTAMINANTS OF EMERGING CONCERN AND THE MICROBIOME IN BIOFILTERS

Abstract

The removal and fate of contaminants of emerging concern (CECs) in water treatment systems is of interest given the widespread occurrence of CECs in water supplies. Biofiltration, which granular media filters are inhabited by viable bacteria, has the potential of providing long-term, sustainable CECs removal in drinking water treatment. Bacterial communities in biofilters can be beneficial through biodegradation of contaminants but also pose potential risks by harboring and releasing detrimental microbes into water distribution systems. In this work, the removal of eight CECs, including atenolol, atrazine, carbamazepine, fluoxetine, gemfibrozil, metolachlor, sulfamethoxazole and tris(2-chloroethyl) phosphate, was investigated in pilot-scale granular activated carbon (GAC)-sand and anthracite-sand biofilters. The effects of water quality and engineering decisions on the biofilter microbiome and the effect of biofilters on the microbiome in filter effluent were evaluated in the aforementioned pilot-scale biofilters. In addition, the geographic patterns of the biofilter microbiome were investigated by sampling filter media from full-scale biofilters at fourteen treatment plants throughout North America. The CECs concentrations in the filter influent and effluent were determined using liquid chromatograph tandem mass spectrometry, and the bacterial abundance and community composition in the biofilters were determined using real-time quantitative polymerase chain reaction (qPCR) and Illumina HiSeq high-throughput sequencing of PCR amplicons. GAC-sand biofilters provided superior CECs removal for all compounds (mean removal efficiencies: 49.1-94.4%) compared to anthracite-sand biofilters (mean removal efficiencies: 0-66.1%) due to a combination of adsorption and biodegradation. Adsorption was determined to be the dominant removal mechanism for most selected CECs in GAC-sand biofilters. A multiple linear regression based empirical relationship considering water quality, engineering decisions, and CECs chemical properties was developed to predict CECs removal in the GAC-sand biofilters. The microbiome in the pilot-scale and full-scale biofilters contained genera that are commonly found in the freshwater environments and water distribution systems, such as Limnohabitans, Flavobacterium, Nitrospira, and, Hydrogenophaga. The microbiome in the pilot-scale biofilters exhibited temporal variations, and varied with media type (GAC vs. anthracite), backwash strategy (chloraminated vs. non-chloraminated), and bed depth. The pilot-scale biofilters effectively removed biomass (~70%) from the water, but only marginally impacted the microbiome in the filter effluent. Significant inter-filter variations were observed in the full-scale biofilter investigation that followed a weak but highly significant distance-decay relationship. The water quality characteristics exhibited a stronger influence on the microbiomes in the full-scale biofilters than the geographic distance according to a multiple regression on matrix analysis. Nitrosomonas oligotropha-like ammonia oxidizing bacteria (AOB) were generally more abundant than ammonia oxidizing archaea in the full-scale and pilot-scale biofilters. The ratios of nitrite oxidizing bacteria to AOB exceeded the theoretical ratio for conventional two step nitrification in most full-scale biofilters (12 of 14 biofilters) and in the pilot-scale biofilters for most of the operation. This work should be beneficial to drinking water treatment facilities in improving CECs removal performance using biofilters, as well as to environmental engineers and scientists in understanding the microbiome in biofilters.