Evaluation of a ceramic membrane bioreactor functioning under non-steady state operational parameters for the removal of pollutants from municipal wastewater plant effluent

Abstract

In South Africa ill-managed municipal wastewater treatment plants limit the quality and quantity of already exploited surface water resources. Elevated nitrate levels cause eutrophication and chemicals (measured as chemical oxygen demand or COD) discharged through effluent water may even irreversibly alter the quality of potable water. Pathogens entering river systems cause high risk especially for rural communities. The status quo of membrane bioreactor technology for microorganism and solid retention is very broad and the development and application thereof is being driven by both fresh water shortage and anticipated stringent environmental regulations. Very little research is done on membrane bioreactors (MBRs) functioning under non-steady state parameters with spontaneous fluctuations in permeate flux. Basic models that provide a holistic understanding of the technology at a fundamental level are therefore of great necessity. It was hypothesized that a simple aerobic-anaerobic external circuit ceramic MBR operated under spontaneous parameter variance could reach high pollutant removal efficiencies, even when shock-loaded with nutrients and microorganisms. A ceramic membrane bioreactor system was constructed and operated under feed-and-bleed conditions (CNP=85:15:1) for 112 days and hydraulic retention time (HRT) of between 4 and 7 days, to determine whether the system could buffer nitrate and COD discharge into river systems. Transmembrane pressure was maintained at 10kPa and flux were allowed to change to obtain non-steady operating conditions. From stabilization (day 47) towards day 63 of operation, average nitrate and COD removal reached 47.64% and 86.55%, respectively the system stabilized after day 47 reaching a COD removal near 99% between days 85 and 101. Shock loading with nutrients (CNP=250:35:3) was done on day 64, where after nitrate and COD removing efficiency decreased to 9.41% and 80.29% respectively within 10 days. COD removal capacity recovered to 99.92% 5 days after the shock loading on day 69. This demonstrates the robustness of the system. However, it took 35 days for nitrate removal to recover to 25.12 %. A major challenge during the operation of the system was fouling which started on day 47. This could be ascribed to the extremely low average solid retention time (SRT) (0.086 days) values. A gas-liquid back flush regime of three times a week was necessary to overcome this problem. On day 146 the system was shock loaded with E.coli transformed with pBR322 for microbial retention analysis. Average retention was 89.15% in the first 5 days during the experiment after breakthrough occurred. Carbon breakthrough was measured as elevated COD. Microbial levels reached ~5.3×105 cfu/mℓ in the effluent. Biofilm samples were taken throughout MBR operation. DNA was extracted and the V3 region of 16S rRNA was amplified by PCR using primers for Bacteria. PCR products were subjected to SSCP and DGGE analysis to generate molecular fingerprinting profiles representative of the community structure associated with the non-steady operational conditions. Bands, representing various species were excised, reamplified and sequenced to determine identities of the bacteria. The PCR-SSCP and PCR-DGGE banding patterns were subjected to Shannon Weaver diversity index analysis. Dendrograms for each of the SSCP and DGGE gel profiles were obtained using Ward‟s method and Euclidean distances. Maximum H' values of 1.11, 1.27 and 1.26 were reached on days 20, 40 and 81 with close correlation between days 40 and 81 with Euclidean linkage distances lower than 3.2. This is indicative of the increased vigor of specific organisms specializing in nutrient removal at high concentrations DGGE fingerprinting suggested a subsequent shift in diversity. Three distinctive shifts in diversity were evident throughout MBR operation. This may have been due to re-organization of the community as the species involved out-competed other species over time. A sudden shift in community was observed during 1) days 6-20 ; 2) days 27 to 40 and 3) days 63 to 81 with ultimate H' values exceeding 1.0 at the end of each phase with clear differences in SRT and flux showing a gradual drop in value for each of the three phases. Results suggest limitations in the surviving capacities of the mixed culture biofilm. PCR-DGGE as well as PCR-SSCP were useful methods to obtain a genetic fingerprint profile for MBR biofilm characterization. PCR-SSCP was the method of choice, due to its sensitivity and expediency. SCCP profiles showed that Aeromonas hydrophila, Delftia spp. and an uncultured bacterial species were the three most evident organisms present, potentially responsible for elevated nitrate and COD removal soon after nutrient shock loading with average nutrient removal (days 63 and 91) of 20.76% nitrate and 82.13% COD. Analyses of an MBR functioning under non-steady state conditions are complex with reference to spontaneous change in a variety of parameters. SSCP is less time consuming but the steps involved are nonetheless complex. We conclude that non-steady state MBRs have limited potential to be used as an add-on to existing municipal wastewater treatment plants to serve as buffer for reducing COD and nitrate levels in wastewater effluent due to the intricate and complex nature of operation. Therefore, steady state MBR operation remains the optimal method of operation for the purpose of studying biofilm characteristics.