Understanding The Mechanistic Contributions To Virus Removal In Membrane Bioreactors

Introduction and Background
Pathogen removal by MBRs has shown to be substantial; however, the credited log reduction values (LRV) are often conservatively downrated due to uncertainty in current and future ability to maintain performance. Based on a recent and extensive case study, Kubota flat sheet membranes have achieved average culturable human virus LRVs of 4.1 for both adenovirus and enterovirus. Protozoa LRVs have averaged >4.2 for Cryptosporidium and 5.2 for Giardia, with no Cryptosporidium detections in the filtrate. The Kubota membranes have an average pore size of 0.2 µm and are therefore reasoned to remove protozoa (typically > 3 µm). Viruses, however, are much smaller in comparison to the pore size (<0.1 µm), but case study results demonstrate significant removal. This study aims to further understand the mechanistic factors that contribute to enhanced virus removal observed in Kubota flat sheet MBR systems.
Methodology
To quantify the contributions of the separate pathogen removal mechanisms, experiments were designed to test mechanisms separately. Single strain male specific (MS2) and somatic (T1) coliphages were used as surrogates to assess virus removal. A two-day trial was conducted to investigate the extent of adsorption and biological degradation on model virus LRV. Activated sludge was sampled fresh from an MBR test plant in Canton, OH. The sludge was sent for characterization of indigenous male-specific and somatic coliphages. Experiments were conducted whereby the concentrations of spiked and indigenous male-specific coliphages were determined after various time periods and with centrifugation. LRV due to adsorption of indigenous coliphages was inferred by comparing the native well mixed activated sludge sample concentrations to centrifuged samples. Whereas, the extent of aerobic biological predation was determined by comparing the supernatant concentrations on day 1 to that of day 2 after 24 hours of aeration. See Figure 1 for an illustration of the experimental approach. Concerning biofilm plus membrane contribution to virus removal, a pilot plant was used to control biofilm growth over a period of time. The membranes in the pilot plant are flat sheet microfiltration style with a 0.2 µm average pore size. Following sufficient biofilm growth (fouling layer deposited on the membranes during operation), mixed liquor and filtrate samples were taken to determine native populations of MS2 and T1 coliphages. The mixed liquor was then spiked with both MS2 and T1 to determine virus removal of the entire system. The mixed liquor was then flushed from the membrane tank and filled with tap water, so as to not disturb the biofilm. This can be seen in Figure 2. Spikes were administered to the system and tap water and filtrate samples were taken; only the membranes and biofilm would be contributing to virus removal.
Results
Significant adsorption occurred rapidly for all spiked indicators; results show that of 10# 7 (MS2) and 10# 6 (T1) spiked organisms, 10# 4 (MS2) and 10# 3 (T1) were associated with sludge flocs. This indicates that initial adsorption may contribute 2.8 to 4.2 LRV. Figure 3 graphically shows that adsorption to sludge increases over time for both coliphages examined, in both anoxic and aerated conditions. Results indicate that biofilm plus membrane removal can contribute from one-fifth to upwards of one-half of the total removal. Average results from all experimental runs are displayed in Table 1. These results show that the biofilm plays a significant role in virus removal in microfiltration MBRs.
Conclusions
This study distinguishes between participating virus removal mechanisms in membrane bioreactors including adsorption to sludge and biofilm plus membrane removal utilizing male specific and somatic coliphages as surrogate microorganisms for human enteric viruses. It was shown that overtime, coliphage adsorption seemingly plateaus after several hours while biofilm contribution can vary, possibly due to environmental factors. Finally, these removal data were compared to an existing pathogen removal data set from a year-long case study of a full-scale Kubota MBR to verify the impact of the contributing mechanisms to actual virus removal upon scale up.