The Critical Role of Chemical Pre-Treatment in Ensuring Cryptosporidium Removal by Filtration of High Quality Source Water

Abstract

Drinking water utilities reliant on surface water utilize chemically-assisted filtration (CAF) as a key barrier against the passage of protozoan pathogens, like Cryptosporidium spp. oocysts, to treated water. The goal of this work was to enable system-specific and potentially dynamic assessment of oocyst removal by CAF by using zeta potential as a tool for rapid operational feedback. Specifically, this work focused on systems utilizing high quality, low turbidity (typically <1 NTU) source water, with relatively low C. parvum oocyst concentrations and applied full scale coagulant doses (typically <5 mg/L). In these systems, the formation of settleable flocs is not a necessity because source water turbidities are already low and frequently meet treated water criteria. Rather, coagulation is used to enable particle removal through physico-chemical (i.e., chemically -assisted) filtration, as indicated by filter effluent turbidities that may or may not be indicative of optimal particle destabilization and removal by CAF. Accordingly, the identification of “optimal” coagulant doses can be challenging, and becomes even more challenging when process performance is being assessed, such as when Cryptosporidium oocysts are added to filter influents to evaluate their removal by CAF processes. In Phase 1 of this work, the role of oocyst coagulation during CAF performance demonstrations was investigated. It was demonstrated that appropriate coagulation of oocyst seed suspensions is critical to reflecting “well-operated” CAF performance. A protocol for ensuring optimal coagulation of oocyst seed suspensions during such performance demonstrations was developed and demonstrated at pilot-scale. Here, zeta potential was useful in identifying the coagulant doses needed for maximal particle destabilization and removal by CAF. This pilot-scale approach was then validated using lower, environmentally relevant oocyst concentrations (and much longer pilot-scale investigations) during which the entire filtered volume of water was evaluated. Using this protocol, it was demonstrated that a minimum of 3-log oocyst removal could be achieved by CAF (essentially direct filtration) at a variety of operational conditions. During Phase 2 of this work, the protocol developed in Phase 1 was used to evaluate oocyst passage through CAF processes with different filter designs (bed depths, water temperature) at various operational conditions (suboptimal coagulation, filter ripening, end-of-run operation, and hydraulic surges). Here, because of the high quality, low turbidity source water, adequate coagulation was the dominant control for risk, in contrast to many reported investigations in which more deteriorated source water was investigated and operational period within the filter cycle was a more dominant control over oocyst passage through the CAF process. Here, with the exception of suboptimal coagulation conditions, the pilot-scale filters consistently achieved >3-log C. parvum oocyst removal in an essentially directly filtration mode. Thus, this work demonstrated the critical importance of (1) appropriate particle destabilization by coagulation prior to CAF of low turbidity, low DOC source waters, (2) coagulation of oocysts prior to their addition to filter influent streams during CAF performance demonstrations, and (3) zeta potential as a useful tool for ensuring adequate particle destabilization in situations (i.e. treatment of low turbidity, low DOC source waters) in which extensive particle settling is not likely. In doing so, this work further highlights that Cryptosporidium spp. oocyst removal credits of >2.5 log may be warranted for “well-operated” direct filtration processes.