Bradley M. Coffey

Effects of aluminum sulfate and ferric chloride coagulant residuals on polyamide membrane performance

Abstract Pretreatment may constitute up to one-fourth of the total costs of a membrane desalting facility. By using preexisting conventional filtration plants for pretreatment, significant cost savings may be realized. However, coagulant residuals from the pretreatment process may negatively affect reverse osmosis (RO) membrane performance. Various RO membranes were tested at three different treatment plants in southern California, using either aluminum sulfate (alum) or ferric chloride coagulants and chloramines. Repeated testing using alum with multiple RO elements revealed rapid deterioration in specific flux (up to 60% over 100 h of operation), as well as progressive reductions in salt rejection (typically 3–4% over 500 h of operation). Microscopic analysis of the fouled membranes revealed that the foulants were primarily aluminum hydroxide and aluminum silicate materials. In contrast to the RO data for alum coagulation, which showed declining membrane flux, the specific flux data using ferric chloride and chloramines increased over time for all membranes. Salt rejection decreased significantly during testing of each membrane. These data suggest that the RO membranes were physically degrading over time. The RO membranes may have been degraded by residual iron catalyzing a chlorine-amide reaction on the membrane surface, despite the fact that chlorine was present as chloramines.

The effect of UV light on the inactivation of Giardia lamblia and Giardia muris cysts as determined by animal infectivity assay (P-2951-01).

This study measured the effect of germicidal ultraviolet (UV) light on Giardia lamblia and Giardia muris cysts, as determined by their infectivity in Mongolian gerbils and CD-1 mice, respectively. Reduction of cyst infectivity due to UV exposure was quantified by applying most probable number techniques. Controlled bench-scale, collimated-beam tests exposed cysts suspended in filtered natural water to light from a low-pressure UV lamp. Both G. lamblia and G. muris cysts showed similar sensitivity to UV light. At 3 mJ/cm2, a dose 10-fold lower than what large-scale UV reactors may be designed to provide, > 2-log10 (99 percent) inactivation was observed. These results, combined with previously published data showing other protozoa and bacteria have similar, high sensitivity to UV light, establish that UV disinfection of drinking water is controlled by viruses which may require over 10-fold more UV dose for the same level of control.

Pilot-scale testing of reverse osmosis using conventional treatment and microfiltration

Abstract In order to meet the growing needs of its consumers, the Metropolitan Water District of Southern California is looking for ways to desalt water from the relatively hard Colorado River. This study evaluated conventional treatment as the pretreatment step for reverse osmosis (RO) desalting. Other pretreatment options studied include microfiltration (MF) and conventional treatment with ozone and biologically active filters. Each of three pretreatment scenarios produced effluent waters generally considered appropriate for use with RO [median turbidity of less than 0.1 Nephelometric turbidity unit (NTU) and median silt density index of less than 3]. Both microfiltered and ozonated/ biofiltered waters gave steady RO performance over 3 months of testing. However, conventional treatment left the RO system vulnerable to organic and biological fouling. Despite maintaining a 2-2.5 mg/L chloramine residual, pretreatment using conventional treatment required more frequent cleanings than either MF or ozonation/biofiltration (O3/BF). The good performance for biofiltered water may have resulted from the stabilization of the natural organic matter through the O3/BF process. Microfiltration, with its superior particle and bacteriological removal characteristics, provided the best RO pretreatment technology.

THE IMPORTANCE OF ROBUSTNESS IN DRINKING-WATER SYSTEMS

It is important that drinking-water systems be as robust as possible. That is, they should deliver excellent quality water under adverse conditions. Robustness is important for each of the five elements that can be considered necessary for providing safe drinking water (a good source, adequate treatment, secure distribution, appropriate monitoring, and appropriate response to adverse monitoring results). However, a given degree of overall system robustness can be achieved in varying ways. The quantification of robustness facilitates its improvement in a rational way. This paper introduces the concept of robustness, and illustrates one way in which it could be quantified by means of an example involving filtration in relation to Cryptosporidium removal. With regard to a serious water contamination incident that occurred in Canada during May 2000, the robustness of each of the five elements (source, treatment, distribution, monitoring, and response) is assessed qualitatively to explain the overall vulnerability of the water-supply system in that town.

The Importance of Coagulation for the Removal of Cryptosporidium and Surrogates by Filtration

This investigation examined the removal of Cryptosporidium and several surrogates by granular medium filtration. The study focused on coagulation impacts on removals. Inactivated C. parvum oocysts and non-inactivated Bacillus subtilis were seeded in two pilot plants. The pilot plant locations were chosen to represent different coagulation regimes. Under optimized coagulation conditions (filter effluent turbidity < 0.1 NTU), approximately 5 log10 removal of Crypto. was obtained in one location, and approximately 3 log10 in the other. This difference is likely related at least in part to the difference in coagulation regimes. Suboptimal coagulation produced a substantial deterioration of Crypto. removal in both locations. The trends seen for Bacillus were generally similar to those observed for Crypto.. Although turbidity may be of value for assessing coagulation impacts on Crypto. removal, particle counts may be a more sensitive parameter in this regard.