Conventional and Advanced Water Treatment Processes for the Removal of Endocrine Disruptors and Pharmaceuticals

A suite of target endocrine disruptor compounds (EDCs) and pharmaceuticals and personal care products (PPCPs) were evaluated for removal by several water treatment processes at bench-, pilot-, and fullscale. Over 60 different EDCs and PPCPs were chosen and simultaneously evaluated based on structure, occurrence, and potential health impacts. Treatment processes evaluated include physical (e.g., activated carbon, membrane filtration, softening, ion-exchange), oxidative (e.g. ozonation, chlorination, ultraviolet irradiation, chloramination), advanced oxidation (UV/peroxide, ozone/peroxide), and biological (e.g. biologically active filtration, soil aquifer treatment, membrane bioreactors, river bank filtration). Bench-scale evaluations were performed as static experiments in discrete conditions, generally as “jar-tests”. Pilot-scale testing was conducted as dynamic, flow-through, experiments by either spiking a head tank or by feeding the spiking solution in-line. Full-scale testing was accomplished without spiking, thus limiting evaluations to compounds present in raw water. It was determined that bench-scale results adequately predicted removal trends observed at pilot and full scale. Summarily, the results show that ozone, chlorine, certain membranes, and powdered activated carbon are effective for removal of many EDCs and PPCPs. However, removal efficiency is compound specific and depends on operational parameters, such as oxidant dosage and contact time. For instance, UV irradiation was ineffective for removing target compounds at typical disinfectant doses of 40 mJ/cm while doses of 1,000 mJ/cm and peroxide addition were effective for many compounds. Molecular structure (e.g., pKa, Kow, molecular weight, UV absorbance) generally dictated whether a treatment process would or would not efficiently remove the compound. As a disinfection process, ozone was more effective for removing target compounds than was chlorine, UV, and chloramination. Removal by membranes was highly selective based on the membrane pore size, surface characteristics, and degree of fouling. Biological processes were effective for some compounds, and nearly ineffective for others. At full-scale, granular activated carbon and reverse osmosis were found to be the most effective treatment processes for removing a wide-variety of target compounds. Based on data obtained from this study, removal predictions are possible for new contaminants based on physico-chemical properties of the contaminant.