Fate of endocrine-disruptor, pharmaceutical, and personal care product chemicals during simulated drinking water treatment processes

The potential occurrence of endocrine-disrupting compounds (EDCs) as well as pharmaceuticals and personal care products (PPCPs) in drinking water supplies raises concern over the removal of these compounds by common drinking water treatment processes. Three drinking water supplies were spiked with 10 to 250 ng/L of 62 different EDC/PPCPs; one model water containing an NOM isolate was spiked with 49 different EDC/PPCPs. Compounds were detected by LC/MS/MS or GC/MS/MS. These test waters were subjected to bench-scale experimentation to simulate individual treatment processes in a water treatment plant (WTP). Aluminum sulfate and ferric chloride coagulants or chemical lime softening removed some polyaromatic hydrocarbons (PAHs) but removed <25% of most other EDC/PPCPs. Addition of 5 mg/L of powder activated carbon (PAC) with a 4-h contact time removed 50% to >98% of GC/MS/MS compounds (more volatile) and 10% to >95% of LC/MS/MS compounds (more polar); higher PAC dosages improved EDC/PPCP removal. EDC/PPCP percentage removal was independent of the initial compound concentration. Octanol-water partition coefficients served as a reasonable indicator of compound removal under controlled PAC test conditions, except for EDC/PPCPs that were protonated or deprotonated at the test pH and some that contained heterocyclic or aromatic nitrogen. Separate chlorine orozone experiments decreased the EDC/PPCP initial concentration by <10% to >90%; EDC/PPCPs were likely transformed to oxidation byproducts. Ozone oxidized steroids containing phenolic moieties (estradiol, ethynylestradiol, or estrone) more efficiently than those without aromatic or phenolic moieties (androstenedione, progesterone, and testosterone). EDC/PPCP reactivity with oxidants were separated into three general groups: (1) compounds easily oxidized (>80% reacted) by chlorine are always oxidized at least as efficiently by ozone; (2) 6 of the ∼60 compounds (TCEP, BHC, chlordane, dieldrin, heptachlor epoxide, musk ketone) were poorly oxidized (<20% reacted) by chlorine or ozone; (3) compounds (24 of 60) reacting preferentially (higher removals) with ozone rather than chlorine. Conventional treatment (coagulation plus chlorination) would have low removal of many EDC/PPCPs, while addition of PAC and/or ozone could substantially improve their removals. Existing strategies that predict relative removals of herbicides, pesticides, and other organic pollutants by activated carbon or oxidation can be directly applied for the removal of many EDC/PPCPs, but these strategies need to be modified to account for charged (protonated bases or deprotonated acids) and aliphatic species. Some compounds (e.g., DEET, ibuprofen, gemfibrozil) had low removals unless ozonation was used. Other compounds had low removals by all the WTP processes considered (atrazine, iopromide, meprobamate, TCEP), and removal processes capable of removing these types of compounds should be investigated.