Passive sampling for PFAS in a drinking water treatment plant: Potential utility, challenges, and pathways forward

Per- and polyfluoroalkyl substances (PFAS) need to be effectively monitored in drinking water treatment plants (DWTPs) to assess treatment performance and manage potential health risks. Grab sampling provides only snapshot concentrations, whereas passive sampling offers time-weighted average concentrations (C_TWA). Polar organic chemical integrative samplers (POCIS) are widely used for polar contaminants, including PFAS, yet their reliability for estimating C_TWA in individual DWTP processes remains unclear.

PFAS passive sampling performance was evaluated in controlled laboratory tests and a 14-day field study at a pilot-scale DWTP. C_TWA values were determined using POCIS consisting of a polyethersulfone membrane and a hydrophilic-lipophilic balanced sorbent. Under controlled laboratory conditions with time-varying concentrations, passive sampling reliably captured PFAS fluctuations, while 33% of the total accumulated mass of perfluorooctanesulfonic acid was found in the membrane, highlighting the importance of whole-sampler uptake for C_TWA estimation. In the field study, passive sampling generally yielded lower PFAS concentrations than grab sampling even after flow-specific sampling rates were applied to each process stage, especially in stages with elevated suspended solids and organic matter. Under DWTP conditions, 21-52% of accumulated perfluorooctanoic acid was retained in the membrane, indicating that water-matrix effects altered sampler uptake.

POCIS provided reliable PFAS concentration estimates under controlled exposure conditions, whereas field application was influenced by water-matrix conditions. These findings identify water-matrix effects as an important source of uncertainty beyond hydrodynamic correction and highlight the need for matrix-aware interpretation of passive sampling results in fields.