Avoiding artifacts in the determination of per- and polyfluoroalkyl substance sorbent-water distribution

Characterizing sorbent affinity for a target compound (described by sorbent-water distribution coefficient, K_sw) is a necessary step in the sorbent selection and performance-testing process in the process of capturing aquatic contaminants. However, no standardized procedure exists to measure K_sw, and studies display significant variations in set-up and performance. For per- and polyfluoroalkyl substances (PFAS), most K_sw determinations employ batch experiments with small-scale water-sorbent mixtures, methanol-based spike of target compound(s), and analysis after assumed equilibrium, but methodological details of the above procedure differ and might cause artifacts in the determination of K_sw. We conducted several batch experiments systematically varying a general procedure to characterize effects of sub-optimal experimental design. Using a selection of PFAS (6-carbon fluorinated chain length with differing functional groups) and two sorbents, we tested variations of solution:sorbent ratio, methanol content, and PFAS initial concentration, and compared derived K_sw values. Each methodological component affected log(K_sw), usually by suppressing the value (by 0-48%) when compared with a "best design" procedure. Thus, we suggest (1) a reference procedure for PFAS and sorbents used here, and (2) general guidelines for batch experiment design with different compounds and sorbents. Additionally, we report well-constrained K_sw values for 23 PFAS and two sorbents.