Passive sampling is useful for monitoring of PFASs in water because of its high representativeness through determining time-weighted average concentration (C_TWA) across the sampler exposure period. The rate of the diffusive transport of PFASs to a passive sampler is represented by sampling rate (R_s), which is used to calculate the aqueous PFAS concentration using the amount of PFASs accumulated in the sampler. Therefore, a reliable determination of C_TWA by passive sampling requires an accurate determination of R_s. Among various environmental factors that affect R_s, the water flow velocity effect has been identified as the most significant. Currently, the correlation between the water flow velocity and R_s is determined using a large-scale (≥ 500 L) device that simulates field-scale water flows. The use of large-scale devices is costly and labor-intensive, and it limits the number of compounds and flow velocity conditions to be experimentally studied. This study aims to address this concern by developing a new experimental approach that utilizes a small-scale module to investigate the correlation between the flow velocity and R_s of a passive sampler for PFASs. A small-scale (500 ml) device was developed by adopting the configuration of a direct-contact membrane distillation module. By applying seven different flow velocities to the module, we accurately determined R_s for 11 PFASs. The variation of flow velocity resulted in up to a 3.1-fold variation in R_s, which verified the significance of the flow velocity effect on R_s for passive sampling of PFAS. Because the flow regime in the small-scale module (Re = 512, laminar flow) was different from that in the field, the module flow velocity (v_module)-R_s correlation may not directly represent the field flow velocity (v_field)-R_s correlation. To resolve this issue, we determined the mass transfer coefficient (k_w) of calcium sulfate in the small-scale device and in a large-scale water tank (1,040 L) under various flow velocity conditions. The experiments showed that v_module = 3.6 v_field + 9.9 satisfies identical k_ws. This empirical relationship enabled extrapolating the v_module-R_s correlation to the v_field-R_s correlation. Our study demonstrates the promise of the new experimental approach to accurately determining the flow velocity effect on R_s of a passive sampler.

Hyun-Ah Kwon

Appearances