LAUSR

Ecotoxicological risks of perfluoroalkyl and polyfluoroalkyl substances (PFASs) in surface waters are difficult to model because data on PFASs distribution in multiple compartments (sediments, suspended particles and aqueous phase) are difficult to predict. This study quantified the distribution of 21 PFASs including PFCAs: C6-C13 perfluoroalkyl carboxylates, C4, C6, C8 and C10 perfluoroalkane sulfonates, 5 perfluorooctane sulfonamide substances (FOSAMs, including EtFOSA, FOSA, MeFOSAA, EtFOSAA, FOSAA), 2 N-alkyl perfluoroalkane sulfonamidoethanols (MeFOSE and EtFOSE), bis (perfluorooctyl) phosphinic acid (C8/C8 PFPIA), and 5:3 fluorotelomer carboxylic acid (5:3 acid) between bulk water and suspended particles in water column, and pore water and benthic sediments from a tropical urban water body. The distribution of PFASs between sorbed and dissolved phase was largely dependent on the perfluoroalkyl chain length (NCF2). PFCAs with NCF2 > 11 and perfluorodecane sulfonate (PFDS, NCF2 = 10) were found predominantly in the suspended particles and sediments. By contrast, short-chain PFASs (NCF2 ≤ 7) were detected predominantly in the dissolved phase. Sediment acts as a sink for long-chain PFASs while short-chain PFASs are more easily transported via the aqueous phase. Compared with benthic sediments, suspended particles, especially those in the top water layer, carried much higher concentrations of PFASs (by a factor of >100), indicating the stronger sorption capability of suspended particles. The wide variation in PFAS concentrations in suspended particles (∑PFASs concentrations: < 26.8-1,284 ng/g d.w.) suggests that some suspended particles were preloaded with different concentrations of PFASs in the water column which could highly affect the distribution of PFASs in the aquatic environment. Pore water contained 1-2 times higher concentrations of PFASs (∑PFASs: <20.25-159.34 ng/L) than overlying bulk water (∑PFASs: <14.2-79.98 ng/L), indicating the accumulation of PFASs in pore water. Distribution coefficients (KD) were calculated using paired solids concentration and dissolved concentration in both water and sediment column (KD-SP and KD-SED respectively) and were compared with values derived from a laboratory batch experiment. The averaged Log KD-SP/Log KD-SED showed significant positive correlation with NCF2, except for short-chain PFASs (NCF2 <5) which presented higher Log KD values than estimated. The discrepancies found between KD (Log KD-SP > Log KD -desorption > Log KD -sorption > Log KD-SED) suggest that the distribution of PFASs in the field, especially between suspended particles and bulk water could not be well represented by lab results and that using the water concentrations in the bottom layer for estimation of pore water concentrations could lead to bias results.