Adsorption Model Evaluation of Perfluoroalkyl Substances (PFAS) in Saturated Media

Abstract

PFAS have been the target of increasing interest in recent years due to their persistence in environmental media and negative health effects on human health. In addition, PFAS is found in numerous environments around the world, where it often displays nonlinear sorption behavior due to its complex composition. This study conducted regression analysis on PFOS sorption results from 28 studies (n = 209) in order to analyze the effects of solid-phase properties on PFOS sorption, and evaluate the performance of 9 proposed sorption models (Kd) at environmental concentrations (0.05 mg/L and 0.1 mg/L). Correlations for the entire dataset showed that organic carbon has the largest contribution to PFOS sorption (R2 = 0.75), with silt and clay content, pH, CEC, and metal oxide content not attributing significant influence at the entire population level. Model estimates for the entire dataset were most accurate (R2 = 0.82) when employing multi-component sorption models that utilize both organic carbon and silt and clay content. Separating the above solid-phase population by OC content indicated that aquifer materials (OC < 0.2%) exhibited the highest sorption influence from silt and clay content (R2 = 0.58), whereas the remaining soils exhibited highest sorption influence by organic carbon content (mineral soils: R2 = 0.11, organic soils: R2 = 0.35). Additional analysis based on organic carbon and silt and clay content population differentiation showed that a two-component sorption model had the highest sorption prediction accuracy (R2 = 91 for high OC-low clay soils, R2 = 0.72 for low OC-high clay soils). Kd models employing components with aqueous calcium concentrations had high estimation accuracy (R2 = 0.94 for the entire dataset) but were limited in contributing sample count (n = 48) due to limited literature reporting and experimentation.