Structure-Based Prediction and Ranking of PFAS Groundwater Mobility for EU Drinking Water Regulation

Abstract The EU Drinking Water Directive (2020/2184) mandates monitoring of 20 priority per- and polyfluoroalkyl substances (PFAS) from January 2026, yet predicting which PFAS structures pose the greatest groundwater contamination risk remains challenging. We present a machine learning framework that simultaneously predicts the three physicochemical properties governing PFAS subsurface mobility─soil sorption (log KOC, R2 = 0.90), water solubility (log Sw, R2 = 0.90), and ionization class (93% accuracy)─from molecular structure alone. The models were trained on 122 PFAS spanning 37 subclasses using 44 molecular descriptors, including nine that encode the segmentation of the fluorinated backbone by ether linkages, and were validated by cross-validation, external test sets, leave-one-subclass-out analysis, Y-randomization, and a formal applicability domain. SHAP analysis indicates that ether-linkage descriptors are associated with lower predicted soil retention, providing a model-based explanation for the high mobility of GenX-type replacements. A composite Groundwater Mobility Index integrating the three end points identifies PFBA, PFPeA, and PFPeS as the most mobile EU-regulated compounds, consistent with field monitoring evidence, and flags several nonregulated PFAS with comparable mobility. An open-access prediction tool with applicability-domain screening accompanies this work.