Atmospheric PFAS Partitioning and Source Attribution Using a Trajectory-Informed Targeted and Non-Targeted Approach: Insights from Seoul

Per- and polyfluoroalkyl substances (PFASs) are a large class of synthetic chemicals that are globally distributed due to extensive industrial use, exceptional chemical stability, and long-range atmospheric transport. Many PFASs have been linked to adverse toxicological effects, including developmental, immunological, and endocrine disruption, raising concerns regarding chronic human and environmental exposure. Despite increasing regulatory attention, substantial knowledge gaps remain regarding the atmospheric occurrence, composition, and source influences of both legacy and emerging PFASs, highlighting the need for continuous monitoring.

This study investigates the atmospheric occurrence, distribution, and source-related characteristics of PFASs in the urban environment of Seoul, Korea, by integrating targeted quantification with non-targeted screening. A total of 21 ambient air samples were collected between February and June 2024 on the rooftop of the Korea Institute of Science and Technology (KIST) and analyzed using high-resolution Orbitrap mass spectrometry. Thirty-two ionic and neutral PFASs were quantified, with total concentrations ranging from 17.0 to 348 pg m⁻³. Short-chain perfluoroalkyl carboxylates and sulfonates, including perfluorobutanoic acid (PFBA) and perfluorobutanesulfonic acid (PFBS), were identified as dominant contributors, consistent with the increasing use of short-chain alternatives.

Gas–particle partitioning of PFASs was dominated by temperature effects. Across the campaign, TSP-normalized log Kp values spanned several orders of magnitude, indicating large compound-to-compound differences in aerosol affinity. For most measured PFASs. For most measured PFASs, log Kp was positively correlated with 1/T, indicating that increasing air temperature shifted gas–particle partitioning toward the gas phase. This temperature dependence was most evident for short- to mid-chain PFCAs (perfluoroalkyl carboxylic acids) and for several PFSAs (perfluoroalkyl sulfonic acids) and precursor compounds. By contrast, longer-chain homologues exhibited weak or nonsignificant temperature dependence, consistent with stronger particulate association. Relative humidity showed no statistically significant influence for most compounds; notably, perfluoroethoxyethanesulfonic acid (PFEESA) was the sole species with a strong positive association with humidity, indicating increased particle-phase partitioning at higher humidity. These results highlight temperature as the key meteorological variable to consider when interpreting and modeling PFAS phase partitioning in urban air.

Non-target screening conducted using the FluoroMatch Modular workflow revealed 43 additional PFAS-like features with annotation confidence levels of D - or higher, indicating the presence of a diverse set of previously uncharacterized compounds. To evaluate potential source influences, air-mass back trajectories were clustered into five distinct groups and further examined using partial least-squares discriminant analysis (PLS-DA). Each cluster exhibited a characteristic PFAS profile, reflecting differences in transport pathways and regional influences. Air masses associated with transport over the Yellow Sea (Clusters 2 and 3) showed the highest numbers of unidentified PFAS features (5 and 34, respectively), suggesting enhanced regional contamination or complex source contributions. Selected formulas, including C₃HF₅O₃ (Cluster 2, B–) and C₇H₈F₆N₂O₂ (Cluster 3, D), were identified as indicative features based on cluster specificity and annotation confidence rather than definitive source markers.

Overall, this trajectory-informed analytical framework improves the understanding of PFAS behavior in urban air and demonstrates the value of combining targeted and non-targeted approaches for identifying emerging PFASs and assessing their potential source regions.