Finding the Fluorine: Non-Target Analysis Illuminates Atmospheric PFAS Deposition Trends 
Quinn Whiting1, Alexander Frie1, Samantha McClung2, Bridget Ulrich3, Martin Schafer4, Emily Sellers4, Iris Bloede4, Jared Kunick4, and Elizabeth O'Gorman4
1 University of Minnesota - Duluth, Minnesota Sea Grant, Duluth, MN, 2 University of Minnesota - Duluth, Duluth, MN, 3 University of Minnesota Natural Resources Research Institute, Duluth, MN, 4 National Atmospheric Deposition Program, Wisconsin State Lab of Hygiene, Madison, WI
Per- and polyfluoroalkyl substances (PFAS) have been detected globally in wet deposition samples suggesting that the atmosphere is one mechanism for global transport of this persistent class of chemicals. Currently, long-term monitoring studies (i.e., years) of precipitation samples for PFAS are minimal. As part of the National Atmospheric Deposition Program-National Trends Network (NADP-NTN) PFAS weekly wet deposition samples were collected across 5 sites in Minnesota and Michigan. Targeted analysis of 33 PFAS was conducted on weekly samples to obtain high resolution temporal trends. Additionally, total extractable organic fluorine (EOF) was used to obtain a "total fluorine" contamination level on monthly composite samples. Target PFAS were found to explain ~2-5% of EOF, suggesting a substantial "fluorine-gap". To investigate other fluorinated species contributing to EOF, an open-source non-target analysis (NTA) workflow was created to identify novel, emerging, and other organic-fluorine containing pollutants. This approach identified ~1200 fluorinated features, ~60% were low confidence (unequivocal molecular formula), ~35% were medium confidence (tentative structure), and ~5% were high confidence features (confirmed with analytical standards). Semi-quantification of probable fluorine containing features increased the explained fluorine by an order of magnitude, reaching ~20-60% of EOF. Features contributing to the additional explained fluorine include agrochemicals, pharmaceuticals, traditional PFAS not screened for in the target method, and unique PFAS precursors and transformation products. The 2-year monitoring duration enabled temporal trends to be investigated across targeted PFAS and NTA features, with estimated concentrations of the sum of fluorine containing features increasing in the spring and summer while decreasing in the winter. Temporal patterns support the use of long-term monitoring to determine the overall input of atmospheric deposition, as sampling for shorter periods (days-months) may under- or over-estimate annual loading.