Shirley Pu, Heather D. Whitehead, James McCord, Timothy J. Buckley, Andri Dahlmeier, Stefan Saravia, Rosie Rushing, Marla P. DeVault, Nathaniel Charest, Valery Tkachenko, A. J. Williams, Jon R. Sobus
Non-targeted analysis (NTA) enables the detection of novel chemicals but is limited in providing quantitative estimates for risk-based interpretation. Quantitative NTA (qNTA) using surrogate chemicals has been previously explored, yielding metrics for qNTA performance and proof-of-principle approach comparisons across matrices. Here, we apply that work in an integrated qNTA workflow, with demonstration on surface and groundwaters impacted by historic PFAS waste. Commercially available surrogate chemicals ( n = 37) spiked into pooled matrix provided initial calibration data and were used for qNTA estimation via a naïve bounded response factor approach. A validation subset, using paired NTA and targeted analysis estimates (16 PFAS and 129 paired values), showed median accuracy within a factor of 2, an uncertainty fold-range of 12, and overall reliability of 85%. Lower than expected reliability indicates an underestimation of uncertainty, likely from suboptimal surrogate selection. The validated qNTA workflow produced concentration and uncertainty estimates for 210 individual PFAS, with legacy and emerging PFAS estimated concentrations as high as parts-per-billion. PFAS with available standards ( n = 22) contributed, on average, to 91% of the estimated sum concentrations. These findings suggest quantitative estimation for chemicals identified via NTA is valuable to ensure that exposure, hazard, and risk assessments consider the total PFAS burden for impacted watersheds.