Sensitivity of PFAS-derived Fluorine Metrics to Target-List Composition, LoQ Distributions, and Non-Detect Handling...

Targeted analytical methods are widely used to quantify per- and polyfluoroalkyl substances (PFAS) in combustion flue-gas. However, interpretation of PFAS emission data is challenged by high shares of non-detects (NDs), heterogeneous limits of quantification (LoQs), and extensive target analyte lists. This study evaluates the influence of ND-treatment, target-list composition and LoQ distributions on PFAS-derived fluorine metrics using two pilot-scale incineration datasets. Both datasets exhibit high ND-shares and analyte-specific LoQs spanning several orders of magnitude. Current analytical target lists appear highly sensitive to the ND-substitution approach. However, sensitivity analyses demonstrated that this apparent sensitivity was governed by a small number of analytes exhibiting elevated LoQs. Excluding these analytes reduced variability among ND-substitution approaches by approximately 95 % in one dataset and by more than 99.9 % in the other, largely independent of the selected descriptive statistical parameter. Target-list reduction also decreased variability but provided little additional improvement. Consequently, the apparent influence of target-list composition was found to be largely attributable to the inclusion of a small number of analytes with exceptionally high LoQs. The findings indicate that apparent sensitivities to ND-treatment are governed primarily by analyte-specific LoQ distributions, and, to a lesser extent, by target-list composition. Robust assessment of PFAS-derived metrics therefore primarily requires critical evaluation of analytes exhibiting elevated LoQs, transparent handling of NDs and the application of lower- and upper-bound estimates. Future method development should prioritize harmonized LoQs, refinement of target lists, and inclusion of volatile PFAS and non-target analysis to improve PFAS emission assessment.