Trifluoroacetic acid occurrence across source and consumer facing water categories in Ghana: targeted LC-MS/MS...

Trifluoroacetic acid (TFA) is a highly mobile and persistent ultra-short-chain PFAS that is seldom included in routine surveillance in under monitored regions. This non-probability, stratified Ghanaian case study screened 260 groundwater, surface water, treated water, packaged water and rainwater samples by direct aqueous LC-MS/MS. TFA identity was supported by calibration with a certified native standard, three MRM transitions and reproducible proximity to a ¹³C₂-labelled internal standard; the observed 0.040 min median retention time offset was treated as supportive rather than standalone identification evidence. A response at the expected difluoroacetic acid (DFA) transition was retained only as an exploratory DFA-like signal because the method used one transition, showed early chromatographic elution and lacked a DFA-specific labelled internal standard. TFA was detected in 243 samples (93.5%); the tentative DFA-like signal occurred in 152 samples (58.5%). Agricultural groundwater had the highest TFA mean (10.05 µg/L), 95th percentile (41.62 µg/L) and maximum (48.99 µg/L), identifying it as the principal follow-up stratum within the sampled design rather than demonstrating an agricultural source. Nineteen samples exceeded the precautionary 10 µg/L screening benchmark for unregulated drinking water substances, whereas none exceeded the TFA-specific 60 µg/L health-based guidelines. Treated and packaged water categories had lower unmatched concentrations than source water categories, but these contrasts do not measure treatment removal. TFA rankings and benchmark classification were unchanged across four treatments of non-detects. The study provides a transferable Ghanaian occurrence screening framework that combines analytical confidence classification, censoring sensitivity and explicit limits on causal, treatment and health interpretation. Furthermore, to address the possibility that the DFA-targeted response could arise from TFA during electrospray ionisation, the chromatographic evidence was re-evaluated: the DFA-like response eluted at a median 1.070 min, clearly earlier than TFA at 1.430 min. This separation argues against a simple post column/in-source TFA-derived product, but a dedicated TFA-only interference challenge monitored at m/z95.2 > 51.1 was not observed; therefore an ESI-related artefact cannot be definitively excluded and the signal remains unconfirmed.