Drinking waters contamination by per‐ and poly‐fluoroalkyl substances of the anthropized area of Fos‐Berre, France Authors: M. Dijoux, A. Piram, S. Augy, A. Austruy, A. Souloumiac, L. De Jong, E. Gonzales‐Camoin, J. Dron, P. Chamaret, X. Moreau, S. Gori, P. Wong‐Wah‐Chung, P. Doumenq Published: 2026 ABSTRACT AbstractPer‐ and poly‐fluoroalkyl substances (PFAS) contamination and geographical distribution were studied in drinking water from the industrial harbor of Fos‐Berre, a highly industrialized area in southern France. An analytical method for 38 PFAS was developed, using large volume injection ultra‐high performance liquid chromatography‐tandem mass spectrometry. To minimize the risk of contamination during preparation, samples were directly injected onto a SCIEX Triple Quad 7500 system after dilution in acidified methanol. Method limits of quantification ranged from 0.081 to 2.2 ng L−1, except for 8:2 fluorotelomer unsaturated carboxylic acid and perfluoropentanoic acid (5.4 ng L−1). The method was applied to 28 tap waters and two groundwater from 28 stations on the Crau groundwater table and Durance River. Eighteen PFAS were quantified in at least one sample, with concentrations from 0.14 for linear perfluorooctane sulfonic acid (PFOS) to 62.85 ng L−1 for 6:2 fluorotelomer sulfonic acid. Short‐chain carboxylic acids (perfluorocarboxylic acids) and sulfonic acids (perfluorosulfonic acids) were the most frequently quantified, with perfluorohexanoic acid showing the highest frequency (73%), followed by perfluorooctane carboxylic acid, branched PFOS, branched perfluorooctane hexanoic acid (PFHxS), linear PFOS and linear PFHxS. Total PFAS ranged from <0.081 to 126 ng L−1 and the sum of the 20 regulated PFAS remained below the 2020/2184 drinking water directive limit of 100 ng L−1 except for one sample (119 ng L−1). Contamination levels and profiles varied across sites, reflecting catchment location, surrounding industrial activities, and groundwater flow, with higher concentrations detected downstream. Abbreviations aqueous film‐forming foams Crau groundwater table certified reference material drinking water directive European Food Safety Authority formic acid raw groundwater limit of quantification method quantification limit classified installations for environmental protection per‐ and poly‐fluoroalkyl substances perfluorocarboxylic acids perfluorosulfonic acids quality control tap water tolerable weekly intake INTRODUCTION Per‐ and poly‐fluoroalkyl substances (PFAS) comprise between 4730 (OECD, 2018) and 14,735 (U.S. EPA, 2022) partially or fully fluorinated aliphatic compounds, differing by their functional groups (Buck et al., 2011; OECD, 2021) and in the perfluorinated chain length. Carboxylic acids (perfluorocarboxylic acids [PFCAs]) with chain length above seven carbons (above six for sulfonic acids—perfluorosulfonic acids [PFSAs]) are called “long‐chain” PFAS, while those with fewer are “short‐chain,” and those with two to three carbons “ultra short.” Synthesized since the 1940s for their water and grease repellence, heat resistance and nonstick properties (3M, 1999), PFAS are used in firefighting foams, electronics, food contact materials, and textiles (Glüge et al., 2020). During production and use, they are released into air, wastewater, solid waste, and landfills (Dauchy et al., 2017). Their amphiphilic nature and persistence—especially for PFCAs and PFSAs—lead to transfer and accumulation in all environmental compartments (Nakayama et al., 2019): air, soils (Brusseau et al., 2020), sediments, waters, drinking water (Boiteux et al., 2012; Teymoorian et al., 2023), crops, and livestock, resulting in widespread exposure of all forms of life. Despite active research, remediation remains costly and energy intensive (Fang et al., 2024). Their ubiquitous presence in environmental matrices raises public health concerns due to their potentially significant toxicity. In 2020, the European Food Safety Authority (EFSA) identified association between PFAS absorption and increased risks of thyroid dysfunction, hypercholesterolemia, cancer, and adverse reproductive outcomes. Accordingly, EFSA established a tolerable weekly intake (TWI) of 4.4 ng kg− 1 body weight (bw) for the combined exposure to perfluorooctane sulfonic acid (PFOS), perfluorooctane carboxylic acid (PFOA), perfluorooctane hexanoic acid (PFHxS), and perfluorononanoic acid (PFNA) (EFSA CONTAM Panel et al., 2020). More recently, the French Agency for Food, Environmental and Occupational Health and Safety defined long‐term toxicological reference values for perfluorobutanoic acid (PFBA) (20 µg kgbw −1 j−1), perfluorohexanoic acid (PFHxA) (20 µg kgbw −1 j−1), and 6:2 fluorotelomer sulfonic acid (FTS) (0.33 µg kgbw −1 j−1) (ANSES, 2025). Furthermore, the European Drinking Water Directive (DWD) (2020/2184; European Parliament, 2020) established a regulatory limit for the sum of 20 PFAS (10 PFCAs and 10 PFSAs), to be enforced from 2026 (see detailed list in Table S1) (European Parliament, 2020). Studying PFAS behavior in the environment requires highly sensitive analytical methods capable of detecting ultra‐trace concentrations (ng L− 1). Given the high risk of sample contamination and compound losses through adsorption or volatilization, rigorous handling procedures are essential, with minimized preparation steps to reduce these risks. This study aimed to assess PFAS occurrence in drinking water from the Gulf of Fos and the Berre Lagoon (Fos‐Berre area), a heavily industrialized region in southern France (“Région Sud”) that hosts no PFAS production facilities. An optimized analytical protocol was therefore developed for 38 PFAS (see Section 2.2) and applied to 28 tap water (TW) samples mainly supplied from collective water catchments of the Crau groundwater table (CGWT), and two groundwater samples. MATERIALS AND METHODS Sampling site description The Crau Plain, located west of the Berre Lagoon and north of the Gulf of Fos in southern France, covers approximately 550 km2 and hosts the CGWT, a major strategic freshwater resource for drinking water, agriculture, and industry. The CGWT supplies about 270,000 inhabitants in 13 municipalities via 18 public well fields and numerous private wells. Approximately 70% of its recharge originates from traditional irrigation of Crau hay meadows, with the remaining 30% derived from precipitation (SYMCRAU, 2019), which produces an artificial annual cycle characterized by high water levels in summer and low levels in winter. In the absence of an overlying impermeable layer, the aquifer is highly vulnerable to contamination from surrounding human activities. To characterize the influence of anthropogenic pressures, the study area included rural, urban (about 10% of the surface), and industrial settings. In the northern and northwestern sectors, agricultural practices may impact groundwater quality through the use of fertilizers and pesticides. The southern and eastern sectors are affected by emissions from the chemical, petrochemical, and metallurgical activities of the Fos‐sur‐Mer industrial harbor, the largest industrial port complex in France, and by waste incineration plants, extensive oil and gas pipeline networks, a military air base, and intense maritime, road, and air traffic. Additional sources of contamination include fuel storage, aircraft cleaning with products containing fluorinated surfactants, and the frequent use of aqueous film‐forming foams (AFFF) on military and airport facilities (SYMCRAU, 2024). Although no PFAS production plants are reported in the area, seven classified installations for environmental protection located directly above the CGWT are subjected to PFAS monitoring in industrial effluents according to the French Decree of June 20, 2023. Core Ideas Thirty‐eight per‐ and poly‐fluoroalkyl substances (PFAS) were investigated in 30 water samples.Eighteen PFAS were quantified.Perfluorohexanoic acid (PFHxA), perfluorooctane carboxylic acid (PFOA), branched (Br)‐perfluorooctane sulfonic acid (PFOS), Br‐perfluorooctane hexanoic acid (PFHxS), linear (L)‐PFOS, and L‐PFHxS were found in >50% of samples.Short‐chain perfluorocarboxylic acids (PFCAs) and perfluorosulfonic acids (PFSAs) were the most frequently quantified PFAS.One sample exceeded the European Directive of 100 ng L−1 for ∑ 20 PFAS. Thirty‐eight per‐ and poly‐fluoroalkyl substances (PFAS) were investigated in 30 water samples. Thirty‐eight per‐ and poly‐fluoroalkyl substances (PFAS) were investigated in 30 water samples. Eighteen PFAS were quantified. Eighteen PFAS were quantified. Perfluorohexanoic acid (PFHxA), perfluorooctane carboxylic acid (PFOA), branched (Br)‐perfluorooctane sulfonic acid (PFOS), Br‐perfluorooctane hexanoic acid (PFHxS), linear (L)‐PFOS, and L‐PFHxS were found in >50% of samples. Perfluorohexanoic acid (PFHxA), perfluorooctane carboxylic acid (PFOA), branched (Br)‐perfluorooctane sulfonic acid (PFOS), Br‐perfluorooctane hexanoic acid (PFHxS), linear (L)‐PFOS, and L‐PFHxS were found in >50% of samples. Short‐chain perfluorocarboxylic acids (PFCAs) and perfluorosulfonic acids (PFSAs) were the most frequently quantified PFAS. Short‐chain perfluorocarboxylic acids (PFCAs) and perfluorosulfonic acids (PFSAs) were the most frequently quantified PFAS. One sample exceeded the European Directive of 100 ng L−1 for ∑ 20 PFAS. One sample exceeded the European Directive of 100 ng L−1 for ∑ 20 PFAS. The study area and sampling locations are shown in Figure 1. In June 2024, 30 water samples (mainly TW) were collected from 28 sites. Sampling points (A–O) are represented by triangles and are named according to their associated catchments, which are shown as diamonds for CGWT and as a square for the Durance River; black triangles indicate private wells, and an asterisk identifies those used for drinking water supply. Map of the 30 water sampling sites (triangles). Colored triangles are linked by lines to their collective catchments (from drinking water treatment plants) (diamonds for the Crau groundwater table (CGWT) and square for the Durance River) while black triangles indicate private catchments (private well from individual household). Sample identity: A–O, with * indicating drinking water supplied by the private catchments. Groundwater flow directions were obtained from the SYMCRAU. GW, raw groundwater; TW, tap water. All 28 TWs were collected from private households. Three TWs were supplied from one collective catchment on the Durance River, 20 TWs from nine collective catchments on the CGWT, and five TWs from individual water catchments located in the vicinity of the CGWT (except for site D*). For two of the five sites using individual water catchment, the groundwater (used as TW source) was also collected, from pipes in the gardens, to evaluate the efficiency of the domestic treatments. Groundwater analysis was not part of the initial experimental design. It was performed only to compare baseline PFAS concentrations with those measured in TW after treatment using domestic purification systems, with the aim of evaluating the removal efficiency of these processes. Additional information on catchment type and the presence of treatment or filtration is provided in Table 1.​ TABLE 1Information on the water samples, catchments (A–O), household individual, and collective treatments.Sample IDWater originT (°C)Samples coord.Household water treatment systemAssociated water catchment and IDCatchment coord.Catchment typeCollective water treatment systemA a Groundwater at the edge of the CGWT17.943.692; 5.045NoIndividualA a 43.692;5.045Individual/BCGWT21.943.645; 4.723NoSt HippolyteB43.637133; 4.750683CollectiveChlorine gasCCGWT24.143.635; 4.813NoVal BoiséC43.641855; 4.809606CollectiveChlorine gasD a CGWT22.443.673; 4.892NoIndividualD a 43.637; 4.847Individual/E1Durance22.343.409; 5.013NoDuranceE43.64023; 5.02513CollectiveOzone and/or chlorineE222.543.396; 5.059NoE321.543.453; 5.015NoFCGWT21.143.609; 5.071NoMary RoseF43.620012; 5.048790CollectiveNo informationG1CGWT2043.589; 5.019NoPuits des Canaux JumeauxG43.605493; 4.958637CollectiveChlorine gasG222.943.549; 5.035NoH1CGWT20.443.490; 5.00120 µm sediment filterLa CaspienneH43.538622; 4.941626CollectiveChlorine gasH223.443.494; 4.990NoH324.143.490; 4.992NoI‐TW a Groundwater at the edge of the CGWT23.843.532; 4.99310 µm filter, softener, osmosis with activated carbon, U.V.IndividualI a 43.532; 4.993Individual/I‐GW a 16.643.532; 4.993NoJ a Groundwater at the edge of the CGWT21.343.528; 5.000NoIndividualJ a 43.528; 5.000Individual/K1CGWT22.443.425; 4.986NoDes TapiesK43.486058; 4.920420CollectiveDirect chlorinationK221.843.425; 4.986NoK321.543.403; 4.981NoK423.243.404; 4.980Old salt softenerK524.743.398; 4.981NoL1CGWT25.843.434; 4.948NoFanfarigoule‐BAL43.484885; 4.948613CollectiveDirect chlorinationL222.343.430; 4.941NoL320.543.459; 4.940NoL424.743.454; 4.935NoMCGWT22.843.490; 4.877NoVentillonM43.478763; 4.926804CollectiveNo informationN1CGWT20.943.389; 4.848No except for hot water + asbestos pipeLa PissaroteN43.490395; 4.846588CollectiveNo informationN221.543.389; 4.805Softener, cottonO‐TW a Groundwater at the edge of the CGWT17.943.463; 4.947Taste filter, textile cartridge (cotton), glyphosate filter (crystals), carbon filterIndividualO a 43.463; 4.947Individual/O‐GW a 17.643.463; 4.947No Note: Sample coord. and Catchment coord. correspond to the latitude and longitude coordinate of sampling sites and the corresponding water catchment. /: no information.Abbreviations: CGWT, Crau groundwater table; GW, raw groundwater; T (°C), temperature of the water sampled; TW, tap water, GW = raw groundwater. a Drinking water supplied by the private catchments.John Wiley & Sons, Ltd. Information on the water samples, catchments (A–O), household individual, and collective treatments. Note: Sample coord. and Catchment coord. correspond to the latitude and longitude coordinate of sampling sites and the corresponding water catchment. /: no information. Abbreviations: CGWT, Crau groundwater table; GW, raw groundwater; T (°C), temperature of the water sampled; TW, tap water, GW = raw groundwater. Drinking water supplied by the private catchments. Previous investigations in the Fos–Berre region have documented trace elements and organochlorine contaminants in soils (Austruy et al., 2021), as well as PFAS in drinking water (ARS, 2025b) and in marine biota, although quantitative PFAS data for this area remain limited. The collection of tap water and groundwater samples in this study was made possible through collaboration with volunteers from the Eco Citizen Institute for the Knowledge of Pollution within the VOCE citizen science network (Volunteers for the Citizens’ Observation of the Environment). Chemicals and standards Thirty‐five PFAS regulated by either the DWD 2020/2184, the ISO Standard 21675, or the French ministerial ruling for classified installations for environmental protection (PCI) were purchased: 13 PFCAs (C3–C18), 10 PFSAs (C4–C13), two FTS, two perfluoroalkyl ether carboxylic acids, one perfluoroalkyl ether sulfonic acid, one fluorotelomer unsaturated carboxylic acid (FTUCA), one polyfluoroalkyl phosphoric acid diester, three perfluoroalkyl sulfonamides (FASAs), and two perfluoroalkyl sulfonamido acetic acids. For that, a 2 mg L−1 mixture of 29 PFAS in methanol was purchased from Accustandard (ISO 21675 R1) (TechLab), and other PFAS were purchased as individual methanolic solutions of 2, 50, or 100 mg L−1 from Wellington laboratories (B.C.P. Instruments), Accustandard, Cambridge Isotope Laboratories Inc (CIL) (TechLab), and Chiron AS (BCP Instruments). PFOS, PFHxS, and N‐ethyl perfluorooctane sulfonamido acetic acid (N‐EtFOSAA) standards contained a mixture of linear (L‐) and branched (Br‐) isomers, making a list of 38 PFAS. Seven isotopically labeled PFAS were purchased from CIL as two commercial mixtures (ES‐5610A: 13C8‐PFOA, 13C8‐potassium perfluoro‐1‐octanesulfonate [PFOS], and d3‐N‐methyl perfluorooctane sulfonamido acetic acid [d3‐NMeFOSAA] at respective concentrations of 1, 3, and 4 mg L−1; and ES‐5632: 13C3‐sodium tetrafluoro‐2‐(heptafluoropropoxy)propanoate [13C3‐HFPO‐DA], 13C6‐sodium perfluoro‐n‐hexanoate [13C6‐PFHxA], 13C9‐sodium perfluoro‐n‐decanoate [13C9‐PFDA], and d5‐NEtFOSAA at respective concentrations of 1, 1, 1, and 4 mg L−1) and used as internal standard (IS) stock solutions. In the rest of the document, we will refer to the acidic form of the compounds. The full list of compounds with their chemical names and abbreviations can be found in Table S1. The standard and isotopically labeled standard solutions were prepared by dilution in methanol and were stored in polypropylene vials at 4°C or −20°C depending on the supplier's recommendation. ULC‐MS grade methanol, water, and ammonium acetate were purchased from Biosolve Chimie. Formic acid (FoA) was obtained from Fisher Scientific SAS. Evian mineral bottled water was used for calibration and quality control (QC) solutions, after assessing the absence of PFAS contamination. Polypropylene vials were purchased from Agilent Technologies France with silicon caps. A 250 g L−1 ammonium acetate (Biosolve Chimie) solution was prepared as preservative solution. Sample collection Samples were collected within 2 week in June 2024 using 250‐mL high‐density polyethylene bottles prewashed with ultrapure water and methanol to minimize PFAS contamination. At each site, taps were flushed for 5 min to stabilize temperature and flow and to remove any tap‐accumulated contamination. Bottles were then rinsed three times with site water before collecting approximately 250 mL. Immediately after collection, 1 mL of ammonium acetate preservative solution was added to each sample to neutralize residual free chlorine and prevent degradation, in accordance with USEPA Method 533 (US EPA, 2019). Field blanks were prepared by filling high‐density polyethylene bottles with ultrapure water on site and exposing them to the sampling environment to assess potential contamination during collection and transport. To minimize contamination, waterproof clothing and food packaging were avoided during sampling, non‐powdered nitrile gloves were worn, samples were placed individually in labeled plastic bags, and no permanent markers were used. All fluoropolymer‐containing equipment was replaced with alternative materials. Samples were transported to the laboratory within 24 h in a cooler containing natural ice. Upon arrival, raw samples were stored at −20°C and analyzed within 2 weeks of collection. Sample preparation In this study, samples were prepared by direct injection after dilution with methanol to limit matrix effects, contamination (cross‐contamination, materials, and reagents), and analyte losses by adsorption or volatilization during preconcentration. Polypropylene and polyethylene are generally recommended as PFAS container materials (ISO 21675, 2019; US EPA, 2019), although glass is also frequently used (Zenobio et al., 2022). Some studies report negligible or lower PFAS sorption to polypropylene compared with glass, while others show stronger sorption to polypropylene (Woudneh et al., 2019; Zenobio et al., 2022). Adsorption depends on matrix composition and adding methanol increases PFAS solubility and reduces sorption (Mancini et al., 2023), especially for long‐chain PFAS. With polypropylene vials, a 50:50 MeOH/H2O ratio gave poor recoveries for the longest‐chain PFCA (perfluorooctadecanoic acid). Testing methanol contents from 50% to 99% showed that FASAs exhibited peak asymmetry above 70% methanol. A MeOH/H2O ratio of 60:40 (v/v) with 0.1% FoA was therefore selected to improve long‐chain PFAS solubility without excessively increasing limits of quantification (LQs). Upon arrival at the laboratory, samples were prepared immediately. After thawing at room temperature, 250 mL samples were homogenized by shaking, and 400 µL aliquots were transferred into polypropylene autosampler vials. Each aliquot was diluted with 600 µL methanolic FoA to reach a final MeOH/sample ratio of 60:40 (v/v) with 0.1% FoA, then spiked with ISs. Polypropylene vials with silicone septa were used to minimize contamination. Prepared samples were stored frozen until analysis to limit transformation of polyfluoroalkyl substances, since biodegradation can occur at ambient temperature (Woudneh et al., 2019). Polyfluoroalkyl substances that can transform into more stable PFAS such as PFCAs and PFSAs are frequently called precursors. Analytical methods The analytical method was developed to quantify 38 PFAS in TW, selected from ISO 21675, the French PCI regulation, and the EU DWD 2020/2184. The 38 target PFAS were analyzed using an Agilent 1290 Infinity II UPLC coupled to a SCIEX Triple Quad 7500 tandem mass spectrometer (ultra‐high performance liquid chromatography‐tandem mass spectrometry). Samples (100 µL) were injected onto a Luna Omega PS C18 column (3 µm, 100 Å, 100 mm × 3.0 mm, Phenomenex) protected by a UHPLC ZORBAX Eclipse Plus C18 guard column (2.1 mm × 5 mm, 1.8 µm, Agilent). To prevent background contamination coming from the analytical system, a Gemini C18 delay column (5 µm, 110 Å, 50 mm × 3.0 mm, Phenomenex) was installed between the solvent mixer and the injector. The needle was washed for 10 s between injections with 50:50 MeOH/H2O to limit carryover. The mobile phase consisted of water with 2 mM ammonium acetate (A) and methanol (B). The gradient at 600 µL min− 1 was 10% B for 0.5 min, 10%–50% B over 1 min, 50%–95% B over 6.5 min, 95% B for 4 min, and then re‐equilibration to 10% B over 2.5 min. The column temperature was 40°C. Detection and quantification were performed by MS/MS in scheduled multiple reaction monitoring (sMRM) mode using two transitions per compound, with the most intense transition for quantification and the second for confirmation. Ionization was operated in negative mode with curtain gas, ion source gas 1, and ion source gas 2 set to 40, 50, and 80 psi, respectively, and a source temperature of 350°C. The ion spray voltage was set at 1500 or 2500 V depending on the compound. Optimization from 1000 to 3000 V showed that long‐chain PFAS signals increased with voltage and plateaued at 2500 V, whereas lower molecular weight compounds exhibited maximum sensitivity at 1500 V. PFAS transitions are described in Table S1. Linear and branched isomers of PFHxS, PFOS, and N‐EtFOSAA were quantified separately by integrating their individual chromatographic peaks. Ten calibration levels from 0.1 to 250 ng L− 1 (0.1, 0.5, 1, 2.5, 5, 10, 25, 50, 100, and 250 ng L− 1) and three QCs (1, 5, and 50 ng L− 1) were prepared in ULC‐MS methanol/Evian water (60:40, v/v) to correct for potential matrix effects in TW (Grüning, 2022), after assessing the absence of PFAS contamination. Quantification of the 38 PFAS was performed using an internal calibration with seven isotope‐labeled standards. For analytes lacking labeled analogues, ISs were attributed following their properties (see Table S1). Quality assurance and QC Due to the widespread use of PFAS in everyday objects and laboratory equipment, stringent precautions were implemented to prevent sample contamination during collection and preparation. All fluoropolymers (e.g., Teflon) were avoided throughout the process. One field blank per sampling site was prepared and analyzed, following the same preparation steps as water samples, to assess procedural background and contamination. Solvent blanks (MeOH/H2O 60:40 v/v with 0.1% FoA) were run before and periodically within each batch to evaluate instrumental background and carryover. Method precision, accuracy, and instrument performance were verified for each analytical run through regular analysis of QCs at three concentration levels. Accuracy was further assessed using water certified reference material (CRM IRMM‐428) and by duplicate analysis of five samples at an accredited laboratory. Data acquisition was performed with SCIEX OS software (v. 3.4.0). PFAS were identified based on retention time (±10 s), the presence of at least two selected MRM transitions, and ion abundance ratios (quantitative/qualitative transition) within 30% of acceptable differences. Seven isotopically labeled PFAS served as ISs for the 38 target analytes, selected based on structural similarity and retention time. Quantification employed quadratic regression with 7–10 calibration levels from 0.1 to 250 ng L− 1. Accuracy criteria were set at ±20% for standards and ±30% for QCs. LQs were defined as the lowest calibration level yielding a signal‐to‐noise ratio of 10. To account for potential PFAS background contamination arising from the overall preparation and analytical procedure, despite the precautions taken, we quantified its contribution to each sample's signal. This was achieved by measuring PFAS concentrations in field blanks and calculating the ratio of the PFAS concentration in the blank to that in the corresponding sample. Samples were subsequently classified info three categories depending on the blank ratios value: quantifiable, detectable and non‐exploitable. The blank signal was subtracted from the sample when it was below 25% of the sample signal. For a percentage of blank between 25% and 50% of the sample signal, PFAS were considered as detected but not quantifiable (identified as DET). Results were considered non exploitable and therefore discarded when blank values exceeded 50% (identified as NE). RESULTS AND DISCUSSION Method performance parameters and data analysis All compounds had determination coefficient of their calibration quadratic curves R 2 above 0.99. Among the 38 compounds, method quantification limits (MQLs) ranged from 0.081 for Br‐PFOS to 5.4 ng L−1 for 8:2 FTUCA. The results obtained for the six compounds included in the certificate of analysis of the CRM (perfluorobutanesulfonic acid, PFHxS, L‐PFOS, perfluoropentanoic acid [PFPeA], PFHxA, and perfluoroheptanoic acid) showed a relative deviation from the certified values of less than 30% (Table S2). To compare our method with the method of the accredited laboratory, five TWs were sampled three times following the three different approaches: (1) developed method used in this study (polypropylene bottle, ammonium acetate) and analyzed by our laboratory; (2) collection in polypropylene bottles from the accredited laboratory, no ammonium acetate added, analyzed by our laboratory; (3) collection in polypropylene bottles from the accredited laboratory, no ammonium acetate added, analyzed by the accredited laboratory (33 PFAS in common with our study). Detailed results can be found in Tables S3 and S4. Briefly, the comparison of methods 2 and 3 showed that 93% of the results obtained from method 2 were compliant with those of method 3 (relative error of less than 30% of the certified values), including 72% of compliance for the quantified results and 98% of compliance for non‐quantified results (see Table S3). This showed that both analytical methods mostly gave the same results. When comparing methods 1 and 2, 97% of the results were in agreement, including an 85% and 98% agreement for quantified and non‐quantified results, respectively (see Table S4). This shows that results were similar no matter the container or the addition of ammonium acetate. Finally, methods 1 and 3 were also compared, showing that 93% of our results were compliant with those of the accredited laboratory, including 72% and 98% of compliance for quantified and non‐quantified results respectively, assessing the precision of our method, no matter the collection and storage procedure or the analytical method (see Table S3). Differences in concentration were mostly obtained for PFAS that were present at high levels in the blanks, due to lab contamination. Indeed, despite the precautions taken, contamination by a few PFAS was observed in the solvent blanks and field blanks. The use of a sensitive device to lower the limits of quantifications can explain the more visible contamination. PFBA (at low levels but above LQ) and 6:2 FTS (at very high levels) were detected in the solvent blanks, suggesting contamination from the solvent themselves or from the laboratory during preparation. Most of the field blanks also contained some PFAS, primarily 6:2 FTS, PFBA, perfluoropentane sulfonic acid, PFHxA, PFOA, and L‐PFOS, with mean concentrations of 49.3, 1.80, 1.39, 0.357, 0.350, and 0.166 ng L−1, respectively. As contamination levels were very similar between field blanks, the sampling site was ruled out as a source of contamination, and sample preparation was found to be the most likely cause of contamination, either from the laboratory environment or the chemicals added (ammonium acetate, MeOH, FoA). We therefore calculated the blank percentage as described in the QA/QC section and replaced the values by DET or NE when necessary. Although these thresholds allowed us to deal with the unpredictable background PFAS contamination, which was more noticeable due to the high sensitivity of the device, we acknowledge that this approach has limitations and that contamination should be either minimized or addressed through stricter decision thresholds. Finally, open‐source software QGIS was used to map sampling locations and the spatial distribution of PFAS. PFAS occurrence and levels in drinking waters The sampling campaign comprised 28 TW samples supplied by the CGWT and the Durance River, with most originating from collective water catchments and two groundwater. Frequency of quantification; frequency of detection; and minimum, maximum, median, and mean concentrations of PFAS in the samples are summarized in Table 2. Compounds were considered as detected but not quantifiable when the signal‐to‐noise ratio was between 3 and 10. Only quantified values were included in the calculations of the mean and median. Detailed concentrations of each PFAS congener in each sampling site are provided in Table S5, and Figure 2a,b graphically shows the sum of PFAS for each sampling site. TABLE 2Summary of per‐ and polyfluoroalkyl substances (PFAS) occurrence in the 30 drinking waters.PFAS nameMQLs (ng L−1)Frequency of quantification (n = 30)Frequency of detection (n = 30)Minimum (ng L−1)Maximum (ng L−1)Mean (ng L−1)Median (ng L−1) Perfluoroalkyl carboxylic acids (PFCAs) PFBA1.1037%70%2.126.743.72 (n = 11)2.85PFPeA5.4140%57%5.4517.711.5 (n = 12)12.3PFHxA0.21873%90%0.63715.45.77 (n = 22)3.85PFHpA1.0747%50%1.384.712.75 (n = 14)2.48PFOA0.21663%87%0.2682.601.22 (n = 19)0.966PFNA1.080%0%