A stochastic approach for parameter optimization of feature detection algorithms for non-target screening in mass spectrometry.

Feature detection plays a crucial role in non-target screening (NTS), requiring careful selection of algorithm parameters to minimize false positive (FP) features. In this study, a stochastic approach was employed to optimize the parameter settings of feature detection algorithms used in processing high-resolution mass spectrometry data. This approach was demonstrated using four open-source algorithms (OpenMS, SAFD, XCMS, and KPIC2) within the patRoon software platform for processing extracts from drinking water samples spiked with 46 per- and polyfluoroalkyl substances (PFAS). The designed method is based on a stochastic strategy involving random sampling from variable space and the use of Pearson correlation to assess the impact of each parameter on the number of detected suspect analytes. Using our approach, the optimized parameters led to improvement in the algorithm performance by increasing suspect hits in case of SAFD and XCMS, and reducing the total number of detected features (i.e., minimizing FP) for OpenMS. These improvements were further validated on three different drinking water samples as test dataset. The optimized parameters resulted in a lower false discovery rate (FDR%) compared to the default parameters, effectively increasing the detection of true positive features. This work also highlights the necessity of algorithm parameter optimization prior to starting the NTS to reduce the complexity of such datasets.

Sorption of per- and poly-fluoroalkyl substances and their precursors on activated carbon under realistic drinking water conditions: Insights into sorbent variability and PFAS structural effects.

Recent stringent drinking water quality standards create challenges for water utilities to meet these standards. Advanced treatment techniques will have to be applied on many drinking water production locations to meet these quality standards. This study investigated the sorption of per- and polyfluorinated-alkyl substances (PFAS) onto granular activated carbon (GAC). The study was performed at environmentally relevant PFAS concentrations and a realistic water-to-GAC ratio, providing a realism often overlooked in existing studies. Three different forms of GAC were evaluated, differing in micropore and mesopore structures. Tap water spiked with 5 ng/L of each of 31 PFAS was used in the sorption experiments, i.e. perfluorocarboxylic acids (C4-C12), perfluorosulfonic acids (PFSA, C5-C10) including linear and branched isomers, and three groups of PFAS precursors (per-/polyfluoroalkyl ether acids, sulfonamides, and sulfonamide acetic acids). The three studied GAC did not exhibit distinct differences in PFAS sorption. The removal of PFAS was below 50 % for most studied PFAS, except for the short-chain PFAS precursors. Sorption was affected by both the carbon chain length and functional groups for PFAS, while this was not observed for PFAS precursors. The presence of ether linkages and sulfonamide groups notably enhanced sorption. Linear and branched PFSA demonstrated similar sorption behavior, whereas branched isomers of the sulfonamide acetic acid precursors exhibited significantly higher sorption. This indicates that sorption was determined by both hydrophobic and electrostatic interactions. Given the relatively low PFAS removal by GAC under environmentally relevant test conditions, further improvements in sorbents are required to ensure that PFAS concentrations in produced drinking water comply with drinking water standards.

Forever legacies? Profiling historical PFAS contamination and current influence on groundwater used for drinking water.

A wide range of PFAS residues were studied in an aquifer used for drinking water production which was affected by historical PFAS contamination from a landfill and military camp. Samples were taken at three monitoring and four pumping wells at different depths ranging from 33 to 147 m below the land surface and analysed for a series of 53 PFAS (C2-C14) and PFAS precursors (C4-C24). A comparison of results to earlier research from 2013, with a more limited range of PFAS, showed decreasing concentrations and migration of PFAS with increasing depth and distance from the contamination source. The PFAS profile and branched/linear isomer ratio are used as source characterization tools. The landfill was confirmed to contaminate the groundwater in both monitoring wells, while the military camp was indicated as a probable source for PFAS observed in the deep sampling points of one of the monitoring wells. Pumping wells used to produce drinking water are not yet affected by these two PFAS sources. In one of the four sampled pumping wells, a different PFAS profile and isomer pattern was observed, which indicated a different but yet unknown source. This work shows the necessity of implementing regular screening to identify potential (historical) PFAS sources to be able to prevent future contaminant migration nearby and towards drinking water abstraction wells.

Occurrence, Fate, and Related Health Risks of PFAS in Raw and Produced Drinking Water.

This study investigates human exposure to per- and polyfluoroalkyl substances (PFAS) via drinking water and evaluates human health risks. An analytical method for 56 target PFAS, including ultrashort-chain (C2-C3) and branched isomers, was developed. The limit of detection (LOD) ranged from 0.009 to 0.1 ng/L, except for trifluoroacetic-acid and perfluoropropanoic-acid with higher LODs of 35 and 0.24 ng/L, respectively. The method was applied to raw and produced drinking water from 18 Dutch locations, including groundwater or surface water as source, and applied various treatment processes. Ultrashort-chain (300 to 1100 ng/L) followed by the group of perfluoroalkyl-carboxylic-acids (PFCA, ≥C4) (0.4 to 95.1 ng/L) were dominant. PFCA and perfluoroalkyl-sulfonic-acid (≥C4), including precursors, showed significantly higher levels in drinking water produced from surface water. However, no significant difference was found for ultrashort PFAS, indicating the need for groundwater protection. Negative removal of PFAS occasionally observed for advanced treatment indicates desorption and/or degradation of precursors. The proportion of branched isomers was higher in raw and produced drinking water as compared to industrial production. Drinking water produced from surface water, except for a few locations, exceed non-binding provisional guideline values proposed; however, all produced drinking waters met the recent soon-to-be binding drinking-water-directive requirements.

Perfluoroalkane substances in national samples from global monitoring plan projects (2017-2019).

The global monitoring plan (GMP) established under the Stockholm Convention on Persistent Organic Pollutants (POPs) had defined ambient air, human milk or blood, and water as core matrices to be analyzed and assessed for spatial and temporal distribution. Within projects coordinated by the United Nations Environment Programme (UNEP), developing countries were offered to have other matrices analyzed for perfluoroalkane substances (PFAS) in one experienced laboratory. A total of 266 samples from 26 countries located in Africa, Asia, and Latin America were collected during 2018-2019 and analyzed for 15 PFAS. The limits of quantification were 6.2 pg/g fresh weight for most PFAS. The statistical assessment of 262 samples, four were excluded due to extreme values, showed that across abiotic and biota samples, perfluorooctane sulfonic acid (PFOS) had the highest detection frequency (80%) and the highest median value (19.2 pg/g), followed by perfluorooctanoic acid (PFOA) with 73% and a median value of 7.67 pg/g. Among the matrices, water (55%), sediment (49%) and fish (44%) had the most complex pattern, i.e., number of PFAS quantified. Dairy products and chicken meat had less PFAS present. From the 137 foods, fish, meat, eggs, analyzed in this study, only two fish samples would exceed one of the limit values proposed by the European Commission. To assess human exposure, we suggest including dairy products and drinking water since selective and sensitive methods would allow quantification of the four proposed PFAS.

Air monitoring with passive samplers for perfluoroalkane substances in developing countries (2017-2019).

To support the global monitoring plan on persistent organic pollutants (GMP) under the Stockholm Convention, 41 countries have exposed passive air samplers equipped with polyurethane foam disks (PUFs) to monitor perfluoroalkane substances (PFAS). The recommended procedure had PUFs exposed for three months during two years; analysis was performed in one expert laboratory to generate harmonized results. Chemical analysis using Soxhlet extraction with methanol or a 60:40 MTBE:methanol mixture, solid-phase extraction and UPLC/MSMS detection posed challenges as to matrix interference and sensitivity. Single PUFs as quarterly samples and combination of up to 4 PUFs as annual samples were analyzed; all concentrations were normalized to one PUF and 3 months exposure. Exceptionally high values (up to 36 000 pg SPFOS/PUF) were observed in nine samples from Zambia where a local source was identified. For the remaining 308 PUFs, PFOA had a higher median value (188 pg/PUF) than SPFOS (125 pg/PUF) whereas the mean value of SPFOS (254 pg/PUF) was slightly higher than the mean value for PFOA (230 pg/PUF). PFHxS concentrations were much lower in concentration and detection frequency and FOSA, as the only relevant PFOS precursor compound, showed a median value of zero. No clear regional differentiation was found.

Regional occurrence of perfluoroalkane substances in human milk for the global monitoring plan under the Stockholm Convention on Persistent Organic Pollutants during 2016-2019.

The human milk monitoring component of the global monitoring plan under the Stockholm Convention on Persistent Organic Pollutants was implemented from 2016 to 2019 and had 44 human milk samples from primiparae collected in 42 countries and analyzed for perfluoroalkane substances (PFAS), at the MTM Research Centre at Örebro University, Sweden. The targeted compounds were restricted to the two listed PFAS (perfluorooctane sulfonic acid, PFOS and perfluorooctane carboxylic acid, PFOA) and the one recommended PFAS (perfluorohexane sulfonic acid, PFHxS). PFOA was quantified in all 44 samples in a quite narrow range (6.20 pg/g-37.4 pg/g); PFHxS was quantifiable in only four samples (max. 111 pg/g), and PFOS in 36 samples across a wide range (<6.2 pg/g-212 pg/g). Branched PFOS isomers on average had a share of 16% of the total PFOS with a maximum of 33%. PFOS was highly correlated with PFHxS (Pearson correlation coefficient R = 0.95) and weakly but still positively with PFOA (R = 0.44). Statistical analysis (all on p<0.05) showed that PFOS and PFOA in European countries were significantly different from those in African and Latin American countries and between high-income and low-income countries. PFOA tends to have higher concentrations in wealthier countries. No correlation was found for population density.

Trace level analyses of selected perfluoroalkyl acids in food: Method development and data generation.

To comprise the future requirements to detect low levels of perfluoroalkyl acids (PFAAs) including branched and linear perfluorooctane sulfonic acid (PFOS), perfluorooctanoic acid (PFOA), and perfluorohexane sulfonic acid (PFHxS) in food items, analytical methods for their determination in five different food matrices (cow milk, butter, chicken meat, beef, and fish) were developed and validated. Analytical method for eggs was only validated for PFOS and PFOA because of interfering substance appeared in some samples. The method applied on foods of animal origin includes alkaline digestion, extraction, and clean-up with solid phase extraction and adsorption on granular carbon where necessary. The method was shown effective to eliminate taurodeoxycholic acid (TDC), a bile acid that is an endogenous interference compound in egg samples causing ionization suppression and false positive result for PFOS when 499 > 80 transition was used for quantification. The validation was performed and resulted in recoveries >70% for all three PFAAs, the limits of quantification (LOQs) in all matrices were 3.1 pg g-1, 3.4 pg g-1, and 4.9 pg g-1 for PFHxS, PFOA, and L-PFOS, respectively. The optimized method was successfully applied to 53 food samples from the Swedish market and from developing countries. PFOS and PFOA were detected in all samples. PFHxS was detected in 76% of the samples. Further method development on separating interfering substance from PFHxS in egg is warranted due to relatively high detection of this compound in other food items. With this method, concentrations in the low pg g-1 range in food samples of animal origin were quantified including the branched PFOS isomers. This method can be applied to enforce potential future limit values for PFOS and PFOA as a consequence of the recent European Food Safety Authority (EFSA) recommendation where the tolerable intakes have been drastically lowered. Further method development is needed for foods of plant origin such as vegetables, flour, nuts, or bread.