Detecting PFAS Contamination in Agricultural Soils: Non-targeted Identification of Precursors, Fluorine Mass Balance, and Microcosm Studies

You recently published a manuscript analyzing per- and polyfluoroalkyl substances (PFAS) contamination in agricultural soils (1). This involved exploring the non-targeted identification of precursors, fluorine mass balance, and microcosm studies. What was the rationale behind this research, and can you clarify the three main components of this study: identification of precursors, fluorine mass balance, and microcosm studies?
The goal of this study (1) was to address the environmental impact arising from elevated PFAS levels in soils of Rastatt-Baden/Baden and Mannheim, southwest Germany. This contamination originated from the historical application of PFAS-contaminated paper fibers as an agricultural soil amendment. This led to the input of so-called PFAS precursors into the topsoil, with their slow but ongoing biotransformation producing mobile perfluoroalkyl carboxylic acids (PFCAs) that contaminate groundwater.

To investigate this issue, we conducted a comprehensive characterization of PFAS in topsoil using both targeted and non-targeted analytical methods. To enhance the interpretation of our findings, we employed complementary techniques such as the total oxidizable precursor (TOP) assay and extractable organic fluorine (EOF) analysis. These approaches allowed us to establish a fluorine mass balance, a critical aspect of PFAS analysis. Moreover, we performed microcosm experiments to determine rate constants for the formation of PFCAs, and provided insights into the soil properties influencing these. With this information, we estimate the timescales of contamination, that is, the duration required for the biotic transformation of PFAS precursors in the topsoil, information that is essential for improving groundwater contamination assessments.

Joel Fabregat-Palau holds a PhD in analytical and environmental chemistry, and is currently a postdoctoral researcher at the faculty of geosciences at the University of Tübingen (Germany

What analytical techniques were used in this study to identify and quantify PFAS in agricultural soils, and how do they compare in terms of sensitivity and accuracy?
We monitored specific target PFAS using the EPA 1633 method (2), which uses isotopic dilution mass spectrometry (IDMS) for quantification. This approach is typically performed with triple quadrupole mass spectrometers. Triple quadrupole mass spectrometry (TQMS) is currently the most accurate and sensitive for analyzing organic pollutants in environmental matrices. This technique offers high precision and accuracy while correcting for matrix effects and potential losses during sample preparation. The method relies on a calibration curve prepared with commercially available native and mass-labeled PFAS standards.

However, for many PFAS present in the environment no commercial standards are available, making their identification and quantification challenging. This is the domain of high-resolution mass spectrometry (HRMS), such as time-of-flight (TOF) instruments, combined with non-target screening workflows. This approach allows for non-target screening of unknown PFAS and potential transformation products. In our study, liquid chromatography (LC)–HRMS data were processed using PFΔScreen, an open-source tool previously implemented for PFAS identification and structural elucidation in environmental samples (3). Identification is typically tentative and obtained on different confidence levels dependent on the information available in databases and mass spectral libraries; for a final and unequivocal validation of the identification, an analytical standard is required. Additionally, we (semi)quantified PFAS precursors using a response factor approach, which involved a one-point calibration based on available PFAS standards selected by similar structures and functional groups. Semi-quantification does not fully account for response differences because of different ionization and gas phase transfer efficiencies, or signal suppression or enhancement effects of the sample matrix. Therefore, this method provides only a rough estimate of PFAS concentrations.

How do non-targeted analysis (NTA) using high performance liquid chromatography (HPLC)-QTOF-MS and target analysis using HPLC–MS/MS analyses complement each other to identify PFAS precursors in the study?
The combination of target and non-target analyses provided a more comprehensive assessment of PFAS contamination in soils. In our study, we monitored a total of 40 target PFAS, including perfluorocarboxylic acids (PFCAs) and perfluoroalkyl sulfonic acids (PFSAs) with fluorinated carbon chains ranging from 3 to 13. Additionally, we analyzed fluorotelomer sulfonic acids (FTSs) and fluorotelomer carboxylic acids (FTCAs) with fluorinated carbon chains ranging from 4 to 8, perfluorooctane sulfonamides (FOSAs), perfluorooctane sulfonamidoacetic acid (FOSAAs), perfluorooctane sulfonamide ethanols (FOSEs), and selected per- and polyfluoroether carboxylic acids and ether sulfonic acids. Many of these PFAS are precursors to PFCAs or PFSAs.

Through non-target screening, we identified and (semi)quantified an additional 23 PFAS, including several homologs of fluorotelomer alkyl phosphate di-esters (diPAPs) and fluorotelomer mercapto alkyl phosphate esters (FTMAPs), both considered PFCAs precursors. Many of these PFAS lack commercial standards, leading to their exclusion from target analyses despite comprising over 80% of the PFAS burden in the studied soils. This highlights the need for non-target screening methods to achieve a comprehensive characterization of PFAS contamination in environmental samples.

Can you explain how the fluorine mass balance was achieved during the TOP assay and why it is considered a critical quality assurance procedure?
To ensure confidence in the quantified PFAS levels, we conducted a TOP assay. This oxidation method, first introduced for aqueous samples by Houtz and Sedlak (2012) (4) and later adapted for soils by Göckener et al. (2020) (5), converts certain overlooked PFAS precursors into measurable PFCAs and, to a lesser extent, PFSAs. The process involves oxidation using hydroxyl radicals under basic conditions. In our study, fluorine mass balances were assessed by converting PFAS concentrations in both oxidized and non-oxidized solid samples to µg F/kg. This allowed direct comparison with EOF, which quantifies fluorine associated with organic compounds and, in some cases, serves as a proxy for total PFAS levels in soils. Other fluorine mass balance approaches against absorbable (AOF) or total (TOF) organic fluorine may also be used. EOF determination involves soil extraction with an organic solvent, followed by fluorine detection via high-resolution continuum source graphite furnace molecular absorption spectrometry (HR-CS-GFMAS). Previous research showed a positive correlation between EOF and the number of PFAS features prioritized by non-target screening (6).

Closing the fluorine mass balance is a critical quality assurance step in PFAS analysis. Since target methods only monitor a subset of PFAS, they may overlook others present at even higher concentrations. Assessing fluorine mass balances helps determine the proportion of PFAS captured by conventional target methods and should also be applied in NTA and oxidation assays to ensure that only a minor fraction of PFAS remains undetected.

What role did the chemical oxidation (TOP assay) play in confirming PFAS precursor concentrations, and how was repeat oxidation used to ensure accurate fluorine mass balance?
By assessing fluorine mass balances, we observed that in certain soils the fluorine levels derived from (semi)quantified PFAS were low. This indicated the presence of still overlooked PFAS precursors in some samples. However, after oxidation, fluorine mass balances showed successful fluorine recovery, suggesting that these previously undetected precursors were converted into PFCAs under the oxidative conditions of the TOP assay.

The assessment of fluorine mass balances during oxidation assays also led to a key finding in our study. Traditionally, a single oxidation step is applied to environmental samples, regardless of their characteristics. However, when we observed low fluorine mass balances, we hypothesized that organic matrix constituents might be consuming hydroxyl radicals, thereby reducing oxidation efficiency and potentially leaving a fraction of the PFAS burden undetected. Alternative hypotheses included the formation of ultra-short-chain PFCAs or other PFAS species, for example, diacid PFAS containing both carboxylate and sulfonate moieties, which would all remain undetected by target methods. By performing sequential oxidation steps, fluorine recovery improved, providing more reliable data for estimating PFCA generation rate constants and confirming the scavenging effect of matrix constituents on hydroxyl radicals in our soils. These data underscore the importance of assessing fluorine mass balances to ensure a more accurate evaluation of the total PFAS burden in environmental samples.

How do soil physicochemical properties, such as organic carbon content, influence PFAS generation and decay rates based on the study's findings?
Soil organic carbon is the most critical factor influencing PFAS sorption in soils, although other parameters, such as pH, ionic strength, and mineral content, also contribute (7). Sorption plays a key role in determining the distribution of PFAS between soil and aqueous phases and affects both PFAS precursors and their terminal biotransformation products. The sorption of PFAS precursors onto soil particles ultimately dictates their bioavailability to microorganisms. As a result of their typically strong sorption, especially in organic carbon-rich soils, their biotransformation tends to be slow. The fraction of PFAS precursors present in the aqueous phase undergoes biodegradation, forming more mobile PFCAs. However, these PFCAs also partition between soil and liquid phases, which ultimately dictates their transport to groundwater.

In our study, we monitored the generation of biodegradation end-products over time resulting from the decay of comingled PFAS precursors. By accounting for the solid-liquid distribution coefficient (K_d) of PFCAs, we were able to assess the total fraction generated, an aspect often overlooked but found to be an important fraction in some of our soils. The K_d values varied depending on the specific PFAS/soil pair and increased with both PFAS molecular weight and soil organic carbon content.

What were the specific microbial or enzymatic factors, for example, acid phosphomonoesterase or microbial biomass, that seemed to enhance PFAS generation rates in the microcosm experiments?
Many of the PFAS precursors identified in our soils, for example fluorotelomer alkyl phosphate diesters (diPAPs), fluorotelomer mercapto alkyl phosphate esters (FTMAPs) and N-ethyl perfluorooctane sulfonamide ethanol-based phosphate diester (diSAmPAP) contain a phosphate moiety, and previous studies have identified the cleavage of P–O bonds as the first biodegradation step. While certain enzymes have been proposed to play a role in PFAS biodegradation, the involvement of phosphatases (enzymes capable of hydrolyzing P–O bonds) has largely been overlooked. We hypothesized that if biodegradation is limited by the release of PFAS precursors from soil particles, a kinetic bottleneck in the cleavage of the P–O bond may occur. Consequently, we explored potential links between PFCA generation rates and phosphatase enzymatic activities or other soil biological properties, such as microbial biomass.

Many of the PFAS precursors identified in our soils, for example fluorotelomer alkyl phosphate diesters (diPAPs), fluorotelomer mercapto alkyl phosphate esters (FTMAPs) and N-ethyl perfluorooctane sulfonamide ethanol-based phosphate diester (diSAmPAP) contain a phosphate moiety, and previous studies have identified the cleavage of P–O bonds as the first biodegradation step. While certain enzymes have been proposed to play a role in PFAS biodegradation, the involvement of phosphatases (enzymes capable of hydrolyzing P–O bonds) has largely been overlooked. We hypothesized that if biodegradation is limited by the release of PFAS precursors from soil particles, a kinetic bottleneck in the cleavage of the P–O bond may occur. Consequently, we explored potential links between PFCA generation rates and phosphatase enzymatic activities or other soil biological properties, such as microbial biomass.

To evaluate this, we conducted a principal component analysis (PCA) and found that both acid phosphomonoesterase activities and, to a lesser extent, soil biomass carbon affected PFCA generation rates. In contrast, phosphodiesterase and basic phosphomonoesterase activities showed no significant effect. These results suggest that PFAS precursors are first released from soil particles and then undergo biotic cleavage of their P–O bonds, ultimately affecting the overall biotransformation kinetics and therefore PFCA generation rates. Supporting this finding, inorganic phosphorus sources appeared to be negatively correlated with PFCA generation rates, possibly because of bacterial uptake of alternative phosphorus sources. However, it is important to note that these findings apply specifically to certain P-containing precursors and may not be applicable to others. Also, further research in well-controlled systems or with well-characterized bacterial strains is needed to confirm these preliminary observations.

How did the study use microcosm incubations to determine the production rate constants of short-chain PFCA, and what does this imply for PFAS behavior in real-world environments?
In our study, we used the methods presented by Röhler et al. (2023) (8) to determine PFCA generation rates. The first step involves soil washing according to DIN 19528 to remove already present mobile PFCA (especially those short-chained) while keeping immobile PFAS precursors in the soil. After, the soil is mixed with water to form a suspension, which is then incubated at 20 °C under oxic conditions. This first step to reduce the levels of mobile PFCA is necessary to better assess their increase due to precursor transformation over time in the suspension.

The generation of PFCA end-products from PFAS precursor biodegradation follows an inverse relationship with a first-order decay model. During the short incubation period (50 days in our case), this translates into a linear increase in PFCA concentrations over time. Consequently, PFCA generation rate constants (1/year)can be estimated by dividing the total generated PFCA concentrations over time (µg/kg/year) by the final yield of transformation products, that is, the total conversion of the PFAS precursor per mass of soil (µg/kg). These parameters were obtained from our microcosm and TOP assay data, respectively.

By combining the final yield of transformation products per mass of soil with the PFCA generation rate constants, we can estimate the time scale required to transform a specific fraction, for example, 90% of PFAS precursors. Our data suggest that under laboratory conditions, PFAS precursors would continue transforming into more mobile PFCA over a period of years to decades. However, the rate constants derived from this method are specific to laboratory conditions (20 °C, water-saturated system, oxygen supply) and are expected to be lower under field conditions, for example, dry soils, cold seasons, and anaerobic environments, which would further extend the contamination time scale. Additionally, these time scales do not apply to long-chained PFCA, whose generation rates may be slower due to the stronger sorption of their precursors to soil particles. Furthermore, long-chained PFCA themselves exhibit higher sorption affinity to topsoil, leading to their accumulation and less efficient transport to groundwater.

Given that short-chain PFAS may leach from soils for decades, how does the study suggest monitoring strategies for long-term PFAS contamination in agricultural soils and nearby water bodies?
While our study does not explicitly suggest specific long-term monitoring strategies, it provides an important indicator for policymakers to focus on long-term groundwater monitoring efforts. Additionally, it highlights the need to develop targeted topsoil management measures to mitigate the long-term contamination in the agricultural areas affected. Finally, it demonstrates the need to prevent such contaminations in the future.

What challenges were encountered in using HPLC–MS/MS and HPLC–QTOF-MS for detecting low-level PFAS precursors in complex soil matrices and how did you overcome them?

The soils we studied had a generally high PFAS load, but we encountered challenges in detecting PFAS precursors at trace concentrations. These challenges were primarily addressed by implementing quality assurance procedures, such as preparing instrumental blanks, soil blanks, and surrogate-spiked soils. However, most of the PFAS precursors in our study were assessed through non-target analysis, which generally is less sensitive compared to target methods and requires sufficient signal intensity to acquire meaningful MS/MS spectra. At trace levels, different ionization efficiencies of various PFAS can complicate their identification. To overcome this, we used a higher soil mass (five grams) and concentrated the extract to a volume of one milliliter. This approach provided adequate concentrations, ensuring a significant signal for PFAS precursors that might otherwise have low ionization efficiency.

What is novel about this research?
The novelty of our study relies on the exhaustive PFAS characterization in agricultural soils by extending conventional PFAS target analysis with non-target screening, as well as the assessment of fluorine mass balances after the oxidation of PFAS precursors, identifying the need to perform sequential oxidations in some matrix-rich soils. Additionally, we considered the partition of PFCA end-products in the solid phase in the generation rates resulting from PFAS precursors decay, and we suggest the role of certain phosphatases in the decay of P-containing PFAS precursors, an enzyme type that has been overlooked in PFAS biotransformation. We also estimate the time scales that PFAS precursors will biotransform into more mobile PFCA, compounds that already pose an environmental risk because of their elevated levels in groundwater.

Are you planning to continue with research in PFAS?
Yes, in our laboratories under the supervision of both Peter Grathwohl and Christian Zwiener we continue to address environmental challenges related to the analysis, sorption, biotransformation, and remediation of PFAS. Future work in this area would include the upgrade of non-target screening methods and extending their applicability, in addition to elucidating the relationship between biotransformation rates, sorption, and soil properties.

References

1. Fabregat-Palau, J.; Zweigle, J.; Renner, D.; Zwiener, C.; Grathwohl, P. Assessment of PFAS Contamination in Agricultural Soils: Non-target Identification of Precursors, Fluorine Mass Balance and Microcosm Studies. J. Hazard. Mat.2025, 490, 137798. DOI: 10.1021/acs.estlett.4c00442

2. U.S. Environmental Protection Agency. (2024). Method 1633A: Analysis of per- and polyfluoroalkyl substances (PFAS) in aqueous, solid, biosolids, and tissue samples by LC-MS/MS (EPA 821-D-21-001)

3. Zweigle, J.; Bugsel, B.; Fabregat-Palau, J.; Zwiener, C., PFΔScreen — An Open-source Tool for Automated PFAS Feature Prioritization in Non-target HRMS Data. Anal. Bioanal. Chem. 2024, 416 (2), 349–362. DOI: 10.1007/s00216-023-05070-2

4. Houtz, E. F.; Sedlak, D. L. Oxidative Conversion as a Means of Detecting Precursors to Perfluoroalkyl Acids in Urban Runoff. Environ. Sci. Technol.2012, 46 (17), 9342–9349. DOI: 10.1021/es302274g

5. Göckener, B.; Eichhorn, M.; Lämmer, R.; et al. Transfer of Per- and Polyfluoroalkyl Substances (PFAS) From Feed Into the Eggs of Laying Hens. Part 1: Analytical Results Including a Modified Total Oxidizable Precursor Assay. J. Agric. Food. Chem.2020, 68 (45), 12527–12538. DOI: 10.1021/acs.jafc.0c04456

6. Zweigle, J.; Simon, F.; Meermann, B.; Zwiener, C. Can Qualitative Nontarget Data Be Indicative of PFAS Contamination? First Evidence by Correlation with EOF in Environmental Samples. Environ. Sci. Tech. Let. 2024, 11 (9), 996–1001. DOI: 10.1021/acs.estlett.4c00442

7. Fabregat-Palau, J., Vidal, M., Rigol, A. Modelling the sorption behaviour of perfluoroalkyl carboxylates and perfluoroalkane sulfonates in soils. Sci. Total Environ. 2021, 801, 149343. DOI: 10.1016/j.scitotenv.2021.149343

8. Röhler, K.; Susset, B.; Grathwohl, P. Production of Perfluoroalkyl Acids (PFAAs) From Precursors in Contaminated Agricultural Soils: Batch and Leaching Experiments. Sci. Total Environ. 2023, 902, 166555. DOI: 10.1016/j.scitotenv.2023.166555

Biography

Joel Fabregat-Palau holds a PhD in analytical and environmental chemistry, and is currently a postdoctoral researcher at the faculty of geosciences at the University of Tübingen (Germany). His research covers the analysis, sorption, biotransformation and remediation of organic pollutants, with a main focus on per- and polyfluoroalkyl substances (PFAS).