Dispersive Solid-Phase Extraction and Solid-Phase Extraction for ppt-level PFAS Analysis in Apples: A Comparison

Per- and polyfluoralkyl substances (PFAS) are found in our food. Sensitive, precise, and accurate analytical methods are needed to estimate human exposure to these chemicals. A comparative study was performed between two extraction and cleanup methods (solid-phase extraction [SPE] and dispersive SPE) for the analysis of PFAS in apples. Both methods showed excellent sensitivity, precision, and accuracy. dSPE has some benefits over conventional SPE, and vice versa. The advantages and disadvantages of both methods are discussed.

Per- and polyfluoralkyl substances (PFAS) are a very large family of man‑made chemicals that can be found in food and drinking water (1). PFAS have received considerable attention over recent years (2,3). The European Food Safety Authority (EFSA) has published a risk-assessment (4), setting very low safety levels (tolerable weekly intake) for the sum of four of these chemicals: PFHxS, PFOS, PFOA, and PFNA. To shed a light on the actual human exposure through food consumption, and to anticipate possible future legislative limits, sensitive and easy‑to-operate analytical approaches are needed for the detection of PFAS in foods at low ppt levels. Fruits were chosen in this study because limited data are available on fruits and vegetables and because they are a significant part of the human diet. As such, data are urgently needed.

In this work, we compared dispersive solid-phase extraction (dSPE) with an approach using SPE columns based on a weak anion exchange (WAX). The latter method was already established in-house for milk, eggs (5), and fishery products (6), but not yet tested on apples. In the current study, we wanted to compare the analysis of apples with both methods.

With the continuous emergence of new PFAS, the need for new cleanup and analysis methods also increases. Conventional SPE and dSPE can both serve this need, but both have their advantages and disadvantages. SPE-WAX cannot capture all relevant PFASs in one processing and/or measurement method; neutral and cationic PFASs are lost during the SPE cleanup, and due to the high range of polarity, multiple separation methods are required. As an alternative to conventional SPE, dSPE ([QuEChERS] quick, easy, cheap, effective, rugged, and safe) is used in pesticide analysis to remove interferants such as fatty acids, sugars, and pigments (7). It has also shown to be very effective as a cleanup for PFAS analysis (8). Several benefits of dSPE over SPE apply: (i) Due to a larger surface area, recoveries of dSPE are, in theory, higher compared with conventional SPE. (ii) The adsorption material is directly added to the extract/matrix so that no compounds are lost due to any washing steps. As the SPE material is added to adsorb interferences rather than the PFAS, neutral and cationic compounds can also be analyzed in a single cleanup. In addition, clogging of SPE columns is omitted. (iii) The cleanup can be much faster and cheaper when pre-mixed dSPE salts are used. The challenge with dSPE, however, lies in achieving similar detection limits as a conventional SPE method. As the cleanup using dSPE is in general less effective, the concentration of the raw extract to a small end volume is also restricted as a result of ionization effects. Moreover, residual salts in the extracts may desolvate in the MS source (and LC column) leading to salt deposits and increased dead volume and peak tailing. Furthermore, due to the strong adsorptive properties of the dSPE-salts, it is a challenge to obtain 100% PFAS-free solid phases. In conventional SPE procedures, an initial washing step is included, which removes most PFAS-residuals present in the SPE-column. This preconditioning step is not possible for dSPE, potentially resulting in higher background concentrations. Another challenging aspect of dSPE is that the required amount of added salts is largely dependent on the water content in the matrix. This makes the method less robust for varying matrices.

Methodology

The PFAS compounds in the study are mentioned in Table 1 and included perfluoro carboxylic acids from C5 (PFPeA) to C14 (PFTeDA) and several sulfonates (C4, C6, C7, C8, and C10). Moreover, a non-ionic precursor (PFOSA) and GenX (successor of PFOA as a fluoropolymer polymerization aid) were added as target analytes to this study. Two sample cleanup procedures were tested: SPE-WAX and dSPE. The SPE-WAX method was based on an existing protocol (5), with adjusted sample intake (10 g) and final extract volume to increase the sample enrichment factor. The dSPE method was optimized as regards sample intake (2 and 5 g) and the composition of the adsorbents and salt. The final composition was magnesium sulfate, sodium chloride, and the sorbents C18, primary secondary amine (PSA), and granulated active carbon.

The instrumental analysis was achieved using ultrahigh-performance liquid chromatography (UHPLC) coupled to a Sciex 7500 tandem mass spectrometer, operated in multiple reaction monitoring (MRM) mode for (tandem mass spectrometry [MS/MS]). Separation was performed on a 100 × 2.1 mm, 1.6-µm Luna Omega Polar C18 100Å column (Phenomenex) with methanol and an ammonium acetate buffer (20 mM) as the mobile phase. The MS/MS conditions were optimized for maximum selectivity and sensitivity and an instrumental limit of detection (iLOD) of 10 femtograms or less on-column for most analytes was achieved, based on a 20-μL injection. All experiments for comparing both methods were performed in triplicate.

Results and Discussion

The SPE (WAX) method showed a low limit of quantification (LOQ) at the pg/g level for most components, as well as a low limit of confirmation (LOC—the level where both diagnostic ions met the ion-ratio criteria).
The trueness, determined at the 50 pg/g spike level, was between 98–114% (except for GenX), with a relative standard deviation (RSD) lower than 10%, which is excellent. However, PFTrDA and PFTeDA did not meet the criteria, which was caused by the lack of an isotopically-labelled internal standard (PFTrDA) and low recoveries due to strong adsorption to surfaces. Moreover, the non-ionic PFOSA could not be retained by the SPE column.

dSPE, on the other hand, showed LOQs of 5 and 2.5 pg/g for about half the compounds. For the other half, high background signals were observed and yielded much higher LOQs. Further efforts are required to resolve the background signals for apples, and, if successful, lower LOQs can likely be achieved for the dSPE method as well. Moreover, excellent trueness values were obtained, as well as good RSDr values for most PFAS. Another advantage of the dSPE method over SPE-WAX was that the non-ionic PFOSA could be measured, as well as PFTrDA and PFTeDA.

Conclusion

Both methods showed advantages and disadvantages. dSPE was easy to operate and allowed the concurrent analysis of both ionic and nonionic PFAS (PFOSA). The LOCs were generally higher due to (i) a lower matrix-to-column ratio and (ii) background signals (by the presence of the actual PFAS in sorbents or unidentified interferences) on some of the PFAS. SPE based on WAX did not allow the analysis of PFOSA, PFTrDA, and PFTeDA at relevant levels, but a more powerful cleanup was achieved, allowing lower LOCs in most cases. Accuracies and RSDr values were excellent. Further optimization of the dSPE method will likely resolve the relatively high LOQs for some of the short-chain carboxylic acids. Both methods can be used for monitoring PFAS in fruits in the future, and, depending on the specific interests, one can choose either SPE or dSPE. These methods will contribute to a better understanding of human exposure to PFAS.

References

1. R.C. Buck, J. Franklin, U. Berger, J.M. Conder, I.T. Cousins, P. de Voogt, A.A. Jensen, K. Kannan, S.A. Mabury, and S.P.J. van Leeuwen, Integrated Environmental Assessment and Management 7(4), 513–541 (2011).
2. E. Thépaut, H.A.A.M. Dirven, L.S. Haug, B. Lindeman, S. Poothong, M. Andreassen, H. Hjertholm, and T. Husøy, Environmental Research 195, 110795 (2021).
3. A. Ramírez Carnero, A. Lestido-Cardama, P. Vazquez Loureiro, L. Barbosa-Pereira, A. Rodríguez Bernaldo de Quirós, and R. Sendón, Foods 10, 1443 (2021).
4. EFSA, Scientific Opinion on the risk to humanhealth related to the presence of perfluoroalkyl substances in food, EFSA Journal 18(9), 6223 (2020).
5. B.J.A. Berendsen, F. Lakraoui, L. Leenders, and S.P.J. Van Leeuwen, Food Additives & Contaminants Pt. A 37(10), 1707–1718 (2020).
6. E. Zafeiraki, W.A. Gebbink, R.L.A.P. Hoogenboom, M. Kotterman, C. Kwadijk, E. Dassenakis, and S.J.P. van Leeuwen, Chemosphere 232, 415-423 (2019).
7. M. Anastassiades, S.J. Lehotay, D. Stajnbaher, and F.J. Schenck, Journal of Aoac International 86(2), 412–431 (2003).
8. M. Al Amin, Z. Sobhani, Y.J. Liu, R. Dharmaraja, S. Chadalavada, R. Naidu, J.M. Chalker, and C. Fang, Environmental Technology & Innovation 19, 100879 (2020).

Marise van der Vegt has obtained a B.Sc. in analytical chemistry. At Wageningen Food Safety Research, she investigated dSPE and SPE extraction and cleanup approaches for PFAS analysis in fruits, and mutually compared them. She has recently accepted a position as Junior Researcher Analytical Chemistry at Enza Zaden, Enkhuizen, The Netherlands.

Ruben Kause is a scientist at Wageningen Food Safety Research (WFSR, Wageningen, The Netherlands). Ruben has a background in analytical chemistry and environmental sciences. His expertise is mainly in emission modelling and analysis of perfluorinated alkyl substances in food and environmental matrices.

Bjorn Berendsen studied analytical chemistry at Amsterdam University and did his Ph.D. at Wageningen University whilst being affiliated to Wageningen Food Safety Research (WFSR). He is an expert in residue analysis of mainly veterinary drugs and perfluoro alkyl substances in food and environmental matrices. He has been working at WFSR for 23 years in several functions and is currently, besides researcher and project manager, the programme manager of the statutory tasks policy research programme and as such part of the institute’s management team.

Stefan van Leeuwen is a senior scientist at Wageningen Food Safety Research (WFSR, Wageningen, The Netherlands). His current focus is on investigating the impact of environmental pollutants like PFAS, on food safety and our dietary exposure. He employs a range of complementary analytical techniques to unravel the complexity of the PFAS issue.