Environmental Fate of Cl-PFPECAs: Accumulation of Novel and Legacy Perfluoroalkyl Compounds in Real-World Vegetation and Subsoils

Abstract

Per- and polyfluoroalkyl substances (PFAS) are globally distributed and potentially toxic compounds. We report accumulation of chloroperfluoropolyethercarboxylates (Cl-PFPECAs) and perfluorocarboxylates (PFCAs) in vegetation and subsoils in New Jersey. Lower molecular weight Cl-PFPECAs, containing 7-10 fluorinated carbons, and PFCAs containing 3-6 fluorinated carbons were enriched in vegetation relative to surface soils. Subsoils were dominated by lower molecular weight Cl-PFPECAs, a divergence from surface soils. Contrastingly, PFCA homologue profiles in subsoils were similar to surface soils, likely reflecting temporal-use patterns. Accumulation factors (AFs) for vegetation and subsoils decreased with increasing CF₂, 6-13 for vegetation and 8-13 in subsoils. In vegetation, for PFCAs having CF₂ = 3-6, AFs diminished with increasing CF₂ as a more sensitive function than for longer chains. Considering that PFAS manufacturing has transitioned from long-chain chemistry to short-chain, this elevated vegetative accumulation of short-chain PFAS suggests the potential for unanticipated PFAS exposure levels globally in human and/or wildlife populations. This inverse relationship between AFs and CF₂-count in terrestrial vegetation is opposite the positive relationship reported in aquatic vegetation suggesting aquatic food webs may be preferentially enriched in long-chain PFAS. AFs normalized to soil-water concentrations increased with chain length for CF₂ = 6-13 in vegetation but remained inversely related to chain length for CF₂ = 3-6, reflecting a fundamental change in vegetation affinity for short chains compared to long.

Keywords: PFAS alternatives; accumulation factors; next-generation chemicals; polyfluorinated alkyl substances.

Figure 1.

A. Structure of perfluorooctanoate (PFOA), one of the legacy PFCAs, for which fluorocarbons commonly vary in number from 3 to 11. B. Structure of the (e,p) = (1,1) congener of the novel Cl-PFPECAs, for which perfluoroethyl (e) and perfluoropropyl (p) groups each can vary from 0 to 4.

Figure 2.

Congener distribution (A–C) and concentration as C9 (D–F) of Cl-PFPECA congeners in surface soils (A and D), vegetation (B and E), and subsoils (C and F) normalized by dry-sample mass. Samples ordered by distance from Solvay. Relative signal in vegetation and surface soils for all 24 samples is 86% vegetation and 14% surface soils. Relative signal strength in vegetation, surface soils, and deep soils for the four cored samples is 84% vegetation, 14% surface soil, and 2% deep soils. Note the greater fraction of the 0,1 congener in vegetation and deep soils (relative to that in surface soils), the increasing fraction of the lower molecular weight congeners with increasing distance from the facility, and the general decreasing trend in total Cl-PFPECAs with distance from the facility. See the text for discussion.

Figure 3.

Homologue distribution (A–C) and concentration (D–F) of PFCAs in surface soils (A and D), vegetation (B and E), and subsoils (C and F) normalized by dry-sample mass. Samples ordered by distance from Chemours. The sample labeled Solvay is Sample 8, the closest sample to the Solvay facility. Relative concentrations in vegetation and surface soils for all 24 samples are 63% vegetation and 37% surface soils. Relative concentration strengths in vegetation, surface soils, and deep soils for the four cored samples is 27% vegetation, 44% surface soil, and 29% deep soils. Note the relatively large contribution of long even-numbered PFCAs (C10 + C12 + C14 + C16) in surface soil closest to Chemours, the relatively large contribution of long odd-numbered PFCAs (C9 + C11 + C13) in surface soils and vegetation near Solvay, and the contrast in deep soils of intermediately sized PFCAs with predominantly lower molecular weight congeners for Cl-PFPECAs. See the text for discussion.

Figure 4:

Vegetative accumulation factor (VAF) decreases with fluorinated carbon chain length for both Cl-PFPECAs (yellow) and legacy PFCAs (blue). Fitting PFCA and Cl-PFPECA data for chain length ≥6 yields trend lines having statistically similar slope (A). When the VAFs of all PFCA chain-lengths are regressed together, the result is nearly identical to the literature-review regression of Lesmeister et al. (28) (Figure S1). Subsoil accumulation factor (SAF) vs chain length for Cl-PFPECAs and PFCAs (B). In these SAFs, an inflection is evident at chain-length 8, perhaps reflecting a complex balance of sufficient relative mobility from the surface to the subsurface sample depth to deliver long chains vs still higher mobility for short chains fostering continued percolation to greater depths. Vegetative accumulation factor (VAFW) normalized to estimated soil-water concentration (using eq 2) vs chain length (C). The inflection at chain length 6 is even more pronounced when normalized to soil water, possibly reflecting a change in relative affinity of vegetation for the short chains. See text for details. Subsoil accumulation factor (SAFW) normalized to estimated soil–water concentration (using eq 2) vs chain length for Cl-PFPECAs and PFCAs (D). When normalized to estimated soil water the inflection observed in SAF at chain length 8 disappears, yielding a continuous linear function from C3 to C12, within the resolution of the data set. See the text for details.

Figure 5:

Log solubility (red; primary y axis) and log relative vegetative affinity (blue; secondary y axis) vs number of fluorinated carbons (CF₂) for PFCAs. Solubility (red) was modeled by Wang et al. (55) Relative vegetative affinity for homologues is estimated by the regressions depicted in Figure 4C as explained in the text. According to this approach, the incremental increase in affinity of vegetation for PFCAs is greater for short chains (SVeg; ${\mathrm{C}\mathrm{F}}_2<6$) than long chains (LVeg; ${\mathrm{C}\mathrm{F}}_2>6$). For example, the relative affinity of perfluorobutane carboxylate (${\mathrm{C}\mathrm{F}}_2=3$) is >700-fold higher than suggested by the long-chain (LVeg) trend. Also noteworthy, the relative vegetative affinity slope is less than that of bulk-water solubility for long-chain PFCAs but greater than the solubility slope for short-chain PFCAs.