Assessment of Mobilization Potential of Per- and Polyfluoroalkyl Substances for Soil Remediation

Abstract

This study investigated the mobilization of a wide range of per- and polyfluoroalkyl substances (PFASs) present in aqueous film-forming foams (AFFFs) in water-saturated soils through one-dimensional (1-D) column experiments with a view to assessing the feasibility of their remediation by soil desorption and washing. Results indicated that sorption/desorption of most of the shorter-carbon-chain PFASs (C ≤ 6) in soil reached greater than 99% rapidlyafter approximately two pore volumes (PVs) and were well predicted by an equilibrium transport model, indicating that they will be readily removed by soil washing technologies. In contrast, the equilibrium model failed to predict the mobilization of longer-chain PFASs (C ≥ 7), indicating the presence of nonequilibrium sorption/desorption (confirmed by a flow interruption experiment). The actual time taken to attain 99% sorption/desorption was up to 5 times longer than predicted by the equilibrium model (e.g., ∼62 PVs versus ∼12 PVs predicted for perfluorooctane sulfonate (PFOS) in loamy sand). The increasing contribution of hydrophobic interactions over the electrostatic interactions is suggested as the main driving factor of the nonequilibrium processes. The inverse linear relationship (R 2 = 0.6, p < 0.0001) between the nonequilibrium mass transfer rate coefficient and the Freundlich sorption coefficient could potentially be a useful means for preliminary evaluation of potential nonequilibrium sorption/desorption of PFASs in soils.

Role of Mineral–Organic Interactions in PFAS Retention by AFFF-Impacted Soil

“…An equilibrium model was well suited to describe retention (or breakthrough) of short-chain PFAS (C ≤ 6) in other cases as well. 34,75 Example BTCs for three PFAS (Figure 2) are presented to illustrate instances where the equilibrium model was a good fit to all four BTCs (perfluorooctanoic acid; PFOA); a poor fit (perfluorooctane sulfonic acid; PFOS) or a partial fit to two of the four BTCs (N-dimethyl ammonio propyl perfluoropropane sulfonamide; AmPr-FPrSA). The BTCs for PFOA from all four soil columns were characterized by Pe numbers > 10, IPWR ranges between 70% and 130% (Table S14), and were symmetrical without any significant tailing during sorption or desorption phases (Figure 2).…”

“…41 Our modeling activity led to the finding that sorbent parameters Fe ox and %OC were statistically related to the 'number of nitrogens' parameter, meaning that they influence the retention of zwitterionic PFAS that possess both positive and negative charges, thus underlining the plausibility of a mechanistic contribution of these structural PFAS features to interactions with charged soil surfaces. 26,34,37 This is noteworthy as Al/Fe oxides contribute positive charge at typical soil pH ranges, while well-decomposed organic carbon carries net negative charge and can serve as a potent cation exchange medium. 84,85 Several considerations may help to constrain the underlying mechanisms of zwitterionic sorption.…”

“…To date, organic carbon (OC) − and the length of the fluorinated carbon chain (as a proxy for molecular hydrophobicity) are viewed as the primary features related to PFAS partitioning to soil, , especially for PFAS with more than six or seven fluorinated carbons. − However, recent analyses of published soil partitioning coefficients for anionic PFAS indicate that, in soils with less than 5% OC, the fluorinated tail–OC relationship is not a robust predictor of sorption, , with one study reporting little change in sorption with increasing OC content . Several investigators also noted that the quality of the OC as opposed to the quantity alone may be a better predictor of anionic PFAS sorption. ,− Other researchers, using column and batch-based experiments suggest that soil constituents and interaction mechanisms other than strict hydrophobic partitioning to OC were likely responsible for PFAS sorption. This is especially true for zwitterionic PFAS, where molecular hydrophobicity and OC are poor predictors of sorption and where nonlinear ,, or nonequilibrium sorption is often observed.…”

“…28 Other researchers, using column 34 and batch-based 35 experiments suggest that soil constituents and interaction mechanisms other than strict hydrophobic partitioning to OC were likely responsible for PFAS sorption. This is especially true for zwitterionic PFAS, where molecular hydrophobicity and OC are poor predictors of sorption 35 and where nonlinear 26,35,36 or nonequilibrium 34 sorption is often observed. This view casts doubt on the validity of the current understanding that OC and number of fluorinated carbons represent the dominant controls on the retention of both anionic and zwitterionic PFAS in soil.…”

Impacts of Environmental and Engineered Processes on the PFAS Fingerprint of Fluorotelomer-Based AFFF

“…While PCA using aqueous PFAS concentrations clearly linked some sites (e.g., Green Bay shoreline) with the fluorotelomer AFFF facility, the role of sediments in fingerprinting is less straightforward. Longer-chain PFAS are more sorptive to sediments, , potentially enabling the detection of long-chain compounds that may not be detectable in water samples. For example, we observed longer-chain compounds (i.e., C ≥ 12 PFCAs and PFDS) in sediment samples but not in water samples (Tables S11 and S14) and observed legacy 8:2 FtTAoS only in sediment (Figure ).…”