Determining air-water interfacial areas for the retention and transport of PFAS and other interfacially active solutes in unsaturated porous media

Environmental fate and transport of PFAS in wastewater treatment plant effluent discharged to rapid infiltration basins

Fate and transport of per- and polyfluoroalkyl substances (PFAS) in wastewater treatment plant (WWTP) effluent discharged to rapid infiltration basins (RIBs) is investigated using data from 26 WWTPs in Michigan, USA. PFAS were found to accumulate in groundwater downgradient from RIBs with median groundwater-effluent enrichment factors for ten commonly detected, terminal-form perfluoroalkyl acids (PFAAs) ranging from 1.3 to 5.2. Maximum contaminant levels for drinking water were exceeded in groundwater at all WWTPs with available PFAS data. Numerical models of unsaturated fluid flow and PFAS transport honoring RIB site properties, such as median vertical separation distance to the water table and a realistic range of area-normalized effluent fluxes, show long-chain PFAS undergo significant delays from air-water interface (AWI) adsorption, requiring up to 15 times longer to reach maximum mass flux to the saturated zone under low-flux conditions, where AWI area is 2.5 times greater. Short-chain PFAS commonly detected in effluent are only minimally affected by AWI adsorption and show little to no attenuation under high-flux conditions. The nonlinear inverse relationship between water content and AWI area highlights the important role of AWI adsorption in modulating unsaturated transport of long-chain PFAS to underlying groundwater due to the broad range of flux rates applied to RIB systems.

Non-Fickian transport processes accelerate the movement of PFOS in unsaturated media: An experimental and modelling study

The transport of per- and polyfluoroalkyl substances (PFASs) through unsaturated source-zone soils is a critical yet poorly understood aspect of their environmental behavior. To date, most experimental studies have only focused on the equilibrium or non-equilibrium partitioning of PFASs to the air-water interface, or solid-phase based equilibrium or non-equilibrium transport. Currently, there are discrepancies between air-water interfacial partitioning (Kia) results measured using a drainage-based column method (which supports a Langmuir isotherm) when compared to measurements from alternative experimental methods (which support a Freundlich isotherm). We hypothesize that this discrepancy is the result of non-Fickian transport conditions developing during column tests using the drainage method, which reduces the magnitude of the apparent Kia (Kia,app) when estimated using the retardation factor correlation from breakthrough curve experiments. To test the validity of this hypothesis, the drainage method was implemented using PFOS in a sand column and compared with prior data collected using a quasi-saturated column method. Results demonstrate that the apparent Kia was reduced by 3 to 123-fold, resulting in up to 123-fold faster breakthrough of PFOS than predicted with the assumption of equilibrium adsorption to the air-water interface. A novel mobile-immobile model (MIM) of PFAS fate and transport was developed, incorporating a term for anomalously adsorbed solute in the mobile zone to explain highly anomalous data. The modelling results using a modified HYDRUS-1D software show that anomalous air-water interfacial adsorption and/or flowpath channelization are plausible mechanisms for accelerated transport of PFOS and support the application of a Freundlich isotherm for PFOS. Overall, non-Fickian transport mechanisms demonstrate the potential to accelerate PFOS transport through the vadose zone by up to a factor of 123 under specific circumstances. This work demonstrates the assumption of equilibrium adsorption to air-water interfaces, even for homogeneous laboratory experiments, is not necessarily valid.

PFAS transport under lower water-saturation conditions characterized with instrumented-column systems

The transport of PFOS and PFOA in well-characterized sand was investigated for relatively low water saturations. An instrumented column was used for some experiments to provide real-time in-situ monitoring of water saturation and matric potential. The results showed that water saturations and matric potentials varied minimally during the experiments. Flow rates were monitored continuously and were essentially constant. These results demonstrate that surfactant-induced flow and other nonideal hydraulic processes did not materially impact PFAS transport for the experiment conditions. Air-water interfacial adsorption was demonstrated to provide the great majority of retention for PFOS and PFOA. Retention was significantly greater at the lower water saturations (0.35-0.45) compared to the higher saturations (∼0.66) for both PFAS, due to the larger extant air-water interfacial areas. Retardation factors were 5 and 3-times greater at the lower water saturations for PFOS and PFOA, respectively. Early breakthrough was observed for the PFAS but not for the non-reactive tracers at the lower water saturations, indicating the possibility that air-water interfacial adsorption was rate-limited to some degree. Independently determined retention parameters were used to predict retardation factors for PFOS and PFOA, which were similar to the measured values in all cases. The consistency between the predicted and measured values indicates that PFAS retention was accurately represented. In addition, air-water interfacial adsorption coefficients measured from the transport experiments were consistent with independently measured equilibrium-based values. Based on these results, it appears that the air-water interfacial adsorption processes mediating the magnitude of PFOS and PFOA retention under lower water-saturation conditions are consistent with those for higher water saturations. This provides some confidence that our understanding of PFAS retention obtained from work conducted at higher water saturations is applicable to lower water saturations.

Numerical modeling of PFAS movement through the vadose zone: Influence of plant water uptake and soil organic carbon distribution

In this study, we investigated the effects of soil organic carbon (SOC) distribution and water uptake by plant roots on PFAS movement in the vadose zone with a deep groundwater table under temperate, humid climate conditions. Two series of numerical simulations were performed with the HYDRUS computer code, representing the leaching of historical PFOS contamination and the infiltration of water contaminated with PFOA, respectively. We considered soil profiles with three distributions of SOC (no SOC, realistic SOC distribution decreasing with depth, and uniform SOC equal to the content measured in topsoil), three root distributions (bare soil, grassland, and forest), and three soil textures (sand, sandy loam, and loam). The SOC distribution had a profound impact on the velocity of PFOS movement. The apparent retardation factor for realistic SOC distribution was twice as large as for the scenario with no SOC and more than three times smaller than for the scenario with uniformly high SOC content. We also showed that the root distribution in soil profoundly impacts the simulations of PFAS migration through soil. Including the root zone significantly slows down the movement of PFAS, primarily due to increased evapotranspiration and reduced downward water flux. Another effect of water uptake by plant roots is an increase of PFAS concentrations in soil water (evapo-concentration). The evapo-concentration and the slowdown of PFAS movement due to root water uptake are more significant in fine-textured soils than in sand.

Vapor-phase transport of per and polyfluoroalkyl substances: Processes, modeling, and implications

An increasing number of studies have demonstrated the presence of per and polyfluoroalkyl substances (PFAS) in the vapor phase. It is therefore important to consider the potential for vapor-phase transport of PFAS in soil and the vadose zone and to investigate the processes impacting the retention and transport of volatile PFAS in soil. It is also critically important to evaluate existing models and develop new models as needed for their application to PFAS vapor-phase transport. The objectives of the present work were to provide an overview of vapor-phase transport processes and modeling, with a specific focus on their relevance for PFAS, and to discuss implications for mass discharge to groundwater, vapor intrusion, and soil vapor extraction. Decades of research have been devoted to the retention and transport of legacy volatile organic contaminants in the vadose zone. This work provides an abundant source of information concerning the many factors and processes of relevance, and insights into the development and application of mathematical modeling. However, given the unique properties of PFAS, there is a need to conduct research to investigate vapor-phase transport of PFAS and to develop PFAS-specific models. We highlight with illustrative examples that vapor-phase transport can be significantly more rapid than aqueous-phase advective transport, which can result in enhanced mass discharge to groundwater.

Air-water interfacial collapse and rate-limited solid desorption control Perfluoroalkyl acid leaching from the vadose zone

Some Per- and polyfluoroalkyl substances (PFAS) are strongly retained in the vadose zone due to their sorption to both soils and air-water interfaces. While significant research has been dedicated to understanding equilibrium behavior for these multi-phase retention processes, leaching and desorption from aqueous film-forming foam (AFFF) impacted soils under field relevant conditions can exhibit significant deviations from equilibrium. Herein, laboratory column studies using field collected AFFF-impacted soils were employed to examine the leaching of perfluoroalkyl acids (PFAAs) under simulated rainfall conditions. The HYDRUS 1-D model was calibrated to estimate the unsaturated hydraulic properties of the soil in a layered system using multiple boundary condtions. Forward simulations of equilibrium PFAS partitioning using the HYDRUS model and simplified mass balance calculations showed good agreement with the net PFAS mass flux out of the column. However, neither were able to predict the PFAS concentrations in the leached porewater. To better understand the mechanisms controlling the leaching behavior, the HYDRUS 1-D two-site leaching model incorporating solid phase rate limitation and equilibrium air-water interfacial partitioning was employed. Three variations of the novel model incorporating different forms of equilibrium air-water interfacial partitioning were considered using built-in numerical inversion. Results of numerical inversion show that a combination of air-water interfacial collapse and rate-limited desorption from soils can better predict the unique leaching behavior exhibited by PFAAs in AFFF-impacted soils. A sensitivity analysis of the initial conditions and rate-limited desorption terms was conducted to assess the agreement of the model with measured data. The models demonstrated herein show that, under some circumstances, laboratory equilibrium partitioning data can provide a reasonable estimation of total mass leaching, but fail to account for the significant rate-limited, non-Fickian transport which affect PFAA leaching to groundwater in unsaturated soils.

High-Resolution Depth-Discrete Analysis of PFAS Distribution and Leaching for a Vadose-Zone Source at an AFFF-Impacted Site

The long-term leaching of polyfluoroalkyl substances (PFAS) within the vadose zone of an AFFF application site for which the depth to groundwater is approximately 100 m was investigated by characterizing the vertical distribution of PFAS in a high spatial resolution. The great majority (99%) of PFAS mass resides in the upper 3 m of the vadose zone. The depths to which each PFAS migrated, quantified by moment analysis, is an inverse function of molar volume, demonstrating chromatographic separation. The PFAS were operationally categorized into three chain-length groups based on the three general patterns of retention observed. The longest-chain (>∼335 cm3/mol molar volume) PFAS remained within the uppermost section of the core, exhibiting minimal leaching. Conversely, the shortest-chain (<∼220 cm3/mol) PFAS accumulated at the bottom of the interval, which coincides with the onset of a calcic horizon. PFAS with intermediate-chain lengths were distributed along the length of the core, exhibiting differential magnitudes of leaching. The minimal or differential leaching observed for the longest- and intermediate-chain-length PFAS, respectively, demonstrates that retention processes significantly impacted migration. The accumulation of shorter-chain PFAS at the bottom of the core is hypothesized to result from limited deep infiltration and potential-enhanced retention associated with the calcic horizon.

PFAS Porewater concentrations in unsaturated soil: Field and laboratory comparisons inform on PFAS accumulation at air-water interfaces

Poly- and perfluoroalkyl substance (PFAS) leaching from unsaturated soils impacted with aqueous film-forming foams (AFFFs) is an environmental challenge that remains difficult to measure and predict. Complicating measurements and predictions of this process is a lack of understanding between the PFAS concentrations measured in a collected environmental unsaturated soil sample, and the PFAS concentrations measured in the corresponding porewater using field-deployed lysimeters. The applicability of bench-scale batch testing to assess this relationship also remains uncertain. In this study, field-deployed porous cup suction lysimeters were used to measure PFAS porewater concentrations in unsaturated soils at 5 AFFF-impacted sites. Field-measured PFAS porewater concentrations were compared to those measured in porewater extracted in the laboratory from collected unsaturated soil cores, and from PFAS concentrations measured in the laboratory using batch soil slurries. Results showed that, despite several years since the last AFFF release at most of the test sites, precursors were abundant in 3 out of the 5 sites. Comparison of field lysimeter results to laboratory testing suggested that the local equilibrium assumption was valid for at least 3 of the sites and conditions of this study. Surprisingly, PFAS accumulation at the air-water interface was orders of magnitude less than expected at two of the test sites, suggesting potential gaps in the understanding of PFAS accumulation at the air-water interface at AFFF-impacted sites. Finally, results herein suggest that bench-scale testing on unsaturated soils can in some cases be used to inform on PFAS in situ porewater concentrations.

Modeling 1-D aqueous film forming foam transport through the vadose zone under realistic site and release conditions

Aqueous film forming foams (AFFFs) have been used to extinguish fires since the 1960s, leading to widespread subsurface contamination by per- and polyfluoroalkyl substances (PFAS), an essential component of AFFF. This study presents 1-D simulations of PFAS migration in the vadose zone resulting from AFFF releases. Simulation scenarios used soil profiles from three US Air Force (USAF) installations, encompassing a range of climatic conditions and hydrogeologic environments. A three-component mixture, representative of major constituents of AFFF, facilitated the exploration of competitive and synergistic effects of co-constituents on PFAS migration. To accurately capture unsaturated transport of PFAS in porous media, the model considers (1) surfactant-induced flow, (2) non-linear sorption to the solid phase, (3) competitive accumulation at the air-water interface, and (4) the moisture-dependence of the air-water interfacial area. Defined PFAS releases were consistent with fire training exercises, emergency responses, and accidental spills of record. Simulation results illustrate the importance of hydrogeologic, climatic, geochemical, and AFFF release conditions on PFAS transport and retention. Comparison of field observations and model simulations for Ellsworth AFB indicate that much of the PFOA and PFOS mass is associated with the air-water interface and the solid phase, which limits their migration potential in the vadose zone. Results also show that rates of migration in the aqueous phase are largely controlled by hydrogeologic properties, including recharge rates and hydraulic conductivity. AFFF spill scenarios varying in volume, concentration, and frequency reveal the importance of release characteristics in determining rates of PFAS migration and concentration peaks. Variability is attributed to non-linear sorption processes, where, contrary to simple linear partitioning formulations, transport is strongly affected by the concentration of PFAS species. Simulations also demonstrate the importance of modeling the AFFF as a mixture since competitive interfacial accumulation effects are shown to enhance the mobility of less surface-active PFAS compounds.

An integrated analytical modeling framework for determining site-specific soil screening levels for PFAS

Soils at many contaminated sites have accumulated a significant amount of per- and polyfluoroalkyl substances (PFAS) and may require remediation to mitigate leaching to groundwater. USEPA's current approaches for determining soil screening levels (SSLs) were developed for non-PFAS contaminants. Because many PFAS are interfacially-active with unique leaching behaviors in soils, the current non-PFAS-specific soil screening models may not be applicable. Following USEPA's general methodology, we develop a new modeling framework representing PFAS-specific transport processes for determining site-specific SSLs for PFAS-contaminated sites. We couple a process-based analytical model for PFAS leaching in the vadose zone and a dilution factor model for groundwater in an integrated framework. We apply the new modeling framework to two typical types of contaminated sites. Comparisons with the standard USEPA SSL approach suggest that accounting for the PFAS-specific transport processes may significantly increase the SSL for some PFAS. For the range of soil properties and groundwater recharge rates examined, while SSLs determined with the new model are less than a factor of 2 different from the standard-model values for less interfacially-active shorter-chain PFAS, they are up to two orders of magnitudes greater for more interfacially-active longer-chain PFAS. The new analytical modeling framework provides an effective tool for deriving more accurate site-specific SSLs and improving site characterization and remedial efforts at PFAS-contaminated sites.