Air Bubbling Assisted Soil Washing to Treat PFAS in High Organic Content Soils

Abstract

The soil-washing technique has been successfully utilized for the remediation of PFAS-contaminated soils. Prior studies have shown that the organic carbon (OC) content and grain size of soil determined the efficiency of PFAS removal during washing. However, most of the past studies have focused on soils with a low OC content, typically ranging from 0–3%. In this study, we explored the use of a novel process where soil washing was combined with air bubbling (or foam fractionation) to aid in the removal of PFAS from high OC-content soils (~4–20%). Treatment with air bubbling of high OC soil (~20%) with perfluorobutane sulfonate (PFBS) and perfluorooctanoate (PFOA) did not enhance their removal, as they featured low surface activity. However, we observed an improvement in the extraction of perfluorooctane sulfonate (PFOS) from 27% to 42% with bubbling, consistent with the higher surface activity of PFOS compared to PFOA and PFBS. Perfluorodecanoic acid (PFDA) was irreversibly adsorbed to the high OC soil and was not removed efficiently by both bubbling and soil washing. A slight improvement in PFDA removal (6–13%) was observed when a co-surfactant (cetyltrimethylammonium chloride) was added and when the OC content was reduced to ~4% by the addition of nonorganic sand to the contaminated soil prior to soil washing. This suggested that the interaction of PFDA with OC was the dominant factor determining its extraction from soil. In conclusion, our results indicated that soil washing alone was sufficient for the removal of short-chain PFAS from soil. Although bubbling had a mild effect on the removal of some long-chain PFAS from the solution, it did not help in the overall removal of PFAS from high OC soils, highlighting the difficulty in the treatment of high OC-content soils and that immobilization of PFAS would be an ideal approach in managing such contaminated sites.

1. Introduction

The importance of soils as potential global per- and polyfluoroalkyl substance (PFAS) reservoirs was highlighted as early as 2012 [1,2]. Samples collected from 60 locations in six countries revealed estimated global soil loadings of PFOA and PFOS at 1860 and >7000 metric tons, respectively [1,2]. Interestingly, the samples were collected from locations far from known PFAS-contaminated locations, revealing their ubiquitous presence in the soil environments. A follow-up study utilized data from previous findings and estimated the average loads of PFOA and PFOS to be 1000 metric tons [2]. These results stated the importance of the vadose zone as a major reservoir for PFAS. PFAS accumulated in this zone can serve as a long-term contamination source to groundwater, surface water, and, eventually, the atmosphere and biota [3,4,5,6].

Soil remediation strategies are chosen based on the treatment goals and the nature and concentrations of the contaminants, and the protection of human life and environment are considered key elements [7]. Broadly, soil treatment technologies can be classified into three categories: (i) physicochemical, (ii) biological, and (iii) thermal. The approaches for soil remediation can be classified as either in situ or ex situ. Gas fractionation is an approach that was demonstrated in a recent study to increase PFAS removal by nine times compared to other soil remediation approaches [8]. This was an in situ approach that was impacted by the gas type, the flow rate, and the fractionation time [8]. Similarly, more in situ approaches have been recently investigated for PFAS-contaminated soils. A recent study showed the successful extraction of an undisturbed, 3 m deep, sandy vadose zone soil that was contaminated with aqueous film-forming foam (AFFF) by utilizing gas-fractionation-enhanced soil washing in Norway, with removal efficiencies ranging from 11–73% [9]. However, the efficacy of an in situ process is affected by the soil permeability, affected by the undisturbed soil structure [10]. PFAS tend to preferentially adsorb to air–water interfaces in soil pores or micropores, and thus, this process can unintentionally permeate PFAS beyond the remediation zone and in the surrounding matrix [10].

Soil washing is an ex situ physicochemical technology that has been historically used to remove organic and inorganic contaminants such as metals, polycyclic aromatic hydrocarbons, and polychlorinated biphenyls [10,11,12,13]. Soil washing can be viewed as the removal of contaminants from soil usually by physical and/or chemical separation to reduce the volume going to waste or incineration. Compared to the in situ approaches, this approach offers better handling and control and can limit the migration of contaminants into the surrounding matrix. For a contaminated site post-excavation, a wash solution is used to aid in the extraction of contaminants from the soil phase. The waste solution becomes concentrated with the chemical of interest, and the treated soil can potentially be transferred back to the original site.

There is increasing research performed on identifying soil-washing techniques for PFAS removal from contaminated soils [2,10,14,15,16]. One study investigated the soil from Eielson Air Force Base in Fairbanks, Alaska, USA, which was treated in bench-scale and full-scale setups. This was done by collecting three 20 L soil samples and then treating them in three subsequent stages: (i) size separation, (ii) attrition, and (iii) chemical extraction using water in a 5:1 liquid/soil slurry [15]. The soil was mainly contaminated with perfluorooctanoate (PFOA), perfluorooctane sulfonate (PFOS) and perfluorobutane sulfonate (PFBS), and post-treatment, enrichment of mainly PFOS and also PFOA was observed on the fines (<0.074 mm) [15]. A recent study also treated soil contaminated with AFFFs using soil washing and studied the influence of the grain size, organic carbon (OC) and organic matter residue (decomposable part of soil) content on PFAS sorption. The partition coefficient (K_d) values increased in the following order: PFCAs < PFSAs < FTS < FOSA, with the fine-grained silt and clay fractions demonstrating the highest sorption potential due to their high OC content [17]. The retardation in soil for PFAS of the same chain length increased in the order PFCAs < PFSAs < FTS < FOSA, which was directly related to the K_d values of the compounds [17]. Another study combined soil washing with foam fractionation and electrochemical treatment as a three-step process to desorb and treat PFAS from contaminated soils [18]. Soils with <0.4% OC content, contaminated predominantly with PFOS, were treated, and the effects of the soil particle size, pH, and soil/liquid ratios were evaluated. Approximately 95% of the PFOS was removed by soil washing, ~95–99% through foam fractionation and aeration, and ~97% removal was achieved using electrochemical oxidation via the three-stage treatment [18]. A simulation study modeled the adsorption of PFCAs and PFSAs in soils and found a strong correlation between their K_d values based only on the soil OC and silt and clay contents and PFAS chain length [19]. The organic carbon content of soils was found to be the primary factor affecting sorption, while sorption at mineral sites, especially for short-chain PFAS, has an important contribution to soils with a low OC and high silt and clay content [19].

The OC content in the previous studies ranged from 0–3% [10,15,16,17,18], classifying these soils as nonorganic/sandy. There is a scarcity of work that assessed the treatability of PFAS in organic carbon-rich soils. Soil washing alone may not be effective in the treatment of PFAS-contaminated soil with high organic matter due to the strong adsorption of PFAS to OC. In this study, we aim to combine soil washing with an aeration/air-bubbling approach to enhance the desorption of PFAS. PFAS are surface active, and foam fractionation is commonly employed to treat contaminated waters [20,21,22]. Air–water interfacial accumulation of PFAS increases with the carbon chain, and hence, the introduction of air bubbles can remove long-chain PFAS from contaminated waters. A recent study showed successful extraction of both short-chain and long-chain PFAS using cationic surfactant-enhanced foam fractionation [20]. In the present study, we hypothesized that the introduction of air bubbles will enhance the desorption of long-chain PFAS from contaminated organic soils, synergically improving the efficiency of soil washing. The primary aim of this work was to combine air bubbling with soil washing to extract PFAS of varying hydrophobicity in organic soils. The specific objectives of this study were to (i) evaluate the effect of bubbling on the extraction of PFAS from contaminated soil with a high OC, (ii) perform a mass balance of PFAS to investigate the compartmentalization of PFAS post-treatment, and (iii) investigate the impact of wash solution chemistry (pH, co-surfactant) on PFAS compartmentalization.

2. Materials and Methods

2.1. Chemicals and Materials

PFAS standards were purchased from Sigma–Aldrich Chemical Co (St. Louis, MI, USA). All solutions used in the experiments were prepared with Milli-Q water supplied by a Milli-Q Gradient-A 10 Millipore (Burlington, MA, USA) (resistance of 18.25 MΩ·cm@25 °C). All chemicals used in the experiments were of reagent grade or higher. Commercial premium topsoil (lawn and garden soil conditioner) was purchased from Home Depot, Marietta, GA, USA, and sand (C33, no OC, specific gravity of 2.5–2.7) was obtained from a wastewater research and innovation facility at Stony Brook, New York, NY, USA.

2.2. Preparation of PFAS-Contaminated Soil

Sand and topsoil were dried in an oven at 70 °C and ground using a mortar and pestle. We utilized commercial topsoil and sand to control the organic content of the soil samples treated in this study. Previous studies have performed their experiments on sandy/less OC soils (0–3%), and thus, we wanted to test the extractability of PFAS from high OC-content soils. This approach also helped us to control the initial concentration of PFAS and to perform a mass balance on PFAS after treatment. Sand and topsoil were mixed in a pre-fixed ratio (75:25 or 25:75) in an HDPE container, and a known amount (200–300 mL) of deionized water (DIW) was added to form a slurry. PFAS was spiked in the slurry from a stock solution, and the container was placed on an orbital shaker overnight. Post mixing, the container was placed in an oven overnight to dry at 70 °C. This procedure was followed to uniformly distribute PFAS in the test soil. The drying process did not alter the chemistry of the soil, as indicated in

Table 1. Characterization of tested sand: soil mixtures and comparison with typical soil values and values utilized by other PFAS studies [14,15,16,18,23].

Analyte	Unit	75 Soil: 25 Sand Non-Dried	75 Soil: 25 Sand Dried	25 Soil: 75 Sand Non-Dried	25 Soil: 75 Sand Dried	Typical Values Observed in Environment	Typical Values Used in Other PFAS Studies
pH		7.47	7.27	7.4	7.39		4–8.2
P	kg/ha	342	862	186	296	0–153
K		ADL	ADL	1132	1636	0–310
Mg		1309	1959	596	774	0–330
Ca		5904	ADL	3069	4049	0–2005	2016–2240
Zn	parts per million	13.11	26.18	5.76	7.82	1–50
Cu		1.77	3.95	0.93	1.13	0.5–20
Mn		20.56	65.24	21.3	33.24	25–100	200–300
B		3.08	5.35	2.27	2.62	0.5–20
Fe		195.19	265.9	140.4	150.19	50–100	14,500–15,000
S		11.9	506.75	12.68	62.48
EC	mOhm/cm	1.13	2.59	0.66	0.99
Organic matter		18.6%	20.6%	3.6%	4.8%	0.1–90% *	0–3%
Gravel content	>2 mm	16%	13.5%	16%	13.6%		0–5 %
Sand		89%	90%	91%	93%		0.3–96.5%
Silt		8%	6%	5%	4%		3–56%
Clay		3%	4%	4%	3%		0–50%
CEC	meq/100 g	18.63	21.37	10.3	13.8	5–25

* low lying, extremely marshy soil. ADL indicates above the detection limit. Columns 3–6 indicate the weight ratio of soil and sand used and whether this mixture was dried or not.

. The increase in nutrients and OC was likely due to a reduction in moisture content after drying.

2.3. Soil Washing Process

The soil/liquid (S:L) ratio is an important treatment parameter that can impact the removal efficiency. Preliminary experiments were conducted to optimize the S/L ratio. The S/L ratios of 1:5, 1:10, 1:20, 1:50, and 1:100 were tested, and it was observed that at ratios below 1:20, the slurry was thick and the mixer paddles were unable to mix the solution uniformly. Previous studies that have utilized soil washing for PFAS remediation have used S/L ratios ranging from 1:5 to 1:25 [14,15,17,18]. For this study, a S/L ratio of 1:20 was chosen. Two ratios of organic soil/sand were chosen (75:25 and 25:75) to simulate the contamination of PFAS in organic soil with a high OC (18–21%) and sandy soils with an OC < 5%, consistent with previous studies.

A fixed mass of dried sand–soil mixture was placed in separate HDPE containers, and a known amount of wash solution was added to keep a constant S/L ratio of 1:20. The containers were placed on a Phipps and Bird’s PB-900™ series programmable jar tester setup (Figure 1). Holes were drilled in the HDPE bottle, and a cylindrical stone diffuser was attached at the bottom; bubbling tubes were connected to the ports at the bottom of the container, and the air-bubbler apparatus was turned on. The mixing speed was set to 100 rpm, and the airflow rate was set at 0.5 L/min to generate microbubbles, and the experiment was run for one hour. Duplicates were run for each sample condition, and a control (no mixing, no bubbling) was also run in duplicates for each batch of operation.

2.4. Sample Processing

After treatment, 3 mL of the aqueous wash solution was drawn using a syringe and transferred to a 5 mL polypropylene (PP) tube. A known amount of surrogate was spiked into the sample, vortexed, and centrifuged at 3000 rpm for 3 min. The supernatant was passed through a 0.22 μm Regenerated Cellulose (RC) membrane filter. The final solution was collected in a separate PP tube, diluted 1:1 with methanol and stored at 4 °C prior to analysis. The remaining wash solution was passed through a 0.45 μm nitrocellulose membrane filter to collect the fines/unsettled solids. The filter was weighed before and after the solution was poured through it, placed in a 50 mL PP tube and dried in the oven at 70 °C. The remaining soil at the bottom was rinsed with DIW and scooped out using a disposable spatula into 50 mL PP tubes, and the resulting treated soil was dried in the oven at 70 °C. Post drying, both the treated and untreated dried soils and the filter containing the unsettled solids were extracted following EPA Method 1633 [24]. Briefly, this involved the extraction of 0.25 g of soil (surrogate spiked) using 0.3% methanolic ammonium hydroxide at different volumes, followed by a water addition to bring up the volume to ~35 mL. To remove the methanol, a N₂ blowdown was performed until the final volume reached ~10 mL, after which ~40 mL of water was added. The pH of the solution was verified to be ~6–7, and solid-phase extraction using Oasis WAX cartridges was performed, and the final eluate was collected in PP tubes and stored for PFAS analysis. For the extraction of unsettled solids, the entire filter and a pristine filter paper (as control) were used for extraction instead of a fixed amount of soil, with the rest of the procedure remaining unchanged.

The HDPE containers, mixing paddles, and stir rods were rinsed with methanol to assess any PFAS absorbed to surfaces. The eluate was collected and stored for further analysis. Finally, the aqueous solution that was transferred into a separate container was passed through a 0.45 μm nitrocellulose filter to filter out non-settleable solids (fines). The filter papers were then extracted using EPA Method 1633, similar to that used for untreated and treated soils. Post SPE, the eluate was stored for further analysis.

2.5. Sample Analysis

An aliquot of the samples was taken out for further dilution (to fall within the calibration), and the pH was adjusted to near neutral using 10% acetic acid (if needed). A total of 10 µL of 100 μg/L of isotopically labeled standards were added prior to analysis to provide recovery-corrected PFAS concentrations. The samples were analyzed using an Agilent (Santa Clara, CA, USA) 6495B triple quadrupole liquid chromatography–tandem mass spectrometer (LC-MS/MS) equipped with electron spray ionization (ESI) in the negative mode. The limit of quantification (LOQ) was 0.01 ppb for the water samples and 0.05 ppb for the soil, and a valid peak was defined as the peak of the analyte with a signal-to-noise (S/N) ≥ 10. Internal standards (IS) were composed of four mass-labeled compounds as ¹³C₃-PFBS, ¹³C₈-PFOA, ¹³C₈-PFOS, and ¹³C₂-PFDA, and the average recoveries ranged from 88–96% in the water samples and 75–90% in the soil samples. Details about the MRM transitions and LC-MS/MS conditions used are provided as Supplementary Information.

3. Results and Discussion

3.1. Treatment of Soil Contaminated with Single Solute PFAS

Soils spiked with single PFAS were prepared at a topsoil/sand ratio of 75:25 and at a concentration of ~0.1 μg of PFAS/g soil. PFOA, PFOS, and PFDA were used as model PFAS compounds to assess the treatment efficiency. Long-chain PFASs were selected for the experiments because short-chain PFASs feature lower sorption to soil and, hence, are expected to mobilize readily during soil washing [18].

As seen in Figure 2, ~4 ± 23% of the PFOA was extracted from the soil to the aqueous phase for the control sample. The extraction efficiencies from the soil phase to the aqueous phase for the ‘only mixing’ and ‘mixing and bubbling’ conditions were higher at 46.5 ± 3% and 46.4 ± 9.3%. Interestingly, bubbling did not improve the extraction efficiency for PFOA. For the tested conditions, the contribution of PFOA from sorption loss and unsettled solids (fines) was <1% each. For PFOS, mixing alone led to ~27.5 ± 19.2% of the PFOS extracted from the soil phase into the aqueous phase, and this value increased to 41.6 ± 13.4% for the samples that were mixed and bubbled. Bubbling, along with mixing, led to a loss of ~26% of PFOS mass, which was likely aerosolized due to the generated air bubbles [20]. Under similar conditions, after one hour of treatment, <1% PFDA compartmentalized in the aqueous phase, and this number was ~1.8 ± 2.9% with only mixing and ~13 ± 0.3% for the mixing and bubbling condition. Thus, bubbling did positively impact the extractability of hydrophobic PFOS and PFDA into the aqueous phase, compared to the control samples and samples that were simply mixed. PFOS and PFDA exhibit higher surface activity compared to PFOA, and hence, the results are consistent with their increased ability to partition onto air–water interfacial areas [25].

The difference in the extraction efficiencies for the soils contaminated with single solute PFAS can be attributed to their adsorption to the organic matter in the soil phase. The hydrophobicity for the three PFASs studied increases in the order of PFOA < PFOS < PFDA [19,26]. This was confirmed experimentally by a recent study investigating soils contaminated due to AFFF, where the K_d values for PFOA, PFOS, and PFDA were ~1.2, 6.5, and 22 L/kg, respectively. The higher the K_d value, the stronger the adsorption to the organic phase, resulting in reduced extractability from the soil. Thus, decreasing the solid-to-liquid ratio can also enhance the mass of PFAS extracted from soil. This observation was demonstrated when comparing the extractability of PFOA and PFOS from soil using a lower solid-to-liquid (S:L) ratio of 1:100 (Figure S1). The removal efficiency was improved during soil washing from 46.5% to 70% for PFOA and from 27.5% to ~41% for PFOS when the S/L ratio was changed from 1:20 to 1:100. These results suggested that sequential washing of contaminated soil or using a larger washing solution volume can improve the removal of PFAS.

We assessed three conditions to improve the extraction of PFDA into the aqueous phase: varying the wash solution pH; adding a cationic surfactant, cetyltrimethylammonium chloride (CTAC); and decreasing the OC content by increasing the sand content. As seen in Figure 3a, varying the wash solution’s pH did not improve PFDA extraction efficiencies, with ~85% PFDA still present in the soil and ~15% unaccounted for in the three conditions tested (likely aerosolization). This is contrary to the data obtained in previous studies [14,18], where the percentages of PFOA, PFOS, and PFDA bound to multiple soil samples decreased as the solution pH increased. This was attributed to the electrostatic interactions between the PFAS and the soil, which decreased as the solution basicity increased. However, these studies treated soils with a low OC content (0.3–3.1%). Additionally, the previous studies did not employ bubbling as performed in the present study with high OC contents in soils. The desorption of PFDA, influenced by the air bubbles introduced in this study, may have masked the effects of the pH observed in the other study.

We also added 1 mg/L of CTAC to the wash solution (DIW, no pH adjustment) to observe the effect on PFDA extraction. Our previous study on foam fractionation (aeration) showed that the addition of a cationic co-surfactant, such as CTAC, enhanced the removal of PFAS from contaminated water [20]. This effect was hypothesized to result from the interaction between anionic PFAS and the cationic functional group of CTAC, which increased the surface activity of the resulting complex and thereby improved their removal efficiency during foam fractionation. As seen in Figure 3b, the extraction efficiency for the control sample was ~1.6 ± 19%, which increased to ~11 ± 28% with mixing. The extraction efficiency for the samples that were mixed and bubbled was ~1.3 ± 14%. Interestingly, we detected PFDA in the extraction solution, with ~13, 7, and 10% of the initial PFDA compartmentalized in the samples that were simply mixed, mixed and bubbled, and for the control sample. Although the addition of CTAC did not lead to a complete extraction of PFDA, it demonstrated a slight increase in the migration of PFDA to the aqueous phase. This may be attributed to the competition between the long-chain (C16) surfactant molecule CTAC and PFDA (C10), which could lead to the substitution and adsorption of CTAC onto sites containing PFDA [27]. This was elucidated in a recent study using GAC that showed that the adsorbed short-chain PFAS were prone to displacement upon the addition of long-chain PFAS [27]. Additionally, CTAC may interact with PFAS to form ion-pair complexes that could contribute to the extractability of PFDA. This latter explanation was observed in previous work with surfactant-enhanced foam fractionation and coagulation processes [20,27]. However, adsorption of the CTAC on soil may also improve the electrostatic interaction of PFAS to soil due to an increase in cationic functional groups, as observed for other adsorbents [28]. Future studies should assess the addition of an anionic or non-ionic surfactant that may enhance the removal of PFAS from the soil due to charge repulsion and enhanced surface activity of the co-surfactant. We also noticed large analytical variability in the CTAC-added samples. Although we did not specifically test the impact of CTAC on the 1633 method, we anticipate interferences caused due to the enhanced surface activity of the added CTAC. Future studies should assess the impact of co-surfactants on analytical performances.

Finally, we tested the addition of sand to contaminated soil to reduce the OC content of the soil. A total of 6.25 g of soil (0.4 μg PFDA/g soil) was mixed with 18.75 g of sand, to which 500 mL of DIW was added without any additives. This was to maintain a concentration of PFDA in the topsoil/sand (25:75 ratio) mixture at 0.1 μg/g. As shown in Figure 3c, compartmentalization of PFDA in the wash solution/aqueous phase was ~11 ± 4.6% and ~11 ± 3.2% for the samples only mixed and mixed and bubbled, respectively, and <3% for the control samples. Thus, varying the soil/sand ratio showed an improvement in the extraction efficiency in the wash solution. These results were similar to the previously tested condition (CTAC addition), where ~10% PFDA compartmentalized into the aqueous phase. Thus, under the tested conditions for high OC soil, we did not observe a significant improvement by the (i) addition of cationic surfactant, CTAC, (b) varying the sand/soil ratios, and (c) combining mixing with air bubbling for PFDA-contaminated soils.

3.2. Treatment of Soil Spiked with Equimolar Mixture of PFAS

We also tested the extraction efficiency of the soil (75 soil, 25 sand) spiked with 0.5 nmol/g of PFBS, PFOA, PFOS, and PFDA each (DIW, pH~6). PFBS was added to the system to represent a short-chain compound to compare its extractability to the long-chains. The spiked mass concentrations for PFBS, PFOA, PFOS, and PFDA were 0.15, 0.20, 0.25, and 0.26 μg/g, respectively. As seen in Figure 4a, the extraction efficiency of PFBS into the aqueous phase was 64 ± 2.8% for the control sample. This increased to ~84.8 ± 2.7% when mixed and was ~76.7 ± 11.3% when mixed and bubbled. For PFOA, without any mixing and bubbling (control), ~13.2 ± 7.4% was extracted into the aqueous phase, and this number increased to ~60% for the samples that were only mixed and the samples that were mixed and bubbled. Similarly, the extraction efficiency for PFOS and PFDA for the samples only mixed and mixed and bubbled was ~30%. However, ~30–40% of PFOS and PFDA were unaccounted for after the treatment, and we did not detect either of those PFASs in the aqueous solution. This could be either due to the aerosolization of hydrophobic PFOS and PFDA during the treatment, adsorption on the materials used in the experiments, or the lack of a representative soil aliquot sample processed and analyzed post-treatment.

The results for both the single solute and equimolar PFAS-contaminated soil experiments demonstrate that under similar conditions, the extraction of PFAS from the soil into the aqueous phase is inversely correlated to their hydrophobicity/K_d values. Compounds such as PFDA, with the highest K_d value amongst those tested, did not display significant extraction from the soil, even with the variations in the pH of the wash solution and CTAC addition. This is contrary to the previous results for both long- and short-chain PFASs in contaminated waters, where the addition of CTAC showed a near-complete removal of both PFBS and PFDA via foam fractionation [29]. This could be attributed to (a) poor extraction from the soil phase, suppressing the effects of bubbling via an air-flotation mechanism and (b) resorption of PFAS onto readily available organic matter present during the treatment of contaminated soils.

3.3. Mass Balance of PFAS After Treatment

To understand the compartmentalization of PFAS after treatment in the various soil and aqueous fractions, a mass balance on the four selected PFAS was performed (Figure 5). The various fractions analyzed were soil, aqueous (wash solution), filtered solids (unsettled fines), and sorption loss to containers and apparatus. As observed in Figure 5, the mass balance was successfully closed for all the compounds with some analytical variability due to the complexity of the soil matrix with a high OC content. The amount of PFAS mass present in the aqueous fraction was as expected and decreased with increasing hydrophobicity of the studied PFAS (PFBS > PFOA > PFOS > PFDA). We did not observe any significant loss in PFAS mass due to sorption onto containers. Interestingly, we noticed a significant mass associated with the unsettled fines for PFOS and PFDA. This observation is consistent with a previous study, where PFAS were shown to accumulate in fines due to their higher OC contents and larger surface areas compared to course materials, thus causing an increased sorption of PFAS [15].

Table 2. Summary of PFAS concentration detected in fines and soil after treatment.

Condition	Initial Concentration in Soil (μg/g)	Concentration in Soil After Treatment (μg/g)	Concentration in Fines After Treatment (μg/g)	Mass of Fines Captured (Underestimated) (g)
Mixing only	0.1	0.09 (~10% ↓)	0.27 (170% ↑)	0.47
Mixing and bubbling	0.1	0.06 (~40% ↓)	0.25 (150% ↑)	0.65
No mixing or bubbling (control)	0.1	0.09 (~10% ↓)	0.08 (20% ↓)	1.31

summarizes the concentration of PFOS in various fractions. The mass of fines captured in this study is underestimated, as a significant fraction settled down prior to sampling. Irrespective of the amount captured, the concentration of PFOS in fines was 150–170% higher compared to the untreated soil. These results suggested that the separation of fines during soil washing is important and should be treated separately, as it may be easier to extract PFAS from course materials.

4. Conclusions

The current study aimed to utilize soil washing assisted with air bubbling as a technique to remove PFAS from contaminated, high OC (organic) soils. This was done by spiking commercially available topsoil with PFAS, followed by soil washing that was assisted with air bubbling at a constant soil/sand ratio. The results from this study indicated that for PFBS and PFOA, air bubbling did not have an effect on the overall compartmentalization. An increased removal for PFOS by ~1.5X was observed when the bubbling was combined with soil washing in a single solute treatment. The current setup was unable to successfully extract PFDA from the contaminated soil. A slight improvement (6–13%) in the migration to the aqueous phase was observed with the addition of CTAC and by mixing nonorganic sand to contaminated soil prior to soil washing. This was due to the induction of a competitive effect by CTAC addition for PFAS migration and due to a decrease in the overall organic content of the soil/sand mixture, preventing the possibility of resorption of desorbed PFDA, respectively. The addition of bubbling along with mixing to interact with the soil did not significantly improve PFDA extraction, even with the addition of CTAC or sand.

However, this effect is expected to be more pronounced if the desorption of PFDA from the soil was made possible by changing the PFAS loadings onto the soil or by the addition of an anionic surfactant to compete for the adsorption sites onto the organic carbon present in the soil. The incomplete removal of PFAS observed in this study was due to the high OC content of the soils tested. Another challenge was that the high OC content of the soil increased the analytical variability in the study, masking the effect of aeration during the soil-washing process. To the best of our knowledge, this is the first study to assess the treatability of PFAS in high organic soils. Our results indicate that it may be more economically practical to immobilize hydrophobic PFAS in organic soils rather than trying to extract using bubbling or soil washing. Thus, future or recommended work on soil remediation should focus on (i) reducing the organic carbon content to observe the effect of the solution’s pH and additives, such as cationic/anionic/non-ionic surfactants in the presence of air bubbles; (ii) observing the compartmentalization and competition effects of a wider suite of PFASs in real-world AFFF-impacted soils; and (iii) separating the fines enriched with PFAS during soil washing and treating them separately.