Sustainable Remediation Strategies and Technologies of Per- and Polyfluoroalkyl Substances (PFAS)-Contaminated Soils: A Critical Review

Abstract

Per- and polyfluoroalkyl substances (PFAS) have been reported to contaminate soil as a result of improper management of waste, wastewater, landfill leachate, biosolids, and a large and indiscriminate use of aqueous film-forming foams (AFFF), posing potential risks to human health. However, their high chemical and thermal stability pose a great challenge for remediation. As a result, there is an increasing interest in identifying and optimizing very effective and sustainable technologies for PFAS removal. This review summarizes both traditional and innovative remediation strategies and technologies for PFAS-contaminated soils. Unlike existing literature, which primarily focuses on the effectiveness of PFAS remediation, this review critically discusses several techniques (based on PFAS immobilization, mobilization and extraction, and destruction) with a deep focus on their sustainability and scalability. PFAS destruction technologies demonstrate the highest removal efficiencies; however, thermal treatments face sustainability challenges due to high energy demands and potential formation of harmful by-products, while mechanical treatments have rarely been explored at full scale. PFAS immobilization techniques are less costly than destruction methods, but issues related to the regeneration/disposal of spent sorbents should be still addressed and more long-term studies conducted. PFAS mobilization techniques such as soil washing/flushing are hindered by the generation of PFAS-laden wastewater requiring further treatments, while phytoremediation is limited to small- or medium-scale experiments. Finally, bioremediation would be the cheapest and least impactful alternative, though its efficacy remains uncertain and demonstrated under simplified lab-scale conditions. Future research should prioritize pilot- and full-scale studies under realistic conditions, alongside comprehensive assessments of environmental impacts and economic feasibility.

1. Introduction

Per- and polyfluoroalkyl substances (PFAS) consist of a wide range of synthetic compounds of variable molecular weight which, due to their peculiar physicochemical properties, have been used in a vast variety of industrial and commercial applications such as aqueous film-forming foams (AFFFs) used in firefighting, waterproof, stain-resistant, and non-stick coatings, food packaging, paints, and technology devices (Figure 1) [1,2,3].

Despite the numerous benefits these applications have provided to societal development, they have contributed to the increasing diffusion of PFAS into the environment, especially into surface water, groundwater, and soil [4]. Especially regarding the latter, PFAS contamination has been reported to be relevant in airports and fire training areas, due to large and indiscriminate use of AFFFs, and urban and industrial contexts, due to inadequate treatment and/or disposal of PFAS-containing waste, landfill leachate and biosolids, and uncontrolled wastewater discharge [5,6].

Over the past few decades, concerns about the occurrence of PFAS in the environment increased as a consequence of their adverse effects on human and other living organisms [7]. In soil, PFAS can potentially compromise soil health and ecosystem functioning, posing risks to microbial communities [8]. Additionally, PFAS can be taken up by plants, mainly for root uptake [9], which may affect plant physiology and growth, and serve as a pathway for PFAS transfer through the food chain, ultimately leading to human exposure [10,11]. In addition, other relevant exposure pathways include direct contact with contaminated soil (via dermal absorption or accidental ingestion) [12], dispersion of PFAS-containing dust particles into the air with potential inhalation [13], and the leaching of hydrophilic PFAS into drinking water sources (surface and groundwater) [14]. Documented adverse effects of PFAS to human health include endocrine disruption, thyroid hormone alterations, reproductive and immune system interferences, increased cholesterol levels, damage to liver and kidneys, and possible association with cancers [15,16,17,18].

PFAS are highly persistent in the environment due to their strong chemical bonds, earning them the name “forever chemicals” [19]. This poses a significant challenge in selecting remediation technologies that are not only effective in removing PFAS but also environmentally sustainable and economically feasible. In this context, treatment technologies to remove PFAS from aquatic and terrestrial environments are being developed and extensive research is ongoing. While the initial focus was on PFAS removal from aqueous matrices, allowing scientific and technological advancements as documented by several review papers [20,21,22,23,24], research on soil remediation by PFAS contamination is at an early stage. Previous reviews on soil remediation by PFAS were specific to individual technologies and processes such as PFAS (im)mobilization-based [25,26], biological [27], and destruction [28] ones, thermal treatment [29], sorption [30], or limited to the emerging treatment technologies [31]. However, more attention should be paid to the environmental and economic impacts of the remediation strategies. Issues like operating costs and energy demands (including pretreatment steps), risks of secondary pollution, and regeneration/reuse/disposal of exhaust materials have not often been addressed. Moreover, several technologies, although promising, have been tested only under controlled and simplified conditions at pilot- or even bench-scale, and assessments regarding their scalability and real-scale implementation are still lacking. The present review aims to fill those gaps providing researchers and stakeholders with a summary and critical discussion of all the different treatment technologies to date explored for remediation of PFAS-contaminated soil for effective decision-making.

The discussion that follows is divided into paragraphs with the objectives of (i) briefly describing the main PFAS physicochemical characteristics which influence their behavior in soil (Section 2), (ii) summarizing, critically comparing, and discussing studies on PFAS remediation through immobilization, mobilization and extraction, and destruction techniques (Section 3, Section 4, and Section 5, respectively), and (iii) providing the state-of-the-art and future perspective of the reported technologies focusing on the balance among effectiveness, in terms of process performance and sustainability, in terms of environmental and economic impacts, and scalability.

In line with those objectives, a traditional review approach was adopted. Relevant publications were identified through databases such as ScienceDirect and Google Scholar, using various keyword combinations related to PFAS soil contamination and remediation technologies. Greater emphasis was placed on recent publications from high-impact journals. Additional sources were included to ensure a comprehensive discussion. This methodology enabled the authors to screen approximately 60 potentially relevant papers, of which 49 were selected for critical analysis, providing a thorough overview of the current state of research.

2. PFAS Characterization and Behavior in Soil

PFAS include those substances made up of a carbon chain whose atoms are bonded to fluorine atoms as expressed by the general formula C_nF_2n+1– [32]. More specifically, when all the carbon atoms of the chain are bonded to fluorine ones, those compounds are classified as perfluoroalkyl substances; otherwise, they are defined as polyfluoroalkyl substances. Furthermore, the carbon chain, widely known as the tail, is bonded to a radical group, such as carboxylic (–COOH) or sulfonic (–SO₃H) referred to as the head [33].

Perfluoroalkyl substances are classified on the basis of the functional group. Specifically, perfluoroalkyl acids (PFAA) include a carboxyl or sulfonic group, whereas perfluoroalkyl sulfonamides (FASA) include a sulfonamide group in their chemical structure. PFAAs are further classified into two main categories: perfluorocarboxylic acids (PFCA) and perfluoroalkanesulfonic acids (PFSA), characterized by the carboxylic and sulfonic groups, respectively. Even FASAs are further subclassified into perfluoroalkane sulfonamide acetic acids (FASAA), N-alkyl perfluoroalkane sulfonamides (N-alkyl FASA), and perfluoroalkane sulfonamido ethanols (FASE), which contain one or more alkyl groups in the head of the molecule, bonded to the sulfonamide spacer (Figure 2) [34]. Additionally, other types of PFAS, such as fluorotelomers, are classified as precursors due to their potential to undergo chemical transformation and degradation, which can result in the formation of more persistent and bioaccumulating PFAS [35]. Another well-established PFAS classification lies on carbon chain length: PFCA with eight or more carbon atoms and PFSA with six or more carbon atoms are defined as long-chain PFAS. In contrast, those with a lower number of carbon atoms are considered as short-chain PFAS (Figure 3) [36]. This latter classification is usually taken into account when discussing PFAS’ occurrence, fate, and remediation since the length of the carbon chain substantially influences PFAS’ solubility, mobility, and bioaccumulation [37]. Finally, polyfluoroalkyl substances differ from perfluoroalkyl substances for the presence of at least one carbon–hydrogen (C–H) bond in the carbon chain. This bond acts as a point of weakness thus making polyfluoroalkyl substances prone to (a)biotic transformation. Consequently, many polyfluoroalkyl substances that transform into terminal PFAAs can be considered as PFAAs precursors [34].

The physicochemical properties of PFAS are mainly influenced by fluorine, carbon–fluorine (C–F) bonds and functional groups and their knowledge is essential to understand PFAS behavior in soil (Figure 4). Fluorine has high electronegativity and ionization potential, small atomic size, and low polarizability [38]. The former results in a strong covalent C–F bond characterized by a high dissociation energy of 116 kcal/mol. This leads to high thermal, chemical, and biological stability, thus making PFAS highly persistent and resistant to environmental degradation [39]. Furthermore, the small atomic size of fluorine increases PFAS chemical stability since fluorine atoms generate a dense shield around carbon atoms giving them electronic and steric protection [40]. Finally, the low polarizability reduces intermolecular interactions, such as Van der Waals forces and surface energy, thus providing PFAS with hydro- and lipophobic properties. However, when the carbon chain is bonded to a hydrophilic functional group, such as a carboxylic one, PFAS exhibit amphiphilic properties (i.e., hydrophobic tail and hydrophilic head) [34].

Generally, long-chain PFAS have higher boiling and melting points than short-chain ones (e.g., perfluorotridecanoic acid, PFTrDA, and perfluorobutanoic acid, PFBA, have melting points of 112–123 and −17.5 °C, respectively). Conversely, vapor pressure is higher in short-chain PFAS than long- chain ones [41]. Those properties influence PFAS’ mobility in soil and their volatility. A low boiling point and high vapor pressure enhance both the efficiencies of thermal and physical remediation processes. Conversely, removal technologies such as air stripping are ineffective when targeting at PFAS characterized by low vapor pressure, namely low volatility [20,42]. Carbon chain length also affects PFAS’ solubility (along with pH, salinity, and molecular aggregation), specifically, the shorter the length the larger the solubility [20,43]. Furthermore, solubility is also influenced by the functional groups since many PFCA and PFSA have higher water solubility than FASAs [44]. The organic carbon to water partition coefficient (K_OC) assesses the extent to which PFAS are bonded to the organic matter in the soil. Studies demonstrated that K_OC increases with the length of the carbon chain. In fact, generally, long-chain PFAS tend to transfer from a liquid to solid phase (i.e., soil, organic matter or other sorbent materials) [45,46,47]. Such characteristics make sorption one of the most effective techniques for PFAS removal. However, PFAS can have different adsorption rates depending on chain length and functional group. Particularly, PFAS with sulfonate groups are more easily adsorbed in the soil. For instance, perfluoropentanoic acid (PFPeA) and perfluorobutanesulfonic acid (PFBS) have the same carbon chain length (four carbon atoms) but the latter is more prone to be adsorbed [46,48,49,50,51,52].

Besides physicochemical properties, PFAS behavior in soil also depends on soil geochemical and physical properties such as texture, organic matter and metal oxide contents, ionic strength, salt composition, pH, and presence of co-contaminant [46,53,54].

3. PFAS Remediation Technologies

Generally, when soil is classified as contaminated according to national/regional regulations, it must be remediated by reducing contaminant concentrations below the respective limits. To achieve this, several cleanup techniques can be applied based on soil and contamination characteristics. In the case of PFAS, features reported in the previous paragraph are crucial in understanding contaminant dynamics and guiding researchers and specialists in the selection of the appropriate remediation technologies [55,56,57].

In this context, several techniques for soil remediation by PFAS contamination have been studied and they can be summarized into three main categories: immobilization, extraction, and destruction. Briefly, immobilization treatments reduce the mobility of contaminants through methods such as sorption [58,59,60] and stabilization/solidification (S/S) [61,62]. Conversely, extraction techniques aim at increasing the mobility of contaminants for their removal from soil, and they include soil washing/flushing [63,64,65], phytoremediation [66], and electrokinetic removal [67,68,69]. Finally, destruction techniques, such as bioremediation [70,71], thermal treatments [72,73], and ball milling [74,75], aim at the PFAS degradation mainly by F–C bonds breaking.

3.1. PFAS Immobilization Technologies

Immobilization-based techniques for the remediation of PFAS-contaminated soils aim to reduce the mobility, leachability, and bioavailability of PFAS, thereby mitigating their exposure to ecological and human receptors [25]. PFAS immobilization can be pursued either in situ or ex situ through the addition of sorbent materials and/or binding agents to the contaminated soil matrix. The addition of sorbents promotes sorption, while binding agents are used in a stabilization/solidification (S/S) technique to form a solid matrix that physically confines the PFAS. More specifically, sorption involves the use of sorbent materials, such as activated carbon (AC), biochar (BC), commercial ones (e.g., RemBind^®, RB), and clays, to enhance the retention of PFAS in the soil matrix. S/S involves the mixing of contaminated soil with binders and/or chemical additives to reduce the mobility and bioavailability of hazardous contaminants by chemically binding them within a solid matrix (stabilization), while enhancing soil’s physical characteristics, such as mechanical strength, compressibility, permeability, and long-term durability (solidification).

In order to evaluate the effectiveness of PFAS immobilization strategies, most of the studies calculated the reduction in PFAS leachability compared to an unamended soil scenario. In contrast, only a few of them determined PFAS removal from the soil. In terms of sustainability, generation/pretreatment of the sorbents, operating costs, and management of exhaust materials have been considered.

The S/S technique is based on a physicochemical process consisting of two phases: stabilization, in which chemical reagents reduce the solubility and hazard of the contaminants, and solidification.

3.1.1. Sorption

The flowchart in Figure 5 illustrates the stages of sorption treatment for PFAS-contaminated soil. Many studies and previous reviews use the term ‘sorption’ to encompass both adsorption (the binding of contaminants to the surface of a solid) and absorption (the diffusion of contaminants into the internal structure of porous materials). In this paper, sorption’ refers to the main technique, while, for the reported studies, the original terminology employed by the respective authors is retained.

Table 1. Comparative table of studies on PFAS-contaminated soil treated by sorption.

Soil Type	PFAS (C-mg/kg)	Experimental Conditions	Results (%)	Ref.
Loamy sand	PFOA, PFOS, PFHxS and PFHxA (C: 10)	Bench scale: Pine-biochar (5% w/w)	88.7 (PFOS removal) 40.5 (PFOA removal) 46.2 (PFHxS removal) 8.9 (PFHxA removal)	[58]
Sandy clay loam			66.3 (PFOS removal) 23.2 (PFOA removal) 9.6 (PFHxS removal) 13.4 (PFHxA removal)
Loamy sand	∑29 PFAS (C: 30–31,730)	Bench scale: RB or Pul Fsorb 400 (1–5% w/w; pH = 5.1 or =9.3)	>99 (PFAS leaching reduction; RB) ~97 (pH 5.1)–>99 (pH 9.3) (PFAS leaching reduction; Pul Fsorb 400)	[59]
Sandy	Σ29 PFAS (C: 0.0004–0.67)	Bench scale: RB or GAC (1–5% w/w; pH = 5.1 or =9.3)	>99 (PFAS leaching reduction) ~99.9 (pH 5.1) and ~99 (pH 9.1) (PFAS leaching reduction; RB)
Loamy sand	∑20 PFAS (C: 37,200)	Bench scale: RB (5% w/w; OP, HA)	>99 (Most PFAS leaching reduction) 92 (PFBA leaching reduction)	[60]
Loamy sand	∑12 PFAS (C: 14,000)	Bench scale: RB (15–30% w/w) or RB+ (15%)	99.99 (PFOS removal; 30% RB) 99.92 (PFHxS removal; 15% RB)	[76]
Sandy	∑12 PFAS (C: 2400)		99.91 PFOS removal (25% RB) 99.30 PFHxS removal (25% RB)
-	PFOA (C: 1000)	Bench scale: Carnitine or choline (2% w/w)	58.0 (carnitine) and 57.9 (choline) (bioavailability reduction)	[77]
	PFOS (C: 1000)	Bench scale: Carnitine or choline (2% w/w)	77.5 (carnitine) and 79.4 (choline) (bioavailability reduction)
Sandy	∑11 PFAS (C: 6.4–2510)	Bench scale: PAC, Mts and compost (3% w/w)	28 and 34 (PFAS leaching reduction; compost soil) 28 and 40 (PFAS leaching reduction; montmorillonite)	[78]
Loamy sand	∑6 PFAS (C: 2.4–1000	Bench scale: Activated biochar (0.5–5% w/w)	0–60 (PFAS leaching reduction; w/w < 5%) 23–100 (PFAS leaching reduction; w/w = 5%)	[79]
Sandy clay loam	∑6 PFAS (C: 7.7–3400)		>90 (Most PFAS) and >57 (PFBA) (leaching reduction; w/w: 0.5%) >98 (Most PFAS) and >79 (PFBA) (leaching reduction; w/w: 1–5%)
Peat	∑9 PFAS (C: 1200)	Bench scale: Fe-char amendment (1–20% w/w)	~0 (PFAS leaching reduction)	[80]
-	∑28 PFAS (C: 3093–32,780)	Field-scale: RB100 and RB300 (1–10% w/w)	~85–99.4 (PFAS leaching reduction; RB100; w/w: 1–10%) >16% (PFAS bioavailability reduction; RB100; w/w: 5%) >14% (PFAS bioavailability reduction; RB300; w/w: 5%)	[81]

C: concentrations; RB: RemBind; Pul Fsorb 400: pulverized granular activated carbon; RB+: Rembind plus; GAC: granular activated carbon; OP: Ortho phosphate; HA: humic acid; Mts: montmorillonite; Fe-Char.

summarizes the main studies addressing PFAS sorption for soil remediation, including used materials, experimental conditions, and achieved results.

Kabiri et al. [59] tested the effects of two commercial sorbent materials, RemBind^® 100 (RB) and Pul Fsorb (granular activated carbon Filtrasorb 400™ grounded to the same particle size as RB), on the reduction in leaching of 29 PFAS in either AFFF-impacted loamy sand and sandy soils. Bench-scale simulation of in situ remediation with RB and Pul Fsorb (dosed at 5% w/w) significantly reduced the desorption and leaching of most PFAS compounds in both soils (e.g., PFOS content in leachate reduced from 1800 to <1 μg/L). Authors also demonstrated the efficacy of the adsorption across the normal pH range (4–9) of most soils, while more acidic and basic conditions resulted in greater leaching of many compounds such as PFBS, PFPeA, and PFHxA. Afterwards, authors reported that presence in soil of competing ions (i.e., orthophosphate and humic acid), variation in soil salinity, extreme temperatures, and repeated leaching had irrelevant impacts on the desorption of PFAS from the RB-remediated soils thus highlighting the stability of RB-mediated PFAS immobilization [60]. Bräunig et al. [76] performed a batch-mode remediation of 12 PFAS in an AFFF-impacted soil sample. The remediation occurs through the amendment with RB and RemBindPlus^® (RB+) at high doses (up to 30% w/w). In general, authors observed greater immobilization with increasing perfluoroalkyl chain length (e.g., leaching of PFOS and PFOA decreased by >90% while there were almost no effects for PFBA). Sulfonate PFAS were more effectively immobilized than carboxylates. Furthermore, RB+ resulted more efficiently than RB in leaching reduction in a short chain, likely occurring due to the differences in sorbent compositions. Short-chain PFAS (e.g., PFBA, PFPeA, PFHxA, PFBS) sorbed onto aluminum hydroxide and clay mineral components (more abundant in RB+) through ligand exchange and electrostatic interactions, whereas hydrophobic interactions with activated carbon components became more dominant with PFAS chain length. Authors also assessed the long-term stability (i.e., three years) of the RB-treatment and the reduction in PFAS bioavailability in terms of uptake into earthworms and wheat grass. Likewise, Hearon et al. [77] reported reductions in PFOA and PFOS bioavailability through soil amendments (at 2% w/w) with calcium montmorillonite (CM) clay modified with the nutrients L-carnitine and choline in lab-scale experiments. Particularly, CM-carnitine and CM-choline reduced soil bioavailability and translocation to plants of PFOA by 58.0 and 57.9%, respectively. For PFOS, these reductions were 77.5 and 79.4%, respectively. Hale et al. [78] investigated the efficiency of the adsorption treatment through powdered activated carbon (PAC), montmorillonite, and compost material in AFFF-contaminated soils. Leaching of PFOS decreased for PAC, montmorillonite, and compost soil by 94–99.9, 28–40, and 28–34%, respectively. For other PFAS, the decreases in their concentrations could not be quantified because they were below the limit of detection both before and after the treatment.

Askeland et al. [58] and Sørmo et al. [79] represent two investigations of reduction in PFAS leaching by using biochar (BC) as sorbent material in aqueous batch shaking tests. The former added (at 0 to 5% w/w) BC derived from pine biomass to PFAS-spiked soils, whereas the latter produced BC from waste timber to be added (0–5% w/w) to AFFF-impacted soils. In Askeland et al.’s study [58], PFOS removal increased from 77.6 (unamended soil) to 88.7% in 5% BC-remediated loamy sand soil. In sandy clay loam soil, BC amendment made a far larger contribution with PFOS removal increased from 0.8 to 66.3%. Furthermore, BC improved PFOS retaining in loamy sand and sandy clay loam soils by 20.1 and 42.7%, respectively. Authors speculated that the higher presence of bridging divalent ions and organic matter in loamy sand soil contributed to PFAS adsorption, in contrast to sandy clay loam soil in which BC amendment resulted more critically. A similar conclusion on the role played by the organic matter was achieved by Sørmo et al. [79] by comparing BC amendment of soils with either low or high total organic carbon (TOC) content. Generally, the effect of BC on reduction in PFAS leaching was considerable in the low-TOC soil and less pronounced in the high-TOC soil. For instance, in the former soil, at a BC dose of 0.5% w/w, leaching of PFBS, PFHxS, PFOS, PFHxA, and PFOA decreased by >90 and 57% for PFBA. By increasing the BC dose to 5% w/w, >98 (for PFBS, PFHxS, PFOS, PFHxA, and PFOA) and >79% (for PFBA) leaching reductions were observed. Conversely, the reduction in PFAS leaching from the high-TOC soil varied between 0 and 60% at BC doses < 5%. By testing different conditions of BC activation, authors concluded that BC-mediated soil PFAS remediation can be tailored to site-specific conditions. However, life cycle assessment (LCA) would be conducted to prove the sustainability of the solution. For instance, to achieve sufficient reduction in PFAS leaching from high-TOC soils, a higher degree of BC activation (≥0.75 oxidant to feedstock C) might be implemented. On the one hand, this would imply higher energy consumption. On the other hand, the assessment would also consider the carbon sequestration by BC and the circular valorization of a waste material (i.e., from the wood industry). Regarding this latter point, Zhang et al. [80] tested the use of mineral industrial by-products, namely iron-rich char and slag, as PFAS sorbents for either high- and low-TOC and acidity soils. The amendment with Fe-char reduced the leaching of PFAS for the low-TOC soil up to 99.7%, whereas for high-TOC soil, no significant reduction was observed. Conversely, slag amendment did not exert any considerable effect on PFAS leaching.

Furthermore, the cost of the sorbent materials (either commercial or custom-made) per amount of remediated soil should be taken into account [76,82].

To sum up, sorption has become a commonly investigated method for reducing PFAS mobility in soils and several sorbent materials were tested. However, such a technique seems at a developmental stage since full-scale and long-term evidence of PFAS bioavailability minimization and immobilization are required. This is necessary to provide a greater level of confidence and identify and compare the environmental impacts of the different solutions [60]. Sorption effectiveness was observed to be influenced by factors such as TOC content, competing ions, and pH, making it site-specific. Economically, the use of either commercial or custom-made materials like, for instance, RB at a selling price of USD 3000 per ton, might not be sustainable for extensive applications. In this context, BC, derived from waste biomass and priced between USD 600 and USD 1300 per ton, could offer a more sustainable and low-energy alternative [82]. It performs well in soils with low-TOC content, but its efficiency can be limited by competition with natural organic matter. This highlights the need for further optimization. Finally, the management of used sorbents poses significant challenges since high energy demands for their regeneration and risk of generating secondary waste are issues of particular concern.

3.1.2. Stabilization/Solidification (S/S)

The flowchart in Figure 6 outlines the in situ S/S process. For clarity and comparison purposes, the studies discussed in this section are summarized in

Table 2. Comparative table of studies on PFAS-contaminated soil treated by stabilization and solidification (S/S).

Soil Type	PFAS (C-mg/kg)	Experimental Conditions	Results (%)	Ref.
Clay	Σ30 PFAS	Pilot-scale: PC–FA–GGBS (15%) + GAC (0.2%) (6-year rains simulation)	28 to 4.5 (PFHxS) 0 to 50 (PFPeA) >97 (PFHxA, PFOA, PFHxS, and PFOS)	[62]
Sand	Σ24 PFAS (C: 7–2282)	Bench-scale: PC + GAC, modified clay and AC–clay blend	87.1–99.9 (GAC)	[83]
Sandy Clay Loam	Σ24 PFAS (C: 6–13,676)		95.4–99.9 (GAC)
Loamy sand	Σ14 PFAS (C: 200–1500)	Bench scale: binders (PC, FA, and GGBS; ratio 1:1:2) + additives (PAC, RB, bentonite, calcium chloride, hydrotalcite, chitosan, zeolite)	>90 (Short-chain PFAS; PAC and RB) 99– 99.9 (Long-chain PFAS; PAC and RB) ~0 (Bentonite, Idroalcite, Chitosano, Zeolite) 50 to 38 (PFOS) 28 to 4.5 (PFHxS)	[84]
Sandy Clay Loam	Σ24 PFAS (C: 6–13,676)	Field-scale: PC (5–15%) + RB or FS (5–10%)	>99% (PFASs)	[85]

C: concentrations; PC: Portland cement; FA: fly ash; GGBS: ground granulated blast-furnace base slag; PAC: powdered activated carbon, RB: Rembind; GAC: granular activated carbon; FS: Fluoro-Sorb.

.

Barth et al. [83] investigated the use of Portland Cement (PC) as binding agent along with either granular activated carbon (GAC), modified clay, or activated carbon (AC)-clay blend as additives, to treat PFAS-contaminated soils from two different contaminated sites. PC and GAC resulted to be the best combination since GAC reduced PFAS leachability of 55.8–99.9% (stabilization). Moreover, due to PC contribution, the solidification phase further reduced the leachability for some of the PFAS up to 87.1–99.9%. In another study, Sörengård et al. [84] investigated the performance of S/S treatment, with and without either conventional or novel additives, in terms of the leaching of 14 PFAS spiked in loamy sand soil (stabilization). They also evaluated the mechanical strength of solidification. Specifically, as binder agent, a constant combination of PC, fly ash (FA), and ground granulated blast-furnace base slag (GGBS) was mixed with the PFAS-spiked soil. Seven commercially available additives were tested at a dose of 2% w/w of the binder weight: PAC, RB, pulverized zeolite, chitosan, hydrotalcite, bentonite, and calcium chloride. Authors observed that only AC-containing sorbents (i.e., PAC and RB) significantly reduced the leachability of PFPeA, PFHxA, PFHpA, PFOA, PFBS, PFHxS, and PFOS. Moreover, they observed increasing performances for long-chain and sulfonate group PFAS as previously reported [59] in Section 3.1. With regard to the strength of the solidified soil, the use of additives reduced the strength, especially RB.

This result highlights the trade-off between mechanical strength and contaminant immobilization needs to be balanced to ensure the long-term performance of the S/S soil treatment.

In order to overcome the limitations of lab-scale conditions, Sörengård et al. [62] evaluated the long-term efficiency of a pilot-scale S/S treatment of 3 tons of AFFF-impacted clay soil through artificial irrigation of 15 mm/day (6 years of precipitation) evaluating the reduction in PFAS leaching. Soil was amended with binder (i.e., mix of PC, FA, and GGBS) and PAC and GAC as additives. After the long-term simulation, in leachate collected from the untreated (reference) soil, 10 PFAS were detected with concentrations ranging from 0.0026 μg/L (PFNA) to 32 μg/L (PFOS), whereas 7 PFAS were detected in S/S treated soil with concentrations up to 0.52 μg/L (PFOS). Subsequently, lab-scale tests were also performed and good correlations with the pilot-scale experiment were observed. However, sorption degrees were up to two orders of magnitude higher in the pilot-scale test, highlighting the environmental significance of long-term effectiveness compared to short-term laboratory outcomes. Besides several laboratory experiments, some case studies of full-scale applications especially using commercial sorbent materials are available [85].

S/S techniques show strong performance in reducing PFAS leaching, but they can alter the mechanical properties of soil, creating a trade-off between immediate effectiveness and long-term usability. S/S has well-established data at both laboratory and pilot-scale, demonstrating high and durable performance over time, as well as good adaptability to different soil types. This technology thus appears promising in terms of scalability and cost (approximately USD 228/ton, according to Ciabattoni et al. [86]). However, the critical trade-off between immobilization effectiveness and the mechanical integrity of the treated soil remains an important consideration for long-term application.

3.2. PFAS Mobilization Technologies

Mobilization aims to increase contaminant mobility and bioavailability to enable their subsequent extraction. Mobilization can be achieved through several techniques, including soil washing/flushing, phytoremediation, and electrokinetic remediation. In soil washing/flushing, appropriate extraction solutions are flushed through the contaminated soil to enhance contaminant solubilization or desorption via chemical interactions. Given the strong affinity of PFAS for soil particles, the use of agents such as acids, bases, surfactants, co-solvents, or complexing agents is often required to improve extraction efficiency.

Phytoremediation employs specific plants to uptake and accumulates contaminants from the soil, offering a low-cost and environmentally friendly solution. On the other hand, it requires specific environmental conditions to support plant growth, it is effective only when contaminants can be reached by plant roots, and time for remediation is long (decades).

Finally, electrokinetic (EK) remediation represents an in-situ technique that relies on the application of an electric field by electrodes embedded in the soil to induce the movement of water and charged species, facilitating the transport and removal of contaminants.

3.2.1. Soil Washing/Flushing

This flowchart in Figure 7 summarizes the soil washing process.

Table 3. Comparative table of studies on PFAS-contaminated soil treated by soil washing/flushing.

Soil Type	PFAS (C-mg/kg)	Experimental Conditions	Results (%)	Ref.
Clay	Σ30 PFAS (C: 0.4–2666.5)	Pilot-scale: Soil washing	94.4–97.1 (PFCA short-chain) 85.7–94.9 (PFSA short-chain) 13.6–51.4 (PFSA long-chain)	[63]
Sand	PFOS (C > 751) PFOS (C > 1291)	Field-scale: soil washing (L/S: 5 L/kg) Field-scale: soil flushing (L/S: 15.3 L/kg)	11–73 23–62	[64]
Sandy loamy silt	Σ35 PFAS (C: 56.1)	Bench-scale: Soil washing + MeOH (5–80%), EtOH (5–80%) or ACN (80%) (L/S = 528)	78 (PFOA)–24 (PFOS) (water) 96 (PFOA)–35 (PFOS) (5% EtOH) 75 (PFOA)–37 (PFOS) (5% MeOH) 86 (PFOA)–59 (PFOS) (80% EtOH) 93 (PFOA)–53 (PFOS) (80% MeOH) 100 (PFOA)–80 (PFOS) (80% ACN)	[65]
Sand	PFOS (C: 1000)	Bench scale: soil flushing + EtOH (50%)	>98 (PFOS)	[87]
Silty clay loam	GenX, PFOA, PFOS, PFDA, 6:2 FTAB (C:200 each)	Bench-scale: soil washing + MeOH (1–50%) + EtOH (1–50%) + HPCD (1–10 mg/g)	92–98 (PFAS; MeOH 50% + HPCD 10 mg/g) 51 (HPCD 1 mg/g))–99 (HPCD 10 mg/g) (PFOS)	[88]
Clay	Σ23 PFAS (C: 24,230)	Bench scale: soil washing + O2 (W:S = 2.0–2.5; t = 5–10 min)	58.1 (PFOS; Ozonated oxygen) 72.9 (PFHxS; Ozonated oxygen) 54.7 (PFOA; Ozonated oxygen)	[89]

EtOH: ethanol; MeOH: methanol; HPCD: 2-hydroxypropyl-β-cyclodextrin; ACN: acetronile, L/S = liquid-to-solid ratio.

presents a comprehensive overview of various soil washing/flushing studies, summarizing the soil types, target PFAS, experimental conditions, and extraction efficiencies to enable an effective comparison of different approaches.

Li et al. [65] investigated different combinations and ratios of water and organic solvents (i.e., ethanol, EtOH, methanol, MeOH, and acetonitrile, ACN) for remediation of clay soil collected from a historical PFAS-contaminated site (35 PFAS) through soil washing with both batch and column tests. The highest extraction efficiency was observed for 80% v/v ACN followed by MeOH (when ≥60% v/v) and EtOH (except for 100% v/v). Based on chain length, it was observed that the highest extraction efficiency of short-chain PFAS was achieved at 80% v/v organic solvent. However, increasing the proportion of organic solvent from 0 to 60% v/v led to a slightly decreased extraction efficiency. Conversely, increasing trends in long-chain PFAS (e.g., PFOS) concentration in the extraction solutions were observed as the proportion of organic solvents ranged from 0 to 80% v/v. During column tests, authors recorded the peak in desorption from soil of PFOA and PFOS after a liquid to soil ratio (L/S) of 1 when using extraction solutions at 80 v/v of organic solvent. However, the peak in desorption was reached at L/S of 3–4 with 0–5% v/v of organic solvent. Especially for the latter, authors estimated that, after prolonged washing time (i.e., L/S = 984 corresponding to 137 days), 5% v/v of EtOH is able to 100% remove PFOA (1700 ng/kg_DM). Considering that the presence of abundant organic solvents in soil poses risks to environmental health (e.g., impacts on plant and microorganism growth and soil fertility [90]), the use of extraction solutions with small proportions of EtOH can represent a suitable strategy for long-term PFAS remediation. Such an approach could save solvent purchasing costs and limit the risks related to solvent high volatility and flammability. However, limitations like high water consumption and treatment of PFAS-concentrated leachate collected after soil washing should also be considered.

In this latter regard, some studies investigated holistic approaches aimed at extracting PFAS through soil washing and treating PFAS leachate through adsorption. Senevirathna et al. [87] tested three different organic solvents, namely, EtOH, MeOH, and propanol (PrOH) diluted with water for the extraction of PFOS. They used clay and sandy soil samples in lab-scale batch and column tests. As a result of the former, EtOH and MeOH showed better washing efficiency than PrOH, especially for clay samples, at the optimum solvent concentration of 50–75% v/v. Subsequently, column tests on sandy soil were performed using solution with 50% v/v EtOH (less toxic than MeOH) and less than 2% of PFOS resulted as retained in the soil sample after five bed volumes of solvent flushing. Finally, authors simulated the treatment of leachate collected after the soil washing step, mimicking the conditions of groundwater receiving the PFOS-extraction solution. By testing different commercially available ion exchange resins and ACs, they reported the effective removal of PFOS especially by PFA694E, even though further tests would be required considering real-world groundwater quality. Besides conventional organic solvents like EtOH and MeOH, Usman et al. [88] tested a biomass (starch)-derived solvent, cyclodextrin (HPCD), characterized by non-toxic nature and biodegradability. In their experiments, a silty loam soil sample was spiked with PFAS (GenX, PFOA, PFOS, PFDA, and 6:2 FTAB) and subjected to both batch and column washing tests. HPCD, at a concentration of 10 mg/g, exhibited >90% removal efficiency for all the target PFAS after L/S = 1 under column conditions. However, authors pointed out that when spiking PFAS in soil, their contact time with soil particles is short compared to historically contaminated samples. This leads to higher desorption efficiencies. For this reason, albeit promising and environmentally friendly, the efficiency of HPCD washing solution should be validated in real-world scenarios. Furthermore, authors reported complete removal of PFAS from washing effluent through either AC or BC adsorption (batch mode). However, also in this case, leachate composition may not reflect real-world conditions and batch tests are not representative of full-scale treatment, so results need to be further validated.

In order to enhance soil washing performance towards PFAS removal, Pang et al. [89] investigated the gas fractionation of a real-world PFAS-contaminated soil preliminary. The soil was subjected to water extraction under the optimum conditions with a water to soil ratio (W/S) of two and an extraction time of 10 min. Among the four different gases tested in the study, such as air, oxygen, ozonated air, and ozonated oxygen, the latter exhibited the highest removal efficiencies for PFHxS, PFOS, and total PFAS (72.9, 58.1, and 55.9%, respectively) with considerable increases compared to conventional soil washing with water. This was due to the fact that the dissolved PFAS will concentrate at the interface between the hydrophobic gas bubbles and water. Subsequently, they fractionate into the foam above the water. The sequence of PFAS affinity with gas bubble interfaces should be ozonated oxygen > ozonated air ≈ oxygen > air.

As one of the first full-scale field applications, Høisæter et al. [64] compared in situ and ex situ soil washing of a sandy soil in a firefighting training site. The study demonstrated the higher efficiency of the in-situ approach, which removed 73% of PFOS with a W/S ratio of 5 L/kg. However, ex situ treatment reached removal efficiency of 62% with 15.3 L/kg. The lower costs of in situ washing (i.e., USD 25,000–110,000 per kg PFOS removed) further support its sustainability, especially when compared with ex situ remediation (USD 10,000–600,000 per site, for 0.4–5.4 kg total PFAS), which involves additional infrastructure such as impermeable liners. The superior performance of the in-situ method is attributed to less preferential flow in undisturbed soil. In contrast, excavated, not-compacted soil in the ex situ setup exhibited greater hydraulic heterogeneity. However, a key limitation of the study is that it was conducted exclusively on high-permeability sandy soil. Thus, the results may not be directly applicable to low-permeability soils, such as clay, where infiltration, sorption, and PFAS mobility behave differently. In this regard, Grimison et al. [63] evaluated a full-scale ex situ soil washing of a high clay-content soil (61%). The process employed a soil washing plant that combined physical separation techniques. These include wet screening, particle attrition, density-based separation, and centrifugation, with chemical desorption. This was performed using high-volume water washing (W/S of 18 L/kg) without chemical additives. The system achieved over 94% removal of carboxylic PFAS and between 13.5 and 95% removal of sulfonic ones. Finally, the washing water was subsequently treated with GAC and ion-exchange resins. This latter represents a crucial point when treating soil with washing liquid solution. For this reason, future research should integrate PFAS-rich wastewater treatment into the soil washing process for a more comprehensive approach to effective and consistent PFAS removal [65]. Grimison et al.’s [63] study shows how an engineered ex situ system can successfully treat soils with high clay content, but at higher costs and operational complexity compared to in situ approaches such as the one proposed by Høisæter et al. [64].

In summary, the application of solvents such as 50% ethanol can extract over 98% of PFOS. However, the washing effluent poses significant challenges to economic and environmental sustainability. It necessitates energy-intensive processes to remove residual contaminants and recover the solvents. Moreover, conventional solvents raise additional concerns related to toxicity and cost. However, more environmentally friendly alternatives like HPCD are still undergoing validation for use in historically contaminated sites. Other methods may involve the investigation of the so-called ‘green’ solvents (such as ionic liquid and deep eutectic solvent, less impactful and highly recoverable [91]). Gas-enhanced treatments, such as ozone-based approaches, can eliminate the need for solvents. Nonetheless, ozone generation is itself energy-demanding, and the success of these techniques heavily depends on soil composition. In terms of widespread application, soil washing/flushing is constrained by environmental, operational, and economic factors. Laboratory experiments often rely on artificially contaminated soil and high solvent concentrations. These scenarios differ from real-world conditions at historically contaminated sites, where PFAS are more strongly bonded to the soil matrix. As a result, pilot- and full-scale applications frequently show reduced efficiency due to the complexities of soil matrices and real-world hydraulic and physicochemical settings [92]. In situ soil washing proves to be more cost-efficient in high-permeability soils, such as sandy soils, with estimated costs ranging between USD 25,000 and USD 110,000 per kilogram of PFOS removed. Conversely, the ex situ approach, while delivering higher removal efficiency, incurs considerably greater costs, up to approximately USD 111,000 per kilogram of PFOS removed, partly due to increased water consumption [64].

3.2.2. Phytoremediation

This flowchart in Figure 8 outlines the phytoremediation process. Due to the knowledge gap about the most promising plants for PFAS phytoremediation, Huff et al. [93] comparatively tested eight herbaceous and seven woody plant species grown in sandy soil with six PFAS compounds introduced via irrigation water (1 mg/L for each compound). PFAS accumulation varied by species and compound. In fact, PFAS concentrations were lower in wood. However, biomass growing with time may significantly contribute to total accumulation over longer treatment periods. Festuca rubra (herbaceous species) was found to hyperaccumulate all the target PFAS having a bioaccumulation concentration factor (BCF) > 10, whereas Schedonorus arundinaceus (herbaceous species) and Betula nigra (woody species) showed BCF > 10 for some PFAS compounds.

Similarly, He et al. [66] conducted a greenhouse study in which seven local weed plant species were chosen to investigate their mechanism of uptake of PFAS. Phyllanthus urinaria (Pu), Aster indicus (Ai), Justicia procumbens (Jp), Alternanthera philoxeroides (Ap), Imperata cylindrica (Ic), Juncus effuses (Je), and Setaria viridis (Sv) were used as extraction plants and the soil was spiked with 11 target PFAS. Authors determined the mass distribution of the target PFAS after 30 days of treatment. A result showed that 0.04–41.4% of PFAF were extracted from the soil and bioaccumulated in the harvestable compartments. Specifically, Je exhibited the highest mass accumulation considering all the PFAS, with 39.2 and 16.8 μg accumulated in shoot and root, respectively. This was mainly due to Je’s high biomass (i.e., 3.85 and 1.65 g_DM of shoot and root, respectively), possibly indicating that greater plant biomass is desirable for phytoextraction. Additionally, bioconcentration factors of the shoot and root (BCF_shoot, BCF_root, respectively) and translocation factors were calculated as key indicators for evaluating the phytoremediation potential of plants. Particularly, BCF_shoot and translocation factor showed a negative correlation with PFAS molecular size and Kow as expected since the plant uptake of short-chain PFAS was greater than that of long-chain homologues (e.g., >98% of the mass of PFOS remained in the soil). With regard to plant properties, both BCF_root and BCF_shoot were positively correlated with the protein and lipid content in the roots, whereas BCF_root resulted as positively correlated with root length also.

Nassazzi et al. [94] evaluated the PFAS phytoremediation potential of sunflower (Helianthus annuus), mustard (Brassica juncea), and industrial hemp (Cannabis sativa) in a greenhouse experiment. The study used organic potting soil spiked with 14 PFAS compounds at a concentration of 1.5 mg/kg each. Moreover, the effects of inorganic fertilizer and microbial mixture as supplements were tested. Authors determined the BCFs for PFAS in leaves, stems, and roots across the different plant species. BCF values varied considerably: up to 2671 in leaves, 656 in stems, and 42 in roots, with the highest accumulation typically occurring in leaves. Carboxylic PFAS showed greater accumulation than sulfonic ones, and BCFs generally decreased with increasing perfluorocarbon chain length. Short-chain PFAS exhibited higher mobility and were more broadly distributed within plant tissues. With regard to supplements, neither fertilizer nor microorganisms had a significant effect in terms of improving PFAS accumulation.

Nason et al. [95] studied several hemp varieties (such as H51, Hlesia, Hlianato, and ChinMa) to remove PFAS from soil in an AFFF-contaminated area. Ten PFAS compounds were detected in the plants, with higher bioaccumulation factors in leaves compared to stems. Bioaccumulation decreased with increasing C–F chain length and was higher for carboxylic acids than for sulfonic acids, as previously observed and reported. The ChinMa variety accounted for 75% of the total PFAS uptake despite covering only 25% of the cultivated area. In total, about 1.4 mg of PFAS were removed, mostly in the leaves (85%), with PFPeA being the most accumulated compound (0.79 mg, 56% of the total). In order to complete PFAS removal, authors addressed the issue related to the disposal of the harvested hemp tissue through hydrothermal liquefaction (HTL). This treatment effectively degraded most carboxylic PFAS (>99%), especially with the addition of Ca(OH)₂ or KOH, which enhanced defluorination and the degradation of 6:2 FTS. However, degradation of sulfonic PFAS was limited and, in some cases, an increase in PFOS was observed, likely due to precursor transformation.

Finally, some authors studied the uptake of PFAS in edible plants with the aim of assessing food safety but also to search for potential PFAS-accumulating species. For instance, Stahl et al. [96] assessed the transfer of PFAS from soil to crops, including wheat, oats, maize, potatoes, and perennial ryegrass under controlled conditions exposed to PFOA/PFOS concentrations from 0.25 to 50 mg/kg. Across all crops, the results consistently showed that PFAS accumulated predominantly in the non-edible parts of the plants, such as straw, husks, and peels, rather than in the grains or tubers. Blaine et al. [97] examined the uptake of PFAS in lettuce and strawberry. For strawberry, the mass distribution among tissues for each PFAS showed that more than 65% of PFBA and PFPeA was in the fruit. However, more than 70% of the long-chain PFAS accumulated in the root, whereas the shoot compartment showed the lowest total accumulation. Phytoremediation represents a potentially sustainable, low-cost, and simple option for PFAS remediation, with a total estimated cost of USD 7.54/m2 when treating heavy metal-contaminated soil [98]. Regarding PFAS remediation,

Table 4. Comparative table of studies on PFAS-contaminated soil treated by phytoremediation.

Soil Type	PFAS (C-mg/kg)	Experimental Conditions	Results (%)	Ref.
-	Σ11 PFAS (C: 16.65–297.52)	Greenhouse: Juncus effusus	39.2 μg/kg (Shoots) 16.8 μg/kg (Roots)	[66]
-	Σ6 PFAS	Greenhouse: Eight herbaceous and seven woody plant species	1–38,121 μg/kg (PFAS, Herbaceous) 41–35,975 μg/kg (PFAS, Foliage woody’s plant)	[93]
-	Σ14 PFAS (C: 1500, each)	Greenhouse: Sunflower, mustard, and hemp	3–11,000 μg/kg (Sunflowers) 5–14,000 μg/kg (Mustards) 70–960 μg/kg (Hemp)	[94]
-	Σ20 PFAS (C: 0.75–22)	Field-scale: H51, Hlesia, Hlianato, and ChinMa	1.4 mg (PFAS, Leaves)	[95]
Silty	PFOA/PFOS (C:250–50,000)	Pilot-scale: Corn, oat, potatoes, wheat, and perennial ryegrass cultivation	15,500 (PFOA)–7900 μg/kg (PFOS) (Corn, Straw) 440 μg/kg (PFOA, Corn, Ears) 217,000 (PFOA)–41,400 μg/kg (PFOS) (Oat, Straw) 1480 μg/kg (PFOA, Oat, Grains) 52 (PFOA)–34 μg/kg (PFOS) (Potatoes, Tuber)	[96]
Sand	Σ9 PFAS (C: 3.3–25,100)	Greenhouse: Lettuce cultivation (0.4% OC)	25,000 μg/kg (PFBA, Leaves)	[97]
		Greenhouse: Strawberry cultivation (0.4% OC)	>10,000 μg/kg (PFBA and PFPeA, Fruits) 5450 μg/kg (PFHxA, Roots)

summarizes the aforementioned studies.

Nonetheless, this approach exhibits considerable limitations in terms of both efficiency and scalability. In fact, phytoremediation appears to be less effective for long-chain and sulfonic PFAS due to their low mobility and poor plant uptake. Species-specific variability and the generally slow rate of contaminant removal further constrain its large-scale application. Moreover, the safe disposal of PFAS-laden biomass is a major challenge, as incomplete degradation may lead to secondary pollution. Future research should focus on identifying high-biomass, high-accumulation species, enhancing uptake through genetic or agronomic strategies, and developing efficient, sustainable biomass treatment methods.

3.2.3. Electrokinetic (EK) Remediation

The flowchart in Figure 9 outlines an electrokinetic remediation process. A summary of the studies mentioned below is presented in

Table 5. Comparative table of studies on PFAS-contaminated soil treated by electrokinetic remediation.

Soil Type	PFAS (C-mg/kg)	Experimental Conditions	Results (%)	Ref.
Kaolin clay	PFOA (100,000)	Bench-scale + SDS, NaC or TW80 (5%) (I: 10–20 mA; t: 7–14 days)	14.5–75.7 (PFOA)	[69]
Loamy sand	PFOA and PFOS (10,000)	Bench-scale (∆V: 30 V; t: 14 days; ∆CC: 5–25%)	75–85 (PFOA) 20–30 (PFOS)	[99]
		Bench-scale (∆V: 30 V; t: 14 days; ∆CC: 50–75%)	95–100 (PFOA) 60–80 (PFOS)
Loamy sand	PFOA (10,000)	Bench-scale (∆V: 30 V; t: 14 days; ∆OM: 5–75%)	70–80 (PFOA) >60 (PFOS)	[100]
Kaolin clay	PFOA (10,000)	Bench-scale + NaC + PRB (iron slag/AC) (P: middle; t: 2–3 weeks)	33 (No PRB) 75 (PRB 50/50) 79 (PRB 70/30) 63.9 (PRB 50/50)	[101]
Kaolin clay	PFOA (10,000)	Bench-scale + NaC + PRB (AC, FeAC) (P: middle, anode; t: 14 days)	59.55 (PRB: FeAC; P: middle) 52.35 (PRB: AC; P: middle) 40.37 (PRB: ReAC; P: middle) 21.96 (PRB: FeAC; P: anode) 20.62 (PRB: ReFeAC; P: middle)	[102]

∆V: voltage; t: time; ∆CC: variation clay content; ∆OM: variation organic matter; SDS: sodium dodecyl sulphate; TW80: Tween80; NaC: sodium cholate; AC: activated carbon; P: position; I: intensity current; FeAC: iron-loaded AC; ReAC: regenerated activated carbon; ReFeAC: regenerated iron-loaded AC.

, providing a clear overview for comparison and analysis.

Abou-Khalil et al. [99] investigated the efficacy of EK remediation for the removal of PFOA and PFOS from soils with different clay contents. Authors observed that the removal efficiency of PFAS (especially PFOS) improved when increasing the clay content of the soil samples through artificial kaolinite. Specifically, after 14 days of treatment, PFOA removal efficiency increased from 75 to almost 100% for clay contents of 5–50 and 75%, respectively. However, PFOS was removed by ~25% in 25% clay soil and increased to >75% in 75% clay soil. Similarly, Dhulia et al. [100] examined the effect of the different OM contents in soils on the removal rate of spiked PFOA and PFOS by lab-scale EK. After 15 days of treatment, PFOA removal efficiency was found to be 70–80% for all the tested OM contents. Removal of PFOS varied with OM contents increasing from 25 to >60% at 5 and 50% of OM contents, respectively, and decreasing to 50 at 75% of OM. Those studies were able to highlight the crucial role of soil composition and properties (e.g., clay and OM contents) in influencing EK remediation for PFAS contamination. However, the use of PFAS-spiked soil at laboratory scale limits the validity of the results, since target PFAS are more prone to be mobilized compared to historically contaminated soils.

In order to optimize PFAS mobilization and removal through the EK process, Ganbat et al. [69] tested anionic and non-ionic surfactants in a PFAS-spiked kaolin clay model soil (namely, without interferences from (in)organic matter). After the experiment duration of a week, PFOA removal efficiencies in the unenhanced EK treatment were 14.5 and 31.0% at electrical currents of 10 and 20 mA, respectively. In comparison, in the enhanced EK processes conducted at 10 mA, PFOA removal was 13.8, 17.7, and 7.7% when sodium dodecyl sulfate (SDS), sodium cholate (NaC), and non-ionic surfactant (TW80), respectively, were added as enhancers. Moreover, when increasing the current up to 20 mA, PFOA removal was 15.7, 32.7, and 7.7%. Authors demonstrated the possibility of improving PFOA removal depending on the surfactant type, with NaC biosurfactant exhibiting the highest removal efficiency. Subsequently, authors coupled the enhanced EK treatment with the adsorption of migrating PFOA onto an iron slag/AC permeable reactive barrier (PRB). PFOA removal efficiency was observed to increase up to 75% when 50:50 iron slag/AC PRB was introduced in the soil middle section, whereas the highest PFOA removal was 94% when the duration of EK + PRB (slag/AC 70:30) treatment was extended from 2 to 3 weeks. To improve the sustainability of the combined technique, the possibility of reusing PRB after regeneration via solvent (methanol) washing was evaluated. This process showed that the performance of the system in terms of PFOA removal decreased with the regenerated PRB. Finally, authors observed that the slag content in the PRB contributed to increase the average electric current and therefore the electric consumption [101]. Ganbat et al. [102] further tested an iron-coated AC (FeAC) PRB and observed that varying the position of PRB within the soil affected the efficiency of the overall treatment. In fact, the EK remediation with the FeAC PRB in the middle section of the system resulted in the highest PFOA removal efficiency of 59.6%. However, when the FeAC PRB was located close to the anode, the removal efficiency of PFOA dropped down to 22.0%.

However, besides the aforementioned limitations related to artificial PFAS contamination, results coming from those studies are limited by the absence in the model soil of naturally occurring (in)organic compound. These compounds could possibly interfere with PFAS mobilization through EK treatment and adsorption onto PBR. Therefore, further research is needed to prove the scalability of the EK application.

EK remediation faces significant challenges due to a combination of technical, economic, and environmental limitations. Its effectiveness is highly dependent on soil composition, in particular, the presence of clay and organic matter, with studies reporting contrasting results. Most research to date has been conducted under laboratory conditions using artificially contaminated soil and a limited range of PFAS compounds, primarily PFOS and PFOA. This reduces the applicability of the findings to complex, historically contaminated sites. Enhancing removal efficiency through the integration of surfactants and PRBs demonstrated potential. However, these modifications significantly increase both energy consumption and system complexity as well as the need for regular PRB maintenance, which should be taken into account. Operational costs, inferred from analogous studies on heavy metal remediation, range from USD 110 to over 1000/m³ of treated soil, underscoring the significant challenges associated with electricity demands and reagent costs. Moreover, the post-treatment management of contaminated effluents represents an additional operational burden [103].

3.3. PFAS Destruction Technologies

Whilst (im)mobilization techniques aim at influencing the mobility of PFAS, other strategies for the remediation of PFAS-contaminated soils are those based on contaminant degradation. Biological treatments rely on biochemical reactions driven by microorganisms, primarily bacteria, which are naturally present in soil and water. These biological agents can degrade various organic compounds and, under favorable environmental conditions, may contribute to the transformation or breakdown of PFAS. Thermal treatments concern the application of high temperatures to degrade or volatilize contaminants present in the soil. Depending on the temperature range and operational conditions, thermal processes can enable the thermal decomposition of PFAS compounds or facilitate their separation from the solid matrix. This category includes technologies such as incineration, pyrolysis, thermal desorption, and other controlled heating systems. These technologies can be often integrated with gas treatment units to prevent contaminants release into the atmosphere. Thermal treatments offer high efficacy in PFAS removal but can entail significant energy costs and require careful management of gaseous emissions. Mechanical treatments, such as ball milling, are based on the application of high-energy mechanical forces to promote the degradation of contaminants. In ball milling, soil is subjected to intense grinding using rotating balls inside a mill, generating mechanical stress and localized heat. This process can break chemical bonds within PFAS molecules, increase their reactivity, or enhance interaction with other reagents. The degradation often occurs through defluorination, and the use of co-grinding agents like KOH or metal oxides can significantly improve the reaction efficiency.

3.3.1. Bioremediation

The flowchart in Figure 10 illustrates the bioremediation process.

Table 6. Comparative table of studies on PFAS-contaminated soil treated by bench-scale bioremediation.

Soil Type	PFAS (C-mg/kg)	Experimental Conditions	Results (%)	Ref.
-	PFOS and PFOA	Aerobic treatment + A6 (t: 100 days)	35 (PFOS) 63 (PFOA)	[70]
Silty sand	Σ15 PFAS (C: 1.2–59)	Aerobic treatment + CB and YM (t: 28 days)	46 (PFOS, CB) 16 (PFOA, CB) 69 (PFOS, YM) 36 (PFOA, YM)	[104]
-	PFOS	Aerobic treatment + HJ4 (t: 48 h)	67	[105]
Silty sand	PFOA	Aerobic treatment + PP (t: 72 h)	32.4 (PFOA 500 mg/L) 12.6 (PFOA 1000 mg/L)	[106]
		Aerobic treatment + PP (t: 72 h) + glucose	48.1
-	PFOS	Aerobic treatment + PP 2.4-D (t: 6 days)	75	[107]
Silty sand	PFOA (C: 500)	EOH + laccase (PO and PS; d: 60 U) + mediator (soybean-meal)	35 (PO) 40 (PS)	[108]
		EOH + laccase (PO and PS; d: 20/4 wk) + mediator (soybean-meal)	29 (PO) 35 (PS)
Silt loam	PFOA	Anaerobic treatment + mmc (substrates: acetate, lactate, ethanol; t: 259 days)	~0	[109]
-	PFOS	MD (aerobic and anaerobic; t: 177 weeks)	0	[110]

C: concentrations; t: time; mmc: mixed microbial communities; EOH: enzymatic oxidative humification; PO: Pleurotus ostreatus; PS: Pycnoporus sp SYBC-L3; d: dosage; 20/4 wk: 20 U of laccase added every 4 weeks; CB: chemoorganoheterotrophic bacteria; YM: yeasts/molds; HJ4: Pseudomonas aeruginosa; A6: Acidimicrobium sp. A6; PP: Pseudomonas parafulva, MD: microbial dehalogenation.

summarizes the key studies analyzed below.

Beškoski et al. [104] investigated the biotransformation of PFOS and PFOA, under aerobic conditions, by chemoorganoheterotrophic bacteria (CB) and yeast and molds (YM) isolated from river sediment taken from PFOS- and PFOA-polluted areas. After 28 days, concentrations of PFOS and PFOA decreased by 46 and 69% for PFOS and by 16 and 36% for PFOA in samples incubated with CB and YM, respectively. However, despite such reduction, no defluorinated PFOS or PFOA was detected, indicating potential limitations for full PFAS mineralization. Nevertheless, that study highlighted the potentiality of microbial communities isolated from PFAS-contaminated environments to conduct biosorption of these contaminants. Previously, Kwon et al. [105] achieved a comparable result showing that 67% of PFOS was degraded after 48 h by Pseudomonas aeruginosa strain HJ4 isolated from samples of sewage sludge and soil without defluorination but producing PFBS and PFHxS as minor products [106]. Similarly, Yi et al. observed that when exposing Pseudomonas parafulva to 500 and 1000 mg/L of PFOA, its concentration decreased by 32.4 and 12.6%, respectively, after 72 hours of incubation. In contrast, Chetverikov et al. [107] and Huang and Jaffé [70] examined the degradation of PFOS and PFOA by Pseudomonas plecoglossicida 2.4-D and Acidimicrobium sp. strain A6, respectively, leading to defluorination and the formation of minor products. The former study achieved a 75% degradation of PFOS after 6 days of treatment, producing PFHpA and releasing fluoride ions. The latter used an autotrophic microorganism responsible for the anaerobic oxidation of ammonium under iron-reducing conditions (Feammox process). Authors observed decreases in PFOS and PFOA concentrations by 35 and 63%, respectively. This occurred with the release of fluoride ions and the production of shorter-chain compounds during degradation, including PFBA, PFPeA, PFHxA, and PFHpA from PFOA, and PFBS and PFBA from PFOS. According to authors, such results suggest that under the soil conditions of acidic pH and iron-rich environment, the Feammox process may be stimulated to achieve PFAS biodegradation in contaminated sediments and groundwater systems.

Luo et al. [108] examined PFOA degradation by enzyme-catalyzed oxidative humification reactions (ECOHRs) utilizing enzymes, such as laccase Pleurotus ostreatus (PO) and Pycnoporus sp. (PS), and natural organic materials as mediators, namely soybean meal (SBM). The laccase PO and PS were tested and dosed in PFOA-spiked soil samples using two different modes. The first mode involved adding 20 U enzyme/g soil every four weeks during incubation. The second mode consisted of one single addition of 60 U enzyme/g soil at the start of incubation. In the former trial, PFOA removal efficiency reached 29 and 35% for PO and PS, respectively. However, in the second one, PFOA removal was 35 and 40% for PO and PS, respectively. As PFOA degradation products by ECOHRs, any short-chain PFAS were detected but partially fluorinated organic compounds.

Liou et al. [109] have investigated the biodegradation of PFOA under anaerobic conditions using five different microbial communities, isolated from a municipal wastewater treatment plant, an industrial site sediment, an agricultural soil and soils from two fire training areas, and organic substrate such as acetate, lactate, and ethanol. However, after 259 days of incubation, the results have shown just a slight PFOA degradation and low defluorination process.

Later, Ochoa-Herrera et al. [110] also observed the recalcitrance of PFAS to the anaerobic biodegradation. In fact, even after an incubation time up to 177 weeks with different inocula (including sediments, anaerobic granular sludge, anaerobically digested sewage sludge, and activated sludge), all the tested PFAS compounds were found to be highly resistant to microbial degradation.

Finally, in a recent study, Harris et al. [111] investigated the use of various microorganisms isolated from PFAS-contaminated soils. Among others, Delftia acidovorans demonstrated good survival in media contaminated with PFOA. The authors assessed PFAS degradation by measuring the release of fluorine ion (F-) upon exposure to PFOA, showing promising results. Two dehalogenase enzymes, DeHa I and II, were identified in D. acidovorans as capable of binding to and degrading PFOA.

To sum up, it can be noted that PFAS bioremediation is hard to conduct even under controlled laboratory conditions and pure cultures of isolated microorganisms. In fact, while certain bacterial and fungal species have demonstrated a partial ability to reduce concentrations of PFOA and PFOS, the complete mineralization of these substances remains generally limited. For this reason, its application in real-world scenarios of PFAS soil contamination is still far from being a suitable alternative to other PFAS remediation technology.

3.3.2. Thermal Treatment

Figure 11 illustrates the thermal treatment ex situ.

Table 7. Comparative table of studies on PFAS-contaminated soil treated by thermal treatment.

Soil Type	PFAS (C-μg/kg)	Treatment (Conditions)	Degradation (%)	Ref.
Clay loam	Σ18 PAFS	Bench-scale: Thermal desorption (air; N2; T: 125–500 °C; t: 30 min)	> 99 (PFAS; T: 400 °C) 60–71 (PFSA; T: 400 °C) 99.8 (HFPO-DA; T: 300 °C)	[72]
Kaolinite clay			> 99 (all PFAS; T: >400 °C) 60–71 (PFSA; T: 400 °C) 99.8 (HFPO-DA; T: 300 °C)
Calcareous soil	Σ14 PAFS (700)	Bench-scale: Thermal desorption (T: 150, 550 °C)	29-36.3 (all PFAS; T: 150°C) 96.4–97.5 (all PFAS; T: 550°C) 97.27 (PFBS; T: 550°C) 32.75 (PFNA; T:150°C)	[73]
-	Σ6 PAFS	Bench-scale: Thermal desorption (T: 100-400 °C; t: 0-600 s)	100 (PFAS; T: 300 °C; t > 300 s) ~99 (PFOS; T: 400 °C; t: 60 s) 23 (PFAS; T: 100 °C, t: 600 s)	[112]
Clay	Σ9 PAFS (4000)	Bench-scale: Thermal desorption (T: 350, 450, 550 °C, t: 15–75 min)	>99 (PFCA/FOSA; T: 350 °C) 51–66 (PFSA; T: 350 °C) >99 (PFAS; T: 450 °C)	[113]
Loamy sand soil			>97 (PFPeA; T: 550 °C) 71–93 (PFCA; T: 550 °C) 99 (all PFSA; T: 450 °C)
Silty Clay	Σ9 PAFS (25)	Bench-scale: Thermal desorption (T: 350, 450, 550 °C, t: 15–75 min)	43 (all PFAS; T: 350 °C)
Sand	Σ6 PAFS (3000–5000)	Bench-scale: Smoldering combustion + GAC (w: 50 g/kg sand; air > 5 cm/s; T > 900 °C)	>98.9 (all PFAS; 16–17% defluorination)	[114]
	Σ6 PAFS (3000–5000)	Bench scale: Smoldering combustion + GAC (w: 15 g/kg sand; air > 5 cm/s; T = 642 °C)	>98.9 (all PFAS; no defluorination)
	Σ3 PAFS (200)	Benche scale: Smoldering combustion + GAC (w: 51 g/kg sand; air > 5 cm/s; T > 1000 °C)	>99.9 (all PFAS; high defluorination)
	PFOS, PFOA (0.6–250)	Field-scale: FH ex situ stockpiles (T: 350, 400 °C)	>95.3 (Average)	[115]

T: temperature; t: time; GAC: granular activated carbon; w: weight, FH: FlexHeater^®.

provides an overview of the analyzed studies on thermal treatment for PFAS remediation.

Al-Amin et al. [112] analyzed the thermal kinetics of PFAS and precursors in real-world contaminated soil using a muffle oven. The tests were conducted at moderate temperatures (100–400 °C) and short time (up to 600 s). They focused on the competition between the release of precursors, as 6:2 FTS, and the degradation of PFOA and PFOS. Authors reported that at 100 °C, the ∑PFAS concentration slightly increased during the early stages of thermal treatment. In particular, the result showed after 10 and 30 s an increase of ~4 and 6%, respectively, compared to the initial level. This may be released from unknown precursors. Up to 300 s of treatment, PFAS concentrations remained relatively stable, likely due to a balance between release and degradation, whereas by 600 s, the total PFAS concentration decreased by ~23%, indicating that degradation became the dominant process over time. When increasing the operating temperature at 300 °C, the concentration of PFAS decreases gradually and with a significant change at 60 s, after which only PFOS is still detected. Finally, at the set process temperature of 400 °C, most of the PFAS compounds were degraded before reaching 30 s and PFOS concentration decreased by about 99%. After 300 and 600 s, no compounds were observed, suggesting a complete conversion to their final products.

Sörengård et al. [113] investigated the thermal desorption of PFAS from soil with the aims of identifying the critical variables of the process such as optimal temperature (150–550 °C) and treatment times (15–75 min). In addition, authors compared the treatment performance for artificially (spiked PFAS) and naturally (AFFF-impacted) contaminated soil. The thermal treatment was conducted in a furnace. As a result, it was observed that high temperatures (>400 °C) were crucial to achieve effective removal, especially for sulfonic PFAS. In fact, at 350 °C, carboxylic PFAS and FOSA were removed more than 99%, whereas PFSA removal efficiencies ranged from 51 (PFHxS) to 66% (PFOS). However, their removal was more than 99% at 450 °C. This outlines that the functional group is an important parameter influencing PFAS thermal desorption. Moreover, the optimal treatment time turned out to be in the range 15–45 min, while longer periods were not necessary. Compared to the artificially contaminated soil, the naturally contaminated one exhibited lower removal efficiencies for all the tested temperatures. This was due to the lower concentration of PFAS and/or the stronger adsorption of PFAS onto soil grains due to the historical contamination and longer contact time than artificial soil. Alinezhad et al. [72] confirmed these findings performing an ex situ thermal treatment of kaolinite clay and clay loam soil samples. The soil sample was heated in air or N₂ (pyrolysis) at several temperatures (125 to 500 °C) for a residence time of 30 min. PFAS were spiked into the soil samples through two different modes that promoted either strong or weak bonding with soil particles. First of all, authors observed that the thermal decomposition efficiency of the target PFAS in soil in the sealed (air) system was similar to that measured in a constant-pressure system under a flow of N₂. This occurred even though performing pyrolysis on solid materials in an inert atmosphere may prevent the generation of dioxins and furans. Secondly, it was reported that carboxylic PFAS remained mostly stable for temperatures lower than 200 °C, then they started to decompose at 200–400 °C. Finally, they were almost completely degraded at >400 °C in 30 min. On the other hand, sulfonic PFAS resulted as more thermally stable, reaching a degradation efficiency of 60–71% in 30 min at 400 °C. The effects of the increasing temperature on PFAS degradation were confirmed also in calcareous soils. However, it was reported that heating PFAS in an oven to 400–500 °C can achieve volatilization but cannot destroy the compounds; instead, it produces shorter-chain PFAS and volatile organic fluorine by-products. Even though some PFAS defluorination can occur at 700 °C, however, 900–1100 °C is likely necessary for a high degree of PFAS destruction and to minimize production of undesired by-products [114,116]. Finally, authors detected various nonpolar products, particularly perfluoroalkenes, which were generated from PFOA and PFOS thermal decomposition. These by-products require further treatment [73]. The issue related to the emission of volatile organofluorine species should be taken into account when performing thermal decomposition of PFAS. In fact, short-chain perfluoroalkanes (such as perfluoromethane and perfluoroethane) are of significant concern as they are greenhouse gases and more thermally stable than PFAS. Furthermore, the release of F radicals may lead to the formation of corrosive hydrogen fluoride representing another technological challenge to thermally treat PFAS-containing materials.

Duchesne et al. [114] tested a smoldering combustion to destroy PFAS from sandy soil with a GAC-like enhancer. This is a flameless oxidation reaction that occurs on the surface of a solid (GAC in that case) or liquid fuel when penetrated by gaseous oxygen and it can be self-sustaining after ignition, meaning no external energy input is needed. With regard to GAC addition, authors observed that >35 g GAC/kg sand resulted in average smoldering temperatures over 900 °C. Therefore, the experiment was performed on a sandy soil enhanced with PFAS-spiked GAC at a dose of 50 g/kg sand and flux air of 5 cm/s to investigate the removal of six PFAS, including PFOA, PFOS, PFHxS, PFHpA, PFBS, and PFNA. The treatment has achieved reduction in all PFAS more than 98.9% with a defluorination efficiency of ~17%. Authors concluded that mixing spent GAC with contaminated soil could be used to enhance destruction of PFAS sorbed to both GAC and soil. They finally suggested the further application of a large ex situ batch treatment system or smaller continuous treatment reactor to perform the smoldering process to be integrated with technologies for the removal of hydrogen fluoride emissions. Albeit this holistic approach needs to be assessed by technical and economic analyses. It would allow us to manage PFAS-laden GAC (typically used in wastewater treatment), remediate PFAS-contaminated soil, and remove gaseous emissions.

Due to the diversity of PFAS compounds and the different thermal treatment and combustion conditions, results from laboratory studies often cannot directly translate to field-scale applications. Furthermore, since most lab experiments involve only one or a few PFAS compounds, the mixture of PFAS contamination in real-world settings needs further investigation [29]. In this context, some companies have conducted field demonstrations of thermal treatment of PFAS-impacted soils. For instance, a thermal conductive heating (TCH) system (FlexHeater^®) was installed into ex situ soil stockpiles of a PFAS-contaminated area. Soil surrounding the heating elements reached a minimum temperature of 350 °C throughout the treatment volume. The predominant PFAS, i.e., PFOS, exhibited a 95.3% average percent reduction from thermal treatment with a post-treatment average concentration of 4.1 μg/kg with the highest levels found at the top of the pile where the soil reached 350 °C. On the other hand, all soil samples heated to 400 °C were below local soil cleanup levels for PFOA and PFOS, namely 1.7 and 3.0 μg/kg, respectively. Soil vapor extraction screens were included in the TCH system to extract from treated soil steam and organic vapors. These were transferred to wet scrubber, condenser, vapor/liquid separator, and heat exchanger units. Finally, condensate and non-condensable air were treated using activated charcoal filtration. As a result of an economic analysis, treatment costs ranging from USD 420.5 to 611.6 per m³ were projected, with lower unit costs for large volumes. Site location (climate and mobilization distance) and local electricity costs were identified as the main cost drivers. From an environmental perspective, authors arose concern about the further use of the soil after the thermal treatment. In fact, treated soil should be augmented with moisture, nutrients, and organics to allow its proper reuse [115].

Thermal treatment methodologies offer a promising yet complex solution for PFAS remediation. These technologies achieve high removal efficiency while raising important concerns regarding energy consumption and environmental safety. Techniques such as thermal desorption and pyrolysis have shown the capacity to effectively decompose PFAS compounds but necessitate substantial energy inputs and present risks associated with the generation of hazardous by-products. Smoldering combustion seems a promising alternative since even PFAS-laden GAC can be used as an enhancer of soil heating if penetrated by gaseous oxygen. However, despite its reported efficacy, the operational complexity of this approach can limit its scalability, so further investigation is required. However, field demonstrations of thermal remediation for PFAS-contaminated soils have been successfully implemented with associated costs of USD ~420–610 per m³ of treated soil [115]. In conclusion, these systems facilitate the treatment of substantial volumes of contaminated soil under controlled conditions. They incorporate advanced emission control mechanisms to mitigate environmental risks. However, this comes at the cost of high operational efforts and complex infrastructure requirements.

3.3.3. Ball Milling

Figure 12 schematically outlines the ball milling process. Regarding PFAS remediation,

Table 8. Comparative table of studies on PFAS-contaminated soil treated by ball milling.

Soil Type	PFAS (C-μg/kg)	Treatment (Conditions)	Degradation (%)	Ref.
Sand (dry)	PFOS and PFOA	Bench-scale: PBM with/without KOH (w: 15 g; t: 15 min–4 h)	70–98 (PFOS, w/o KOH)	[74]
			82–96 (PFOA, KOH)
		Bench-scale: PBM with/without KOH (w: 40 g; t: 15 min–4 h)	90–99 (PFOS, w/o KOH)
		Bench-scale: HBM without/with KOH (V: 25 L, t: 3 h)	94–99 (PFOA, KOH)
Sand (saturated)		Bench-scale: PBM with/without KOH (w: 15 g; t: 15 min–4 h)	29–38 (PFOA, w/o KOH)
		Bench-scale: HBM without/with KOH (V: 25 L, t: 3 h)	74–83 (PFOA, KOH)
		Bench-scale: PBM with/without KOH (w: 40 g; t: 15 min–4 h)	97–99 (PFOS, w/o KOH)
			88–92 (PFOA, KOH)
NSS	PFOS (6.1) and 6:2 FTSA (2.4)	Bench-scale: HBM without/with KOH (V: 1 L, t: 3 h)	19 (w/o KOH), 43 (KOH) (PFOS; t: 3 h)	[75]
			97 (w/o KOH), >88 (KOH) (6:2 FTSA)
Sand	PFOS and 6:2 FTSA (~1.4 each)	Bench-scale: HBM without/with KOH (V: 1 L, t: 3 h)	15 (w/o KOH), 69 (KOH) (PFOS)
		Bench-scale: HBM without/with KOH (V: 25 L, t: 3 h)	67 (w/o KOH), 61 (KOH) (6:2 FTSA)
Clay	PFOS and 6:2 FTSA (~0.6 each)	Bench-scale: HBM without/with KOH (V: 1 L, t: 3 h)	5 (w/o KOH), 81 (KOH) (PFOS)
		Bench-scale: HBM without/with KOH (V: 25 L, t: 3 h)	12 (w/o KOH), 85 (KOH) (6:2 FTSA)
Silica sand NSS Calcite Marble	PFOS (18.3) and PFOA (16.1) PFOS (12.7) and PFOA (18) PFOS (17.8) and PFOA (19) PFOS (20.2) and PFOA (25.5)	Bench-scale: PBM (t: 4 h)	99 (PFOS), 99 (PFOA)	[117]
			95 (PFOS), 96 (PFOA)
			91 (PFOS), 93 (PFOA)
			83 (PFOS), 92 (PFOA)
NSS	6:2 FTS (~3.5), PFOS and PFOA (~5)	Pilot-scale: HBM without/with KOH (V: 267 L; t: 10–120 min)	~0 (6:2 FTS), 69 (PFOS), 70 (PFOA) (w/o KOH)	[118]
			97 (6:2 FTS), 70 (PFOS), 74 (PFOA) (KOH)
Silty sand	6:2 FTS, PFOS and 8:2 FTS (2.6–5.6)	Pilot-scale: HBM without/with KOH (V: 267 L; t: 10–120 min)	46 (6:2 FTS), 50 (PFOS), 55 (PFOA) (w/o KOH)
			46 (6:2 FTS), 31 (PFOS), 33 (PFOA) (KOH)

NSS: nepheline syenite sand; HBM: horizontal ball milling; w/o: without; PBM: planetary ball milling, w: weight; V: volume; t: time.

summarizes the relevant studies found in literature.

Turner et al. [74] tested the destruction of PFAS using planetary ball milling (PBM) and studying the effect of sand mass, deionized water saturation, and KOH supplementation. An amount of 15 g of PFOS- and PFOA-spiked sand amended with 10 g of KOH was subjected to PBM treatment. The result showed that PFOS and PFOA concentrations were reduced by 73 and 90% and 81 and 96%, respectively, after 1 and 4 h. For both compounds, the least degradation efficiencies were recorded for saturated sand. This indicates that water saturation was a significant hindrance on the mechanochemical destruction of PFOS and PFOA. PBM was further conducted on soil samples (40 g amended with 10 g of KOH) from a firefighting training area. After 6 h of treatment at 275 rpm, PFOS concentrations were observed to decrease up to 96%. Afterwards, Turner et al. [117] performed the same experiment on different media, such as silica sand (SS), nepheline syenite sand (NSS), calcite, and marble. These media were sieved to achieve similar initial particle sizes to investigate the effect of grinding on process performance. The study showed a reduction in the concentrations of PFOA and PFOS in all four media. Particle size reduction was observed to be positively correlated to PFOA and PFOS destruction in silica sand and NSS. Furthermore, silicates resulted as more effective than carbonates in the destruction of PFAS, achieving a 95% removal in 4 h for SS and NSS, whereas in carbonates the degradation was lower, namely 91 (PFOS) and 93% (PFOA) in calcite, and 83 and 92%, respectively, in marble. In carbonate soil samples, an agglomeration phenomenon (caking) after 1.5 h of grinding was observed, compromising the effectiveness of the process. Short-chain carboxylic PFAS, with PFHpA as the main compound, were identified among the degradation by-products.

Even though the aforementioned studies are promising and some of them are derived from real-world contaminated samples, the small size of the experimental setup makes the development of PBM technology still at an early stage.

Battye et al. [75] preliminarily investigated the viability and effectiveness of horizontal ball milling (HBM) for PFAS destruction. It is an already existing technology at commercial/industrial sizes from the mining, metallurgic, and agricultural industries. Authors used NSS, spiked separately with PFOS, 6:2 FTSA, and a mixed formulation AFFF, and AFFF-impacted sandy and clay soils. Each sample was treated in either 1 or 25 L cylinders to simulate real-world applications. PFOS degradation in NSS reached 19 and 43% without and with KOH, used as a co-milling reagent, in 3 h, whereas 6:2 FTSA degraded to 97% with KOH, producing PFPeA as a by-product. In trials with AFFF-impacted soil samples, clay exhibited PFOS degradation of 5 and 81% without and with KOH, respectively, whereas in sand, 15 and 69%, respectively. When scaling the HBM to 25 L, clay soil maintained its effectiveness (i.e., PFOS degradation of 85% with KOH), demonstrating the industrial potential of HBM. Afterwards, Battey et al. [118] aimed at evaluating PFAS remediation by HBM at an industrial scale with a cylinder of 267 L in a perspective of future on-site applications. After the same sample preparation of the previous study, authors observed that in spiked NSS, 6:2 FTS degradation reached 89 and 97% after 10 and 120 min in the trials with KOH, whereas it resulted as negligible without KOH. As in the previous study, PFPeA was identified as a by-product, while no other carboxylic PFAS were identified. With regard to PFOS, the same degradation was recorded in tests with and without KOH addition (i.e., 69–70%), contrarily to previous findings. When testing historically (60 years) AFFF-impacted soils, more PFAS destruction was surprisingly achieved without KOH addition (i.e., degradation efficiencies of 31, 17, and 33% for PFOS, 6:2 FTS, and 8:2 FTS, respectively, with KOH, while 50, 46, and 55%, respectively, without KOH). Authors speculated that in the presence of KOH, the non-target PFAS contained in real-world contaminated soil were converted to PFOS, 6:2 FTS, and 8:2 FTS. For this reason, free-fluoride analysis was performed, and, as a result, extractable fluorine was observed to be larger in KOH-supplemented HBM (i.e., 6.0 mg/kg versus 3.4 mg/kg). This proved that more PFAS was indeed degraded in the presence of KOH, as previously observed.

Challenges include soil heterogeneity, complexity of the PFAS mixture, supplementation of chemical reagents, and management of the residual alkalinity, but results are open to large-scale sustainable remediation.

Ball milling has gained significant attention as a promising method for PFAS degradation. This technology offers high removal efficiency without requiring elevated temperatures or generating associated emissions. The reported studies highlight the role of soil type and chemical additives in optimizing PFAS destruction through ball milling. This sets a foundation for evaluating the sustainability of a scaled-up ball milling process, especially if HBM was successfully applied to large volumes of real-world contaminated soils, maintaining high degradation efficiencies. Despite some challenges, such as soil variability, the complexity of contaminants, and the handling of residual materials, ball milling is already widely used in industries like mining. This highlights its potential for broader environmental applications. Economically, operational costs for this method have been estimated at around USD 76/m³ of treated soil, with a processing capacity of roughly 19 tons per hour, according to Cagnetta et al. [119]. These factors position ball milling, particularly HBM, as one of the most viable and cost-effective solutions for large-scale PFAS remediation. However, challenges include soil heterogeneity, complexity of the PFAS mixture, supplementation of chemical reagents, and management of the residual alkalinity prior to industrial implementation.

4. PFAS Soil Remediation Technologies: Effectiveness, Sustainability, and Future Directions

Sustainable cleanup of PFAS-contaminated soil is one of the most challenging issues in the field of environmental remediation. PFAS remediation technologies must be evaluated based on two primary criteria: effectiveness and sustainability, which include high PFAS removal efficiency but minimize ecological impact and optimize energy use. Figure 13 summarizes the positioning of various technologies within a two-dimensional framework based on effectiveness and sustainability, according to the results and the critical discussion reported in this paper.

The technologies are grouped into three categories: immobilization, mobilization and extraction, and destruction. Immobilization techniques, such as sorption with BC, commercial sorbents (e.g., RB), and AC, and S/S methods, predominantly stay in the higher-left quadrant of the framework. This suggests these methods generally achieve moderate effectiveness with varying degrees of acceptable sustainability. These approaches limit PFAS mobility within soils, though their long-term sustainability requires careful further examination. In fact, despite satisfied results found in the literature, the energy demands associated with regenerating or disposing of spent sorbents, coupled with secondary waste generation, reduce the overall sustainability of this method. In this regard, BC, derived from waste biomass, offers a lower-energy alternative. Similarly, S/S methods show strong immobilization potential but can alter soil structure and mechanical properties, thereby raising concerns about the usability of treated sites in the long term.

Mobilization technologies tend to appear closer to the center of the graph, indicating a balance between effectiveness and sustainability. These approaches aim to extract PFAS from soil but are often challenged by managing secondary by-products. Solvent-based soil washing, for example, generates effluent streams that require decontamination and solvent recovery. Alternatively, gas-enhanced techniques employing ozone avoid solvent use and achieve removal efficiencies exceeding 70%. However, ozone generation is energy-intensive and sensitive to soil properties. This limits wider applicability. Finally, EK remediation, to date, faces the challenge of considerable high cost and strong dependence on soil composition. Destructive technologies can be divided into biological and thermal/mechanical ones. The former (using either aerobic or anaerobic microorganisms) seem unsuitable for soil remediation by PFAS contamination in terms of effectiveness. However, their environmental impacts are limited. The latter focusing on breaking down PFAS compounds using thermal and mechanical processes are in the right part of the framework, highlighting their high effectiveness. Thermal treatments like desorption and pyrolysis achieve significant PFAS degradation when soil is heated up to high temperatures (>400 °C). However, considerable consumption of energy and treatment of gaseous emissions limit their sustainability. In this regard, smoldering combustion using even spent GAC can represent a lower impactful strategy. It was demonstrated to only partially convert PFAS (16–17%) and its operational complexity makes large-scale deployment challenging. On the other hand, ball milling, either planetary or horizontal, can degrade up to 97% of PFAS at ambient temperatures without generating emissions. However, PBM has significant scalability limitations, requiring further technological advancements for widespread adoption.

Based on the reviewed studies, sorption technologies are still primarily at the bench scale, and further research is needed at pilot-scale and full-scale, particularly in historically contaminated soils. S/S, ball milling, and thermal treatment emerge as the most promising options for large-scale applications, as they are already supported by successful industrial and full-scale results. In particular, thermal treatment has demonstrated consistent performance in field conditions, proving it to be a mature and readily applicable technology. Soil washing/flushing has been implemented at full scale, but efficiency often declines when transitioning from controlled laboratory conditions to real-world field applications. EK remediation, although promising in laboratory settings, requires additional pilot- and field-scale validation to assess whether its technical and economic scalability is feasible, considering its complexity and potential cost escalation. Phytoremediation studies remain limited to small- or medium-scale experiments. Full-scale trials are needed to evaluate their practicality and cost-effectiveness in real-world scenarios. Bioremediation, while showing some potential in lab conditions, appears less viable at larger scales due to limited effectiveness and the complexity of PFAS mixtures typically found in contaminated sites.

The body of literature analyzed in this paper indicates that no single, universally applicable solution currently exists. The significant heterogeneity of contaminated soils, coupled with the complexity of PFAS mixtures, often originating from historical sources, requires a multidisciplinary approach. For this reason, future research should prioritize pilot- and full-scale investigations that reflect real-world site conditions. Simultaneously, the evaluation of environmental impacts through life cycle assessment and financial feasibility through techno-economic analysis or life cycle cost analysis are key methodologies. These should be included in studies aimed at providing scalable and long-term technologies for remediation of PFAS-contaminated soils. Technologies such as HBM, some S/S approaches, and the use of sustainable sorbents appear to offer a favorable balance between performance and feasibility, though further validation is needed. In addition, hybrid or combined treatment strategies could enhance removal efficiencies when treating complex matrices. Finally, increased attention must be directed toward the management of secondary residues to ensure long-term safety. This involves focusing not only on improving treatment efficacy but also on gaining deeper insights into post-treatment behavior and system dynamics.

5. Conclusions

Per- and polyfluoroalkyl substances (PFAS) continue to pose significant environmental issues due to their chemical stability and resistance to natural degradation, along with their ability to move through the subsurface. This persistence makes PFAS contamination particularly difficult to remediate and poses a major challenge in developing effective and sustainable remediation strategies.

This review has explored a variety of remediation technologies, grouped into immobilization, mobilization and extraction, and destructive treatments, with the aim to assess their applicability and long-term sustainability to remediate PFAS-contaminated soils.

What emerged is that immobilization techniques allow them to confine PFAS into a solid matrix, but they face the challenges of managing spent sorbents (through regeneration or disposal), and/or monitoring the long-term performance of the treatment and how it can affect usability of the treated soil. Mobilization techniques, such as soil washing and electrokinetic remediation, showed good potential for PFAS extraction. However, both require adequate treatment of the PFAS-laden wastewater, and especially the latter has not been tested at pilot- and full-scale. Phytoremediation and bioremediation would represent the most sustainable and cost-effective techniques of remediation. However, the former demonstrated good PFAS removal performances only for low and restricted PFAS contamination, whereas the latter has been tested only in controlled bench-scale experiments and artificial PFAS contamination conditions. Finally, PFAS destruction techniques currently represent the most promising choices to achieve higher and long-term PFAS removal efficiency. Nevertheless, their applications should be optimized by reducing energy consumption and potential by-products formation (thermal treatment) and conducting more full-scale investigations (mechanical treatment).

In this context, future research should (i) investigate remediation technologies under more realistic experimental conditions (up to field demonstrations) encompassing soil heterogeneity, historical and complex PFAS contamination, and (ii) provide the assessment of the environmental impacts of the tested strategies and their financial feasibility through techno-economic analysis or life cycle cost, with the ultimate aim of developing scalable, low-energy solutions for practical implementation.