Review of Per- and Poly-fluoroalkyl Treatment in Combustion-Based Thermal Waste Systems in the United States

Abstract

Per- and poly-fluoroalkyl substances (PFAS) are a class of synthetic chemicals known for their widespread presence and environmental persistence. Carbon-fluorine (C-F) bonds are major components among PFAS and among the strongest organic bonds, thus destroying PFAS may present significant challenge. Thermal treatment such as incineration is an effective and approved method for destroying many halogenated organic chemicals. Here, we present the results of existing studies and testing at combustion-based thermal treatment facilities and summarize what is known regarding PFAS destruction and mineralization at such units. Available results suggest the temperature and residence times reached by some thermal treatment systems are generally favorable to the destruction of PFAS, but the possibility for PFAS or fluorinated organic byproducts to escape destruction and adequate mineralization and be released into the air cannot be ruled out. Few studies have been conducted at full-scale operating facilities, and none to date have attempted to characterize possible fluorinated organic products of incomplete combustion (PICs). Further, the ability of existing air pollution control(APC (APC) systems, designed primarily for particulate and acid gas control, to reduce PFAS air emissions has not been determined. These data gaps remain primarily due to the previous lack of available methods to characterize PFAS destruction and PIC concentrations in facility air emissions. However, newly developed stack testing methods offer an improved understanding of the extent to which thermal waste treatment technologies successfully destroy and mineralize PFAS in these waste streams.

1. INTRODUCTION

Per- and poly-fluoroalkyl substances (PFAS) are a synthetically derived chemical class with varying degrees of fluorinated aliphatic and aromatic moieties (OCED, 20222). PFAS have been used in industrial processes and commercial products since the 1940s (Gaines, 2023). The success of these products in repelling oil, grease, and water and providing nonstick, stain-resistant, heat-resistant, nonreactive, and fire-retardant properties has driven the synthesis of thousands of PFAS compounds of various carbon chain lengths and configurations in the decades since (Gaines, 2023; Glüge et al., 2020). Due to the extensive history of PFAS usage and their recalcitrant the carbon-fluorine (C-F) bonds, PFAS are now commonly found throughout the environment (Miner et al., 2021; Salvatore et al., 2022; Kurwadkar et al., 2022). Significant concern has been placed on the health and environmental impacts of the more common PFAS, such as perfluorooctanoic acid (PFOA) and perfluorooctanesulfonic acid (PFOS) (Sunderland et al., 2019; Fenton et al., 2021; Kwiatkowski et al., 2020; Sinclair et al., 2020

The C-F bonds cause PFAS to be recalcitrant in nature with respect to biological mechanisms of destruction (O’Hagan, 2008; Tsang et al., 1998). Thus, sufficient energy is required to break these C-F bonds to convert PFAS to inert or more easily treatable species. Furthermore, matrix impacts such as co-contaminants within PFAS-containing wastes may affect their treatment (Shields et al., 2023).

Several thermal and non-thermal destruction technologies are in various stages of research, development, demonstration, and commercialization. Many of these technologies show significant promise for treating PFAS and have been discussed in previous review papers (Berg et al., 20222; Meegoda et al., 2022; Verma et al., 2022). In this review, we investigate the available literature reporting on the performance of combustion-based thermal treatment technologies that treat a variety of PFAS-containing wastes.

In the United States, nonhazardous and hazardous wastes containing wide-ranging concentrations of PFAS are managed via a variety of thermal treatment technologies including municipal waste combustors (MWCs), hazardous waste incinerators (HWIs), and sewage sludge incinerators (SSIs). Industries that use high temperature processes (cement kilns, shale kilns, etc.) may obtain permitting to process hazardous wastes containing PFAS. Thermal processes are sometimes used for site remediation to remove and destroy PFAS from soils and other environmental matrices (e.g., thermal desorption) (Anderson, 1993). In addition, many industrial processes include thermal oxidizers (TOs) as part of their APC systems to destroy pollutants (including PFAS) and process gases.

There is limited information for both the extent to which conventional combustion-based thermal treatment technologies adequately destroy (i.e., mineralize) PFAS and their ability to minimize the formation and emission of fluorinated products of incomplete combustion (PICs) or products of incomplete destruction (PIDs). These data gaps are a result of limited testing at operational facilities, as well as challenges and limited methodologies in accurate PFAS measurement, that lead to uncertainty in characterizing PFAS release from thermal treatment facilities. This article reviews the state of the science with respect to PFAS destruction in commercial-scale thermal waste treatment systems.

2. BACKGROUND ON THERMAL DESTRUCTION OF ORGANIC AND HALOGENATED CHEMICALS

Thermal treatment units use high-temperature combustion, to destroy organic materials and control organic pollutants. In the United States, the Clean Air Act (CAA), Resource Conservation and Recovery Act (RCRA), Toxic Substances Control Act (TSCA), and Comprehensive Environmental Response, Compensation and Liability Act (CERCLA) regulate industrial air emissions as well as the reporting, management, disposal and cleanup of solid and hazardous materials and wastes, and the remediation of abandoned or uncontrolled hazardous waste sites. Thermal oxidation and incineration are effective methods for destroying volatile organic compounds (VOCs) and wastes including halogenated organic wastes such as chlorinated solvents, polychlorinated biphenyls, dioxin-laden wastes, brominated flame retardants, refrigerants, and ozone-depleting substances (Oppelt, 1987).

In the 1980s, investigators at the University of Dayton (Graham et al.,, 1986; Taylor et al., 1990; Dellinger et al., 1993) developed experimental methods to evaluate the relative thermal stability of individual (mostly organic) compounds and this led to the publication of a Thermal Stability Index for 320 species based on temperature required to achieve 99% destruction efficiency (DE) in 1 second, also known as T₉₉ (U.S. EPA, 1989). These compounds were ranked into seven classes with Class 1 compounds being most thermally stable and Class 7 being least thermally stable. This ranking is used to help identify difficult-to-destroy organic species for use during trial burns at hazardous waste incineration facilities. At the time, chlorinated solvents were of particular interest and the Class 1 species are largely comprised of polyaromatic hydrocarbons (PAHs) and chlorinated aromatics. Chlorinated aromatics and chlorinated alkanes (as well as aromatic amines) contribute heavily to Class 2 species. No PFAS are listed. The only fluorocarbons listed are chlorinated fluorocarbons (CFCs). Of the 320 species evaluated, 113 are chlorinated and 8 are fluorinated. Additionally, the thermal stability ranking does not consider the formation of PICs.

Investigators at the National Institute of Standards and Technology (NIST) were among the first to examine the combustion properties of fluorocarbons (Westmoreland et al., 1993; Burgess et al., 1994; Westmoreland et al., 1994; Babushok et al., 1994; Daniel et al., 1994; Burgess et al., 1995; Babushok et al., 1995; Linteris and Truett, 1996).This was originally driven by the need to identify potential replacements for the flame suppressant, Halon 1301 (trifluorobromomethane), and later characterize the flame safety for several new fluorinated refrigerants as replacements for those identified as ozone-depleting substances (ODS). The NIST researchers developed a detailed chemical mechanism for fluorocarbon destruction and their participation in and influence on hydrocarbon flame chemistry. Burgess et al. (1994, 19951995) presents a review of this mechanism development work, and Tsang et al. (1998) discusses some of the unique challenges associated with the destruction of highly fluorinated Class 1 and Class 2 species.

2.1. Factors Affecting the Mineralization, Formation of Intermediates, and Products of Incomplete Combustion

Understanding how PFAS are composed is important to ensuring destruction through incineration. In general, the chemical structure of PFAS plays a big role in the destruction of the substance For instance, the presence of nonfluorinated functional groups (head groups) will assist in the initial decomposition of many PFAS. As a result, the temperatures required to initiate decomposition decrease as follows: fully fluorinated carbon compounds (aka perfluorocarbons, PFCs) > perfluoroacyl fluorides > perfluoroalkane sulfonic acids (PFSAs) > perfluoroalkyl carboxylic acids (PFCAs) > perfluoroether carboxylic acids (Xiao et al., 2020). Additionally, the unimolecular decomposition of PFCs is affected by the length of the carbon chain, with shorter-chained PFCs, such as CF₄ and C₂F₆, being the most stable (Tsang et al., 1998; Steunenberg and Cady, 1952).

During attempted destruction, in the absence of high temperatures and in incorrect conditions, low molecular weight PFCs may be released as air emissions, which are well known for high greenhouse potential (Burkholder et al., 2023). In addition, there are concerns that if fluorotelomer alcohols (another type of PFAS), and 1H-perfluorocarbons are released as PICs, they may convert to perfluorocarboxylic acids (PFCAs) in the atmosphere, furthering PFCA contamination via wet and dry deposition (Ellis et al., 2001, 2004).

2.2. PFAS Destruction, Transformation, and PICs

The importance of a source of hydrogen (e.g., H atoms, H bonds, CH₄,H₂O) for the incineration of PFAS-containing wastes is crucial as PFAS will decompose to formCF₂ carbenes and CF₃ radicals. These species, in the absence of hydrogen, will recombine or combine with F atoms to form stable PFCs such as CF₄ and C₂F₆ (hexafluorethane). Thus, the mineralization of PFAS-containing wastes requires the presence of a source of hydrogen to minimize the formation of stable, highly fluorinated PICs.

3. THERMAL TREATMENT FOR PFAS DESTRUCTION AND EMISSIONS

PFAS come from numerous sources associated with most waste materials (e.g., municipal solid waste, sewage sludge, hazardous solid/liquid waste) sent to thermal treatment and other waste management facilities (Buck et al., 2011; Favreau et al., 2017; Fiedler et al., 2021; Glüge et al., 2020; Hammel et al., 2022; ITRC, 2023; OECD, 2022; Sanborn, Head, & Associates, 2019; U.S. FDA, 2022; Whitehead et al., 2021).

A wide variety of thermal technologies are used to treat these nonhazardous and hazardous wastes with the following main thermal treatment types:

• HWIs, which include stand-alone facilities accepting hazardous solid and liquid wastes;

• MWCs, which include stand-alone facilities accepting nonhazardous solid wastes (i.e., municipal solid wastes [MSW]);

• SSIs, which include dedicated facilities for dewatered sewage sludge treatment;

• Gasification and pyrolysis processes that intentionally operate under substoichiometric or reducing conditions to generate carbonaceous solid or synthesis gas products from wastes; and

• TOs, which include a wide variety of units typically installed at industrial facilities for process gas treatment and air pollution control (namely volatile organic compounds; VOCs).

Several studies have outlined PFAS-containing household, commercial, and industrial products such as cosmetics and personal care products, carpeting and rugs, packaging, grease- and water-resistant paper products, furniture, upholstery, and textiles as likely PFAS sources in waste materials (e.g., sewage sludge, hazardous solid/liquid waste) sent to thermal treatment and other waste management facilities (Buck et al., 2011; Favreau et al., 2017; Fiedler et al., 2021; Glüge et al., 2020; Hammel et al., 2022; ITRC, 2023; OECD, 2022; Sanborn, Head, & Associates, 2019; U.S. FDA, 2022; Whitehead et al., 2021). The types of PFAS compounds and concentrations found in materials contained in thermal treatment facility feedstocks varyy. Figure 1 presents maximum PFAS concentrations measured in various consumer and industrial products and waste matrices as compared to those measured in ashes from MWCs, biosolids from wastewater treatment plants (WWTPs), and other common materials such as AFFFs that may be in the feedstock to thermal waste treatment facilities.

Figure 1. Highest Reported PFAS Concentrations and Compositions Measured in Various Products, Wastes, and the Environment Compared to Thermal Treatment Byproducts.

Note: Numbers in parentheses refer to sources. Numbers to the right of the bars are the number of PFAS analytes.

*Leachate sampled from an eighteen-year-old MSW incinerator ash monofill in Florida operated at boiler temperatures of 930 to 980°C. The fly ash included residuals from spray dryer absorption, activated carbon, nitrogen oxide reduction, and bag house filtration air pollution control devices.

**Maximum PFAS concentrations with published compositions from a MSW incinerator in China. Total PFAS concentrations ranged from 1,460 to 87,600 ppt for fly ash and 3,120 to 77,400 for bottom ash from three facilities.

***Grown near a PFAS production plant. Romaine lettuce grown farther from the PFAS production plant showed < MDL for all 16 tested PFAS.

Sources: 1-Wood Environment & Infrastructure Solutions (2021); 2-Choi et al. (2019); 3-Venkatesan & Halden (2013); 4-CalEPA (2019); 5-CalEPA (2021); 6-Ministry of Environment and Food of Denmark, Environmental Protection Agency (2018); 7-Lang et al. (2017); 8-Allred et al. (2014); 9-Cousins et al. (2022); 10-Sanborn, Head & Associates (2019); 11-Liu et al. (2021); 12-Weiner et al. (2013); 13-FDA (2020); 14-Ruffle et al. (2020); 15-Solo-Gabriele et al. (2020)

3.1. Current Use of Thermal Treatment to Manage PFAS-Containing Wastes

There are 126 operating waste combustion facilities in the United States including 72 stand-alone MWCs, 27 HWIs, 9 medical waste incinerators, and 18 industrial combustors that accept outside wastes (U.S. EPA, 2023a). Each year these facilities directly treat 32,000,000 tons (U.S. short tons) of nonhazardous MSW, 681,000 tons of hazardous waste, and 633,000 tons (dry) of sewage sludge (U.S. EPA, 2020b; U.S. EPA, 2021a; U.S. EPA, 2023b). There are also more than 200 SSIs and more than 4,000 thermal ([VOC] control) units within commercial and industrial facilities.

Table 1 summarizes common sources of PFAS-laden input materials, typical environmental controls, and process emissions and byproducts by primary treatment type. For all treatment types, the primary emissions are combustion gases and byproducts collected within air pollution control devices (APCDs). Discharges from APCDs often include ash residuals from particle control devices and liquid discharges from wet acid gas scrubbers. Other residuals may also be generated depending on the APCD components used. Pollution controls at thermal treatment facilities are generally designed to address criteria and hazardous pollutants as per regulatory requirements. However, the ability of existing APCDs to adequately control PFAS air emissions (e.g., PICs) has not been determined.

Table 1. Summary of Thermal Treatment Technologies, PFAS-containing Feedstock and Pollution Controls.

Thermal Treatment Type	Description	Common Sources of PFAS Input	Environmental Controls	Byproducts Generated
MWC	Multiple high-temperature combustion technologies (nonhazardous—NSPS subparts CCCC, DDDD) for treating PFAS-containing waste mixed with MSW; fluidized bed, rotary kiln, moving grate.	PFAS-laden solids	Baghouse; wet scrubber; SCR; ammonia injection; afterburner	Air emissions; PICs; ash
HWI	Multiple high-temperature hazardous waste combustions technologies (NESHAP EEE) for treating PFAS-containing waste and other hazardous waste; fluidized bed, rotary kiln, moving grate. High-temperature cement kilns firing PFAS-containing waste and other hazardous waste as fuel.	PFAS-laden solids and liquids	Baghouse; wet scrubber; SCR; ammonia injection; afterburner	Air emissions; PICs; ash
SSI	Multiple sewage sludge incineration technologies in current use: multiple hearth furnaces, fluidized bed, rotary kiln, moving grate, liquid injection.	PFAS-laden sewage sludge	Baghouse; wet scrubber; granulated activated carbon (GAC)	Ash; air emissions
TO	Thermal oxidizers include multistage scrubbing systems on industrial processes in a wide variety of configurations (e.g., direct fired, indirect, regenerative recuperative, catalytic, noncatalytic, flameless). Some are specifically designed for high concentrations of halogenated wastes.	Manufacturing processes; other incomplete destruction devices (e.g., kilns)	Post-combustion APCD such as SCR	Air emissions; PICs
	Carbon (GAC) reactivation units; thermal desorption of GAC and combustion of gases.	PFAS-laden GAC (from adsorption)	Additional treatment of gaseous PFAS	Air emissions; PICs; spent GAC (~10%)
	Pyrolysis and gasification units with TOs; converts solids into combustible gases at high temperatures.	PFAS-laden solids	Baghouse; wet scrubber; GAC	Air emissions; PICs; biochar
	In- and ex-situ smoldering treatment via diffusion of air through contaminated media; with and without supplemental fuel.	PFAS-contaminated soils	Post-smoldering APCD	Air emissions; PICs

In addition, few studies have been conducted at full-scale operating units. We present reported values found in the literature for the PFAS DEs and destruction and removal efficiencies (DREs) via computational methods, laboratory studies, and testing at different types of commercial thermal treatment units.

3.2. Analytical Approaches for PFAS Thermal Destruction

Understanding the available data and data gaps for PFAS thermal destruction requires an understanding of the analytical approaches for measuring PFAS. Many analytical approaches are used for measuring PFAS, which include three broad approaches of targeted, non-targeted, and screening analyses. However, many of these approaches focus on the measurement of PFAS in both solids and liquids (e.g., water). The relative lack of standardized methods to characterize PFAS or their possible fluorinated organic byproduct concentrations in combustion flue gases poses a significant problem for closing data gaps.

Recently, the U.S. EPA attempted to address this limitation by developing two other test methods (OTM). The first was OTM-45, which measures nonvolatile polar (and semi-volatile) PFAS (U.S. EPA, 2021b). The second is OTM-50, which measures the nonpolar volatile PFAS (U.S. EPA, 2024). In combination, these two methods can offer significant information about the amounts of PICs especially at low concentrations, which other measuring techniques cannot obtain.

In addition to the EPA OTMs, researchers at laboratory and pilot scales have used Fourier transform infrared spectroscopy (FTIR) and chemical ionization mass spectrometry (CIMS) for non-target analyses to measure fluorinated products in the gas phase (Baker et al., 2023; Krug et al., 2022; Mattila et al., 2024; Shields et al., 2023; Wang et al., 2023; Weber et al., 2021, 2022a, 2022b, 2023a, 2023b, 2024).

3.3. Computational Analysis

PFAS thermal decomposition can be mathematically modeled using quantum chemical calculations to simulate a PFAS molecule such as perfluorobutanoic acid (PFBA). This is usually done in software such as Gaussian 09 or 16 (Frisch et al., 2016a, 2016b) where use of a variety of different levels of theory (usually higher level such as G4) are employed to model PFAS molecular behavior in the gas phase. This provides crucial information into the likely decomposition pathway.

In the case of perfluorocarboxylic acids (PFCAs), Altarawneh (2012), Altarawneh et al. (2022), Khan et al. (2022), and Weber et al. (2023a) all found the lowest barrier pathway is through an α-lactone and the elimination of hydrogen fluoride (HF)) The lactone is unstable and will quickly decompose into a perfluoroaldehyde and CO. In the case of PFBA, it will decompose into the C₄ lactone and release a HF molecule. This C₄ lactone will decompose into the C₃ perfluoroaldehyde and CO. Alternatively, the C₄ lactone can decompose into CO₂ and C₃ perfluoroalkene instead. Other routes directly from PFCAs were also suggested, but all required significantly higher barriers such as the direct route to a 1H-perfluoroalkane.

Instead of a lactone and the release of HF, perfluorosulfonic acids (PFSA) will form an unstable α-sultone and release a HF where the α-sultone will rapidly decompose into a perfluoroaldehyde and SO₂ ( (Altarawneh, 2021,, Khan et al., 2020, Weber et al., 2021), For decomposition to occur, these computational studies show PFCA and PFSA need temperatures above 700°C to achieve an acceptable level of DRE/DE. Similar studies have also been completed on HFPO-DA (sometimes referred to as “GenX”) (Adi and Altarawneh, 2022).

AO significant drawback of using quantum chemical simulations is their limited applicability outside the gas phase, which excludes the influence of reactor surfaces. Additionally, quantum chemical simulation studies are limited to the level of theory, due to the number of heavy atoms (any atoms other than hydrogen) present in PFAS. In addition, these simulation studies are further limited to shorter chained PFAS, which may be ill-suited as proxies for the more abundant longer PFAS (Altarawneh, 2012, 2021; et al., 2022; Weber et al., 2021, 2023a).

Some of these PFAS were ranked based on the incinerability index of Taylor et al. (1990), which ranges from Class 1 for the most thermally stable chemicals to Class 7 for the most thermally labile chemicals. Branching substantially reduces the required incineration temperatures, while chain length has no impact on PFCA decomposition rates (Blotevogel et al., 2023). Bimolecular reactions with secondary species such as atoms and radicals have lower activation barriers than unimolecular decomposition, but their second-order kinetics depend on concentration and thus waste stream composition. When simulating radical concentrations under typical incineration conditions, the determined free energies of activation suggest an incinerability index for linear PFOA in Class 3, for branched PFOA in Class 4, and for hexafluoropropylene oxide-dimer acid (HFPO-DA) and both PFOA and HFPO-DA ammonium salts in Class 5 (Blotevogel et al., 2023). Collectively, the temperatures needed to destroy 99.99% of PFOA, HFPO-DA, and other PFCAs in 2 seconds of gas residence time are well below those of other commonly incinerated organic compounds such as chlorobenzene (Class 1). However, while Blotevogel et al. (2023) provided insights into the predicted incinerability of some PFAS, they did not address the formation of PICs. This is a significant limitation of the incinerability index and rankings, as they only consider the effectiveness of thermal treatment on the target compounds and do not account for the formation of byproducts or PICs. These byproducts could fall into an incinerability class that is much harder to destroy, as is likely the case for CF₄ or C₂F₆ (Krug et al., 2022).

3.4. Laboratory Studies of PFAS Destruction

Waste destruction in commercial HWIs is characterized by the calculation of DREs based on the mass feed and mass air emission of one or more organic waste species denoted as a principal organic hazardous constituent (POHC). HWIs include mass removed from the combustion gases by APCDs in the calculation as these discharges are treated as hazardous wastes. Laboratory- and pilot-scale studies often perform the same calculation but do not include waste mass removed by APCDs. In these cases, DEs rather than DREs are used. DEs and DREs of a limited set of PFAS are often the only data available. Yamada and Taylor (2003) investigated PFOS incineration in a bench-scale quartz reactor at two temperatures (600°C and 900°C). While DE was not directly reported, PFOS remaining in the reactor system/transfer lines and samples suggest ~99.56% removal at 600°C and ~99.91% removal at 900°C with a 2 second residence time. The authors concluded “[t]he data from this laboratory-scale incineration study indicates that properly operating full-scale incineration systems can adequately dispose of PFOS and the C8 perfluorosulfonamides” with further research needed to accurately determine the minimum temperature for 99.99% DE. Subsequent work demonstrated PFOA was not detected at a temperature of 1000°C for a 2 second residence time under conditions representative of waste incineration of fluorotelomer polymers and fluoropolymers (Yamada et al., 2005; Taylor et al. 2008, 2009, 2014). Taylor et al. (2014) report limits of detection for PFOA equivalent to 54 ng/dsm³.

Weber et al. (2021) determined Arrhenius kinetics for PFOS decomposition in a stainless-steel flow reactor over a temperature range of 400°–615°C. Using these values, a T₉₉ (temperature required to achieve 99% DE in 1 second) for PFOS of 690°C was calculated for 2 second residence times typical of full-scale thermal treatment systems.

Taylor (2022) suggested T₉₉ values for perfluorosulfonic and perfluorocarboxylic acids of less than 850°C. These data include F mass balances in addition to thermal destruction of the parent compounds in the incinerability assessment. Incinerability studies to provide additional data to better pinpoint the T₉₉ values for these classes of PFAS compounds are forthcoming.

Weber et al. (2022b) conducted bench-scale experiments providing crucial insights on PFOS decomposition between 450 and 1000°C in inert pyrolysis conditions. They presented information on possible PICs formation, which included C₂F₆, C₈F₁₆O, C₂F₄ and COF₂.

More bench-scale studies completed by Weber et al. (2022a,2023b) added water vapor as a source of excess hydrogen and found that the combination with air (O₂) (humid air conditions) and temperatures above 900°C promoted the mineralization of PFOS with 99% of F into HF and 100% of C into CO₂. Additionally, direct reaction between O₂ and H₂O with PFOS was not mentioned by Weber et al. (2023b). The research completed by Weber et al. (2022a, 2023b) indicates that thermal treatment of PFAS in the presence of H₂O as a source of hydrogen can lead to the mineralization of PFAS into HF and CO₂, after which HF can be removed by conventional acid gas control technologies.

Wang et al. (2023) and Weber et al. (2023a) found that under inert pyrolysis conditions PFCAs including PFOA (Weber et al., 2023a) and perfluoropropionic acid (PFPrA) and PFBA (Wang et al. 2023) will decompose into a range of stable PFCs including the very stable perfluoro-1-alkenes, C₂F_6, C₂F₄, and COF₂. Similar PICs were also observed in the thermal treatment of PFAS-laden spent adsorbents (e.g., granulated activated carbon [GAC]) (Alinezhad et al., 2022, 2023; Koster van Groos, 2021; Xiao). Additionally, SiF₄ was a commonly found product in the presence of PFAS products and Si-based reactors (Weber et al., 2024; Wang et al. 2023).

Shields et al. (2023) also reported PIC data associated with PFAS incineration and examined the relationship among DE, PICs, and temperatures. Shields et al. (2023) used an EPA pilot-scale research furnace and atomized a PFOS-dominant AFFF at various furnace locations and combustion environments. These included atomization through flame with natural gas auxiliary fuel and at various post-flame locations so the AFFF would experience decreasing peak temperature exposures ranging from ~1200° to ~800°C. Stack measurements included a solvent extraction test method for semi- and nonvolatile polar PFAS (designated OTM-45), and a recently published canister-based gas chromatography/mass spectrometry (GC/MS)-based test method for nonpolar volatile PFAS. Analysis of the AFFF and OTM-45 samples was used to calculate DEs for 10 PFAS quantified in the AFFF. A precursor to OTM-50 was used to quantify concentrations of 30 PFAS PICs. Results indicated that several operating conditions with peak temperatures above 1080°C resulted in high DEs (>99.99%) and low PIC emissions near the method detection limits. However, several conditions below 1000°C produced DEs >99.99% for quantifiable PFAS, especially the PFSAs, and emission concentrations on the order of mg/m³ for several nonpolar volatile PFAS. PICs included all eight C1-C8 1H-perfluoroalkanes and other PFAS. The study concluded DE alone may not be the best indication of total PFAS destruction.

Mattila et al. (2024) through a combination of CIMS and the EPA pilot-scale research furnace found that PICs including PFCAs and trifluoroacetic acid (TFA) were readily forming. The formation of PFCAs and TFA was observed to occur predominantly at temperatures below 760°C. This further suggests the importance of operating conditions.

Together, the current literature performed at laboratory-scale suggests calculated destruction of 99––99.99% for PFOA and PFOS are achievable at temperatures of 650––800°C and indicate that PFAS can be considered as Class 3 to Class 5 POHCs if fluorocarbon PICs are not considered. The theoretical and experimental studies outlined above suggest some PFAS are not thermally stable. However, destruction of the recalcitrant C-F backbone associated with PFAS will require higher temperatures and free radical chemistry associated with combustion conditions. Complete mineralization of PFAS is possible but careful consideration needs to be taken for temperatures and other operating conditions such as the need for excess hydrogen from fuel sources or moisture and flame conditions for production of reactive OH radicals, and hydrogen atoms for rapid conversion of C-F bonds to HF and CO₂ and to minimize the formation of stable PICs such as CF₄.

3.5. Testing at Operating Thermal Treatment Facilities

3.5.1. Hazardous Waste Incinerators

HWIs are facilities specifically regulated under RCRA to treat materials regulated as hazardous. In the United States, HWI operating conditions and requirements are set per regulation based on the materials being treated (e.g., RCRA, CERCLA) and the results of a trial burn to document POHC destruction. Commercial HWIs are often rotary kilns that treat solid or containerized hazardous wastes. Retention times for solids in rotary kilns range from 0.5 to 1.5 hours, and gases have a residence time of approximately 2 seconds. Temperatures range from 650° to 1650°C in the initial combustion chamber (rotary kiln) and 1100° to 1370°C in the afterburner or secondary combustion chamber (U.S. EPA, 2020a).

HWIs use the secondary combustion chamber for treatment of vaporized gases. These secondary combustion chambers can include an afterburner with additional heating value delivered, or they may be over-fire air to facilitate oxidation in transient substoichiometric conditions. HWIs use post-combustion APCDs. Fabric filters and electrostatic precipitators (ESPs) control particulates (including fly ash) and wet or dry scrubbers are used to control acid gases. These controls may generate separate waste streams (i.e., ash and scrubber discharges) that may contain residual PFAS from thermal treatment. The effectiveness of post-combustion APCDAPCDs to prevent release of PFAS compounds is unknown and fluorinated PICs are not well characterized in any thermal system treating PFAS wastes.

HWIs are required by EPA to conduct testing to determine DRE performance (40 CFR 63.1219I). Time, temperature, and turbulence ensure good combustion and high DREs. The purpose of DRE testing is to show a percentage that represents the mass of a compound destroyed or removed in an incinerator relative to the mass that entered the system. For HWIs, and depending on the waste, EPA requires a minimum DRE of 99.99% (RCRA) or 99.9999% (CERCLA). During DRE tests, POHCs are designated by EPA per the waste feed. The POHCs are chosen based on factors such as incinerability (difficulty to destroy by incineration) and concentration in the waste feed and must be representative of the most difficult-to-destroy organic compounds in the hazardous waste feed stream (40 CFR 63.1205(c)(3)(ii)).

POHCs are selected from an index developed by Taylor et al. (1990); however, only a few one and two-carbon (C₁ and C₂) fluorinated organic compounds were assessed during development of the incinerability index. PFAS are chemically and structurally different from the chlorinated organic compounds listed in the incinerability index due to the highly electronegative F in the C-F bonds and tendency to reform fluorinated alkyl chains from CF₂ radicals and fluorinated PICs from F radicals compared to other halogenated compounds (Phelps, 2020).

Additional research is needed to validate estimates of the relative incinerability of PFAS compounds and formation and destruction of PICs. Limited studies have published testing for PFAS destruction in commercial and captive HWIs. Table 2 provides a summary of test results found in the literature for PFAS destruction in HWIs. Only one study identified evaluated native (non–PFAS spiked) mixed hazardous waste. Other published studies are narrow in scope, targeting specific PFAS materials, such as AFFF and contaminated soil.

Table 2. Published Test Results for PFAS Destruction in HWIs.

Study	PFAS Input	Technology and Operating Conditions	Samples Tested	PFAS Tested	Input Concentrations	Outlet Concentrations	DE
Japan Ministry of Environment, 2013	Fire-extinguishing foam (0.67% PFOS content)	Commercial rotary kiln; alkali adsorber, wet electric dust collector; 1100°C (primary) - 900°C (secondary); 8 seconds (combined)	Flue gas, residues, effluent, sludge, treatment water	PFOS	Not Reported	Not Reported	99.999891%
NRC Alaska, 2019	Contaminated soil	Commercial rotary kiln; baghouse; wet scrubber; 815°C (primary) - 1200°C; residence time unspecified	Post-treatment soil, baghouse fines; air emissions	24 total PFAS compounds	Highest sample was 0.0765 mg/kg (PFOA) and 7.24 mg/kg (PFOS)	Soil/baghouse: ND except for PFHxS in 2 baghouse fines samples (0.000372 and 0.000383 mg/kg) and PFOS (0.0003–0.00220 mg/kg) Stack Air: 10/28 PFAS detected. PFOA 0.0185 mg/hr and PFOS was ND. Total air PFAS avg. of 0.0791 mg/hr.	Not Reported
EA Engineering, 2021	Mixedhazardous waste (including AFFF) spiked with PFOA, PFOS, PFHxS, and HFPO-DA (GenX); or additional AFFF (depending on test run)	Commercial rotary kiln with afterburner chamber; baghouse; wet scrubber; 1034°−1098°C (kiln), 1122°−1154°C (afterburner); 2–3 second gas residence time	Hazardous wastewaste input, treatment chemicals, process residues (slag, baghouse dust, spray dryer solids), and stack gas	49 target PFAS compounds	Not reported	Baghouse/filter: ND except PFOS, PFHxA, and FBSA. Stack Air: up to 25/49 PFAS detected with individual emissions rates of 10−9 to 10−7 lb/hour. Total stack PFAS emissions on the order of 10−6 lb/hr (0.45 mg/hr).	>99.9999% for legacy terminal PFAAs (PFOA, PFOS, PFHxS, and GenX)

NRC Alaska (2019) analyzed a 17 million British thermal unit (Btu) (approximately 18 million kJ) per hour kiln used to treat PFAS-contaminated soil in Alaska. Soil samples were tested for 24 PFAS using SGS Laboratories’ method 537M with 21 PFAS present in the pretreatment samples, including PFOS and PFOA. The highest PFOS concentration detected in the pretreatment soils was 7.24 mg/kg and the highest PFOA concentration was 0.0765 mg/kg. The operating temperature in the kiln was 815°C in the primary combustion unit. Gases from the kiln then passed into an 8 million Btu/hr (8.4 million kJ/hr) secondary combustion unit operating at 1200°C.

Total PFAS DRE was not reported. Post-treatment, soils and baghouse fine particulates did not show any PFAS compounds above the detection limits except for PFOS in the treated soil and baghouse fines (0.000300 mg/kg to 0.002200 mg/kg) and perfluorohexanesulphonic acid (PFHxS) detected in two baghouse fines samples (0.000372 mg/kg and 0.000383 mg/kg). Air samples from the kiln exhaust showed 10 of the 21 PFAS compounds found in the soil after extraction using XAD resin tubes and Method 537. The total combined PFAS air emissions from the kiln were on average 0.0791 mg/hr, including PFOA (0.0185 mg/hr). Later work with the system found a source of potential contamination of PFAS in the testing of the flue gases indicating that more studies need to be conducted to quantify air emissions. Based on the results of this test, a wet scrubber was added to the process prior to commercial startup of operations to control HF emissions.

The Japan Ministry of Environment (2013) conducted a study treating AFFF in an operating commercial-scale HWI and reported high levels of PFOS destruction (DRE of 99.999891%) at 1100°C in the primary furnace and 900°C in the secondary furnace for 8 seconds of combined gas residence time. This measurement was taken by comparing the amount of PFAS in the simulated waste and in the post-combustion air, liquid, and solid waste streams. A possible unintended consequence of the batch introduction of AFFF into the combustion unit is the potential for upsetting the steady-state operation and steady high temperatures, which can reduce DRE and lead to an increased PIC formation.

EA Engineering (2021) conducted a study treating hazardous waste spiked with four commercial PFAS samples (PFOA, PFOS, PFHxS, HFPO-DA) or AFFF in an operating commercial-scale HWI in the United States (Clean Harbors’ HWI in Aragonite, Utah). They reported high destruction levels for PFOA (DRE of 99.999943%), PFOS (DRE of 99.999955%), PFHxS (DRE of 99.999977%), and GenX (DRE of 99.999979%) at >1000°C in both the kiln and afterburner units with a 2- to 3-second residence time in the afterburner chamber. DRE measurements were made through a mass balance of the amount of PFAS in native and spiked waste materials and the amount of PFAS in the post-treatment combustion air, liquid, and solid waste streams. Samples taken from the stack gas were collected using sequential XAD resin traps and Method OTM-45 and analyzed via liquid chromatography with tandem mass spectroscopy (LC/MS/MS). Total stack PFAS emissions were observed on the order of 10⁻⁶ lb/hr (0.45 mg/hr), with up to 25 of 49 PFAS detected. One compound (PFOS) was detected at low levels in slag, spray dryer solids, and baghouse dust, with an additional two compounds (PFHxA and nonafluoro-butane-1-sulfonic acid amide [FBSA]) detected in spray dryer solids and baghouse dust. Volatile non-polar PICs were not characterized. Additional analysis quantified “hidden” feedstock PFAS from high molecular-weight precursors that break down into the targeted compounds. Comparison of total oxidizable precursor assay results indicated concentrations of PFAS in AFFF-containing waste increased by one to two orders of magnitude after subjection to thermal and chemical oxidation conditions due to the breakdown of higher molecular weight precursors, which are not target analytes, into smaller molecules (total F and C mass do not change).

3.5.2. Municipal Waste Combustors

There are many sizes and designs for MWCs that generally include mass burn, refuse-derived fuel, and modular systems. MWCs accept bulk MSW that includes a wide range of PFAS-containing materials. The amount and chemistries of PFAS treated in most MWCs have not been well characterized to date; however, Table 3 and Figure 1 provide some background into MWCs and PFAS. Besides the difficulties that the heterogeneity of MSW poses, the complexity of analytical techniques along with the plethora of measurable PFAS, and their currently unmeasurable PFAS precursors, add further challenges to estimating the PFAS content of MSW. A value of 50 μg/kg of PFAS content was used as a conservatively low estimate by Coffin et al. (2023) based on total leachable PFAS concentrations of 26 PFAS analytes using a modified EPA Method 1316 from MSW screenings by Liu et al. (2022). Assuming this average concentration, conservatively approximately 1500 kg of PFAS are treated annually via MWCs in the United States using 2018 data (U.S. EPA, 2020b).

Table 3. Published Test Results for PFAS Destruction in MWCs.

Study	PFAS Input	Technology and Operating Conditions	Sample Tested	PFAS Test	Input Concentration	Outlet Concentrations
Lemieux et al., 2007	Carpet (with and without stain-resistant coating)	Pilot-scale (0.73 kW) rotary kiln; no APCD; operating at <1000°C; no specified gas residence time	Air emissions	14 total PFC compounds	2.72–27.2 mg/hr PFAS based on 6 lb/hr carpet feed	Most common compounds detected were PFHxA and PFOA (all <1 ug/m3). Others were ND. No statistical difference between natural gas only or adding carpet.
Aleksandrov et al., 2019	Wood pellets with PTFE pellets	Operating 1.5 MW rotary kiln; fabric filter, 2 scrubbers, SCR; 4s (870°C) - 2.7s (1020°C)	Flue gas	N/A	300 g/hr PTFE	Too low to be quantified. No statistical difference in PFAS concentration between control and PTFE samples.
Sandblom, 2014	Bulk MSW	Four operating MWCs with carbon injection; filter and electrochemical filter; wet scrubber; operating at >850°C; no specified gas residence time	Slag, fly ash, sludge, condensate, wastewater (no air emissions)	PFAA	Unknown	Solids: single-digit to sub ng/g range; Liquids: single-digit ng/L to below MDL. Total estimated PFAS deposited to landfills from ash is <3 kg/year for each chemical.
Awad et al., 2021	Bulk MSW	Thirty-one active MWCs (24 grate type, 4 FBF, 1 rotating, 2 both grate + FBF); APCD varies; operating ranges 850°−1125°C; no specified gas residence time	Bottom and fly ash, condensate (no air emissions)	27 PFAS and precursors	Unknown	Bottom ash: 9/31 had PFAS (0.22–12.76 ug/kg) Fly ash: 15/31 had PFAS (0.18–37.71 ug/hg) Condensate: 13/31 had PFAS (0.28–182.95 ng/L).
Bjorklund et al., 2023	Bulk MSW with and without Sewage Sludge	Full-scale (20 tph) moving-grate boiler MWC; operating ranges 850°−1100°C; 2 second residence time	Bottom ash, APCD residue, treated process water, gypsum, and flue gas	18 native PFAS	Unknown	Flue gas: PFBA (2.5–27 ng/m3), PFHxA (0.39–2.0 ng/m3) and PFOA (0.23–0.79 ng/m3) Bottom ash: PFBA (0.81 – 1.5 ng/g), PFHxA (0.16 ng/g) and PFOA (0.54 ng/g) ACPD residue: up to 1.3 ng/g total PFAS Treated process water: total PFAS up to 220 ng/L (sludge) and 190 ng/L (no sludge).

Typical operating temperatures for MWCs range from 750° to 1100°C. Giraud et al. (2021) reported that MWC temperatures for municipal waste-to-energy facilities in the United States ranged from 884°C to as high as 1226°C with gas residence times greater than 2 seconds. The temperatures presented by Giraud et al. (2021) reflect a combination of temperature measurements, publicly presented data, and estimated operating conditions; and as such, may not be representative of typical operating conditions due to temporal variation caused by variable feed rates and feedstock compositions and spatial variation within MWCs.

Combustion ash is the main solid byproduct of the combustion process and is a potential source of PFAS that is not destroyed in the combustion chamber, with short chain PFAS such as PFBA usually detected in the ng/g range (Solo-Gabriele et al., 2020; Liu et al., 2021).

Table 3 provides a summary of test results found in the literature for PFAS destruction in MWCs. Although the results can be viewed as representing high levels of PFAS destruction, a review of the technologies, analysis methods, and results highlight the challenges of gaining a complete picture of PFAS destruction from one study. Measurement of PFAS destruction is limited for air emissions, with most studies testing a subset of outlet streams and none including the measurement of PICs. Three of the studies in Table 3—Bjorklund ((2023),), Awad et al. (2021),), and Sandblom (2014)—were performed at commercially operating and pilot-scale facilities, while two others—Aleksandrov et al. (2019) and Lemieux et al. (2007)—were performed using separate pilot-scale rotary kilns. In general, the MWCs and kilns included rely on multiple combustion stages and varied pollution control treatment trains.

Air emissions of PFAS are difficult to quantify because of challenges measuring PFAS in gaseous streams. Bjorklund (2023) presented the first published PFAS measurements in flue gas at operating MWCs. Sandblom (2014) and Awad et al. (2021) did not include sampling and testing for PFAS in air emissions but rather the ash and condensate byproducts. Lacking a validated EPA method for the measurement of PFCs from combustors, the remaining studies employed various sample collection and analysis methods to examine emissions. PIC measurements were not included in any of the studies.

Lemieux et al. (2007) may be among the first studies to characterize PFAS air emissions from incineration environments. With respect to PFAS compounds tested, Lemieux et al. (2007) targeted14 perfluorinated compounds, specifically PFCAs of chain length 4–12; PFSAs of chain length 4, 6, 8, and 10; and PFOSA measured in exhaust emitted from the incineration of carpet samples spiked with commercially available PFCs. Lemieux et al. (2007) compared results using four sampling methods: Tenax, methanolic extracts and impinger, water impinger samples analyzed via high-performance liquid chromatography and mass spectrometry (HPLC/MS/MS), and stack samples analyzed via LC-MS/MS. As noted in Table 3, Lemieux et al. found PFAS contamination in background measurements to be highly elevated. Awad et al. (2021) tested for 27 PFAS compounds and PFAS precursors in the bottom ash, fly ash, and condensate collected from active MWCs. Samples were tested using LC-MS/MS for PFCAs (chain lengths 4–14, 16, and 18), PFSAs (chain length 4, 6, 8, and 10), five PFSA precursors, and five PFCA precursors. Aleksandrov et al. (2019) tested for 31 PFAS compounds including 13 perfluoro-carboxylic acids anticipated to be potential combustion products of the PTFE pellet feedstock; and employed two sampling methods (one based on EPA Method 5 and one based on VDI 2471) with samples analyzed via LC-MS/MS. Bjorklund et al. (2023) performed targeted analysis for PFCAs (chain lengths 4–8, 10, 12, 14), PFSAs (chain lengths 4–8, 12), fluorotelomer sulfonic acids (6:2 and 8:2), and polyfluroroalkyl phosphoric acid diesters (diPAPs) (6:2 and 8:2), and 9 isotope-labeled PFAS. Samples were collected from bottom ash, APCD residue, treated process water, gypsum, and flue gas, and were extracted using weak anion exchange solid-phase extraction (WAX-SPE) for liquid chromatography mass spectrometry (LC-MS) analysis. Sandblom (2014) focused on perfluoroalkyl acid (PFAA) only. Bjorklund et al. (2023) identified PFAS in all residual streams, with short-chain PCFAs (mainly PFBA and PFHxA) being dominant.

Solo-Gabriele et al. (2020) examined the PFAS content of leachate from five landfill sites that received different types of waste, including an MWC ash monofill. Samples collected were tested for 11 PFAS compounds: PFOA, PFOS, perfluorohexanoic acid (PFHxA), PFBA, perfluoropentanoicp acid (PFPeA), perfluoroheptanoicp acid (PFHpA), perfluorohexanesulfonicp acid (PFHxS), perfluorononanoicp acid (PFNA), perfluorodecanoicp acid (PFDA), fluorotelomer carboxylic 5:3 acid (5:3 FTCA), and perfluorobutanesulfonic acid (PFBS). Total leachate PFAS concentrations from the ash monofill was 2,820 ng/L, as compared to leachate samples from landfills containing mixed waste (including MSW, C&D, lower temperature combustion ash) of 19,970 ng/L. The monofill received only ash from an MWC that operated at a temperature of 930°–980°C, whereas the mixed waste landfills received non-combusted mixed waste and ash generated from combustors operating at lower temperatures. While these results indicate reduction of PFAS concentration in solids from higher temperature thermal destruction units, it cannot be concluded whether this is the result of complete mineralization of PFAS at these units or conversion to PICs. Further direct measurement of flue gas emissions including PICs is needed.

Together, these studies reflect the limited current state of the science for PFAS destruction in MWCs and provide some indication of the presence or concentration of PFAS compounds in outlet streams. Measurement is most limited for air emissions, with only one study characterizing flue gas concentrations (Bjorklund et al., 2023). Most studies tested a subset of outlet streams and none included the measurement of PICs. Thus, the current understanding of PFAS destruction via direct measurement at operating MWCs is limited.

3.5.3. Sewage Sludge Incinerators

SSIs include units designed to treat sludge from WWTPs. In general, SSIs can be classified as multiple hearth furnaces (MHFs) and fluidized bed combustors (FBCs), but other configurations exist. Depending on the hearth, temperatures can range from 425425° to 925925°C throughout the furnace (U.S. EPA, 2023c). MHFs may be equipped with an afterburner or TO, to help ensure that VOCs are fully oxidized. In the context of PFAS, dried biosolids sent to SSIs have been shown to contain PFAS, including one study that showed PFAS concentrations in biosolids used for compost to be over 15 ng/g, with the largest component of that being PFOS (Williams, 2021). For comparison, a study to evaluate the presence of PFAS in municipal wastewater and associated residuals (sludge/biosolids) in Michigan found an average of 195 μg/kg for PFOS in sewage sludge (Michigan EGLE, 2020). Bjorklund et al. (2023) demonstrated when sewage sludge was added to the waste fuel mix at MWCs the total concentration of extractable PFAS more than doubled compared to MSW-only feedstocks. Operating temperatures for SSIs are generally lower than those for MWCs and HWIs and range from about 925°C for MHF (in the combustion hearth) and 7500°−9257500925°C for FBCs, which are the predominant design types in the United States (U.S. EPA, 2023c). SSIs may also employ TOs to ensure total organic oxidation.

No current studies were found that tested SSIs for gaseous emission of PFAS. Instead, most studies focus on sampling and testing of sewage sludge for PFAS. For example, Sun et al. (2011) found the highest recorded PFOA content in sewage sludge across 12 wastewater facilities in the United States was 241 ng/g and the highest PFOS content was 993 ng/g. Of the 12 facilities, one used an SSI for sludge management and noted considerable reduction in PFCs in the facility’s solid wastes (Loganathan et al., 2007). Kim et al. (2015) analyzed the PFOA and PFOS content of commercially available biochar material made from sewage sludge and other sources, such as oak and rice husk, and found overall, biochars from sewage sludge contained 10.64–12.03 ng/g of PFOA and 4.82–6.28 ng/g of PFOS while biochars from non-sewage sludge sources did not contain detectable levels of PFOA or PFOS.

3.5.4. Thermal Oxidizers

TOs have long been used in industrial settings to remove organic compounds from process gas waste streams and can include direct- or indirect-fired, regenerative, recuperative, catalytic, and flameless configurations.

Fluorinated organics are routinely disposed of in TOs, combustors, and incinerators of various designs (Chemours Company FC, 2020). TOs are commonly used by the semiconductor industry and by fluorocarbon manufacturers to destroy fluorocarbon process waste gases. In 1991, EPA’s Office of Research and Development participated in a study to evaluate the thermal destruction of CFC-11 (trichlorofluoromethane) and CFC-12 (dichlorodifluoromethane) using a T-Thermal, Inc. Sub-X thermal oxidizer. Results indicated high levels of CFC destruction and minimal formation of volatile and semi-volatile organic byproducts (U.S. EPA, 1993).

In the context of PFAS, TOs are used to treat fluorinated organics. Many short-chain PFAS are treated in TOs per requirements of the Montreal Protocol, including refrigerants, flame retardants, and PFAS and fluoropolymer precursors (UNEP, 2018). TOs also treat process gases that contain precursors and byproducts during PFAS manufacturing, and TOs may be installed at industrial facilities that utilize PFAS materials. TOs can also be included as part of GAC regeneration and reactivation furnaces, pyrolysis/gasification processes, and soil remediation processes that may treat PFAS-containing materials.

TOs have been identified as a method to control PFAS emissions from industrial sites and add-on controls to other PFAS mitigation technologies to remove PFAS from concentrated byproduct streams (Chemours Company FC, 2020). Modern TOs have control efficiencies for VOCs higher than 99% to meet regulatory limits. Operating temperatures for TOs range are typically from 760° to 1200°C and residence times from 0.5 to 2 seconds. TOs may use catalysts to reduce the temperature needed for control of the organic constituents of the process gas (Lewandowski, 2017). Additional control devices such as fabric filters or acid gas scrubbers may be used in systems with TOs to control other pollutants.

Table 4 provides a summary of test results found in the literature for PFAS destruction in other thermal treatment methods, including TOs. While the results can be viewed as representing high levels of PFAS destruction, they do not provide a complete picture of PFAS destruction.

Table 4. Published Test Results for PFAS Destruction in Other Thermal Treatment Methods.

Study	PFAS Source	Design and Operating Conditions	Sample Tested (Method)	PFAS Tested	Input Concentration	Outlet Concentration	Reported DE
Yamada, 2005	Textiles treated with fluorotelemer-based acrylic polymer	Small-scale combustion; no APCD; 725°C; 2 seconds	Combustion gases (in-line GC/MS)	PFOA	Not Reported	Not Reported	99.90%
Taylor et al., 2014	Fluorotelomer-based polymers	Small-scale combustion, no APCD, 1000°C, 2 seconds	Impingers (LC/MS/MS)	PFOA	Not fed	Not detected at a detection limit of less than or equal to 54 ng/dscm	Not Reported
Major (SERDP), 2019	PFAS absorbed in GAC	Smoldering No APCD; 908°–950°C; residence time unspecified	Soil (EPA Method 537 Rev 1)	PFOA, PFOS, PFHxS	PFOA - 510–590 mg/kg PFOS - 120–140 mg/kg PFHxS - 220– 220 mg/kg	Soil - All compounds ND (0.4 ug/kg)	Not Reported
Major (SERDP), 2019	PFAS-contaminated soil with GAC as smoldering fuel	Smoldering no APCD; 1016°C; Residence time unspecified	Soil and air emissions (EPA Method 537 Rev 1 and XAD tubes)	PFOA, PFOS, PFHxS	PFOA - 13.41 mg/kg PFOS - 23.3 mg/kg PFHxS - 16.86 mg/kg	Soil - All compounds ND (0.5 ug/kg) 82% fluorine captured as HF	Not Reported
Major (SERDP), 2019	PFAS- contaminated soil with GAC as smoldering fuel	Smoldering No APCD 1064°C Residence time unspecified	Soil and GAC (EPA Method 537 Rev 1)	PFOA, PFOS, PFHxS, PFNA, PFBS, PFHpA	PFOA - 14.91 mg/kg PFOS - 10.87 mg/kg PFHxS - 10.84 mg/kg PFBS - 3.19 mg/kg PFHpA - 13.32 mg/kg PFNA - 28.73 mg/kg	Soil - All compounds ND (0.5 ug/kg)	Not Reported
Chemours Company FC, 2020	Process Emissions	10 MMBtu/hr Thermal Oxidizer HF scrubber 1922°F; Residence time unspecified	Air emissions (Method 18)	HFPO, HFPO-DA, HFPO-DAF, COF2, Fluoroether E-1	Total PFAS inlet: Run 1: 41809 g/hr Run 2: 35508 g/hr Run 3: 43286 g/hr	Total PFAS outlet (sum of 5 PFAS) Run 1: <0.0733 g/hr Run 2: <0.0907 g/hr Run 3: <0.0607 g/hr	99.99981%

Chemours Company FC (2020) is the only current study found that performed testing on a TO used to treat industrial process gases. Emission testing of a group of five PFAS compounds from the processes (HFPO, HFPO-DA, HFPO-DAF, COF₂, and fluoroether E-1) showed a DE of greater than 99.999% during all three runs performed (Chemours Company FC, 2020). The control system converted the fluorine in the PFAS into calcium fluoride for offsite disposal. Although multiple PFAS compounds were tested for across the studies, PICs were not included.

Discussion of specific TO technology types—including GAC reactivation, pyrolysis and gasification, and smoldering—are provided in the following sections.

3.5.5. Granulated Activated Carbon Reactivation and Smoldering

GAC is used to remove PFAS from contaminated water. Once the GAC is used to adsorb the PFAS, it must be regenerated and eventually reactivated before reuse (McNamara et al., 2018). Regeneration refers to a simple low temperature vaporization of adsorbed PFAS, while reactivation refers to a higher temperature process to reopen and reactivate internal pores and surface sites (typically using steam) to reestablish adsorption properties. GAC regeneration and reactivation are done in large variable-temperature furnaces. Eventually, when the internal surfaces can no longer be regenerated and reactivated, the GAC is disposed of, often by incineration (DiStefano et al., 2022; Kempisty2019; Sonmez Baghirzade et al., 2021).

GAC generally undergoes four series of reactions during reactivation: drying, desorption, pyrolysis, and oxidation at temperatures up to 800°C. Residence time varies between units but can be up to 120 minutes at some facilities. PFAS are vaporized into the gas phase during the reactivation. The vapors are then typically sent to a TO for combustion of organic compounds and other APCDs for the removal of constituents such as particulates and acid gases. Experimental data suggest a DE of PFAS of 99.99% can be achieved in GAC reactivation MHFs equipped with afterburners to treat off-gases (DiStefano et al., 2022). Additional studies such as Watanabe et al. (2018) have shown that fluorine mineralization can be achieved during GAC reactivation, although it is much less (51––91% recovery as mineralized fluorine) than the destruction seen in TOs for VOC control.

An alternative option to incineration or the reactivation of GAC is smoldering. Smoldering is a method involving a self-sustaining combustion reaction to remove organic compounds embedded within a nonflammable matrix such as PFAS-contaminated soils or GAC. Major (2019) examined the use of smoldering to remove PFAS from a sand matrix mixed with GAC to simulate contaminated soil. Tests were conducted to examine whether PFAS had been adsorbed into the GAC prior to the test or added to the sand and GAC matrix. Results showed combustion conditions sustained temperatures greater than 900°C. In all cases, the target PFAS substances in the trial each measured below the detection limits of 0.4 μg/kg and 0.5 μg/kg after smoldering in the soil. Air exhaust from the system was captured using XAD tubes and found to contain 82% of available fluorine as HF.

3.5.6. Pyrolysis and Gasification

Pyrolysis and gasification are technologies have long been used as an alternative to direct combustion or incineration of hydrocarbon-containing materials. Pyrolysis and gasification involve heating material under controlled conditions that limit the amount of oxygen into the reactor. The resulting products are a carbon-rich solid stream and a hydrocarbon-rich liquid or gaseous stream. Pyrolysis is conducted with little to no oxygen at temperatures ranging from 300° to 850°C. Products of pyrolysis are typically short- and intermediate-chain hydrocarbons and other solid combustion byproducts (char). A fraction of these hydrocarbons can be condensed as a pyrolysis oil and utilized as fuels or feedstocks for chemical synthesis. The non-condensable fraction is often used as fuel to provide heat for the pyrolysis process. Gasification occurs at higher temperatures, around 800° to 1000°C, with substoichiometric amounts of oxygen or air added to the reaction chamber to encourage the formation of gaseous products. Gasification produces a synthesis gas (primarily CO, H₂, and CO₂) that can be used as a fuel. Steam can be added to shift the gas-phase reactions to optimize CO and H₂ formation (Pahnila et al., 2023; Shah et al., 2023). One application that has recently been studied is the use of pyrolysis and gasification at sewage treatment plants to treat sewage sludge that often contains PFAS. A number of these facilities have had operational challenges that limited the technology’s use. Currently, three such facilities are operational (Winchell et al., 2021).

Bamdad et al. (2022) examined PFAS in biosolids generated from three U.S. water resource recovery facilities that underwent pyrolysis in a bench-scale pyrolysis reactor operating at 500° to 700°C. The reactor consisted of a mechanically fluidized vessel with a feed rate of 25 g/min. The incoming dried biosolids were tested using Method 537 for 29 PFAS compounds, with 9 detected. The average PFAS concentration of the biosolids was 36.7 ppb; the PFAS compound with the highest initial concentration was PFOS at 13.7, 16.6, and 26.6 ppb for samples from each water resource recovery facility, respectively. The generated biosolids, biooil (sample 3 only), and pyrogas (sample 3 only) were all tested post-pyrolysis for the same compounds. No PFOS was detected in any of the resulting products, although detection limits were not presented. Pyrolysis at 500°C was found to drive the PFAS compounds out of the solid phase to below detectable limits, but 72.9% by weight of influent PFAS were detected in the pyrogas, mostly in the form of PFHxA. At 700°C, pyrolysis destroyed 88% (by weight) of the influent PFAS and only 12% was found in the pyrogas. Overall, 10 PFAS compounds were detected in the products, including PFHxA, PFHxS, and PFOA. PFOA was the only chemical detected in the products but not the input material. It is unclear if the detected PFOA was generated during pyrolysis from precursors (such as sulfonamide substances) or increased by concentration, resulting from pyrolysis of the sample. This test showed an additional PFAS control device for the pyrogas stream may be needed depending on the temperature at which the pyrolysis occurs and the ultimate use of the pyrogas.

Pyrolysis and gasification results from both full scale and laboratory scale have found that PFAS with a functional group can be readily treated. However, due to the lack of hydrogen and oxygen balance in pyrolysis and gasification, it is likely to lead to significant amounts of PIDs.

4. SUMMARY, CONCLUSIONS, AND RESEARCH NEEDS

Demonstrating the destruction of PFAS materials via thermal treatment is important for several reasons: (1) PFAS are commonly found in thermal treatment feeds; (2) the strong C-F bonds in PFAS result in environmental persistence and require high amounts of energy to cleave; and (3) combustion and incineration may lead to formation of PICs that may also be environmentally harmful. However, demonstrating PFAS destruction via thermal treatment is challenging due to: (1) the large number and variability of PFAS present in our waste, (2) the high variability of waste feeds and operating conditions in thermal systems, and (3) the limited standard measurement methods available have resulted in few thermal systems having been subjected to measurements to characterize PFAS destruction and emissions of PFAS byproducts.

Thus, to date, very few studies have been published, especially those that include a characterization of the air emissions due primarily to that lack of available methods to characterize PFAS destruction and the formation and emissions of fluorinated PICs. High destruction rates of PFAS media sampled have been reported; however, sufficient data do not exist to support the conclusion of nearly complete mineralization of PFAS by existing commercially available thermal treatment technologies.

The temperature and residence time for waste in thermal treatment can be favorable to the destruction of PFAS, but the possibility for PFAS to escape destruction and be released into the air or process byproduct streams cannot be ruled out completely. Properly operating HWIs that use supplemental fossil fuels to maintain optimum combustion conditions (high temperature, sufficient gas-phase residence time, reactant mixing, and sources of excess hydrogen) and MWCs consistently operating above 1000°C likely promote nearly complete mineralization of PFAS (formation of CO₂ and HF). However, non-optimum operations of these systems, and other thermal technologies that may be designed to operate at lower temperatures (<1000°C), such as some SSIs, may not provide a thermal environment that promotes complete PFAS mineralization. TOs, depending on thermal conditions and waste feed composition, may or may not provide adequate conditions leading to complete mineralization of PFAS. Concerns exist about the emissions of PICs from incinerators.

Computational chemistry studies through quantum chemical simulations have given insight into how the initial thermal decomposition of PFCA and PFSA may occur. Experimental results at the laboratory scale have confirmed how PFCA and PFSA decomposes at elevated temperatures; however, they imply that fluorinated PICs will readily form when hydrogen is present. When excess hydrogen (through water vapor) and O₂ are present, experimental results indicate that temperatures above 900°C can cause the complete mineralization of PFOS. Although a notable limitation is that existing mechanistic studies do not adequately represent full-scale treatment systems, which have operational variations significantly different from bench-scale systems. Knowledge about the efficiency of full-scale thermal treatment processes for both PFAS degradation and mineralization is also lacking.

Research using a pilot-scale furnace and a legacy AFFF concluded injection temperatures above 1090°C resulted in high PFAS destruction efficiencies and few detectable fluorinated PIC emissions. However, several operating conditions below 1000°C resulted in destruction efficiencies >99.99% and mg/m³ emissions of several nonpolar PFAS PICs, including many 1H-perfluoroalkanes. The PIC analysis was limited to 30 nonpolar volatile PFAS for which calibration standards were available and should not be construed to be a comprehensive evaluation of all PFAS PICs.

Thermal treatment of PFAS materials is known to generate PICs; however, minimal measurements of PICs have been made at full-scale processes. The type and amount of PIC formation depends on the starting PFAS-containing material, specific PFAS compound(s) present, operating temperature, and treatment method. In general, destruction is often measured by conducting a fluorine balance to compare the input fluorine (within PFAS) to the output fluorine. The output fluorine may be contained within undegraded PFAS, HF, PICs, or other chemical compounds or it may be contained within other streams (i.e., scrubber wastewater) or adsorbed onto reactor surfaces for which PFAS levels have not been adequately quantified. The number of PFAS compounds and the different phases PFAS can partition make determining the complete fluorine balance very difficult. For example, a study may determine a high level of DE for the original PFAS compounds in contaminated soil when compared to post-treatment soil, but not determine the full fluorine balance because of a lack of testing for PICs or air emissions.

Through this review, we have found the following major observations and gaps in data and information regarding PFAS control and destruction in thermal waste treatment:

1. Based on general empirical evidence, PFAS compounds with functional groups such as PFOA, branched PFOA, and HPFO-DA are classified as Class C 3 to 5 POHCs and can be effectively destroyed (~99.99%) in thermal treatment systems at temperatures above 850°C with a residence time of at least 2 seconds in the combustion unit/reactor. Although the destruction of these types of PFAS is achievable, it does not consider the formation of byproducts such as PICs. Therefore, without considering the complete mineralization of these PFAS compounds, the incinerability index is inadequate for establishing the effectiveness of thermal treatment.

2. Experimental evidence suggests high PFAS DEs (>99.99%) can be associated with measurable PFAS PIC emissions. This indicates PFAS can lose their original molecular identity but not be fully mineralized to CO₂ and HF. However, further study is needed to develop methods to identify PFAS PICs.

3. Minimal data exist on the concentrations of PFAS in gaseous emissions from thermal treatment. The data that are available are not comprehensive, as they are limited to characterizations of target PFAS compounds for select feedstock and technology types.

4. Further research is needed to examine the formation and measurement of PICs in full-scale treatment facilities.

To improve the understanding of the fate of PFAS during thermal treatment, we would encourage the scientific community to pursue:

1. Joint experimental and computational studies of the combustion of relevant PFAS both individually and in mixtures. There is limited information on PFAS decomposition pathways under high-temperature oxidizing conditions, and computational mechanism development is needed to accompany experimental studies. Experimental studies under representative conditions are also needed to provide guidance for full-scale performance testing. These experimental studies should focus on using a full complement of analytical techniques to fully assess the potential sinks for fluorine while determining fluorine mass balances. The combined experimental and computation studies may lead to a chemical kinetics model and computational fluid dynamics studies to model the full-scale processes.

2. Development and validation of instrumentation, sampling, and analytical methods capable of measuring PFAS decomposition products and PICs. Although progress is being made in this area for PFAS compounds, measurement of decomposition products lags and is required for validation of conditions for PFAS mineralization.