Fate of perfluoroalkyl and polyfluoroalkyl substances (PFAS) through two full‐scale wastewater sludge incinerators

Perfluoroalkyl and polyfluoroalkyl substances (PFAS) are an emerging issue in wastewater treatment. High‐temperature thermal processes, incineration being time‐tested, offer the opportunity to destroy and change the composition of PFAS. The fate of PFAS has been documented through wastewater sludge incinerators, including a multiple hearth furnace (MHF) and a fluidized bed furnace (FBF). The dewatered wastewater sludge feedstock averaged 247‐ and 1280‐μmol targeted PFAS per sample run in MHF and FBF feed, respectively. Stack emissions (reportable for all targeted PFAS from MHF only) averaged 5% of that value with shorter alkyl chain compounds comprising the majority of the targeted PFAS. Wet scrubber water streams accumulated nonpolar fluorinated organics from the furnace exhaust with an average of 0.740‐ and 0.114‐mol Fˉ per sample run, for the MHF and FBF, respectively. Simple alkane PFAS measured at the stack represented 0.5%–4.5% of the total estimated facility greenhouse gas emissions.


INTRODUCTION
Perfluoroalkyl and polyfluoroalkyl substances (PFAS) have become an important environmental contaminant of concern.Scientists have well documented these chemicals persist in the environment and, given their broad use, have contaminated far reaches of the globe (Buck et al., 2011).Mass production of PFAS began in the 1950s (Kissa, 2001), and governments have since implemented bans on specific compounds (UNEP, 2009;UNEP, 2018;UNEP, 2019) and manufacturers voluntarily discontinuing PFAS production (3M News Center, 2022).The properties that make PFAS attractive for anthropogenic use also cause them to persist in the environment.Thus, these "forever chemicals"-both those that have been discontinued and their contemporary replacements-will continue to be present indefinitely (Ghisi et al., 2019).
Not surprisingly, the pervasiveness of PFAS contamination impacts all water cycles, including wastewater and the solids produced during treatment (Winchell et al., 2022).The solids-historically referred to as sludge and if treated and reused beneficially referred to as biosolids-are collected at water resource recovery facilities (WRRFs).In the United States, sludge or biosolids are typically land applied (43%), landfilled (42%), and incinerated (14%), along with other minor practices for disposition of the remaining percentages (USEPA, 2021b).Documentation of agricultural crop (Ghisi et al., 2019;Zhang et al., 2020) and groundwater (Navarro et al., 2018;Pepper et al., 2023) contamination related to land application and landfill releases (Ahrens et al., 2011;Harrad et al., 2020) have placed pressure on these methods of sludge and biosolids disposition.Thus, the potential for wastewater sludge incinerators (WSIs) to address PFAS contamination has generated interest into whether they provide a viable solution for destruction of these compounds or whether they exacerbate the problem through incomplete mineralization (Winchell, Ross, et al., 2021).
PFAS can be thermally degraded.Winchell, Ross, et al. (2021) summarized literature available on the topic and identified the conditions required for PFAS destruction, that is, mineralization to thermodynamic endpoints of carbon dioxide (CO 2 ), water vapor (H 2 O), hydrogen fluoride (HF), and other compounds depending on the makeup of PFAS in feedstocks.In some cases, the PFAS destruction guidelines, and often citations of them, have focused on temperature, typically quoted as 1000 C or greater, as the key controlling factor for destruction, without mention of the other critical combustion efficiency parameters like residence time and turbulence (Winchell, Ross, et al., 2021).Winchell, Ross, et al. (2021) made a case that the often-superior conditions of enhanced time and turbulence in WSIs (compared with other types of incineration) may overcome the lower operating temperatures (700-1000 C) in combustion areas.If PFAS compounds are not completely destroyed, products of incomplete combustion (PIC) may include shorter alkyl chain versions of the longer chains that are anticipated to pose a lower toxicological concern (ASTDR, 2021).Moreover, air pollution control (APC) equipment used with WSIs may remove PFAS from the flue gas and limit direct emissions to the environment.Thus, while WSIs are expected to degrade PFAS, the destruction and removal efficiency (DRE) and PIC generation potential have not been documented.Seay et al. (2023) attempted to develop a "PFAS" DRE for a WSI, but data used for the calculation suffered from laboratory quality control issues.Beyond that, Loganathan et al. (2007) sampled dewatered wastewater sludge and ash from a WSI, observing 26%-97% reduction of six targeted PFAS, while two other compounds increased in concentration.A third study, led by the United States Environmental Protection Agency (USEPA), is currently underway but has yet to be published (Potter, 2022).Finally, a fourth study was recently presented examining laboratory-scale results of PFAS within a WSI technology being offered in the market, although with limited installations (Orr et al., 2023).In this laboratory study, researchers observed a 99.9% DRE for four PFAS, but at temperatures of 1150 C for 4 s, which does not compare well with existing WSI installations in the United States (Winchell, Ross, et al., 2021).
Given the lack of published data regarding the fate of PFAS through WSIs, this research evaluated the fate of PFAS through two full-scale WSI facilities.Both a multiple hearth furnace (MHF) and fluidized bed furnace (FBF) WSIs were sampled to track PFAS using a state-of-the-art sampling and analytical program with the objective of elucidating PFAS destruction capabilities of existing WSIs to support the industry and scientific community with information that could be used to develop policy and regulatory plans and actions.

METHODOLOGY
This project characterized the fate of PFAS, to the extent practical, through furnace technologies currently employed in the US municipal wastewater industry.

Sampling sites
One MHF and one FBF WSI were sampled; configurations of these incinerator installations are depicted in Figures S1 and S2, respectively.Additional details are provided on the facilities in the Supporting Information (SI), Section S1.
The facilities selected for this study represented installations classified as "existing" and are therefore not subjected to more stringent emissions limits of "new" facilities (Standards of Performance for New Stationary Sources and Emission Guidelines for Existing Sources, 2011).Facilities classified as "new" require additional APC equipment to meet the stricter emission limits, often including activated carbon to remove mercury, which would potentially remove PFAS, like use of activated carbon for drinking water treatment applications (USEPA, 2019).As such, facilities with more advanced APC can expect greater PFAS removal than measured here.
At each site, all major inputs and outputs (e.g., dewatered wastewater sludge, combustion air, ash, flue gas, scrubber water supply, and drain) were sampled in triplicate (Figure 1).Refer to SI Section S2 for a more in-depth description of the individual sample locations.

General analytical approach
A comprehensive analytical workflow that aspired to achieve a mass balance of PFAS across gaseous, solid, and aqueous matrices encountered in the two types of WSIs was designed.The sampling and analytical workflow were implemented, and while a strict mass balance is not possible with current analytical techniques (Winchell, Wells, et al., 2021), a mass balance approach was used to assess targeted, quantifiable PFAS listed in SI Sections S3.3 and S3.4.PFAS can differ in polarity (polar or nonpolar), charge state (neutral, anionic, cationic, or zwitterionic), and volatility (volatile, semivolatile, or nonvolatile) (Winchell, Ross, et al., 2021).To most fully characterize fate of PFAS in these systems, in addition to well-accepted methods for polar PFAS, it was necessary to address challenges with currently available analytical methodologies, which were used to improve the recovery of nonpolar PFAS (Phelps & Ryan, 2021) in this research based on the principles of separation science (National Academies of Sciences, Engineering, and Medicine, 2019).Thus, the project team developed tailored separation schemes for each of the matrices (air, water, solids) encountered in WSIs (Figures S3-S6).
For this study, two generalized sample preparation schemes-polar and nonpolar-were used to capture PFAS to the extent practical.Subsequent analyses included targeted and total organic fluorine (TOF) using combustion ion chromatography (CIC).TOF generically refers to adsorbable organic fluorine (AOF) or extractable organic fluorine (EOF), separately applied depending on the sample preparation.In situ Fourier transform infrared spectrometry (FTIR) was also used to measure fluorinated greenhouse gases within the stack gas.Nontargeted analysis (NTA), as described in Winchell, Wells, et al. (2021), was also deployed to identify unknown PFAS precursors and PICs as part of this project, as suggested in Figures S3-S6, but the results are presented separately by this research group.

Sampling, sample preparation, and analysis
On-site sampling for this study was supervised by operations personnel from each test site with project investigators.Two specialty subcontractors-Eurofins Test America (ETA), Knoxville, TN, and Alliance Source Testing (AST), Carrollton, TX-assembled specialized sampling equipment and collected the samples.

Source air emissions (stack)
The source air (stack) emissions sample apparatus was assembled (Figure S7) according to the "Other Test Method 45 (OTM-45) Measurement of Selected Per-and Polyfluorinated Alkyl Substances from Stationary Sources" (USEPA, 2021a).Source air sampling methodology is presented in Table S2.Refer to SI Section S3.1 for additional details on the four fractions-front half, back half, impinger condensate, and breakthrough XAD resin-collected from the sampling train.Analytical results of each fraction are summed to represent the stack PFAS levels.
Samples collected by the OTM-45 sample train were extracted using an OTM-45/Modified M0010 sequential extraction procedure.OTM-45, which is designed for extraction of polar PFAS only, was modified to capture additional PFAS by implementing a sequential extraction of nonpolar and polar analytes and a parallel extraction of nonpolar analytes only (Figure S3).Surrogate and isotope dilution standards were spiked into samples before laboratory extraction.Field rinses of the sample train components were performed with methylene chloride (CH 2 Cl 2 ) followed separately by ammonium hydroxide (NH 4 OH) in methanol (CH 3 OH) and were collected in separate sample containers.Each rinse was subsequently combined with the extract of like solvent.For the liquid impinger samples, nonpolar analytes were extracted first with CH 2 Cl 2 ; then polar analytes were extracted with a mixed mode solid phase extraction/weak anion exchange (SPE/WAX) sorbent.Samples, with exception of impinger condensates, were sequentially extracted with CH 2 Cl 2 followed by an NH 4 OH in CH 3 OH solution.
Sequentially extracted samples were generated by combining an aliquot of the initial CH 2 Cl 2 extract after undergoing a solvent exchange and the secondary CH 3 OH/NH 4 OH extract (except for liquid impinger fractions for which polar extraction was performed using SPE/WAX).The combined sequential extract was treated to precipitate inorganic fluoride and analyzed by CIC for TOF (AOF or EOF as appropriate).Figure S3 outlines the sample preparation scheme for the modified OTM-45 method applied in this research.

Combustion air
The combustion air sampling methodology is presented in Table S2 and described in SI Section S3.2.An XAD sorbent was used for collecting the combustion intake air sample (Figure S8).Samples were collected using the same apparatus and sample preparation as used for the breakthrough XAD module in the source air emissions (stack) samples (Figure S7).

Dewatered wastewater sludge or bottom ash
Grab samples of dewatered wastewater sludge and bottom ash (MHF only) were collected.The FBF dewatered wastewater sludge was collected prior to injection of lubrication water, as described in the SI.Grab samples were composited over the 4-h duration of the flue gas sampling.Equal volumes of the sample matrix were collected at the initiation of the sample run, and every 30 min thereafter.Individual grab samples were combined and homogenized.Subsequently, aliquots were packed with ice and transported to various laboratories for analyses.
Both the dewatered wastewater sludge and bottom ash contained no filterable liquid phase.These two types of samples were processed in the same manner (Figure S5).Surrogate and isotope dilution standards were spiked into samples prior to extraction.
An aliquot of raw dewatered wastewater sludge was collected for total solids (TS), volatile solids (VS), and ultimate analyses for carbon, hydrogen, nitrogen, oxygen, and sulfur composition.Raw samples were subjected to parallel extraction procedures: a nonpolar CH 2 Cl 2 extraction and a polar, basic, CH 3 OH extraction using a 0.4% potassium hydroxide (KOH), and CH 3 OH solution subjected to shaking and sonication with final pH adjustment.The polar extract was cleaned using SPE/WAX.PFAS analytes were eluted using 0.3% NH 4 OH and CH 3 OH solution.
The polar extracts were analyzed chromatographically by polar targeted and polar NTA.The nonpolar extracts were analyzed by nonpolar targeted and nonpolar NTA.A portion of each of the polar and nonpolar extracts were treated separately with calcium hydroxide (Ca(OH) 2 ) to precipitate fluoride and processed for TOF (AOF or EOF as appropriate).

Wet scrubber supply or drain
Grab samples of the wet scrubber supply and drain were collected in a similar manner (every 30 min and composited) as dewatered wastewater sludge and bottom ash.Notably, wet scrubber supply water source was undisinfected treated effluent for the FBF WSI, and effluent disinfected with sodium hypochlorite for the MHF WSI.
The wet scrubber supply and drain samples were each composed of two-phase samples, liquids and solids, which required separate sample preparation schemes.The generalized procedure is presented in Figure S6.Aliquots were separated for total suspended solids (TSS) and inorganic fluoride analysis.Samples were filtered to separate the liquid and solid phases.Surrogate and isotope dilution standards were spiked into samples prior to extraction.
Filtered liquid samples were subjected to a parallel extraction procedure.An aliquot of the CH 2 Cl 2 nonpolar extract was analyzed for nonpolar targeted compounds.A second aliquot of the nonpolar extract underwent additional processing with Ca(OH) 2 to precipitate inorganic fluoride prior to TOF analysis.Filtered liquid samples were processed by SPE/WAX; the analyte was eluted using 0.3% NH 4 OH and CH 3 OH solution.An aliquot of the resulting polar extract was analyzed by polar targeted and NTA.A second aliquot of the polar extract was further processed with Ca(OH) 2 to remove inorganic fluoride prior TOF analysis.
Solid-phase sample preparation employed the same parallel nonpolar/polar extraction for the filtered solids as the liquid filtrate, except the polar extraction used a 0.4% KOH and CH 3 OH solution with shaking and sonication with a final pH adjustment.An SPE/WAX cleanup was used on extracts as previously described, and sample aliquots were analyzed identically to the liquid phase samples.

Analytical methods
At the time this project was proposed (Fall 2020) and funded (Spring 2021), the USEPA had promulgated two methods (537.1 and 533; Shoemaker & Tettenhorst, 2020) for quantifying PFAS, mainly polar compounds, in drinking water.However, as the project progressed, new draft methods emerged: USEPA Draft Method 1633 (to test for 40 PFAS compounds in wastewater, surface water, groundwater, soil, biosolids, sediment, landfill leachate, and fish tissue), first proposed in August 2021, is currently published (USEPA, 2024a).Also, the USEPA Draft Method 1621 (for Adsorbable Organic Fluorine) was not published until April 2022 (USEPA, 2022) and now finalized (USEPA, 2024b).Consequently, PFAS analytical knowledge progressed in parallel with this project.
Furthermore, some of the analytical methods applied in this research were published and considered standard while the others were developed specifically for this project, as described herein.The following identifies the standard methods, modifications thereof, or emerging techniques utilized in this research.

Targeted-polar
Quantitative targeted polar analyses utilized known reference standards, based on standard regulatory methods (Shoemaker & Tettenhorst, 2020) and extended methods established by the analytical vendor referred to here as Method 537 (modified).The list of PFAS examined is presented in Table S3.Refer to this table for the names and acronyms of individual PFAS compounds and the PFAS families to which they belong, hereafter referred to by their acronyms (SI Section 3.3).

Targeted-nonpolar
Five targeted nonpolar fluorotelomers (FTOH) known to be perfluorinated alkyl acid (PFAA) precursors were investigated in the nonpolar CH 2 Cl 2 extracts from each site using gas chromatography/mass spectrometry/mass spectrometry (GC/MS/MS) in the selected ion monitoring (SIM) mode; see Table S4 (SI Section S3.4).The FTOH compounds were analyzed using a method developed by the analytical vendor guided by the regulatory method USEPA SW-846 8270E: semivolatile organic compounds by gas chromatography/mass spectrometry (GC/MS), which has provisions allowing use of triple quad (MS/MS) technology (USEPA, 2018).TOF TOF analyses are important in revealing whether unidentified PFAS are present.TOF is a screening tool intended to elucidate the total amount of PFAS in a sample by comparison with standardized targeted analyses.Similar results between TOF and targeted analyses indicate absence of substantial concentrations of precursors, whereas divergence in results suggests that precursors, undetectable by standard quantitative analyses, are present and important in an overall PFAS mass balance.Analysis using the TOF assay is recommended if the PFAS product composition is unknown, where known PFAS composition extends beyond the list of compounds in the standard targeted analytes, where there is likely to be transformation of PFAS, or where precursors are unknown (Winchell, Wells, et al., 2021).For TOF analysis and calculated results, refer to SI Section S3.5.
For this study, organic fluorine analysis are generically referred to as TOF unless reference material from elsewhere is noted.

FTIR
FTIR is a spectroscopic method used to measure fluorinated greenhouse gases within the stack gas.A discussion of FTIR methodology is provided in SI Section S3.1.

Fluoride
Two methods for measuring fluoride were used.Wet scrubber supply and drain water filtrate samples were analyzed using USEPA SW-846 Test Method 9056A (USEPA, 2007) for fluoride characterization.Dewatered wastewater sludge samples collected after the primary sampling campaigns were analyzed for fluoride by USEPA Method 300.0 Part A (USEPA, 1993).

Solids characteristics
The dewatered wastewater sludge and solids from the wet scrubber water streams underwent solids characterization analyses as described in SI Section S3.6.The dewatered wastewater sludge was analyzed for moisture and ash content using ASTM-D482 (ASTM, 2019), sulfur using ASTM-D129 (ASTM, 2018), and ultimate composition per laboratory developed methods using combustion analysis (carbon, hydrogen, nitrogen) and pyrolysis product characterization (oxygen).TSS was measured in all wet scrubber supply and drain samples by weighing the filter used before and after filtering the sample.

Sampling events
Both sites were sampled within 4 months of each other.The MHF, sampled in April of 2022 (Table S5), followed the FBF, sampled in December of 2021 (Table S6).Details of the sampling events and samples collected are provided in SI Section S3.7.

Data quality standards
Project analytical data were assessed for data quality in a four-step process outlined in Table 1.The final step, data usability, is critical for advancing scientific knowledge based on sound information.The researchers established judgment criteria on which data usability was assessed (Table S7).This project screened data based on qualifications made by the laboratories to have the highest confidence in the results presented.SI Section S3.8 contains more details on the treatment of the data and basis for the screening approach employed.

WSI operating conditions
The fate of PFAS through WSIs is dictated by system operating conditions.Primary conditions dictating destruction or degradation in combustion systems include temperature, residence time, turbulence, and oxygen availability (WEF, 2009).Temperature and oxygen are typically measured in WSIs, but residence time and turbulence are not.While turbulence is difficult to quantify in these systems, residence time can be calculated.
Heat and materials balances for each WSI were developed based on the conditions recorded by permanently installed instrumentation and equipment dimensions (Tables S8-S21).Balances were created for each triplicate sampling run (Figures S9-S14).Refer to SI Section S4.1 for details on development of the balances.
The main driver for developing heat and materials balances was to assess parameters required for PFAS mass balances along with the furnace residence time.For the MHF, the wet scrubber drain flow, combustion air flow, and furnace gas-phase residence times are not directly measured.Likewise, for the FBF, the wet scrubber drain flow and furnace gas-phase residence times were calculated.
The quantifiable temperatures and residence times PFAS experienced through the MHF furnace are presented in Table 2; similar thermal treatment equipment does not exist in the WSI train and thus are not presented.Briefly, an MHF consists of a cylindrical shell with several levels or hearths where combustion occurs.Winchell, Ross, et al. (2021) provided a more detailed description of MHF technology.
Turbulence is not characterized in Table 2 and not easily quantified.From a qualitative perspective, turbulence in an MHF consists of frequent turning of the solids, dewatered wastewater sludge or ash, by the rabble arms and the flow of various gases through the furnace.While turbulence of the gas phases can be evaluated by computational fluid dynamics, this effort was not conducted as part of this research.Additional information on characterizing turbulence within a furnace can be found elsewhere (Baukal et al., 2001).Winchell, Ross, et al. (2021) provided a qualitative turbulence comparison of MHF and FBF, amongst other furnace technologies.
Solids residence times in MHFs depend on the furnace radius, number of rabble arms, center shaft speed, and rabble arm plow configuration.These details were not readily available for the MHF sampled.General design information for this style of furnace suggests the detention time on hearths with four rabble arms is approximately 6.5 min and 13.5 min with two rabble arms operating at one revolution per minute (Zimpro, 1977).Other sources suggest a 10-min retention T A B L E 1 Analytical data management summary compiled from USEPA ( 2002) and ITRC (2022a and 2022b).

Verification of data
The process of evaluating the completeness, correctness, and conformance/compliance of a specific data set against the method, procedural, or contractual requirements.

Validation of data
A formal analyte/sample specific review process that extends beyond verification to determine the analytical quality of a specific data set.

Data quality
Indicators of precision, accuracy/bias, representativeness, comparability, completeness, and sensitivity are defined.Data integrity, unambiguity, consistency, completeness, and correctness are established.

Data usability
Data usability is determined by the project team after verification, validation, or any other data quality review is complete, and the overall quality of the collected data is known.Simply, a data usability assessment determines whether or not the quality of the analytical data is fit for its intended use.
time on hearths with four rabble arms and 20 min with two rabble arms (Zimpro, n.d.).Overall, the solids residence time in the furnace could range from 80 to 120 min.
The conditions expected to thermally degrade PFAS within the FBF are summarized in Table 3, and like the MHF, no other thermal treatment equipment was present in the WSI train.Briefly, an FBF consists of two combustion zones-the sand bed and the freeboard.Dewatered wastewater sludge is injected into the fluidized sand bed where primary combustion occurs, and any volatized material escaping to the overlying open space, or freeboard, will combust there.Winchell, Ross, et al. (2021) provided a more detailed description of FBF technology.
Quantifying turbulence within the FBF sand bed or freeboard was not part of this research.On a qualitative basis, the turbulence within an FBF exceeds other furnace technologies, including an MHF (Winchell, Ross, et al., 2021).This characteristic should result in greater PFAS destruction or degradation, comparatively.
Solid or gas residence time within the sand bed cannot easily be quantified.Estimated solids residence times are less than 1 min before all combustible materials have reacted or volatilized (Winchell, Ross, et al., 2021).A gas residence time estimate requires an assumption on where material becomes volatile, and the bulk flow rate depends on yet another assumption regarding where combustion occurs.The sand bed gas phase residence times were not estimated for the sampled system, but general design approaches assume a nominal 2-s residence time for design.Whether the volatilized PFAS stay within the sand bed for this duration cannot be stated with certainty, but at a minimum, the compounds will be subject to the entire freeboard, given all material entering an FBF must pass through that area.

RESULTS
Results from analyses of the samples collected from the two WSIs and the derived operating conditions using the site installed instrumentation are presented in this section.Data presented here are those that have passed the data quality review process and are deemed usable for interpretive purposes.Refer to SI Section S5.2 (Tables S22-S50) for the results of the data quality exercise identifying the final disposition of each measurement recorded and PFAS naming conventions (Tables S3  and S4).

Source PFAS and fluorine based compounds
Potential sources of PFAS to an incineration system primarily consist of the combustion materials and water supplied to the wet scrubber.Combustion materials include dewatered wastewater sludge and combustion air.Table 4 contains reportable results from the sample analyses and the following sections provide interpretation of the PFAS and related results for each WSI site.

Dewatered wastewater sludge
The dewatered wastewater sludge serves as the primary feedstock to the furnace and often has PFAS contamination (Thompson et al., 2022).Both WRRF sites yielded reportable values of PFAS.

MHF
The MHF site provided a preliminary dewatered wastewater sludge-targeted PFAS analysis obtained from a separate research effort.Sampling and operating condition details other than PFAS results were not disclosed for this site.

WATER ENVIRONMENT RESEARCH
Each polar PFAS family evaluated (Table S3) contained at least one reportable compound, except for compounds in the ether families (PFECA and PFESA) when analyzed.Six of the 13 PFCAs, two of the eight PFSAs, and two of the four FTSA compounds were reported in at least one of the three runs; with fewer reportable values in the remaining PFAS families analyzed.The field duplicate of Run 2 and the parent Run 2 sample yielded similar results.All reportable PFAS in the parent sample were also reported in the field duplicate, except PFNA, which was also reported in the latter.Differences between the parent and field duplicate ranged from three to 81%, with only PFDS and NMeFOSA exceeding 43% difference in concentrations (76% and 81%, respectively), and no compound differing by more than approximately 2 ng/dry g.None of the samples contained reportable levels of TOF or the nonpolar FTOHs.In total, the targeted PFAS in the dewatered wastewater sludge averaged 247 μmol over the three sample runs.
Fluoride was measured in the dewatered wastewater sludge feed to the MHF facility over a 7-day sampling campaign from May 3 to 9, 2023, using EPA Method 300.0 Part A, after the main sampling campaign in April 2022.The average dewatered wastewater sludge fluoride concentration in seven samples run in duplicate was 46 mg/dry kg.However, this value is suspected to represent the water-soluble content of the dewatered wastewater sludge rather than the total fluoride.Sample preparation for EPA Method 300.0 Part A calls for mixing a solid sample with distilled (DI) water prior to filtering and analyzing via ion chromatography.As such, the quality control reference material used by the laboratory allowed for acceptance limits 10 times lower than the certified value, with the mean recovery reported by 52 different laboratories at 13.5% of the certified value.

FBF
Preliminary and primary sampling event dewatered wastewater sludge samples were collected from the FBF WRRF, both of which are reported in Table 4. WRRF personnel collected a set of three preliminary samples in July 2021 prior to initiation of the main project, which were analyzed for targeted analysis alone.In all samples, PFOS was again a dominant PFAS contaminant, as was seen in the MHF samples; however, in the primary samples, several FASAA and FASE compounds were also measured in comparable molar amounts.
Each polar PFAS family targeted, except the ethers, was represented by a reportable compound during the December 2021 sampling event.Ten of the 13 PFCAs analyzed produced reportable values, but PFTrDA and PFHxDA only produced reportable results for a single run during the main sampling event.Half of the PFSA compounds analyzed were reportable in the preliminary samples collected in July 2021, while PFHxS and PFDoS were unreportable in the main sampling runs conducted December 2021.All the FASA, FASAA, and FASE compounds analyzed produced reportable results except NMe-FOSA in the July 2021 samples.The field duplicate of Run 2 from the primary samples yielded similar results with all compound detections being shared with the parent sample.Differences in the Run 2 parent and field duplicate samples ranged from zero to 33% with most results differing by less than 2 ng/dry g except NEtFOSE at just over 5 ng/dry g difference.Neither sample consistently reported higher values than the other.Like the MHF, no reportable TOF or FTOHs were observed in any samples.The total targeted PFAS averaged 1280 μmol over the three runs.

Combustion air
No reportable targeted PFAS were reported in combustion air samples, collected at both sites.TOF analyses resulted in reportable values for each site and in all sampling runs.The units for TOF refer to fluoride concentrations, not PFAS.Because no reportable targeted PFAS were detected in the combustion air, none of the TOF can be described by specific PFAS.Interestingly, the combustion air contained roughly the same concentration of TOF at both sites.The combustion air sample locations were outside of the building housing the WSI process.Thus, the sampled air reflected the general atmosphere around the WRRF at both sites, albeit at a single point on the facility grounds.

Wet scrubber supply water
The wet scrubbers at both sites used treated wastewater effluent as a water supply, where it was discharged or sprayed into various sections of the wet scrubber to contact flue gases, capture particulates, and neutralizes acid gases.Treated effluent from WRRFs has been documented to contain PFAS (Thompson et al., 2022).When used to supply a wet scrubber, this treated water acts as another source of PFAS to the incineration mass balance.
In the MHF wet scrubber supply water, PFOS was the highest reportable compound (mass and molar), followed by PFHxA (Table 4).The PFAS content of the wet scrubber water supply shifts to the short carbon chain analogs compared with the dewatered wastewater sludge.Solids from the wet scrubber supply water contained no reportable PFAS, except one result in the third run, which had a reportable concentration of 0.339 ng/g dry PFOS (reporting limit 0.167 ng/g dry).
The FBF wet scrubber water supply differed from the MHF in that PFBS had the highest concentration (mass and molar) instead of PFOS.Otherwise, the PFAS concentrations reported resembled those from the MHF site except the PFOS concentration was roughly one-third.Further, the wet scrubber supply water PFAS content also favored smaller chain compounds compared with the dewatered wastewater sludge.The solids from the wet scrubber supply water contained no reportable PFAS except for the third run, which produced 0.151 ng/g dry 6:2 FTS (reporting limit 0.108 ng/g dry).
From the MHF samples, the only sample extract resulting in reportable TOF results was the nonpolar filtrate.None of the FTOH compounds were reportable, indicating that the nonpolar TOF is indicative of other forms of fluorinated organics.The only reportable value of TOF from the wet scrubber supply samples from the FBF occurred in Run 1, with a nonpolar filtrate (2.65 μg/L fluoride).
Inorganic fluoride in the wet scrubber supply water filtrate was not reportable for either MHF or FBF samples.The values were above the method detection limit but less than the reporting limit.In the following discussion, regarding the amount of fluoride gained across the wet scrubber, the outcome does not change if the values in the wet scrubber supply assume the value reported by the commercial laboratory.
The FBF site utilizes a minimal amount of water to alleviate high pressures in the dewatered wastewater sludge pumping system feeding the furnace.Lubrication water comes from the same source as wet scrubber water supply.All data presented for the FBF wet scrubber supply water also applies the lubrication water.The MHF system does not use any lubrication water.

Emissions of PFAS and fluorine based compounds
Potential emissions of PFAS from an incineration system generally include air emissions from the stack and the ash.In the case of this investigation, the wet scrubber drain also serves as a potential PFAS outlet, though the drain is recycled back to the head of the wastewater treatment process.Table 5 contains all the reportable emissions for both sites.

Bottom ash
Only the MHF produces bottom ash, which is handled separately as it leaves the incineration system.The FBF pushes all the ash with the flue gas to the wet scrubber where it is captured and flows out in the drain as a slurry.
The MHF bottom ash contained no reportable quantities of targeted PFAS or TOF.

Wet scrubber drain
At both sites, water sent to the wet scrubber exits via a drain at the bottom of the vessel after treating the flue gas.Water vapor in the flue gas also condenses but adds less than 3% of flow in the drain.The composition of the wet scrubber drain includes particulates captured and constituents dissolved from the flue gas, which can lead to a drop in pH.The researchers' experience at several WSI facilities suggest pH in the water may drop from approximately 7.0 to as low as 2.0.The temperature of the water increases across the wet scrubber as heat transfers from the flue gas.The water temperature may increase by a factor of two in the drain compared with the supply, based on the researchers' experience at other WSI facilities.The pH and temperature changes through the wet scrubber were not evaluated as part of this research.This research does not consider these parameters in depth across the wet scrubber because they are not anticipated to impact changes in PFAS degradation; however, these parameters may impact whether PFAS stays in the flue gas or partitions to the wet scrubber water as discussed later.
The targeted PFAS composition of the wet scrubber drain filtrate largely reflected the supply water as described previously.In contrast, solids, mostly ash, from the wet scrubber drain contained no reportable PFAS, except in the third run, which produced 0.165 ng/g dry PFOS for the MHF and 0.0939 ng/g dry 6:2 FTS for the FBF, correlating with detections in the wet scrubber supply water.The field duplicate of the second run solids also had no reportable PFAS.
The TOF analysis provided interesting results.At both facilities, TOF was reportable in the nonpolar filtrate extract, while the polar filtrate and both solids' extracts did not produce reportable results at either site.The MHF wet scrubber drain collected on average 0.901 mol of fluorine per sample run compared with 0.168 mol of fluorine for the FBF.
Fluoride was reportable in the wet scrubber drain and the load exiting during each sampling event run may suggest partial or complete destruction of fluorinated organics in the furnace.The liberated fluorine is then captured in the wet scrubber and measured in the drain as fluoride.

Source air (stack)
The cleaned flue gas eventually flows out of the stack and to the atmosphere.This represents the single direct     point of environmental discharge from the incineration process.
The multiple sample fractions of the OTM-45 resulted in four reported values (Figure S7).At the MHF facility, the back half and breakthrough fractions produced no reportable results.While the research team anticipated to detect most PFAS in the back half sample, comparatively higher reporting limits, analyte interferences, and laboratory control qualifications resulted in no useable data.Unfortunately, proof blank samples were lost in processing and therefore could not be used to qualify the data.Fortunately, field blank samples were analyzed, with no detectable targeted PFAS.The front half fraction would be more prone to capture particulate bound compounds on the filter, or those that can adhere to the walls of the sample tubing.In these samples, PFOA, PFOS, and 6:2 FTS were present in the front half sample.PFBA was present in front half and impinger samples, but there were higher concentrations in the latter.PFPeA and PFHxA were reportable only in the impingers.Notably, none of the consistently reported PFAS contain more than eight carbons.The molar loads emitted during each run were calculated using the measured stack flow rates.
Samples representing the FBF stack were collected from the duct between the sorbent polymer composite (SPC) module and secondary heat exchanger; see Figure S2.The secondary heat exchanger transfers heat into the flue gas prior to discharging from the stack.Given the secondary heat exchanger increases the flue gas temperature from roughly 19 C to 125 C, there are no anticipated impacts on PFAS composition.Unfortunately, the field and proof blanks were contaminated with the same PFAS measured in the run samples.Following the FBF sampling event, AST and ETA subsequently revised glassware cleaning procedures prior to the successful MHF sampling.This, and other quality control considerations, limited the amount of useable data to intermittent detections of five compounds.Levels of PFBA and PFPeA were reportable from the breakthrough fraction of Run 3, suggesting the relatively small alkyl chain PFAS passes upstream fractions of the OTM-45 train, and some may not have been captured for analysis.Their detection also raises questions on whether the same PFAS profile would have been observed as in the MHF, which favored smaller compounds, if not for sample contamination.The presence of PFUnA and PFDoA in the back half of Run 2 and two instances of HFPO-DA offer little insight, given the limited results.
The TOF analyses did not produce any reportable results for both sites.In all OTM-45 fractions, values were below reporting limits or contamination detected at similar levels in blank samples.
During each sampling run, a continuous stream of flue gas was pulled from the stack and analyzed in real time for tetrafluoromethane (CF 4 ), hexafluoroethane (C 2 F 6 ), and octafluoropropane (C 3 F 8 ).See Table S51 for the method detection limits achieved.The molar load of fluorine from the three analytes was calculated using the measured stack flow rates as shown in Table 5.
The detection of C 3 F 8 in Run 3 of the MHF sampling skews the calculated load higher when compared with the other two runs (runs were measured sequentially).The measured C 3 F 8 during Run 3 exceeded the method detection limit by several parts per million by volume wet (ppm vw ) for most of the sampling run until the measurement stabilized at, or below, the limit for the last 20 min.The other two sampling runs never produced C 3 F 8 above the method detection limit.
The detection of C 2 F 6 in Runs 2 and 3 for the FBF skews the load high compared with Run 1.The measured C 2 F 6 during Run 3 exceeded the method detection limit by several ppm vw for most of the sampling period but dropped to the method detection limit, or below, at different times for several minutes during the last third of the run.Run 2 consistently produced reportable levels of C 2 F 6 .Run 1 never produced C 2 F 6 above the method detection limit.

DISCUSSION
The sampling and analytical programs used for this study produced sufficient data to explore several concepts related to the fate of PFAS through the incineration processes.The following discussion details these concepts.

Targeted PFAS fate
When studying pollutants through a process designed or suspected to result in pollutant reduction from the emissions, the performance is generally characterized from a DRE perspective.However, DRE for PFAS must be carefully evaluated and described to avoid misleading characterizations.SI Section S6.1 and Tables S52-S56 detail the major limitations for publishing DRE values with the data collected for this project.Full-scale PFAS DRE remains indefinable when applying the data usability guidelines referenced here.Confidence in DRE evaluations will require PFAS spiked feedstocks unless a highly contaminated dewatered wastewater sludge feeding a WSI is identified.
Figure 2 depicts the fate of the targeted PFAS through the MHF and FBF, summarizing the data presented in Tables 4 and 5 and represents the nearest approach to DRE discussed here.The values in the figure represent the average load during the three sampling runs of all targeted PFAS summed.Very little of the targeted PFAS associated with the dewatered wastewater sludge leaves the system.Only 12 μmol escape the MHF via the stack, while none leaves the FBF though contamination eliminated consideration of several key PFAS (PFBA, PFPeA, and PFHxA) observed in the MHF stack samples.The PFAS entering the wet scrubber with the water supply essentially remained in that liquid phase and exited via the drain; see following discussion on wet scrubber implications.Combustion air and bottom ash from the MHF did not contain reportable targeted PFAS.As presented, the WSIs reduce the amount of targeted PFAS released to the environment by over 95%.

Targeted mass emissions
A comparison of the three typical WRRF emissions streams (effluent, dewatered wastewater sludge, and air emissions from incineration) is presented here as a perspective on potential PFAS sources to the environment.This approach allows the emissions measured here to be placed into perspective with other release points at WRRFs.Table 6 presents average targeted PFAS emissions; reportable levels were found in all three samples of the wet scrubber supply, dewatered wastewater sludge, and stack.The emissions were normalized as follows.For the wet scrubber supply, which was used to estimate the WRRF effluent characteristics, the calculated emission values universally assumed a flow of 3.785 ML per day (ML/day).The dewatered wastewater sludge calculation used 0.907 dry metric tons per day (DMT/day) converted to a wet basis assuming 23% and 26% TS for the MHF and FBF respectively and based on the quality measured during the sampling runs.The flow and solids production values are roughly equivalent to US customary units of one million gallons per day of flow, which produces roughly one dry ton per day of solids, a general rule of thumb to estimate solids production absent other data in the authors' experience.The stack flows used to estimate the emissions came from data measured for the project, normalized to dry feed rate basis.
No single PFAS compound was reportable in all emissions to compare the two sites.At the MHF site, PFHxA, PFOA, and PFOS all had reportable values, although the latter only had two results that were used to calculate the dewatered wastewater sludge load.The FBF did not emit PFOA or PFOS from the stack like the MHF but may have released PFPeA and PFHxA if not for the sample contamination.In all cases where data can be reported, the emissions load from the effluent exceeded the dewatered wastewater sludge, which in turn exceeded the stack for the targeted PFAS measured in this study.The trend of effluent to dewatered wastewater sludge loads has been observed by others (Schaefer et al., 2023;Tavasoli et al., 2021;Weston & Sampson, 2023).All three studies also showed the tendency of the longer alkyl chain PFAS to preferentially partition to wastewater sludge compared with the primarily liquid effluent, consistent with the MHF and FBF results.
The reportable results in Table 6 for the MHF suggest that PFAS characteristics shift to shorter alkyl chain compounds in the plant effluent or stack compared with the dewatered wastewater sludge.Except for PFHxA, the dewatered wastewater sludge PFAS content consisted of 8-carbon compounds, or greater.In the stack, PFBA and PFPeA represented roughly 80% of the molar PFAS concentrations reported with no compounds over 8-carbon in length.The FBF followed similar trends, but the measurements of the shorter chain PFCA compounds in the stack were not useable due to contamination.Björklund et al. (2023) found PFBA to be the most abundant PFAS detected in stack samples from a municipal solid waste (MSW) incinerator while processing MSW alone or when co-processing dewatered wastewater sludge.Unfortunately, PFBA and PFPeA analyses can be compromised by interferences (Bangma et al., 2023), so results reported here must be viewed with some skepticism until the previously noted nontargeted analyses are completed.Other shorter PFCAs were detected as well, with the general trend agreeing with the MHF results.
The SPC module installed in the FBF train may also remove some PFAS.Certain PFAS may accumulate on the media over time; see Table S57.Refer to SI Section-S6.2 for further discussion.

Air emissions assessment
Several states within the United States have developed air emissions guidelines or screening levels for specific PFAS.Table 7 contains the threshold values applied by these states and compares them to the measured air emissions from the stack of the MHF.For all instances, the MHF stack emissions fell below the most stringent requirements.
A similar table for the FBF was not prepared given there are no consistently reportable values from the stack; see Table 5.Like the MHF, the FBF would also be considered compliant with these emissions requirements but noting the results for PFBA and PFHxA were screened due to contamination.

Wet scrubber impacts
The wet scrubbers play an important role in determining the fate of PFAS within WSIs.Both facilities sampled for this project had measurable PFAS loads introduced to the incineration system through the wet scrubber supply water.Figure S15 shows that the wet scrubber supply water introduces similar amounts of organically bound fluorine from the targeted PFAS into the system as the dewatered wastewater sludge for the MHF and 15%-20% for the FBF.
The PFAS entering the wet scrubber in the supply water largely left via the drain.Table S58 shows loads of PFAS in either stream.When averaging the total mass processed during each sampling run, there was no statistical difference (two tailed, α = 0.05, paired) between the supply and drain for either the MHF or FBF, except PFOS for the MHF came close to statistical significance (p value = 0.06).Basically, targeted PFAS that went in with the supply water came out in the drain.Seay et al. (2023) observed a similar trend at an FBF WSI where the wet scrubber supply and drain water PFAS content differed by no more than 23% for five compounds; see Table S59.Table S60 contains the results of another study where no differing trend was observed (Weston & Sampson, 2023).
The wet scrubber may impact PFAS stack detections given the levels in the supply water.Interestingly, the predominant molar stack emissions from the MHF (PFBA, PFPeA, and PFHxA) were primarily collected in the impinger section of the OTM-45 sampling train; see Table 5.This may suggest that any of these compounds emitting from the furnace would be captured in the wet scrubber, but the previous discussion suggests none of these species are sequestered in this compartment.This leaves wet scrubber supply water as a potential source of these compounds to the stack, which is evaluated in the following paragraph.Unfortunately, the FBF stack samples exhibited contamination in the field and proof blank OTM-45 trains; thus, the following discussion focuses on the MHF results.
The PFAS fingerprint of the MHF wet scrubber supply water and stack samples share a resemblance in compounds reported; see Figure S16 and associated discussion.The amount of water vapor collected as part of the MHF stack sample cannot consistently explain, if ignoring any water vapor contribution from the furnace, the levels of PFAS emitted to the atmosphere-see Table S61 and related discussion.Concomitantly, the PFAS entering the incineration system via the wet scrubber supply water may partition into the flue gas.If the PFAS reported in the MHF stack came from the wet scrubber supply, the analytical sensitivity of the  wet scrubber water samples would not register the loss for all but PFBA, as shown in Table S62.Those compounds showing a loss across the wet scrubber water stream, though not statistically significant, would require much higher reportable stack values than observed per Table S63.Similarly, a cursory evaluation of the air-to-water partition coefficient (D iaw for compound i) does not indicate the transfer of PFAS from the wet scrubber supply water into the flue gas given the current state of the science.Using theoretical values at 25 C (Equations S2 and S3), the resulting D iaw suggests PFAS in the wet scrubber supply water will remain in the aqueous phase per calculated values in Table S64.Current physicochemical data do not correct for the wet scrubber water stream temperature rise, from 19 C to 47 C; however, a temperature adjustment must be profound to shift the calculated equilibrium toward the gas phase.Further, impacts on D iaw from water salinity, gas-phase pressure drop, and localized conditions within the multiple stages of the wet scrubber cannot be estimated with the currently available data.The scientific community must develop robust physicochemical data to confidently evaluate the fate of these PFAS through multiphase systems.
The wet scrubber results in Table 8 indicate an increase of TOF from the water supply to the drain for both the MHF and FBF, specifically in the nonpolar extract.The reportable or estimated TOF values from other inputs or outputs to the incineration system have little impact on the overall balance compared with the wet scrubber water balance.For example, if the targeted PFAS organofluorine content in the dewatered wastewater sludge (average 0.00422 and 0.0215 mol of fluorine in the MHF and FBF, respectively) transformed into a nonpolar form and was captured in the wet scrubber water, it would not explain the increase in TOF.The accumulation of organic fluorine in the nonpolar wet scrubber drain water implies nonpolar compounds are emanating from the furnaces.As previously described, five nonpolar FTOH PFAS compounds were targeted, and no samples yielded reportable results.
Another explanation proposed relates to compounds categorized as hydrophobic ionogenic organic chemicals (HIOCs), which have a dual hydrophobic, yet ionized, character that varies according to their pK a and the pH of the medium.Some PFAS, but not all, can be categorized as HIOCs.As ions, these compounds will partition to the wet scrubber water from the flue gas.Further, even when Note: --= result less than reporting limit, screened during quality control, or not present in all samples.DWS = dewatered wastewater sludge.PFAS emissions normalized to 0.907 DMT/day of dewatered wastewater sludge and 3.785 ML/day of wastewater flow.Alternative units without changing the values shown would be mg/3.785MLÁday À1 (mg per million gallons per day) or mg/0.907DMTÁday À1 (mg per dry ton per day), see note below for stack flows.a Assumed same targeted PFAS levels as wet scrubber supply.b Stack emissions, sum of all four fractions, normalized by the average stack flow per DMT/day of cake feed observed during sampling events (MHF-10.4DSCM per minute/DMTÁday À1 , FBF-7.69 DSCM per minute/DMTÁday À1 ).Or, in alternative units, mg per 406 dry standard cubic feet per minute per dry ton per day for the MHF and 299 for the FBF in similar units.fully ionized, these compounds are highly hydrophobic (log D ow > 3; Wells, 2006) and may partition into the nonpolar CH 2 Cl 2 extract, thereby contributing to the TOF results observed here.Schaefer et al. (2022) found that two families, proposed here as potential HIOCs-the polyfluoroalkyl phosphate diesters (diPAPs) and perfluorophosphinates (PFPis)-dominated biosolids samples from seven WRRFs when analyzed for 41 PFAS.
If diPAP and PFPi compounds were present in the MHF and FBF dewatered wastewater sludge, their presence may have gone unnoticed.The TOF reporting limit in the dewatered wastewater sludge nonpolar extracts was nearly 300 ng/g wet compared with 2 μg/L, essentially equivalent to ng/g, in filtrate from the wet scrubber water stream.These suspected nonpolar acting compounds may be detected in the wet scrubber water stream, assuming they passed through the furnace untransformed.Alternatively, the parent diPAP and PFPi compounds may degrade through the furnace, perhaps like the ether-based PFAS decomposition described in Alinezhad et al. (2023), into smaller nonpolar versions.However, even when assuming the reporting limits for the dewatered wastewater sludge TOF, a clear trend cannot be discerned, as only the FBF data explain the TOF gains in the wet scrubber.The nonpolar TOF captured in the wet scrubber water stream cannot be explained by the diPAP and PFPi concentrations measured at other facilities (Schaefer et al., 2022), some other fluorinated organic compounds must therefore be contributing to this TOF concentration increase.If, during the combustion process in the furnace, fluorine-carbon bonds are broken, fluoride would be expected to accumulate in the water stream from the wet scrubber.The liberated fluorine would likely become HF in the flue gas, which would dissolve and dissociate in the wet scrubber stream, resulting in an increase of fluoride ions.A mass balance using HF is unrealistic in fullscale systems given the tendency for this acid to react with the refractory lining of the furnace and ductwork before the wet scrubber (Aleksandrov et al., 2019).Seay et al. (2023) observed 44,000 times larger fluoride outflows from the incineration system, largely from the gain across the wet scrubber, compared with the fluorine measured in the targeted PFAS analyses.The fluoride concentrations in the dewatered wastewater sludge may explain the increase in wet scrubber water stream observations.Table S65 contains the calculated dewatered wastewater sludge fluoride content required to explain the gain in the wet scrubber water stream compared with that measured.While the results as presented in Table S65 suggest the possibility of substantial fluorinated organic degradation, the potential underreporting of the dewatered wastewater sludge fluoride could be the source of this discrepancy, as previously described.Table S66 contains dewatered wastewater sludge fluoride measurements from other WRRFs for perspective, the range encompassing the value measured at the MHF WRRF.Further research is necessary to establish concentrations of fluoride in dewatered wastewater sludge.

Combustion air organic fluorine
Other researchers have documented targeted PFAS air contamination at WRRFs (Hamid & Li, 2016).The nonreportable values in this research may lie in the volume of air sampled.Each sample represented a nominal volume of three dry standard cubic meters (DSCM), consistent with the OTM-45 methodology (USEPA, 2021a).This sample volume facilitated a straightforward comparison to stack samples to determine significance of the combustion air contamination on the mass of PFAS moving through an incinerator.For comparison, a recent study that sampled air around a WRRF collected 100 times as much volume and detected several PFAS (Seay et al., 2023).The highest reported value not complicated by qualifications or control contamination exceeding sample values was PFBA at 16 pg/m 3 (25 C and 101.325Pa).If this concentration were delivered to the MHF and FBF studied in this research, at the calculated combustion airflow rates in Tables S15 and S21 with zero DRE through the incineration process, the amount of PFAS exiting the process from the stack sampling point would not have been reportable for either site.
The reportable levels of TOF from the combustion air were unexpected; see Table 4.The sampling apparatus described previously was intentionally designed to closely resemble the OTM-45 sampling system, specifically the back half portion (compare Figures S7 and S8).Unfortunately, any TOF detections in the stack OTM-45 samples generally corresponded with measured contamination in the proof and field blank controls.As a result, the research team has low confidence in the stack TOF concentration data and did not evaluate it further.Due to the perceived diminutive organic fluorine content of the combustion air, no controls were executed beyond laboratory method blanks and spikes.The data collected here must be viewed with skepticism; however, there are some observations worth discussing.
Given the lack of reportable targeted PFAS in the combustion air, in theory, the TOF reported represents other forms of fluorinated organic compounds.Unfortunately, like this work, the body of literature covering the organic fluorinated compounds airborne within WRRF sites focuses on targeted compounds (Ahrens et al., 2011;Seay et al., 2023;Shoeib et al., 2016 andWeinberg et al., 2011).Moreover, results of these studies measured PFAS in pg/m 3 whereas TOF reported from the MHF and FBF samples ranged from approximately 0.5-2 ug/m 3 , suggesting that fluorinated organic compounds were something other than targeted PFAS.
Data collected here cannot rule out that combustion air TOF originated from emissions from the stack.Although screened out for data usability, the OTM-45 samples from both the MHF and FBF contained similar levels of TOF.The stack outlet and combustion air sampling points lie within a hundred meters of each other, line of sight.Wind speed or direction was not recorded during sampling, but review of publicly available data from local weather stations for each site indicated wind speeds (8-32 km/h) in the general direction toward the combustion air sample site from the stack.Additional research is required to explore this possibility, including control samples at both sample points, before prospective conclusions can be drawn.

Volatile PFAS evolution
As noted previously, the targeted PFAS reported in the stack of the MHF represented short alkyl chain compounds and ultra-short chain volatile alkane PFAS (CF 4 , C 2 F 6 , and C 3 F 8 ).This suggests that longer alkyl chain targeted PFAS, and perhaps other fluorinated compounds, in the dewatered wastewater sludge break down, but not completely.CF 4 represents the most thermodynamically stable PFAS and arrestment of the thermal decomposition at this compound is logical (Altarawneh et al., 2022;Winchell, Ross, et al., 2021).Any production of CF 4 , C 2 F 6 , and C 3 F 8 in the system would represent a transformation of other fluorinated organics given their volatile nature and only logical entry to the incineration system with the combustion air.Atmospheric concentrations of these compounds remain in the parts per trillion range (Mühle et al., 2010) and therefore the combustion air content would not influence the stack measurements based on parts per million.
Stack FTIR results suggest roughly two orders of magnitude more organically based fluorine enters the incineration process than reported in the dewatered wastewater sludge targeted PFAS (Table S67).The amount of undescribed organic fluorine translates into dewatered wastewater sludge targeted PFAS levels in the parts per million level.Levels like this have yet to be documented on nonindustrial impacted WRRF solids streams (Rainey, 2019).This stark imbalance calls into question the veracity of analysis, like the previous wet scrubber water fluoride discussion.Thoma et al. (2021) questioned FTIR results from the stack of a WRRF solids pyrolysis system where several fluorinated species were targeted including CF 4 .Two FTIR analyzers and canister samples subsequently analyzed with GC/MS reported varying values.One FTIR analyzer provided reportable values, exceeding the MDL (5 parts per billion [ppb]) by more than 50%.The other FTIR analyzer and canister analysis resulted in intermittent or nondetection, respectively, at similar MDL (6 ppb).The authors questioned the accuracy of the FTIR measurements due to contamination from the dynamic spiking required by the methodology.At the MHF and FBF facilities, sulfur hexafluoride was used for dynamic spiking and was not expected to interfere with the CF 4 , C 2 F 6 , and C 3 F 8 spectra.Shields et al. (2023) published the results of processing aqueous film forming foam (AFFF) through a pilotscale furnace and measured volatile fluorinated organics in the flue gas.The researchers took samples with canisters and analyzed CF 4 , C 2 F 6 , and C 3 F 8 , plus others, on a GC/MS.CF 4 was nondetect at all experimental temperatures though the detection limit was comparatively high (180 μg/m 3 , 50 ppb by volume [ppbv]).Both C 2 F 6 and C 3 F 8 (2.82 and 3.84 μg/m 3 , 0.5 ppbv, detection limit, respectively) were reported at lower temperatures but disappeared at 1090 C and higher.While the CF 4 detection limit was on par with MHF and FBF results (20 and 10 ppbv), the C 2 F 6 and C 3 F 8 were two orders of magnitude lower.These results show that volatile fluorinated byproducts result from the thermal processing of AFFF and perhaps in WSIs processing PFAS-contaminated dewatered wastewater sludge, although comparison of the two studies is tenuous from a feedstock and temperature regime perspective.At a minimum, the lower detection limits achieved for C 2 F 6 and C 3 F 8 would validate the intermittent detections in the MHF and FBF, and the approach should be considered for future research efforts.
FTIR results provide a glimpse into potential greenhouse gas emissions from fluorinated compounds processed through the incineration systems sampled.Based on the global warming potential (GWP), Table S68, the CF 4 , C 2 F 6 , and C 3 F 8 results were converted into equivalents of CO 2 and gasoline powered passenger vehicles shown in Table 9. Results were normalized to the estimated annual dewatered wastewater sludge processed at each site.As indicated by the results, when either C 2 F 6 or C 3 F 8 values are reportable, the emissions increase significantly compared with CF 4 alone.
The estimated greenhouse gas emissions from incomplete destruction of fluorinated organic compounds may be relatively small but warrant further investigation.The United States generates about 129 billion liters of wastewater per day (USEPA, 2023b), emitting around 42 million metric tons of CO 2 equivalents of greenhouse gases (USEPA, 2023c).That equates to approximately 0.89 metric tons of CO 2 equivalent per mega-liter of wastewater treated.Applying this emission rate to the estimated treatment flow at the test site facilities, the MHF and FBF respectively emit roughly 31,000 and 124,000 metric tons of CO 2 equivalent per year.At these levels, the measured volatile PFAS emissions represent from 0.5% to 4.5% and 0.5% to 2.8% from the MHF and FBF, respectively, of the total facility emissions.While this evaluation is overly simplistic and must be calculated based on site-specific emissions for localized understanding, it broadly indicates that these emissions would represent a sufficient portion of a WRRF's total emissions.

CONCLUSIONS AND RESEARCH NEEDS
This research project documented PFAS in all major inputs and outputs of two WSIs representing technologies used in the United States.An extensive, state-of-the-art sampling and analytical program allowed researchers to track PFAS through the process.Even so, results supporting DRE determinations were impractical after applying data quality standards, insufficient detections, and skewing from secondary streams.
Nevertheless, at processing temperatures, residence times, and turbulence typical of these technologies, significant reductions and changes in the composition of PFAS were observed.The dewatered wastewater sludge feedstock contained reportable amounts of targeted PFAS averaging 247-and 1280-μmol PFAS per sample run, MHF (Runs 2 and 3 were averaged because the large contribution from PFOS was not reportable in Run 1) and WATER ENVIRONMENT RESEARCH FBF, respectively.Of those amounts, 81-and 1157-μmol PFAS came from compounds reported in the dewatered wastewater sludge but not at the primary environmental release point at the stack.For the MHF, with uncontaminated stack samples, the targeted PFAS stack emissions averaged 5% of the dewatered wastewater sludge value and favored shorter alkyl chain compounds, suggesting transformation of larger influent PFAS.While reduced, some targeted PFAS escaped the system, albeit primarily as shorter alkyl chain compounds including the volatile versions.The human health and environmental impacts of these emissions must be assessed to determine whether they represent concerning levels.
Comparing the performance of the MHF and FBF is difficult with the limited amount and quality of data collected.Winchell, Ross, et al. (2021) suggested the comparably higher turbulence and longer gas residence time at elevated temperature in an FBF would translate into greater PFAS DRE.Providing a definitive comparison with this research is not possible.Even tenuous conclusions must be qualified given the FBF stack sample contamination of PFBA, PFPeA, and PFHxA, leaving that data unusable, were the primary (molar) compounds found in the MHF stack samples.However, the fact that the FBF dewatered wastewater sludge contained on average over 1000-μmol PFAS (targeted) more per sampling run than the MHF and did not consistently emit any reportable amounts from the stack suggests comparably better performance.If stakeholders care to further distinguish these two furnace technologies, additional research is needed.This research will also need to consider the plethora of operating conditions and configurations possible with these furnaces to arrive at an objective conclusion.
Various samples also indicated significant amounts of organic fluorine exist in the system that is not explained by targeted PFAS results.The TOF concentration increases in the wet scrubber water streams are virtually unexplained by the targeted PFAS, likewise for the combustion air results.The FTIR data from the stack and the fluoride increases in the wet scrubber water stream also support the conclusion that a significant portion of organic fluorine flows through the process.Further, these measurements support the conclusion that some PFAS are degraded and potentially destroyed.Additional research also needs to include determination of what types of substances represent this organic fluorine and if the observed concentrations are concerning.That research will be answered, in part, when NTA results are completed on the samples collected as part of this project.
Several lines of research needs have been identified by this project in addition to those already noted.First, if the levels of PFAS observed in the samples collected pose concern and defensible DREs required, a controlled experiment with known quantities of select PFAS spiked into the dewatered wastewater sludge must be performed.This provides a known quantity of PFAS that can also be used to evaluate the veracity of the TOF, FTIR, and fluoride results observed in this project.The results of this experiment will lead to the evaluation of whether existing WSIs require modification, and if regulatory authorities determine that modifications are required to increase PFAS DRE, changes can be made to increase temperature and residence time but must be weighed against the cost and benefit.
The nonpolar TOF results from the wet scrubber water stream imply that focus must be given to that class of compounds.Ideally, the NTA work being completed will help identify the specific compounds mainly responsible for the observations.Subsequently, these compounds can be the focus of future efforts, including the development of the physical-chemical properties of PFAS to determine the mechanisms creating and partitioning these compounds within the process and during analysis.
Advancements in the analytical techniques that lower reporting limits for TOF and the volatile PFAS are required to verify the observations here.Better characterization of the fluoride content in the samples is needed as well.The significant amount of organic fluorine suggested by these analyses warrants additional work to increase confidence in the data.Additionally, the targeted compounds in this study were all nonpolymeric PFAS, and the near-total lack of information on the fate and transport of polymeric PFAS in the environment and WSIs is a serious knowledge gap.
Overall, the two existing WSIs sampled as part of this project are degrading fluorinated organics, including PFAS.The degree of degradation remains elusive, but the data presented indicate significant reductions when applying established data quality procedures and removing bias of wet scrubber water streams.Whether the degradation achieved meets the desires of all stakeholders can be debated, but the data presented shows that WSIs are degrading, and potentially destroying, PFAS.

F
I G U R E 1 Wastewater sludge incinerator (WSI) sampling point schematic.

T
A B L E 3 Quantifiable FBF operating conditions conducive to PFAS degradation.= temperature.GRT = gas residence time.NE = not estimated.TA B L E 4 Organic fluorine sources to the MHF and FBF.
= less than or equal to method detection limit.Q = data screened during quality control.a Values represent the total moles of analyte exiting the incineration process in the sample point during each sampling run.b Percent of values detected greater than the MDL.
= less than or equal to method detection limit.Q = data screened during quality control.a Values represent the total moles of analyte exiting the incineration process in the sample point during each sampling run.b Percent of values detected greater than the MDL.

F
I G U R E 2 Average targeted PFAS) fate in the sampled wastewater sludge incinerators (WSIs).
Quantifiable MHF operating conditions conducive to PFAS degradation.
T A B L E 2Note: T = temperature.GRT = gas residence time.SRT = solids residence time.NA = not applicable, no solids are processed on this hearth.
Emissions from MHF and FBF.
T A B L E 5 T A B L E 6 Targeted PFAS emissions normalized to unit mass of dewatered wastewater sludge (mg/day).
Stack samples yielded reportable results but quality control blanks demonstrated contamination.dOnly one sample at the stack yielded a reportable result.Only two samples of dewatered wastewater sludge from the MHF site yielded reportable results.
c e T A B L E 8 TOF gains across sampled systems.Calculated balance across WSI system.Reportable inputs included combustion air, wet scrubber supply water, dewatered wastewater sludge water content, lubrication water (FBF only).Reportable output limited to wet scrubber drain water.Average of triplicate analyses for 6:2, 8:2, 10:2, and variations of 8:2/6:2 and 10:2/8:2 diPAPs from seven WRRFs applied to dewatered wastewater sludge TS and loading rates for the MHF and FBF.Values in parentheses represent the concentration of fluorine in the dewatered wastewater sludge (ng fluorine/g wet dewatered wastewater sludge).
c d Combined polar and nonpolar reporting limits.e T A B L E 9 Estimated greenhouse gas emissions from volatile PFAS.Normalized by converting the emissions measured during the sampling run to the generalized annual dewatered wastewater sludge processed at each site assuming treated wastewater flows of 95 ML/day and 379 ML/day for the MHF and FBF, respectively, and 0.907 DMT/ML.Gasoline powered passenger vehicle equivalent, 4.49 MT CO 2 equivalent per vehicle per year (USEPA, 2023a).
a e