Per- and Polyfluoroalkyl Substances Presence, Pathways, and Cycling through Drinking Water and Wastewater Treatment

Per- and polyfluoroalkyl substances (PFAS) present a unique challenge in water and wastewater treatment (Vo et al. 2020). Conventional physicochemical and biological treatment methods cannot fully degrade or mineralize PFASs in drinking water treatment (Belkouteb et al. 2020; Kim et al. 2020; Boone et al. 2019; Dauchy 2019; Page et al. 2019; Vughs et al. 2019; Hopkins et al. 2018) or wastewater treatment (Chen et al. 2018; Szabo et al. 2018; Arvaniti and Stasinakis 2015) and may generate PFAS transformation products. The treatment goal of mineralization relative to PFAS is defined by Horst et al. (2020) as complete defluorination regardless of whether the carbon is fully oxidized to carbon dioxide. In contrast to conventional treatment methods, Ahmed et al. (2020a) published a critical overview of different advanced degradation methods—advanced oxidation, PFAS defluorination, advanced reduction, and thermal and nonthermal processes—for PFAS removal from water and wastewater. However, the authors noted that most of these methods are laboratory studies at this stage that show promise but have not been tested in real commercial-scale plants.

Fluorine, the lightest halogen, is the most electronegative element; as such, the C–F bond is among the strongest of known covalent bonds (Kissa 2001). Properties of PFAS are useful for industrial and commercial purposes, and mass production has occurred since the early 1950s (Kissa 2001). The same properties, such as thermal stability and chemical inertness, that make them useful also make them stable and environmentally persistent. As production continued into this century, environmental accumulation of PFAS from intentional and nonintentional releases has occurred. For instance, PFAS utilized in specific firefighting foams are distinct sources of environmental PFAS (Dauchy 2019). More broadly, the USEPA states that manufacturing plants, fire-training areas, and fire-suppression activities at airports, refineries, and military installations are primary sources of PFAS release to the environment (USEPA 2021a).

The USEPA maintains an online, continuously growing (currently $>\mathrm{9,000}$ chemicals) master list of PFAS (USEPA 2021b). Multiple categories of perfluoroalkyl (fully fluorinated carbon atoms) and polyfluoroalkyl (partially fluorinated carbon atoms) compounds exist. Each PFAS has a linear or branched alkyl chain (Kissa 2001) and a perfluoroalkyl group (${\mathrm{C}}_{\mathrm{n}}{\mathrm{F}}_{2\mathrm{n}+1}$) (Buck et al. 2011). Detailed definitions of PFAS-related terminology are available in Buck et al. (2011), Dauchy (2019), Pancras et al. (2016), and Wang et al. (2017). PFAS diversity is tabulated in Winchell et al. (2021a), and representative chemical structures are illustrated in Winchell et al. (2021b). PFAS comprise a large family of chemicals cataloged across various nomenclatures such as (1) polymeric (including fluoropolymers, perfluoropolyethers, and side-chain fluorinated polymers) and nonpolymeric [including perfluoroalkyl acids and per-/polyfluoroalkylether acids (PFAA), PFAA precursors, and others], most of which are further subcategorized; (2) legacy, transformation products, and emerging PFAS; (3) ultrashort-, short-, and long-chain PFAS; and (4) polar/nonpolar, nonvolatile/semivolatile/volatile, and neutral/anionic/cationic/zwitterionic PFAS. PFAA precursors are fluorinated chemicals that can be transformed abiotically or biotically into terminal perfluoroalkyl carboxylic acid (PFCA) or perfluoroalkyl sulfonic acid (PFSA) products (Casson and Chiang 2018).

In 2006, the USEPA initiated the 2010/2015 PFOA Stewardship Program, inviting eight major PFAS industries to work toward the elimination of perfluorooctanoic acid (PFOA), PFOA precursor chemicals, and related higher homolog chemicals. The program is one of the main drivers of the reformulation of PFAS-based products. All companies have met the PFOA Stewardship Program goals (USEPA 2021c). The Stockholm Convention has restricted the production of perfluorooctanesulfonic acid (PFOS) (UNEP 2019a) and banned the production of PFOA (UNEP 2019b). However, as will be confirmed in this review, these PFAS still occur in the water cycle because of their persistence and because importation from other global markets remains a route for material entering the domestic chain of commerce.

In response to increased regulation of long-chain so-called legacy PFAS [eight or more carbons (C8) in the alkyl chain and no ether bonds], industrial applications have shifted toward short-chain (C4–C7) and ultrashort-chain (C2 and C3) PFAS alternatives, with or without ether bonds (Wang et al. 2019; Mulabagal et al. 2018). As reported by the Interstate Technology and Regulatory Council (ITRC 2020), alternative PFAS-based chemical replacements are being marketed and some are appearing in the environment (Munoz et al. 2019; Gustavsson et al. 2018; Hopkins et al. 2018; Wang et al. 2015b), especially several subclasses of ether-PFASs, including the most well-known F-53B [major component (9-chlorohexadecafluoro-3-oxanonane-1-sulfonic acid) or 9Cl-PF3ONS and minor component (11-chloroeicosafluoro-3-oxaundecane-1-sulfonic acid) or 11Cl-PF3OUdS], Gen-X (hexafluoropropylene oxide dimer acid or HFPO-DA), and ADONA (4,8-dioxa-3H-perfluorononanoic acid) compounds (named in their acid forms). These PFAS replacement chemicals are now included in the EPA Analytical methods 533 and 537.1 for PFAS in drinking water (USEPA 2019). Unfortunately, many recently adopted fluorinated alternatives and precursors are more persistent and more environmentally mobile than legacy PFAS (Ghisi et al. 2019).

Broad environmental PFOS contamination has been documented. Giesy and Kannan (2001) were the first to report global PFOS contamination in wildlife. For example, PFOS concentrations in the Midwestern US was as high as $\mathrm{2,570}\mathrm{ng}/\mathrm{mL}$ in the blood plasma of bald eagles and $\mathrm{3,680}\mathrm{ng}/\mathrm{g}$ wet weight in liver samples from mink. Animals tested from urban areas yielded higher concentrations than more remote locations; even wild polar bears contained detectable levels of PFOS. PFOS was measured at higher concentrations in predator species than prey, suggesting bioaccumulation of PFOS. A follow-up study demonstrated bioaccumulation in aquatic species (Giesy et al. 2010) and estimated water quality criteria for the protection of aquatic organisms and wildlife.

Several studies across the globe have recently been published to illustrate the broad environmental contamination of PFAS. Muir et al. (2019) documented the wide presence of PFASs throughout the Arctic Ocean and near-shore environments, but they noted that “most long-term time series show a decline from higher concentrations in the early 2000s.” Ali et al. (2021) measured Saudi Arabian Red Sea PFAS concentrations in sediments and edible fish and pointed to wastewater effluents as the main source of these compounds. Bai and Son (2021) performed a study investigating the presence of 17 specific PFAS compounds in 6 surface water and sediment locations in the US state of Nevada, and found that short-chain PFAS were more prevalent in the aqueous phase, while long-chain PFAS were more prevalent in soil (sediments). Cao et al. (2019) investigated soil, sediment, and aqueous-phase PFAS in the Yuqiao reservoir in Tianjin, China, while Chen et al. (2021) investigated similar environmental samples in the Pearl River in southern China. Additional local studies across the globe regarding PFAS environmental contamination include Florida (Cui et al. 2020), Kampala, Uganda (Dalahmeh et al. 2018), Three Gorges Reservoir, China (Jin et al. 2020), the Asan Lake region of South Korea (Lee et al. 2020), South African estuaries (Olisah et al. 2021), Beibu Gulf, South China (Pan et al. 2021), Gangetic Plain, Patna, India (Richards et al. 2021), Mexico City (Rodríguez-Varela et al. 2021), Xiamen Bay, China (Wang et al. 2020), and the Loess Plateau, China (Zhou et al. 2021).

In soils, Brusseau et al. (2020) compiled data from more than 30,000 samples from over 2,500 sites globally; the maximum reported PFOS concentrations reached up to several hundred milligrams per kilogram, even in remote regions, where sites were far from PFOS sources. Zacs and Bartkevics (2016) measured PFOA and PFOS in surface water, wastewater, biota, sediments, and sewage sludge in the Baltic region. They noted a prevalence of PFOS over PFOA in surface water and biota samples, whereas mean concentrations of PFOA were greater than PFOS in wastewater, sediments, and sewage sludge. Other, recent, reviews focused on fate and transport within soils and sediments, such as Willemsen and Bourg (2021), who studied the adsorption kinetics of various PFAS structures. Ahmed et al. (2020b) noted decreasing adsorption onto activated carbon as the carbon chain length increased in PFAS, although longer-chain PFAS have larger partition coefficient values than shorter-chained versions. Using numerical studies of PFOS transport within unsaturated soil, researchers have concluded that groundwater contamination could occur from decades to centuries after an initial surface spill (Mahinroosta et al. 2021; Sánchez-Soberón et al. 2020).

A growing body of literature has demonstrated that PFAS are found in nearly all environments and in the organisms living therein, including humans. There are now thousands of research reports on the occurrence of PFAS in humans, with data being reported on PFAS levels in human blood, urine, milk, and other tissues (hair and nails) (De Silva et al. 2021; Liu et al. 2020; Jian et al. 2018). While most studies, to date, have focused on human blood, more recent research indicates the distribution of PFAS in various tissues is a function of both PFAS chain length and physiological characteristics (Jian et al. 2018).

Water and diet are major potential pathways of PFAS exposure in humans along with food packaging, cookware, air, and air-suspended dust (Ghisi et al. 2019). Elevated PFAS water contamination from localized sources can impact raw water, and thus consumption of contaminated drinking water is a major pathway for human exposure (Tröger et al. 2021; Appleman et al. 2014) if PFAS-specific drinking water treatment is not implemented. The USEPA has established the drinking water health advisory level (HAL) at 70 parts per trillion (combined concentrations of PFOA and PFOS) (USEPA 2016). If drinking water meets the USEPA HAL, the daily intake of PFOS and PFOA from water is less than 20% of the estimated total average human intake per day (USEPA 2016).

PFAS contamination of food may occur via irrigation and biosolid application in agriculture (Ghisi et al. 2019) and home food production (Huset and Barry 2018) or farmed and hunted animals (Death et al. 2021), resulting in additional human exposure. Ramakrishnan et al. (2021) documented the presence of PFAS even within organic farming systems and produce production. They noted the use of PFAS-containing composts and feeds, and report that, in some instances, organic meats may contain even higher concentrations of select chemicals than conventionally grown produce. With ubiquitous contamination and bioaccumulation tendencies, growing public concern about human exposure to PFAS stems from an increasing body of evidence of adverse toxicological effects (Sunderland et al. 2019; ATSDR 2021).

Given the concern and broad environmental impact of PFAS, this manuscript aims to establish the connections within and between water resource recovery facilities (WRRFs) and drinking water treatment plants (WTPs) including their role in the fate and transport of PFAS cycling in the environment in a manner to inform the industry on this issue. The term water resource recovery facility is adopted here instead of wastewater treatment plant (WWTP) to more broadly reflect the ability to recover valuable resources from wastewater as advocated by the Water Environment Federation (WEF) (WEF 2014). Evidence for PFAS cycling is presented to inform and educate students, professional engineers, scientists, lab managers, treatment plant operators, and decision makers interested in the status of PFAS water quality impacts.

To understand potential human exposure to PFAS, it is necessary to thoroughly evaluate the PFAS cycle, including WTPs and WRRFs (AWWA 2020). While significant research has focused on the occurrence of PFAS in portions of the water cycle, comprehensive discussions of the complete cycle are limited, as discussed herein. Vo et al. (2020) recently contributed a review of PFAS sources and their occurrence, transformation, fate, migration, and remediation that addresses technologies in both water and wastewater.

In this manuscript, a different approach to investigating the water/wastewater cycle through the environment is made by examining the three phases of matter (i.e., liquid, solid, and gaseous) in which PFAS are transported. Each phase may contribute to overall human exposure via different routes prior to returning to the WTP or WRRF.

The production, use, and disposal of PFAS connects to the water cycle via aquifers/surface water/sediment, drinking water treatment, domestic/industrial water use, domestic/industrial wastewater treatment, the atmosphere, land application, landfills, thermal treatment of wastes, and food production (Fig. 1). In Fig. 1, outputs from PFAS primary sources are depicted by densely dotted lines. Each arrowhead in Fig. 1 represents an input to a different environmental compartment.

Here we focus on WTPs and WRRFs as an integrated anthropogenic system composed of multiple unit operations that result in aqueous (loosely dotted lines), solids (solid lines), and gaseous (dashed lines) discharges of legacy, replacement, and transformed PFAS to the environment (Fig. 1). In addition to cycling within these phases, phase-transfer processes, i.e., solids–air, solids–water, and air–water partitioning, are also important in the transport and fate of PFAS (Brusseau 2019a, b).

Liquid-phase PFAS-containing inputs continuously enter WTP and WRRF facilities (liquid-phase path, loosely dotted lines, Fig. 1). WTPs are fed from groundwater and surface water. WRRFs are fed used WTP water, wastewater from untreated domestic, commercial, and industrial sources—including from PFAS production, internal recycling of spent streams, inflow and infiltration, and landfill leachate discharges to sanitary sewers. Dry and wet deposition can result in PFAS contamination (Pike et al. 2021). Chen et al. (2019) collected water from precipitation events in urban China and measured 22 PFAS reporting that trifluoroacetate (TFA), PFOA, and PFOS were the predominant PFAS.

WTP outputs, which often become WRRF inputs, directly impact human exposure by supplying drinking water for human consumption and irrigation for home garden food production (Huset and Barry 2018). WRRFs return treated effluents to surface water and indirectly to aquifers that become inputs to WTPs, completing the urban water cycle. Once in the aquatic environment, bioaccumulation and biomagnification of these compounds can lead to human exposure through consumption of fish and wildlife (Fair et al. 2019).

Residuals and solids (solid-phase path, solid lines, Fig. 1) generated at WTPs and WRRFs, respectively, can be land applied for beneficial reuse, sent to a landfill, or subjected to thermal treatment or incineration. Residuals from water treatment have been shown to contain PFAS (Coggan et al. 2019; Lazcano et al. 2019; Nakayama et al. 2019; Rainey 2019; Seo et al. 2019; Eriksson et al. 2017; Gallen et al. 2017; Kwon et al. 2017; Mailler et al. 2017; Armstrong et al. 2016; Pan et al. 2016; Ulrich et al. 2016; Alder and van der Voet 2015; Wang et al. 2015b; Yan et al. 2012; Navarro et al. 2011; Sun et al. 2011; Guo et al. 2010; Loganathan et al. 2007). Sediments from surface waters have been shown to contain PFAS that have been linked to WRRF effluent (Mussabek et al. 2019; Bach et al. 2017; Qi et al. 2017). Ash from inefficient incineration processes (Loganathan et al. 2007) may also be another potential input to PFAS cycling in the environment.

Gas-phase emissions of PFAS (gas-phase path, dashed lines, Fig. 1) have been documented across unit processes in WTPs and WRRFs. Both WTPs and WRRFs are in open contact with the atmosphere, and PFAS exist in atmospheric gaseous and particulate matter associated with these processes (Yu et al. 2020b; Barber et al. 2007; Jahnke et al. 2007). Thermal treatments and incineration of wastewater solids also are potential sources of direct gas emissions of PFAS to the atmosphere (Winchell et al. 2021a). Once in the atmosphere, FTOHs likely oxidize to yield PFCAs initiated by reaction with OH radicals (Ellis et al. 2004).

As part of the gas cycle, deposition from the atmosphere also occurs. Global atmospheric PFCA deposition was recently modeled by Thackray et al. (2020). These authors found that the ratio of long-chain to short-chain PFCA increased with distance from sources. This ratio can inform the understanding of PFCA formation due to degradation of precursors.

While there are both voluntary and regulatory restrictions and bans on the production of some PFAS globally, destructive treatment methods that can break the PFAS cycle, such as plasma technologies, electrochemistry technologies, and other advanced reduction methods (Ross et al. 2018), represent critical research areas. Some of these technologies, though they hold promise, may be years from commercialization, and PFAS continue to enter and exit WTPs and WRRFs in flows that interconnect in the urban water cycle, with the result that these facilities are important in PFAS cycling into the total environment.

Current water and wastewater treatment methods, such as GAC, IX, and RO, that remove PFAS from one medium and transfer them to another only extend PFAS cycling among liquid-, solid-, and gas-phase paths. Alternative thermal treatment technologies, such as pyrolysis, gasification, and hydrothermal liquefaction of WRRF solids, have demonstrated partial mineralization of PFAS (Winchell et al. 2021a; Yu et al. 2020a). An understanding of the mechanisms and extent of destruction has not been fully developed and deserves further research.

While investigations of aspects of PFAS cycling in water and wastewater treatment and the treatment methods by which to interrupt the PFAS cycle are essential research needs, they are complicated by several factors. Kotthoff and Bücking (2018) identified four major trends challenging future PFAS research: mobility, substitution of regulated substances, increase in structural diversity of existing PFAS molecules, and unknown so-called dark matter.

PFAS have been shown to contaminate all aspects of the global ecosystem, and WTPs and WRRFs are no exception. Researchers have detected and documented the mobility of PFAS through all three phases (i.e., gas, liquid, and solid) of all inputs and outputs around WTPs and WRRFs. Neither type of treatment facility frequently utilizes processes to remove and destroy PFAS, though physical separation methods exist for PFAS removal from the liquid phase, and thermal processes applied to residual streams from WTPs and WRRFs offer a potential destruction outlet. Moreover, owing to their recalcitrance, PFAS released to the environment, or even disposed of in landfills, can return to the treatment facilities in an endless cycle. For PFAS already present in the environment, physical separation and combustion provide a commercially relevant destruction technology, while pyrolysis and gasification systems are emerging at full scale, and other wet thermal and nonthermal processes are being examined in laboratory settings for solid- and liquid-phase treatment. Because WTPs and WRRFs are not producers of PFAS but are often publicized as sources of contamination, source reduction of the influent PFAS loads to these facilities will be critical to bringing this global issue under control. To better understand how PFAS move through these engineered systems and identify sources to collaboratively reduce the influent contamination, better analytical capabilities are required to measure the broad spectrum of the chemical class in the complex environmental matrices and at toxicologically relevant concentrations. As regulatory and public pressures mount on PFAS contamination, WTPs and WRRFs offer a logical opportunity to stop the environmental cycling of these chemicals, but much work is required to do so in an efficient, economical, and socially just manner.

No data, models, or code were generated or used during the study.