An Overview of the Formation of PFOA and PFOS in Drinking-Water and Wastewater Treatment Processes

Per-and polyfluoroalkyl substances (PFAS) are synthetic surfactants with a fluorinated alkyl moiety of varying chain length and varying hydrophilic functional groups attached to the fluorinated chain. They are oleophobic and resistant to strong acids/bases, and have been massively produced for decades for various industrial and commercial products (Paul et al. 2009; Post et al. 2012; Prevedouros et al. 2006; Wang et al. 2017). Once released to the environment, PFAS distribute themselves among different environmental compartments, and are transported to drinking-water sources. The widespread occurrence of PFAS in the environment coupled with their known toxicity to humans has aroused great concern in both the scientific community and the public at large (Blum et al. 2015; Buck et al. 2011; Vestergren and Cousins 2009).

Two primary PFAS, namely perfluorooctanoic acid (PFOA) (${\mathrm{C}}_7{\mathrm{F}}_{15}\mathrm{COOH}$) and perfluorooctane sulfonate (PFOS) (${\mathrm{C}}_8{\mathrm{F}}_{17}{\mathrm{SO}}_3\mathrm{H}$), have been measured in drinking water around the globe at concentrations ranging from a few to several tens of ng/L (Andrews and Naidenko 2020; Ericson et al. 2009; Eschauzier et al. 2012; Hoffman et al. 2011; Hölzer et al. 2008; McDonough et al. 2021; Quinones and Snyder 2009; Takagi et al. 2011; Xiao 2017). They are not readily removed by conventional drinking-water treatment processes and are stable against physical and chemical degradation at circumneutral pH (6–9) (Eschauzier et al. 2012; Rahman et al. 2014; Schröder and Meesters 2005; Takagi et al. 2011; Xiao et al. 2013). Both chemicals have been observed in $>95\%$ of the blood samples collected during US national surveys (NHANES 2014) at health-relevant concentrations (Grandjean et al. 2012; Hoffman et al. 2010; Ji et al. 2012; Joensen et al. 2009; Kim et al. 2011; Melzer et al. 2010; Steenland et al. 2010). The US EPA plans to regulate PFOA and PFOS under the Safe Drinking Water Act (USEPA 2021) to reduce the direct human exposure to these chemicals from drinking water.

Much less attention has been given to the indirect source of PFOA and PFOS in current regulatory efforts. Accumulating evidence has shown that both PFOA and PFOS can be generated from precursor compounds in chemical and biological processes (Dasu et al. 2012; Jin et al. 2020; Martin et al. 2006; Mejia-Avendano et al. 2016; Wallington et al. 2006; Xiao et al. 2018a; Zhu et al. 2019). There is also a concern that the in vivo metabolic transformation of precursors may be a significant source of indirect exposure to PFOA and PFOS (D’eon and Mabury 2011; Vestergren et al. 2008; Wang et al. 2009; Yeung et al. 2013). This editorial provides an overview of the de novo formation of PFOA, PFOS, and other PFAS chemicals in water and wastewater treatment processes and discusses the challenges and research opportunities for technological approaches to counter these challenges.

PFAS have been frequently detected in municipal wastewater, and the effluent from wastewater treatment plants (WWTPs) is a major source of PFAS in receiving water bodies (Boulanger et al. 2005; Clara et al. 2009, 2008; Guo et al. 2010; Houtz et al. 2016; Huset et al. 2008) where the compounds can bioaccumulate and biomagnify in aquatic food webs. The contamination of drinking-water sources by PFAS can occur as upstream WWTPs discharge effluent into rivers or lakes that serve as water supplies (Hu et al. 2016; Quinones and Snyder 2009). A number of studies have documented that the removal of PFOA and PFOS is insignificant by conventional wastewater treatment processes (Loganathan et al. 2007; Schultz et al. 2006b; Xiao et al. 2012; Yu et al. 2009). In fact, compelling evidence has shown that PFOA and PFOS can be generated during biological wastewater treatment, which is likely due to the biotransformation of precursor compounds (Guo et al. 2010; Murakami et al. 2009; Schultz et al. 2006a; Xiao et al. 2012). For example, a significant increase in the wastewater concentrations of perfluorohexanoic acid (PFHxA, C6) and PFOA (C8) from influence to effluent was observed in 22 out of 37 WWTPs (59%); in certain WWTPs concentrations of PFOA and PFOS have been found to increase by up to 1,200% (Fig. 1) (Xiao et al. 2012).

Previous researchers have attempted to understand biological degradation mechanisms of a few anionic and non-ionic precursor compounds, including 8:2 fluorotelomer alcohol (FTOH), 6:2 FTOH, and N-ethyl perfluorooctane sulfonamidoethanol (EtFOSE). However, the generation of PFOA and PFOS from these precursors were slow and/or insignificant. For example, the yield of PFOA from 8:2 FTOH is typically less than 6% (Dinglasan et al. 2004; Wang et al. 2005a, b); one study reported a yield of only 2.1% after 28-day incubation of 8:2 FTOH in an activated sludge reactor (Wang et al. 2005b). The biodegradation of 6:2 FTOH generates shorter-chained PFAS (e.g., PFHxA); however, the yield is $<5\%$ after incubation of 6:2 FTOH for as long as 3 months (Wang et al. 2011; Zhao et al. 2013). Researchers have also investigated EtFOSE, a component present in many PFAS-relevant surfactant products, as a possible PFOS precursor compound (Benskin et al. 2013; Boulanger et al. 2005; Mejia Avendano and Liu 2015; Rhoads et al. 2008). The yield of PFOS from EtFOSE is less than 1% after incubation of EtFOSE in an activated sludge reactor for 4–10 days (Boulanger et al. 2005; Rhoads et al. 2008).

The minimum yields of PFOA and PFOS from legacy precursor compounds (FTOHs and EtFOSE) suggest the presence of other unidentified precursors that can readily biodegrade to PFOA and PFOS during wastewater treatment with a hydraulic retention time of a few hours (Guo et al. 2010; Murakami et al. 2009; Schultz et al. 2006a; Xiao et al. 2012). Recently, a number of cationic and zwitterionic PFAS have been identified in PFAS-based surfactant products (D’Agostino and Mabury 2014; Xiao et al. 2017), aqueous film-forming foams (AFFFs) (Barzen-Hanson et al. 2017; D’Agostino and Mabury 2014; Place and Field 2012), and AFFF-impacted sites (Nickerson et al. 2021; Place and Field 2012). These compounds are structurally similar to PFOS and PFOA, except that the perfluoroalkyl chain is attached to a nonfluorinated moiety through a polar group. Biotransformation of cationic/zwitterionic precursor compounds to PFOA and PFOS during wastewater treatment have not been quantified. A few studies have found that cationic/zwitterionic PFAS can transform to PFOA and PFOS in chemical (Xiao et al. 2018), biological (Mejia-Avendano et al. 2016), and low-temperature thermal (Xiao et al. 2021) processes.

The generation of PFOA and PFOS has also been observed in the drinking-water disinfection processes (Appleman et al. 2014; Boiteux et al. 2017; Dauchy et al. 2012) [see more studies in Table 3 of Rahman et al. (2014)]. Boiteux et al. found a significant increase in the water concentration of perfluoroalkyl carboxylic acids (PFCAs), including PFOA, after disinfection by ozone or chlorine [see Fig. 4 in Boiteux et al. (2017)]. The authors concluded that 18%–77% of the mass of PFCAs after disinfection was caused by the transformation of unidentified precursors in surface water (France) other than legacy precursor compounds such as 8:2 FTOH (Boiteux et al. 2017). In a survey of 15 US water treatment plants, Appleman et al. found that the concentration of PFOA and PFOS in water was consistently higher after chemical disinfection treatments [see Fig. 1 of Appleman et al. (2014)]. Similarly, negative removals of PFOA and PFOS in drinking-water treatment processes have been reported in Japan, which was attributed to the transformation of precursor compounds (Takagi et al. 2008).

The mechanisms underlying the secondary formation of PFOA/PFOS from precursor compounds in disinfection processes are not well understood. It has been demonstrated that PFOA, PFOS, or both can be formed from polyfluoroalkyl amides or sulfonamides during drinking-water disinfection with chlorine or ozone (Xiao et al. 2018). During conventional and booster chlorination, polyfluoroalkyl amides transform to PFOA through a Hofmann-type rearrangement by reacting with chlorine. Upon ozonation, polyfluoroalkyl sulfonamides transform to both PFOS and PFOA through direct oxidation and a radical-mediated pathway (Xiao et al. 2018).

Furthermore, granular activated carbon (GAC) adsorption is a frequently used approach for treatment of PFAS-contaminated water at pilot- and full-scale operations (Rahman et al. 2014). Some studies have shown that bituminous coal-based reagglomerated GAC is better than coconut-based direct GAC for removing anionic PFAS species (e.g., PFOA and PFOS) from water (McNamara et al. 2018). Spent or exhausted GAC can be thermally reactivated or regenerated, where the carbon is heated with inert gases (e.g., ${\mathrm{N}}_2$), ${\mathrm{CO}}_2$, or steam (Marsh and Rodríguez-Reinoso 2006). Heating PFAS-laden GAC at high temperatures (${\ge 500}^{\mathrm{o}}\mathrm{C}$) is highly effective for decomposition of PFAS, including PFOA and PFOS (Duchesne et al. 2020; Sasi et al. 2021; Watanabe et al. 2018; Xiao et al. 2020). However, at low temperature conditions (${<400}^{\mathrm{o}}\mathrm{C}$), PFAS can transform to shorter-chained homologues or other PFAS species (Sasi et al. 2021; Xiao et al. 2021).

PFOA and PFOS account for only a small fraction of the total organic fluorine present in water and wastewater samples. Certain polyfluoroalkyl substances, including FTOHs, EtFOSE, and cationic/zwitterionic PFAS, can transform to PFOA and PFOS during chemical, biological, and thermal treatments. However, there are also unidentified precursors present in drinking water (Boiteux et al. 2017). The removal of precursors of PFOA and PFOS is the key to safeguarding drinking-water quality. The source water chemistry, the dose and type of disinfectants, and adequate pretreatments prior to the biological treatment or chemical disinfection are likely the main factors for controlling the formation of PFOA and PFOS in water and wastewater treatment processes.

Various advanced technologies have been developed to remove PFOA, PFOS, and their precursor compounds from drinking water. Although some of these treatments have been successful in research labs [see references in Espan et al. (2015)], most of these approaches have major economic or design challenges prior to implementation for treatment of natural waters containing colloidal particles and dissolved organic matter (NOM) at circumneutral pH. At present, GAC adsorption appears to be one of the very few scalable treatment options for PFAS. The author believes that PFOA, PFOS, and their precursor compounds could be controlled by preoxidation via ozone or advanced oxidation processes to transfer precursor compounds to perfluoroalkyl substances (e.g., PFOA and PFOS), followed by adsorption of PFAS by GAC and thermal treatment of PFAS-laden GAC at 700°C or above (Xiao 2019).

Conversely, although the thermal treatment holds promise for degradation of PFAS chemicals and reactivation of PFAS-laden GAC (Duchesne et al. 2020; Sasi et al. 2021; Watanabe et al. 2018; Xiao et al. 2020), carbon loss, energy cost, and volatile decomposition products of PFAS are potential concerns. Further studies are suggested to investigate alternative reactivation methods, such as chemical (Huling et al. 2005) or electrochemical (Karimi-Jashni and Narbaitz 2005) approaches, for PFAS-spent GAC.

Nanofiltration (NF) and reverse osmosis (RO) as advanced separation technologies are capable of removing PFAS from water (Appleman et al. 2013; Tang et al. 2006), although not every community can afford to include NF and RO in the water treatment system. The main residual produced from an NF or RO system is brine containing elevated levels of NOM and ionic strength. The removal of PFOA, PFOS, or their precursor compounds from NF/RO brine is rarely studied.

Furthermore, only limited data are available on the reaction of the cationic and zwitterionic precursor compounds with chemical disinfectants (Xiao et al. 2018). Systematic studies are needed to understand the fate and transformation of cationic and zwitterionic PFAS in water and wastewater treatment processes.

Treated water leaving the drinking water treatment plants is usually disinfected and stored prior to distribution. The reaction between precursor compounds and residual chlorine may contribute to a further concentration increase of PFOA and PFOS during water storage. Future studies are suggested to understand the fate and transformation of precursor compounds in storage facilities and distribution systems.

This work was supported by the US EPA Science to Achieve Results (STAR) Program (RD839660), US National Science Foundation Faculty Early Career Development Program (CAREER) Program (2047062), and Strategic Environmental Research and Development Program (SERDP) (ER21-1185).