Occurrence of perfluoroalkyl and polyfluoroalkyl substances in the water environment and their removal in a water treatment process

Perfluoroalkyl and polyfluoroalkyl substances (PFASs) such as perfluorooctanoic acid (PFOA), and perfluorooctane sulfonate (PFOS) have been recognized as emerging environmental pollutants (Giesy & Kannan 2001) because of their ubiquitous occurrence in aquatic systems such as municipal wastewater, surface water and tap water. PFASs can also be found in snow and rain water (Liu et al. 2009; Dreyer et al. 2010). In PFASs, the hydrogen atoms attached to the alkyl chain carbons have been replaced by fluorine (Schuetze et al. 2010). The carbon–fluorine bonds are very strong and have thermal and chemical stability. This stability makes the fluorinated compounds desirable for commercial use and also makes them persistent environmental contaminants due to their resistance to natural breakdown processes (Key et al. 1997); in this context, PFOS was listed as a persistent organic pollutant (POP) under the Stockholm Convention on POPs (Wang et al. 2009).

PFAS compounds have been used for many industrial purposes and consumer-related products such as firefighting forms, mining, photolithography, electronics chemicals, hydraulic fluid additives, floor polishes, photographic film, denture cleaners, soap, shampoos, cosmetics, coating of cooking pans, oil repellents for leather, paper, and textiles. Paul et al. (2009) suggest that 45,000 tons of PFOS was released into the global environment between 1970 and 2012. During that time, the 3M Company was the primary global producer of PFOS and PFOS-related substances, and although 3M phased out production of these chemicals in 2002, contamination from PFOS is expected to continue because of its own persistence and that of its transformation products.

The relative importance of the various pathways is not well understood; however, contamination of aquatic systems by industrial products like PFASs can only be through a limited number of mechanisms, e.g. through atmospheric transport and deposition, run-off from industrial sites, breakdown and run-off from domestic uses, and disposal of PFASs to sewer. Moreover, these chemicals are introduced into the aquatic environment as a result of degradation in the environment of precursor compounds or reaction impurities (Boulanger et al. 2005). PFASs have already been found to cause a variety of toxicological effects in humans (Suja et al. 2009). The concentration of PFOS and PFOA in cord blood and maternal pregnancy serum were negatively associated with human birth weight, ponderal index, and head circumference of new born babies (Apelberg et al. 2007). There have been several reviews on the toxicity and bioaccumulation of PFASs (Lau et al. 2004, 2007; Suja et al. 2009), so this review focuses on the contamination of the aquatic environment by PFASs and their removal by water treatment processes.

Several studies have reported on PFASs found in the water environment, which is a crucial environmental concern. PFOA and PFOS were detected in tap water in Georgia, where several secondary manufacturers are located, which produce non-woven, household additives, apparel, carpet, and household textiles (3M 2001). Lai et al. (2009) reported higher PFOS concentrations in tap water in the Kansai region than in the other 47 regions of Japan. Mak et al. (2009) studied tap water from different cities of China and they found that PFAS contamination decreased in the following order Shanghai > Wuhan > Nanjing > Shenzhen > Xiamen > Shenyang > Beijing. Atkinson et al. (2008) monitored the presence of PFOA and PFOS in tap water from different sites in England. They found that the highest levels of PFOS (162 ng/L) were observed south of Cambridge near an airstrip, indicating that airstrips are a potential source of perfluorinated chemicals (PFCs) to the environment. The PFOS and PFOA concentrations in the water environment throughout the world are shown in Table 1.

A comprehensive study on the occurrences of PFOS (0.89–5.773 ng/L) and PFOA (0.97–21.5 ng/L) at 79 sampling sites in Japanese rivers revealed widespread occurrences of these compounds (Saito et al. 2004). Hansen et al. (2002) found that the PFOS and PFOA concentration was 50 times higher in Tennessee River than in the Elbe River. This was because of fluorochemical manufacturing facility discharge into the Tennessee River bodies. Jin et al. (2009) found that the level of PFAS contamination was greater in urban rivers than in rural areas in China. Seasonal changes have a significant effect on the loading of PFASs in surface water. So et al. (2004) found that higher concentrations were detected in winter than in summer. A similar observation was reported by Tsuda et al. (2010) in Japan, and is attributed to both the effects of different sampling seasons and the presence of algal blooms in summer, which may affect the distribution of different PFAS forms.

PFOA have been found in the coastal water of Japan in different places such as Hiroshima Bay, Kin Bay, and Lake Shikatsu (Taniyasu et al. 2003). Hart et al. (2008) found that Tokyo Bay had 2–3 times greater PFOS concentration in coastal water than in offshore locations of Japan. So et al. (2004) found that the concentration of PFOS in coastal waters of Hong Kong was up to 3.1 ng/L, which is greater than those observed in Gyeonggi Bay (South Korea) but approximately 10 times less than those observed in Lake Shihwa (China). Finally, Kim & Kannan (2007) found that the PFOA concentrations in lake water were significantly greater than the concentrations found in rainwater, suggesting that snowfall meltwater is the source of PFOA contaminations in lake water in the spring.

PFASs have been observed in the effluent from municipal wastewater treatment plants (WWTPs). For instance, Loganathan et al. (2007) found that the concentration of PFOA and PFOS in secondary effluent was 227 ng/L and 22 ng/L, respectively. Another study in Singapore by Yu et al. (2009a) found that the concentration of PFOA and PFOS in WWTP secondary effluent was 1,057 ng/L and 461 ng/L, respectively. Pan et al. (2014) found that PFOA concentrations in Lake Taihu (China) were higher than PFOS concentrations (136 ng/L and 29.2 ng/L, respectively), perhaps due to the higher water solubility of PFOA.

The ability of wastewater treatment technologies to remove or degrade PFASs depends on the water treatment process. For instance, Ochoa-Herrera & Sierra-Alvarez (2008) reported that granular activated carbon (GAC) could effectively remove PFOS from aqueous solutions (Table 2). Flores et al. (2013) found that GAC removed PFOS and PFOA concentration with removal efficiency of 64 ± 11% and 45 ± 19%, respectively. The explanation for this was that the molecular size of PFOS is slightly larger than the PFOA which is slightly more hydrophilic. Takagi et al. (2011) reported that GAC could effectively remove the PFOS and PFOA when it was used for less than 1 year. In contrast, the effluent concentration of those compounds was higher than the influent concentration when it used for a longer time (>1 year), possibly due to growth of the biofilm into the pores and surface of the carbon, exhausting the adsorption capacity and possibly interaction between those compounds and between those compounds and the biofilm, producing lowered concentrations initially and higher concentrations after significant use.

The adsorption capacity of the GAC is also related to the temperature of the water sample. Takagi et al. (2008) found that the removal of PFOA was 36–56% in summer and 31–58% in winter. Knepper & Lange (2012) made a statement that the adsorption capacity of PFASs was low for GAC treatment because of the presence of organic matter in the water sample, considering the organic fractions of the dissolved organic matter adsorb more strongly than PFOS and PFOA compounds. The removal efficiency of PFOA and PFOS by GAC also depends on the volumes of activated carbon. Lampert et al. (2007) found in their batch test that more than 90% of both PFOA and PFOS were removed when the activated carbon was 0.1047 g or greater at 7 days of contact time. However, at a similar contact time, the reduction efficiency was decreased to about 50% for PFOA and 82% for PFOS when 0.0587 g GAC was used. Therefore, use of GAC with greater volume and suitable regeneration regimes appear to be the important parameters in the efficient removal of PFASs.

Deng et al. (2011) found that coagulation (polyaluminium chloride) could remove the PFOA from water. This was because some PFOA transferred from aqueous phase to solid phase. They also found that the addition of powdered activated carbon (PAC) before the coagulation process significantly enhanced the PFOA removal efficiency. This was explained as the negative PFOS adsorbing onto the PAC via electrostatic interaction, resulting in removal with the precipitate form in the coagulation process. A similar observation was made by Yu et al. (2009b) that PAC could effectively remove the PFOA concentration through the electrostatic interaction and hydrophobic interaction between them. A greater coagulation dosage (>60 mg/L) and lower pH (4.5–6.5) can enhance the PFAS removal (Xiao et al. 2013). Knepper & Lange (2012) mentioned that the combination of PAC with microfiltration/ultrafiltration is a promising advance for conventional PAC treatment for the removal of PFASs, because of improving sorption kinetics.

There are many studies demonstrating that the high pressure membranes such as those in nanofiltration (NF) and reverse osmosis (RO) can effectively remove the PFAS compounds (Tang et al. 2006; Zhao et al. 2013). Tang et al. (2006) investigated the use of RO membrane in removing PFOS from wastewater with the concentration range of 0.5 to 1,500 mg/L and they found that >90% reduction was achieved by RO. Tang et al. (2007) further studied the use of RO and NF to removal PFOS from wastewater and they got >99% and 90–99% removal for RO and NF membranes, respectively. Knepper & Lange (2012) mentioned that 99% rejection of PFOS was achieved by RO over a wide range of feed concentrations of 1–1,000 mg/L. They explained that the passage of up to 1% PFOS through RO membranes was by diffusion through the polyamide separation layer. Zhao et al. (2013) studied the NF membrane for removing PFOS from simulated surface water and found that the membrane can effectively reduce the concentration. They also investigated the effects of PFOS concentration, pH and calcium concentration on PFOS rejection. They found pH leads to an increase in the PFOS rejection when the pH increases from 3 to 9 which leads to an increase in the PFOS rejection from 86 to 95% and 93 to 97% in the presence of 0.1 mM Ca²⁺ and 1 mM Ca²⁺ at 0.4 MPa, respectively, because the increasing pH generally increases electrostatic interactions which play a role in PFOS rejection. Similarly, increasing calcium concentration decreased the permeate PFOS concentration because calcium ions bridge PFOS and the negatively charged membrane surface, which enhances the adsorption of PFOS on the membrane. Another study shows that average rejections of the PFASs were 99.3% for virgin RO membranes, but 95.3% for fouled RO membranes (Appleman et al. 2013). The trans-membrane pressure was not increased to maintain a constant flux across the membrane's surface for the fouled membranes. Although Tang et al. (2006) recommended that high flux RO membranes should be avoided when treating water with high concentrations of PFOS (>30 mg/l PFOS), these membranes normally have a low rejection effect and the advantage of a high flux cannot be maintained for a long time.

UV photolysis has been demonstrated to be effective at degrading PFOS and PFOA (Hori et al. 2004; Chen et al. 2007). Fujii et al. (2007) demonstrated that photocatalysis (reaction time up to 3 days) and advanced oxidation (with high temperature and pressure) could effectively degrade the PFOS and PFOA to CO₂ and ‘hydrogen fluoride’. Lampert et al. (2007) demonstrated more than 99% removal of PFOS and PFOA using anion exchange (AIX) resins (Table 2).

Conventional water treatment processes are not effective in removing PFASs. For instance, Zhang et al. (2011) found that PFASs are partly reduced by a drinking water treatment system in China where surface water was the source sample. Another study by Eschauzier et al. (2010) shows that river bank filtration did not remove the PFASs. Skutlarek et al. (2006) also clearly demonstrated that PFASs were not removed by water treatment steps and the concentrations in the surface waters corresponded to PFAS levels in drinking water. Recently Quinones & Snyder (2009) confirmed that the PFAS concentration of influent and effluent was similar in drinking water treatment plants in the USA, suggesting that treatment systems were ineffective in removing these compounds. Sand filtration and ozonation processes were also ineffective in removing PFOS and PFOA during drinking water treatment (Takagi et al. 2011). A similar result was found by Thompson et al. (2011a) in a water reclamation plant in South East Queensland, Australia. Appleman (2012) investigated the efficiency of PFAS removal using coagulation followed by sedimentation or dissolved air flotation (DAF) and/or filtration. They found that coagulation (aluminium sulfate) followed by sedimentation did not lead to PFAS removal, but where sedimentation was replaced by DAF, a 49% removal of PFOS was observed.

PFASs are globally distributed in the aquatic environment, including drinking water and wastewater. The clear pathways are still poorly understood for this group although the manufacture of PFASs and use of products containing PFASs are direct sources for the release of these compounds to the aquatic environment. Developed countries are more highly affected by PFAS contamination than developing countries due to their increased use of PFAS-containing goods. Several treatment processes are used in the removal of PFAS compounds. Conventional water treatments are ineffective in removing these compounds. Activated carbon may be a good choice to reduce the concentration if the activated carbon is used for a year. High pressure membrane systems are the most suitable processes for reducing these chemicals from water and wastewater and the extent of their reduction efficiency depends on the solution chemistry of the water sample.

PFASs are listed as chemicals of concern, so regular monitoring of the contamination level of these chemicals in the water environment is required for regulatory purposes. Although biological processes are mostly ineffective in removing these chemicals, there is a possible biological conversion of these compounds during water treatment processes. Therefore further investigation regarding the mechanisms of the biological degradation of these chemicals is required. Moreover, other treatment processes such as GAC and/or high pressure membranes are essential to equip the water treatment processes to remove these chemicals for drinking purposes and for safe water disposal to the environment.