Occurrence of Poly-and-Perfluoroalkyl Substances in United States Water Resource Recovery Facilities

Learning Objectives:
1. Understand the occurrence of PFAS at US water resource recovery facilities (WRRFs)
2. Understand typical temporal fluctuations
3. Understand partitioning behavior and role of precursors through of treatment.
Introduction
The presence of poly- and perfluoroalkyl substances (PFAS) in water resource recovery facilities (WRRFs) has been widely reported. However, comprehensive quantitative data on specific PFAS compounds, their fate and phase partitioning through WRRF treatment processes, and the factors that control PFAS distribution in finished biosolids compared to the effluent outfall remain poorly understood. The absence of this fundamental information is a critical barrier for utilities to effectively manage and respond to a rapidly evolving public perception and regulatory climate related to PFAS. A significant body of research on PFAS fate and transport in the environment has accumulated in recent years and many of these recently developed tools and insights can be leveraged to investigate much-needed information specific to WRRFs.
Status:
Sampling has been completed. Results have been compiled for the occurrence survey at 38 WRRFs and temporal variability assessment at 14 WRRFs. The detailed mass balance assessment at 9 WRRFs has been conducted and data analysis is currently underway.
Methodology:
Thirty four utility partners representing 38 facilities were involved spanning most US geographies and climates. In the first phase of the study, these 38 facilities collected 24-hour flow-weighted composite samples from the WRRF influent and effluent, in addition to a single-point biosolid grab sample. Samples were collected by the facilities following strict quality control procedures and adhering to US EPA SOP EMAB0113.01. For the second phase of the study, 14 facilities continued to sample their influent and effluent on a monthly basis and their biosolids quarterly for 12 months. Samples were analyzed for 32 target and 1,500 suspect compounds using high HPLC-QToF-MS/MS with a mass spectral library (6,7,8). Analytical replicates were conducted in duplicate for 80% of the samples and triplicate for 20% of the samples. In addition to the PFAS data analyzed, utilities provided information regarding their treatment processes and process flow diagrams, historical flows, collection system point source dischargers, and water and solids quality data on the day of sampling. Finally, a detailed mass balance investigation was conducted at 9 WRRFs to characterize PFAS partitioning through various unit processes through treatment. This included influent, primary treatment, secondary treatment, effluent and biosolids at the 9 facilities, as well as additional treatment trains specific to each facility. Aqueous, solids, colloidal and volatile PFAS samples were collected and analyzed for 32 target PFAS compounds, total oxidizable precursors, and/or extractable organic fluorine.
Findings:
The sum of target PFAS in the influent and effluent samples typically ranged from 50 ng/L to 250 ng/L, though a couple outliers had concentrations greater than 500 ng/L (Figure 2). The sum of target PFAS was not significantly higher in the influent samples relative to the effluent samples (t(74) = 0.797, p = 0.427). Wastewater effluents were generally enriched in perfluorinated carboxylic acid (PFCAs) relative to the influent, and, in particular, short chain (C4-C6) PFCAs. PFCAs are known preferential degradation products from PFAA precursor compounds (2). Precursor concentration largely decreases between influent and effluent samples. For example, 6:2 fluorotelomer sulfonamide (FtS) was a precursor that was present at the highest concentrations at the influent and effluent samples and was largely responsible for this discrepancy. As shown in Figure 3, biosolids were enriched with precursors when compared to influent and effluent samples, suggesting that there is potential for transformation of precursors once biosolids are land applied. Notably, long chain PFAAs were present at higher concentrations than short chain PFAAs in biosolids. This is likely due to longer chain PFAAs being more strongly associated with solids due to hydrophobic interactions of the longer perfluorinated carbon chain relative to short chain PFAAs. To further examine the fate of target PFAS detected in WRRFs, PFAS mass loading expressed as fluoride was calculated for each facility based on influent and effluent flowrates on the day of sampling and average annual biosolids production. The results presented in Figure 3 suggest that the majority of target PFAS leaving WRRFs were associated with the aqueous effluent, rather than partitioning into biosolids. Sixteen WRRFs had an increase in total detected PFAS in the effluent relative to influent. PFCA concentrations generally increased in the effluent as opposed to the influent, while precursors decreased in the effluent. Transformation of precursors into short chain PFCAs might be a key element in the efforts to close mass balance of PFAS in these WRRFs. For example, at Facility 38, 29,640 mg/day of 6:2 FtS entered the influent, but only 33% (9,710 mg/day) was detected in the effluent and none was detected in the biosolids, suggesting precursor transformation occurred. Relatively, low mass concentrations of the C4 to C6 PFCAs (e.g., PFBA, PFPeA and PFHxA) were observed in the influent (8,340 mg/day), but the total mass for these compounds more than doubled in the effluent (17,609 mg/day). However, the decreased concentration of 6:2 FtS in the effluent did not appear to solely account for the mass of short chain PFCAs present in the effluent. Other precursors including fluorotelomer carboxylic acids (FTCA) and other suspect analytes were detected that may have contributed to the increased PFCAs. FTCAs are precursor compounds that form short-chain PFCAs. Their total mass concentration in biosolids were 23,561 mg/day and they were not detected in the effluent, suggesting transformation of these precursors was an important pathway. Of note was enrichment of precursor compounds in biosolids for the non-target compounds. Of those that were analyzed, 38 were detected in greater than 10% of samples. Perfluoroethane phosphonic acid (PFPA) and polyfluoroalkyl phosphoric acid diesters (di-PAPs) were detected in more than 90% of WRRFs and thus represent two classes that warrant attention. di-PAPs have been reported as precursors that may transform into short chain PFCAs (5,9). Analysis of suspect PFAS in the influent and effluent of the 38 WRRFs showed no presence of di-PAPs in any aqueous phase samples, thus agreeing with reports suggesting that biosolids are major sink for di-PAPs.
Significance:
This study showed that PFAS ranged between 50 and 250 ng/L and that were detected at all of the 38 WRRFs sampled. Within the WRRF treatment schemes, the decrease of short chain PFCAs from influent to effluent suggests that transformation may be occurring. The result from this study provides a benchmark for comparison of PFAS occurrence in US WRRFs and a context for understanding phase partitioning and precursor transformation. Ultimately, these results from this study may be used as the basis to develop appropriate methods and to develop appropriate methods and tools to decrease PFAS release to the environment.