The 18 linear PFAS measured in this study are listed in Table 1 and standards were purchased from Wellington Laboratories, USA. These 18 PFAS of interest include perfluoroalkyl ether carboxylic acids (PFECA), polyfluoroalkyl ether sulfonates (PFES), perflourinated sulfonic acid (PFSA), perfluorinated carboxylic acids (PFCA), and PFAS precursors. The three isotope-labeled internal standards (IS): d₃-NMeFOSAA, ¹³C₄ -PFOS, and ¹³C₂-PFOA, and four isotope-labeled surrogates ¹³C₂-PFHXA, ¹³C₃-HFPO-DA, ¹³C₂-PFDA, and d₅-NEtFOSAA were also obtained from Wellington Laboratories, USA. Standards and surrogates were added in a concentrated methanol solution. All reagents, unless otherwise noted, were of the highest grade offered by manufacturers (≥97% purity) and care was taken to minimize background contamination during sample preparation. High purity methanol (≥99.99%, Honeywell CHROMASolv LC MS) was purchased from Midland Scientific Corporation, USA. Tetra-butylammonium hydrogen sulfate (TBAS), sodium carbonate (Na₂CO₃), and sodium hydroxide (NaOH) were purchased from Aldrich Chemical (Milwaukee, WI, USA). Methyl tert-butyl ether (MTBE), formic acid, and acetonitrile (ACN) were purchased from Fisher Scientific (Pittsburg, Pennsylvania, USA). Highest purity nitrogen and argon gas (≥99.99%) were used.

The stainless-steel holders, cages, and POCIS were purchased from Environmental Sampling Technologies (EST) Inc., Missouri, USA. Hydrophilic-lipophilic balance (HLB)-type POCIS (poly[divinylbenzene]-co-N-vinylpyrrolidone; 200 mg; 30 μm; Waters, USA) were used for WWTP and surface water monitoring. Surface runoff grab samples were collected and extracted with Sep-Pak PS2 cartridges (styrene-divinylbenzene (SDVB) copolymer; 500 mg; 80 μm particle size; Waters, USA). Polypropylene (PP) filters housed in PP (0.45 μm pore size; 25 mm), 60 mL Sep-Pak reservoir, reservoir adaptors, QuEChERS cleanup salts (6 g MgSO₄; 1.5 g NaCH₃COO), and 15 mL dispersive solid phase extraction (SPE) tubes (900 mg MgSO4; 300 mg primary secondary amine, PSA) were purchased from Waters Corporation (Milford, Massachusetts, USA). A polypropylene (PP) manifold cover housing stainless steel needles was purchased from Millipore Sigma (St. Louis, Missouri, USA) and used exclusively for this method. Empty 6 mL PP SPE cartridges with 20 μm polyethylene (PE) frits and 30 mL capacity PP funnels were obtained from Millipore Sigma (St. Louis, Missouri, USA). Three hundred micro liter PP conical vials and pre-slit polytetrafluoroethylene silicone septum screw caps were used to store final extracts (Waters; Milford, Massachusetts, USA).

The WWTP serviced an average of 80,000 m³ day⁻¹ from a highly urbanized setting in Lincoln, Nebraska. After securing permission from the WWTP facility, POCIS were deployed in both the influent and effluent, as well as upstream of the effluent discharge and downstream in the effluent mixing zone (EMZ) in a stream with an average discharge of 215,617 m³ per day. The upstream site was 22 m above discharge and the EMZ site was 21 m below WWTP discharge. The EMZ is the area where the WWTP effluent is initially diluted within the receiving surface water. POCIS were deployed at the four sites for two sampling periods in 2020, sampling period one ranged from September 17th to October 1st and September 17th to October 15th, and sampling period two was from October 1st to October 15th. The WWTPs ultra-violet radiation (UV) disinfection period was from September 17th to October 1st and disinfection ceased after October 1st. POCIS deployment periods ranged from 14 to 28 days. Effluent POCIS deployed from September 17th to October 1st during UV disinfection was destroyed due to turbulent flow. Influent and surface water sites were collected during disinfection and after while effluent was only collected after disinfection ceased. POCIS were collected using nitrile gloves and immediately transported directly to the Water Sciences Laboratory in Lincoln, Nebraska. Exposed POCIS were stored at −20°C in virgin polypropylene bags until extraction.

Six agricultural test plots (University of Nebraska Rogers Memorial Farm) were established 2020. Each 25.2 m² plot had a slope of 4–6%. Grain sorghum was planted in each plot May 2020. Biosolids produced by the WWTP plant were collected for the land application study and were applied to plots in early August 2020 due to COVID-19 restrictions. Two biosolid plots received 177 kg of municipal biosolids and two reference plots received no amendments. Two hundred and fifty milliliter grab samples of runoff were collected after precipitation events between August-October 2020 in new PP bottles. Samples were stored frozen at −20°C until extraction. Due to COVID-19 pandemic restrictions on field and laboratory activities, runoff sampling was limited to 3 precipitation-induced runoff events and sample storage time ranged from 10 to 11 months. Four hundred and fifty-four gram of soil and a fully grown grain sorghum plant were sampled from each plot. Plant tissues were collected from each plot and the top 10 cm of soil was sampled at the conclusion of the growing season in 2020. Samples were collected using stainless steel apparatus and nitrile gloves. All biosolid, plant, and soil samples were frozen in polyethylene zipper bags and stored at −20°C.

Uptake rates (R_s) used to estimate TWA concentrations for 11 of the 18 analytes of interest quantified in POCIS extracts were published by Gobelius et al. (2019). The TWA concentration of each analyte that was provided a sampling uptake rate (R_s) was calculated as follows:

Naderi Beni et al. (2020). Equations used to estimate mass flow are described by Pan et al. (2016) and are supplied in Supplementary material.

Processing procedures for POCIS samples follow Gobelius et al. (2019) and included steps to minimize potential for cross-contamination. Briefly, POCIS devices were thawed before extraction. An empty 6 mL PP SPE cartridge with 20 μm PE frits were washed with 5 mL of methanol and 10 mL of purified reagent water. A 20 μm PE frit was placed tightly at the cartridge bottom and a 30 mL PP funnel was placed on top. The PES membranes of each POCIS were carefully separated and the HLB sorbent carefully emptied into the funnel. The funnel was rinsed with 30 mL of purified reagent water to ensure all resin particles were transferred. A second PE frit was then placed on top of the resin. The resin was then spiked with 4 ng of surrogate spike. Quality control samples received 1 ng of analyte spike to measure recovery (see Supplementary material for details). Residual water in the cartridge was removed by pulling air through for 2–30 min (<5 in Hg). A PP manifold cover was attached to elution manifold and 15 mL PP tubes were eluted with 20 mL of methanol. Eluent was slowly evaporated to dryness by heating (60°C) under a stream of nitrogen and then spiked with 2 ng of internal standard mix (see Supplementary material for details).

Water extraction methods generally followed procedures described by Shoemaker and Tettenhorst (2020). Samples were thawed and mixed by hand for 5 min. Turbid samples were filtered with a PP filter (0.45 μm pore size, 25 mm) and PP syringe. The sample was weighed and spiked with 4 ng of surrogate spike. Quality control samples, prepared by weighing purified reagent water, received 1 ng of analyte spike. Each cartridge was conditioned by passing 25 mL of high purity methanol and then 30 mL of purified reagent water through the cartridge without allowing the cartridge to go dry. A 60 mL Sep-Pak reservoir was attached to the cartridge and 60 mL of sample was poured into the reservoir. The sample was pumped by aspiration through cartridge at ~3 mL/min until all sample was extracted. The sample container was rinsed with two aliquots of 7.5 mL of purified reagent water and drawn through the cartridge. The cartridge was dried by pulling room air for 5 min (10–15 in. Hg), and the empty container weighed to determine total sample size.

Solid samples (biosolids and soil) were solvent extracted following procedures described by Rankin et al. (2016). Soil and biosolids were air dried and crushed into a fine homogonous mix using a clean mortar and pestle. A 5-gram (0.3 g for biosolids) sample was weighed into a 15 mL PP tube and spiked with 4 ng of surrogate spike. QC samples received 2 ng of analyte spike. Four hundred microliter of 2 M sodium hydroxide (NaOH) solution and 8.5 mL of 90:10 acetonitrile/water was added to each sample. Solutions were vortexed to mix, sonicated in an ice bath for 1 h, and then shaken on a rotary shaker for 15 h followed by centrifuging at 3,000 rpm for 15 min. Supernatant was decanted and this process was repeated by mixing solids with an additional 400 μL NaOH solution and 8.5 mL of 90:10 acetonitrile:water. Extracts were evaporated to near dryness (100–300 μL) under a stream of dry nitrogen gas. Final cleanup using TBAS ion-pairing was accomplished by mixing the extract with 4 mL TBAS ion solution (0.25 M Na₂CO₃, 0.5 M TBAS). Extract solution was vortexed with 5 mL MTBE and this final solution was separated at −20°C overnight. The supernatants were decanted. Aqueous extracts were finally filtered with a PP filter (0.45 μm pore size, 25 mm). Final solution cleanup step was repeated. The supernatants were combined and blown to dryness by heating under a stream of nitrogen gas. Once dry, each tube was spiked with 2 ng of IS (see Supplementary material for details).

Lastly, the following describes plant tissue extraction. All plant tissue samples were lyophilized, crushed, and homogenized before processing using procedures described by Organtini et al. (2021). A 0.5 g sample was weighed into a 50 mL PP centrifuge tube and spiked with 4 ng of surrogates. Fortified samples received 4 ng analytes to measure recovery (see Supplementary material for details). Tissue was mixed with 5 mL purified reagent water, 10 mL acetonitrile, and 150 μL formic acid by vortex and shaking for 1 min. QuEChERS salts were added, vortexed, and shaken for an additional 5 min. Mixtures were centrifuged at 3,300 g for 15 min, and supernatant decanted into 15 mL dispersive SPE tubes. Mixtures were vortexed, shaken for 5 min and centrifuged again at 3,300 rpm for 15 min. Supernatants were decanted into 15 mL PP tubes and evaporated to dryness by heating under a stream of nitrogen. Each tube was spiked with 4 ng of IS prior to dissolving in mobile phase (see Supplementary material for details).

Dried extracts from all methods were dissolved in 300 μL 96:4% methanol:water and then transferred into a 300 μL conical PP autosampler vial using a PE transfer pipette. Each extract was analyzed for the 18 PFAS analytes using liquid chromatography tandem mass spectrometry (Table 1).

Chromatographic separation was performed using an Acquity H-Class Plus ultrahigh pressure liquid chromatography (UPLC) system with a 1.7 μm Premier BEH C18 column and equipped with a 2.1 × 50 mm isolator column interfaced to a Xevo TQ-S Micro triple quadrupole mass spectrometer system equipped with an Unispray™ source operating in a negative ion detection mode (Waters Corporation, Milford, MA). Gradient separation was accomplished at 35°C with a mobile phase flow rate of 0.30 mL/min. Mobile phase solvents: (I) 2 mM ammonium acetate in water; (II) 2 mM ammonium acetate in methanol. Initial conditions of 95%I and 5%II, hold until 0.5 min, then change to 75%I and 25%II, hold until 3 min, at 3 min through 6 min 50%I and 50%II, 6–6.5 min switch to 15%I and 85% II, 6.5–9 min 5%I ad 95%II, and 9–12 min back to initial conditions of 95%I and 5%II. Total analysis time is 12 min. Mass spectrometer settings were: collision gas: argon at 4.0 × 10⁻³ Torr; desolvation gas: N₂ at 900 L/h; desolvation temperature: 400°C; cone gas: N₂ at 10 L/h; source temperature: 150 °C; and capillary voltage: 0.5 kV. Collision energies, cone voltages, ion transitions, and retention times are listed in Supplementary Table 1. Injection volume was 5 μL. Peak integration and calculations of concentrations against the standard curve were performed using Waters Masslynx software v4.2.

A total of 7 standards spanning 0- to 20-ng/mL were used for the instrumental calibration curve. Using the standard deviation of the lowest calibration standard instrument detection limits ranged from 0.03 to 0.5 picograms (pg) on-column (Table 2). Method detection limits (MDLs) of water, solids and plant tissue samples were calculated using replicate analysis of a low-level fortified matrix (USEPA, 1986).

Statistical analyses included t-statistics and one-way ANOVA comparisons followed by Tukey's HSD. Significance level was set at α equal to 0.05. Data was normalized by natural log transformation on a need-by-need basis to improve residual plots. All recorded and detectable data values were utilized for statistical analysis (Helsel, 1990). Values below detection limits were allocated as “not detected, ND.” Sample sizes (n) varied throughout environmental matrices ranging from 1 to 4 for each sampling collection. All statistical computations were performed using statistical software (MiniTab Statistical Software, State College, PA; R-Studio, Bos, MA).

Previously reported extraction methods for PFAS in water (Shoemaker and Tettenhorst, 2020), POCIS (Gobelius et al., 2019), solids (Rankin et al., 2016), and plant tissue (Organtini et al., 2021) use offline SPE procedures, liquid extraction with organic solvents utilizing sonication and shaking, QuEChERs and dispersive SPE employing liquid extraction, and extraction with organic solvents. Offline SPE procedures use surrogates to monitor recovery of analytes throughout sample prep and concentration steps. The use of deuterated internal standards helps correct for ion suppression and enhancement effects influenced by co-extracted contaminants as well as improves accuracy of analyte quantitation as final extracts may contain a suite of co-extracted contaminants which will affect ionization efficiency and may possibly influence recoveries (Snow et al., 2013). Field sample surrogate recoveries ranged widely between the four matrices evaluated indicative of complex matrices and the co-extraction of non-target contaminants. The average and standard deviation of surrogate recoveries for non-potable water, solids, POCIS, and crop samples were 93.6 ± 63, 63.1 ± 28, 101 ± 75, and 200 ± 121 per cent, respectively. Each sample matrix was validated. Offline SPE, IPE, QuEChERS salts, and dSPE were utilized for extraction and cleanup, and sensitivity of each compared for environmental matrices. Calculated MDL values and percent recoveries of 18 analytes of interest are reported in Table 2. MDLs values in this study for non-potable water, solids, and crop ranged from 0.308 to 0.750 ng L⁻¹, 0.0279–0.687 ng g⁻¹, and 1.79–9.74 ng g⁻¹, respectively (Table 2). Average recoveries for 18 PFAS analytes for non-potable water, solids, and crop were 105 ± 4, 78.0 ± 38, and 114 ± 18 per cent, respectively (Table 2).

Of the 18 PFAS measured, 14 were detected at least once. 11Cl-PF3OUdS, 9Cl-PF3ONS, ADONA, and HFPO-DA were not detected within any POCIS extracts or municipal biosolids (Supplementary Table 3, Table 3). The majority of detected PFAS followed an increasing concentration trend from the influent wastewater to the effluent wastewater. Of the 14 measured PFAS, in ng/POCIS 12 out of the 14 (85.7%) were at measured at higher levels in in the effluent POCIS in comparison to the influent wastewater. Only two PFAS (PFOS, PFHxS) showed no observable significant differences between the influent and effluent concentrations (p-value > 0.05).

Using uptake rates provided by Gobelius et al. (2019), we calculated TWA concentrations for 11 of the 18 PFAS of interest. The average TWA Σ₁₁ PFAS found in WWTP influent and effluent sites, respectively, were 27.9 and 132 ng L⁻¹ across both sampling periods (Table 4). Within the WWTP there were no differences in the average Σ₁₁ PFAS between sampling periods. These results suggest that TWA concentrations varied widely within the WWTP. In the influent average TWA concentrations ranged from a low of 0.0403 ng L⁻¹ for PFDoA to a high of 12.5 ng L⁻¹ for PFOS (Table 4). Effluent TWA concentrations ranged from 0.621 ng L⁻¹ for PFDoA to 39.5 ng L⁻¹ for PFHpA (Table 4).

The mass of dissolved PFAS entering and leaving the WWTP was estimated using calculated TWA obtained from the POCIS measurements. In the influent flow the average mass loading of the Σ₁₁ PFAS was 2.60 g day⁻¹, increasing to 10.5 g day⁻¹ in the effluent. Represented in Figure 1, the influent mass loading for the Σ₁₁ PFAS slightly decreased from sampling period 1 (3.39 g day⁻¹) to sampling period 2 (1.81 g day⁻¹), likely influenced by a minor decrease in influent flow (Supplementary Figure 1) All 14 PFAS detected within the WWTP were detected at the two surface water sites, with a trend of increasing concentrations from upstream to the EMZ. Twelve out of 14 PFAS were detected at higher concentrations in ng/POCIS in the EMZ compared to upstream. PFNA has no significant differences between sites (p-value > 0.05), while PFHxS was at higher concentrations upstream than within the EMZ.

Average TWA Σ₁₁ PFAS found upstream and in the EMZ were 37.7 and 71.4 ng L⁻¹ through both sampling periods (Table 4). At upstream site, TWA during sampling period 1 was >2 and in the EMZ sampling period 2 was >1. The average TWA concentrations between the two surface water sites was similar. Upstream ranged from 0.21 ng L⁻¹ for PFUnA to 17.8 ng L⁻¹ for PFHxS and EMZ ranged from 0.482 ng L⁻¹ for PFUnA to 16.3 ng L⁻¹ for PFHpA. The median TWA for the EMZ (5.29 ng L⁻¹) was higher compared to the median at the upstream (1.98 ng L⁻¹) site.

Estimated mass loading of detected compounds (Σ₁₁ PFAS) upstream was 5.06 g day⁻¹, while the downstream EMZ was roughly 3× higher at 15.5 g day⁻¹. Upstream and EMZ mass loading showed slight temporal differences during our four-week sampling campaign. Upstream, mass loading decreased from sampling period 1 (6.63 g day⁻¹) to sampling period 2 (3.52 g day⁻¹), as shown in Figure 1. In contrast, the EMZ mass loading rate increased from sampling period 1 (11.6 g day⁻¹) to sampling period 2 (19.3 g day⁻¹). Short and long chain PFAS occurrence was variable between the four aqueous sites, as shown in Table 5. In the influent, effluent, upstream, and EMZ short chain PFAS compromised 27.8, 53.6, 60.7, and 37.8 percent, respectively, while long chain PFAS and 2 precursors compromised 72.1, 46.4, 39.3, and 62.2 percent, respectively.

Land applied municipal biosolids had detectable concentrations of 13 PFAS, and Σ₁₃ PFAS in biosolids was 22.9 ng g⁻¹ dw. Detectable biosolid concentrations ranged from a low of 0.0382 ng g⁻¹ dw for PFTrDA to a high of 9.40 ng g⁻¹ dw for PFOS (Table 3). Class type and chain length of PFAS sequestered within biosolids may be important in handling and use of municipal biosolids (Higgins and Luthy, 2006; Venkatesan and Halden, 2013; Tavasoli et al., 2021). In this study, PFSAs compromised 45.2% of the Σ PFAS in biosolids, with 41% of the Σ PFAS detected being contributed by PFOS. Long chain PFAS contributed to 77.5% of the Σ PFAS in municipal biosolids. Mass loading of PFAS into biosolids at the WWTP was estimated (equation in Supplementary material) and results are shown in Figure 2. The mass load of the Σ₁₃ PFAS in dewatered biosolids was 72.8 g day⁻¹ dw, based on 12,701 kg of biosolids produced daily. The individual PFAS with the highest mass loading rate was PFOS at 29.9 g day⁻¹ dw, followed by PFHxA at 13.3 g day⁻¹ dw (Figure 2). Of the 18 PFAS measured, 12 PFAS were detected between the surface soil and runoff samples (Table 3). No PFAS were detected in the grain sorghum plant tissue grown on the plots or surrounding field (controls).

In surface runoff samples from test plots treated with biosolids, the average Σ₁₂ PFAS in August was 19.0 ng L⁻¹ and in September it was 7.95 ng L⁻¹. In the soil on biosolid treated plots, the Σ₁₂ PFAS in the top 10 cm in October was 0.176 ng g⁻¹ dw⁻¹. The total measured Σ₁₂ PFAS concentrations significantly decreased between runoff events sampled in August and September (p-value < 0.05). During the August sampling event PFOA was the highest detected PFAS at 4.22 ng L⁻¹. While in September, PFOS was found at the highest concentration at 2.23 ng L⁻¹. An estimated total of 4.06 mg of PFAS was land applied in 176 kg of municipal biosolids resulting in 0.126 mg of Σ₁₂ PFAS in surface soil and an average loss of ~0.0005 mg of Σ₁₂ PFAS in runoff. In the surface soil, individual PFAS mass ranged from 0.00100 mg for PFDoA to 0.0358 mg for PFHxA. Residual or background contamination by PFAS was indicated in control plots with detectable concentrations of PFAS. Runoff water in August on control plot had an average Σ₁₂ PFAS of 9.57 ng L⁻¹ and in September the concentration decreased to 3.15 ng L⁻¹. In the soil control plots the average Σ₁₂ PFAS was 0.130 ng g⁻¹ dw.

Based on limited sampling in the present study, 14 of the 18 measured PFAS occurred in raw and treated wastewater samples. Σ₁₄ PFAS concentrations significantly increased in concentration from the influent to effluent. As with many trace organics, removal efficiency of PFAS in WWTP is low and many individual compounds can increase within the WWTP, consistent with other findings (Pan et al., 2016; Gallen et al., 2018; Tavasoli et al., 2021). Past studies have reported that PFAAs may increase in WWTPs after secondary treatment stages possibly due to the biodegradation of PFAS precursors, similar to findings stated here (Sinclair and Kannan, 2006; Lenka et al., 2021; Tavasoli et al., 2021). For example, the 8:2 fluorotelomer has degradational byproducts that can be a suite of PFBA, PFPeA, PFHxA, PFHpA, or PFOA (Mueller and Yingling, 2018). Our study found Σ₁₁ PFAS in the effluent was more than four times the average TWA concentration detected in the influent implying that the treatment plant was a source of PFAS in effluent.

Concentrations and occurrence of PFAS in this study are similar to those reported in other studies. For example, Tavasoli et al. (2021) reported a max Σ of PFAS at 198 ng L⁻¹ in a study of a New Hampshire WWTP. In six Californian WWTPs sampled by Houtz et al. (2016), they report an average concentration ranging from 3.5 ng L⁻¹ for PFDA to 26 for PFHxA ng L⁻¹. Our study reported a slightly wider range of average concentrations (0.621 ng L⁻¹ for PFDoA; 39.5 ng L⁻¹ for PFHpA). In California following an annual FAA required testing of airport firefighting foams, Houtz et al. (2018) detected PFBS at higher concentrations in effluent (~2–10 μg L⁻¹) than detected in this study (30 ng L⁻¹).

Estimated daily mass flow rates of 11 PFAS within the WWTP suggest that average flow of the Σ₁₁ PFAS increased ~7.90 g day⁻¹ from influent to effluent. The daily maximum loading rate for the Σ₁₁ PFAS in the influent was 3.68 g day⁻¹ and in the effluent was found to be 11.1 g day⁻¹. Differences in community and industrial uses of PFAS laden products may affect predominant PFAS chain length and classes which occur in influent. Although industry is slowly phasing out long chain PFAS and replacing with short chain PFAS (KEMI, 2015), our results suggest that long chain PFAS still predominate. Tavasoli et al. (2021) found effluent to contain more shorter chain PFAS, similar to results indicated in this study.

Comparable levels of 14 PFAS detected within the WWTP were also detected upstream and in the EMZ of the adjacent surface water system. This finding suggests similar sources of PFAS may be introduced to surface waters and wastewater. The occurrence of low levels of PFAS upstream of the discharge may be a result of previous deposition from storm run-off. For example, Codling et al. (2020) identified stormwater runoff as a contributor of PFAS to respective surface water systems that receive this runoff. The Σ₁₄ PFAS occurred at higher levels within the EMZ compared to upstream. This finding suggests that effluent discharge may be a significant source of PFAS to our surface water systems. Ahrens et al. (2009) indicated that levels of PFAS in adjacent surface waters may increase after receiving inputs from municipal effluent discharge and are also influenced by industrial discharge and surface runoff.

Variation in human usage of PFAS products may influence seasonal variations in the detection of individual PFAS. For example, in this study the Σ PFAS at our EMZ in time period 2 was 34.1 ng L⁻¹ higher than EMZ time period 1. Our limited sampling time frame allowed for a short glance into temporal variations of PFAS reaching our surface waters.

The PFAS measured in the surface waters in this study have been found in other studies of surface water. PFOS, PFHpA, PFHxA, PFDA, and PFOA were found in the Yangtze and Pearl River in China (So et al., 2007). PFBS, PFOS, and PFOA in the Mississippi River (Nakayama et al., 2010) and Cape Fear River Basin in North Carolina (Heinze et al., 2008). PFOS and PFOA were found at highs of 245- and 125-ng L⁻¹ in Mississippi River Basin, higher than concentrations (PFOS: 3.94–15.8 ng L⁻¹; PFOA: 3.31–3.77 ng L⁻¹) presented here (Nakayama et al., 2010). PFOA, PFHxA, and PFHpA in Rhine watershed in Switzerland (Gobelius et al., 2018), and PFOS in Vaal River in South Africa (Groffen et al., 2018). So et al. (2007) found low PFHpA and PFOS concentrations (0.074–9.2 ng L⁻¹, <0.01–14 ng L⁻¹) in the Yangtze and Pearl River in China which were similar to those in the present study. PFAS classes vary widely in their production and environmental release.

Gobelius et al. (2018) found PFCAs and PFSAs to be dominate PFAS in Swedish surface waters, similar to our findings shown in Table 5. At the upstream site, 71.2% of the detected compounds were in the group PFSAs and this decreased to 61% within the EMZ. Indicating that PFSAs dominated compounds detected in Nebraskan surface water in this study. Contrastingly, Ahrens et al. (2009) determined that PFCAs were the dominating group detected in their study. Most surface water studies to date are in a highly urbanized area impacted storm runoff, WWTP effluent, and partially industrialized areas (Zushi and Masunaga, 2009, 2011).

Supplementary Figure 4 presents the ambient temperature and quantity of precipitation for sampling locations. Surface water levels can be influenced by ambient air temperature and runoff from local precipitation events, which might have affected PFAS concentrations. Especially during EMZ-2 as it contained highest quantity of precipitation and greatest number of precipitation events.

Municipal biosolids are known to contain a variety of PFAS, often at temporally varying concentrations (Venkatesan and Halden, 2013). Sampling in this study took place in August while sampling of influent and effluent wastewater took place over a month later. PFBS was not detected in biosolids but was found in influent and effluent. Comparatively, NEtFOSAA, PFTrDA, and PFUnA were detected in biosolids and effluent, but not within the influent flow. These results are possibly indicative of temporal variation in inflow PFAS concentrations into the WWTP, precursor biodegradation (Sinclair and Kannan, 2006), differences in adsorption potentials between chain lengths and PFAS groups (Higgins and Luthy, 2006; Pan et al., 2016; Tavasoli et al., 2021), and desorption rates.

The estimated accumulation rate of PFAS within biosolids occurred at a mean rate of 72.8 grams day⁻¹. PFOS was the most abundant PFAS detected in our biosolids. This finding was similar to findings in a mass survey which incorporated United States biosolids from 94 WWTPs (Venkatesan and Halden, 2013). Their study reported that the average PFOS concentration in the United States was 403 ng g⁻¹ dw, nearly 14 times greater than the concentration reported in this study (Venkatesan and Halden, 2013). It is likely that PFOS was a dominant PFAS due to the understanding that long chain PFAS have greater potential to adsorb to particulates and organic matter due to increased hydrophobic properties, and the sulfonate moiety of PFSAs increases adsorption relative to PFCAs (Higgins and Luthy, 2006; Pan et al., 2016; Tavasoli et al., 2021).

Land application of PFAS laden municipal biosolids resulted in PFAS at detectable levels in the surface runoff of our biosolid treated plots. The results in this study suggest that PFAS may be transported through runoff from biosolids-treated soils. Run-off concentrations of the total Σ₁₂ PFAS residue decreased over time (August to September) suggesting that losses will diminish as easily-mobilized contaminants are lost from treated soils on a temporal scale due to rainfall events. Three rainfall events occurred in August (30.7 mm) and September had 3 events (62.5 mm), with only 1 event producing enough volume to sample each month. Gottschall et al. (2017) reported PFOS in soil throughout a yearlong study, and Chu and Letcher (2017) also suggested that soils applied with municipal biosolids may instigate the discovery of PFAS in these amended soils.

Figure 3 shows that PFHxA (0.0358 mg) and NMeFOSAA (0.0351 mg) comprised the highest masses in the 0–10 cm sampled portion of the soil column (Figure 3). Based on our estimates, transport of nearly 3.93 mg or 96.9% of PFAS mass applied may have been retained in biosolids on soil surface or lost through other pathways such as degradation, infiltration, or volatilization. Previous studies have reported possible PFAS movement through infiltration (Pepper et al., 2021) after land application of biosolids.

No detectable levels of the 18 PFAS were found in grain sorghum samples grown on test plots. Similarly, Gottschall et al. (2017) did not detect PFAS in crops (maize, soybean, wheat) on fields applied with municipal biosolids. Other studies have reported conflicting results, noting detectable concentrations of PFAS in differing crop types (lettuce, tomato, wheat, pumpkin) (Blaine et al., 2014; Lee et al., 2014; Wen et al., 2014). Our result of no detection with grain sorghum may have been influenced by the late application of municipal biosolids on to our biosolid plots. Rankin et al. (2016) reported widespread occurrence of PFAS in surface soils (0–10 cm) throughout the US. In comparison, the control plots contained low concentrations of 10 PFAS in the surface runoff and in the soil column. Degradation of PFAS precursors, atmospheric deposition, contaminated precipitation, and pesticide application are possible sources of detected background levels of PFAS (Kim and Kannan, 2007; Gottschall et al., 2017; Arinaitwe et al., 2021; Nguyen, 2021).