Aqueous film forming foam and associated perfluoroalkyl substances inhibit methane production and Co-contaminant degradation in an anaerobic microbial community

Nicole J. M. Fitzgerald *a, Hanna R. Temme a, Matt F. Simcik b and Paige J. Novak *a
aDepartment of Civil, Environmental, and Geo- Engineering, University of Minnesota, 500 Pillsbury Drive SE, Minneapolis, Minnesota 55455, USA. E-mail:;; Tel: +1 847-791-2844 Tel: +1 612-626-9846
bSchool of Public Health, University of Minnesota, 420 Delaware St. S.E., Minneapolis, MN 55455, USA

Received 22nd May 2019 , Accepted 14th August 2019

First published on 27th August 2019

Aqueous film forming foams (AFFF) can contain gram per liter concentrations of per- and polyfluoroalkyl substances (PFAS) and are often released in large quantities directly to the environment as they are used to fight fires. AFFF composition is complex and contains many unknown PFAS in addition to ingredients such as hydrocarbons, solvents, and corrosion inhibitors. While biological effects of single PFAS have been studied, the effects of PFAS-containing mixtures, such as AFFF, are unknown. The effect of PFAS on microorganisms is also not well understood; nevertheless, we rely on microorganisms in locations containing elevated PFAS concentrations to perform certain functions, such as carbon cycling and co-contaminant degradation. This study focused on determining the functional consequences of AFFF and PFAS exposure in a microbial community in both the presence and the absence of a co-contaminant. AFFF, select PFAS, and a PFAS mixture were tested to determine the effect of AFFF on an anaerobic microbial community and the characteristics of the PFAS that drive toxicity in such mixtures. To study this, anaerobic digester communities were exposed to PFAS and a co-contaminant (2,4-dichlorophenol, DCP); methane production, as an indicator of toxicity and the community's ability to cycle carbon, and co-contaminant degradation were monitored. Results showed that PFAS and AFFF can alter the toxicity of DCP, inhibit DCP degradation, decrease the number of methanogens present, and change the microbial community structure. DCP was also able to decrease the toxicity of the PFAS perfluorooctane sulfonate (PFOS), possibly by changing the sorption of PFOS to the microorganisms present. Additionally, it was determined that while PFOS was responsible for AFFF toxicity, no single PFAS or simple PFAS mixture accurately accounted for the inhibition of DCP degradation caused by AFFF exposure.

Environmental significance

Aqueous film forming foams (AFFF) contain many perfluoroalkyl substances (PFAS) as well as other organic ingredients. The effects of AFFF and PFAS on microbial communities are not well documented and these communities are essential for processes including contaminant degradation and nutrient cycling. Moreover, many of the PFAS components in AFFF are difficult to quantify, thus, it is pertinent to determine the components that are primarily responsible for AFFF's biologic effects. Herein it was determined that long-chain PFAS and AFFF constrain methane production and perfluorooctane sulfonate was the driver of AFFF toxicity. Long-chain PFAS also inhibited co-contaminant degradation, though no single PFAS or simple PFAS mixture accurately described the inhibition of co-contaminant degradation caused by AFFF exposure.


The use of aqueous film forming foams (AFFF) containing per- and polyfluoroalkyl substances (PFAS) to suppress fires has resulted in wide-spread and persistent pollution globally. Significant contamination of surface water, groundwater and biota proximate to AFFF use has been reported.1–4 Areas contaminated with AFFF can contain up to mg L−1 levels of PFAS;1,3 the fluorinated chains of which are generally resistant to degradation and therefore persist in the environment.2,5–8 In addition to high levels of PFAS, AFFF also contains solvents, corrosion inhibitors, and hydrocarbon-based surfactants (e.g., 2) To complicate matters, AFFF formulations are proprietary and have varied over the years of their manufacture, with over 100 PFAS associated with AFFF use.9,10

The PFAS components of AFFF have caused environmental concern because of their effect on biota4,5,11–13 and persistence in the environment.2,5–8 Of particular concern, they have been found to alter the response of organisms to other contaminants present in the environment, so called co-contaminants. Many changes in co-contaminant toxicity in the presence of PFAS have been linked to PFAS-caused changes in cell membrane permeability. Indeed, PFAS, specifically carboxylates with 7–8 fluorinated carbons and sulfonates with 4–8 fluorinated carbons, were shown to increase cell membrane permeability, and as a result, the quorum sensing response in a pure bacterial culture.14 Others have shown that cellular uptake and the genotoxicity of cyclophosphamide increased in human lung cells exposed to 12 mg L−1 perfluorooctane sulfonate (PFOS).15 Similarly, the exposure of human liver cells to either 50 mg L−1 PFOS or 41 mg L−1 perfluorooctanoate (PFOA) increased the uptake and toxicity of pentachlorophenol.16 Although these effects were presumably a result of PFAS-induced cell permeability increases, the toxicity of atrazine and diuron decreased when algal cells were exposed to 10–40 mg L−1 PFOS,17 complicating a simple extrapolation of the risk associated with mixtures of PFAS and other contaminants.

AFFF and some of its fluorinated components have also been observed to alter microbial function and transport in ways that cannot easily be attributed to changes in cell permeability. Simulated AFFF formulations with concentrations of total PFAS of 110 mg L−1 were shown to reduce the degradation rate of trichloroethene and alter microbial community structure in microcosms enriched for Dehalococcoides mccartyi strains, organisms responsible for the dechlorination of chlorinated contaminants.18 Additionally, PFAS components of AFFF were found to decrease the rate of trichloroethene degradation, though dependent on PFAS structure and concentration, while the biodegradable components were found to simulate trichloroethene degradation.19 In other experiments, simulated AFFF did not change the rate or extent of aerobic toluene degradation in a pure culture of bacteria, but did increase biofilm formation and the expression of stress response factors.20 Finally, PFOA exposure at 100 μg L−1 increased the transport of a pure bacterial culture in simulated aquifer material.21 Although these studies indicate that AFFF can have an ecological impact, the chemical components responsible for these effects and reasons for this impact are not clear. In addition, the complex composition of AFFF makes it particularly difficult to understand which components of this mixture dominate from a risk perspective.

The research presented herein was based on our previous research on the partitioning of individual PFAS to microorganisms22 and the impact of individual PFAS on microbial membrane permeability and associated functions, such as quorum sensing.14 This research was expanded to examine the effects of AFFF on a broader range of microbial functions, both in the absence and presence of co-contaminants, and to ascertain why these effects occurred (physical-chemical or microbiological reasons). We also sought to discern the characteristics of the PFAS most likely to be of concern in AFFF formulations and whether these effects could be predicted from the behavior of the dominant PFAS in the AFFF formulation studied, offering a simplified approach to risk management and assessment.



Six different PFAS were studied: perfluorobutanoate (PFBA, CAS 375-22-4), PFOA (CAS 335-67-1), perfluorononanoate (PFNA, CAS 375-95-1), perfluorobutane sulfonate (PFBS, CAS 375-73-5), perfluorohexane sulfonate (PFHxS, CAS 355-46-4), and PFOS (CAS 1763-23-1). These PFAS were chosen to ascertain the role that functional group and fluorinated chain length have in biological effects of PFAS. Two functional groups, carboxylic acid and sulfonic acid were tested. Additionally, three fluorinated chain lengths of each functional group were tested. PFNA and PFOS have the same fluorinated chain length; therefore, these results could be compared to tease out effect of functional group. We also chose PFAS that are commonly found in the environment. Other than PFOS, all PFAS were purchased from Sigma Aldrich. PFOS was purchased from Santa Cruz Biotechnology. PFAS were dissolved in methanol and stored in the freezer at −20 °C. Light Water™ AFFF manufactured by 3 M (Lot 624 packed in April, 1990) was obtained from a local municipal fire station and stored in a sealed polypropylene tube. The AFFF was analyzed for perfluoroalkyl substances using liquid chromatography – mass spectrometry (LC-MS); triplicate AFFF aliquots were diluted in methanol and analyzed for PFAS as described below. PFOS and PFHxS were the dominant PFAS present in triplicate samples of the AFFF, at concentrations of 13 ± 3 and 2.3 ± 0.2 g L−1 AFFF, respectively. The co-contaminant 2,4-dichlorophenol (DCP) was purchased from Sigma Aldrich at 99% purity.

Microbial culture

For use in these experiments, a complex community containing organisms with both diverse and redundant metabolic functions was desired, in addition to a community with important engineering functions that might be susceptible to PFAS exposure. An anaerobic digester community was therefore used to assess effects of AFFF and PFAS on microbial function, as it has a complex structure, high biomass levels, was expected to have members with both highly redundant and diverse metabolic capabilities, and could be exposed PFAS via normal operation. It was also thought that the response of this community might also be relevant to subsurface environments that had been exposed to AFFF through fire training activities, as AFFF has an extremely high organic carbon content (>500[thin space (1/6-em)]000 mg L−1 as COD)2 and is likely to cause microbial use of any oxygen present and therefore create anaerobic conditions with high biomass levels during AFFF loading. Digester seed was obtained from the Empire Wastewater Treatment Plant (Farmington, MN) and was used to inoculate four 1.5 L reactors. Reactors had an equal solids and hydraulic retention time (SRT and HRT, respectively) of 18 days, were fed a 50/50 by volume mixture of thickened waste activated and primary sludge (obtained from Empire Wastewater Treatment Plant), and maintained at a temperature of 37 °C. These were also the conditions under which the digester was operated in the field. Total gas production and pH were monitored in reactors 1–2 times a week. Reactors were operated for over a year before digester material was removed for experiments. Gas production and pH over this period were relatively stable at 1.65 ± 0.51 L per day and 7.18 ± 0.15, respectively (Fig. SI.1 and SI.2).

Experimental design

Immediately after the digesters were fed and mixed, biosolids were collected for use in experiments, and therefore contained electron donor (sludge). Biosolids were diluted by ten with reduced anaerobic mineral media (RAMM) in an anaerobic glovebag (Coy). RAMM was prepared as described by Shelton and Tiedje23 then sterilized by boiling under nitrogen gas for over an hour prior to use. pH was adjusted to 7.0 using pH paper and hydrochloric acid or sodium hydroxide inside an anaerobic glovebag. pH did drift and vary, however, during experiments.
Co-contaminant toxicity. All experimental treatments were prepared and analyzed in triplicate. Three types of experiments were performed, all of which were amended with either PFAS only or PFAS plus DCP: (1) experiments in which PFAS were added to a final concentration of 50 mg L−1 to mimic the high concentrations observed in AFFF-impacted areas1,3, (2) experiments to study the PFAS responsible for AFFF toxicity, with neat AFFF added directly to serum bottles such that the targeted final concentrations of PFOS and PFHxS were 50 mg L−1 and 8.8 mg L−1, respectively, and subsequent PFAS-only controls containing those same PFOS and PFHxS concentrations, and (3) a final type of experiment in which PFOS, PFOA, or PFNA were amended at 5 mg L−1. DCP was chosen as a model co-contaminant because it was expected to partition into cell membranes where it has been shown to cause uncoupling,24,25 similar to the expected partitioning behavior of PFAS.26 Because the mechanism of toxicity is known for DCP, choosing this compound as the co-contaminant was more likely to clarify the effect of PFAS on the microorganisms present. DCP is also quite soluble (4500 mg L−1), making it available to the cultures. The concentration of DCP added (1 mg L−1) was chosen because in preliminary experiments this concentration was only slightly inhibitory to the microbial community, resulting in a <20% decrease in methane production when amended.

Diluted biosolids were divided into two portions: one portion was amended with 1 mg L−1 DCP via a methanol stock solution and the other received the equivalent amount of methanol only, both to a final methanol concentration of 18.5 mg L−1. The diluted biosolids, with or without DCP added, were then aliquoted into: (1) bottles already containing longer-chain PFAS; (2) empty bottles to which PFBS or PFBA were added immediately after biosolids addition; or (3) empty bottles used for no-PFAS controls.

PFNA, PFOS, PFOA, and PFHxS were added to empty 100 mL serum bottles via a methanol stock solution, after which the methanol was allowed to evaporate for at least 48 hours prior to the addition of diluted biosolids as described above. The four-carbon PFAS (PFBS and PFBA) were added directly to serum bottles already containing diluted biosolids with or without added DCP and were added as a methanol stock solution. In treatments amended with AFFF, neat AFFF was added directly to 100 mL serum bottles, such that the targeted final concentrations of PFOS and PFHxS were respectively 50 mg L−1 and 8.8 mg L−1. Negative controls for all treatments were prepared identically to the PFAS-amended treatments, and therefore either were amended with methanol that was then allowed to evaporate for 48 hours prior to addition of diluted biosolids or contained the equivalent concentration of methanol added to the biosolids that had been added with the PFBS or PFBA stock.

Time zero of the experiments was considered to be the time of diluted biosolids (either with or without added DCP) addition to the serum bottles. After filling, bottles were immediately capped with butyl rubber caps and aluminum crimps. All treatments were shaken at 150 rpm at 37 °C. Although unrealistic, this was done for experimental convenience and to maintain the conditions under which the microbial community was cultured. Methane production was used as an indicator of carbon cycling ability and therefore general toxicity and was measured twice daily for approximately 3 days. The pH and volatile solids of the diluted biosolids were measured at Day 0. In the treatments containing DCP, the chemical oxygen demand (COD) was measured on Days 0, 3, and 25 using HACH COD digestion vials following manufacturers standards (Digestion Method 8000, Hach Company, Loveland, CO).

DCP degradation. Experiments were set up as described above. Daily samples (2 mL) were removed from each bottle in the anaerobic glovebag for DCP analysis for the first five days. After five days, DCP was analyzed every 2–3 days in each bottle until the end of the experiment at Day 25, or until the DCP was no longer detected. Methane and biogas samples were measured immediately before samples for DCP analysis were taken. COD was monitored in the initial diluted biosolids and in PFAS/AFFF-amended treatments on Days 0, 3, and 25. Samples (1 mL) were taken from each bottle on Days 0 and 25, weighed, centrifuged for 5 min at 2500 rcf, and decanted. DNA was extracted from the biomass pellet using the FastDNA Extraction Kit (MP Biomedicals) according to the manufacturer's instructions. Once extracted, the DNA was stored at −20 °C until further analysis, as described below.
Sorption. The quantity of DCP and PFAS sorbed to diluted biosolids was also determined. The following PFAS were evaluated in quadruplicate: AFFF (targeted at a concentration such that PFOS and PFHxS concentrations were 50 mg L−1 and 8.8 mg L−1, respectively), PFOS (targeted concentration 50 mg L−1), PFHxS (targeted concentration 8.8 mg L−1), and PFOS+PFHxS (50 mg L−1 and 8.8 mg L−1, respectively). Biosolids (1 mL) diluted with 9 mL of RAMM made without resazurin or sodium sulfide were placed into glass centrifuge vials on the bench top (aerobic conditions). PFAS and 1 mg L−1 DCP were added to the vials. Vials were sealed with aluminum foil-covered screw caps and mixed using a Glas-Col rotator (18 rpm) at room temperature. After mixing for 24 hours, two 4 mL samples were taken. Of these two samples, one, for DCP analysis, was aliquoted into a 15 mL glass centrifuge tube; the other, for PFAS analysis, was aliquoted into a 15 mL polypropylene centrifuge tube. All samples were centrifuged at 2500 rcf for 15 min and the supernatants were sampled (1 mL), analyzed as described below, and then decanted. The remaining solids were then analyzed as described below. PFAS-free controls containing DCP, used to determine DCP partitioning in the absence of PFAS, were also prepared and analyzed identically to those described above with the exception that no PFAS were present. Likewise, DCP-free controls, to determine PFAS partitioning in the absence of DCP, were prepared and analyzed identically to those described above with the exception that no DCP was present. All vials contained equivalent amounts of methanol and water.

Analytical methods

Biogas and methane. Total biogas production was measured by water displacement in a 50 mL glass burette. Methane was measured in headspace samples (200 μL) via a locking gas-tight syringe on a gas chromatograph (GC) (Hewlett Packard 6890) coupled with a thermal conductivity detector. A 3 m × 3.2 mm × 2.1 mm 45/60 MOL SIEVE 13X Supelco column was used for methane separation. The oven temperature was constant (75 °C) with a carrier gas (Ultra Pure Grade nitrogen, >99.99%) flow rate of 19.3 mL min−1. The inlet was set at 150 °C with a total flow of 20 mL min−1 nitrogen gas; the detector was held at 210 °C with a reference flow of 45 mL min−1 of nitrogen gas.
DCP. DCP was extracted from both liquid and solid samples with hexane. Briefly, 2 mL hexane was added to a vial containing the liquid or solid sample and shaken for approximately 60 s. Vials were then settled for 20 min and approximately 1 mL of the hexane was drawn off and shaken with 0.5 g sodium sulfate. Dried hexane samples were then analyzed on a Hewlett Packard 6890 GC coupled with a micro electron capture detector and fit with a RESTEK Rtx-1701 column (30 m × 0.25 mm and 0.25 μm film thickness). The GC was operated with a carrier gas flow rate of 2 mL min−1, an oven temperature program of 105 °C for 10 min followed by a 5 °C min−1 ramp to 140 °C, a detector temperature of 350 °C, and an inlet temperature of 280 °C. DCP standards were made with sterilized anaerobic biosolids and were prepared identically to samples; therefore, the DCP extraction efficiency was accounted for in the standards. The DCP detection limit was 0.035 mg L−1.
PFAS. PFAS were analyzed as described previously.22 Briefly, liquid samples were diluted into optima grade methanol and an internal standard was added. PFAS were extracted from solids by adding 1 mL of optima grade methanol, shaking, and sonicating in a bath sonicator for 15 min. An aliquot (0.5 mL) of this methanol was then was added to an HPLC vial with internal standard. Samples were analyzed on a Hewlett Packard 1100 LC-MS, as described elsewhere.22 Methanol blanks were run between each sample and an analytical standard was run every 6 samples. Experimental and method blanks were run every 8 samples. The concentration of a given PFAS was calculated using the relative response factor method.
Volatile solids. To measure volatile solids, the mass difference between a sample after drying and after heating overnight in a 550 °C furnace was determined. A minimum of six replicates were used to determine the volatile solids concentration in a sample.

Microbial analysis

Illumina sequencing. To determine the effect of PFAS on bacterial communities, communities were analyzed by Illumina sequencing of the 16S rRNA gene. Amplification and sequencing was completed by the University of Minnesota Genomics Center (UMGC). The V4–V6 region of the 16S rRNA gene was amplified with primers F-GTGCCAGCMGCCGCGGTAA and R- GACRRCCATGCANCACCT and sequenced using Illumina MiSeq paired end sequencing (2 × 300). Methodological details are provided in the ESI. Whole community data and data from which the Archaea were excluded (i.e., Bacteria only) were both analyzed and suggested that the Archaea were not adequately represented with the Illumina primers selected. Beta diversity analysis was completed in R with the Phyloseq package. The beta diversity between the samples was calculated using Bray Curtis and plotted using principle coordinate analysis. The Kruskal–Wallis test was performed to determine if any OTUs present in the samples at a relative abundance above 0.5 were statistically different between the different treatments.
Quantitative polymerase chain reaction. Methanogen populations were quantified using qPCR and previously published methods.27–29 Methanogens were of particular interest in this research, because their function in the carbon cycle is critical and they may be more sensitive to other organisms to uncoupling from PFAS or DCP as a result of the low quantity of energy gained from their electron donor/acceptor couple. Details of the method are provided in the ESI. To determine whether differences in methanogen numbers between samples were significant, t-tests between all treatments were performed with P-values corrected via the Tukey's correction.

Data analysis

Toxicity. The rate of methane production was calculated for each bottle by linear regression of the data from the first three days of the experiment (Fig. SI.3). Rates were then averaged among the triplicate bottles in a given treatment. Outliers were discovered using Grubbs' correction with α = 0.05. Rates were compared statistically within experiments performed at the same time with the same digester biosolids via an ANOVA test with P-values adjusted using the Tukey correction. Absolute values of the rate of methane production between experiments were more variable, presumably because of changes to the digester community or the wastewater solids fed to the digester community over time, or the pH in each experiment. In all cases a P < 0.05 was considered significant. Pooling data from multiple experiments to accomplish a more global statistical comparison was not attempted.

Antagonistic interactions were defined as when the predicted combination effect (eqn (1), see below) was significantly less than the measured effect; a synergistic interaction was defined as when the predicted combination was significantly greater than the measured effect. When predicted and measured effects were not statistically different, the interaction was said to be additive. Predicted methane production in treatments containing both DCP and PFAS were calculated by subtracting the effect of the PFAS and DCP treatments from the corresponding experimental control, as shown below (eqn (1)).

Predicted rate = rate control − (rate control − rate PFAS) − (rate control − rate DCP treatment)(1)
where rate control = experimentally determined rate of methane production in control bottles, rate PFAS = experimentally determined rate of methane production in bottles amended with PFAS only, and rate DCP = experimentally determined rate of methane production in bottles amended with DCP only.

Values in Table 1 were obtained from the EPIWEB software30 unless specifically referenced. The fraction of molecules in the neutral form (α) were calculated according to Schwarzenbach et al.,31 assuming an ideal solution at equilibrium. All of the compounds in the table were either weak acids or were largely neutral at pH 7. The octanol–water partition coefficients were corrected by assuming that the neutral fraction is responsible for partitioning and hence the partitioning coefficients were multiplied by the percent neutral fraction.31

Table 1 Characteristics of co-contaminants and changes in their toxicity in the presence of PFAS
Log[thin space (1/6-em)]Kowa pKa Molecular weight Percent neutralb Adjusted log[thin space (1/6-em)]Kow Change in KD Observed change in toxicity
a Log[thin space (1/6-em)]Kow values were obtained using EPI WEB.39 Experimental values were taken when possible, with predicted values were used in absence of experimental values. b The percent of molecules that are neutral were calculated by assuming an experimental pH of 7. Eqn (1) was used to calculate (as described in the Methods above).40
2,4-Dichlorophenol (DCP) 3.06 7.85 (ref. 31) 163 87.62% 3.0 Decrease (this research) Decrease toxicity (this research)
2,4,6-Trichlorophenol 3.69 6.19 (ref. 31) 197.45 13.4% 2.82 NA Increase toxicity (algae35) dose dependent
Pentachlorophenol 5.12 4.83 (ref. 31) 266.34 0.67% 2.95 Increase (liver cells16 and algae17) Increase toxicity (liver cells16 and algae17)
Atrazine 2.61 NA (ref. 31) 215.68 100% 2.6129 Decrease (algae17) Decrease toxicity (algae17)
Diuron 2.68 NA (ref. 31) 233.1 100% 2.68 Decrease (algae17) Decrease toxicity (algae17)
Triclosan 4.76 7.9 (ref. 36) 289.54 75.97% 4.64 NA Decrease toxicity (algae35) increase toxicity (anaerobic bacteria37)
Bezafibrate 4.25 (calc.) 3.61 (ref. 36) 361.8 0.04% 0.86 NA Decrease toxicity (algae35)
Gemfibrozil 4.77 (calc.) 4.7 (ref. 38) 250.34 0.50% 2.47 Decrease toxicity (algae35)

Degradation. For a given treatment, DCP concentrations were averaged at each time point and concentrations were plotted against time. PFAS or AFFF additions did not appear to affect degradation rate, though they did delay the onset of DCP degradation. Hence, degradation was modeled by the Gompertz equation32,33 (eqn (2)) to compare the lag period prior to the onset of DCP degradation (k) in the different treatments.
f(t) = ae−ec(tk)(2)
where a = initial DCP concentration, k = time to inflection point, termed lag, c = fitting parameter associated with degradation rate.

The initial DCP concentration (a), was held constant at 1 mg L−1. The decay rate (c) was held constant for experiments performed at the same time with the same digester biosolids and was obtained using the Solver function in Microsoft Excel to fit the data of the control treatments (minimization of the sum of the square of the differences). The Solver function in Microsoft Excel was also used to solve for lag (k) in all treatments. The lag in PFAS-amended treatments was compared statistically to the appropriate experimental control using multiple t-tests followed by the Dunnet correction. A spearman correlation test was performed to access relationship between the number of fluorinated carbons and the effect of PFAS on lag (k) in DCP degradation.

Sorption. Partition coefficients (Koc) were defined as the ratio of a compound's (DCP or PFAS) concentration in the solid phase (mg kg−1) to the compound's (DCP or PFAS) concentration in the liquid phase (mg L−1) (eqn (3)).
image file: c9em00241c-t1.tif(3)
where Cs = volatile solids concentration, Cw = aqueous concentration.

Both solid and liquid concentrations of DCP and PFAS were measured directly. The quantity of solids per sample was calculated using the measured volatile solids. Outliers were determined using Dixon and Grubb's tests, with one outlier identified for DCP sorption in the PFOS treatment, which was subsequently discarded.

Results & discussion

AFFF and PFAS toxicity

With the exception of one replicate experiment, PFOS-containing treatments, including those amended with PFOS, PFOS+PFHxS, or AFFF, were toxic/inhibitory to mixed anaerobic cultures (Fig. 1 and 2, Table SI.1). In all but one case (Table SI.1) PFOS-, PFOS+PFHxS-, and AFFF-amended treatments all had a significantly lower methane production rate than the PFAS-free control treatment (P < 0.01 for all) or the PFHxS-amended treatment (P < 0.01 for all of the PFOS-, PFOS+PFHxS-, and AFFF-amended treatments). Indeed, the effect that AFFF had on methane production could be estimated by only that of the dominant PFAS in the AFFF formula: PFOS (Fig. 1). In addition to PFOS, AFFF contains PFHxS, a large number of unidentified PFAS,34 solvents, corrosion inhibitors, and hydrocarbon-based surfactants (e.g., 2). Nevertheless, our results showed that PFOS alone appeared to be the driver of AFFF toxicity to the microorganisms in this anaerobic community, rather than those other known and unknown ingredients, with the methane production rate in the PFOS-amended treatments equivalent to that in the AFFF-amended treatments (P = 0.91). The presence of PFHxS did not appear to play a role in the toxicity of AFFF to this mixed anaerobic culture, with the PFOS+PFHxS-amended treatments affecting methane production similarly to the treatments amended with only PFOS (P = 0.88).
image file: c9em00241c-f1.tif
Fig. 1 Rate of methane production within the first three days of exposure to AFFF or its primary PFAS constituents. Error bars represent the standard deviation. Letters show treatments with rates of methane production that are statistically similar. Note: the control treatments labeled as “PFAS+DCP” contained DCP only, as all control treatments lacked PFAS.

image file: c9em00241c-f2.tif
Fig. 2 Effect of PFAS on methane production as compared to appropriate controls.

PFAS with fewer than eight fully fluorinated carbons did not alter the methane production rate in the anaerobic community, with PFBS-, PFHxS-, PFBA-, and PFOA-amended cultures producing methane at a similar rate as the control (Fig. 2, Table SI.1). In addition, although the longer-chain PFNA did appear to be toxic and decrease the methane production rate in the digester culture, the decrease was not consistently statistically significant (P = 0.30, P < 0.01, Fig. 2). As mentioned above, in replicate experiments, PFOS toxicity to the community also varied, and in one case did not appear to be toxic to the digester culture (Fig. 2). Others have also observed variable effects of PFOS on microorganisms and microbial communities, the reasons for which are not entirely known.17–20,35

Effects of AFFF and PFAS on co-contaminant toxicity

Because PFAS, and in particular PFOS, have been shown to partition to lipid bilayers and change microbial cell permeability,14,22 a co-contaminant was added to AFFF- and PFAS-amended treatments to determine whether co-contaminant toxicity was altered in the presence of PFAS. DCP was chosen as a model co-contaminant to test how the addition of two chemicals that accumulate in cell membranes behave. Of interest was whether the combination of PFAS and DCP might enhance uncoupling, and therefore toxicity, as cited by others,16,17,25 or whether the two chemicals might compete for available sites near or within the membrane and moderate observed toxicity.

Changes in toxicity were observed in the presence of PFOS (Fig. 1), with the amendment of PFAS+DCP (as PFOS only, PFOS+PFHxS, or AFFF) less toxic to the culture than the amendment of PFAS alone (P < 0.01 for each PFOS, PFOS+PFHxS, and AFFF). DCP alone had a slight, but significant (P = 0.05) effect on methane production in PFAS-free controls (Fig. 1), suggesting that when amended by itself, DCP may have partitioned to the microorganisms present. Additionally, changes in toxicity when DCP was added as a co-contaminant were only observed for PFOS-containing treatments. The results therefore also suggest that DCP perhaps displaced PFOS from the cells, moderating PFOS toxicity. Interestingly, the effect of AFFF+DCP exposure could not be approximated by PFOS+DCP exposure (P < 0.01) but could be approximated by PFOS+PFHxS+DCP exposure (P = 0.99). This indicates that as the system became more complex with the presence of both a co-contaminant (DCP) and PFAS, the behavior of the AFFF-amended treatment began to deviate from the behavior of the treatment amended with the dominant PFAS in AFFF (PFOS) (Fig. 1).

Similar results have been observed by others. Indeed, researchers have found that the toxicity of mixtures of PFAS and chlorophenols deviate from a purely additive model and they also alter the way that these chemicals are taken up by cells. For instance, the combination of 2,4,6-trichlorophenol and PFOS was antagonistic or synergistic towards green algae dependent on the dose.35 In contrast, however, both PFOS and PFOA increased cellular uptake of pentachlorophenol and mixtures behaved synergistically towards liver cells.16 A synergistic model of toxicity was also observed, coupled with increased cellular uptake of PFOS and pentachlorophenol, in green algae exposed to PFAS-containing chemical mixtures.17 The effect of PFAS on co-contaminant toxicity and co-contaminants on PFAS toxicity therefore appears to be complex, with the hydrophobicity, ionic state, and individual toxicity of the compounds likely playing a role, along with the identity of the organism(s) involved (e.g., algae vs. bacteria vs. human cells). It is impossible to point to a single parameter (i.e., hydrophobicity, charge, molecular weight), however, that allows prediction of which co-contaminants are likely to react synergistically or antagonistically with PFAS (Table 1), making the specific effect of co-contaminants on PFAS toxicity a challenge to predict.

AFFF and PFAS effects on secondary metabolism

PFAS also had an impact on the anaerobic community's ability to degrade the co-contaminant DCP, increasing the lag period prior to the onset of DCP degradation (Fig. 3 and 4, SI.4 and SI.5). Treatments containing 50 mg L−1 PFOS as individual compounds and as a mixture increased the lag period prior to DCP degradation (for PFOS P < 0.01, PFOS+PFHxS P < 0.01, AFFF P < 0.01). Indeed, lag, normalized to the control, was positively correlated to fluorinated chain length (P = 0.02, Fig. 4), suggesting that longer chain-length PFAS were more likely to influence co-contaminant degradation. The perfluorinated carboxylates also appeared to have a larger effect than the sulfonates. For example, PFNA, another eight fluorinated-chain PFAS, increased the lag period prior to DCP degradation to a greater extent than PFOS and also increased the lag period when amended at concentrations of both 50 mg L−1 and 5 mg L−1 (P < 0.01, P < 0.01, respectively); PFOS did not delay degradation at a concentration of 5 mg L−1 (P = 0.85, Fig. SI.5). PFHxS (amended at 50 mg L−1) also significantly delayed the onset of DCP degradation (P < 0.01), but only by approximately 15 hours (Fig. SI.4). The other PFAS tested (PFBS, PFBA, PFOA) did not significantly delay DCP degradation at concentrations of 50 mg L−1 (PFBS P = 0.89, PFBA P = 0.34, PFOA P = 0.31, Fig. SI.4).
image file: c9em00241c-f3.tif
Fig. 3 Degradation of DCP in the presence of AFFF and its major PFAS constituents. Error bars show the standard deviation.

image file: c9em00241c-f4.tif
Fig. 4 The lag period before DCP degradation in an anaerobic community exposed to sulfonate and carboxylate PFAS of varying fluorinated chain length. Error bars show the standard deviation of triplicate treatments.

In contrast to the toxicity results, AFFF caused a much greater lag prior to DCP degradation than that caused by PFOS alone or PFOS+PFHxS. A component of AFFF other than PFOS or PFHxS therefore had an effect on the lag prior to DCP degradation; it is possible that the high COD content of AFFF could result in diauxy, with the organisms degrading the other organic compounds present in AFFF prior to degrading DCP (Fig. SI.7, Table SI.2). It is also possible that the high COD content caused a shift in the microbial community, enriching for organisms that were not capable of DCP degradation.

The results from this work support the hypothesis that PFOS drives toxicity with respect to methane production, while AFFF, likely as a result of the high COD content of the AFFF, has a unique impact on other aspects of the community's function, such as DCP degradation ability. This is of great importance when thinking about bioremediating a contaminated site that contains both AFFF and other contaminants (i.e., fuel or solvents) and may result in much longer remediation times or altered approaches.

AFFF and PFAS effects on co-contaminant behavior and microbial community

To understand whether the functional responses observed upon AFFF, PFAS, and DCP exposure were a result of a physical–chemical phenomenon (e.g., chemical displacement near the cell surface), a microbial phenomenon (e.g., diauxy or a community shift), or both, additional sorption (Fig. 5) and microbial community data (Table 2, Fig. SI.6) were obtained.
image file: c9em00241c-f5.tif
Fig. 5 (A) Effect of PFAS addition on DCP partitioning. (B) Effect of DCP addition on PFOS partitioning. Error bars represent standard deviation.
Table 2 Number of methanogens in DCP-degradation experiment on Day 25
Treatment Average ± standard deviation (log copies/ml sample)
Initial (Day 0) 7.31 ± 0.12
Control (Day 25) 7.98 ± 0.37
PFHxS (Day 25) 8.09 ± 0.66
PFOS (Day 25) 7.51 ± 0.04
PFOS+PFHxS (Day 25) 7.28 ± 0.07
AFFF (Day 25) 7.33 ± 0.19

The presence of PFOS or PFOS+PFHxS appeared to decrease the sorption of DCP to solids when compared to DCP only, but the difference was not statistically significant (Fig. 5). Mass balances for DCP across all treatments were 88 ± 16% (n = 18), suggesting that DCP sorption to the glass or foil top was not significant in these experiments. PFOS has been shown to affect sorption of co-contaminants in the literature. In studies completed with green algae, the presence of PFOS increased the sorption of pentachlorophenol, but decreased the sorption of atrazine and diuron.17 Increased uptake of pentachlorophenol was also observed in liver cells.16 Researchers proposed that differences in hydrophobicity and contaminant structure could be responsible for differences observed in uptake, though based on Table 1, this does not appear to be entirely straight-forward. These studies were completed with eukaryotes, whereas this study was completed with a mixed community of prokaryotes. In mixtures of AFFF and DCP, no differences in DCP sorption to the solids were observed (Fig. 5). Because of the complex chemical nature of AFFF, components of AFFF could increase solubility of DCP, change interactions between DCP and PFOS outside of the membrane, or alter the interaction of PFOS with the microbial culture in an unpredicted way. Though the data are limited, it appears that sorption mirrors toxicity (Table 1). Greater partitioning to a microbial community increases toxicity, and conversely, less partitioning decreases toxicity.

Similar sorption trends were observed with PFOS sorption in the presence of DCP. In this case, DCP appeared to decrease the sorption of PFOS. Sorption was only significantly different from the control (PFOS only without DCP) in the PFOS+DCP and PFOS+PFHxS+DCP treatments. PFOS sorption in the presence of AFFF+DCP was not significantly different (Fig. 5). This is similar to what was observed with DCP partitioning and is likely a result of the increased complexity of the mixture. It is possible that PFOS and PFHxS compete with, and potentially displace, DCP in the cell membrane and DCP in turn displaces PFOS; more work, however, is needed to determine whether this is statistically supported. Other researchers have hypothesized that organic carbon competes for binding sites with PFAS on activated carbon.42–44 Additionally, PFAS can form hemi-micelles at 0.1 to 1% of the critical micelle concentration (on the order of gram per liter concentrations).40,43,45 It is possible therefore, that reductions in DCP binding could be a result of either binding site competition or DCP dissolving into PFOS-derived micelles. It should be noted that the organic carbon–water partitioning coefficients reported in this manuscript are lower, though still in the same order of magnitude, as what has been previously reported in the literature. For example, the minimum organic water partition coefficient reported in the review by Zareitalabad et al. is 251 L kg−1.41 Regardless, the effect is not dramatic and is not likely to be the only reason for the moderation of PFOS toxicity in the presence of DCP or the long lag period observed prior to DCP degradation in the presence of PFOS or PFNA.

Changes in the microbial community after exposure to PFAS, DCP, and AFFF were also observed (Table 2, Fig. SI.5). Table 2 shows the number of methanogens present, as measured by qPCR, in initial microcosm samples and in samples taken on Day 25 from the DCP-degradation experiments. In the presence of DCP, methanogens were at a similar abundance at the end of the 25 day experiment in both the control and PFHxS-amended treatments (P = 0.98) (Table 2). There were fewer methanogens present in the treatments amended with PFOS+DCP, PFOS+PFHxS+DCP, or AFFF+DCP, though the reduction was not statistically significant for the PFOS+DCP treatment (P = 0.17, P = 0.01, P = 0.03, for PFOS+DCP, PFOS+PFHxS+DCP, or AFFF+DCP, respectively). Additionally, methanogen numbers were statistically equivalent in PFOS+DCP-, PFOS+PFHxS+DCP-, and AFFF+DCP-amended treatments (P = 0.89, P = 0.76, P > 0.99 respectively, for comparisons between AFFF+DCP and PFOS+DCP, PFOS + PFHxS+DCP and PFOS+DCP, and AFFF+DCP to PFOS+PFHxS+DCP). This supports the toxicity/inhibition results and shows that even with the moderating effect of DCP (Fig. 1), the methanogen numbers are negatively affected by PFOS amendment. Although the structures of the Bacteria communities were initially similar, the structures of each treatment changed over time and with PFAS exposure (Fig. SI.6). The addition of PFOS, again as PFOS alone, PFOS+PFHxS, or AFFF, altered the community structure away from that of the PFHxS+DCP- or DCP-amended treatments (Fig. SI.6); nevertheless, these changes could not be statistically verified and the control community in one of the replicate treatments was very different from the other two treatments, containing a higher abundance of Anaerolinea and a lower abundance of Clostridia (Fig. SI.6), making it impossible to say anything definitive regarding the effect of PFAS on the Bacteria community structure.

Finally, with respect to the unique impact of AFFF on co-contaminant degradation and community structure, there was evidence of possible substrate competition in the system (Fig. SI.7). These observations are likely a result of the complex chemical make-up of AFFF, including the high levels of organic material in the AFFF (COD = 670[thin space (1/6-em)]000 ± 34[thin space (1/6-em)]000 mg L−1, SI, Table SI.2). It is possible that the high COD content of the AFFF led to at least some of the changes observed in the microbial community and the subsequent inhibition of DCP degradation.

These results suggest that when evaluating the risk of AFFF to microbial function, the effect of the long-chain PFAS present (e.g., PFOS) need to be considered from a basic toxicity perspective (Fig. 1), as do the presence of co-contaminants (e.g., DCP) and their moderating or exacerbating effects on toxicity16,17,35,37 (Fig. 1, Table 1). When placing this work in the context of prior research (Table 1), however, it is difficult to identify specific characteristics, in terms of molecular size, charge, or hydrophobicity, that indicate when a co-contaminant is likely to be more or less toxic in the presence of PFAS or moderate the toxicity of PFAS themselves. In addition, the presence and effect of compounds that might not initially be considered problematic, such as hydrocarbons, must also be considered, as these could change the solubility and partitioning of PFAS and co-contaminants (Fig. 5) and may also cause community shifts or substrate competition (Fig. SI.6), and as a result, delay remediation of the site (e.g., Fig. 3). Some disruptions in ecological function caused by AFFF contamination can be predicted to some degree by the dominant long-chain PFAS present in the AFFF formulation; nevertheless, prediction of the complete effect that AFFF might have on multiple microbial functions of interest does not appear to be predictable.

It is also important to note that both in the research presented herein and in much of the work presented in the literature,18–20 effects on microbial function are often investigated at PFAS concentrations greater than or equal to 50 mg L−1, indicating that only microorganisms in areas directly affected by fire-training activities are the ones that are likely to be impacted. In addition, the work presented herein relied upon toxicity assays performed with anaerobic digester cultures at 37 °C. Therefore, straight-forward extrapolation to environmental situations should be viewed with caution. Some microbial communities are likely to be more sensitive than others and the effect of PFAS will need to be assessed in specific situations if there is a reason for concern. Finally, the impact of AFFF and PFAS on co-contaminant degradation is likely dependent on the identity of the co-contaminant with few obvious trends emerging from examination of the chemical properties of co-contaminants. What is becoming clear, however, is the possible impact that PFAS, particularly in complex mixtures of AFFF, could have on microbial function.


When evaluating the risk of AFFF to microbial function, the effect of both the long-chain PFAS present (e.g., PFOS) and the presence of co-contaminants need to be considered with respect to their inhibitory effect on function, and in the case of co-contaminants, their moderating or exacerbating effects on inhibition. Co-contaminants and other non-PFAS compounds in AFFF formulations can also alter the solubility and partitioning of PFAS, the degradation of co-contaminants, cause changes in certain populations (e.g., methanogens), making it difficult to accurately predict the effects of AFFF exposure on microbial function based only on the anticipated impacts of the dominant PFAS present in the AFFF formulation.

Conflicts of interest

There are no conflicts of interest to declare.


We would like to thank Amy Prok for initially establishing the anaerobic digester cultures. Special thanks to Michael McCarty for some of the PFAS analysis. The work was supported by the Environment and Natural Resources Trust Fund as recommended by the Legislative Citizen Commission on Minnesota Resources.


  1. C. A. Moody and J. A. Field, Determination of Perfluorocarboxylates in Groundwater Impacted by Fire-Fighting Activity, Environ. Sci. Technol., 1999 Aug, 33(16), 2800–2806,  DOI:10.1021/es9813552B.
  2. C. A. Moody and J. A. Field, Perfluorinated Surfactants and the Environmental Implications of Their Use in Fire-Fighting Foams, Environ. Sci. Technol., 2000, 34(18), 3864–3870 CrossRef CAS.
  3. E. V. a. Arias, M. Mallavarapu and R. Naidu, Identification of the source of PFOS and PFOA contamination at a military air base site, Environ. Monit. Assess., 2015, 187(1), 4111 CrossRef ,
  4. S. B. Gewurtz, S. P. Bhavsar, S. Petro, C. G. Mahon, X. Zhao and D. Morse, et al., High levels of perfluoroalkyl acids in sport fish species downstream of a firefighting training facility at Hamilton International Airport, Ontario, Canada, Environ. Int., 2014, 67, 1–11,  DOI:10.1016/j.envint.2014.02.005.
  5. G. Ding and W. J. G. M. Peijnenburg, Physicochemical Properties and Aquatic Toxicity of Poly- and Perfluorinated Compounds, Crit. Rev. Environ. Sci. Technol., 2013, 43(6), 598–678,  DOI:10.1080/10643389.2011.627016.
  6. J. S.-C. Liou, B. Szostek, C. M. DeRito and E. L. Madsen, Investigating the biodegradability of perfluorooctanoic acid, Chemosphere, 2010, 80(2), 176–183 CrossRef CAS PubMed ,
  7. D. O'Hagan, Understanding organofluorine chemistry. An introduction to the C–F bond, Chem. Soc. Rev., 2008, 37(2), 308–319 RSC ,
  8. S. Vaalgamaa, A. V. Vähätalo, N. Perkola and S. Huhtala, Photochemical reactivity of perfluorooctanoic acid (PFOA) in conditions representing surface water, Sci. Total Environ., 2011, 409(16), 3043–3048 CrossRef CAS PubMed ,
  9. L. A. D. Agostino and S. A. Mabury, Identification of Novel Fluorinated Surfactants in Aqueous Film Forming Foams and Commercial Surfactant Concentrates, Environ. Sci. Technol., 2014, 48, 121–129 CrossRef PubMed.
  10. W. J. Backe, T. C. Day and J. A. Field, Zwitterionic, cationic, and anionic fluorinated chemicals in aqueous film forming foam formulations and groundwater from U.S. military bases by nonaqueous large-volume injection HPLC-MS/MS, Environ. Sci. Technol., 2013, 47(10), 5226–5234 CrossRef CAS.
  11. M. Fernández-Sanjuan, M. Faria, S. Lacorte and C. Barata, Bioaccumulation and effects of perfluorinated compounds (PFCs) in zebra mussels (Dreissena polymorpha), Environ. Sci. Pollut. Res. Int., 2013, 20(4), 2661–2669 CrossRef ,
  12. S. R. de Solla, a O. De Silva and R. J. Letcher, Highly elevated levels of perfluorooctane sulfonate and other perfluorinated acids found in biota and surface water downstream of an international airport, Hamilton, Ontario, Canada, Environ. Int., 2012, 39(1), 19–26 CrossRef CAS ,
  13. M. Houde, J. W. Martin, R. J. Letcher, K. R. Solomon and D. C. G. Muir, Biological Monitoring of Polyfluoroalkyl Substances: A Review, Environ. Sci. Technol., 2006, 40(11), 3463–3473 CrossRef CAS.
  14. N. J. M. Fitzgerald, M. F. Simcik and P. J. Novak, Perfluoroalkyl Substances Increase the Membrane Permeability and Quorum Sensing Response in Aliivibrio fischeri, Environ. Sci. Technol. Lett., 2018, 5, 26–31,  DOI:10.1021/acs.estlett.7b00518.
  15. S. Jernbro, P. S. Rocha, S. Keiter, D. Skutlarek, H. Färber and P. D. Jones, et al., Perfluorooctane Sulfonate Increases the Genotoxicity of Cyclophosphamide in the Micronucleus Assay with V79 Cells, Environ. Sci. Pollut. Res. Int., 2007, 14(2), 85–87 CrossRef CAS.
  16. G. Shan, M. Ye, B. Zhu and L. Zhu, Enhanced cytotoxicity of pentachlorophenol by perfluorooctane sulfonate or perfluorooctanoic acid in HepG2 cells, Chemosphere, 2013, 93(9), 2101–2107 CrossRef CAS ,
  17. W. Liu, Y.-B. Zhang, X. Quan, Y.-H. Jin and S. Chen, Effect of perfluorooctane sulfonate on toxicity and cell uptake of other compounds with different hydrophobicity in green alga, Chemosphere, 2009, 75(3), 405–409 CrossRef CAS ,
  18. T. S. Weathers, K. Harding-Marjanovic, C. P. Higgins, L. Alvarez-Cohen and J. O. Sharp, Perfluoroalkyl Acids Inhibit Reductive Dechlorination of Trichloroethene by Repressing Dehalococcoides, Environ. Sci. Technol., 2016, 50, 240–248 CrossRef CAS PubMed.
  19. K. C. Harding-Marjanovic, S. Yi, T. S. Weathers, J. O. Sharp, D. L. Sedlak and L. Alvarez-Cohen, Effects of Aqueous Film-Forming Foams (AFFFs) on Trichloroethene (TCE) Dechlorination by a Dehalococcoides mccartyi-Containing Microbial Community, Environ. Sci. Technol., 2016, 50, 3352–3361 CrossRef CAS PubMed.
  20. T. S. Weathers, C. P. Higgins and J. O. Sharp, Enhanced Biofilm Production by a Toluene-Degrading Rhodococcus Observed after Exposure to Perfluoroalkyl Acids, Environ. Sci. Technol., 2015, 49, 5458–5466 CrossRef CAS PubMed.
  21. D. Wu, M. Tong and H. Kim, Influence of Perfluorooctanoic Acid on the Transport and Deposition Behaviors of Bacteria in Quartz Sand, Environ. Sci. Technol., 2016, 50(5), 2381–2388 CrossRef CAS.
  22. N. J. M. Fitzgerald, A. Wargenau, C. Sorenson, J. Pedersen, N. Tufenkji and P. J. Novak, et al., Partitioning and Accumulation of Perfluoroalkyl Substances in Model Lipid Bilayers and Bacteria, Environ. Sci. Technol., 2018, 52(18), 10433–10440 CrossRef CAS.
  23. D. R. Shelton and J. M. Tiedje, General Method for Determining Anaerobic Biodegradation Potential, Appl. Environ. Microbiol., 1984, 47(4), 850–857 CAS.
  24. G. Chen, H. Yu, H. Liu and D. Xu, Response of activated sludge to the presence of 2,4-dichlorophenol in a batch culture system, Process Biochem., 2006, 41, 1758–1763 CrossRef CAS.
  25. L. Song, J. Wenju, T. Qiong and L. Yaozhong, Impact of a metabolic uncoupler, 2,4-dichlorophenol on minimization of activated sludge production in membrane bioreactor, Water Sci. Technol., 2010, 62(6), 1379–1385 CrossRef PubMed.
  26. a. a. Starkov and K. B. Wallace, Structural determinants of fluorochemical-induced mitochondrial dysfunction, Toxicol. Sci., 2002, 66(2), 244–252 CrossRef CAS PubMed ,
  27. D. K. Nelson, T. M. Lapara and P. J. Novak, Effects of Ethanol-based Fuel Contamination: Microbial Community Changes, Production of Regulated Compounds, and Methane Generation, Environ. Sci. Technol., 2010, 44(12), 4525–4530 CrossRef CAS PubMed.
  28. D. K. Nelson, T. M. Lapara and P. J. Novak, Structure and Function of Assemblages of Bacteria and Archaea in Model Anaerobic Aquifer Columns: Can Functional Instability Be Practically Beneficial?, Environ. Sci. Technol., 2012, 46, 10137–10144 CAS.
  29. S. E. Hook, K. S. Northwood, A. G. Wright and B. W. Mcbride, Long-Term Monensin Supplementation Does Not Significantly Affect the Quantity or Diversity of Methanogens in the Rumen of the Lactating Dairy Cow, Appl. Environ. Microbiol., 2009, 75(2), 374–380 CrossRef CAS.
  30. US EPA, Estimation Programs Interface Suite for Microsoft Windows, v 4.11, 2014 Search PubMed.
  31. R. P. Schwarzenbach, P. M. Gschwend and D. M. Imboden, Environmental Organic Chemistry, John Wiley & Sons, Inc., Hoboken, New Jersey, 2nd edn, 2003 Search PubMed.
  32. G. C. Okpokwasili and C. O. Nweke, Microbial growth and substrate utilization kinetics, Afr. J. Biotechnol., 2005, 5(4), 305–317 Search PubMed.
  33. G. Hallett LeFevre, Fate and Degradation of Petroleum Hydrocarbons in Stormwater Bioretention Cells, University of Minnesota, 2012 Search PubMed.
  34. E. F. Houtz, C. P. Higgins, J. A. Field and D. L. Sedlak, Persistence of perfluoroalkyl acid precursors in AFFF-impacted groundwater and soil, Environ. Sci. Technol., 2013, 47(15), 8187–8195 CrossRef CAS PubMed.
  35. K. Boltes, R. Rosal and E. García-Calvo, Toxicity of mixtures of perfluorooctane sulphonic acid with chlorinated chemicals and lipid regulators, Chemosphere, 2012, 86(1), 24–29 CrossRef CAS ,
  36. D. M. Revitt, T. Balogh and H. Jones, Sorption behaviours and transport potentials for selected pharmaceuticals and triclosan in two sterilised soils, J. Soils Sediments, 2015, 15, 594–606 CrossRef CAS.
  37. P. J. Mcnamara, T. M. Lapara and P. J. Novak, The Effect of Perfluorooctane Sulfonate , Exposure Time , and Chemical Mixtures on Methanogenic Community Structure and Function, Microbiol. Insights, 2015, 8(S2), 1–7 Search PubMed.
  38. L. Fernandezndez, R. Machın, A. Zornoza, I. Velaz, C. Martin and M. C. Martinez-Oharriz, Mechanism of sorption and release of a weak acid from b-cyclodextrin polymers, J. Inclusion Phenom. Macrocyclic Chem., 2011, 69, 411–415 CrossRef.
  39. US EPA, Estimation Programs Interface Suite™ for Microsoft®Windows, United States Protection Agency, Washington, D.C., 2012 Search PubMed.
  40. R. Schwarzenbach, P. Gschwend and D. Imboden, Environmental Organic Chemistry, John Wiley & Sons, Inc., Hoboken, New Jersey, 2nd edn, 2003 Search PubMed.
  41. P. Zareitalabad, J. Siemens, M. Hamer and W. Amelung, Perfluorooctanoic acid (PFOA) and perfluorooctanesulfonic acid (PFOS) in surface waters, sediments, soils and wastewater – A review on concentrations and distribution coefficients, Chemosphere, 2013, 91(6), 725–732 CrossRef CAS PubMed ,
  42. C. Eschauzier, E. Beerendonk, P. Scholte-Veenendaal and P. D. Voogt, Impact of Treatment Processes on the Removal of Perfluoroalkyl Acids from the Drinking Water Production Chain, Environ. Sci. Technol., 2012, 46, 1708–1715 CrossRef CAS PubMed.
  43. P. Mccleaf, S. Englund, A. Ostlund, K. Lindegren, K. Wiberg and L. Ahrens, Removal efficiency of multiple poly- and perfluoroalkyl substances (PFASs) in drinking water using granular activated carbon (GAC) and anion exchange (AE) column tests, Water Res., 2017, 120, 77–87 CrossRef CAS PubMed.
  44. D. Zhao, J. Cheng, C. D. Vecitis and M. R. Hoffman, Sorption of Perfluorochemicals to Granular Activated Carbon in the Presence of Ultrasound, J. Phys. Chem., 2011, 115, 2250–2257 CrossRef CAS.
  45. R. L. Johnson, A. J. Anschutz, J. M. Smolen, M. F. Simcik and R. L. Penn, The Adsorption of Perfluorooctane Sulfonate onto Sand, Clay, and Iron Oxide Surface, J. Chem. Eng. Data, 2007, 52, 1165–1170 CrossRef CAS.


Electronic supplementary information (ESI) available. See DOI: 10.1039/c9em00241c
Current Address: Department of Civil and Environmental Engineering, Colorado School of Mines 1500 Illinois Street, Golden, CO 80401, USA.

This journal is © The Royal Society of Chemistry 2019