Article pubs.acs.org/est
Generation of Perfluoroalkyl Acids from Aerobic Biotransformation of Quaternary Ammonium Polyfluoroalkyl Surfactants Sandra Mejia-Avendaño,† Sung Vo Duy,‡ Sébastien Sauvé,‡ and Jinxia Liu*,† †
Department of Civil Engineering, McGill University, Montreal, Quebec H3A 0C3, Canada Department of Chemistry, Université de Montréal, Montreal H3C 3J7, Canada
‡
Environ. Sci. Technol. 2016.50:9923-9932. Downloaded from pubs.acs.org by UNIV OF WINNIPEG on 10/29/18. For personal use only.
S Supporting Information *
ABSTRACT: The aerobic biotransformation over 180 days of two cationic quaternary ammonium compounds (QACs) with perfluoroalkyl chains was determined in soil microcosms, and biotransformation pathways were proposed. This is the first time that polyfluoroalkyl cationic surfactants used in aqueous film-forming foam (AFFF) formulations were studied for their environmental fate. The biotransformation of perfluorooctaneamido quaternary ammonium salt (PFOAAmS) was characterized by a DT50 value (time necessary to consume half of the initial mass) of 142 days and significant generation of perfluoroalkyl carboxylic acid (PFOA) at a yield of 30 mol % by day 180. The biotransformation of perfluorooctane sulfonamide quaternary ammonium salt (PFOSAmS) was very slow with unobservable change of the spiked mass; yet the generation of perfluorooctanesulfonate (PFOS) at a yield of 0.3 mol % confirmed the biotransformation of PFOSAmS. Three novel biotransformation intermediates were identified for PFOAAmS and three products including perfluorooctane sulfonamide (FOSA) for PFOSAmS through high-resolution mass spectrometry (MS) analysis and t-MS2 fragmentation. The significantly slower PFOSAmS biotransformation is hypothesized to be due to its stronger sorption to soil owing to a longer perfluoroalkyl chain and a bulkier sulfonyl group, when compared to PFOAAmS. This study has demonstrated that despite overall high stability of QACs and their biocide nature, the ones with perfluoroalkyl chains can be substantially biotransformed into perfluoroalkyl acids in aerobic soil.
1. INTRODUCTION Perfluoroalkyl and polyfluoroalkyl substances (PFAS) represent a major class of contaminants of emerging concern (CECs) found in groundwater, surface water, soil, sediment, and biota globally.1−5 Being designed to be chemically stable, PFAS species are highly persistent in the environment and some species are considered potential threats to human and environmental health.6,7 In recent years, it has been recognized that practices of uncontrolled releases of PFAS-based aqueous film-forming foams (AFFFs) used for firefighting activities have resulted in many AFFF-impacted sites with elevated levels of PFASs.8−10 Although there is extensive research on PFASs in the global scientific community, the research on AFFF contamination is still in its infancy. Consequently, regulatory frameworks for managing pollution associated with AFFFs mostly do not exist. Whereas the most prominent PFAS contaminant, perfluorooctanesulfonate (PFOS), is listed under the Stockholm Convention on Persistent Organic Pollutants for restricted production and use, numerous polyfluoroalkyl surfactants with various functionalities could be present in AFFF-impacted sites. Environmental fate, behaviors, and effect of these chemicals have not been sufficiently investigated. A major challenge in characterization and management of AFFF-impacted sites is the lack of information about identity and © 2016 American Chemical Society
concentration of PFASs present in proprietary AFFF formulations. Thus, several recent studies have focused on identifying specific PFASs in major formulations using advanced mass spectrometry techniques.11−13 These studies have proposed structures of a large suite of cationic, anionic, and zwitterionic PFAS surfactants made by both Simon’s electrochemical fluorination (ECF) and fluorotelomerizaton chemistry. It is notable that all the newly identified PFASs are polyfluoroalkyl compounds, which have perfluoroalkyl carbon chain lengths varying from 4 to 12 and possess functionalities such as sulfonyl, thioether, tertiary amine, quaternary ammonium, carboxylate, sulfonate, amine oxide, and betaine, etc. The extensive use of polyfluoroalkyl rather than perfluoroalkyl substances in AFFFs is probably due to a variety of functions that polyfluoroalkyl compounds can serve (e.g., as foaming, wetting, and dispersant agents), as well as high aqueous solubility, low volatility, and low acid strength.14 The identified cationic PFAS surfactants in these studies contain either tertiary amine or quaternary ammonium groups. Received: Revised: Accepted: Published: 9923
January 10, 2016 July 25, 2016 July 31, 2016 August 1, 2016 DOI: 10.1021/acs.est.6b00140 Environ. Sci. Technol. 2016, 50, 9923−9932
Article
Environmental Science & Technology
biotransformation of ECF-based AFFF surfactants will differ from previously reported fluorotelomer AFFF components. Because there is no information on the transformation potential of ECF-based or cationic PFAS surfactants, the present study was conducted to fill such a knowledge gap. Two such ECF-based cationic surfactants were selected, which appear in several patents for their use in AFFFs: perfluorooctane sulfonamide quaternary ammonium salt [PFOSAmS, F(CF2)8SO2NH(CH2)3N+(CH3)3] and perfluorooctaneamido quaternary ammonium salt [PFOAAmS, F(CF2)7CONH(CH2)3N+(CH3)3].15−19 Both compounds differ from the previously identified polyfluoroalkyl amines identified by D’Agostino and Mabury in the ammonium group: the identified compounds are tertiary amines, rather than quaternary ammonium salts.13 The corresponding perfluoroalkyl sulfonamide amine was also identified by Place and Field, and Backe et al. in AFFF formulations.11,12 As previously stated, the QACs are the effective ingredients whereas the tertiary amine compounds are probably synthesis impurity or breakdown products of QACs; therefore QACs are studied instead of the tertiary amine compounds. The primary objective of the present study is to determine the aerobic biotransformation potential of the PFAS-based QACs in soil microcosms and to examine biotransformation pathways and products. In addition, PFOSAmS and PFOAAmS mainly differ in the functional group that is immediately adjacent to the perfluoroalkyl chain, being sulfonyl and carbonyl groups, respectively; PFOAAmS has a C7 perfluoroalkyl chain whereas PFOS has C8. Therefore, it is of the interest of the authors to examine if the two cationic PFAS-based QACs have a similar potential to yield PFAAs as other negatively charged PFAS surfactants, despite their differences in structure, physicochemical properties, and bioavailability. The knowledge generated in this study is expected to improve the understanding of the environmental fate of cationic PFAS surfactants and provide more elements for a sound management in AFFF-impacted sites.
Quaternary ammonium compounds (QACs) are functioning surfactants due to the presence of the permanent charge on the -NR4+ group, whereas tertiary amines are only precursor materials used for synthesizing QACs.14 Therefore, it is possible that the tertiary amines identified in these studies11−13 are only impurities carried from synthesis processes rather than effective ingredients, or breakdown products of corresponding QACs. The extensive use of PFAS-based QACs in AFFF formulations or as fire suppressants is documented in many patents.15−19 Thus, the question arises: what is the significance of cationic PFASbased QACs, for which no data are available regarding environmental fate and effect, in AFFF-impacted sites? Another great challenge for characterizing and managing AFFF-impacted sites is inadequate understanding of the fate of PFAS surfactants once they are released into soil, surface, and groundwater environment. The primary environmental processes that can greatly complicate site investigation include microbially mediated degradation or abiotic transformation. It has been reported that, despite the general persistence of PFASs, the polyfluoroalkyl compounds with non-fluorinated functionalities can undergo partial microbial transformation. Under aerobic conditions, many PFASs can be microbially degraded to a complex mixture of products including perfluoroalkyl acids (PFAAs) such as PFOS and perfluorooctane carboxylic acid (PFOA).20−23 Understanding biotransformation potential is thus essential for performing site investigation and conducting remediation if the latter is deemed necessary. Biotransformation products could have different physicochemical properties, mobility, and toxicity from their parent compounds. For instance, Phillips et al. demonstrated that the fluorotelomer carboxylic acids pose higher acute toxicity to aquatic organisms than their parent compounds of fluorotelomer alcohols.24 Knowing key biotransformation products or biomarkers could also reveal the sources of PFAAs, whether they are directly from AFFF products or from natural biotransformation of other PFAS surfactants. Several AFFF components, all of which are fluorotelomerbased, have recently been examined for their biotransformation potential: 6:2 fluorotelomer sulfonate (6:2 FTS), fluorotelomer thioether amido sulfonates (FTAoS), and 6:2 fluorotelomer sulfonamido betaine (6:2 FTAB).25,26 In contrast, no ECF-based AFFF components have been investigated to date. HardingMarjanovic et al. found that, under aerobic conditions, FTAoS can be readily biodegraded through a few intermediates to form FTSs.26 The later biotransformation of FTSs produced a range of characteristic fluorotelomer products that were previously identified by Wang et al., including x:2 fluorotelomer unsaturated carboxylic acids, x:3 fluorotelomer carboxylic acids, and perfluoroalkyl carboxylic acids (C4−C8).27 6:2 FTAB was also found to degrade aerobically in activated sludge following similar pathways as demonstrated by Harding-Marjanovic et al.26,28 Since biotransformation pathways of these highly fluorinated compounds do not always follow known biotransformation pathways of non-fluorinated compounds, analysis by highresolution mass spectrometry was indispensable in determining empirical molecular formula and structures of degradation products. Previous research on the biotransformation of PFAS has demonstrated differences between ECF-based compounds (non-AFFF components) and fluorotelomer derivatives. In both cases the non-fluorinated moiety undergoes degradation, but fluorotelomer derivatives undergo partial defluorination while ECF-based derivatives do not show any degradation of the perfluoroalkyl moiety.29 For this reason, it is expected that
2. MATERIALS AND METHODS 2.1. Standards and Reagents. The model cationic surfactants of perfluorooctane sulfonamido ammonium iodide [PFOSAmS, CAS No. 335-90-0, F(CF2)8SO2NH(CH2)3N+(CH3)3I, 98%] and perfluoroctaneamido ammonium iodide [PFOAAmS, CAS No. 1652-63-7, F(CF2)7CONH(CH2)3N+(CH3)3I, 98%] were custom-synthesized at Beijing Surfactant Institute (Beijing, China). Three polyfluoroalkyl compounds that were used for identifying biotransformation products or as internal standards for quantitative chemical analysis were also synthesized: perfluoroctane sulfonamidoamido amine [PFOSAm, CAS No. 13417-01-1, F(CF2)8SO2NH(CH2)3N(CH3)2, 96%], perfluoroctane sulfonamido amine oxide [PFOSNO, CAS No. 30295-51-3, F(CF2)8SO2NH(CH2)3N(CH3)2O, 97%], and perfluoroctaneamido amine oxide [PFOANO, CAS No. 3029553-5, F(CF2)7CONH(CH2)3N(CH3)2O, 96%]. The synthesis processes for these compounds are briefly described in the Supporting Information (SI). A standard mixture of PFAAs (2 μg mL−1, >98% purity), which included the PFOS (linear isomer) and PFOA (linear isomer) were obtained from Wellington Laboratories (Guelph, Canada). Perfluorooctane sulfonamide acetate (FOSAA, 50 μg mL−1, >98% purity) was also from Wellington Laboratories. Perfluorooctane sulfonamide (FOSA, 98% purity) and N-ethyl perfluorooctane sulfonamide (EtFOSA, 98% purity) were purchased from Advanced Technology & Industrial Co. (Hong Kong, China). 9924
DOI: 10.1021/acs.est.6b00140 Environ. Sci. Technol. 2016, 50, 9923−9932
Article
Environmental Science & Technology
and the cap was resealed. The whole bottle was sonicated for 30 min and shaken for 120 min on a horizontal shaker (240 rpm) at ambient temperature. The extract was separated from the soil by centrifugation at 1600g, and the supernatant was pipetted out. This extraction step was repeated two more times. The three extracts were combined, evaporated to dryness, and reconstituted with 10 mL of methanol with 0.1% acetic acid and cleaned up with ENVI-Carb using the procedure developed by Powley et al.33 Extracts were stored at −20 °C before chemical analysis. Recoveries of the parent compounds and the quantifiable biotransformation products were found to be satisfactory in a spike recovery test (Table SI-2 in the SI). Methanol or acetonitrile alone could only recover no more than 40% of the dosed parent compounds (data not shown). Acetonitrile with sodium hydroxide, which provided complete recovery of perfluoroalkane sulfonamido derivatives from soil in our previous study,20 was deemed problematic because of possible amide hydrolysis and Stevens rearrangement and therefore not used.34 2.3. Quantitative LC-MS/MS Analyses. Quantitative analysis was performed using a Shimadzu UHPLC system coupled to an AB Sciex 5500 Qtrap mass spectrometer (LC-MS/ MS), working in multiple reaction monitoring (MRM) mode. Separation of PFOA, PFOS, EtFOSA, FOSA and FOSAA was achieved using an Ascentis Express F5 column (2.7 μm, 100 × 2.1 mm, Sigma-Aldrich, Oakville, ON, Canada), and the MS detection was under negative electrospray ionization. Separation of PFOAAmS, PFOSAmS, and PFOSAm was performed with a Kinetex C18 column (2.6 μm, 50 × 3.0 mm, Phenomenex, Torrance, CA, USA), and ionization was achieved with both positive (for ammonium salts) and negative (for amine) electrospray ionization. In both cases, a Kinetex EVO C18 column (5 μm, 50 × 3.1 mm2, Phenomenex) installed upstream of the UHPLC autosampler was used as a delay column to separate the PFASs leaching from polytetrafluoroethylene parts of the instrument, particularly the degasser. Immediately before the chemical analysis, the extracts were spiked with internal standards at 0.4 ng/mL. As no isotope-labeled standards were available for PFOAAmS, PFOSAmS, and PFOSAm, structurally similar compounds PFOANO and PFOSNO were used as the internal standards for PFOAAmS and PFOSAmS/PFOSAm, respectively. The details on chromatographic methods, mobile phases, monitored transitions, and calibration methods can be obtained from the SI. 2.4. Identification of Biotransformation Products. The soil extracts after concentration by nitrogen evaporation and the headspace extracts without concentration were analyzed under full scan to determine the exact mass of suspected biotransformation products. The analysis was performed on a Dionex UHPLC system coupled to a Q-Exactive Orbitrap mass spectrometer working in full scan mode (R, 70,000 at m/z = 200) with positive and negative heated electrospray ionization. Chromatographic separation was achieved with a Thermo C18 Hypersil aQ Gold column (1.9 μm, 100 × 2.1 mm2). For structure confirmation of proposed degradation products, soil extracts were further analyzed in t-MS2 positive and negative ionization mode (normalized collision energy, NCE = 20−70%) using the same UHPLC-MS system. Qualitative analysis was based on a method previously developed by Munoz et al.35 The level of confidence in the structures identified is reported following the system established by Schymanski et al.36 2.5. Determination of Biotransformation Kinetics. Single first-order (SFO) kinetics model has been preferred in
The information on other chemicals and reagents can be found in SI. Test Soil. The soil used for setting up microcosms was collected in 2014 from a nonimpacted area in proximity to a contaminated site (Quebec, Canada). As shown in Figure SI-1 in the SI, the soil was obtained from the zone next to Rivière Chaudière, upstream of the downtown area of Lac-Mégantic, where about 33,000 L of AFFF concentrates were deployed for controlling crude oil fire in a major train derailment accident in July 2013.30 The top 20 cm layer was collected, sieved via a 2 mm sieve immediately after collection, and stored at 4 °C until use. It is a sandy loam (59.2% sand, 32.2% silt, and 8.6% clay) with 3.1% organic matter (OM), pH of 5.0, CEC of 7.2 meguiv/(100 g), 11 ppm phosphorus-P, and 64 ppm nitrate-N. As the levels of 18 selected PFAS species were found in the low levels of 0.02−0.77 ng/g (listed in Table SI-1), comparable to other soils that had no known PFAS exposure history,31 the soil was deemed not impacted by the AFFF deployment. Before the commencement of the biotransformation experiment, part of the sieved soil was sterilized through three cycles of autoclaving (60 min per cycle at 121 °C and 15 psi) with intermittent incubation at ∼22 °C for 24 h between autoclaving cycles. Three antibiotics (chloramphenicol, kanamycin, and cycloheximide) were spiked into the soil to reach an approximate concentration of 100 mg/(kg of soil) to prevent the revival of microbial activity during long-term incubation.21 Soil Microcosms. Closed test vessels were created by using 50 mL amber serum bottles fitted with crimp-sealed natural rubber stoppers, so as to minimize loss of potential volatile products. Similar closed systems have been previously used for evaluating the biodegradability of fluorotelomer alcohols.21,32 Approximately 5 g of soil (dry weight (dw) equivalent at 103 °C) was added to each vessel, adjusted to a gravimetric moisture content of 23% (∼80% of water holding capacity at 1/3 bar) and was preincubated at ∼22 °C for 5 days. Then the soil was dosed with 10 μL of a methanol stock solution of a single component (or methanol only for untreated soil) to result in a starting concentration of 1 μg/g (dw). The dosed soil was mixed manually using a sterile polypropylene spatula. Three treatments were prepared for each model compound: (a) untreated (matrix) live soil with 10 μL of methanol; (b) treated live soil, and (c) treated sterile soil. A total of 30 vessels were prepared for each treatment and 90 vessels for each individual compound. At every sampling point, three vessels of each treatment were sacrificed for analysis. 2.2. Sampling and Sample Preparation. Sampling points were chosen to be 0, 3, 7, 14, 28, 60, 90, 120, and 180 days, and at each sampling point, the whole vessels of live, sterile, and matrix samples were sacrificed. The concentration of headspace O2 was measured using a Quantek oxygen and carbon dioxide analyzer 902D (Grafton, MA, USA) before sampling. When the level of O2 was below 10% saturation, all remaining bottles that were to be sampled on a future date were flushed with sterile air (filtered with 0.3 μm air filter) to reaerate. The headspace was first sampled by flushing with air through a Maxi-Clean C18 cartridge to capture potential volatile products. The cartridge was later eluted with acetonitrile, and the eluent was stored at −20 °C before chemical analysis. Then the bottle’s crimp cap was opened to allow whole-bottle extraction and the original crimp cap was also extracted at the same time. The solvent extraction method was modified from Houtz et al.,8 and validated before the commencement of biotransformation experiments. Briefly, 10 mL of methanol with 0.1% NH4OH was added to each bottle 9925
DOI: 10.1021/acs.est.6b00140 Environ. Sci. Technol. 2016, 50, 9923−9932
Article
Environmental Science & Technology
FOMC) that are commonly used in simulating biodegradation and biotransformation kinetics of pesticides in soil.39 All four models passed the test of goodness of fit with 95% confidence level whereas DFOP showed the smallest χ2 error, as detailed in the SI. DT50 value of PFOAAmS was estimated to be 142 days in the DFOP model and 127 days (or half-life) in the SFO model. For other known PFOA precursors reported in the literature, which are largely fluorotelomer alcohol derivatives with C8 perfluoroethyl chain, their half-lives (as determined using SFO model) in aerobic soils ranged from
