Certain Perfluoroalkyl and Polyfluoroalkyl ... - ACS Publications

Subscriber access provided by READING UNIV

Article

Certain Perfluoroalkyl and Polyfluoroalkyl Substances Associated with Aqueous Film Forming Foam Are Widespread in Canadian Surface Waters Lisa Anne D'Agostino, and Scott Andrew Mabury Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b03994 • Publication Date (Web): 07 Nov 2017 Downloaded from http://pubs.acs.org on November 10, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 30

1 2 3

Environmental Science & Technology

Certain Perfluoroalkyl and Polyfluoroalkyl Substances Associated with Aqueous Film Forming Foam Are Widespread in Canadian Surface Waters

4

Lisa A. D’Agostino and Scott A. Mabury*

5

Department of Chemistry, University of Toronto, 80 St George Street, Toronto, M5S

6

3H6, ON, Canada

7

[email protected]

8

TOC Art

9

ACS Paragon Plus Environment


10

Page 2 of 30

Abstract:

11

The presence of perfluoroalkyl and polyfluoroalkyl substances (PFASs)

12

commonly associated with aqueous film forming foams (AFFFs) at sites without known

13

AFFF contamination is a largely unexplored area, which may reveal widespread

14

environmental

15

chromatography-tandem mass spectrometry (LC-MS/MS) screening for 23 classes of

16

PFASs, followed by quantitative analysis was used to investigate surface waters from

17

rural, urban, and AFFF-impacted sites in Canada. The PFASs detected included

18

perfluorohexane sulfonamide (FHxSA), 6:2 fluorotelomer sulfonamide (FTSAm),

19

fluorotelomer sulfonamide alkylbetaines (FTABs), fluorotelomer betaines (FTBs), 6:2

20

fluorotelomer

21

fluorotelomerthiohydroxyl ammonium sulfoxide (FTSHA-SO), 6:2 fluorotelomer

22

sulfonamide alkylamine (FTAA) and C3 to C6 perfluoroalkane sulfonamido amphoterics.

23

Detection of FHxSA in all urban and AFFF-impacted sites (0.04–19 ng/L) indicates the

24

widespread presence of rarely considered perfluorohexane sulfonate (PFHxS) precursors

25

in Canadian waters. FTABs and FTBs were especially abundant with up to 16–33 ng/L of

26

6:2 FTAB in urban and AFFF-impacted water suggesting it may have additional

27

applications, while FTBs were only in AFFF-impacted sites (qualitative; ΣFTBs 80

28

ng/L). The distributions of PFASs moving downstream along the AFFF-impacted

29

Welland River and between water and sediment suggested differences in the persistence

30

of various AFFF components and enhanced sorption of long-chain fluorotelomer

31

betaines. Total organofluorine combustion-ion chromatography (TOF-CIC) revealed that

32

fluorotelomer betaines were a substantial portion of the organofluorine in some waters

33

and 36–99.7% of the total organofluorine was not measured in the targeted analysis.

34

Introduction

contaminants

requiring

mercaptoalkylamido

further

sulfonate

investigation.

sulfone

Sensitive

(FTSAS-SO2),

liquid

6:2

35

Since 1999, aqueous film forming foams (AFFFs) used to fight liquid-fuelled fires

36

have been implicated in local environmental contamination with perfluoroalkyl and

37

polyfluoroalkyl substances (PFASs),1 particularly perfluoroalkyl carboxylates (PFCAs),1–

38

10

perfluoroalkane sulfonates (PFSAs),2–7,9–11 and fluorotelomer sulfonic acids

2 Environment ACS Paragon Plus

Page 3 of 30


39

(FTSAs).4,12 Extremely high concentrations of PFASs have been associated with AFFF,

40

including up to 2,210,000 ng/L of perfluorooctane sulfonate (PFOS) in Etobicoke Creek

41

immediately after an AFFF release from Toronto Pearson Airport, Ontario in 20003 and

42

up to 920,000 ng/L of perfluorohexane sulfonate (PFHxS) and 14,600,000 ng/L of 6:2

43

FTSA in groundwater at Tyndall Air Force Base, Florida.4 Inputs of AFFF are also

44

implicated in less extreme contamination of surface waters over longer time frames,

45

including 290 ng/L of PFOS in Etobicoke Creek in 2009 near the AFFF release site from

46

2000,7 and 45–64 ng/L of PFOS in Lake Niapenco downstream of Hamilton Airport,

47

Ontario in 2010.5

48

However, while PFOS is a major component of AFFFs prepared from

49

electrochemical fluorination products and PFHxS may be found in these products at

50

concentrations less than 15% of the PFOS concentration, most of these PFCAs, PFSAs,

51

and FTSAs are not major PFASs in AFFFs relative to the primary PFAS active

52

ingredients.14–16 In the past five years, over twenty additional classes of PFASs in AFFFs

53

and commercial fluorinated surfactant concentrates17,18 and several additional degradation

54

products of AFFF components14,19–21 have been identified. Reports of these recently

55

identified AFFF components in the environment are limited and usually focus on sites

56

known to have received substantial inputs of AFFF. Efforts to determine which of the

57

AFFF-related PFASs can be detected at trace levels in surface waters without known

58

AFFF impacts, such as urban and rural surface waters, are lacking. Only a study that

59

estimated concentrations of fluorotelomer sulfonamide alkylamines (FTAAs) and

60

fluorotelomer sulfonamide alkylbetaines (FTABs) in French river sediments at 0.055–

61

13.5 ng/g in total investigated sites without significant known impacts from AFFF or

62

AFFF component manufacturing.13

63

The highly impacted sites that have been investigated in terms of novel AFFF-

64

related PFASs include on and around military bases with long histories of AFFF use,14–19

65

near airport and other firefighting training areas,19,20 the areas surrounding large oil fires

66

that were fought with AFFF,19,21–23 and near a location of PFAS manufacturing for

67

AFFF.24,25 These studies have determined that a variety of AFFF-related PFASs are

68

present in environmental samples from these sites, with one or more congeners of the

69

following fluorotelomer PFAS classes found in various environmental media:



FTAAs,17,19–23,25

Page 4 of 30

70

FTABs,17,19–25

71

(FTSASs),14,21,23

72

SO2s),18 FTSAS-related thioether/sulfone-linked carboxylic acids,18 fluorotelomer

73

sulfonamides (FTSAms),21 fluorotelomerthiohydroxyl ammoniums (FTSHAs),21,23

74

FTSHA-sulfoxides (FTSHA-SOs),21,23 demethyl-FTAAs,21 and fluorotelomer betaines

75

(FTBs).21,23 Perfluoroalkane sulfonamido substances found in AFFFs or their likely

76

degradation products have also been measured on highly AFFF impacted U.S. military

77

bases, including C4–C6 perfluoroalkane sulfonamido alkylamine acids (FASAAAs),14

78

C4–C6 perfluoroalkane sulfonamido alkylamines (FASAAms),14,15 and perfluorohexane

79

sulfonamide (FHxSA).15,16 Many of these studies provided quantitative or semi-

80

quantitative results for the novel AFFF-related PFASs in environmental matrices.14–

81

17,19,21–25

FTSAS-sulfoxides

fluorotelomer

mercaptoalkylamido

(FTSAS-SOs),21,23

FTSAS-sulfones

sulfonates (FTSAS-

82

However, by focusing investigations on highly AFFF impacted sites, previous

83

studies were not able to address how widespread these contaminants are in the

84

environment and whether the patterns in their detection and measured or estimated

85

concentrations may suggest that they have applications outside of AFFF. For example,

86

according to a patent, 6:2 FTAB may be used to coat ceramic surfaces to prevent

87

deposits.26 In addition, a "betaine partially fluorinated surfactant," Capstone FS-50, is

88

marketed for various applications, including cleaning and floor care, and has identical

89

listed properties to Capstone 1157, which is marketed for use in AFFFs and is

90

presumably a renamed Forafac 1157, which contained mostly 6:2 FTAB.20,27,28 Herein,

91

by screening for 23 classes of PFASs with high sensitivity and quantifying, as possible,

92

the AFFF-related PFASs found in surface water samples from rural, urban, and AFFF

93

impacted locations in Canada, information about the environmental occurrence of PFASs

94

typically considered AFFF-related is obtained. This shows which of these PFASs are

95

relatively widespread in the environment due to their widespread use, high persistence,

96

and/or environmental transport, which can help prioritize future research into the most

97

environmentally relevant of these PFASs.

98

In addition to the occurrence of AFFF-related PFASs in surface waters, the total

99

organofluorine content of surface water extracts was determined by total organofluorine

100

combustion ion chromatography (TOF-CIC). Comparison of results of this technique


Page 5 of 30


101

with targeted, quantitative analysis can show when a substantial amount of the total

102

organofluorine is made up of additional components containing organofluorine that were

103

not quantified. It has been applied to numerous matrices, including sediments,29 water,30

104

and AFFF concentrates31 demonstrating that substantial amounts of organofluorine are

105

not accounted for by targeted analysis. For example, the PFCAs and PFSAs measured in

106

the acidic fraction of Lake Ontario surface sediment extracts accounted for only 2 to 44%

107

of the total organofluorine.29 Using TOF-CIC, the contribution of novel AFFF-related

108

PFASs to the total organofluorine in the surface water samples and the size of the

109

remaining unknown fraction was investigated.

110

Another important aspect of the behaviour of AFFF-related PFASs in surface

111

water environments is sorption to sediment, because increased sorption can reduce the

112

transport of PFASs downstream.32,33 Neither laboratory nor field determinations of the

113

sorption behaviour of the novel AFFF-related PFASs are available. With this gap in

114

knowledge, although the focus of this investigation is surface water, sediment samples

115

were collected concomitantly with water samples from a subset of sites along the AFFF-

116

impacted Welland River5,34 and at one nearby reference site. By comparing the PFAS

117

profiles in corresponding sediment and water samples, preliminary insights into the

118

sorption behaviour of the novel AFFF-related PFASs found in the Welland River were

119

obtained. Where PFASs were quantified in both sediment and water at a site, field-

120

derived sediment-water distribution coefficients (logKd) and organic carbon normalized

121

distribution coefficients (logKOC) were determined. The determined field-derived logKOC

122

values for PFAAs were compared to previous field derived values in river and coastal

123

sediments,8,33,35 while field derived logKOC values and sediment concentrations of novel

124

AFFF-related PFASs provide an indication of their sorption behaviour.

125

Materials and Methods

126

Materials

127

6:2 FTAA, 6:2 FTAB, and 6:2 FTSAm were synthesized in house. As described

128

in detail previously,36 6:2 fluorotelomer sulfonyl chloride (FTSO2Cl) was prepared by a

129

one step synthesis from 6:2 fluorotelomer thiol, using KNO3 and SO2Cl2. The 6:2



130

FTSO2Cl was reacted with N,N-dimethylamino-1-propylamine to form 6:2 FTAA or with

131

NH3 in tetrahydrofuran to form 6:2 FTSAm. The 6:2 FTAA was used to prepare 6:2

132

FTAB by refluxing it with sodium chloroacetate.36 A listing of the other solvents,

133

chemicals, and standards used is in the Supporting Information (SI).

134

Samples

135

Welland River (sample points 1 to 9) are sequential sites along the AFFF-

136

impacted Welland River. Because it was not possible to sample upstream of the former

137

firefighting training site on the Welland River, since this site is in the headwaters, Big

138

Creek (2 sample locations) and Welland River Reference, which are nearby rural sites not

139

downstream of the former firefighting training site at Hamilton Airport, were sampled as

140

rural reference sites. Water was collected in submerged 500 mL wide-mouth

141

polypropylene bottles (Nalgene, Penfield, NY) on October 22, 2015. Water samples from

142

two AFFF-impacted Arctic lakes, Resolute Lake and Meretta Lake, were collected in

143

August 2012 and 2014 in 1 L polyethylene bottles (Kartell, Noviglio, Italy) by

144

Environment and Climate Change Canada (ECCC) personnel. Rivers in Southern Ontario

145

were sampled in conjunction with Ontario Ministry of the Environment and Climate

146

Change (MOECC) sampling programs in 2015 with 500 mL polyethylene terephthalate

147

jars used routinely for monitoring perfluoroalkyl acids (PFAAs). Of these rivers, Perch

148

Creek, Thames River, Grand River, Humber River, and Don River are classified as urban;

149

Etobicoke Creek has known prior AFFF-impacts. Additional rural samples were collected

150

using 500 mL wide-mouth polypropylene bottles (Nalgene) from Little Rouge Creek in

151

2016 and Lake of Bays in 2015 some of which were used for matrix spike and recovery.

152

Samples were refrigerated at 4°C prior to extraction. Sampling dates and locations are in

153

SI Figure S1 and Table S1. Sediment samples were obtained in 500 mL wide-mouth

154

polypropylene bottles (Nalgene) from four Welland River and one Big Creek site at the

155

same time and place as water samples. Sediments were frozen at -20°C and lyophilized

156

prior to extraction.

157

Because the Welland River upstream of the Binbrook Dam flows very slowly,

158

these sediments were likely at or near equilibrium with the surrounding waters. The

159

month of October 2015, prior to sampling, was relatively dry and water levels in the


Page 6 of 30

Page 7 of 30


160

reservoir (Lake Niapenco) remained below the holding level of 650.5 feet throughout

161

September and October 2015, prior to the sampling.37 Therefore, the residence time of

162

water likely was quite long prior to sampling and the water was relatively stagnant.

163

Observations at the time of sampling support this assessment as the water was not

164

observed to flow and the flow out of the Binbrook Dam was a trickle.

165

Water Extraction

166

Water samples were extracted in the Advanced Laboratory for Fluorinated and

167

Other New Substances in the Environment (ALFONSE) clean laboratory (Class 100A)

168

using Oasis WAX solid phase extraction (SPE) cartridges (6 mL, 200 mg, 30 µm; Waters,

169

Milford, MA). The extraction protocol used to extract the approximately 500 mL water

170

samples was a modification of existing methods38 and is described fully in the SI. The

171

cartridges were eluted with methanol (fraction containing neutrals, zwitterions, bases, and

172

cations) followed by 0.1% NH4OH in methanol (fraction containing acids).

173

Sediment Extraction

174

Approximately 0.5 g subsamples of lyophilized sediment were weighed into

175

polypropylene centrifuge tubes and extracted using 0.1% NH4OH in methanol based on

176

Houtz et al.15 with clean-up using 1mL/100mg Supelclean ENVI-Carb SPE cartridges

177

(Supelco, Bellefonte, PA) in the ALFONSE clean laboratory. The full extraction

178

procedure and the method for determination of organic carbon content are described in

179

the SI.

180

LC-MS/MS

181

Analysis of sample extracts for PFASs by LC-MS/MS was performed using an

182

Acquity UPLC–Xevo TQ-S system (Waters, Milford, MA). An Acquity UPLC BEH C18

183

column (1.7 µm, 2.1 mm×75 mm) and a mobile phase gradient of 10 mM aqueous

184

ammonium acetate and methanol were used. The Xevo TQ-S was operated in positive

185

and negative electrospray (ESI) mode. Further LC-MS/MS method details are in the SI.

186

Transitions for AFFF components (SI Table S2) were optimized by infusing dilutions of



Page 8 of 30

187

suitable AFFF extracts prepared previously.39 Screening LC-MS/MS runs were used to

188

determine which PFASs may be present in sample extracts and included transitions for

189

additional AFFF-components that were never detected.

190

The suite of PFAAs analyzed is limited to PFOS or PFOA and shorter in order to

191

focus on those PFAAs most associated with AFFFs, which generally contain primarily

192

PFOS and C6 or C8 perfluoroalkane sulfonamido substances or 6:2 fluorotelomer

193

surfactants. The shortest PFCA analyzed was PFPeA because PFBA was poorly retained

194

on the reverse phase column and suffered from background contamination.

195

TOF-CIC

196

Surface water extracts from each sampling location were analyzed by total

197

organofluorine-combustion ion chromatography (TOF-CIC) using previously published

198

procedures.31,40 The acid extracts were also analyzed for inorganic fluoride by combining

199

100 µL of extract with 6.5 mL of deionized water and directly analyzing by ion

200

chromatography using the TOF-CIC system.

201

Quality Assurance of Data

202

Analyses by LC-MS/MS were completed in at least duplicate (n=2–4).

203

Quantification of PFASs was performed in several ways depending on the availability of

204

standards.

205

perfluoroheptanoate (PFHpA), perfluorooctanoate (PFOA), PFHxS, PFOS, 4:2 FTSA,

206

6:2 FTSA, 8:2 FTSA, N-ethyl perfluorooctane sulfonamido acetic acid (EtFOSAA), and

207

perfluorooctane sulfonamide (FOSA) were quantified using internal calibration with

208

isotopically labeled standards. Perfluorobutane sulfonate (PFBS), perfluoropentane

209

sulfonate (PFPeS), perfluoroheptane sulfonate (PFHpS), FHxSA, and 6:2 FTSAm were

210

quantified using structurally related surrogate mass-labeled internal standards: mass-

211

labeled PFHxS for PFBS and PFPeS, mass-labeled PFOS for PFHpS, and mass-labeled

212

FOSA for FHxSA and 6:2 FTSAm. Two multiple reaction monitoring transitions were

213

used to confirm detection of all PFASs, except PFPeA, PFHxA, EtFOSAA, FHxSA, and

214

FOSA, where a single transition was used. For 6:2 FTAB and 6:2 FTAA, matrix matched

215

calibration was used along with mass-labeled FOSA internal standard. The data are

Perfluoropentanoate

(PFPeA),

perfluorohexanoate


(PFHxA),

Page 9 of 30


216

considered quantitative for the PFASs analyzed as described above with commercial

217

analytical standards or standards prepared from neat materials. Estimation of the

218

concentrations for qualitative analytes (novel PFASs) for which isolated standards were

219

unavailable was performed by using matrix matched calibration curves for 6:2 FTSA for

220

sulfonic acid-containing compounds, EtFOSAA for carboxylic acid compounds, 6:2

221

FTAA for amine or quaternary ammonium compounds, and 6:2 FTAB for amphoteric

222

compounds. To confirm the detection of qualitative AFFF components, two MS/MS

223

transitions were used and retention times were matched to AFFF extracts containing

224

those components from a previous study.39

225

The calibration curves had a minimum of five points and were usually constructed

226

from 0.02 ng/mL to 10 ng/mL in the vial (range: 0.006–0.03 ng/mL to 10 ng/mL). This

227

corresponds to 0.08 ng/L to 40 ng/L in a 500 mL water sample and 0.08 ng/g to 40 ng/g

228

in a 0.5g sediment sample. A quality control sample spiked at 1 ng/mL in the vial with all

229

the PFASs being quantified was analyzed in duplicate during each analysis and analyses

230

were accepted if the concentrations determined were within 20%. Where the

231

concentration of a PFAS was above the highest calibration point in a sample, that sample

232

was rerun with a suitable dilution factor. The limit of detection (LOD) was defined as the

233

concentration giving a signal-to-noise ratio of 3 in the sample matrix and the limit of

234

quantitation (LOQ) was the concentration giving a signal-to-noise ratio of 10 in the

235

sample matrix. Matrix LODs and LOQs are given in SI Table S4. During each water

236

extraction, two cartridge blanks were subjected to all portions of the extraction procedure

237

except sample loading to evaluate contamination of the extraction solvents and contained

238

no PFASs above the LOD. Two field blanks of deionized water that were opened to air

239

during the collection of one Welland River sample were also extracted and found to

240

contain no PFASs above the LOD. For the sediment extraction, two solvent blanks

241

containing no sediment were extracted along with the sediments and contained no PFASs

242

above the LOD. Inter-day duplicate extractions were performed on Welland River 1, 2, 5,

243

6, 7, and 9; Welland River Reference; and Big Creek 1 water samples, while intra-day

244

duplicate extractions were performed on all sediments and Little Rouge Creek water.

245

Recoveries for water extraction were determined by spiking surface water

246

samples (~500 mL) from relatively clean environments (n=5) with 2.5 ng of FTSAs,



247

EtFOSAA, FHxSA, FOSA, 6:2 FTSAm, 6:2 FTAA, and 6:2 FTAB and 25 ng of PFCAs

248

and PFSAs in methanol, swirling to mix, and extracting the following day. The recoveries

249

were 71–96% for all PFASs, including recoveries of 88±24% for 6:2 FTAB (zwitterionic)

250

and 71±12% for 6:2 FTAA (cationic at surface water pH; SI Table S3). Recoveries from

251

freeze dried sediment were determined by spiking sediment from Big Creek 1 (~0.5 g,

252

n=4) with 1 ng each of PFCAs, FTSAs, EtFOSAA, FOSA, FHxSA, 6:2 FTSAm, 6:2

253

FTAB, and 6:2 FTAA, and 5 ng each of PFSAs in 100 µL of methanol, vortexing to mix,

254

and extracting the sediment the following day. Recoveries from sediment were 76–94%

255

for the PFASs, except 6:2 FTAB (31±2%; SI Table S3). Sediment concentrations are

256

reported without correction.

257 258

Statistical Analysis

259

Statistical comparison of the concentrations of PFASs found at rural (n=5), urban

260

(n=5), and AFFF-impacted (n=14) sampling sites were performed using the two-tailed

261

Mann-Whitney U test. For the statistical analysis, values determined between LOQ and

262

LOD were used as determined from the calibration and concentrations below the

263

detection limit were substituted with half the LOD. The Mann-Whitney U test was only

264

used for PFASs detected in at least 21 out of 24 samples or 17 out of 19 urban and AFFF

265

impacted samples to limit the impact of non-detects. The threshold for statistical

266

significance used was p

Certain Perfluoroalkyl and Polyfluoroalkyl ... - ACS Publications

Recommend Documents