Assessment of the Influence of Soil Characteristics and Hydrocarbon

Subscriber access provided by Van Pelt and Opie Library

Article

Assessment of the influence of soil characteristics and hydrocarbon fuel co-contamination on the solvent extraction of per- and polyfluoroalkyl substances Sandra Mejia-Avendaño, Gabriel Munoz, Sébastien Sauvé, and Jinxia Liu Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.6b04746 • Publication Date (Web): 20 Jan 2017 Downloaded from http://pubs.acs.org on January 23, 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.

Analytical Chemistry 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 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

1

Assessment of the influence of soil characteristics and

2

hydrocarbon fuel co-contamination on the solvent extraction

3

of per- and polyfluoroalkyl substances

4 Sandra Mejia-Avendañoa, Gabriel Munoza,b, Sébastien Sauvéb, Jinxia Liua*

5 6 7

a

Department of Civil Engineering, McGill University, Montreal, Quebec, H3A 0C3, Canada

b

Department of Chemistry, Université de Montréal, Montreal, Quebec, H3C 3J7, Canada

8 9 10

*Corresponding author. Tel: +1 514 398 7938; fax: +1 514 398 7361

11

E-mail address: [email protected]

12 13 14

Page 1 of 27

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 32

15

Abstract

16

Sites impacted by the use of Aqueous Film-Forming Foams (AFFFs) present elevated

17

concentrations of per- and polyfluoroalkyl substances (PFAS). The characterization of the

18

PFAS contamination at such sites may be greatly complicated by the presence of

19

hydrocarbon co-contaminants, and by the large variety of PFAS potentially present in

20

AFFFs. In order to further a more comprehensive characterization of AFFF-contaminated

21

soils, the solvent extraction of PFAS from soil was studied under different conditions.

22

Specifically, the impact of soil properties (textural class, organic matter content) and the

23

presence of hydrocarbon contamination (supplemented in the form of either diesel or crude

24

oil) on PFAS recovery performance was evaluated for two extraction methods

25

(MeOH/NaOH and MeOH/NH4OH). While both methods performed satisfactorily for

26

perfluoroalkyl acids and fluorotelomer sulfonates, the extraction of newly identified

27

surfactants with functionalities such as betaine and quaternary ammonium was improved

28

with the MeOH/NaOH-based method. The main factor that was found to influence the

29

extraction efficiency were the soil properties; a high organic matter or clay content was

30

observed to negatively affect the recovery of the newly identified compounds. While the

31

MeOH/NaOH solvent yielded more efficient recovery rates overall, it also entailed the

32

disadvantage of presenting higher detection limits and substantial matrix effects at the

33

instrumental analysis stage, requiring matrix-matched calibration curves. The results

34

discussed herein bear important implications for a more comprehensive and reliable

35

environmental monitoring of PFAS components at AFFF-impacted sites.

Page 2 of 27


Page 3 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Introduction

36 37

Per- and polyfluoroalkyl substances (PFAS) have unique physicochemical properties that

38

have made them highly suitable for a variety of industrial uses and consumer products for

39

more than 60 years. After the recognition of their negative environmental and health

40

effects,1-5 long-chain perfluoroalkyl acids (PFAAs) have been banned or phased out in

41

several countries, including the United States,6 Canada,7 and the European Union.8 The

42

Swedish Chemicals Agency estimates that more than 3000 PFASs are in circulation

43

globally.9 Certain PFAS remain in use for applications such as fighting hydrocarbon fuel

44

fires, due to their ability to create a thin film over the fuel preventing oxygen contact and fire

45

reignition.10-12 Such firefighting formulations are known as aqueous film-forming foams

46

(AFFFs).

47 48

AFFFs are complex mixtures usually composed of fluorosurfactants as the key active

49

component, as well as hydrocarbon-based surfactants and other additives including foam

50

stabilizers and corrosion inhibitors. The actual identity and concentration of most PFAS

51

species in AFFFs is rarely known due to the proprietary nature of their formulations. As the

52

use of long-chain PFAAs such as perfluorooctane sulfonate (PFOS) and perfluorooctanoate

53

(PFOA) in AFFFs is on the decline, there has been a progressive shift of the

54

fluorosurfactant industry towards other fluorinated alternatives.13 Current or new

55

formulations of AFFFs favor the use of short-chain PAAS because of their lower

56

bioaccumulative potential as compared to long-chain perfluoroalkyl analogs.14 New

57

formulations also make use ofpartially fluorinated PFAS (i.e. fluorotelomer derivatives), for

58

which there is a general lack of information regarding bioaccumulation or toxicity. In recent Page 3 of 27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 32

59

years, several studies have sought to identify the composition of AFFF formulations using

60

high-resolution mass spectrometry.15-17 Such research resulted in the discovery of dozens

61

of PFAS classes with different functionalities and chain lengths. A few examples of the

62

newly identified PFAS classes in firefighting formulations include betaine-based PFAS (e.g.,

63

6:2 fluorotelomer sulfonamide betaine, 6:2 FTAB, Figure 1), amine oxides (e.g.,

64

Perfluoroalkyl sulfonamidoalkyl amine oxide, PFOSNO, Figure 1), quaternary ammonium

65

PFAS (e.g., perfluorooctaneamide ammonium salt, PFOAAmS, Figure 1) and fluorotelomer

66

thioether derivatives (e.g., 6:2 Fluorotelomer thioether amido sulfonate, 6:2 FTSAS). Many

67

PFAS species that are employed in the new AFFFs have the potential to yield PFAAs of

68

various chain lengths due to environmental biotic or abiotic transformation processes.

69

Previously, some AFFF components (e.g., 6:2 FTSAS and PFOAAmS) have been

70

confirmed through experimental work to be precursors of PFAAs.18-20

71 72

The co-existence of such a high number of PFAS components greatly complicates the

73

assessment and management of AFFF-contaminated sites. The lack of authentic standards

74

for most of the newly-identified PFAS is one major limiting factor to allow for proper

75

investigations of the environmental occurrence and fate of these chemicals in natural or

76

engineered systems impacted by AFFFs. When only a limited number of compounds

77

(typically PFOS, PFOA and a few other PFAAs) are monitored, the total amount of PFAS

78

present at AFFF-impacted sites could be greatly underestimated.16 In some cases, the

79

environmental risk assessment at AFFF-impacted sites may even yield conflicting results.

80

For instance, it is possible that a contaminated site could display increasing PFAAs levels

81

over time due to the biotransformation of non-monitored PFAS to PFAAs. Page 4 of 27


Page 5 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


82 83

Following fire-fighting training activities or AFFF deployment in the context of an emergency

84

response, the soil is commonly the first environmental compartment with which AFFFs will

85

have contact. Soil plays a key role in PFAS sequestration and transfer to groundwater or

86

surface water bodies,21-22 as well as in the possible transformation of PFAS into PFAAs23.

87

Investigating the environmental fate of PFAS in soils from AFFF-impacted sites therefore

88

requires paramount attention. To date, a number of analytical methods have addressed the

89

analysis of PFAS in sediments or soils,24-26 but none has yet been specifically assessed for

90

soil matrixes at AFFF-impacted sites. At these sites, PFAS commonly co-exist with

91

petroleum hydrocarbons, which originally fuelled the fires and are the reason for the use of

92

PFAS-based AFFFs in the first place. As petroleum hydrocarbons are generally present in

93

much higher abundance (e.g., in the order of hundreds to thousands parts per million) than

94

PFASs, chemical extraction and analysis of PFAS in soils is arguably impacted by the co-

95

existence of hydrocarbons. While the available PFAS methods may have performed well in

96

terms of recovery or detection limits for sediments or soils with little or no petroleum

97

hydrocarbon burden, it has yet to be established if similar performances could be obtained

98

when petroleum hydrocarbons co-occur at significant concentration levels. Other factors

99

that are also often overlooked in analytical method development include inherent soil

100

characteristics such as soil texture and organic matter (OM) content. For the sake of cost-

101

effectiveness, analysts traditionally apply a single preparation procedure and instrumental

102

method to a set of sediment or soil environmental samples, relying on internal standard

103

correction to compensate for variable matrix effects or recoveries between samples. To

104

integrate matrix effects at the quantification stage, a matrix-based calibration curve may be Page 5 of 27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 32

105

used. However, some samples could greatly diverge in terms of soil texture, OM or co-

106

contaminant content from the model sample, which may ultimately question the reliability of

107

the analytical results generated.

108 109

In this context, the objective of the present work was to investigate the factors affecting the

110

recovery of different classes of PFAS from a soil matrix in the presence of hydrocarbon co-

111

contaminants. A methanolic solvent extraction under basic conditions was selected as

112

representative of reported extraction methods for PFAS in soils.20, 26-28 The influence of soil

113

texture, two types of co-contaminants, and concentration of co-contaminants on the

114

recovery performance of 36 PFAS species was investigated. The model PFAS under

115

investigation represent a vast array of PFAS chemistries, including PFAAs of a wide span

116

of perfluoroalkyl chain lengths (C4–C16), fluorotelomer sulfonates and carboxylates (FTSAs

117

and FTCAs, respectively), perfluorooctane sulfonamide derivatives (FOSA, EtFOSA and

118

FOSAA), fluorotelomer unsaturated acids (FTUAs), as well as 9 zwitterionic or cationic

119

PFAS representing several classes of newly-identified PFAS. To the knowledge of the

120

authors, this is the first study to comprehensively integrate the influence of petroleum co-

121

contaminants on the analytical chemistry of pollutants of emerging concern.

122

Materials and methods

123 124

Standards and Reagents

125

Chemical structures of the thirty-six model PFAS for which authentic standards are

126

available are presented in Figure 1. Perfluoroalkyl acids (PFCAs, C4-C14 and C16),

127

perfluoroalkyl sulfonates (PFSAs, C4, C6, C7, C8, C10), fluorotelomer unsaturated acids (6:2 Page 6 of 27


Page 7 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


128

FTUA and 8:2 FTUA), 8:2 fluorotelomer sulfonate (8:2 FTSA) and perfluorooctane

129

sulfonamido acetic acid (FOSAA) were obtained from Wellington Laboratories (Guelph, ON,

130

Canada). N-ethyl perfluorooctane sulfonamide (EtFOSA) and perfluorooctane sulfonamide

131

(FOSA) were purchased from Advanced Technology & Industrial Co (Hong Kong, China).

132

Other fluorotelomer sulfonates (4:2 FTSA, 6:2 FTSA) and fluorotelomer carboxylic acids

133

(5:3 FTCA and 7:3 FTCA) were obtained from Synquest Laboratories (Alachua, FL, USA).

134

PFOSAm, PFOSAmS, PFOAAmS, PFOSNO, PFOANO, PFOSB and PFOAB were custom-

135

synthesized at the Beijing Surfactant Institute (Beijing, China); these surfactants are either

136

PFOS or PFOA derivatives, with different terminal functionalities: tertiary amine, ammonium

137

salt, amine oxide or betaine. The difference in functionalities allows them to be used as

138

model compounds to study their extraction from soil matrices. A similar approach was used

139

by Munoz et al., who used these compounds to develop an analytical method to analyze

140

them in sediment matrices.29 6:2 FTAB and 6:2 FTNO were obtained from Shanghai

141

Kingpont Industrial Company, Ltd (Shanghai, China). Isotope-labelled internal standards

142

(ISs) including MPFBA, MPFHxA, MPFOA, MPFNA, MPFDA, MPFUdA, MPFDoA,

143

MPFHxS, MPFOS, d-EtFOSA-M, M6:2 FTUA, M8:2 FTUA, M6:2 FTSA, M8:2 FTSA were

144

all obtained from Wellington Laboratories (Guelph, ON, Canada). Further details on the

145

structure and nomenclature of native analytes and isotope-labelled internal standards can

146

be found in the Supplemental Information (SI). HPLC-grade solvents including acetonitrile

147

(ACN), methanol (MeOH), LC/MS-grade water and acetic acid (HAc), as well as Optima-

148

grade ammonium hydroxide (NH4OH), certified sodium hydroxide (10 N) and certified

149

hydrochloric acid (8 N) were purchased from Fisher Scientific (Ottawa, ON, Canada).

150 Page 7 of 27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 32

151 152

Figure 1. Model native PFAS analytes used in the various recovery experiments (n refers to

153

the number of perfluorinated carbon atoms).

154 155

Soil and oil samples

156

Three textural types of soils were used to examine the analytical method performance:

157

sandy loam, clay loam and loam. Details on soil physical quality parameters, including

158

sand, silt and clay percentages and OM content, can be found in Table SI-2 of the

159

Supporting Information (SI). Two types of petroleum products were considered for possible

160

effect of co-contaminants on the extraction method performance: diesel oil and weathered

161

Bakken crude oil. The diesel oil was obtained from a local gas station and the weathered

162

Bakken crude oil was provided by Centre d'expertise en analyse environnementale du

163

Québec (Laval, QC, Canada). The extraction method was tested in each soil type as Page 8 of 27


Page 9 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


164

received, as well as in soils supplemented with co-contaminant -either diesel or Bakken oil-

165

at two concentration levels (5,000 and 40,000 mg/kg). The oil concentration was selected to

166

represent the medium and high levels of total petroleum hydrocarbons found in

167

contaminated soil available for analysis. Soil supplemented with hydrocarbon co-

168

contaminant was left to age for 72 hours at 4°C; a longer aging period was avoided to

169

minimize hydrocarbon biodegradation.

170 171

Extraction and sample preparation

172

The extraction procedure was adapted from previously published methods and two

173

extraction solvents were tested.20, 27-28, 30 Briefly, two grams of soil were placed in a 15mL

174

polypropylene centrifuge tube and 2.5 mL of extraction solvent, either Methanol + 0.1%

175

NH4OH or Methanol + NaOH 200 mM, were added. The mixture was sonicated for 30 min

176

and shaken in a rotational shaker for 60 min, it was then centrifuged for 15 min at 4,000 g

177

and the supernatant was retrieved. The procedure was repeated twice more. In the case of

178

the MeOH/NaOH extraction solvent, an equivalent amount of HCl was used to neutralize

179

the extracts. The combined extracts were concentrated to 2 mL under a gentle stream of

180

nitrogen at 45°C. The concentrated extract was cleaned with ENVI-Carb graphite (Sigma-

181

Aldrich Canada) by combining the graphite powder with the methanolic extract, vortexing

182

for one minute and centrifuging for 10 min at 20,000 g. The supernatant was subsequently

183

recovered; internal standards were added to the samples for a final concentration of 2

184

ng/mL. Procedural blanks -consisting of clean solvent which underwent the same extraction

185

procedure as the soil- were prepared along with each set of samples extracted; PFAS

186

remained below detectable levels in procedural blanks, or at low and reproducible levels. Page 9 of 27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 32

187 188

Instrumental analysis

189

Quantitative analysis was performed with a Shimadzu UHPLC system coupled to an AB

190

Sciex 5500 Qtrap mass spectrometer working in multiple reaction monitoring (MRM) mode,

191

with positive and negative electrospray ionization. Separation was achieved using an

192

Agilent Zorbax SB-C8 column (3.5 µm, 100 x 2.1 mm). A Kinetex EVO C18 column (5 µm,

193

50 x 3.1 mm) was placed upstream of the UHPLC autosampler. Details on LC-MS/MS

194

monitored transitions as well as other instrumental conditions are provided in Tables SI-3

195

and SI-4.

196 197

Assessment of the matrix effect

198

The “raw” matrix effect was first analyzed by comparing the absolute responses of isotope-

199

labelled internal standards in each matrix to those in the matrix-free solvent reference. For

200

each combination of soil type and hydrocarbon co-contaminant, the matrix was subjected to

201

the extraction procedure. Native PFAS analytes and isotope-labelled ISs were spiked post-

202

extraction to create matrix-matched calibration curves. The slopes of the resulting

203

calibration curves were then compared to those prepared in clean solvent to assess the

204

matrix effects at the instrumental stage.

205 206 207

Determination of the recovery fraction

208

For each combination of matrix type, oil type and oil concentration, native PFASs were

209

spiked to the soil samples at a concentration of 10 ng/g and after two hours they were Page 10 of 27


Page 11 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


210

subjected to the aforementioned extraction and clean-up procedure (see Extraction and

211

Sample Preparation). In parallel, soil samples not initially fortified with PFAS were also

212

extracted in the same fashion, and the resulting soil extracts were spiked post-sample

213

preparation with an equivalent amount of native PFAS. In both cases, preparations were

214

performed in triplicate, and isotope-labelled ISs were added at the end of the preparation

215

procedure.31 The recovery fraction was then expressed as the concentration determined in

216

the soil extract spiked before extraction, divided by that in the corresponding soil extract

217

spiked at the end of the preparation procedure. In order to evaluate whole-method recovery

218

rates, soil samples fortified at the start of the preparation procedure were compared with

219

corresponding matrix-matched extracts supplemented with PFAS post-extraction and clean-

220

up, thereby eliminating possible bias due to matrix effects.

221 222

Statistical treatment of data

223

Statistical analyses were performed on the R statistical software (R Core Team, 2016).32

224

Statistical significance was set at p < 0.05. The raw matrix effect was assessed by

225

comparing the absolute response of each internal standard in each of the matrices (n=9) by

226

a Tukey’s Honestly Significant Difference (HSD) test. A one-way ANOVA test was used to

227

qualitatively identify treatments for which the recovery of individual PFASs spiked in soil

228

was not complete, i.e. where there is a significant difference between the recovered analyte

229

and the matrix-matched reference. A linear model was later applied to estimate the

230

parameters (namely, clay content, OM content, oil type and oil concentration) that may

231

result in an incomplete recovery of the targeted analytes.

232 Page 11 of 27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 32

233

Application to the analysis of contaminated field samples

234

The MeOH/NH4OH extraction method was used to analyze six soil samples collected from

235

a site in Québec, Canada, which was contaminated with both hydrocarbon and AFFFs and

236

whose soil characteristics were similar to the sandy loam model soil.

237

Results and Discussion

238 239

Instrumental method performance

240

LC-MS/MS analytical method. The herein described analytical method performed

241

satisfactorily with chromatographic conditions providing suitable retention and separation of

242

the 36 target analytes. Notably, positive mode and negative mode analytes were

243

simultaneously acquired over a single 10-min analytical run, all the while maintaining

244

excellent instrumental detection limits, linearity, accuracy and precision. The limit of

245

detection (LOD) was defined as the smallest concentration that would yield a detectable

246

chromatographic peak with a signal to noise ratio S/N > 3.33 Instrumental detection limits

247

ranged from 0.005–2.0 ng/mL. Details on compound-specific LODs are provided in the

248

supplemental information. All points in calibration curves presented accuracy between 80

249

and 120%. Intra-day precision –relative standard deviation on replicate analysis (n=5)– was

250

between 1.2–13% while inter-day precision –relative standard deviation on replicate

251

analysis (n=5) over three different days– was between 1.9–17% (RSD), depending on the

252

particular analyte.

253 254

Instrumental matrix effect and use of internal standards. The observed matrix effect

255

differed significantly between the two extraction methods assayed. To reduce the matrix Page 12 of 27


Page 13 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


256

effect, the extracts were cleaned with ENVI-Carb before analysis (25mg/mL and 75mg/mL

257

for the MeOH/NH4OH and MeOH/NaOH extraction solvent, respectively).

258 259

For the extraction method with ammonium hydroxide, some enhancement of the absolute

260

signal of the internal standard was detected, especially in the soils that were added with the

261

crude oil. Table SI-6 presents a summary of the matrices and internal standards,

262

highlighting in dark color the cases for which a significant difference in the internal standard

263

absolute response was observed. The increase in all cases was not larger than 30%. Even

264

though the absolute response increased as a result of the matrix, isotope-labelled internal

265

standards were found to adequately correct the native analyte instrumental matrix effect for

266

the MeOH/NH4OH extracts.

267 268

In the case of the MeOH/NaOH extraction method, however, the magnitude of the

269

instrumental matrix effect was considerably larger. The internal standard absolute response

270

was significantly lower in all matrices when compared to that in the matrix-free solvent, with

271

the signal decreasing between 16 and 78%, as detailed in Table SI-7. The use of isotope-

272

labelled ISs was not sufficient to compensate the instrumental matrix effects for all target

273

native analytes, as the slopes of the matrix-matched calibration curves differed significantly

274

from that in the matrix-free solvent. For both extraction methods, there was no significant

275

difference in the absolute IS signal if the native PFAS were spiked before or after

276

extraction, meaning that the matrix effect was not exacerbated by the fact that there were

277

fluorinated surfactants present in the soil.

278 Page 13 of 27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 32

279

It is noteworthy that the effect on the calibration curve is accentuated when the structure of

280

the isotope-labelled IS differs significantly from the structure of the native analyte, such as

281

for newly-identified surfactants. In the latter case, the internal standard selected was that of

282

the corresponding acid from which they are derived (e.g. MPFOS was used for PFOSB,

283

MPFOA for PFOAB, and M6:2 FTSA for 6:2 FTAB). Such internal standards may not

284

effectively correct for instrumental matrix effects such as those occurring at the electrospray

285

ionization stage, and would rather only compensate for small variations in the injection

286

volume through isotopic dilution.

287 288

Extraction method performance

289

The performance of the extraction using Methanol/NH4OH was evaluated by comparing the

290

recovery fraction across different treatments for each compound. The Tukey HSD test was

291

applied to compare each combination with its reference (see Table SI-8 of the SI). An

292

illustration for the case of the loam soil is provided in Figure 2. The recovery fraction of

293

PFSAs, PFCAs and FTSAs in loam soil fell within the 70–120% range, regardless of the

294

presence or absence of the two types of oil co-contaminants and their concentration.

295

Similar results were obtained for PFSAs, PFCAs and FTSAs in the case of the sandy loam

296

and the clay loam soils (SI Fig SI-1 and SI-2), indicating the robustness of the present

297

method when applied to soils of variable textural classes and petroleum co-contamination

298

burden.

299 300

The new surfactants, however, differed from these trends and tended to present

301

significantly lower recovery fractions with the MeOH/NH4OH method across some of the Page 14 of 27


Page 15 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


302

treatments. For instance, the recovery fraction of betaine-based PFAS (namely, PFOSB,

303

PFOAB and 6:2 FTAB) was in the 30–60 % range across treatments in the sandy loam and

304

the loam soils, with some fluctuations with oil type or oil concentration, while even lower

305

recovery rates were observed in the clay loam soil regardless of the hydrocarbon co-

306

contamination (see Figure SI-2). Similar results were observed for other positively-charged

307

compounds such as quaternary amine-based compounds (PFOAAmS, and, to a lesser

308

extent, PFOSAmS). Based on these results, it can be inferred that clay would provide a

309

significant amount of negative charges, which could interact with positively charged

310

compounds and result in lower extraction recovery rates. The recovery of other analytes,

311

such as FTUAs, FTCAs, and FOSAA tended to be lower than those of PFAAs or FTSAs,

312

particularly for the soil with the highest clay content (Figure SI-2). Equally noteworthy, in the

313

two soils with the lower organic carbon content (clay loam and sandy loam, 2.9–3.1% OM

314

respectively), the recovery rates of novel surfactants such as tertiary amine (PFOSAm) and

315

amine-oxide based PFASs (PFOSNO, PFOANO and 6:2-FTNO) displayed acceptable

316

recovery rates across treatments (range = 70–120 %; see SI Figure SI-1 and SI-2). The

317

recovery for these surfactants dropped to the 30–60 % range in the soil with the highest

318

organic carbon content (12.6 %), especially when petroleum co-contaminants were

319

supplemented (Figure 2), suggesting possible hydrophobic interactions.

320 321

The factors affecting the extraction efficiency of those analytes that displayed recoveries

322

lower than 70% were evaluated as described in the section Statistical treatment of data.

323

The factors analyzed were soil type (by comparing clay content and OM content), oil type

324

(diesel or crude oil) and oil concentration (40,000 mg/kg, 5,000 mg/kg, and no oil added). A Page 15 of 27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 32

325

summary of the factors significantly affecting the recovery fraction is provided in Table SI-9

326

of the SI. It can be concluded therefrom that the most critical underlying factor affecting the

327

PFAS recovery performance was the soil type, specifically soil texture. It was found that

328

both OM content (significant for 11/13 compounds) and clay content (significant for 12/13

329

compounds) have strong influence on the recovery. The type of oil (8/13) and its

330

concentration (4/13) were also significant in some instances, mostly for quaternary or

331

tertiary amine-based fluorosurfactants –PFOSAm, PFOSAmS and PFOAAmS–, albeit to a

332

lesser extent than soil type.

333 334

Figures 2, SI-1 and SI-2 show the recovery fractions of all targeted PFAS in loam, sandy

335

loam and clay loam, respectively, in the presence and absence of supplemented

336

hydrocarbon co-contaminants and using the MeOH/NH4OH-based extraction method. As

337

can be observed in such figures, for each compound the recovery fraction across all

338

conditions within a certain soil type remained within a rather limited range of values

339

regardless of oil type and concentration, consistent with the above-mentioned observation

340

that soil type is a significant controlling factor of the recovery.

341

Page 16 of 27


Page 17 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


342 343

Figure 2. Recovery fraction of all studied PFAS in the loam soil with Methanol/NH4OH as

344

extraction solvent in the presence and absence of hydrocarbon co-contaminants. Circles

345

refer to diesel as a co-contaminant, triangles refer to Bakken crude oil as a co-contaminant

346

and squares represent the recovery in soil not supplemented with hydrocarbon. Note that

347

the first 21 compounds from left to right (from PFBS to EtFOSA), as well as both n:2 FTUAs,

348

have either matched or closely matched (different chain length) isotope-labelled internal

349

standards, which could be a factor in better method performance.

350

Page 17 of 27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 32

351 352

Figure 3. Recovery fraction of model new surfactants in sandy loam soil in the presence and

353

absence of hydrocarbon co-contaminants. Circles refer to diesel as a co-contaminant,

354

triangles refer to crude oil (Bakken oil) as a co-contaminant and squares represent the

355

recovery in soil not supplemented with hydrocarbon. Black symbols represent extraction

356

with Methanol/NH4OH as extraction solvent while red symbols refer to extraction with

357

Methanol/NaOH as extraction solvent.

358

Also from Figures 2, SI-1 and SI-2, it can be observed that the recovery of the newly

359

identified PFAS, as well as fluorotelomer unsaturated acids (FTUAs), fluorotelomer acids

360

(FTCAs), perfluoroctanesulfonamide (FOSA) and perfluorooctanesulfonamido acetic acid

361

(FOSAA), tends to be lower, particularly for the soil with a high clay content (Figure SI-2).

362

To further improve the recovery of such analytes, the extraction was also performed with a

363

stronger extraction solvent (i.e., Methanol/NaOH 200mM). Given that the previous analysis

364

showed that the most important factor affecting analyte recovery was the soil type, and to a Page 18 of 27


Page 19 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


365

lesser extent, the oil type, the extraction was performed with the three types of soil, at one

366

co-contaminant concentration only (40,000 ppm).

367 368

Figure 3, as well as figures SI-3 and SI-4, presents the comparative data of extraction

369

efficiencies using both solvents. The extraction efficiency for PFAAs, FTSAs, EtFOSA and

370

FOSA is not shown in such figures since no significant differences arose in terms of

371

recovery efficiency, although the MeOH/NaOH solvent generally presented higher RSDs.

372

Only the compounds for which the extraction was below 70% with the Methanol/NH4OH

373

extraction solvent are therefore depicted in Figures 3, SI-3 and SI-4. The recovery fraction

374

generally increased for these analytes when using the MeOH/NaOH method. One

375

particularly salient example is the case of betaines in the clay loam soil for which recovery

376

rates were improved from ~5–10 % with MeOH/NH4OH (irrespective of the oil condition) to

377

40–60 % or higher with MeOH/NaOH. There are, however, some analytes for which the

378

recovery fraction in fact decreased. Such is the case of FTUAs in the clay loam soil (Figure

379

SI-3) and FTUAs and amine oxides in the sandy loam soil (Figure 3).

380 381

As previously discussed in the section on Instrumental method performance, the method

382

based on the Methanol/NaOH extraction solvent presented a much stronger matrix effect.

383

This matrix effect made it necessary to further clean the extracts, and signalled to the

384

importance of matrix-matched calibration. Even with these precautions, it was observed that

385

the analytical precision was less favorable than the Methanol/NH4OH-based method.

386

Furthermore, analyte detection could be hampered at lower concentrations, since the

Page 19 of 27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

387

instrumental signal suppression also translated into an increase in instrumental detection

388

limits by a factor of ~10-50.

Page 20 of 32

389 390

Analysis of PFAS in AFFF-impacted soil samples co-contaminated with petroleum

391

hydrocarbons

392

The analytical method using Methanol/NH4OH as the extraction solvent and subsequent

393

ENVI-Carb graphite clean-up was applied to analyze soil samples collected from Quebec,

394

Canada. Due to the high concentration of certain PFAS in the samples, a 1:50 dilution was

395

applied to adequately quantify those compounds exceeding the upper limit of quantification

396

(uLOQ). Such dilution would considerably reduce the instrumental matrix effect if it was

397

present.

398 399

Figure 4. PFAS concentration in the AFFF-impacted soil samples. Stacked bars represent the

400

concentration of the different analyzed PFAS (left axis), while the solid blue bars represent

Page 20 of 27


Page 21 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


401

the total C10-C50 hydrocarbon concentration (mg/kg) measured in each sample (right axis)

402

according to the reference method from the Center of Expertise in Environmental Analysis of

403

Quebec.34

404 405

The sum of the concentrations of target PFASs (ΣPFAS) varied more than 70x between soil

406

samples (~60 to ~4,400 ng/g). The analytes that presented the highest concentrations are

407

6:2 FTAB, 6:2 FTSA and 8:2 FTSA. 6:2 FTAB comprised at least 45% and up to 74% of the

408

ΣPFAS in this set of samples. The sum of fluorotelomer sulfonates accounted for 7.1 to

409

42% of ΣPFAS, making them the second most abundant PFAS class detected. The sum of

410

all perfluoroalkyl carboxylates also represented a significant proportion of the PFAS

411

composition profiles (3.9-24% of ΣPFAS). The extent of total PFAS contamination is likely

412

to be considerably higher than presented in Figure 4, since this study only focuses on the

413

target compounds for which an authentic standard available could be used for quantitation,

414

therefore neglecting the possible presence of many other analytes.16-17 The implications of

415

such a significant proportion of a newly identified analyte, such as 6:2 FTAB, are of great

416

importance: not only could the extent of PFAS contamination be severely underestimated

417

when quantifying only PFAAs and legacy compounds, but the potential biotransformation of

418

partially fluorinated surfactants could also lead to significantly higher concentrations of

419

PFAAs overtime.23, 35 The analysis of this set of samples further illustrates the fact that

420

measuring only the concentration of PFOS and PFOA, or even a suite of PFCAs and

421

PFSAs of different chain lengths, would be insufficient when attempting to evaluate the

422

extent of the PFAS contamination at an AFFF-impacted site.

423

Page 21 of 27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 32

Conclusion

424 425

The performance of two extraction methods for PFAS in soil in the presence of crude oil or

426

diesel as a co-contaminant was evaluated. The MeOH/NH4OH extraction method yielded

427

excellent results for all PFAAs and FTSAs, under all soils and co-contaminants conditions,

428

yet could not recover betaine-based or quaternary amine-based PFAS as efficiently. The

429

factors affecting the recovery fraction of each compound were statistically analyzed. It was

430

evident that while the presence or absence of oil may somehow affect the recovery, the

431

most important controlling factor was soil type. It is hypothesized that both OM and clay can

432

affect the recovery through strong interactions with PFAS; while OM can be a source of

433

hydrophobic interactions, clay can provide a significant amount of negative charges, which

434

potentially interact with positively charged compounds (ammonium salts, betaines, and

435

amines and amine oxides at low pH).

436 437

The recovery for the newly identified PFAS was generally not complete with the

438

MeOH/NH4OH solvent. For this reason, the extraction was also assayed using a stronger

439

extraction solvent (MeOH/NaOH). Even though the recovery fraction of most of the

440

compounds that showed a lower recovery generally improved with the stronger solvent, this

441

method presented its own limitations. Notably, matrix effects were more prevalent, making it

442

compulsory to use a matrix-matched calibration approach to compensate. Additionally, the

443

limits of detection increased considerably and the accuracy was lower, especially for those

444

compounds for which no matched isotope labelled internal standard was available.

445

Page 22 of 27


Page 23 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


446

While both methods yielded satisfactory results overall, especially for PFAAs or FTSAs, the

447

extraction approach using a milder solvent has the advantage of providing excellent limits of

448

detection and moderate matrix effects. Meanwhile, the stronger extraction method yielded

449

better recovery rates for novel PFAS, yet led to higher limits of detection and lower

450

instrumental accuracy. Both methods could potentially underestimate the concentration of

451

many other AFFFs components.

452 453

The focus of the present work was on whether currently used extraction methods (such as

454

those with basic solvent) would be transferable to the analysis of an extended suite of

455

PFAS in co-contaminated soil. It is, however, possible that acidic conditions could provide

456

different results. A lower pH could result in decreased interaction of zwitterionic and cationic

457

compounds with soil, possibly aiding their extraction; it could also affect long-chain PFAAs

458

through increased sorption of their neutral fraction. These are both possibilities that are

459

worth investigating in future work for the optimization of a comprehensive method for both

460

positively and negatively charged fluorosurfactants.

461 462

It has been previously highlighted that a major challenge when managing and remediating

463

an impacted site is the characterization of the chemical contamination. The examination of

464

the extraction recovery under different soil and oil type combinations pinpointed how

465

common AFFF components can be underestimated by current PFAS analytical methods.

466

The authors also reiterate the need for more authentic PFAS standards and isotope-

467

labelled internal standards for a more accurate characterization of the compounds

Page 23 of 27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

468

commonly occurring at AFFF-impacted sites and encourage study-specific method

469

performance evaluation whenever dealing with matrices of unknown behaviours.

Page 24 of 32

470 471

Supporting Information. Details on soil used for analysis, instrumental methods, raw

472

matrix effect, summary of recovery results across different conditions, factors influencing

473

recovery.

474 475

Acknowledgments

476

The authors acknowledge the funding support by Canada Foundation for Innovation and

477

Fonds de recherche du Québec – Nature et technologies (FQRNT) Team Grant, and McGill

478

MEDA Scholarship and J.W McConnell Memorial Fellowship awarded to S. Mejia

479

Avendaño. The authors also acknowledge the support of Mélanie Desrosiers who provided

480

the contaminated soils, and the Centre d’expertise en analyse environnementale du

481

Québec (CEAEQ) for their support to obtain the Bakken crude oil.

Page 24 of 27


Page 25 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


482

References

483 484 485

1. M. K. So, N. Yamashita, S. Taniyasu, Q. Jiang, J. P. Giesy, K. Chen, P. K. S. Lam, Environmental science & technology 2006, 40. 2924-2929.

486

2.

487 488

3. K. Steenland, T. Fletcher, D. A. Savitz, Environmental Health Perspectives 2010, 118. 11001108, DOI: 10.2307/27822994.

489 490

4. M. E. Andersen, J. L. Butenhoff, S.-C. Chang, D. G. Farrar, G. L. Kennedy, C. Lau, G. W. Olsen, J. Seed, K. B. Wallace, Toxicological sciences 2008, 102. 3-14.

491 492

5. B. J. Apelberg, F. R. Witter, J. B. Herbstman, A. M. Calafat, R. U. Halden, L. L. Needham, L. R. Goldman, Environmental health perspectives 2007. 1670-1676.

493

6.

494 495

7. Canada Gazette Part II Vol. 142 No. 12, in Perfluorooctane sulfonate and its salts and certain other compounds regulations, Registration SOR/2008-178. Ottawa, 2008.

496

8.

497 498

9. KEMI Occurrence and use of highly fluorinated substances and alternatives; 361 164; Swedish Chemicals Agency: Stockholm, 2015.

499

10.

500 501

11. H. Ranjbar, B. H. Shahraki, Chemical Engineering & Technology 2013, 36. 295-299, DOI: 10.1002/ceat.201200401.

502 503

12. Y. Wei, W. Peng-xiang, T. Liang, F. Xue-cheng, B. Zhi-ming, Fire Science and Technology 2010, 6. 021.

504 505 506

13. UNEP Technical paper on the identification and assessment of alternatives to the use of perfluorooctane sulfonic acid in open applications; UNEP/POPS/POPRC.8/INF/17; Stockholm Convention on Persistent Organic Pollutants: Geneva, 2012.

507 508

14. Danish Ministry of the Environment Short-chain polyfluoroalkyl Substances (PFAS), Environmental project No. 1707, 2015; Environmental Protection Agency: 2015.

509 510

15. B. J. Place, J. A. Field, Environmental Science & Technology 2012, 46. 7120-7127, DOI: 10.1021/es301465n.

511 512

16. W. J. Backe, T. C. Day, J. A. Field, Environmental Science & Technology 2013, 47. 52265234, DOI: 10.1021/es3034999.

J. L. Domingo, Environment international 2012, 40. 187-195.

USEPA, in PFOA Stewardship Program, Docket ID: EPA-HQ-OPPT-2006-0621. 2006.

Stockholm Convention on Persistent Organic Pollutants Annex B, decision SC-4/17; 2009.

A. R. Sontake, S. M.Wagh, Chemical Engineering and Science 2014, 2. 11-14.

Page 25 of 27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 32

513 514

17. L. A. D’Agostino, S. A. Mabury, Environmental Science & Technology 2013, 48. 121-129, DOI: 10.1021/es403729e.

515 516

18. B. Weiner, L. W. Y. Yeung, E. B. Marchington, L. A. D’Agostino, S. A. Mabury, Environmental Chemistry 2013, 10. 486-493, DOI: http://dx.doi.org/10.1071/EN13128.

517 518

19. K. C. Harding-Marjanovic, E. F. Houtz, S. Yi, J. A. Field, D. L. Sedlak, L. Alvarez-Cohen, Environmental Science & Technology 2015, 49. 7666-7674, DOI: 10.1021/acs.est.5b01219.

519 520

20. S. Mejia-Avendaño, S. Vo Duy, S. Sauvé, J. Liu, Environmental Science & Technology 2016, 50. 9923-9932, DOI: 10.1021/acs.est.6b00140.

521 522

21. C. Eschauzier, K. J. Raat, P. J. Stuyfzand, P. De Voogt, Science of The Total Environment 2013, 458–460. 477-485, DOI: http://dx.doi.org/10.1016/j.scitotenv.2013.04.066.

523 524

22. L. Vierke, A. Möller, S. Klitzke, Environmental Pollution 2014, 186. 7-13, DOI: http://dx.doi.org/10.1016/j.envpol.2013.11.011.

525 526

23. J. Liu, S. Mejia Avendaño, Environment International 2013, 61. 98-114, DOI: http://dx.doi.org/10.1016/j.envint.2013.08.022.

527 528

24. C. R. Powley, S. W. George, T. W. Ryan, R. C. Buck, Analytical Chemistry 2005, 77. 63536358, DOI: 10.1021/ac0508090.

529 530

25. C. P. Higgins, J. A. Field, C. S. Criddle, R. G. Luthy, Environmental Science & Technology 2005, 39. 3946-3956.

531 532

26. J. W. Washington, W. M. Henderson, J. J. Ellington, T. M. Jenkins, J. J. Evans, Journal of Chromatography A 2008, 1181. 21-32, DOI: 10.1016/j.chroma.2007.12.042.

533 534

27. C. Liu, J. Liu, Environmental Pollution 2016, 212. 230-237, DOI: 10.1016/j.envpol.2016.01.069.

535 536

28. E. F. Houtz, C. P. Higgins, J. A. Field, D. L. Sedlak, Environmental Science & Technology 2013, 47. 8187-8195, DOI: 10.1021/es4018877.

537 538

29. G. Munoz, S. V. Duy, P. Labadie, F. Botta, H. Budzinski, F. Lestremau, J. Liu, S. Sauvé, Talanta 2016, 152. 447-456, DOI: http://dx.doi.org/10.1016/j.talanta.2016.02.021.

539 540

30. S. Mejia Avendaño, J. Liu, Chemosphere 2015, 119. 1084-1090, DOI: http://dx.doi.org/10.1016/j.chemosphere.2014.09.059.

541 542

31. G. Munoz, S. Vo Duy, H. Budzinski, P. Labadie, J. Liu, S. Sauvé, Analytica chimica acta 2015, 881. 98-106, DOI: http://dx.doi.org/10.1016/j.aca.2015.04.015.

543

32.

R Core Team. Vienna, Austria, 2016.

Page 26 of 27


Page 27 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


544 545

33. P. Araujo, Journal of Chromatography B 2009, 877. 2224-2234, DOI: http://dx.doi.org/10.1016/j.jchromb.2008.09.030.

546 547 548

34. Centre d'expertise en analyse environnementale du Québec, in MA. 400 - HYD. 1.1. Rev. 3. Ministère du Développement durable, de l;Environnement, et Lutte contre les changements climatiques du Québec, 2016, p 17.

549 550 551 552

35. M. K. Moe, S. Huber, J. Svenson, A. Hagenaars, M. Pabon, M. Trümper, U. Berger, D. Knapen, D. Herzke, Chemosphere 2012, 89. 869-875, DOI: http://dx.doi.org/10.1016/j.chemosphere.2012.05.012.

Page 27 of 27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

TOC Abstract 262x169mm (150 x 150 DPI)


Page 28 of 32

Page 29 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 1. Model native PFAS analytes used in the various recovery experiments (n refers to the number of perfluorinated carbon atoms) Figure 1 177x122mm (300 x 300 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 2. Recovery fraction of all studied PFAS in the loam soil with Methanol/NH4OH as extraction solvent in the presence and absence of hydrocarbon co-contaminants. Circles refer to diesel as a co-contaminant, triangles refer to Bakken crude oil as a co-contaminant and squares represent the recovery in soil not supplemented with hydrocarbon. Note that the first 21 compounds from left to right (from PFBS to EtFOSA), as well as both n:2 FTUAs, have either matched or closely matched (different chain length) isotope-labelled internal standards, which could be a factor in better method performance. Figure 2 141x95mm (300 x 300 DPI)


Page 30 of 32

Page 31 of 32

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. Recovery fraction of model new surfactants in sandy loam soil in the presence and absence of hydrocarbon co-contaminants. Circles refer to diesel as a co-contaminant, triangles refer to crude oil (Bakken oil) as a co-contaminant and squares represent the recovery in soil not supplemented with hydrocarbon. Black symbols represent extraction with Methanol/NH4OH as extraction solvent while red symbols refer to extraction with Methanol/NaOH as extraction solvent Figure 3 141x199mm (300 x 300 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 4. PFAS concentration in the AFFF-impacted soil samples. Stacked bars represent the concentration of the different analyzed PFAS (left axis), while the solid blue bars represent the total C10-C50 hydrocarbon concentration (mg/kg) measured in each sample (right axis) according to the reference method from the Center of Expertise in Environmental Analysis of Quebec.34 Figure 4 113x126mm (300 x 300 DPI)


Page 32 of 32

Assessment of the Influence of Soil Characteristics and Hydrocarbon

Recommend Documents