Identification and quantification of linear and branched isomers of perfluorooctanoic and perfluorooctane sulfonic acids in contaminated groundwater in the veneto region

G Model

ARTICLE IN PRESS

CHROMA-359092; No. of Pages 12

Journal of Chromatography A, xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Identification and quantification of linear and branched isomers of perfluorooctanoic and perfluorooctane sulfonic acids in contaminated groundwater in the veneto region Alessandro Pellizzaro a,∗ , Alessandro Zaggia b , Massimo Fant a , Lino Conte b , Luigi Falletti b a b

Acque del Chiampo S.p.A. – Servizio Idrico Integrato, Via Ferraretta 20, 36071 Vicenza, Italy Department of Industrial Engineering, University of Padua, Via Marzolo 9, 35030 Padua, Italy

a r t i c l e

i n f o

Article history: Received 21 July 2017 Received in revised form 13 December 2017 Accepted 13 December 2017 Available online xxx Keywords: PFAAs Mass spectrometry Linear and branched isomers Isomer profile

a b s t r a c t Perfluoroalkylated acids (PFAAs) are ubiquitous xenobiotic substances characterized by high persistency, bioaccumulation potential and toxicity. They have generated global concern because of their widespread presence both in water and biota compartments. In the past four years, alarming levels of these pollutants have been found in both surface and groundwater collected in an area covering more than 150 square kilometers in the south-western part of the province of Vicenza (Veneto region, Italy). One of the sources of the contamination recognized by local authorities is a fluorochemicals production plant that produced PFAAs since late sixties by electrochemical fluorination involving the obtainment of a complex mixture of linear and branched isomers. Branched isomers account for a significant part of total long chain homologues (22%–35%). Because of the potential threat to public health and the absence of specific limits set for these pollutants by Directive 98/83/EC, local authorities have established the following performance limits for drinking water: 90 ng L−1 for PFOA + PFOS, (reduced to 40 ng L−1 in the most contaminated municipalities), 30 ng L−1 for PFOS and 300 ng L−1 for the sum of all other PFAAs. Given the non-negligible incidence of branched isomers, it appears very important to correctly identify and quantify their contribution to total PFAAs. A liquid chromatography-electrospray ionization tandem spectrometry LC–MS/MS method, coupled with solid phase extraction, was developed to identify and quantify 25 PFAAs including six branched isomers of PFOS and four branched isomers of PFOA. Expanded uncertainty, recovery and precision were determined and found to agree with the reference EPA method 537:2009. The quantification limit is comprised in the 1–5 ng L−1 range. © 2017 Elsevier B.V. All rights reserved.

1. Introduction In the past four years, high levels of PFAAs have been detected in both surface and underground water sampled in an area covering approximately 150 square kilometers in the south-western part of the province of Vicenza (Veneto region, Italy) [1]. Because of the potential threat to public health and the absence of specific limits set for these pollutants by Directive 98/83/EC, local Authorities have established the following performance limits for drinking water: 90 ng L−1 for PFOA + PFOS, (reduced to 40 ng L−1 in the most contaminated municipalities), 30 ng L−1 for PFOS, and 300 ng L−1 for the sum of all other PFAAs.

∗ Corresponding author. E-mail address: [email protected] (A. Pellizzaro).

Surveys conducted by competent authorities identified one of the sources of contamination in a fluorochemical production plant that manufactured PFAAs since late sixties by electrochemical fluorination (ECF). Further investigations on both surface and groundwater sampled in the polluted area revealed the presence of significant quantities of branched and linear isomers for both PFOA and PFOS. The PFAA production process must be considered when seeking to explain the presence of not negligible quantities of isomers. Electrochemical fluorination and telomerization are the major manufacturing methods used to produce both long and short chain PFAAs [2]. While telomerization produces primarily straight chain PFAAs, electrochemical fluorination produces a complex mixture of linear and branched isomers. Electrochemical fluorination is an electrochemical reaction that replaces hydrogen atoms with fluorine atoms in an organic substrate (usually a fluoride or a chloride of an organic acid) dissolved

https://doi.org/10.1016/j.chroma.2017.12.036 0021-9673/© 2017 Elsevier B.V. All rights reserved.

Please cite this article in press as: A. Pellizzaro, et al., Identification and quantification of linear and branched isomers of perfluorooctanoic and perfluorooctane sulfonic acids in contaminated groundwater in the veneto region, J. Chromatogr. A (2017), https://doi.org/10.1016/j.chroma.2017.12.036

G Model CHROMA-359092; No. of Pages 12 2

ARTICLE IN PRESS A. Pellizzaro et al. / J. Chromatogr. A xxx (2017) xxx–xxx

in anhydrous hydrogen fluoride. According to the radical-cation mechanism proposed by Rozhkov [3,4], the formation of stable secondary carbocation intermediates leads to fragmentation, rearrangement and finally isomerization of the carbon skeleton resulting in a complex mixture of linear and branched isomers [5]. As a general rule, increasing the number of carbon atoms in the fluorinated chain increases the number of branched isomers. A study by Vyas [6] using 19 F NMR spectroscopy on PFBS (perfluorobutane sulfonyl) and PFOS derivatives obtained by electrochemical fluorination showed that while commercial samples of PFBS contained no detectable isomers, PFOS samples have a complex isomer profile. Increasing the length of the perfluorinated carbon chain beyond C3 leads to a corresponding rapid increase in the number of potential isomers reaching 89 congeners theoretically possible for C8 [7]. Reagen [8] reported the isomeric profiles of commercial grade PFOA and PFOS obtained by electrochemical fluorination showing that branched isomers account for 22% of PFOA and 29.3% of PFOS content. Internal monomethyl substituted for dimethyl substituted isomers are predominant in technical grade PFAS. The average percentage of isomers reported by Vyas [6] for commercial grade potassium perfluorooctansulfonate is 16.3 ± 2.1 (internally mono-methyl branched PFOS), 2.6 ± 0.1 (␣-methyl branched PFOS), and 10.1 ± 0.4 (isopropyl branched PFOS). Jiang [5] reported the following isomer profile for technical grade PFOA: 11.5% internally mono-methyl branched PFOA, and 9% isopropyl branched PFOA. PFAS derived from electrochemical fluorination are present in the environment as a mixture of linear and branched isomers. The review of Benskin [9] on isomer profiling of PFAAs examined the large body of work demonstrating that branching patterns affect properties such as environmental transport and degradation, partitioning, bioaccumulation pharmacokinetics and toxicity. Branched PFOA isomers have been reported as concentrating in both soil and sediment in higher amounts than linear PFOA. This probably indicates that as PFOA contaminated sites age, branchedchain isomers tend to constitute progressively larger fractions of the total PFOA due to branched-chain sorption together with leaching of linear PFOA [10]. All major branched PFOS isomers are excreted via human urine preferentially over linear PFOS [11]. In the case of FOSA (perfluorooctanesulfonamide, a precursor of PFOS), studies have shown preferential excretion of linear FOSA compared to branched isomers [12]. Although branched isomer content in human serum varies widely between 17% and 52% of total PFOS, their percentage is generally higher than a typical ECF isomer pattern [13,14]. Peng [15] reported that the preferential metabolism of branched isomers of (N-ethyl perfluorooctanesulfonamido)-ethanol (FOSE) -based phosphate diester (diSPAP, a precursor of PFOS) led to the enrichment of branched PFOS in Japanese Medaka. Accurate isomer profiling of PFAS in water and other environmental matrices has the potential to gain information on the emission source. Jin [16] and Benskin [17] linked the presence of high concentrations of a single isopropyl branched PFPeA isomer in water and soil samples to a telomerization production process. Branching also plays a crucial role in bioaccumulation [18–22] and transport potential [23–25]. Tissue-specific isomer patterns found in polar bears suggest isomer-specific pharmacokinetics, probably due to differences in protein affinity [26]. Further, in the case of drinking water contamination, it is worth noting that in the presence of legal limits on PFAS concentration, erroneous quantification of isomers could lead to an improper judgment on water potability. Partitioning, bioaccumulation, metabolism, toxicity and regulatory limitations are not accounted for in many cases despite the significant role played by branching in PFAA environmental dispersion. The quantification of the PFAA load or analytical methods

developed for drinking water fail to address their contribution, leading to an incorrect estimation of total PFAA load. For the first time, a liquid chromatography-electrospray ionization tandem spectrometry LC–MS/MS method, coupled with a solid phase extraction, was used to identify and quantify 14 linear PFAAs, five branched perfluoro monomethyl substituted isomers of PFOS, three branched perfluoro monomethyl substituted isomers of PFOA, one branched perfluoro dimethyl substituted isomer of PFOS and one perfluoro dimethyl substituted isomer of PFOA. 2. Material and methods 2.1. Chemicals and consumables LC–MS grade methanol was purchased from Applichem (Darmstadt, Germany), water was produced by an Elga purification system from Veolia (Saint Maurice, France), ammonium acetate ® was provided by Carlo Erba (Milan, Italy) and Trizma hydrochloride was provided by Fischer-Scientific (Loughborough, UK). Weak anion exchange StrataTM X-AW SPE (200 mg/6 ml) cartridges were obtained from Phenomenex (Torrance, CA, USA) and Oasis WAX SPE (150 mg/6 ml) were purchased from Waters (Millford, MA, USA). 2.2. Reference standards All non-labeled (linear and branched isomers) and labeledstandard (surrogates and internal standards) solutions (standard purity >98%) were purchased from Wellington laboratories (Guelph, ON, Canada) except for perfluoropropanoic acid (PFPA) standard solution, which was purchased from Acros Organics (Geel, Belgium). Single branched isomer standards for P4HpA, P4HpS, P5HpA, P5HpS, P6HpA and P6HpS were not available, so mixtures of P4HpA and P4HpS, P5HpA and P5HpS, P6HpA and P6HpS were used. Two certificates of analysis for the single isomer and isomer mixture purchased from Wellington are supplied as supplementary material. Terminology and acronyms for the identification of perfluoro-compounds following the indications reported by Buck [27] are reported in Table 1. 2.3. Sample shipment and storage Samples were collected in 100 ml polypropylene bottles fitted with a polypropylene screw-cap. The preservation reagent ® (Trizma Hydrochloride) was added to each sample as a solid (0.5 g/100 ml) prior to sample collection. During shipment, the samples were chilled to a temperature below 10 ◦ C and stored at or below 6 ◦ C until extraction. The extraction step was performed within 14 days of conditioning at room temperature for one hour. Extracts were stored at room temperature and analyzed within 15 days. 2.4. Surrogates, internal standard and blank Both the variability of the extraction process and of the analytical instrumentation must be adequately and separately compensated in order to accurately quantify PFAAs using LC–MS/MS. A stable isotope standard solution of 13 C4 -PFBA, 13 C4 -PFOA, 13 C -PFOS, 13 C -PFHxS was used as surrogate analytes and added 4 3 to the sample prior to the extraction process. The labeled surrogates were chosen to encompass all functional groups and cover the water solubility range for the 25 PFAAs studied. Another isotope solution comprising 13 C2 -PFHxA, 13 C3 -PFBA, 13 C -PFOA, 13 C -PFOS, 13 C -PFBS was used as internal standard and 8 8 3 added to the samples after the concentration step and before injection. During the ionization step, source instability causes a stronger

Please cite this article in press as: A. Pellizzaro, et al., Identification and quantification of linear and branched isomers of perfluorooctanoic and perfluorooctane sulfonic acids in contaminated groundwater in the veneto region, J. Chromatogr. A (2017), https://doi.org/10.1016/j.chroma.2017.12.036

G Model CHROMA-359092; No. of Pages 12

ARTICLE IN PRESS A. Pellizzaro et al. / J. Chromatogr. A xxx (2017) xxx–xxx

3

Table 1 List of PFAAs analyzed, acronyms, chemical name, CAS number and molecular formula. Item

Acronym

Reference material

CAS number

Molecular formula

Linear isomers 1 2 3 4 5 6 7 8 9 10 11 12 13 14

PFPA PFBA PFPeA PFHpS PFOS PFBS PFDA PFNA PFDoA PFHpA PFHxA PFUnA PFOA PFHxS

Perfluoropropanoic acid Perfluorobutanoic acid Perfluoropentanoic acid Perfluoroheptane sulfonic acid Perfluorooctane sulfonic acid Perfluorobutane sulfonic acid Perfluorodecanoic acid Perfluorononanoic acid Perfluorododecanoic acid Perfluoroheptanoic acid Perfluorohexanoic acid Perfluoroundecanoic acid Perfluorooctanoic acid Perfluorohexane sulfonic acid

442-64-0 375-22-4 2706-90-3 3965-99-9 1763-23-1 375-73-5 335-76-2 375-95-1 307-55-1 375-85-9 307-24-4 2058-94-8 335-67-1 355-64-4

C3 F5 O2 H C4 F7 O2 H C5 F9 O2 H C7 F15 SO3 H C8 F17 SO3 H C4 F9 SO3 H C10 F19 O2 H C9 F17 O2 H C12 F23 O2 H C7 F13 O2 H C6 F11 O2 H C11 F21 O2 H C8 F15 O2 H C6 FSO3 H

Perfluoro monomethyl substituted branched isomers P1-HpS Perfluoro-1-methylheptane sulfonic acid 15 P3-HpS Perfluoro-3-methylheptane sulfonic acid 16 P4-HpA Perfluoro-4-methylheptanoic acid 17 18 P4-HpS Perfluoro-4-methylheptane sulfonic acid P5-HpA Perfluoro-5-mehylheptanoic acid 19 P5-HpS Perfluoro-5-metilheptane sulfonic acid 20 P6-HpA Perfluoro-6-methylheptanoic acid 21 P6-HpS Perfluoro-6-methylheptane sulfonic acid 22

n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.

C8 F17 SO3 H C8 F17 SO3 H C8 F15 O2 H C8 F17 SO3 H C8 F15 O2 H C8 F17 SO3 H C8 F15 O2 H C8 F17 SO3 H

Perfluoro dimethyl substituted branched isomers P55-HxA 23 24 P35-HxS

Perfluoro-5,5-dimethylhexanoic acid Perfluoro-3,5-dimethylhexane sulfonic acid

n.d. n.d.

C8 F15 O2 H C8 F17 SO3 H

Mass-labeled surrogates solutions 13 C4 -PFBA 25 13 C4 -PFOA 26 13 C4 -PFOS 27 13 28 C3 -PFHxS

Perfluoro-n-[1,2,3,4-13 C4 ] butanoic acid Perfluoro-n-[1,2,3,4-13 C4 ] octanoic acid Perfluoro-1-[1,2,3,4-13 C4 ] octane sulfonic acid Perfluoro-n-[1,2-13 C2 ] hexane sulfonic acid

n.d. n.d. n.d. n.d.

13

Mass-labeled internal standards solutions 13 29 C2 -PFHxA 13 C3 -PFBA 30 13 31 C8 -PFOA 13 32 C8 -PFOS 13 33 C3 -PFBS

Perfluoro-n-[1,2-13 C2 ] hexanoic acid Perfluoro-n-[1,2,3–13 C3 ] butanoic acid Perfluoro-n-[13 C8 ] octanoic acid Perfluoro-1-[13 C8 ] octane sulfonic acid Perfluoro-n-[1,2,3-13C3 ] butane sulfonic acid

n.d. n.d. n.d. n.d. n.d.

13

suppression of short chain PFAAs compared to long chain. The internal standard solution was enriched with short chain 13 C2 -PFHxA and 13 C3 -PFHxS was replaced with 13 C3 -PFBS in order to better compensate for short chain suppression compared to the surrogate solution. Each analyte was quantitated using the internal standard specified in the last column of Table 2. A set of chromatograms for the standard used is reported as supplementary material. While the use of surrogate analytes accounts for variability in sample extraction, internal standards compensate for matrix effects during electrospray source ionization. The use of differently labeled isotopes for surrogates and internal standards makes it possible to separate the effects of sample preparation (extraction) and analysis. Due to the ubiquity of PFAAs, and their presence in the laboratory environment, reagents, bottles, vials and instrumentation were tested by treating DI laboratory water as a sample with subsequent analysis (blank).

2.5. Sample clean up and concentration: PFAA extraction Before extraction, weak anion exchange SPE cartridges were conditioned with 5 ml of methanol and then with 5 ml of 18 M  purified water. Then, 100 ml aliquots of water samples were added with 25 ␮L of 13 C labeled surrogate mix solution, gently homogenized by hand and then loaded into the cartridge at a flow rate of 2 ml min−1 . PFAAs were eluted from the cartridges using a solution 10 ml methanol with 5% (v/v) of aqueous ammonium hydroxide

C4 F7 O2 H C4 12 C4 F15 O2 H C4 12 C4 F17 SO3 H 13 C2 12 C8 F19 O2 H 13 13

C2 12 C4 F11 O2 H C3 12 CF7 O2 H 13 C8 F15 O2 H 13 C8 F17 SO3 H 13 C3 12 CF7 O2 H 13

28%. Extracts were than concentrated to about 200 ␮L or to dryness using a centrifugal vacuum concentrator (Genevac Ltd, Ipswich, UK). They were then reconstituted to 1 ml by a solution comprising methanol: water at a 30:70 ratio (v/v). Last, 5 ␮L of internal standard mix solution was added.

2.6. UPLC method: instrumental conditions Chromatographic separation was performed on an Acquity Hclass system from Waters (Millford, MA, USA) equipped with a quaternary pump, sample manager and autosampler. A trapping column (isolator column Acquity UPLC BEH C18 , 50 × 2.1 mm, 1.7 ␮m) was installed between the eluent mixer and the injector to trap and delay perfluoro-compound contamination from the UPLC system. PEEK solvent lines (Waters “PFC kit”) were installed to eliminate interferences due to traces of PFAAs present in standard solvent lines. Separation was achieved using a Cortecs UPLC C18, 2.1 × 150 mm, with a particle size of 1.6 ␮m (Waters) fitted with an Acquity In-Line Filter including a frit (2.1 mm, 0.2 ␮m). A 20 mm ammonium acetate solution containing 5% (v/v) of methanol was used as mobile phase A. Methanol was used as mobile phase B. The elution process is composed of the following steps: isocratic step (1 min., 60% mobile phase B and 40% of mobile phase A), followed by a linear gradient step (mobile phase B up to 90% in 34 min.). Depending on the concentration of branched isomers, the duration of the linear gradient step was lengthened up to 53 min. in order to reach sufficient peak resolution. Constant aliquots of 15 ␮L

Please cite this article in press as: A. Pellizzaro, et al., Identification and quantification of linear and branched isomers of perfluorooctanoic and perfluorooctane sulfonic acids in contaminated groundwater in the veneto region, J. Chromatogr. A (2017), https://doi.org/10.1016/j.chroma.2017.12.036

G Model

ARTICLE IN PRESS

CHROMA-359092; No. of Pages 12

A. Pellizzaro et al. / J. Chromatogr. A xxx (2017) xxx–xxx

4

Table 2 Instrumental parameters, multiple reaction transition and internal standard (IS) used for all PFAAs considered. Listing follows analytes retention time. Analyte

Mass transition

Cone voltage (V)

Collision energy (V)

IS

PFPA PFBA 13 C4 -PFBA 13 C3 -PFBA PFPeA PFBS

163 > 119 213 > 169 217 > 172 216 > 172 263 > 219 299 > 80 299 > 99 302 > 99 313 > 119 313 > 269 315 > 270 363 > 119 363 > 169 399 > 80 399 > 99 402 > 80 413 > 219 413 > 219 413 > 169 413 > 119 413 > 369 413 > 169 417 > 372 421 > 376 449 > 80 449 > 99 499 > 280 163 > 119 163 > 119 499 > 280 499 > 330 499 > 280 499 > 169 499 > 99 499 > 99 499 > 80 499 > 99 503 > 80 507 > 80 513 > 219 513 > 269 513 > 169 563 > 219 563 > 269 563 > 519 613 > 169 613 > 319 613 > 569

18 18 18 18 15 54 54 54 16 16 15 16 16 40 40 40 16 16 16 16 16 16 15 15 74 74 60 20 20 60 60 60 60 60 60 60 60 78 78 20 20 20 20 20 20 25 25 25

10 10 10 10 9 30 28 30 17 10 15 20 16 40 40 40 10 16 10 10 10 16 10 10 56 34 39 18 16 39 30 39 39 39 39 39 38 48 48 16 16 16 20 18 20 15 15 15

13

13

C3 -PFBS PFHxA 13

C2 -PFHxA PFHpA PFHxS 13

C3 -PFHxS P55-HxA P5-HpA P6-HpA P4-HpA PFOA 13

C4 -PFOA C8 -PFOA PFHpS 13

P35-HxS PFNA P5-HpS P4-HpS P3-HpS P6-HpS P2-HpS P1-HpS PFOS 13

C4 -PFOS C8 -PFOS PFDA 13

PFUnA

PFDoA

were injected and the mobile phase flow rate was kept constant at 0.2 ml min−1 , and the column was thermostated at 45 ◦ C.

2.7. ESI source and MRM method: instrumental conditions PFAAs were quantified in dynamic Multiple Reaction Monitoring anion mode using a triple quadrupole mass spectrometer (MS/MS), Xevo TQD (Waters). A minimum of ten scans across the chromatographic peak were required to ensure adequate precision and at least two transitions per each analyte were recorded when possible. The mass spectrometer was used in negative electrospray mode (ES-) with a capillary needle voltage of 0.8 kV. The source temperature offset was 150 ◦ C. Nitrogen was generated in-situ by a Zefiro 0–40 LC–MS generator from Cinel (Padua, Italy), then used as desolvation and cone gas at flow rates of 1000 l h−1 and 15 l h−1 respectively, and as nebulizing gas. The desolvation gas temperature was 450 ◦ C. Argon was used as collision gas at a flow rate of 3.5 ml min−1 . Compound specific mass spectrometric parameters such as cone voltage, collision energy and mass transitions are listed in Table 2.

C-PFBA C3 -PFBA 13 C3 -PFBA – 13 C3 -PFBA 13 C4 -PFBS 13

− 13

C2 -PFHxA

− 13

C2 -PFHxA

13

C2 -PFHxA

13

C2 -PFHxA C8 -PFOA 13 C8 -PFOA 13 C8 -PFOA 13 C8 -PFOA 13 C8 -PFOA 13

13

C8 -PFOA

− 13

C8 -PFOS

13

C8 -PFOS C8 -PFOA

13

13

C8 -PFOS C8 -PFOS 13 C8 -PFOS 13 C8 -PFOS 13 C8 -PFOS 13 C8 -PFOS 13 C8 -PFOS 13

13

C8 -PFOS

− 13

C8 -PFOS

13

C8 -PFOS

13

C8 -PFOS

The first transition, corresponding to the most abundant product ion, was used for quantification, whereas the second transition was used for confirmation except for PFOA and PFOS branched isomers. The mass spectrometric parameters for each perfluoroalkyl acid were optimized using 1 mg L−1 solution in methanol 40%. The scan range was 70–700 a.m.u. The concentration of the analytes in the samples was calculated using internal standard quantification (surrogates were used only to check recovery for the solid phase extraction step). Data acquisition and processing were performed using MassLynx, ver. 4.1 and TargetLynx Application Manager. 2.8. Calibration curves Calibration curves were drawn separately for linear isomers, branched isomers, surrogates and internal standards. Calibration standard levels and standard mixes are summarized in Table 3. Calibration curves for linear isomers were obtained using a single mix of 14 linear isomers at nine different concentration levels (LEV A – LEV I). Although the instrumental apparatus offers the possibility of quantitating lower levels of linear isomers, the minimum quantification level is set at 100 ng l−1 (corresponding to 1 ng l−1 as

Please cite this article in press as: A. Pellizzaro, et al., Identification and quantification of linear and branched isomers of perfluorooctanoic and perfluorooctane sulfonic acids in contaminated groundwater in the veneto region, J. Chromatogr. A (2017), https://doi.org/10.1016/j.chroma.2017.12.036

G Model CHROMA-359092; No. of Pages 12

ARTICLE IN PRESS A. Pellizzaro et al. / J. Chromatogr. A xxx (2017) xxx–xxx

5

Table 3 Calibration standard solution for different PFAAs concentration level (ng L−1 ). PFAAs

Calibration standard solutions concentration (ng L−1 )

Linear isomers MIX

PFPA PFBA PFPeA PFHpS PFOS PFBS PFNA PFHpA PFHxA PFOA PFHxS PFDA PFUnA PFDoA

Branched isomers (monosubstituted)

P3HpS MIX P1-P4

LEV A

MIX P5-P6

Branched isomers (disubstituted)

P55HxA P35HxS

Surrogates MIX

13

LEV C

LEV D

LEV E

LEV F

LEV G

LEV H

LEV I

97 100 100 95,2 95,6 88,4 100 100 100 100 94,6 100 100 100

194 200 200 190,4 191,2 176,8 200 200 200 200 189,2 200 200 200

485 500 500 476 478 442 500 500 500 500 473 500 500 500

970 1000 1000 952 956 884 1000 1000 1000 1000 946 1000 1000 1000

1940 2000 2000 1904 1912 1768 2000 2000 2000 2000 1892 2000 2000 2000

4850 5000 5000 4760 4780 4420 5000 5000 5000 5000 4730

9700 10000 10000 9520 9560 8840 10000 10000 10000 10000 9460

200 440 200 200 392 620 200 200

500 1100 500 500 980 1550 500 500

1000 2200 1000 1000 1960 3100 1000 1000

2500 5500 2500 2500 4900 7750 2500 2500

5000 11000 5000 5000 9800 15500 5000 5000

390

975 250

1950 500

4875 1500

2500

P4HpA P1HpS P4HpS P5HpA P6HpA P5HpS P6HpS

C4 -PFBA C4 -PFOA C4 -PFOS 13 C3 -PFHxS

5911 14778 1766 5592

7813 19351 2334 7420

9709 24272 2900 9229

11583 28958 3460 11022

13

6000 2994 2940 2784 2782

7968 3984 3904 3697 3705

9950 4975 4875 4617 4263

11928 5964 5845 5535 5546

13 13

Internal standards MIX

LEV B

C2 -PFHxA C3 -PFBA 13 C8 -PFOA 13 C8 -PFOS 13 C3 -PFBS 13

minimum reporting level) which is considered sufficiently cautious for legal purposes considering the limits sets by the Italian Health Ministry (30 ng L−1 for PFOS and 500 ng L−1 for PFOA, PFBA, PFBS and 500 ng ng L−1 for the sum of other PFAAs). Calibration curves for branched isomers were obtained using three single isomers (P3HpS, P55HxA, P35HxS) and two mixes (P1P4 and P5-P6) at five different concentration levels (LEV B – LEV F). In the case of branched isomers, different isomers may co-elute, subsuming each other and reducing peak resolution drastically if calibration is set at overly high concentrations. For these reasons, unlike linear isomers, calibration levels must be differentiated and optimized for each branched isomer or mixture of branched isomers in order to obtain a calibration curve centered on real sample concentration, the maximization of peak resolution and sufficiently precise and accurate instrumental response for quantification. Calibration standard solutions were kept in polypropylene vials at room temperature. A separate calibration curve for linear isomers, branched isomers, labeled surrogates and labeled internal standards maximizes the expression of each analyte in terms of instrumental response (peak area) while the concentration is kept low enough to avoid significant ionic suppression. This is demonstrated by the low residual standard deviation of linear and non-linear second order calibration which for an instrumental reading of 3000 ng L−1 of PFOS (corresponding to the limit set by the Italian Health Ministry) is 16.7%. For this reason, calibration levels were not determined using the IS technique: ESI conditions, LC conditions (chromatographic gradient), method analytes concentration in calibration standard solutions and IS concentration made it possible to avoid

24250 25000 25000 23800 23900 22100 25000 25000 25000 25000 23650

48500 50000 50000 47600 47800 44200 50000 50000 50000 50000 47300

suppression in the ESI source during unlabeled analyte calibration. Furthermore, an initial calibration check, continuing calibration check (by calibration standards) and quality control sample check (method analytes solution obtained from a source different from the source of calibration) made it possible to verify periodically the accuracy of the existing calibration for all analytes (method analytes, surrogates and internal standards). IS was added only to field samples in order to assess suppression depending on matrix effects and on overly high concentration analyte(s). Calibration curves for each analyte were approximated by linear or quadratic functions which were forced to zero.

3. Results and discussion 3.1. Quantitative analysis of PFOA and PFOS branched isomers: method development The analytical identification of PFOA and PFOS branched isomers was carried out by UPLC separation, and different transitions were used to discriminate the response of each PFOA and PFOS branched isomer. A rough separation was obtained using a column Acquity UPLC BEH C18. 50 × 2.1 mm, 1.7 ␮m as a preliminary step. Mobile phases A and B consisted of 20 mm ammonium acetate solution and methanol respectively. UPLC separation was done running a gradient under the following conditions: 40% of B, followed by a linear gradient of eluent B up to 90% in 10 min. Fig. 1 shows the chromatograms obtained for PFOS and PFOA in the rough separation test for a field sample of contaminated groundwater. Two transitions were considered for each molecule: the first for quantification (499 > 80 for PFOS, 413 > 369 for PFOA) and the second for

Please cite this article in press as: A. Pellizzaro, et al., Identification and quantification of linear and branched isomers of perfluorooctanoic and perfluorooctane sulfonic acids in contaminated groundwater in the veneto region, J. Chromatogr. A (2017), https://doi.org/10.1016/j.chroma.2017.12.036

G Model CHROMA-359092; No. of Pages 12 6

ARTICLE IN PRESS A. Pellizzaro et al. / J. Chromatogr. A xxx (2017) xxx–xxx

Fig. 1. Rough separation peaks obtained for PFOS (1.a) and PFOA (1.b) obtained with a column Acquity UPLC BEH C18, 50 × 2,1 mm, 1,7 ␮m for the contaminated groundwater field sample 1.

confirmation (499 > 99 for PFOS and 413 > 169 for PFOA). For both PFOS (Fig. 1a) and PFOA (Fig. 1b), two other peaks were observed in addition to the linear isomer peak (most abundant). In order to investigate the nature of the compounds comprising each additional peak, chromatographic separation was improved using longer columns. Cortecs UPLC C18, Acquity UPLC BEH C18 (Waters), Kinetex PFP and Kinetex Biphenyl (Phenomenex) columns—all measuring 2.1 × 150 mm—were tested. Best separation was achieved on Cortecs UPLC C18, 2.1 × 150 mm, which uses core-shell technology particles with a nominal diameter of 1.6 ␮m. More specific transitions were found through infusion of individual branched isomer solutions (about 0.1 mg L−1 , 50%:50% methanol:water): these additional transitions proved useful in determining and quantifying branched isomers for PFOA (413 > 219, 413 > 119) and PFOS (499 > 280, 499 > 330, 499 > 169). Figs. 2 and 3 refer to the improved separation step for two different field samples of contaminated groundwater. Samples were

taken at different depths of the aquifer in the contaminated area in the south-western part of the province of Vicenza. The data show that by optimizing the chromatographic gradient conditions and increasing the number of transitions considered, different isomers can be resolved and definitely quantitated. Linear, monosubstituted and disubstituted isomer chromatographic peaks can be resolved and integrated using different transitions. In the case of PFOS (Figs. 2a and Figure 3a), transition 499 > 280 was used to quantify isomers P3-HpS, P5-HpS and disubstituted P35-HxS; transition 499 > 330 was used for isomer P4-HpS; transition 499 > 99 for isomers P2-HpS and P1-HpS; and transition 499 > 169 for isomer P6-HpS. For PFOA (Figs. 2b and Figure 3b), transition 413 > 219 was used to quantify P5-HpA and disubstituted P55-HxA; transition 413 > 169 for P6-HpA and transition 413 > 119 for P4-HpA. The C18 column phase separated PFOS and PFOA isomers according to their lipophilic/hydrophilic balance. Lipophilicity is

Please cite this article in press as: A. Pellizzaro, et al., Identification and quantification of linear and branched isomers of perfluorooctanoic and perfluorooctane sulfonic acids in contaminated groundwater in the veneto region, J. Chromatogr. A (2017), https://doi.org/10.1016/j.chroma.2017.12.036

G Model CHROMA-359092; No. of Pages 12

ARTICLE IN PRESS A. Pellizzaro et al. / J. Chromatogr. A xxx (2017) xxx–xxx

7

Fig. 2. Improved chromatographic separation for both PFOS (2.a) and PFOA (2.b) for contaminated groundwater field sample 1 revealing the presence of perfluoromethyl and perfluorodimethyl substituted isomers with higher retention times compared to the linear homologues.

Please cite this article in press as: A. Pellizzaro, et al., Identification and quantification of linear and branched isomers of perfluorooctanoic and perfluorooctane sulfonic acids in contaminated groundwater in the veneto region, J. Chromatogr. A (2017), https://doi.org/10.1016/j.chroma.2017.12.036

G Model CHROMA-359092; No. of Pages 12 8

ARTICLE IN PRESS A. Pellizzaro et al. / J. Chromatogr. A xxx (2017) xxx–xxx

Fig. 3. Improved chromatographic separation for both PFOS (3.a) and PFOA (3.b) for contaminated groundwater field sample 2.

Please cite this article in press as: A. Pellizzaro, et al., Identification and quantification of linear and branched isomers of perfluorooctanoic and perfluorooctane sulfonic acids in contaminated groundwater in the veneto region, J. Chromatogr. A (2017), https://doi.org/10.1016/j.chroma.2017.12.036

G Model

ARTICLE IN PRESS

CHROMA-359092; No. of Pages 12

A. Pellizzaro et al. / J. Chromatogr. A xxx (2017) xxx–xxx

9

Table 4 PFOA concentration estimate accounting separately linear and branching isomers (column b) and totaling Linear and branched isomers (Column c).

SAMPLE 1 SAMPLE 2 SAMPLE 3 SAMPLE 4 SAMPLE 5

Column a L-PFOA (LINEAR ISOMER) ng/L

Column b L-PFOA + BRANCHED ISOMERS ng/L

Column c L-PFOA + BRANCHED ISOMERS, ng/L (transition m/z 413 ->369)

130 3426 1093 328 358

150 4084 1248 438 464

157 4260 1304 450 483

Table 5 PFOS concentration estimate accounting separately linear and branched isomers (column b), totaling Linear and branched isomers applying the transition m/z 499 ->80 (column c) and the transition m/z 499->99 (column d).

SAMPLE 1 SAMPLE 2 SAMPLE 3 SAMPLE 4 SAMPLE 5

Column a L-PFOS (LINEAR ISOMER) ng/L

Column b L-PFOS + BRANCHED ISOMERS, ng/L

Column c L-PFOS + BRANCHED ISOMERS, ng/L (transition m/z 499 -> 80)

Column d L-PFOS + BRANCHED ISOMERS, ng/L (transition m/z 499 -> 99)

409 196 66 14 15

506 241 79 17 19

642 308 92 23 26

508 245 75 18 20

Fig. 4. Peak areas used for total PFOA concentration estimation using transition 413 > 369. Peak areas used for total PFOS concentration estimation using respectively transition 499 > 80 and 499 > 99.

Please cite this article in press as: A. Pellizzaro, et al., Identification and quantification of linear and branched isomers of perfluorooctanoic and perfluorooctane sulfonic acids in contaminated groundwater in the veneto region, J. Chromatogr. A (2017), https://doi.org/10.1016/j.chroma.2017.12.036

G Model CHROMA-359092; No. of Pages 12

100 100 100 100 100 100 100 100 100 100 100 100 100 100 500 500 500 500 200 200 200 200 200 200 200 – – – – – – – – – – – – – – 94 114 107 – – – – – – 117 – 2 1.8 1.6 5.3 2.8 2.6 2.4 4.6 4.9 2.1 2.5 5.2 5.7 6.6 6.6 4.7 6.7 3.9 8.5 8.6 2.3 9.1 8.3 7.7 11.3 74 105 97 90 95 106 108 91 97 109 102 81 100 105 101 104 86 106 87 92 93 99 91 108 98 3.7 6 7.1 8.3 9.7 5.5 14.6 9.6 13.3 10.6 7.5 13.4 8.1 8.7 14.7 2.3 6.4 8.9 9.2 10.1 9 14 10.7 10.5 11.5

Recovery %

73 102 102 90 84 122 125 103 98 127 104 81 91 129 88 124 95 109 103 103 119 84 105 103 89 4.6 2.1 6.7 9 16.8 5.5 19.6 13.5 14.6 13.9 3.7 17.6 12.6 13 – – – – – – – – – – – 101 121 115 70 75 85 76 84 101 97 128 73 94 95 – – – – – – – – – – –

– – – – – – – – – – – – – – 6.1 4.8 7.3 – – – – – – 6.6 –

80 99 97 103 111 96 – 107 – 104 101 – 94 96 – – – – – – – – – – –

Recovery %

2.4 2 3.1 3.7 2.4 2.9 – 2.9 – 3.1 2.6 – 2.4 2.7 – – – – – – – – – – –

Minimum Reporting Level (ng L−1 ) Instrumental quantification limit (ng L−1 ) 25000 ng L−1

Recovery % Recovery % RSD%

500 ng L−1

RSD% 200 ng L−1

influenced by the position of branching points and by the overall number of carbon atoms in the fluorinated backbone: linear isomers are more lipophilic and are thus retained more strongly than branched isomers, which have progressively shorter retention times as a function of the increasing number of branching points (from perfluoro monomethyl substituted to perfluoro dimethyl substituted isomers) [28]. All other PFAAs analyzed are reported as supplementary material for both field samples chromatograms. The method described has been applied since 2013 to analyze PFAAs in groundwater and drinking water samples collected in the contaminated area in the south-western part of the province of Vicenza. In general, PFBA, PFBS, PFPeA, PFHxA and PFOA (and relative branched isomers) were the dominant contaminants in the samples. PFOA and PFOS branched isomers were about 22% and 35%, respectively, of total branched and linear isomers. In the case of multiple isomers, EPA method 537:2009 [29] suggests integrating all the chromatographic peaks observed for each isomer by totaling the peak areas and relying on the initial calibration obtained with the linear isomer quantitative standard. When applying this approach to groundwater and drinking water samples collected in the contaminated area, the integration of linear and branched isomers led to an overestimate of PFAA load (in particular for the sum of PFOS isomers using the mass transition m/z 499 ->80). Five different groundwater samples were analyzed and the results were compared in order to characterize the overestimate produced by applying the approach of the totalization of linear and branched isomers versus the quantification of single isomers. Column a) of Table 4 reports the concentration estimation of the linear isomer of PFOA alone. Column b) reports the concentration estimation obtained by considering the contribution of linear and branched isomers separately according to the method described. Finally, column c) reports the concentration estimation obtained as suggested by the EPA by totaling linear and branched isomers using only transition 413 > 369, specifically for the quantification of the linear isomer of PFOA. Comparing columns b) and c), the method proposed by the EPA shows an average overestimation of 3.2% compared to the method presented in this paper. Fig. 4 graphically shows peak areas used for total PFOA concentration estimation using transition 413 > 369. Column a) of Table 5 reports the concentration estimation of the linear isomer of PFOS alone. Column b) reports the concentration estimation obtained by considering the contribution of linear and branched isomers separately according to the method described in this paper. Columns c) and d) report the concentration estimation obtained by totaling linear and branched isomers using transitions 499 > 80 and 499 > 99, respectively. It is worth noting that the PFOS concentration estimation obtained using transition 499 > 80 (which is specific for linear isomer quantification) shows an average overestimation of 21.3% when compared with the estimation obtained using the present method. The overestimation is reduced significantly if transition 499 > 99 (i.e. the primary confirmation ion) is used. Fig. 4 graphically shows peak areas used for the total PFOS concentration estimation when using transition 499 > 80 and 499 > 99, respectively. The EPA recently recommended that laboratories analyzing samples for PFOA using EPA Method 537 quantify both linear and branched isomers.

PFPA PFBA PFPeA PFHpS L-PFOS PFBS PFDA PFNA PFDoA PFHpA PFHxA PFUnA L-PFOA PFHxS P4-HpA P5-HpA P6-HpA P5,5-HxA P1-HpS P2-HpS P3-HpS P4-HpS P5-HpS P6-HpS P3,5-HxS

3.2. Method validation

Analyte

Concentration level

Recovery % n = 4-7

RSD% 10000 ng L−1

RSD% 5000 ng L−1

RSD%

1 1 1 1 1 1 1 1 1 1 1 1 1 1 5 5 5 5 2 2 2 2 2 2 2

A. Pellizzaro et al. / J. Chromatogr. A xxx (2017) xxx–xxx

10

Table 6 Relevant method performance parameters: recovery, relative standard deviations, instrumental quantification and minimum reporting Level for each of the 24 PFAAs studied.

ARTICLE IN PRESS

The initial demonstration of capability, quality control requirements, specifications, frequencies and acceptance criteria were applied in accordance with EPA method 537:2009. The instrumental quantification limit was determined checking repeatability and

Please cite this article in press as: A. Pellizzaro, et al., Identification and quantification of linear and branched isomers of perfluorooctanoic and perfluorooctane sulfonic acids in contaminated groundwater in the veneto region, J. Chromatogr. A (2017), https://doi.org/10.1016/j.chroma.2017.12.036

G Model CHROMA-359092; No. of Pages 12

ARTICLE IN PRESS A. Pellizzaro et al. / J. Chromatogr. A xxx (2017) xxx–xxx

recovery at low concentrations: seven replicates were fortified, extracted and analyzed. Then, mean (m), standard deviation (SD) and half range for the prediction interval of results HRPIR (=3.963SD) were calculated. Upper and lower limits for the Prediction Interval of Results (PIR = Mean ± HRPIR ) were confirmed at ≤ 150% and ≥ 50%, respectively. The application of EPA method 537:2009 for analysis of perfluoroalkyl acids in drinking water was certified in 2015 according to ISO 17,025. Detection and quantification limits together with other relevant method performance parameters for each of the PFAAs studied are listed in Table 6. 3.3. Quality control and quality assurance Procedural blank and instrumental blank were checked periodically for background contamination and for carryover during the extraction phase. Initially, recovery and relative standard deviation of all analytes were determined in four to seven replicates of groundwater samples fortified near the midrange calibration concentration: mean recovery ranged from 70 to 130% and RSD was lower than 20%, in accordance with EPA method 537:2009 (Table 6). For the sake of completeness of the calibration step, a quality control on the calibration curve was performed using standards (at low, mid and high concentration) belonging to different production batches compared to those used for drawing the curve. The results were required to be within 70–130% of true value. Surrogate standard recoveries must be within 70–130% of the true value, and internal standard must be within 70–140% of the true value. Calibration standards were calculated as unknown samples using the calibration curve: the analyte was within 70–130% of the true value for all the standards except the lowest one (quantification limit), which was within 50–150% of the true value, in complete agreement with EPA method 537:2009. The peak asymmetry factor for the first two eluting chromatographic peaks in a mid-level calibration standard must be between 0.8–1.5. The method was periodically tested using LGC Aquacheck Proficiency Testing Schemes. Nine laboratories participated in the first single round 480 (February 2015): assigned values for PFOA and PFOS were 7.69 ␮g l−1 and 2.55 ␮g l−1 , respectively, and two results were excluded. The UPLC-MS/MS method presented in this paper produced quantitative results for PFOA and PFOS of 7.46 ␮g L−1 (generating a z-score of -0.30) and 1.76 ␮g L−1 (generating a z-score of −2.85), respectively. Fourteen laboratories participated in single round 492 (September 2015): assigned values for PFOA and PFOS were 4.92 ␮g L−1 and 2.21 ␮g L−1 , respectively, and three results were excluded. The UPLC-MS/MS method presented in this paper produced quantitative results for PFOA and PFOS of 5.44 ␮g L−1 (generating a z-score of 0.89) and 2.41 ␮g L−1 (generating a z-score of 0.82), respectively. Fourteen laboratories took part in the last single round 524 (May 2017): assigned values for PFOA and PFOS were 4.55 ␮g L−1 and 2.45 ␮g L−1 , respectively, and five results were excluded. The UPLC-MS/MS method presented in this paper produced quantitative results for PFOA and PFOS of 4.70 ␮g L−1 (generating a z-score of 0.33) and 3.10 ␮g L−1 (generating a z-score of 2.42), respectively. According to ISO 13528, the absolute value of z-score should be less than 3 for acceptable performances of a single round. 3.4. Uncertainty estimation The combined standard uncertainty uc (y) was evaluated by a metrological approach in accordance with Eurachem/Citac Guide CG4 [30] for each analyte at different concentration levels; uc (y) was the positive square root of the combined variance uc 2 (y) which was the sum of the following contributions: repeatability (precision, the largest), recovery (trueness) in accordance with VAM Project 3.2.1 [31] and the residual standard deviation of linear

11

and non-linear second-order calibration function in accordance with ISO 8466-1/2. The expanded uncertainty U% was determined by multiplying uc (y) with a coverage factor k = t(eff ) where eff was the number of degrees of freedom obtained from the WelchSatterthwaite formula and t was the t-Student variable (the k value defined an interval with a level of confidence of approximately 95%). For all analytes, U% varied from about 50% for the instrumental quantification limit to 30% for the higher calibration level. 4. Conclusions Although linear isomers represent the predominant form for both PFOA and PFOS in contaminated ground water, perfluoro monomethyl substituted and perfluoro dimethyl substituted branched isomers account for a significant part of the total PFOA and PFOS burden (22–35%). The integration of all chromatographic peaks related to different isomers of a single compound can lead to an appreciable overestimation, especially for PFOS. For this reason, a solid and sensitive UPLC-MS/MS method has been developed, enabling the simultaneous determination and quantification of 14 linear PFAAs, five branched perfluoro monomethyl substituted isomers of PFOS, three branched perfluoro monomethyl substituted isomers of PFOA, one branched perfluoro dimethyl substituted isomer of PFOS and one perfluoro dimethyl substituted isomer of PFOA. Further, the application of the method described makes it possible to total PFOA and PFOS isomers according to the approach suggested by EPA method 537:2009 (which is practical for routine drinking water analyses) using transition 413 -> 369 for PFOA isomers and transition 499 -> 99 for PFOS isomers. Acknowledgments The authors greatly thank the laboratory at Acque del Chiampo S.p.A. for the analytical determination of PFAAs. A. Pellizzaro and M. Fant also thank D. Cracco and A. Breda for the SPE extraction and Waters S.p.A. and Phenomenex S.r.l. for their technical support. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.chroma.2017. 12.036 References [1] World Health Organization, Keeping Our Water Clean: the Case of Water Contamination in Veneto Region, Italy, World Health Organization, 2017, ISBN 9789289052467. [2] R.C. Buck, Toxicology data for alternative “short-chain” fluorinated substances, in: J.C. Jamie (Ed.), Toxicological Effects of Perfluoroalkyl and Polyfluoroalkyl Substances, Springer, Switzerland, 2015, pp. 451–475. [3] I.N. Rozhkov, Radical-cation mechanism of the anodic fluorination of organic compounds, Russ. Chem. Rev. 45 (1976) 615–629. [4] G.P. Gambaretto, M. Napoli, L. Conte, A. Scipioni, R. Armelli, The electrochemical fluorination of organic compounds: further data in support of the ECb ECn mechanism, J. Fluor. Chem. 27 (1985) 149–155. [5] W. Jiang, Y. Zhang, L. Yang, X. Chu, L. Zhu, Perfluoroalkyl acids (PFAAs) with isomer analysis in the commercial PFOS and PFOA products in China, Chemosphere 127 (2015) 180–187. [6] S.M. Vyas, I. Kania-Korwell, H.J. Lehmler, Differences in the isomer composition of perfluorooctansulfonyl (PFOS) derivatives, J. Environ. Sci. Health, Part A Toxic/Hazard Subst. 42 (2007) 249–255. [7] S. Rayne, K. Forest, H.J. Friesen, Congeneric-specific numbering system for the environmentally relevant C4 through C8 perfluorinated homologues groups of alkyl sulfonates, carboxylates, telomere alcohols, olefins and acids, and their derivatives, J. Environ. Sci. Health A 43 (2008) 1391–1401. [8] W.K. Reagen, K.R. Lindstrom, C.B. Jacoby, R.G. Purcell, T.A. Kestner, R.M. Payfer, J.W. Miller, Environmental characterization of 3M electrochemical fluorination derived perfluorooctanoate and perfluorooctanesulfonate

Please cite this article in press as: A. Pellizzaro, et al., Identification and quantification of linear and branched isomers of perfluorooctanoic and perfluorooctane sulfonic acids in contaminated groundwater in the veneto region, J. Chromatogr. A (2017), https://doi.org/10.1016/j.chroma.2017.12.036

G Model CHROMA-359092; No. of Pages 12 12

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

ARTICLE IN PRESS A. Pellizzaro et al. / J. Chromatogr. A xxx (2017) xxx–xxx

Milwaukee, WI, USA, 11–15 November, Society of Environmental Toxicology and Chemistry 28th North American Meeting (2007). J.P. Benskin, A.O. De Silva, J.W. Martin, Isomer profiling of perfluorinated substances as a tool for source tracking: a review of the early findings and future applications, Rev. Environ. Contam. Toxicol. 208 (2010) 111–160. S. Rayne, K. Forest, Estimated pKa values for the environmentally relevant C-1 through C-8, J. Environ. Sci. Health, Part A: Toxic/Hazard Subst. 51 (2016) 1018–1023. Y. Zhang, S. Beesoon, L. Zhu, J.W. Martin, Biomonitoring of perfluoroalkyl acids in human urine and estimates of biological half-life, Environ. Sci. Technol. 47 (2013) 10619–10627. Y. Liu, A.S. Pereira, S. Beesoon, R. Vestergren, U. Berger, G.W. Olsen, A. Glynn, J.W. Martin, Temporal trends of perfluorooctanesulfonate isomer and enantiomer patterns in archieved Swedish and American serum samples, Eviron. Int. 75 (2015) 215–222. W.A. Gebbekin, U. Berger, I.T. Cousins, Estimating human exposure to PFOS isomers and PFCA homologues: the relative importance of direct and indirect (precursor) exposure, Environ. Iner. 74 (2015) 160–169. Y. Zhang, S. Beesoon, L. Zhu, J.W. Martin, Isomers of perfluorooctanesulfonateand perfluorooctanoate and total perfluoroalkyl acids in human serum from two cities in North China, Environ. Int. 53 (2013) 9–17. H. Peng, S. Zhang, J. Sun, Z. Zhang, J.P. Giesy, J. Hu, Isomer-specific accumulation of perfluorooctanesulfonate from (N-ethyl perfluorooctanesulfonamido)ethanol-based phosphate diester in Japanese medaka (Oryzias, latipes), Environ. Sci. Technol. 48 (2014) 1058–1066. H. Jao, Y. Zhang, L. Zhu, J.W. Martin, Isomer profiles of perfluoroalkyl substances in water and soil surrounding a Chinese fluorochemical manufacturing park, Environ. Sci. Technol. 49 (2015) 4946–4954. J.P. Benskin, L.W. Yeung, N. Yamashita, S. Taniyasu, P.K.S. Lam, J.W. Martin, Perfluorinated acid isomer profiling in water and quantitative assessment of manufacturing sources, Environ. Sci. Technol. 44 (2010) 9049–9054. Z. Zhou, Y. Shi, R. Vestergren, T. Wang, Y. Liang, Y. Cai, Highly elevated serum concentrations of perfluoroalkyl substances in fishery employees from Tangxun lake, China, Environ. Sci. Technol. 48 (2014) 3864–3874. S. Beesoon, G.M. Webster, M. Shoeib, T. Harner, J.P. Benskin, J.W. Martin, Profiles of Perfluorochemicals in matched maternal, cord, and house dust samples: manufacturing sources and transplacental transfer, Environ. Health Perspect. 119 (2011) 1659–1664.

[20] S. Fang, Y. Zhang, S. Zhao, L. Qiang, M. Chen, L. Zhu, Bioaccumulation of perfluoroalkyl acids including the isomers of perfluorooctane sulfonate in carp (Cyprinus carpio) in a sediment/water microcosm, Environ. Toxicol. Chem. 35 (2016) 3005–3013. [21] R.L. Sharpe, J.P. Benskin, A.H. Laarman, S.L. Macleod, J.W. Martin, C.S. Wong, G.G. Goss, Perfluorooctane sulfonate toxicity, isomeric specific accumulation, and maternal transfer in Zebrafish (Danio rerio) and rainbow trout (Oncorhyncus mykiss), Environ. Toxicol. Chem. 29 (2010) 1957–1966. [22] J.O.’ Brien, S. Kennedy, S. Chu, R. Letcher, Isomeric-specific accumulation of perfluorooctane sulfonate in the liver of chicken embryos exposed in ovo to a technical mixture, Environ. Toxicol. Chem. 30 (2011) 226–231. [23] Jana Johansson, PhD Dissertation Stockholm, 2017, 978-91-7649-700-5; 978-91-7649-701-2 http://urn.kb.se/resolve?urn=urn:nbn:se:su:dive142116. [24] G. Shan, Z. Wang, L. Zhou, P. Du, X. Luo, Q. Wu, L. Zhu, Impact of daily intakes on the isomeric profiles of perfluoroalkyl substances (PFASs) in human serum, Environ. Int. 89–90 (2016) 62–70. [25] X. Chen, L. Zhu, Occurrence, fluxes and sources of perfluoroalkyl substances with isomer analysis in the snow of northern China, J. Hazard. Mater. 299 (2015) 639–646. [26] A.K. Greaves, R.J. Letcher, Linear and branched perfluorooctane sulfonate (PFOS) isomer patterns differ among several tissues and blood of polar bears, Chemosphere 93 (2013) 574–580. [27] R.C. Buck, J. Franklin, U. Berger, J.M. conder, I.T. Cousins, P. de Voogt, A.A. Jensen, K. Kannan, S.A. Mabury, S.P.J. van Leeuwen, Perfluoroalkyl and polyfluoroalkyl substances in the environment: terminology, classification and origins, Integr. Environ. Assess. Manag. 7 (4) (2011) 513–541. [28] M. Houde, G. Czub, J.M. Small, S. Backus, X. Wang, M. Alaee, D.C. Muir, Fractionation and bioaccumulation of perfluorooctanesulfonate (PFOS) isomers in a Lake Ontario food wed, Environ. Sci. Technol. 42 (24) (2008) 397–403. [29] EPA Method 537 rev 1.1 Determination of selected perfluoroalkylated alkyl acids in drinking water by solid phase extraction and liquid chromatography/tandem mass spectrometry (L/MS/MS), Sept. 2009. [30] Eurachem-CITAC guide CG4, Quantify Uncertainity in Analytical Measurement, third edition, Ellison and Williams, UK, 2012. [31] V.J. Barwick, S.L.R. Ellison, WAM Project 3.2.1. Development and Harmonization of Measurement Uncertainity Principles, Vers. 5.1, 2000.

Please cite this article in press as: A. Pellizzaro, et al., Identification and quantification of linear and branched isomers of perfluorooctanoic and perfluorooctane sulfonic acids in contaminated groundwater in the veneto region, J. Chromatogr. A (2017), https://doi.org/10.1016/j.chroma.2017.12.036