Session B: Sources, pathways, analysis

Transcription

1 Session B: Sources, pathways, analysis B Simultaneous determination of perfluoroalkyl carboxylates (PFCAs), sulfonates (PFSAs) and phosphonates (PFPAs) in food cauldrons Shahid Ullah, Robin Vestergren, Tomas Alsberg, Urs Berger Department of Applied Environmental Science (ITM), Stockholm University, Stockholm, Sweden B A matrix effect free method for ultra trace analysis of perfluoroalkyl acids in dietary samples Robin Vestergren, Shahid Ullah, Ian T. Cousins and Urs Berger Department of Applied Environmental Science (ITM), Stockholm University, Svante Arrhenius väg 8, SE 69 Stockholm, Sweden B3 PFAS Comparison of MS MS and MS/TOF techniques & two extraction methods on lean and fatty fish Sandra Huber, Ondrej Lacina, Petra Hradkova, Jana Pulkrabova, Dorte Herzke, Roland Kallenborn 3 and Jana Hajslova Norwegian Institute for Air Research (NILU), FRAM Centre, Hjalmar Johansens gate 4, NO 996 Tromsø, Norway Institute of Chemical Technology, Prague, Department of Food Chemistry and Analysis, Technicka 3, 66 8 Prague 6, Czech Republic 3 Norwegian Institute for Air Research (NILU), Instituttveien 8, NO 7 Kjeller, Norway B4 A validation program for analytical methods for PFASs in food Stefan P.J. van Leeuwen, Kees Swart Institute for Environmental Studies (IVM), VU University, De Boelelaan 87, 8 HV Amsterdam B Isotope dilution technique in the analysis of PFOS and in water samples. Sami Huhtala, Noora Perkola, Petra Kosubova Finnish Environment Institute, Laboratories, Research and Innovation Laboratory, Helsinki, Finland Central Institute for Supervising and Testing in Agriculture, Brno, Czech Republic B6 Simple and high throughput method for quantitation of perfluoroalkyl substances using LC MS/MS Jani M. Koponen, Päivi Ruokojärvi, Hannu Kiviranta National Institute for Health and Welfare (THL), Department of Environmental Health, Kuopio, Finland B Polyfluoroalkyl compounds in the atmosphere at a wastewater Treatment Plant Lena Vierke, Lutz Ahrens, Mahiba Shoeib, Eric J. Reiner 3, Guo Rui 3, Wolf Ulrich Palm 4, Tom Harner, Ralf Ebinghaus Federal Environment Agency, Germany Environment Canada, Canada 3 Ontario Ministry of the Environment, Canada 4 Leuphana University of Lüneburg, Germany Helmholtz Zentrum Geesthacht, Germany) B Estimating physicochemical properties of poly and perfluorinated alkyl substances (PFAS) with a quantum chemistry based model

2 Zhanyun Wang, Matthew MacLeod, Ian Cousins, Martin Scheringer, Konrad Hungerbühler Institute for Chemical and Bioengineering, Swiss Federal Institute of Technology, ETH Zurich, CH 893 Zurich, Switzerland Department of Applied Environmental Science (ITM), Stockholm University, SE 69 Stockholm, Sweden B Inhalation anaesthetics and climate change Ole J. Nielsen, Mads P. Sulbaek Andersen *, Stanley P. Sander, Deborah S. Wagner 3, Theodore J. Sanford Jr. 4 Department of Chemistry, University of Copenhagen, Universitetsparken, DK Copenhagen, Denmark Jet Propulsion Laboratory, California Institute of Technology, 48 Oak Grove Drive, Mail Stop 83 9, Pasadena, CA 99, USA 3 Department of Clinical Sciences, College of Pharmacy, University of Michigan, 48 Church Street, Ann Arbor, Michigan 489 6, USA 4 Department of Anesthesiology, University of Michigan Medical School, East Medical Center Drive, Ann Arbor, Michigan , USA B Atmospheric Chemistry of CF 3 CH OCH 3 Freja From Østerstrøm, Ole John Nielsen, Mads P.S. Andersen CCAR, University of Copenhagen, Universitetsparken, Copenhagen Ø, Denmark, Jet Propulsion Laboratory, California Institute of Technology, 48 Oak Grove Drive, Mail Stop 83 9, Pasadena, CA 99, USA B6 Development of a method to combine the extraction of both PFCs and legacy POPs from human serum. Christian Bjerregaard Olesen, Eva Cecilie Bonefeld Jørgensen Centre for Arctic Environmental Medicine & Unit of Cellular and Molecular Toxicology, School of Public Health, Aarhus University, Denmark Bartholins Allé, 8 Aarhus C, Denmark.

3 (counts) (counts) (counts) SIMULTANEOUS DETERMINATION OF PERFLUOROALKYL CARBOXYLATES (PFCAs), SULFONATES (PFSAs) AND PHOSPHONATES (PFPAs) IN FOOD CAULDRONS Shahid Ullah, Robin Vestergren, Tomas Alsberg, Urs Berger Department of Applied Environmental Science (ITM), Stockholm University, 69 Stockholm, Sweden Introduction Perfluoroalkyl carboxylates (PFCAs), sulfonates (PFSAs), and phosphonates (PFPAs), together referred to as perfluoroalkyl acids (PFAAs), are used in a wide range of products and industrial applications, due to their inertness and exceptional surface tension lowering potential. As a consequence, some PFAAs have been found ubiquitously in humans and wildlife. Nevertheless, exposure pathways are still not well understood. Dietary intake was suggested to be one of the major routes of human exposure to perfluorooctanoate and perfluorooctane sulfonate (Vestergren et al., 8). PFCAs and PFSAs have been detected in food from various countries (Ericson et al., 8; Ostertag et al., 9). However, more reliable and sensitive analytical methods are needed to accurately quantify the low levels of PFAAs occurring in food, which is a prerequisite for tracking the sources of food contamination and quantifying the relative importance of diet to the total human exposure. In this study a multi-chemical method based on ultra performance liquid chromatography coupled to quadrupole time-of-flight high resolution mass spectrometry (UPLC-qToF-HRMS) was developed and validated for the determination of PFCAs (C4-), PFSAs (C4,6,8,) and PFPAs (C6,8,) in food cauldrons. Materials and Methods For method development and validation a composite baby food matrix was used, which was found to be free from detectable levels of PFAAs. Homogenized samples (aliquots of g) were extracted with acetonitrile:water (9:) followed by enrichment of the analytes on a CUQAX6 solid phase extraction (SPE) cartridge (UCT). -methyl piperidine (-MP) was used in the elution solvent for the SPE cartridge. Analyte separation and quantification was achieved by UPLC-qToF-HRMS (Acquity UPLC, Waters, and Q-ToF Premier, Micromass). Quantification was performed using the internal standard (IS) method. UPLC conditions UPLC column Acquity UPLC BEH C8 (.7 µm particles,. mm) as separation column. The same type of column was also inserted upstream of the injector and used as trapping column for instrumental PFAA contamination. Run time Gradient with total run time min. Mobile phase composition - Water, methanol, acetonitrile, ammonium acetate, -MP Figure : Inter-day method recovery (n = 3) of PFAAs at two different concentrations spiked to a baby food sample. Table : Method detection limits (MDLs) and method limits of quantification (MLQs) for all PFAAs. Perfluoroalkyl carboxylates (PFCAs) PFBA PFPeA PFHxA PFHpA PFNA PFDA PFUnDA PFDoDA MDL (pg/g) MLQ (pg/g) Perfluoroalkyl sulfonates (PFSAs) & phosphonates (PFPAs) PFBS PFHxS PFOS PFDS PFHxPA PFOPA PFDPA Results The use of -MP significantly improved the chromatography of PFPAs (Figure ) as well as the instrumental sensitivity and SPE recovery for PFPAs, PFCAs and PFSAs. Good whole method recoveries for all analytes were obtained (Figure ). Recoveries for PFHxPA and PFOPA were overestimated due to a matrix effect in ionization. Method detection limits (MDLs) and limits of quantification (MLQs) were in the low pg/g range for all analytes (Table ). Excellent linearity for the whole method (r values >.996) was obtained for all analytes over a spike range of.3. ng/g in baby food samples (6 data points). food duplicate samples collected in Germany were successfully analyzed. PFOS was found as the dominating PFAA with levels up to 84 pg/g (Table ). MDL (pg/g) MLQ (pg/g) Table : Concentrations (pg/g) of PFAAs in food duplicate samples collected in Germany. For PFOS the percentage of the linear isomer and sum of branched isomers is additionally given. nd = not detected, <MLQ = detected but below the method limit of quantification. Sample ID Concentrations of PFCAs (pg/g) PFBA PFPeA PFHxA PFHpA PFNA PFDA PFUnDA PFDoDA Sample nd nd <MLQ <MLQ 8 <MLQ <MLQ <MLQ nd Sample nd <MLQ <MLQ <MLQ <MLQ <MLQ nd Sample 3 nd nd <MLQ <MLQ <MLQ <MLQ <MLQ <MLQ nd Sample 4 nd nd <MLQ nd <MLQ nd nd nd nd Sample nd nd <MLQ 9 <MLQ nd <MLQ nd nd Sample 6 nd nd <MLQ <MLQ nd Sample 7 nd nd <MLQ <MLQ 9. <MLQ nd nd nd Sample 8 nd nd 9 <MLQ 6.3 <MLQ nd nd nd Sample 9 nd nd.4 <MLQ <MLQ <MLQ <MLQ nd nd Sample nd nd <MLQ <MLQ 9.3 <MLQ nd nd nd Sample ID Concentrations of PFSAs & PFPAs (pg/g) PFBS PFHxS PFOS (lin/ branched) PFDS PFHxPA PFOPA PFDPA Sample <MLQ (64/36) <MLQ nd nd nd Sample <MLQ (8/) nd nd nd nd Sample 3 <MLQ <MLQ 6.7 (7/6) nd nd nd nd Sample 4 <MLQ <MLQ 4. (8/8) nd nd nd nd Sample <MLQ <MLQ.3 (64/36) nd nd nd nd Sample 6 <MLQ <MLQ 8.9 (73/7) nd nd nd nd Sample <MLQ 3.43 (7/) nd nd nd <MLQ Sample 8 <MLQ <MLQ <MLQ nd nd nd Sample 9 <MLQ <MLQ.67 (7/8) nd nd nd <MLQ Sample <MLQ <MLQ.7 (73/7) nd nd nd nd Conclusions This study includes to describe the use of -MP in HPLC-MS analysis of perfluorinated surfactants resulting in significantly better chromatography and increased sensitivity. to monitor PFPAs along with PFCAs and PFSAs in food cauldron. PFPA contamination in food duplicates seems not to be a major problem. Figure : Extracted mass chromatograms of (A) perfluoroalkyl phosphonates (PFPAs), (B) perfluoroalkyl sulfonates (PFSAs) and (C) perfluoroalkyl carboxylates (PFCAs) spiked at. ng/g to a baby food sample. References Vestergren R, Cousins IT, Trudel D, Wormuth M, Scheringer M. (8); Chemosphere, 73, Ericson I, Marti-Cid R, Nadal M, Van Bavel B, Lindstrom G, Domingo JL. (8); J. Agric. Food Chem. 6, Ostertag SK, Chan HM, Moisey J, Dabeka R, Tittlemier SA. (9); J. Agric. Food Chem. 7, (min) Acknowledgements The authors gratefully acknowledge the European Union for the financial support through the PERFOOD project (KBBE-7), Anne-Sofie Kärsrud for preparing the baby food cauldron, Martin Schlummer and Hermann Fromme for sharing the food duplicate samples. Shahid.ullah@itm.su.se

4 A matrix effect-free method for ultra trace analysis of perfluoroalkyl acids in dietary samples Robin Vestergren, Shahid Ullah, Ian T. Cousins and Urs Berger Department of Applied Environmental Science (ITM), Stockholm University Introduction Perfluoroalkyl acids (PFAAs) are a class of highly persistent and potentially toxic compounds that have been found ubiquitously in human sera throughout the industrialized world. Exposure modeling studies have indicated that human exposure of the background population occurs primarily via dietary intake. In order to better constrain exposure estimates there is a need for lower detection limits and improved accuracy in quantification of PFAAs in dietary samples. Here, we present an ultra sensitive, matrix effect-free analytical methodology for accurate quantification of C6-C perfluorinated carboxylates (PFCAs) as well as C6 and C8 perfluorinated sulfonates (PFSAs) in complex food mixtures. Analytical method Homogenized food samples (, and g respectively) were extracted in methyl tert-butyl ether (MTBE) using an ion-pair reagent solution (tetrabutylamonium hydrogen sulfate) adjusted to ph ). Enrichment and clean-up was achieved on a Florisil and graphitized carbon solid phase extraction (SPE) column. After conditioning and loading, the column was washed with MTBE and eluted with a 3/7 mixture of MeOH/MTBE. Extracts were kept in a freezer overnight and centrifuged prior to analysis. Analysis was performed on a UPLC separation module (Acquity) coupled to a triple quadrople MS instrument (Xevo-TQS) supplied by Waters Corporation. %response of spiked extract Evaluation of matrix effects Duplicate diet Baby food Vegetable - Recovery 6 Recovery % Duplicate diet baby food vegetable meat fish %response of spiked extract Meat Fish - Figure. Total method recoveries of 3 C-labelled PFCAs (C6, C7-C) and PFSAs (C6 and C8) Figure 3a and 3b. Observed response in prepared food extracts compared to a standard solution prepared in MeOH. % corresponds to the response of a pure standard solution. log peak area Method linearity M y =.337x +.6 R² =.998 PFAA R MPFHxA.999 MPFHxS.996 M.998 MPFOS.988 MPFNA.997 MPFDA.998 MPFUnA MPFDoA.974 log [M] Figure. Whole method linearity of a duplicate diet composite sample spiked at five different concentrations ranging from to pg/g. Detection and quantification MDLs and MQLs were defined from the quantifiable concentrations in procedural blank samples (n=). MDL=mean(blank)+3St.dev.(blank) MLQ=mean(blank)+St.dev.(blank) Table. Method detection limit (MDL) and method limit of quantification (MLQ) of of PFCAs (C6, C7-C) and PFSAs (C6 and C8) in baby food, vegetable, fish and meat samples PFHxA PFHpA PFHxS PFOS PFNA PFDcA PFUnA PFDoA Baby food and vegetables ( g sample) MDL (pg/g) MLQ (pg/g) Fish and meat (. g sample) MDL (pg/g) MLQ (pg/g) Table. Concentrations (pg/g) of PFCAs (C6, C7-C) and PFSAs (C6 and C8) in baby food, vegetable, fish and meat samples Mean concentration (pg/g)± standard deviation PFHxA PFHpA PFHxS PFOS PFNA PFDcA PFUnA PFDoA PFTriA Duplicate diet (n=3) ±3 ± ± 9± ±4 n.d. n.d. n.d. n.d. n.d. Baby food (n=3) n.d. 9± 37± 8± 8±4 ± n.d. ±. n.d. n.d. Vegetable (n=3) 3±7 9±3 9± 8± 6± n.d. n.d. n.d. n.d. n.d. Fish (n=3) n.d. 4± 39±8 8±3 77±8 ± 9± 3±3 ± 48± Meat (n=3) ±3 9±6 8±3 4± ±7 7±4 8±3 3±.4 3±.4 n.d. Conclusions Good method recovery (-78%) was obtained for all analytes (Figure ). Matrix effects on the ionization efficiency were successfully removed in duplicate diet, baby food and vegetable samples (Figure 3a). Enhancement of the ionization efficiency was observed for meat and fish samples (Figure 3b). Method detection limits (MDLs) and limits of quantification (MLQs) were in the low pg/g range for all analytes (Table ). 68% of the analytes were positively detected in the test matrices. PFOS, PFHxS and long-chain PFCAs were the dominant homologues in food samples of animal origin. PFHxS, PFOS and short-chain PFCAs were the dominant homologues in water rich food composites of mixed origin. Acknowledgements The authors gratefully acknowledge the financial support from E.I. DuPont de Nemours & Co., Inc. (unrestricted research grant) and PERFOOD project (KBBE- 7) robin.vestergren@itm.su.se

5 Average Recovery % Average Recovery % PFAS - Comparison of MS/MS and MS-TOF techniques & two extraction methods on lean and fatty fish Sandra Huber, Ondrej Lacina, Petra Hradkova, Jana Pulkrabova, Dorte Herzke, Roland Kallenborn 3 and Jana Hajslova Norwegian Institute for Air Research (NILU), FRAM Centre, Hjalmar Johansens gate 4, NO-996 Tromsø, Norway Institute of Chemical Technology, Prague, Department of Food Chemistry and Analysis, Technicka 3, 66 8 Prague 6, Czech Republic 3 Norwegian Institute for Air Research (NILU), Instituttveien 8, NO-7 Kjeller, Norway INTRODUCTION Expanding monitoring activities as well as advances in available instrumentation have resulted in the detection of various xenobiotics in the human environment which have been escaping attention for decades. Perfluoroalkylated substances (PFAS) represent one group of emerging contaminants which are of high concern. They are generally persistent in the environment, they can be found over a broad concentration range and within the most parts of the food web in both aquatic and terrestrial organisms. Human food items, produced from natural ingredients (wild or farmed), is likely to be contaminated with PFAS as well, giving rise to human exposure. In terms of monitoring the food contamination, most European countries, as Czech Rep. and Norway, carry out national monitoring programs in order to access the daily intake of persistent organic pollutants. To date, only very few international studies focused on PFAS in food and the assessment of dietary intake has been published in Europe. As examples for highly consumed lean and fatty fish species, trout and salmon filet was analysed comparing two fast extraction methods and two detection techniques (low resolution MS/MS and high resolution TOF-MS).To compare the effect of matrix on the quantification results, solvent based and matrix based standards were applied. COMPOUNDS Analytes: 8 different PFAS substances: - carboxylates (C 4 C 4 ) - sulfonates (C 4, C 6, C 8, C ) - perfluorooctane sulfonamid - N-alkylsulfonamides (Me/EtFOSA) Internal standards: 3 C 4 - and 3 C 4 -PFOS Recovery standard: brpfdca Solvent and matrix-matched standard calibration curves:.,.,.,.,.,.,,.,,,, pg/µl MATERIAL Lean fish Trout Fatty fish - Salmon SAMPLE PREPARATION EXTRACTION (6 ml methanol) CLEAN-UP (34 mg activated charcoal) FILTRATION (. µm centrifuge filter) SAMPLE ( g fish material) ANALYSIS UPLC-MS/MS or TOF-MS EXTRACTION (8 ml acetonitrile) VOLUME REDUCTION (to 4 ml) CLEAN-UP (ml + mg ENVI-carb + µl glacial acetic acid) ANALYSES LC-system: Waters Acquity UPLC Column: Waters HSS T3 (.mm,.8 µm) Column oven temperature: 4ºC Sample temperature: ºC Injection volume: µl Mobile phase: A methanol and B mm NH 4 OAc in water Gradient: initial.3 ml/min % A;. min, 4% A; 7 min,.4 ml/min, % A; min,.7 ml/min, % A;. min,.4 ml/min, % A. Ionisation mode: ESI negative Low resolution MS-system: AB Sciex Q-TRAP tandem mass spectrometer with a Turbo V TM ion source High resolution MS-system: Waters LCT premiere XE high resolution time-of-flight mass spectrometer with a Z- spray ion source ACCURACY AND REPEATABILITY OF THE SAMPLE PREPARATION METHODS Spike concentration: ng/g fish fillet (N = 6) Quantification with solvent (SS) and matrix matched standards (MMS) Analysed with the MS/MS instrument EtFOSA MeFOSA FOSA PFBS PFHxS PFOS PFDS PFBA PFPeA PFHxA PFHpA PFNA PFDA PFUdA PFDoA PFTrDA PFTeDA ICT Trout MMS NILU Trout MMS ICT Salmon MMS NILU Salmon MMS ICT Trout SS NILU Trout SS ICT Salmon SS NILU Salmon SS INSTRUMENT LINEARITY AND ACCURACY OF SELECTED STANDARD SOLUTIONS MS/MS TOF MS/MS TOF MS/MS TOF MS/MS TOF MS/MS TOF MS/MS TOF MS/MS TOF MS/MS TOF MS/MS TOF MS/MS TOF MS/MS TOF MS/MS TOF solvent standards MM Trout ICT MM Salmon ICT solvent standards MM Trout ICT MM Salmon ICT solvent standards MM Trout ICT MM Salmon ICT solvent standards MM Trout ICT MM Salmon ICT. ng/ml. ng/ml. ng/ml. ng/ml PFOSA N-Me-FOSA N-Et-FOSA PFBS PFHxS PFOS PFDcS PFBA PFPA PFHxA PFHpA PFNA PFDcA PFUnA PFDoA PFTriA PFTeA Regression coefficient Linear range (ng/ml) Solvent standard MMS Trout ICT MMS Salmon ICT Solvent standard MMS Trout ICT MMS Salmon ICT MS/MS TOF MS/MS TOF MS/MS TOF MS/MS TOF MS/MS TOF MS/MS TOF PFOSA N-Me-FOSA N-Et-FOSA PFBS PFHxS PFOS PFDcS PFBA PFPA PFHxA PFHpA PFNA PFDcA PFUnA PFDoA PFTriA PFTeA LOQ = lowest calibration level At very low concentration levels the instrument accuracy is poor at both instruments, even if the levels are within the linear range of the instrument Better selectivity for low chain PFCAs on the TOF instrument (only transition available for detection) CONCLUSIONS Simple high throuput sample preparation methods are suitable for lean and fatty fish filets ( samples per hours) Use of solvent standards overestimates longer chain PFCAs (from C -PFCA), therefore the application of matrix matched standards is recommended if only 3 C 4 - and 3 C 4 -PFOS are used as internal standards In most cases LOQs are lower when using the MS/MS instrument The MS/MS instrument shows slightly wider linear ranges and slightly better regression coefficients compared to the TOF-MS Decreased linearity for PFSAs in TOF measurements but only little effect on the PFCAs Fatty fish samples cause lower analyte signal in MS/MS spray issue?

6 The PERFOOD validation program for analytical methods for PFASs in food and drinking water incl. interlaboratory study Stefan P.J. van Leeuwen, Kees Swart, Ike van der Veen Introduction Limited data on PFASs in food is available. More data is needed to carry out exposure assessments and ultimately, risk assessments. Analytical methods need to be developed for that purpose. This is challenging because of (i) low PFASs levels in food (often at the fg/g to sub-ng/g level), and (ii) the diversity of matrices (e.g. dairy, vegetables, cereals, meat, fruits and drinking water). Within the EU project PERFOOD, methods have been developed that meet these challenges (i.e. diversity of matrices, low levels). In addition, methods should be simple and easy transferable to interested food analysis laboratories. The compounds of interest are perfluorocarboxylates (PFCAs), perfluorosulfonates (PFSAs) and perfluorophosphonates (PFPAs). Aims The aims of the validation program are:. To test and validate the methods developed within PERFOOD for large scale food, beverage and drinking water monitoring;. To test the performance of laboratories world wide on the analysis of food and drinking water samples; 3. Thereby providing insights in the quality of state-of-the-art methods for food analysis Phase Approach The methods have first been validated in the laboratory where they were developed or implemented. These laboratories then evaluated these methods in terms of accuracy, precision, robustness and sensitivity). The next step is a rigorous -round validation scheme (see scheme below) to which methods are subjected. Targeted PFAS compounds Perfluorocarboxylates (PFCAs), C4 to C4 Perfluorosulfonates (PFSAs), C4, C6, C8 and C, including branched isomers Perfluorophosphonates (PFPAs), C8, C and C PFOSA th International interlaboratory study* Phase Test aims: Instrument calibration, precision, accuracy, robustness Sample types: Solutions with undisclosed concentrations of PFCAs, PFSAs and PFPAs, including interferences. Extract of vegetables and herring PERFOOD validation of methods developed within the project Timeline Phase validation (Perfood partners only) December June Phase validation (open for all interested laboratories) herring June Announcement and invitation 3 July Shipment of samples October Deadline for returning results December Final report Test aims: Accuracy, precision, robustness (inc matrix effects) Sample types: Real food samples including fish tissue, pig liver, drinking water, standard solution with undisclosed PFASs concentrations, vegetables** Production of reference materials Another aim of the validation program is to produce and deliver two reference materials (RMs). The selected matrices to be produced are fish and drinking water. This work is undertaken in collaboration with KWR, Fraunhofer-IVV and the Institute for Reference Materials and Measurements. * The international interlaboratory study will be organised in collaboration with QUASIMEME Laboratory Performance Studies ( ** The vegetable sample is only for Perfood members Acknowledgement : The European Commission is gratefully acknowledged for granting the PERFOOD project. Wellington Laboratories is gratefully acknowledged by providing analytical standards to the PERFOOD consortium. IVM INSTITUTE FOR ENVIRONMENTAL STUDIES VU University Amsterdam De Boelelaan 8 (visiting address) De Boelelaan 87 (postal address) 8 HV Amsterdam t f e info@ivm.vu.nl i

7 ISOTOPE DILUTION TECHNIQUE IN THE ANALYSIS OF PFOS AND IN WATER SAMPLES Sami Huhtala, Noora Perkola, Petra Kosubová ) Finnish Environment Institute, Laboratories, Research and Innovation Laboratory, Helsinki, Finland ) Central Institute for Supervising and Testing in Agriculture, Brno, Czech Republic Abstract The use of isotope dilution mass spectrometry (IDMS) has become more common with the increase of mass spectrometric instrumentation and applications. The usage of isotopic labelled standards, commonly 3 C- or deuterium, enhances the accuracy of measurements, especially in difficult sample matrix. Nonetheless, the nomenclature of internal standard solutions is inconsistent and the purpose of the internal standard not readily apparent. In this study, we present the IDMS procedure to the determination of two most common perfluorinated substances (PFCs) perfluoro octane sulfonate (PFOS) and perfluoro octanoic acid () in water samples. Introduction The PFCs are widely spread to the nature. The attention has focused to their occurence, effects and environmental fate in various environmental compartments thus presenting challeges to analytical work. To study the effect of IDMS three different water samples (Milli-Q water, natural water and waste water) were analysed as four replicate measurements repeated three times on different days. The calculations of and PFOS concentrations are done varying the functions of applied mass labelled internal standards Table. Terms and applications of internals standards Internal standard Surrogate (Other terms: Recovery standard, Extraction standard) Instrument standard (Other terms: Performance standard, Injection standard A compound added to a sample in known concentration to facilitate the qualitative indentification and/or quantitative determination of the sample components (IUPAC) In our study ( 3 C PFHxA, 3 C 4, 3 C 4 PFOS and 3 C PFDA, ng ml - ) To quantificate the analytes To evaluate recovery (extraction, clean up) In our study ( 3 C 8 -PFOS and 3 C 8 -, ng ml - ) To calculate recovery of surrogate standard Materials & methods Analytes were extracted by Oasis HLB SPE cartridges. Prior to SPE, surrogate standards and. ml NH 4 OAc were added to the samples. The analytes were eluted with MeOH. For LC-MS analysis,.3 ml of extract was transferred to a PP vial, and.7 ml of Milli-Q water and instrument standards were added. The diluted extracts were analysed using UPLC-MS/MS. An isolator column was placed before the injector to delay signals originating from the instrument. The analytes were determined with TQ MS (Xevo, Waters) using negative ESI and MRM mode. Results & discussion The applied surrogate standard affects the results. is more sensitive to the used surrogate standard than PFOS. The longer chain mass labelled surrogate standard (PFDA) gave significantly bigger results for both and PFOS than analog or shorter chain (PFHxA) standards. When applying PFOS mass labelled standards for determination of the variation of results were larger than with mass labelled standards. The differences (deviation from reference value, variation of results) were larger at lower concentrations. The more difficult sample matrix increases the differences in the results. The significance of the applied standards is higher when analysing complex matrix like WWTP sludge. The differences in sludge results were 3- times higher than in water sample results when non-identical standards were employed. With correct standards (mass labelled standards of analyte in question) the calculation method doesn t have a significant effect on the results. However, quantification with surrogate standards give readily recovery corrected results. Table. Results of different water samples determined by altering the usage of internal standards PFOS PFOS M4/M8 PFOS M8/M4 M4/M8 M8/M4 PFHxA M / PFOS M8 PFDA M / PFOS M8 MilliQ low, mean (ng/l) (+3%) 4.9 (-6%) 4. (-6%) 4.67 (+8%) 3.6 (-%) t MilliQ low, RSD (ng/l).49 (%).6 (4%).63 (%).9 (%).67 (4%).73 (3%) MilliQ high, mean (ng/l) 89. (+%) 97.(+9%) t 9.3 (+%) t 9. (+6%) (+%) t MilliQ high, RSD (ng/l) 8. (9%) 8.77 (9%) 6.74 (7%) 7.99 (9%).77 (6%) 9.6 (%) River W. low, mean (ng/l) (+%) 4.43 (-%) 4.9 (+3%) 4.64 (+4%) 4.3 (-%) River W. low, RSD (ng/l).9 (%). (3%).9 (3%). (%).7 (%).8 (9%) River W. high, mean (ng/l) (+%) 86.4 (+%) 79.4 (-7%) 87.9 (+%) 4 (+%) t River W. high, RSD (ng/l).9 (4%).7 (3%) 9.8 (%).7 (4%) 7.76 (9%).8 (%) Waste W. dil., mean (ng/l).. (+3%). (-%).4 (+3%). (-%).4 (+3%) Waste W. dil., RSD (ng/l).33 (%).66 (3%).46 (%).73 (4%).99 (8%) 3.96 (3%) Waste W. orig., mean (ng/l) (+3%).3 (-7%) 6.9 (+%).8 (-8%) 73.3 (+3%) Waste W. orig., RSD (ng/l).7 (%).4 (%) 6.79 (3%) 3. (%) 8. (6%) 8.6 (39%) f Sludge, mean (ng/g).3.9 (-9%).9 (-9%).38 (+3%) 7. (+3%) Sludge, mean (ng/g) (-6%) 6. (-%). (+84%) 3.3 (+4%) M4/M8 M8/M4 PFOS M4/M8 PFOS M8 / M4 PFHxA M / M8 PFDA M / M8 Milli-Q low, mean (ng/l) (+6%) 3.4 (±%).78 (-%) 4. (+7%) t.4 (-3%) t Milli-Q low, RSD (ng/l).4 (3%).4 (%).6 (9%). (8%).6 (6%).4 (%) Milli-Q high, mean (ng/l) 4 (-3%) 9. (-8%) t (+6%) t (-%) 7 (+%) Milli-Q high, RSD (%) 4.7 (%) 4.9 (4%) 8.3 (9%) 4.7 (4%) 3.9 (3%) 9. (9%) f River W. low, mean (ng/l) (±%) 8.9 (+%) t 8.7 (±%) 8. (+4%) 9. (+3%) t River W. low, RSD (ng/l). (6%). (7%).8 (3%) f. (6%).6 (7%).97 () f River W. high, mean (ng/l) (+%) (%) (-9%) t (-%) 3 (+9%) t River W. high, RSD (%) 7.7 (7%) 6.7 (6%). (4%) f 6.7 (6%) 9.37 (8%).9 (%) Waste W. dil., mean (ng/l) (+%) 4 (+3%) (+8%) (%) 3 (+%) t Waste W. dil., RSD (ng/l). (%). (%) 6.7 (6%) 3. (%) 6.9 (7%) 7.3 (%) f Waste W. orig., mean (ng/l) (+%) 46 (+9%) 46 (+9%) t 4 (-%) 63 (+48%) t Waste W. orig., RSD (ng/l) 3. (8%) 34.7 (8%) (4%) 37.4 (8%) 34.4 (8%) 3 (49%) f Sludge, mean (ng/g).64.6 (-6%).6 (-%). (+7%).46 (+8%) Sludge, mean (ng/g)..9 (-%).6 (-). (+%).3 (+7%) t = significant difference in mean values (t-test) f = significant difference in deviations (f-test) The attained recoveries were: for 93 ± %, 4 ± 9 % and 8 ± 8 % for Milli-Q water, River water and Municipal waste water for PFOS 94 ± 7 %, 4 ± % and 89 ± % for Milli-Q water, River water and Municipal waste water Within-day variation: of.6 9 % of PFOS. 8 % (depending on concentration and sample matrix) Between-day variation: of.7 4 % of PFOS. 4 % % (depending on concentration and sample matrix) Conclutions The use of isotopically labelled internal standards improves the quality of analytical results in water samples. It is recommended to use isotopically labelled internal standards (analog to the analyte) as surrogates when possible. The means that internal standards are used should be reported when publishing quantitative data to make the data more comparable. The benefits achieved with the correct application of mass labelled internal standards are clear: data handling is simplified, the accuracy and the quality of reported results are increased. References Powley C.R., George S.W., Ryan T.W., Buck R.C., Anal. Chem.,, 77, Tomy G.T., Halldorson T., Danell R., Law K., Arsenault G., Alaee M., MacInnis G., Marvin C.H., RCM,, 9, Acknowledgements The authors wish to thank the association of Maa- ja Vesitekniikan Tuki ry and NordFluor Network for financial support. In addition, Ms. Anne Markkanen is acknowledged for her technical help with the analyses.

8 Koponen Jani M. Ruokojärvi Päivi Kiviranta Hannu contact: Simple and High Throughput Method for Quantitation of Perfluorinated Compounds Using LC-MS/MS Why was this analytical method developed? Instrumentation Perfluoroalkyl substanes (PFASs) are found to be ubiquitous in wildlife and human populations worldwide. Laboratory animal experiments have shown that PFASs are weakly carcinogenic, affect slightly on learning and behaviour, and possess immuno-, developmental and reproductive toxicity. This indicates that these compounds may have negative influence on the human health and welfare. For large-scale epidemiological studies with multitude of samples and limited sample volume simple and high throughput method for analysis of PFASs is needed. The PFASs were analyzed with Dionex Ultimate 3 RSLC coupled with Finnigan TSQ Quantum Discovery Max mass spectrometer LC parameters analytical column: Purospher Star C8 (3 x. mm, µm) trap column: Purospher Star C8 ( x. mm, µm) Sample pre-treatment Simple and fast pre-treatment for serum samples were developed. A. ml or. ml sample was precipitated in neutral and acidic conditions with methanolic ammonium acetate and formic acid, respectively. Prior to LC-MS/MS analysis the samples were filtered through a. µm syringe filter. mobile phase: % aqueous methanol (eluent A) and 9% aqueous methanol with -methylpiperidine (eluent B) eluent program: nonlinear gradient from 3 to % eluent B in minutes injection volume: µl MS parameters Method suitability for large-scale epidemiological studies number of native componds analyzed Perfluorosulfonates Perfluorocarboxylates 9 4 To study the high throughput performance of the validated method, 8 human serum samples were analyzed. The analysis of all the 8 samples was performed from the number of isotopelabelled componds analyzed sample pre-treatment to the quantitated results in four working days demonstrating the polarity negative negative high throughput performance of the method. The main PFASs found in the samples spray voltage - 3. kv - 3. kv were perfluorooctanesulfonate (PFOS) and perfluorooctanoate () ranging from precursor ion [M-H] - [M-H] -.4 to 3 ng/ml and from.3 to 6.9 ng/ml, respectively. Besides the high throughput product ion [M-COOH] - [FSO 3 ] - performance, the developed method is suitable for studies where the sample volume is collision energy V 4 V limited since a low sample volume only from. to. ml serum is needed. Method validation specifications RT = 7.4 A = S/N = 7 SRM = m/z 43/369 M RT = 7.4 A = 99 S/N = 976 SRM = m/z 47/37 PFOS RT = 9. A = 4 S/N = 46 SRM = m/z 499/99 MPFOS RT = 9.3 A = 397 S/N = 4 SRM = m/z 3/99 PFDoA RT =.9 A = 4486 S/N = 7 SRM = m/z 63/69 MPFDoA RT =.99 A = 394 S/N = 3793 SRM = m/z 6/7 Figure. LC-MS/MS chromatograms of perfluorooctanoate (). Perfluorooctanesulfonate (PFOS) and perfluorododecanoate (PFDoA) in. ml serum spiked with. ng of each PFASs. Limit of quantification (LOQ) for. ml serum sample was.3 ng/sample for each analyte. With few exceptions the accuracy at spiked levels of 4 and ng/ml were between 7 %. Coefficient of variation in the inner- and inter-batch precisions varied between 4 7 % and 7 8 %, respectively. Based on the available results from inter-laboratory studies the inter-laboratory reproducibility for PFOS and was highly acceptable. Terveyden ja hyvinvoinnin laitos Institutet för hälsa och välfärd National Institute for Health and Welfare Department of Environmental Health, Chemical Exposure Unit Neulaniementie 4, PL/PB/P.O. Box 9, FI-77 Kuopio, Finland, puh/tel

9 Particle-associated fraction (%) Particle-phase (pg m -3 ) Particle-phase (pg m -3 ) Particle-phase (pg m -3 ) Particle phase (pg m -3 ) Gas-phase (pg m -3 ) Gas-phase (pg m -3 ) Gas-phase (pg m -3 ) Polyfluoroalkyl Compounds in the Atmosphere at a Wastewater Treatment Plant Lena Vierke (Lena.Vierke@uba.de),, Lutz Ahrens 3, Mahiba Shoeib 3, Eric J. Reiner 4, Rui Guo, Wolf-Ulrich Palm, Tom Harner 3, Ralf Ebinghaus 6 Federal Environment Agency, Germany Leuphana University of Lüneburg, Germany 3 Environment Canada, Canada 4 Ontario Ministry of the Environment, Canada University of Toronto, Canada 6 Helmholtz-Zentrum Geesthacht, Germany Introduction Aim of the study Transport pathways and sources for per- and polyfluorinated compounds (PFCs) in the environment are not fully understood Long-range transport of PFCs occurs via ocean currents and in the atmosphere Wastewater treatment plants (WWTPs) are a source for PFCs into the atmosphere, Investigation of PFC emissions from WWTPs into the atmosphere Comparison of PFC concentrations above the aeration tank (aeration might cause aerosol-mediated transfer into the atmosphere) and secondary clarifier (no aeration, calm surface) Calculation of the partitioning of PFCs between gas- and particlephase Sampling Extraction and Analysis Passive air sampling (6 weeks) sorbent-impregnated PUF (SIP) disks Aeration Tank Secondary Clarifier High volume air sampling ( 4 h, ~4 m 3 ) GFFs and PUF/XAD/PUF cartridges Atmospheric Concentrations Fig. WWTP in Toronto, Canada, in Spring Fluorotelomer alcohols (FTOHs), perfluorooctane sulfonamides (FOSAs), perfluorooctane sulfonamidoethanols (FOSEs) Perfluorooalkyl carboxylates (PFCAs), perfluorooalkyl sulfonates (PFSAs) Particle-phase Sonication Dichlormethane Methanol Gas-phase Soxhlet Petroleum ether Methanol GC-PCI-MS LC-( )ESI-MS/MS A Aeration tank Secondary clarifier Fig. (A) FTOHs, FOSAs, FOSEs, (B) PFSAs, PFCAs concentrations for the gas- (top) and particle- (bottom) phase from high volume air sampling (n = ) Concentrations at the aeration tank were 9 times higher compared to the secondary clarifier Results from active and passive air sampling showed good agreement 3 No correlations of PFC concentrations with B Aeration tank Secondary clarifier sampling parameter were found (i.e. T air, OC particles ) Gas-Particle Partitioning Conclusion PFSAs PFCAs Fig. 3 Particle associated fraction in % (n = 4) Particle-associated fraction increases with increasing chain length for PFCAs Aeration process causes elevated atmospheric gas- and particlephase concentrations which might be caused by aerosol-mediated transport PFCAs and PFSAs are mainly particle-associated but also present in the gas-phase Direct atmospheric transport of PFCAs and PFSAs (in addition to their precursors FTOHs, FOSAs and FOSEs) should be considered as a possibility for long-range transport of PFCs Influence of sampling artefacts needs to be further investigated References Weinberg I, Dreyer A, Ebinghaus R () Environ. Pollut. 9: -3. Ahrens L, Shoeib M, Harner T, Lee S C, Guo R, Reiner E J () Environ. Sci. Technol. In press. 3 Vierke L, Ahrens L, Shoeib M, Reiner EJ, Guo R, Palm WU, Ebinghaus R, Harner T () Environm. Chem. In press. Environment Canada Environnement Canada

10 Physicochemical Properties of Poly- and Perfluorinated Alkyl Substances (PFAS) Calculated Using COSMOtherm [ ] Zhanyun Wang¹, Matthew MacLeod², Ian T. Cousins², Martin Scheringer¹, Konrad Hungerbühler¹ ¹ Institute for Chemical and Bioengineering, Swiss Federal Institute of Technology (ETH), CH-893 Zurich, Switzerland ² Department of Applied Environmental Science (ITM), Stockholm University, SE-69 Stockholm, Sweden Contact: zhanyun.wang@chem.ethz.ch Introduction Methods Results and Discussions 4: FTOH. Model evaluation against experimental data of FTOHs and some other PFAS Estimated properties show good self-consistency (e.g. log K OW = log K OA + log K AW ) and agree well with Within a homologous series of PFAS with the same functional group, both K AW and K OW increase with F T polarity of the functional group, leading to lower K AW carbon chain length and functional group on K -6-6 AW and K OW. A comparison substances is shown on Wang, Z. et al.,. Using COSMOtherm to Predict the left; a comparison 9 9 nated 7 Alkyl Substances X = CnFn+ (PFAS). 7 Environ. Chem. alkyl substances is 8 8 PFFs, X-F 6 6 X = PFA Is, X-I CnFn+ shown on the right. PFCAs, X -COOH PFCAs, X -COOH PFSA s, X-SO3H FTCAs, X-CHCOOH 4 PFSiA s, X-SOH 4 PFA Is, X-I FA SA s, X-SONH FTIs, X-CHCHI 3 FA SEs, X-SONHCHCHOH 3 PFSA s, X-SO3H PFPAs, X-PO3H FTSA s, X-CHCHSO3H PFPiA s, X-POH-Y XFTSA, X-CHSO3H FTOHs, X-CHCHOH Log K AW Log K OW perfl uorinated chain length 3. Branched isomers versus linear isomers perfl uorinated chain length Log K AW Log K OW Log K Log S OA W [mol/l] Log P L [Pa]. PFBA PFHpA PFBA 7. PFHpA 4. PFPeA PFPeA 3. PFH A PFH A -. PFHpA PFH A PFH A. PFH A PFPeA PFPeA PFHpA PFHpA PFPeA PFBA PFBA. Environ. 3. PFBA Chem. ranched isomers The partitioning behavior of all theoretically possible branched isomers can vary considerably. Wang, Z. et al.,. Using COSMOtherm to Predict Physicochemical linear isomers linear isomers linear isomers linear isomers linear isomers * K AW K OW - octanol-water K OA y monomethyl + monoethyl ( ) ) of PFCAs (C4-C8) versus modeled property data only for the linear isomer of the same PFCAs (x Conclusions. application to a large set of PFAS (3 individual chemicals).. The property estimates from this work provide a useful basis for further environmental modelling and an insight into the relationship between chemical structure and physicochemical properties of PFAS. References.. Limitations & Outlooks Our estimates represent only the partitioning behavior of neutral monomeric PFAS. To fully understand the overall partitioning behavior of PFAS in reality, further studies on the surfactant-like nature of PFAS, the acid dissociation Acknowledgement Environ. Chem., in press. Environ. Chem., in press. [iii] Arp et al., 6. Predicting the Partitioning Behavior of Various Highly Fluorinated Compounds. Environ. Sci. Technol. 4, this research at ETH Zurich.

11 , 4 and 3 data otained in this work..33, based om the data otained in this work (4..4) (4..4) T (Co) Sevoflurane E/R = 9 A =.8 x k (cm molecule s ) This work Brown et al. 989 Langbein et al. 999 Relative rate data, this work. DeMore /T (K-) 4. 4.

12 m C O I N P S E T N I H T A U G T E N O C F E (I N N T S E E R R F T O N R A A M T E M ) O S P H E R I C R E S E A R C H D E U N P I A V R E R T S M I T E Y N O T F C O O F P E C N H A E G M E NI S T R Y U N I V E R S I T Y O F C O P E N H A G E N The chlorofluorocarbons (CFCs) have long been recognized to be a major participant in chlorine initiated photochemical reactions depleting the stratospheric ozone layer. [,] Furthermore these compounds are green house gases (GHGs) and contribute to the global warming. Because of this, there has been a substantial effort to find environmentally acceptable alternatives. The first generation CFC alternatives contained hydrogen atoms (HCFCs), while the second generation alternatives do not contain chlorine atoms (HFCs). Third generation CFC alternatives are the so called hydrofluoroolefins (HFOs) and hydrofluoroethers (HFEs). Atmospheric chemistry of CF 3 CH OCH 3 Freja From Østerstrøm, Ole John Nielsen, Mads P. Sulbæk Andersen Copenhagen Center for Atmospheric Research, Department of Chemistry, University of Copenhagen, Universitetsparken, DK Copenhagen Ø, Denmark Jet Propulsion Laboratory, California Institute of Technology, 48 Oak Grove Drive, Mail Stop 83 9, Pasadena, CA 99, USA. Introduction. Experimental: Relative rate (OH/Cl + CF 3 CH OCH 3 ) and product study This study concerns the atmospheric chemistry HFEs. Like the CFCs and other replacement compounds, the HFEs are used in applications such as electronics cleaning, fire extinguishing, lubricant deposition, heat transfer and refrigeration. To be able to use the HFEs, the environmental chemistry must be known. Thus, in this study we want to determine the reactivities of Cl atoms and OH radicals towards CF 3 CH OCH as well as the atmospheric degradation products, the atmospheric lifetime, the IR spectrum and the global warming potential for the title compound, CF 3 CH OCH, as a representative for the HFE class of compounds. This study was conducted using the smog chamber at Ford Motor Company. All experiments were performed in a 4 litre Pyrex reactor connected to a Mattson Sirus FTIR spectrometer. [3,4] Twenty two black light lamps surrounded the reactor (GE F4BLB) and were used to photochemically initiate the experiments. Cl atoms were produced by photolysis of Cl, and OH radicals by photolysis of CH 3 ONO in the presence of NO in air. All experiments were performed at 96 ± K. Hydrogen atom abstraction can occur at the CH and CH 3 groups, respectively. For reaction with Cl atoms there is (±) % and (79±3) % at the CH and CH 3 groups, respectively. For reaction with OH radicals there is (4±3)% and (7±) % at the CH and CH 3 groups, respectively. Infrared spectrum of CF 3 CH OCH 3. CH 3 ONO + hv CH 3 O + NO CH 3 O + O HO + HCHO HO + NO OH + NO Cl + hv Cl The laboratory at Ford Motor Company. 3. Atmospheric lifetime With the measured rate constant for the reaction of CF 3 CH OCH 3 with OH radicals, k OH = (4.66 ±.4) 3 cm 3 molecules s, and the global weighted average concentration of OH radicals estimated to [OH] = 6 molecules cm 3 [] we can calculate the global atmospheric lifetime of CF 3 CH OCH 3 to days. The local concentrations of OH can vary a lot, so the local lifetime of CF 3 CH OCH 3 might be different from the one given here. The integrated absorption cross section (6 6 cm ) is (. ±.8) 6 cm molecule. The radiative efficiency of CF 3 CH OCH 3 is then. W m ppb and the AGWPs for CO are estimated by the World Meteorological Organization (WMO) to.97,.676 and.3 W m ppm [6] for the, and year horizons respectively. The following formula is used to calculate GWP: GWP x 4. GWP τ x τx F (t) e F x CO (t) R(t)dt We use the above calculated atmospheric lifetime of days to calculate the GWP. The GWP of CF 3 CH OCH 3 is determined to be 9, 8 and for the, and year time horizons, respectively. The determined value of GWP for CF 3 CH OCH 3 is a factor of thousand smaller than the GWP value for CFC for the year time horizon. [7] The climate effect of CF 3 CH OCH 3 is insignificant. t τx dt. Data for fluorinated ethers and reference compounds Compound k Cl k OH GWP Reference CF 3 CH OCH 3 (.8 ±.3) (4.66 ±.4) 3 8 This work CF 3 CH OCH 3 (.8 ±.9) (.7 ±.8) 3 Not calculated Oyaro et al. [8] C F CH OCH 3 (. ±.37) (.78 ±.) 3 6 Thomsen et al. [9] C F CH OCH 3 (4. ±.8) (6.4 ±.33) 3 Oyaro et al. [] n C 4 F 9 CH OCH 3 (9.7 ±.4) Wallington et al. [] CFC, CFCl 3 47 WMO [7] All rate constants given in the table are in units of cm 3 molecule s. 6. Conclusion The fluorinated ether studied in the present work will not contribute significantly to the radiative forcing of climate change. 7. References : M. J. Molina and F. S. Rowland, Nature, 974, 49, 8 : J. D. Farman, B. G. Gardiner and J. D. Shanklin, Nature, 98, 3, 7 3: T. J. Wallington and S. M. Japar, J. Atmos. Chem., 989, 9, 399 4: J. Sehested, T. Ellermann, O. J. Nielsen, T. J. Wallington and M. D. Hurley, Int. J. Chem. Kinet., 993,, 7 : R. G. Prinn, J. Huang, R. F. Weiss, D. M. Cunnold, P. J. Fraser, P. G. Simmonds, A. McCulloch, P. Salameh, S. O Doherty, R. H. J. Wang, L. Porter and B. R. Miller, Science,, 9, : World Meteorological Organization (WMO). Scientific Assessment of Ozone Depletion: 6, Global Ozone, Research and Monitoring Project Report, Geneva, Switzerland, 7 7: WMO, Scientific Assessment of Ozone Depletion:, Global Ozone Research and Monitoring Project Report No., 6 pp., Geneva, Switzerland, 8: N. Oyaro, S. R. Sellevåg and C. J. Nielsen, J. Phys. Chem. A,, 9, , 9: D. L. Thomsen, V. F. Andersen, O. J. Nielsen and T. J. Wallington, Phys. Chem. Chem. Phys.,, 3, : N. Oyaro, S. R. Sellevåg and C. J. Nielsen, Environ. Sci. Technol., 4, 38, : T. J. Wallington, W. F. Schneider, J. Sehested, M. Bilde, J. Platz, O. J. Nielsen, L. K. Christensen, M. J. Molina, L. T. Molina and P. W. Wooldridge, J. Phys. Chem. A, 997,, Acknowledgements: We acknowledge financial support from the Danish Natural Science Research Council, the Villum Kann Rasmussen Foundation and EUROCHAMP. This work was performed partly at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration. M.P.S.A. is supported by an appointment to the NASA Postdoctoral Program, administered by Oak Ridge Associated Universities through a contract with NASA.

13 DEVELOPMENT OF A METHOD TO COMBINE THE EXTRACTION OF BOTH PFCs AND LEGACY-POPs FROM HUMAN SERUM Christian Bjerregaard Olesen, B.Sc., and Eva Cecilie Bonefeld-Jørgensen, Prof., Dir. Centre for Arctic Environmental Medicine & Unit of Cellular and Molecular Toxicology School of Public Health, Aarhus University, Denmark. Abstract Background: Animal studies have shown that perfluorinated compounds (PFCs) are easily absorbed by oral intake, inhalation and to a smaller extent through dermal contact. The elimination of the PFCs is in general considered very slow with average human serum half-lives of 8.8 years for PFHxS,.4 years for PFOS and 3.8 years for. However, the PFCs are not bioaccumulated in the fatty tissue as most of the legacy persistent organic pollutants (legacy-pops). Instead the PFCs are distributed to organs such as liver and kidney through the bloodstream. In vitro studies have suggested that PFCs, as legacy-pops, also affect the sex hormone receptor functions. The aim of the present study is to develop a method to combine the extraction of both PFCs and legacy-pops from human serum. This method will be used to determine ex vivo the total mixture effect on sex hormone receptor functions. Methods: A method for extracting legacy-pops from human serum was developed by Philip S. Hjelmborg et al. []. This method includes a solid phase extraction (SPE), and a solvent extraction followed by normal-phase high-performance liquid chromatography (NP- HPLC). The method was tested and modified with the aim for application to extract PFCs as well as legacy-pops in the same SPE- HPLC run. Results: The SPE-HPLC legacy-pop method in its original form was found not to fit for extraction of PFCs. However, the SPE-HPLC method showed promising results by an omission of the solvent extraction after SPE, before HPLC. Further adjustments of the HPLC method are required. Materials & Methods Removal of unwanted interferences originating from the serum matrix: Solid phase extraction (SPE) was used to extract xenoestrogens from human Figure : Solid phase extraction method serum. The OASIS HLB extraction cartridges (Waters, Milford, MA, US) were placed in a vacuum manifold (Varian, Harbor City, CA, US) for easy collection of waste. Serum was introduced to the preconditioned cartridges, and after removal of the serum matrix by water, the lipophilic compounds were eluted by methanol and ethyl acetate (see figure ). A solvent extraction with hexane was used to remove any interfering hydrophilic compounds left in the SPE extract. (This step was omitted for the PFCs). Cleanup of the SPE-extract using HPLC for the removal of endogenous hormones: The SPE-extract was injected onto a Si-6, m x 4.6 mm normal phase HPLC column operated at C at a flow rate of. ml/min. Most legacy-pops are more lipophilic than the natural hormones. With this HPLC method, the legacy-pops elute within the first. minutes (see figure ). Pregnenolone is the first natural hormone to elute, at.8 minutes. The legacy-pops are collected in the first fraction (F), and the natural hormones are collected in later fractions. The effect on sex hormone receptors can then be tested for each of the fractions. Figure : HPLC method and elution profile. From []. Conclusion: The SPE-HPLC method for legacy-pop serum extraction was not applicable to include PFC extraction. Minor adjustments of the method may however have a positive influence on the extraction and recovery of PFCs in the same SPE-HPLC run. n= Table : PFC recoveries in the SPE methanol extract PFHxS PFOS PFNA PFDA PFUnA Average 4. % 48. % 47.7 %.9 % 47.7 %.3 % 48. % 48.4 % 44. % 49. % (.6 ng/ml) (46.7 ng/ml) (4.9 ng/ml) (3. ng/ml) (9. ng/ml) (3.4 ng/ml) (7. ng/ml) (.7 ng/ml) (.8 ng/ml) 4.3 % (8.4 ng/ml) 7.6 % (6.7 ng/ml) 97. % (. ng/ml) >. % (.6 ng/ml) >. % (8.4 ng/ml) 9. % (9.6 ng/ml) PFDoA >. % (.6 ng/ml) 93.7 % (. ng/ml) >. % (4. ng/ml) >. % (. ng/ml) >. % (.6 ng/ml) >. % (3.6 ng/ml) >. % (4. ng/ml) 79.8 % (49.4 ng/ml) >. % (4.4 ng/ml) >. % (44. ng/ml) >. % (46.7 ng/ml) 86.7 % (47.7 ng/ml) >. % (4. ng/ml) 83. % (43. ng/ml) >. % (3.4 ng/ml) >. % (38.9 ng/ml) >. % (4.9 ng/ml) 97. % (4.7 ng/ml) >. % (3.4 ng/ml) 73. % (37. ng/ml) >. % (3.3 ng/ml) >. % (33.4 ng/ml) >. % (3. ng/ml) 98.6 % (3.8 ng/ml) 97.9 % (3.4 ng/ml) 84. % (3.9 ng/ml) >. % (.3 ng/ml) >. % (7.8 ng/ml) >. % (9. ng/ml) >. % (9.8 ng/ml) 99.9 % (.3 ng/ml) 78. % (4.7 ng/ml) >. % (. ng/ml) >. % (. ng/ml) >. % (3.4 ng/ml) >. % (3.9 ng/ml) >. % (. ng/ml) 8. % (8. ng/ml) >. % (. ng/ml) >. % (6.7 ng/ml) >. % (7. ng/ml). % (7.8 ng/ml) >. % (. ng/ml) 87. % (.3 ng/ml) >. % (. ng/ml) >. % (. ng/ml) >. % (.7 ng/ml) 4. % (.9 ng/ml) >. % (. ng/ml) n= Table : PFC recoveries in HPLC fraction of the SPE methanol phase PFHxS PFOS PFNA PFDA PFUnA PFDoA 63. % (6. ng/ml) >. % (. ng/ml) >. % (.6 ng/ml) 44.8 % (.8 ng/ml) 4.8 % (6. ng/ml) 8.9 % (. ng/ml) 8. % 99.7 %. % 94. % 8. % 98. % Average.% (6.4 ng/ml) 6.% (6.ng/mL) 4.7% (.3ng/mL) 6.9% (7.ng/mL) 7.3% (.4ng/mL).8% (.4ng/mL).6% (.9ng/mL).% (6.9 ng/ml).% (7.7ng/mL) 6.3% (.ng/ml).% (8.7ng/mL) 4.% (3.3ng/mL) 7.7% (.4ng/mL).% (.6ng/mL).% (.6 ng/ml) 7.% (.ng/ml) 43.% (8.7 ng/ml) 9.% (3.4ng/mL) 3.7% (9.9 ng/ml).% (8.7 ng/ml) 7.% (9. ng/ml).% (. ng/ml) 7.3% (4.3ng/mL).7% (7.9 ng/ml) 9.% (.3 ng/ml) 6.% (8.7 ng/ml).% (8. ng/ml) 8.% (7. ng/ml) 7.9% (4.6 ng/ml) 8.3% (.ng/ml) 47.8% (6. ng/ml) 7.8% (.7ng/mL) 3.9% (7.3 ng/ml) 4.3% (7.4 ng/ml).8% (6. ng/ml).8% (3.8 ng/ml) 4.% (.3ng/mL) 48.% (.9 ng/ml) 3.9% (9. ng/ml) 6.4% (6.3 ng/ml) 4.% (6.4 ng/ml) 8.9% (.3 ng/ml) 6.4% (3.6 ng/ml).7% (.9ng/mL) 6.% (.4 ng/ml) 3.9% (8.6 ng/ml) 8.4% (.9 ng/ml).% (6.7 ng/ml).% (4.6 ng/ml) 8.% (.6 ng/ml).7% (.ng/ml) 86.9% (4. ng/ml) 34.% (6. ng/ml) 8.% (4. ng/ml).% (. ng/ml) 9.% (3.4 ng/ml) 4.% (. ng/ml) 4.% (9. ng/ml) >.% (.9 ng/ml) 36.% (4. ng/ml).% (. ng/ml) 4.4% (.4 ng/ml) 8.6% (.4 ng/ml).4% (. ng/ml) 3.8% (7.7 ng/ml) >.% (.7 ng/ml) 9.% (3. ng/ml) 7.9% (. ng/ml).9% (.6 ng/ml) 3.6% (.9 ng/ml).% 3.4% 8.4% 7.% 3.6%.% 9.% Results & Discussion Solid phase extraction of spiked human serum: The SPE-HPLC method descibed above, was tested for its application to extract PFCs from serum. Except for PFHxS, the SPE method showed good results for the extraction of PFCs (see figure and table ). The PFCs were lost in the hexane extraction following the SPE. The extraction step was omitted due to the low PFC recoveries. Unfortunately, the evaporation of MeOH is slow. Cleanup of the SPE-extract using HPLC: After evaporation, the SPE extract was ready for injection onto the HPLC normal phase column. The recovery in HPLC fraction (F) was acceptable, but the recovery for the rest of the PFCs were too low (see figure and table ). Some tests with PFC standards in MeOH show that an extention of the F time-interval can improve the recovery, significantly. Addition of formic acid and/or ammonium acetate to the SPE extracts may improve the recoveries. The retention times and fraction of elution for the legacy-pops and natural hormones should be determined for the changed method. Conclusion & Perspectives The recoveries of most of the PFCs in the SPE extract was excellent. PFHxS had a low but acceptable SPE recovery. The hexane extraction upon SPE must be omitted, because most of the PFCs are lost in this step. The evaporation of methanol in the SPE extract is very slow. Another more volatile eluent is being considered. The HPLC method showed low PFC recoveries in F. Some minor adjustments like extending the time-interval of F and using additives like formic acid or ammonium acetate might have a positive effect on the elution profile and recoveries. The retention times and fraction of elution of the legacy-pops and the natural hormones must be determined again, because the solvent in the extract is now methanol instead of hexane.. Hjelmborg, P. S., Ghisari, M., and Bonefeld-Jorgensen, E. C., SPE-HPLC purification of endocrine-disrupting compounds from human serum for assessment of xenoestrogenic activity. Anal. Bioanal. Chem., 6. 38(): p