Gas chromatography triple quadrupole tandem mass...

RESEARCH PAPER

Gas chromatography–triple quadrupole tandem massspectrometry: a powerful tool for the (ultra)trace analysisof multiclass environmental contaminants in fish and fish feed

Kamila Kalachova & Jana Pulkrabova & Tomas Cajka &

Lucie Drabova & Michal Stupak & Jana Hajslova

Received: 31 January 2013 /Revised: 12 April 2013 /Accepted: 17 April 2013 /Published online: 10 May 2013# Springer-Verlag Berlin Heidelberg 2013

Abstract A new method for rapid determination of 73 targetorganic environmental contaminants including 18 polychlori-nated biphenyls, 16 organochlorinated pesticides, 14 bromi-nated flame retardants and 25 polycyclic aromatichydrocarbons in fish and fish feed using gas chromatographycoupled with triple quadrupole tandem mass spectrometry(GC–MS/MS) was developed and validated. GC–MS/MS inelectron ionization mode was shown to be a powerful tool forthe (ultra)trace analysis of multiclass environmental contam-inants in complex matrices, providing measurements withhigh selectivity and sensitivity. Another positive aspect char-acterizing the newly developed method is a substantial sim-plification of the sample preparation, which was achieved byan ethyl acetate QuEChERS (quick, easy, cheap, effective,rugged and safe) based extraction followed by silica minicol-umn clean-up. With use of this sample preparation approachthe sample laboratory throughput was increased not onlybecause six samples may be prepared in approximately 1 h,but also because all the above-mentioned groups of contam-inants can be determined in a single GC–MS/MS run. Underthe optimized conditions, the recoveries of all target analytesin both matrices were within the range from 70 to 120 % and

the repeatabilities were 20 % or less. The method quantifica-tion limits were in the range from 0.005 to 1 μg kg–1 and from0.05 to 10 μg kg–1 for fish muscle tissue and fish feed, respec-tively. The developed method was successfully applied to thedetermination of halogenated persistent organic pollutants andpolycyclic aromatic hydrocarbons in fish and fish feed samples.

Keywords Gas chromatography–tandemmass spectrometry .

Triple quadrupole . Halogenated persistent organic pollutants .

Polycyclic aromatic hydrocarbons . Fish . Fish feed

Introduction

Consumption of fish is considered to be an irreplaceablecomponent of a balanced human diet [1], which in the lastfew decades has led to increased fish consumption worldwide,including regions where fish has never formed part of atraditional diet [2]. As a consequence, the wild fish populationhas been decreasing, and the use of new fish farming practicesmaking possible an increase and streamlining of fish bodymass yields has been growing [3]. Although approximatelytwo thirds of the fish consumed in the EU are still wild, selectedspecies such as salmon, rainbow trout and carp are predomi-nantly being farmed [1]. In any case, fish is undoubtedly animportant source of long-chain n-3 polyunsaturated fatty acids(such as docosahexaenoic acid and eicosapentaenoic acid),iodine, selenium and vitamins A and D. On the other hand,fish can also significantly contribute to dietary exposure tovarious contaminants. These hazardous chemicals may includehalogenated persistent organic pollutants (POPs), methylmer-cury and organotin compounds [1], which all pose a risk forhumans, and therefore need to be monitored. Although controlof the diet of wild fish, which is the main source of exposure to

Published in the topical collection Rapid Detection in Food and Feedwith guest editors Rudolf Krska and Michel Nielen.

Electronic supplementary material The online version of this article(doi:10.1007/s00216-013-7000-4) contains supplementary material,which is available to authorized users.

K. Kalachova : J. Pulkrabova : T. Cajka : L. Drabova :M. Stupak : J. Hajslova (*)Department of Food Analysis and Nutrition, Faculty of Food andBiochemical Technology, Institute of Chemical Technology,Prague, Technicka 3,166 28 Prague 6, Czech Republice-mail: [email protected]

Anal Bioanal Chem (2013) 405:7803–7815DOI 10.1007/s00216-013-7000-4

http://dx.doi.org/10.1007/s00216-013-7000-4

these chemicals, is beyond human control, a great chance existsto monitor directly feed and feeding practices in the case offarmed fish, and thus reduce some contaminant levels [1].

A wide range of natural and anthropogenic contaminantshave been identified in fish muscle tissue [1], and even thoughthey are commonly present at (ultra)trace concentrations (μgkg-1), long-term exposure may trigger endocrine disruptions,neurotoxicity, cancer and other adverse health effects [4].Consequently, residues of environmental contaminants infood and feed and food safety overall have become an issueof high priority for many governments, and the call for dataregarding these chemicals and identification of potential risksto consumers has grown in recent years [5, 6]. Although thereare no maximum limits for the presence of brominated flameretardants (BFRs) and organochlorinated pesticides (OCPs) infood and feed in EU legislation, both dioxin-like (congenersno. 77, 81, 126, 169, 105, 114, 118, 123, 156, 157, 167 and189) and so-called major (non-dioxin-like) polychlorinatedbiphenyls (PCBs; congeners no. 28, 52, 101, 138, 153 and180) are regulated. In the case of dioxin-like PCBs, the max-imum limit was set for their sumwith dioxins at 0.0065μg kg-1

wet weight WHO-PCDD/F-PCB-TEQ (World Health Orga-nisation-PolyChlorinated DibenzoDioxins/Furans-PCB-ToxicEQuivalent) for fish meat of both wild and farmed origin.For major PCBs, maximum limits of 75 and 125 μg kg-1 wetweight were set for meat of wild and farmed fish, respectively[7]. For polycyclic aromatic hydrocarbons (PAHs), there areno maximum limits for fresh fish meat; however, muscle meatof smoked fish should not contain more than 5 μg of benzo[-a]pyrene (BaP) per kilogram and a total of 30 μg of BaP,benz[a]anthracene (BaA), benzo[b]fluoranthene (BbFA) andchrysene (CHR) per kilogram [8]. To fulfil all the above-mentioned tasks, the availability of relevant analytical tools isimportant to make possible effective control and eliminationof potential sources of contamination. However, both usuallyextremely low concentrations of PCBs, OCPs, BFRs andPAHs in fish muscle tissue/fish feed and the complexity ofsuch matrices, in which the presence of a high amount of co-extracted interferences affects the determination of target ana-lytes, make this analysis somewhat difficult [2, 4].

Gas chromatography (GC) coupled with mass spectrome-try (MS) in various set-ups is nowadays the technique ofprimary choice when analysing the above-mentioned groupsof contaminants. The most commonly used mass spectrome-ters are unit-resolution single quadrupole instruments operat-ed in both electron ionization (EI) and negative chemicalionization mode. This technique offers rugged, reliable andsimple operation at relatively low cost; however, in somecases it suffers from limitations in terms of selectivity andsensitivity, mainly because of interfering matrix components.High-resolution instruments, on the other hand, offer highsensitivity, fewer interferences and thus improved accuracyof the measurement but at much higher cost [9]. To overcome

these limitations, tandemmass spectrometers with triple quad-rupole (QqQ) mass analysers have recently been introducedrepresenting highly selective instruments. In general, MS/MSallows one to minimize matrix component interferences, andat the same time, thanks to the possibility of selecting suitableprecursor and product ions, makes possible identification andquantification of the above-mentioned contaminants even at(ultra)trace concentrations [4, 9–12].

Notwithstanding that a highly selective QqQ mass spec-trometer is used, since GC–MS instruments are generallyrather intolerant to non-volatile matrix impurities, the choiceof an appropriate sample preparation strategy is also importantto avoid poor ionization, background noise and contaminationof the whole GC–MS system [4]. The isolation of halogenatedPOPs and PAHs is usually based on non-selective isolation ofa lipid portion, for instance in a Soxhlet apparatus, followedby elimination of co-extracts using gel permeation chroma-tography or solid-phase extraction (SPE) with different sorb-ents. Alternatively, microwave-assisted extraction, sonication,pressurized liquid extraction [5] or supercritical fluid extrac-tion may be used [13–15]. Most recently a QuEChERS(quick, easy, cheap, effective, rugged and safe) method orig-inally developed for pesticide residue analysis in fruit andvegetables [16, 17] has been applied for the determination ofvarious environmental contaminants in different matrices,including those of animal origin [6, 18–22].

Regarding method development, various studies based onthe use of GC–MS/MS with a QqQ mass analyser have beenreported. With limited exceptions [4, 6, 23, 24], they mostlyfocused only on one or two groups of contaminants [2, 3, 11,12, 21, 25–28]. The main aim of the study presented here wasto introduce a multiclass strategy for the determination ofvarious groups of contaminants (PCBs, OCPs, BFRs andPAHs) with a wide range of physicochemical properties in fishmuscle tissue and fish feed. Effort was focused on the samplepreparation method, which should increase the overall samplelaboratory throughput, decrease time and cost requirements andat the same time be environmentally friendly. In parallel, theapplication potential of GC–MS/MSwith a QqQmass analyserfor the identification and quantification of (ultra)trace levels ofthe above-mentioned contaminants was evaluated. Finally, abatch of real-life fish (including different species, both wild andfarmed) and fish feed samples was analysed using this newlydeveloped strategy and the variations in the levels of contam-inants and their profiles were assessed.

Experimental

Standards

Certified standards of individual polybrominated diphenylether (PBDE) congeners (nos. 28, 37, 47, 49, 66, 77, 85, 99,

7804 K. Kalachova et al.

100, 153, 154, 183, 196, 197, 203, 206, 207 and 209), 13C-labelled BDE 209, hexabromobenzene (HBB), pentabromoto-luene (PBT), pentabromoethylbenzene (PBEB), bis(2,4,6-tri-bromophenoxy)ethane (BTBPE), octabromo-1-phenyl-1,3,3-trimethylindane (OBIND) and decabromodiphenyl ethane(DBDPE) (all with declared purity greater than 98 %) weresupplied by Wellington Laboratories (Guelph, Ontario, Can-ada). PCB standards (nos. 28, 52, 77, 81, 101, 105, 114, 118,123, 126, 138, 153, 156, 157, 167, 169, 180 and 189) (all withdeclared purity greater than 97 %), certified standards of indi-vidual OCPs—aldrin, o,p′-dichlorodiphenyldichloroethane(o,p′-DDD), p,p′- dichlorodiphenyldichloroethane (p,p′-DDD), o,p′-dichlorodiphenyldichloroethylene (o,p′-DDE),p,p′- dichlorodiphenyldichloroethylene (p,p′-DDE),dieldrin, o,p′-dichlorodiphenyltrichloroethane (o,p′-DDT),p ,p ′-dichlorodiphenyltrichloroethane (p ,p ′-DDT),α-endosufan, β-endosufan, endosulfan sulfate, endrin, α-hexachlorocyclohexane (α-HCH), β-hexachlorocyclohexane(β-HCH), γ-hexachlorocyclohexane (γ-HCH), hexachloro-benzene (HCB), heptachlor, cis-heptachloroepoxide (cis-HEPO), trans-heptachloroepoxide (trans-HEPO), cis-chlor-dane and trans-chlordane (all with declared purity greater than96 %)—and PAHs—acenaphthene (AC), acenaphthylene(ACL), anthracene (AN), BaA, BaP, BbFA, benzo[c]fluorene(BcFL), benzo[e]pyrene (BeP), benzo[j]fluoranthene (BjFA),benzo[k]fluoranthene (BkFA), benzo[ghi]perylene (BghiP),CHR, cyclopenta[cd]pyrene (CPP), dibenz[ah]anthracene(DBahA), dibenzo[ae]pyrene (DBaeP), dibenzo[ah]pyrene(DBahP), dibenzo[ai]pyrene (DBaiP), dibenzo[al]pyrene(DBalP), fluoranthene (FA), fluorene (FL), indeno[1,2,3-cd]pyrene (IP), naphthalene (NA), phenanthrene (PHE), pyr-ene (PY), triphenylene (TRI), 1-methylchrysene (1MC), 1-methylnaphthalene (1MN), 1-methylphenanthrene (1MPH),1-methylpyrene (1MP), 2-methylanthracene (2MA), 2-methylnaphthalene (2MN), 3-methylchrysene (3MC) and 5-methylchrysene (5MC) (all with declared purity greater than98 %)—were purchased from Dr. Ehrenstorfer (Augsburg,Germany). The standards 13C-labelled CB 101 and 13C-la-belled CB 77, the certified standard solution of isotopicallylabelled PAHs—US EPA 16 PAH Cocktail (AC-13C6,ACL-13C6, AN-

13C6, BaA-13C6, BaP-

13C4, BbFA-13C6,

BkFA-13C6, BghiP-13C12, DBahA-

13C6, FA-13C6, FL-

13C6,CHR-13C6, IP-

13C6, NA-13C6, PHE-

13C6, PY-13C3)—and cer-

tified standards of DBaiP-13C12 and DBaeP-13C6 were supplied

by Cambridge Isotope Laboratories (Andover, MA, USA).For optimization of the MS/MS method (transition), in-

dividual standards of all compounds were prepared in iso-octane (10.000 ng mL–1) and stored at –18 °C. Calibrationsolutions containing the BFRs, PAHs, PCBs and OCPsmentioned above at concentrations of 0.05, 0.1, 0.25, 0.5,1, 2.5, 5, 10, 25, 50 and 100 ng mL-1 were prepared inisooctane and stored as mentioned above. Each calibrationlevel contained internal standards BDE 37 and 13C-labelled

CB 77 at 10 ng mL–1 and syringe standards BDE 77, 13C-labelled BDE 209, 13C-labelled CB 101 and 13C-labelledPAHs at 5, 50, 40 and 2 ng mL–1, respectively.

The standard reference material Lake Michigan Fish Tis-sue, SRM 1947 (10.4±0.5 % w/w fat), for selected PCBs,OCPs and PBDEs and standard reference material of musseltissue, SRM 1974b (fat content not provided), for selectedPAHs, PCBs and OCPs were supplied by NIST (Gaithers-burg, MD, USA).

Chemicals, reagents and other material

n-Hexane, dichloromethane and isooctane were supplied byMerck (Darmstadt, Germany). Ethyl acetate was purchasedfrom Sigma-Aldrich (Steinheim, Germany). All solventswere of analytical grade. Silica (0.063–0.200 mm) suppliedby Merck was activated by heating it at 180 °C for 5 h andwas then deactivated by adding 2 % deionized water andshaking the mixture for 3 h; it was stored in a desiccator for16 h before use. The magnesium sulfate and sodiumchloride needed for the QuEChERS-like extraction weresupplied by Sigma-Aldrich and Lach-ner (Neratovice,Czech Republic), respectively. Pasteur pipettes (D812,230-mm length) and glass wool were obtained fromPoulten & Graf (Wertheim, Germany) and Merck, re-spectively. The glass column (1-cm inner diameter) foradsorption chromatography was obtained from Merci(Brno, Czech Republic).

Instruments

A tissue grinder was supplied by Retsch (Haan, Germany).Rotavapor R-114 and R-200 rotary vacuum evaporatorswith a heating bath were obtained from Büchi (Flawil,Switzerland). A Rotina 35R centrifuge was supplied byHettich Zentrifugen (Tuttlingen, Germany).

Tested material

For the quality assurance (QA)/quality control (QC), freshfish represented by pangasius (1 % w/w fat) from the Czechretail market and fish feed (8 % w/w fat) were used. The fishfeed used for QA/QC was composed of fish meal (48.8 %),fish oil (5.7 %), wheat (17.4 %), wheat by-products (8.8 %),soya (13.4 %) and other components (5.9 %). Before theQA/QC experiments, both matrices were tested for the pres-ence of all target analytes. In the follow-up experiments, thenewly developed analytical method was used for the analy-sis of 50 fish samples including different fish species (trout,pangasius, salmon, whiting and cod), both wild (n=20) andfarmed (n=30). Detailed information regarding the real-lifefish samples is given in Table 1. All fish samples were skin-free fillets and after pooling and homogenization were kept

GC–MS/MS for analysis of organic pollutants in fish and fish feed 7805

Table 1 Overview of fish samples analysed in the pilot study and thelevels (μg kg-1) determined of the sum of dioxin-like PCBs (ΣDL-PCBs;congeners no. 77, 81, 126, 169, 105, 114, 118, 123, 156, 157, 167 and

189), major PCBs (Σ6 PCBs; congeners no. 28, 52, 101, 138, 153 and180), benzo[a]pyrene (BaP) and the sum of benz[a]anthracene, BaP,benzo[b]fluoranthene and chrysene (PAH4)

Sample code Species Country of origin Wild or farmed Fat (%) ΣDL-PCBs Σ6 PCBs BaP PAH4

F1 Trout Czech Republic Farmed 4.5 0.35 1.61 <0.025 <MQL



F4 Trout Czech Republic Farmed 8.1 0.87 3.89 0.05 <MQL

F5 Trout Germany Farmed 3.4 0.53 2.56 <0.025 <MQL

F6 Trout Spain Farmed 5.0 0.44 2.21 <0.025 <MQL





F11 Trout Denmark Farmed 4.9 0.56 2.44 <0.025 <MQL




F15 Trout Denmark Farmed 3.3 0.36 1.52 <0.025 0.03

F16 Trout Denmark Farmed 10.6 1.02 4.30 <0.025 0.05

F17 Pangasius Vietnam Farmed 0.5 0.05 0.04 0.03 0.03

F18 Pangasius Vietnam Farmed 3.2 0.02 0.16 <0.025 <MQL

F19 Pangasius Vietnam Farmed 0.7 0.06 0.05 0.03 0.03

F20 Pangasius Vietnam Farmed 0.9 <MQL 0.02 <0.025 <MQL

F21 Pangasius Vietnam Farmed 0.8 <MQL <MQL <0.025 <MQL

F22 Pangasius Vietnam Farmed 1.0 <MQL <MQL <0.025 <MQL



F25 Pangasius Vietnam Farmed 1.2 <MQL 0.05 <0.025 <MQL

F26 Salmon Scotland Farmed 10.9 1.04 4.56 <0.025 0.10

F27 Salmon Norway Wild 7.1 0.24 0.70 <0.025 <MQL

F28 Salmon Norway Farmed 7.9 0.41 1.91 <0.025 0.03

F29 Salmon Norway Farmed 7.2 0.55 2.48 <0.025 <MQL

F30 Salmon France Farmed 15.1 0.75 3.41 <0.025 <MQL

F31 Salmon France Farmed 12.6 0.81 3.64 <0.025 <MQL

F32 Whiting Spain Wild 0.8 4.98 36.97 <0.025 <MQL








F40 Cod Netherlands Wild 0.1 <MQL <MQL <0.025 0.03

F41 Cod Spain Wild 0.1 0.08 0.45 <0.025 <MQL



F44 Cod Baltic Sea Wild 0.1 0.02 0.01 <0.025 <MQL





at –18 °C until the analysis. Additionally, three samples offish feed (sampled in the Czech Republic) were also ana-lysed (see Table 2).

Extraction and clean-up

Fish muscle tissue

The fish samples were prepared by a method describedby Kalachova et al. [29] which was originally designedfor the analysis of several PAHs, PCBs and selectedlower brominated PBDEs. In this study, the potentialof this sample preparation method for the analysis of awide range of BFRs, including highly brominated ones,OCPs and other PAHs was verified. Briefly, 10 g of fishtissue homogenate (with internal standards BDE 37 and13C-labelled CB 77—10 ng absolute) was mixed with5 mL of distilled water and shaken vigorously with10 mL of ethyl acetate in a polypropylene centrifugetube for 1 min. Subsequently, 4 g of anhydrous magne-sium sulfate and 2 g of sodium chloride were added tothe mixture. The tube was shaken for another 1 min andcentrifuged, and an aliquot of 5 mL was removed fromthe upper organic layer. The solvent (5 mL) was care-fully eliminated to the last drop under a gentle streamof nitrogen.

The evaporated extract was redissolved in 1mL of n-hexaneand purified using a handmade silica minicolumn. The fatdetermination and the choice of the size of the silica minicol-umn according to the fish muscle fat content are describedelsewhere [29]. Briefly, a silica minicolumnwith 1 g of sorbentis suitable for fish with fat content up to 2 % (up to 0.1 g of fat,absolute, can be loaded on a 1-g minicolumn). For fish with fat

content higher than 2 %, a silica minicolumn with 5 g ofsorbent should be used (up to 0.8 g of fat, absolute, can beloaded on a 5-g minicolumn). If necessary, the extract can becleaned up using two minicolumns. The eluate collected wascarefully evaporated using a vacuum rotary evaporator and theresidual solvent was removed under a gentle stream of nitro-gen. Residues were finally redissolved in 0.5 mL of isooctane(the final concentration of the matrix was 10 g mL–1 for fishand 1 g mL–1 for fish feed) containing BDE 77 (5 ng mL–1),13C-labelled BDE 209 (50 ng mL–1), 13C- labelled CB 101(40 ng mL-1) and 13C-labelled PAHs (2 ng mL-1) used assyringe standards.

Fish feed

For fish feed sample preparation, almost the same method asfor fish muscle tissue was used; however, some modifica-tions had to be made. First, only 1 g of fish feed wasweighed and after addition of the internal standards (BDE37 and 13C-labelled CB 77–10 ng absolute) was mixed with14 mL of distilled water and then the mixture was left for20 min. Subsequently, 10 mL of ethyl acetate was added andthe rest of the procedure followed the protocol for fishtissue. Regarding the amount of sorbent used for the clean-up step, since the weight of the sample used for the samplepreparation is 1 g (ten times less than with fish), fish feedwith fat content up to 20 % can be loaded on the silicaminicolumn with 1 g of sorbent. For fish feed with fatcontent higher than 20 %, a 5-g silica minicolumn shouldbe used.

Elution profile of target analytes on a silica minicolumn

Since new analytes were added to the method, the elutionprofiles of targeted compounds on a silica minicolumn wereverified using 1 mL of the blank fish extract (prepared asdescribed earlier) spiked with selected analytes at a concen-tration of 50 ng mL-1. The spiked extract was loaded onto theconditioned silica minicolumn (1 g) and the analytes wereeluted with 20 mL of n-hexane–dichloromethane (3:1, v/v).Fractions of 1 mL (20 in total) were collected and evaporated,and the residual solvents were removed under a gentle streamof nitrogen. In the case of the 5-g silica minicolumn, 40 mL ofn-hexane–dichloromethane (3:1, v/v) was used for elution and

Table 1 (continued)

Sample code Species Country of origin Wild or farmed Fat (%) ΣDL-PCBs Σ6 PCBs BaP PAH4




MQL method quantification limit

Table 2 Overview of fish feed samples analysed in the pilot study andlevels (μg kg-1) determined of the sum of dioxin-like PCBs (ΣDL-PCBs; congeners no. 77, 81, 126, 169, 105, 114, 118, 123, 156, 157,167 and 189), major PCBs (Σ6 PCBs; congeners no. 28, 52, 101, 138,153 and 180), BaP and PAH4

Sample code Fat (%) ΣDL-PCBs Σ6 PCBs BaP PAH4

FF1 30 0.43 5.04 0.36 0.62

FF2 30 0.17 7.02 0.29 1.22

FF3 19 0.65 7.55 <0.25 0.65


fractions of 2 mL (20 in total) were collected. In both cases,residues were redissolved in 0.5 mL of isooctane containingthe aforementioned syringe standards.

GC–MS/MS analysis

All GC–MS experiments were performed using a TRACEGC Ultra gas chromatograph (Thermo Fisher Scientific, SanJose, CA, USA) coupled to a TSQ Quantum GC QqQ massspectrometer (Thermo Fisher Scientific) operated in EImode. The GC system was equipped with a TriPlus™autosampler (Thermo Fisher Scientific). For the separation,an Rxi-17Sil MS capillary column (30 m×0.25-mm innerdiameter×0.25-μm film thickness; Restek, Bellefonte, PA,USA) with a mid-polarity Crossbond® silarylene stationaryphase was used. A volume of 1 μL was injected usingprogrammed temperature vaporizer injection in splitlessmode with a 2-min splitless period and the following inlettemperature programme: 95 °C (0.05 min), 14.5 °Cs-1 to200 °C (1 min) and 4.5 °C s-1 to 320 °C (3 min). A baffleliner (2 mm×2.75 mm×120 mm, Siltek-deactivated;Thermo Fisher Scientific) was used. Helium was used as acarrier gas at a flow rate of 1.3 mL min-1. The oven tem-perature programme was 80 °C (2 min), 30 °C min-1 to240 °C and 10 °C min-1 to 340 °C (20 min).

The QqQ mass spectrometer was operated in selectedreaction monitoring mode detecting two transitions per ana-lyte, which are listed together with the particular collisionenergies in Tables S1–S4. The temperature of the transferline and that of the ion source were kept at 320 and 270 °C,respectively. The electron energy and the emission currentwere set to 70 eV and 50 μA, respectively. The scan timewas 0.3 s and the peak width of both quadrupoles was 0.7Dafull width at half maximum. Argon was used as a collisioncell gas at a pressure of 1.5 mTorr.

The Xcalibur™ processing and instrument control soft-ware program and QuanLab™ Forms 2.5 for data analysisand reporting (Thermo Fisher Scientific) were used. For theprincipal component analysis (PCA), the software programSIMCA (version 13.0, 2011, Umetrics, Umeå, Sweden;http://www.umetrics.com) was used.

Quality assurance/quality control

First, with use of the final instrument set-up, the repeatabil-ity of GC–MS/MS (EI) measurements was tested on thestandard mixture of all target compounds in isooctane at aconcentration of 100 ng mL-1, which corresponds to 10 and100 μg kg-1 of fish muscle tissue and fish feed, respectively.In an injection volume of 1 μL, 100 pg of each standard wasinjected into the system, and the repeatability for all targetcompounds expressed as the relative standard deviation(RSD; %) was calculated.

Consequently, the entiremethod including the QuEChERS-based sample preparation strategy with SPE silica minicolumnclean-up and the optimized instrumental determination stepprovided by GC–MS/MS (EI) was evaluated in the validationstudy. Blank fresh fish muscle tissue and fish feed wereartificially contaminated (spiked) with all target compounds.Spiking levels were chosen with regard to real-life contamina-tion levels and method quantification limits (MQLs) of 0.1, 1and 5 μg kg–1 for fish tissue and 1, 5, and 10 μg kg–1 for fishfeed. The lowest spiking level was selected to cover the con-centrations close to the MQL; however, since a wide range ofcontaminants were included in the study, for some the MQLswere still below this concentration and for others they werewell above this concentration. For both matrices, six replicatesat each spiking concentration were conducted. Samples werespiked and left at room temperature to equilibrate overnightbefore the analysis. The trueness of the method was finallydemonstrated through the analysis of SRM 1947 and SRM1974b.

To control the entire sample preparation for real-lifesamples, the recovery of the internal standards (BDE 37and 13C-labelled CB 77 in this particular case) added tothe sample prior to the extraction was monitored.

With regard to the wide concentration range of target ana-lytes (e.g. trace levels of dioxin-like PCBs compared withrelatively high abundance of major PCBs), it was necessaryto use an extensive scale of working standard solutions forcalibration (0.05–100 ng mL–1). Weighted linear regression(1/x) was used and the regression coefficient (R2) was calcu-lated for the calibration curves from theMQL up to the highestcalibration point (100 ng mL–1). R2 values higher than 0.99were obtained for all calibration ranges tested, which meantthat the linearity criterion was met.

To eliminate potential injection inaccuracies and matrixeffects, the peak area ratio of the target analyte and the partic-ular syringe standard was applied for the final quantification.The syringe standards used were as follows: BDE 77 for BDE28–183, HBB, PBT, PBEB and BTBPE, 13C- labelled CB 101for all PCBs and OCPs, and corresponding 13C-labelled ana-logues for PAHs. For those PAHs for which there is not acorresponding 13C-labelled standard, the following 13C-la-belled PAHs were used for the quantification: NA-13C6 for1MN and 2MN, AN-13C6 for 1MPH and 2MA, PY-13C3 for1MP and BcFL, BaP-13C4 for BjFA, DBaiP-

13C12 for DBahPand DBalP and CHR-13C6 for CPP, 1MC, 3MC and 5MC.

Results and discussion

Optimization of the GC–MS/MS parameters

In the GC–MS/MS (EI) experiments, appropriate GC condi-tions together with suitable mass transitions at the detection


http://www.umetrics.com

side had to be determined for all target compounds. Theoptimization of the MS/MS method consisted of (1) acquisi-tion of respective MS spectra in full-scan mode (m/z 100–1,000 mass range), (2) selection of precursor ions, (3) production scans at different collision energies (5, 10, 15, 20, 25, 30,35 and 50 eV) and (4) final tuning of the collision energy inselected reaction monitoring mode. For each compound, twoMS/MS transitions were chosen to fulfil the generally appliedidentification criteria: according to the SANCO document[30], one precursor ion with two product ions or two precursorions with one product ion should be available for unbiasedidentification of the target analyte. The set of QC require-ments to which we refer here was originally designed forpesticide residue analysis, but is now also commonly ap-plied to other organic food contaminants. An overview ofthe quantitative and confirmation MS/MS transitions andthe collision energies selected for each compound in EImode is given in Tables S1–S4.

For BFRs, with use of full-scan mode, the major (the mostintense) ions were chosen from the [M]+· isotopic pattern andidentified as suitable precursors, with the exceptions of PBTand BTBPE, for which ions from the [M–Br]+ and [M–C6H2Br3O]

+ clusters were also taken as suitable precursors.The fragmentation of the selected precursor ion led subse-quently to the formation of product ions. The most selectiveand sensitive MS/MS transitions applicable for quantificationand confirmation purposes were obtained by loss of Br2 fromthe respective precursor ion. In the case of PBT and BTBPE,the loss of Br, CH3 and CH3CH2 was also observed; for moredetails, see Table S1. To obtain the best precursor ion toproduct ion transition signal, various collision energies(between 5 and 50 eV) were tested. For BDE 196–209,OBIND and DBPDE, although theMS/MS transitions for theiridentification and quantification were optimized, they couldnot be included in the multiclass GC–MS/MS (EI) methodowing to the unsatisfactory sensitivity, which was unsuitablefor the (ultra)trace analysis. For the analysis of these com-pounds, a shorter capillary column (10–15m) and a faster oventemperature programme is required. However, since the real-life samples may contain not only the target PCBs and PBDEsbut also some compounds from the total number of 209 con-geners, a relatively slow temperature ramping was required, asdiscussed below, to achieve chromatographic resolution ofisomeric compounds with the same mass spectra [14, 25].

For PCBs, the most intense precursor ions were chosenfrom the [M]+· cluster and the product ions were consequentlyformed mainly by the loss of one or two chlorine atoms; seeTable S2. In the case of OCPs, precursor ions were formed viadifferent fragmentation pathways (Table S3), including, forinstance, dechlorination of the molecular ion as observed inthe case of HCB, cis-HEPO, trans-HEPO, cis-chlordane,trans-chlordane, o,p′-DDE and p,p′-DDE. For α-HCH,β- HCH, γ-HCH, o,p′-DDD and p,p′-DDD, the precursor ions

were selected from the [M–HCl2]+ isotopic pattern, and in the

case of heptachlor, aldrin, o,p′-DDT and p,p′-DDT and endo-sulfan sulfate, the loss of C5H5Cl, C5H6Cl, CCl3 andC4H6O4S, respectively, was observed. The product ions forall OCPs were formed by subsequent dechlorination of therespective precursor ion.

PAHs are well known for their high stability, which makesthem difficult to fragment even when relatively high collisionenergies are applied [4, 15]. In this study, the molecular radicalcation [M]+· was used as a suitable precursor ion for most ofthe PAHs, with the exception of 1MN, 2MN, AC and FL(highly volatile PAHs), for which the [M–H]+ precursor ionwas also observed (Table S4). Since these precursor ions areusually very stable and difficult to fragment by MS/MS [15],the product ions are very often formed only by the loss of oneor two hydrogen atoms. However, in the case of some PAHs,the relatively high voltage (50 eV) applied in the collision cellled to the formation of specific product ions. For instance,those formed via the loss of C2H2, C2H3, C2H4 and CH2 wereobserved.

Following the development of the MS/MS detectionstrategy, the injection and chromatographic parameters wereoptimized. The optimization of the oven temperature gradi-ent was based on the combination of the temperature pro-grammes used for individual separation of BFRs, PCBs,OCPs and PAHs, and was further optimized to obtain thebest chromatographic separation and elution peak shape forall target analytes. A relatively slow ramping was used, asalready mentioned, for the separation of all target and non-target PCBs that could be present in the sample extracts.This was also important for the proper separation of isomer-ic groups of PAHs—(1) BbFA, BjFA and BkFA, (2) IP,BghiP and DBahA and (3) BaA, CHR and CPP—whichhad to be separated chromatographically since they haveidentical MS/MS transitions. Beyond the target analytes,also the critical co-elution of BaP with non-target BeP andCHR with non-target TRI had to be resolved. Unfortunately,such an oven temperature gradient was not suitable for thedetermination of higher BFRs represented, for example, byBDE 209 and DBDPE. These compounds suffer fromthermodegradation and signal suppression under theconditions used and therefore had to be excluded fromthis multiclass method. Final heating up to 340 °C wasused for the elution of a group of heavier dibenzopyr-enes which usually strongly interact with the stationaryphase, resulting in peak broadening and lower sensitiv-ity for these four compounds (DBaeP, DBahP, DBaiPand DBalP) [15].

In spite of the use of a slow oven temperature gradient,some co-elutions of target and non-target isomeric PCBsoccurred, including CB 28 and 31, CB 84 and 101 and CB138 and 163. In the case of target CB 118 and 123, whichwere not baseline-separated, chromatographic resolution of


0.6 was achieved and their quantification was thereforefeasible. For BFRs, baseline separation was achieved forall compounds except for BDE 37 and PBT, which werefinally efficiently resolved using their specific MS/MS tran-sitions. For a critical group of PAHs, three isomeric fluo-ranthenes were not baseline-separated, as shown in Fig. 1a.However, since chromatographic resolution of 0.7 and 1.3was achieved for BbFA and BkFA and BkFA and BjFA,respectively, their quantification was feasible. In Fig. 1b, a

complete co-elution of DBahA and IP is shown. Eventhough these two analytes have different quantificationMS/MS transitions, taking into account their MS spectra, abaseline separation is needed for their proper quantification;this is discussed in more detail by Kalachova et al. [22].DBahA and IP could be quantified; however, it has to betaken into consideration, especially in the case of DBahA,that the results might be highly overestimated by the pres-ence of IP and should be taken only as preliminary data. In

Fig. 1 An example of chromatographic separation of critical groups ofpolycyclic aromatic hydrocarbons (PAHs): a benzo[b]fluoranthene(BbFA), benzo[j]fluoranthene (BjFA) and benzo[k]fluoranthene (BkFA);b indeno[1,2,3-cd]pyrene (IP), benzo[ghi]perylene (BghiP) and

dibenz[ah]anthracene (DBahA); c benz[a]anthracene (BaA), chrysene(Chr) and cyclopenta[cd]pyrene (CPP); d BaA, Chr, CPP and triphe-nylene (Tri); and e benzo[a]pyrene (BaP) and benzo[e]pyrene (BeP).SRM selected reaction monitoring


the case of BaA, CHR and CPP, chromatographic co-elutionof CHR and CPP occurred (see Fig. 1c), but the analytes canbe quantified according to their different MS/MS transitions.Besides CPP, CHR was also co-eluted together with TRI(see Fig. 1d), which had the same MS/MS transition, andthus these two analytes had to be quantified as a sum. Thelast two compounds that were potentially co-eluted, BaAand BeP, were baseline-separated (see Fig. 1e).

Sample preparation

In this study, the sample preparation procedure for the multi-class analysis of POPs and PAHs in fish previously developedwithin the frame of the European project CONffIDENCE ofthe Seventh Framework Programme [22, 31] and furtherbroadened for the determination of selected BFRs [22] wasused and its application potential for the analysis of OCPs andother PAHs, including their methylated analogues, was eval-uated. The extraction was based on the transfer of hydropho-bic target analytes from an aqueous sample suspension(partition supported by inorganic salts) into ethyl acetate fol-lowed by a clean-up (fat removal) of the organic phase on asilica minicolumn. Since the clean-up procedure might be themost critical step, elution profiles of the newly included ana-lytes together with recoveries on silica minicolumns weretested first. With use of 10 mL of elution solvent (30 mL inthe case of the 5-g minicolumn), recoveries in the range from70 to 121 %were obtained for all analytes, except for dieldrin,endrin, endosulfan sulfate, endosulfan (α and β isomers), AC,ACL, PHE, NA, 1MN and 2MN. For these analytes, recov-eries close to zero were achieved, and even when 20 mL(40 mL in the case of the 5-g minicolumn) was used for theelution, they did not increase and beyond that, fat started to beeluted from the minicolumn. The low recoveries might becaused by specific steric interactions between the sorbentand the analyte molecule or in the case of highly volatilePAHs, also the evaporation which follows the silica minicol-umn clean-up might cause loss of these compounds. Since therest of the analytes were eluted using 10mL of elutionmixture(30 mL in the case of the 5-g minicolumn), the clean-upprocedure was used in the original set-up and no additionalchanges were made. For highly brominated flame retardants(BDE 196, 197, 203, 206, 207 and 209, OBIND andDBDPE), the suitability of this clean-up has already beenproved [22]; however, since the instrumental set-up used inthis study was not applicable for these compounds (as de-scribed in “Optimization of the GC–MS/MS parameters”),they were not finally included in themethod validation process.

For fish feed, the same method as for fish tissue was tested.However, when 10 g of fish feed sample and 5 mL of waterwere used, all the water and extraction solvent was absorbedinto the matrix and it was not possible to collect any aliquots.Different sample amounts (1, 2.5, 5 and 10 g) and water

volumes (10, 12, 14, 16, 18 and 20 mL) were therefore tested,and the best results were achieved using 1 g of fish feed with14 mL of water. Such a small amount of sample (1 g) waschosen not only because of the lowmoisture content of the fishfeed (approximately 10 % water), but also because of the highamount of matrix co-extracts, which were isolated togetherwith the target analytes and were not (compared with fishtissue) eliminated from the extract using a bigger silica mini-column. Generally, farmed fish need a mixed diet and the fishfeedstuff typically includes not only fish meat and oil, but alsocereals (wheat, corn), oilseed meals (soybean meal), corngluten meal and premix (mixture of minerals, trace elements,vitamins, single cell proteins and other feed additives) [1],which might cause problems during the preparation of thiskind of matrix. Moreover, the feedstuff composition is alsostrongly dependent on whether it is intended for omnivorous orcarnivorous fish species [1].

In the next step, the extraction efficiency was tested forboth matrices for all compounds except for the aforemen-tioned eight highly brominated flame retardants, five OCPsand six volatile PAHs. In the next step, the validation of theentire method was performed considering the acceptablerecoveries obtained by this sample preparation during thepreliminary experiments (see “Method validation”).

Regarding the relevance of the potential presence of thosecompounds which were finally excluded from the method inreal fish/fish feed samples, their significance in terms of thefrequency of their presence and contamination levels is ratherminor. On the basis of data on food contamination by PBDEs,it was proved that the levels of BDE 209 were the highest inalmost all the food categories except for fish and other seafood,where BDE 47 was the congener with the highest levels [32].The presence of other highly brominated compounds (BDE196, 197, 203, 206, 207 and 209, OBIND and DBDPE) in fishis usually uncommon. Volatile PAHs (AC, ACL, PHE, NA,1MN and 2MN) are typical air pollutants, and their presence inwater and subsequently in aquatic organisms indicates acuteenvironmental contamination, e.g. by an oil spill. Moreover,even though they are present in the environment, these low-weight PAHs do not present such a high risk to humanscompared with heavy (six rings or more) PAHs (e.g. dibenzo-pyrenes) [33], and thus monitoring of those PAHs that areincluded in the method is sufficient. In the case of OCPs, in2005, the European Food Safety Authority adopted an opinionon endosulfan as an undesirable substance in animal feed thatshould be monitored; however, in 2006, higher sensitivity offish exposed to endosulfan through water in comparison withexposure via feed was proved [34]. On the basis of our previousexperiments concerning the occurrence of these five OCPs—-dieldrin, endrin, endosulfan sulfate, α-endosulfan and β-endosulfan—in fish, they are not a significant source of humanexposure. At the same time, their levels are generally rather lowcompared with those of other OCPs (e.g. HCB and p,p′-DDE).


Method validation

As described in “Quality assurance/quality control”, therepeatability (expressed as the RSD) of GC–MS/MS (EI)measurements was calculated first and was in the range from3 to 20 % for all target compounds. In the next step, theentire method including the QuEChERS-based sample prep-aration strategy with SPE silica minicolumn clean-up andthe optimized instrumental determination step provided byGC–MS/MS (EI) was evaluated in the validation study. Anoverview of the validation data including mean recovery,repeatability and MQL for fresh fish tissue and fish feed isgiven in Table S5 and S6, respectively. Mean recovery isexpressed as an average percentage of recovery not cor-rected for the recovery of the internal standard (REC; %),and repeatability was calculated as the RSD. The recoveriesand the RSD for both matrices were in the following ranges:74–119 % (RSD 1–19 %) for PCBs, 72–120 % (RSD 3–20 %) for OCPs, 73–116 % (RSD 3–19 %) for BFRs and70–119 % (RSD 1–20 %) for PAHs. At each spiking level,

REC and RSD were calculated only for analytes for whichthe MQLs were lower than or equal to the particular leveltested. The key performance characteristics documentedthrough the validation protocol outlined above met the cri-teria applied in the EU in control of food and feed contam-inants (SANCO document no. 12495/2011) [30], which wasoriginally designed for pesticide residue analysis but iscommonly applied also for other organic food contami-nants), i.e. the recoveries were in the range from 70 to120 % and the repeatabilities were lower than 20 %. More-over, for BaP the method performance criteria were alsotested against Commission Regulation (EC) No. 333/2007(limit of quantification below 0.9 μg/kg, recovery 50–120 %, precision expressed as Horwitz ratio less than 2)[35]. The results of the analysis of SRM 1947 and SRM1974b are shown in Tables S7 and S8. The concentrationsdetermined for all analytes, for which the certified/referencevalues were available, were in accordance with the certifie-d/reference values (except for BDE 100 in SRM 1947, forwhich a slightly underestimated value was determined).

Fig. 2 Principal componentanalysis score (a) and loading (b)plots of the contaminant patternof various fish species. 1 farmedtrout (Czech Republic), 2, 3 wildwhiting (Spain). The firstcomponent t[1] explains 32 % ofthe variation and the secondcomponent t[2] explains 25 %,i.e. the model describes 57 %variation; p[1] = t[1], p[2] = t[2].BDE - brominated diphenylether, CB - chlorinated biphenyl,PAH - polycyclic aromatichydrocarbon


For the correction of matrix effects, quantification based onthe peak area ratio of the target analyte and particular syringestandard was used, as described in “Quality assurance/qualitycontrol”. Quantification of PAHs was highly influenced bymatrix effects (caused mainly by active sites in the injector),which were corrected using the peak area ratio of the targetanalyte and particular syringe standard. In addition to peakintensities, the matrix also influenced the retention times of theanalytes, which can complicate the identification of a partic-ular compound. Using 18 13C-labelled PAHs therefore allowsone to control retention time shifts along the entire chromato-gram, i.e. the retention time was another parameter supportingidentification of the analytes (the retention time is the same for13C-labelled and native compounds). In the case of chlorinat-ed compounds (PCBs, OCPs), almost no matrix effects weredetermined, and use of 13C-labelled CB 101 serves mainly forthe elimination of potential injection inaccuracies. Applicationof only one 13C-labelled CB was therefore enough. ForPBDE, in the first phase of the method development, twosyringe standards were used to cover the front (up to BDE183) and back (from BDE 196 to BDE 209) parts of thechromatogram. However, since highly brominated com-pounds were not finally included in the method, BDE 77was adequate for correct quantification. Additionally, it shouldalso be emphasized that use of a QqQ mass analyser elimi-nates the presence of potential matrix interferences comparedwith, for example, a single quadrupole mass analyser, asmentioned in “Introduction”.

The MQLs were defined as the lowest concentration ofeach analyte in the matrix at which the quantification andidentity confirmation transitions provided a signal-to-noiseratio greater than 6 for the quantitative transition and greaterthan 3 for at least one confirmation transition [30]. On thebasis of preliminary measurements using matrix samplescontaminated at low levels, the MQLs of target analyteswere in the range from 0.005 to 1 μg kg–1 and from 0.05to 10 μg kg–1 for fish muscle tissue and fish feed, respec-tively; for more details, see Tables S5 and S6. Higher MQLswere obtained, as expected and already explained in “Opti-mization of the GC–MS/MS parameters”, for higher bromi-nated compounds (BDE 183 and BTBPE) and heavy PAHs(dibenzopyrenes). The MQLs achieved using the methoddeveloped enabled us to determine levels of dioxin-likePCBs and major PCBs well below the maximum limits setby EU legislation [7, 8]. Also in case of BaP and the sum ofBaP, BaA, BbFA and CHR, satisfactory MQLs of 0.025 and0.1 μg kg-1, respectively, were achieved.

Analysis of real-life samples

In the final phase of this work, the validated method was usedto analyse 50 fish and three fish feed samples in a pilot studyto further evaluate its performance and applicability. The

samples analysed included five fish species (trout, pangasius,salmon, whiting and cod) of different geographical origin(Czech Republic, Germany, Spain, Denmark, Vietnam, Nor-way, France) and breeding (wild and farmed). The test resultsare presented in Tables 1 and 2, where levels of legislativelyregulated contaminants (sum of dioxin-like PCBs, sum ofmajor PCBs, BaP and sum of BaP, BaA, BbF and CHR) aresummarized. In total, all samples tested were ‘positive’, con-taining at least one target analyte; however, no sampleexceeded the limits set in EU legislation [7, 8]. CB 101(86 %), CB 138 (86 %) and CB 153 (86 %) were the mostfrequently found contaminants. Regarding contamination offish species, the levels of pollutants decreased in the followingorder: whiting > salmon ≈ trout > pangasius ≈ cod. Whencomparing samples from the different countries of origin, wefound fish from the Baltic Sea were the most contaminated. Tovisualize the outputs of the analyses, PCA was used. One ofthe main attractive features of PCA is the ability to easilyproject particular data from a higher-dimensional to a lower-dimensional space and then reconstruct them without anypreliminary assumptions about their distribution. Figure 2shows PCA score and loading plots. Three strong outliersamples in terms of contamination pattern were observed. Adeeper inspection showed that these samples were farmedtrout (Czech Republic; Fig. 2a, outlier 1) and wild whiting(Spain; Fig. 2a, outliers 2 and 3). Whereas the sample offarmed trout was contaminated mainly by PAHs, the wildwhiting contained higher concentrations of PCBs (mainlyCB 138, 153 and 180).

Conclusions

An analytical method for the (ultra)trace determination ofmulticlass environmental contaminants in fish and fish feedusing GC–MS/MS has been developed, and its main fea-tures and benefits can be summarized as follows:

& Altogether, 73 target compounds including 18 PCBs, 16OCPs, 14 BFRs and 25 PAHs can be easily determinedin fish and fish feed using a sample preparation strategybased on the transfer of hydrophobic target analytesfrom an aqueous sample suspension into ethyl acetatefollowed by silica minicolumn clean-up. The determina-tion of target analytes was achieved using GC–MS/MS.

& GC coupled with QqQ MS performed in EI mode wasshown to be an effective tool for the analysis of multi-class environmental contaminants in complex matrices,providing measurements with high selectivity and sen-sitivity. With use of this type of mass analyser, accuratedetermination of even (ultra)trace concentrations, whichmight be of concern under certain conditions, e.g. withintotal diet studies, is feasible.


& The optimized sample preparation followed by GC–MS/MS enabled us to achieve very low MQLs for alltarget analytes, and these were in the range from 0.005to 1 μg kg–1 for fish muscle tissue and from 0.05 to10 μg kg–1 for fish feed. The recoveries of all targetanalytes in both matrices were within the acceptablerange of 70–120 % and the repeatabilities of the analyt-ical procedure were 20 % or less at all three spikinglevels.

& The newly developed method could not be applied forseveral OCPs—dieldrin, endrin, endosulfan sulfate andendosulfan (α and β isomers)—and PAHs—AC, ACL,PHE, NA, 1MN and 2MN—for which the silica minicol-umn clean-up is not suitable owing to specific stericinteractions between the sorbent and the analyte. More-over, highly volatile PAHs might be lost also during theevaporation step which follows the clean-up. Highly bro-minated compounds (BDE 196, 197, 203, 206, 207 and209, OBIND and DBDPE), were not finally included inthe method because when the GC–MS/MS set-up wasused, rather high limits of quantification were achieved.

& From the batch of 50 fish samples analysed, noneexceeded the limits set in EU legislation. CB 101(86 %), CB 138 (86 %) and CB 153 (86 %) were themost frequently found pollutants, and the contaminationof fish species decreased in the following order: whiting> salmon ≈ trout > pangasius ≈ cod. Fish from the BalticSea were the most contaminated.

Acknowledgments Financial support by the European Commissionthrough the Seventh Framework Programme (contract no. FP7–211326–CP – CONffIDENCE) and the Ministry of Education, Youthand Sports of the Czech Republic (MSM6046137305 and contract no.IDS 7E08068) is gratefully acknowledged.

References

1. EFSA (2005) Opinion of the Scientific Panel on Contaminants in theFood Chain on a Request from the European Parliament Related tothe Safety Assessment of Wild and Farmed Fish. EFSA J 236:1–118

2. Nácher-Mestre J, Serrano R, Hernández F, Benedito-Palos L, Pérez-Sánchez (2010) Gas chromatography-mass spectrometric determina-tion of polybrominated diphenyl ethers in complex fatty matricesfrom aquaculture activities. Anal Chim Acta 664:190–198

3. Nardelli V, dell’Oro D, Palermo C, Centonze D (2010) Multi-residue method for the determination of organochlorine pesticidesin fish feed based on cleanup approach followed by gas chroma-tography-triple quadrupole tandem mass spectrometry. J Chroma-togr A 1217:4996–5003

4. Sánchez-Avila J, Fernandez-Sanjuan M, Vicente J, Lacorte S(2011) Development of multi-residue method for the determinationof organic micropollutants in water, sediment and mussels usinggas chromatography-tandem mass spectrometry. J Chromatogr A1218:6799–6811

5. Sun H, Ge X, Lv Y, Wang A (2012) Application of acceleratedsolvent extraction in the analysis of organic contaminants,

bioactive and nutritional compounds in food and feed. J Chroma-togr A 1237:1–23

6. Sapozhnikova Y, Lehotay SJ (2013) Multi-class, multi-residueanalysis of pesticides, polychlorinated biphenyls, polycyclic aro-matic hydrocarbons, polybrominated diphenyl ethers and novelflame retardants in fish using fast, low-pressure gas chromatogra-phy-tandem mass spectrometry. Anal Chim Acta 758:80–92

7. Commission regulation (EU) No 1259/2011 of 2 December 2011amending Regulation No 1881/2006 as regards maximum levelsfor dioxins, dioxin-like PCBs and non dioxin-like PCBs in food-stuffs (2011) Official Journal of the European Union 320:18–23

8. Commission regulation (EU) No 835/2011 of 19 August 2011amending Regulation (EC) No 1881/2006 as regards maximumlevels of polycyclic aromatic hydrocarbons in foodstuffs (2011)Official Journal of the European Union 215:4–8

9. Mackintosh SA, Pérez-Fuentetaja A, Zimmerman LR, Pacepavicius,Clapsadl M, Alaee M, Aga DS (2012) Analytical performance of atriple quadrupole mass spectrometer compared to a high resolutionmass spectrometer for the analysis of polybrominated diphenyl ethersin fish. Anal Chim Acta 747:67–75

10. Leandro CC, Fussell RJ, Keely BJ (2005) Determination of prior-ity pesticides in baby foods by gas chromatography tandem quad-rupole mass spectrometry. J Chromatogr A 1085:207–212

11. Frenich AG, Vidal JLM, Sicilia ADC, Rodríguez MJG, BolanosPP (2006) Multiresidue analysis of organochlorine and organo-phosphorus pesticides in muscle of chicken, pork and lamb bygas chromatography-triple quadrupole mass spectrometry. AnalChim Acta 558:42–52

12. Hernández F, Portóles T, Pitarch E, López FJ, Beltrán J, Vázquez C(2005) Potential of gas chromatography coupled to triple quadru-pole mass spectrometry for quantification and confirmation oforganohalogen xenoestrogen compounds in human breast tissues.Anal Chem 77:7662–7672

13. Papachlimitzou A, Barber JL, Losada S, Bersuder P, Law RJ(2012) A review of the analysis of novel brominated flame retard-ants. J Chromatogr A 1219:15–28

14. Król S, Zabiegała B, Namieśnik J (2012) PBDEs in environmentalsamples: Sampling and analysis. Talanta 93:1–17

15. Plaza-Bolanos P, Frenich AG, Vidal JLM (2010) Polycyclic aro-matic hydrocarbons in food and beverages. Analytical methodsand trends. J Chromatogr A 1217:6303–6326

16. Anastassiades M, Lehotay S, Stajnbaher D, Schenck FJ (2003)Fast and easy multiresidue method employing acetonitrile extrac-tion/partitioning and "dispersive solid-phase extraction" for thedetermination of pesticide residues in produce. J AOAC Int86:412–430

17. http://www.quechers.com. Accessed 27 Jan 201318. Forsberg ND, Wilson GR, Anderson KA (2011) Determination of

parent and substituted polycyclic aromatic hydrocarbons in high-fat salmon using a modified QuEChERS extraction, dispersiveSPE and GC-MS. J Agric Food Chem 59:8108–8116

19. Norli HR, Christiansen A, Deribe E (2011) Application of QuECh-ERSmethod for extraction of selected persistent organic pollutants infish tissue and analysis by gas chromatography mass spectrometry. JChromatogr A 1218:7234–7241

20. Angioni A, Porcu L, Secci M, Addis P (2012) QuEChERS methodfor the determination of PAH compounds in Sardinia Sea Urchin(Paracentrotus lividus) Roe, using gas chromatography ITMS-MSanalysis. Food Anal Methods 5:1131–1136

21. Johnson YS (2012) Determination of polycyclic aromatic hydrocar-bons in edible seafood by QuEChERS-based extraction and gaschromatography-tandem mass spectrometry. J Food Sci 77:131–137

22. Kalachova K, Pulkrabova J, Cajka T, Drabova L, Hajslova J(2012) Implementation of comprehensive two-dimensional gaschromatography-time-of-flight mass spectrometry for the simulta-neous determination of halogenated contaminants and polycyclic


http://www.quechers.com

aromatic hydrocarbons in fish. Anal Bioanal Chem 403:2813–2824

23. Wang N, Kong D, Shan Z, Shi L, Cai D, Cao Y, Liu Y, Pang G(2012) Simultaneous determination of pesticides, polycyclic aro-matic hydrocarbons, polychlorinated biphenyls and phtalate estersin human adipose tissue by gas chromatography-tandem massspectrometry. J Chromatogr B 898:38–52

24. Medina CM, Pitarch E, Portolés, López FL, Hernández F (2009) GC-MS/MS multi-residue method for the determination of organochlo-rine pesticides, polychlorinated biphenyls and polybrominateddiphenyl ethers in human breast tissues. J Sep Sci 32:2090–2102

25. Kalachova K, Cajka T, Sandy C, Hajslova J, Pulkrabova J (2013)High throughput sample preparation in combination with gaschromatography coupled to triple quadrupole tandem massspectrometry (GC-MS/MS): A smart procedure for (ultra)traceanalysis of brominated flame retardants in fish. Talanta 105:109–116

26. Labadie P, Alliot F, Bourges C, Desportes A, Chevreuil M (2010)Determination of polybrominated diphenyl ethers in fish tissue bymatrix solid-phase dispersion and gas chromatography coupled totriple quadrupole mass spectrometry: Case study on European eel(Anquilla anguilla) from Mediterranean coastal lagoons. AnalChim Acta 675:97–105

27. Huo LG, Li HD, Zhao CL, Wang WB, Chen ZL, Ding RY, DongZ, Wang FE, Yang GS, Lu X, Aboul-Enein HY (2012) The determi-nation of PCBs in meat and sea food by GC-QqQ-MS/MS. FoodAnal Methods 5:1481–1491

28. Martínez Vidal JL, Garrido Frenich A, Barco Bonilla MdeL,Romero-González R, Padilla Sánchez JA (2009) Simultaneous

extraction of polycyclic aromatic hydrocarbons and polychlori-nated biphenyls in agricultural soils by pressurized liquid extrac-tion and determination by gas chromatography coupled to tandemmass spectrometry. Anal Bioanal Chem 395:1551–1562

29. Kalachova K, Pulkrabova J, Drabova L, Cajka T, Kocourek V,Hajslova J (2011) Simplified and rapid determination of poly-chlorinated biphenyls, polybrominated diphenyl ethers, and poly-cyclic aromatic hydrocarbons in fish and shrimps integrated into asingle method. Anal Chim Acta 707:84–91

30. Document no SANCO/12495/2011: method validation and qualitycontrol procedures for pesticides residues analysis in food and feedhttp://ec.europa.eu/food/plant/protection/pesticides/docs/qualcon-trol_en.pdf. Accessed 27 Jan 2013

31. European research project CONffIDENCE (FP7-211326-CP).http://www.conffidence.eu/. Accessed 27 Jan 2013

32. EFSA (2011) Scientific Opinion on Polybrominated DiphenylEthers (PBDEs) in Food. EFSA J 9(5):2156

33. European Commission (2002) Opinion of the Scientific Committeeon Food on the risks to human health of polycyclic aromatic hydro-carbons in food. SCF/CS/CNTM/PAH/29. http://ec.europa.eu/food/fs/sc/scf/out153_en.pdf

34. EFSA (2011) Statement on oral toxicity of endosulfan in fish.EFSA J 9(4):2131

35. Commission regulation (EC) No 333/2007 of 28 March 2007laying down the methods of sampling and analysis for the officialcontrol of the levels of lead, cadmium, mercury, inorganic tin, 3-MCPD and benzo(a)pyrene in foodstuffs (2007) Official Journal ofthe European Union 88:29–38


http://ec.europa.eu/food/plant/protection/pesticides/docs/qualcontrol_en.pdf

http://ec.europa.eu/food/plant/protection/pesticides/docs/qualcontrol_en.pdf

http://www.conffidence.eu

http://ec.europa.eu/food/fs/sc/scf/out153_en.pdf

http://ec.europa.eu/food/fs/sc/scf/out153_en.pdf

Analytical and Bioanalytical Chemistry

Electronic Supplementary Material

1

Gas chromatography–triple quadrupole tandem mass spectrometry: a powerful tool for the (ultra)trace

analysis of multiclass environmental contaminants in fish and fish feed

Kamila Kalachova, Jana Pulkrabova, Tomas

Cajka, Lucie

Drabova, Michal Stupak, Jana

Hajslova*

Institute of Chemical Technology, Prague, Faculty of Food and Biochemical Technology, Department of Food

Analysis and Nutrition, Technicka 3, 166 28 Prague 6, Czech Republic

*Corresponding Author E-mail: [email protected]

Tel. number: +420 220 443 185

Fax number: +420 220 443 186

mailto:[email protected]

2

Table S1 Optimised MS/MS transitions used for identification and confirmation of target BFRs and internal

standards

Analyte RT (min) Precursor ion Product ion CE RT Window

(min) (Mrel) BDE 28

10.5 405.8 [M]+

245.9 [M–Br2]+ 15

10.0 11.5 (406.9) 407.9 [M]+

247.9 [M–Br2]+ 18

BDE 37 10.9

405.8 [M]+ 245.9 [M–Br2]

+ 15 10.0 11.5

(406.9) 407.9 [M]+ 247.9 [M–Br2]

+ 18

PBT 10.9

406.7 [M–Br]+ 246.7 [M–Br3]+ 20

10.3 11.8 (486.6) 485.5 [M]+

324.8 [M–Br2]+ 30

PBEB 11.1

499.7 [M]+ 485.1 [M–CH3]

+ 10 10.5 12.0

(500.7) 501.5 [M]+ 487.0 [M–CH3]

+ 15

BDE 49 12.0

485.8 [M]+ 325.8 [M–Br2]

+ 18 9.5 16.5

(485.8) 487.8 [M]+ 327.8 [M–Br2]

+ 18

BDE 47 12.3

485.8 [M]+ 325.8 [M–Br2]

+ 18 9.5 16.5

(485.8) 487.8 [M]+ 327.8 [M–Br2]

+ 18

HBB 12.5

551.6 [M]+ 470.6 [M–Br]+ 25

11.9 13.4 (551.5) 551.6 [M]+

389.6 [M–Br2]+ 30

BDE 66 12.6

485.8 [M]+ 325.8 [M–Br2]

+ 18 9.5 16.5

(485.8) 487.8 [M]+ 327.8 [M–Br2]

+ 18

BDE 77 13.0

485.8 [M]+ 325.8 [M–Br2]

+ 18 9.5 16.5

(485.8) 487.8 [M]+ 327.8 [M–Br2]

+ 18

BDE 100 13.8

561.8 [M]+ 401.8 [M–Br2]

+ 18 12.5 16.5

(564.7) 565.8 [M]+ 405.8 [M–Br2]

+ 18

BDE 99 14.2

561.8 [M]+ 401.8 [M–Br2]

+ 18 12.5 16.5

(564.7) 565.8 [M]+ 405.8 [M–Br2]

+ 18

BDE 85 15.4

561.8 [M]+ 401.8 [M–Br2]

+ 18 12.5 16.5

(564.7) 565.8 [M]+ 405.8 [M–Br2]

+ 18

BDE 154 15.5

641.7 [M]+ 481.7 [M–Br2]

+ 18 14.3 17.3

(643.6) 645.7 [M]+ 485.7 [M–Br2]

+ 18

BDE 153 16.3

641.7 [M]+ 481.7 [M–Br2]

+ 18 14.3 17.3

(643.6) 645.7 [M]+ 485.7 [M–Br2]

+ 18

BDE 183 20.2

721.8 [M]+ 561.8 [M–Br2]

+ 20 15.5 20.5

(722.5) 723.7 [M]+ 563.8 [M–Br2]

+ 20

BTBPE 21.0

356.7 [M–C6H2Br3O]+ 277.4 [M–C6H2Br4O]+ 15 17.8 19.3

(687.6) 356.7 [M–C6H2Br3O]+ 328.4 [M–C8H7Br3O]+ 15

3

Table S2 Optimised MS/MS transitions used for identification and confirmation of target PCBs and internal

standards


(min) (Mrel) CB 28

8.2 255.9 [M]+

219.9 [M–Cl]+ 20 7.6 9.1

(257.5) 257.9 [M]+ 150.9 [M–Cl3]

+ 35

CB 52 8.4

291.9 [M]+ 221.9 [M–Cl2]

+ 22 8.0 11.0

(291.9) 291.9 [M]+

256.9 [M–Cl]+ 15

CB 101 9.2

323.8 [M]+

253.8 [M–Cl2]

+ 30 9.0 12.0

(324.6) 325.8 [M]+

290.8 [M–Cl]+ 14

13C12-CB 101 9.2

338.8 [M]+

265.8 [M–Cl2]

+ 30 9.0 12.0

(336.6) 336.8 [M]+

302.8 [M–Cl]+ 14

CB 81 9.7

291.9 [M]+

221.9 [M–Cl2]

+ 22 8.0 11.0

(291.9) 291.9 [M]+

256.9 [M–Cl]+ 15

CB 77 9.9

291.9 [M]+

221.9 [M–Cl2]

+ 22 8.0 11.0

(291.9) 291.9 [M]+

256.9 [M–Cl]+ 15

13C12-CB 77 9.9

302.9 [M]+

233.9 [M–Cl2]

+ 22 9.0 12.0

(303.9) 302.9 [M]+

268.9 [M–Cl]+ 15

CB 123 10.0

323.8 [M]+

253.8 [M–Cl2]

+ 30 9.0 12.0

(324.6) 325.8 [M]+

290.8 [M–Cl]+ 14

CB 118 10.1

323.8 [M]+

253.8 [M–Cl2]

+ 30 9.0 12.0

(324.6) 325.8 [M]+

290.8 [M–Cl]+ 14

CB 153 10.2

357.8 [M]+

287.9 [M–Cl2]

+ 25 9.0 13.0

(360.9) 359.8 [M]+

289.9 [M–Cl2]

+ 25

CB 114 10.3

323.8 [M]+

253.8 [M–Cl2]

+ 30 9.0 12.0

(324.6) 325.8 [M]+

290.8 [M–Cl]+ 14

CB 105 10.6

323.8 [M]+

253.8 [M–Cl2]

+ 30 9.0 12.0

(324.6) 325.8 [M]+

290.8 [M–Cl]+ 14

CB 138 10.8

357.8 [M]+

287.9 [M–Cl2]

+ 25 9.0 13.0

(360.9) 359.8 [M]+

289.9 [M–Cl2]

+ 25

CB 126 10.9

323.8 [M]+

253.8 [M–Cl2]

+ 30 9.0 12.0

(324.6) 325.8 [M]+

290.8 [M–Cl]+ 14

CB 167 11.0

357.8 [M]+

287.9 [M–Cl2]

+ 25 9.0 13.0

(360.9) 359.8 [M]+

289.9 [M–Cl2]

+ 25

CB 156 11.5

357.8 [M]+

287.9 [M–Cl2]

+ 25 9.0 13.0

(360.9) 359.8 [M]+

289.9 [M–Cl2]

+ 25

CB 157 11.6

357.8 [M]+

287.9 [M–Cl2]

+ 25 9.0 13.0

(360.9) 359.8 [M]+

289.9 [M–Cl2]

+ 25

CB 180 11.6

391.8 [M]+

321.8 [M–Cl2]

+ 25 11.0 13.0

(395.3) 393.8 [M]+

323.8 [M–Cl2]

+ 25

CB 169 12.0

357.8 [M]+

287.9 [M–Cl2]

+ 25 9.0 13.0

(360.9) 359.8 [M]+

289.9 [M–Cl2]

+ 25

CB 189 12.5

391.8 [M]+

321.8 [M–Cl2]

+ 25 11.0 13.0

(395.3) 393.8 [M]+

323.8 [M–Cl2]

+ 25

4

Table S3 Optimised MS/MS transitions used for identification and confirmation of target OCPs


(min) (Mrel) HCB

7.4 248.8 [M–Cl]+ 213.9 [M–Cl2]

+ 20 6.8 8.3

(284.8) 283.8 [M]+ 213.9 [M–Cl2]

+ 20

HCH-α 7.5

216.9 [M–HCl2]+ 180.9 [M–HCl3]

+ 15 7.0 8.5

(290.8) 218.9 [M–HCl2]+ 182.9 [M–HCl3]

+ 15

HCH-γ 7.9

216.9 [M–HCl2]+ 180.9 [M–HCl3]

+ 15 7.0 8.5

(290.8) 218.9 [M–HCl2]+ 182.9 [M–HCl3]

+ 15

HCH-β 8.0

216.9 [M–HCl2]+ 180.9 [M–HCl3]

+ 15 7.0 8.5

(290.8) 218.9 [M–HCl2]+ 182.9 [M–HCl3]

+ 15

Heptachlor 8.1

271.9 [M–C5H5Cl]+ 236.9 [M–C5H5Cl2]+ 15

8.0 12.0 (373.3) 273.9 [M–C5H5Cl]+ 238.9 [M–C5H5Cl2]

+ 12

Aldrin 8.5

262.9 [M–C5H6Cl]+ 192.9 [M–C5H6Cl3]+ 32

7.5 10.5 (364.9) 262.9 [M–C5H6Cl]+ 227.9 [M–C5H6Cl2]

+ 32

HEPO-cis 9.0

352.8 [M–Cl]+ 262.9 [M–C3HCl2O]+ 15 8.5 10.0

(389.3) 354.8 [M–Cl]+ 264.9 [M–C3HCl2O]+ 15

HEPO-trans 9.1

288.9 [M–CHCl2O]+ 218.9 [M–CHCl4O]+ 15 8.5 10.0

(389.3) 352.8 [M–Cl]+ 252.9 [M–CHCl3O]+ 15

Chlordane-trans 9.3

276.9 [M–C2H2Cl3]+ 203.9 [M–C2H2Cl5]

+ 16 8.8 10.3

(409.8) 372.8 [M–Cl]+ 265.9 [M–Cl4]+ 15

o,p'-DDE 9.3

246.0 [M–Cl2]+ 176.0 [M–Cl4]

+ 25 8.8 10.3

(318.0) 317.9 [M]+ 245.9 [M–Cl2]

+ 20

Chlordane-cis 9.4

372.8 [M–Cl]+ 265.9 [M–Cl4]+ 18

8.8 10.3 (409.8) 409.8 [M]+

374.8 [M–Cl]+ 5

Endosulfan-α 9.5

240.9 [M–CH3Cl2O3S]+ 205.9 [M–CH3Cl3O3S]+ 20 8.0 12.0

(406.9) 264.9 [M–C2H2ClO3S]+ 192.9 [M–C2H4Cl3O3S]+ 22

p,p'-DDE 9.6

246.0 [M–Cl2]+ 176.0 [M–Cl4]

+ 25 8.8 10.3

(318.0) 317.9 [M]+ 245.9 [M–Cl2]

+ 20

Dieldrin 9.9

262.9 [M–C5H6ClO]+ 192.9 [M–C5H6Cl3O]+ 26 7.5 10.5

(380.9) 262.9 [M–C5H6ClO]+ 227.9 [M–C5H6Cl2O]+ 5

o,p'-DDD 10.0

235.0 [M–CHCl2]+ 165.0 [M–CHCl4]

+ 20 9.5 12.5

(320.0) 237.0 [M–CHCl2]+ 165.0 [M–CHCl4]

+ 20

Endrin 10.3

262.9 [M–C5H3ClO]+ 190.9 [M–C5H3Cl3O]+ 25 7.5 10.5

(380.9) 280.9 [M–CHCl2O]+ 244.9 [M–CHCl3O]+ 12

p,p'-DDD 10.3

235.0 [M–CHCl2]+ 165.0 [M–CHCl4]

+ 20 9.5 12.5

(320.0) 237.0 [M–CHCl2]+ 165.0 [M–CHCl4]

+ 20

o,p'-DDT 10.4

234.9 [M–CCl3]+ 165.0 [M–CCl5]

+ 15 9.0 14.0

(354.5) 236.9 [M–CCl3]+ 165.0 [M–CCl5]

+ 20

Endosulfan-β 10.7

240.9 [M–CH3Cl2O3S]+ 205.9 [M–CH3Cl3O3S]+ 20 8.0 12.0

(406.9) 271.9 [M–C4H6O3S]+ 236.9 [M–C4H6ClO3S]+ 18

p,p'-DDT 10.8

234.9 [M–CCl3]+ 165.0 [M–CCl5]

+ 20 9.0 14.0

(354.5) 236.9 [M–CCl3]+ 165.0 [M–CCl5]

+ 20

Endosulfan sulfate 11.3

271.9 [M–C4H6O4S]+ 236.9 [M–C4H6ClO4S]+ 15 8.0 12.0

(422.9) 273.9 [M–C4H6O4S]+ 238.9 [M–C4H6ClO4S]+ 15

5

Table S4 Optimised MS/MS transitions used for identification and confirmation of target PAHs and internal

standards


(min) (Mrel) 13C6-NA

5.0 134.0 [M]+

83.0 [M–C4H3]+ 30

4.3 5.8 (134.2) 134.0 [M]+

107.0 [M–C2H3]+ 20

NA 5.0

128.0 [M]+ 77.0 [M–C4H3]

+ 30 4.3 5.8

(128.2) 128.0 [M]+ 102.0 [M–C2H2]

+ 20

1MN 5.6

141.0 [M–H]+ 115.0 [M–C2H3]+ 15

5.0 6.5 (142.2) 142.0 [M]+

115.0 [M–C2H3]+ 25

2MN 5.7

141.0 [M–H]+ 115.0 [M–C2H3]+ 15

5.0 6.5 (142.2) 142.0 [M]+

115.0 [M–C2H3]+ 25

13C6-ACL 6.5

158.0 [M]+ 130.0 [M–C2H4]

+ 30 5.8 7.3

(158.2) 158.0 [M]+ 156.0 [M–H2]

+ 20

ACL 6.5

152.0 [M]+ 102.0 [M–C4H2]

+ 30 5.8 7.3

(152.2) 152.0 [M]+ 126.0 [M–C2H2]

+ 20 13C6-AC

6.6 159.0 [M–H]+ 158.0 [M–H2]

+ 20 5.8 7.3

(160.2) 160.0 [M]+ 159.0 [M–H]+ 20

AC 6.6

153.0 [M–H]+ 126.0 [M–C2H4]+ 40

5.8 7.3 (154.2) 153.0 [M–H]+ 151.0 [M–H3]

+ 40 13C6-FL

7.1 171.0 [M–H]+ 145.0 [M–C2H3]

+ 30 6.4 7.9

(172.2) 171.0 [M–H]+ 169.0 [M–H3]+ 30

FL 7.1

165.0 [M–H]+ 139.0 [M–C2H3]+ 30

6.4 7.9 (166.2) 165.0 [M–H]+ 163.0 [M–H3]

+ 30 13C6-PHE

8.1 184.0 [M]+

156.0 [M–C2H4]+ 30

7.5 9.0 (184.2) 184.0 [M]+

182.0 [M–H2]+ 30

PHE 8.1

178.0 [M]+ 152.0 [M–C2H2]

+ 20 7.5 9.0

(178.2) 178.0 [M]+ 176.0 [M–H2]

+ 20 13C6-AN

8.1 184.0 [M]+

156.0 [M–C2H4]+ 30

7.5 9.0 (184.2) 184.0 [M]+

182.0 [M–H2]+ 30

AN 8.1

178.0 [M]+ 152.0 [M–C2H2]

+ 20 7.5 9.0

(178.2) 178.0 [M]+ 176.0 [M–H2]

+ 20

1MPH 8.6

192.0 [M]+ 165.0 [M–C2H3]

+ 30 8.1 9.6

(192.3) 192.0 [M]+ 189.0 [M–H3]

+ 30

2MA 8.8

192.0 [M]+ 165.0 [M–C2H3]

+ 30 8.1 9.6

(192.3) 192.0 [M]+ 189.0 [M–H3]

+ 30 13C6-FA

9.7 208.0 [M]+

206.0 [M–H2]+ 30

9.3 10.8 (208.3) 208.0 [M]+

180.0 [M–C2H4]+ 30

FA 9.7

202.0 [M]+ 176.0 [M–C2H2]

+ 30 9.3 10.8

(202.3) 202.0 [M]+ 200.0 [M–H2]

+ 30 13C3-PY

10.1 205.0 [M]+

203.0 [M–H2]+ 30

9.3 10.8 (205.3) 205.0 [M]+

204.0 [M–H]+ 30

PY 10.1

202.0 [M]+ 176.0 [M–C2H2]

+ 30 9.3 10.8

(202.3) 202.0 [M]+ 200.0 [M–H2]

+ 30

BcFL 10.7

216.0 [M]+ 189.0 [M–C2H3]

+ 30 10.3 11.8

(216.3) 216.0 [M]+ 215.0 [M–H]+ 20

1MP 11.0

216.0 [M]+ 189.0 [M–C2H3]

+ 30 10.3 11.8

(216.3) 216.0 [M]+ 215.0 [M–H]+ 20

13C6-BaA 12.3

234.0 [M]+ 208.0 [M–C2H2]

+ 30 11.8 13.3

(234.3) 234.0 [M]+ 232.0 [M–H2]

+ 30

BaA 12.3

228.0 [M]+ 202.0 [M–C2H2]

+ 30 11.8 13.3

(228.3) 228.0 [M]+ 226.0 [M–H2]

+ 30 13C6-CHR

12.5 234.0 [M]+

208.0 [M–C2H2]+ 30

11.8 13.3 (234.3) 234.0 [M]+

232.0 [M–H2]+ 30

CPP 12.5

226.0 [M]+ 200.0 [M–C2H2]

+ 30 12.3 13.8

(226.3) 226.0 [M]+ 224.0 [M–H2]

+ 30

CHR 12.5

228.0 [M]+ 202.0 [M–C2H2]

+ 30 11.8 13.3

(228.3) 228.0 [M]+ 226.0 [M–H2]

+ 30

1MC 13.1

242.0 [M]+ 226.0 [M–CH2]

+ 30 12.8 14.3

(242.3) 242.0 [M]+ 240.0 [M–H2]

+ 30

6

5MC 13.3

242.0 [M]+ 226.0 [M–CH2]

+ 30 12.8 14.3

(242.3) 242.0 [M]+ 240.0 [M–H2]

+ 30

3MC 13.4

242.0 [M]+ 226.0 [M–CH2]

+ 30 12.8 14.3

(242.3) 242.0 [M]+ 240.0 [M–H2]

+ 30 13C6-BbFA

14.6 258.0 [M]+

230.0 [M–C2H4]+ 30

14.2 15.7 (258.3) 258.0 [M]+

256.0 [M–H2]+ 30

BbFA 14.6

252.0 [M]+ 226.0 [M–C2H2]

+ 30 14.2 15.7

(252.3) 252.0 [M]+ 250.0 [M–H2]

+ 30 13C6-BkFA

14.7 258.0 [M]+

230.0 [M–C2H4]+ 30

14.2 15.7 (258.3) 258.0 [M]+

256.0 [M–H2]+ 30

BkFA 14.7

252.0 [M]+ 226.0 [M–C2H2]

+ 30 14.2 15.7

(252.3) 252.0 [M]+ 250.0 [M–H2]

+ 30

BjFA 14.8

252.0 [M]+ 226.0 [M–C2H2]

+ 30 14.2 15.7

(252.3) 252.0 [M]+ 250.0 [M–H2]

+ 30 13C4-BaP

15.7 256.0 [M]+

230.0 [M–C2H2]+ 30

15.0 16.5 (256.3) 256.0 [M]+

256.0 [M]+ 30

BaP 15.7

252.0 [M]+ 226.0 [M–C2H2]

+ 30 15.0 16.5

(252.3) 252.0 [M]+ 250.0 [M–H2]

+ 30 13C6-IP

19.3 282.0 [M]+

254.0 [M–C2H4]+ 30

17.1 18.6 (282.3) 282.0 [M]+

280.0 [M–H2]+ 30

IP 19.3

276.0 [M]+ 248.0 [M–C2H4]

+ 40 16.8 18.8

(276.3) 276.0 [M]+ 274.0 [M–H2]

+ 40 13C6-DBahA

19.3 284.0 [M]+

256.0 [M–C2H4]+ 30

17.1 18.6 (284.3) 284.0 [M]+

282.0 [M–H2]+ 30

DBahA 19.3

278.0 [M]+ 252.0 [M–C2H2]

+ 30 17.1 18.6

(278.3) 278.0 [M]+ 276.0 [M–H2]

+ 30 13C12-BghiP

20.8 288.0 [M]+

260.0 [M–C2H4]+ 30

17.9 19.4 (288.3) 288.0 [M]+

286.0 [M–H2]+ 30

BghiP 20.8

276.0 [M]+ 248.0 [M–C2H4]

+ 40 17.6 19.6

(276.3) 276.0 [M]+ 274.0 [M–H2]

+ 40

DBalP 27.9

302.0 [M]+ 276.0 [M–C2H2]

+ 40 20.5 23.5

(302.4) 302.0 [M]+ 300.0 [M–H2]

+ 30 13C6-DBaeP

30.7 308.0 [M]+

282.0 [M–C2H2]+ 40

20.5 23.5 (308.4) 308.0 [M]+

306.0 [M–H2]+ 30

DBaeP 30.7

302.0 [M]+ 276.0 [M–C2H2]

+ 40 20.5 23.5

(302.4) 302.0 [M]+ 300.0 [M–H2]

+ 30 13C12-DBaiP

32.5 314.0 [M]+

288.0 [M–C2H2]+ 40

20.5 23.5 (314.4) 314.0 [M]+

312.0 [M–H2]+ 30

DBaiP 32.5

302.0 [M]+ 276.0 [M–C2H2]

+ 40 20.5 23.5

(302.4) 302.0 [M]+ 300.0 [M–H2]

+ 30

DBahP 33.5

302.0 [M]+ 276.0 [M–C2H2]

+ 40 20.5 23.5

(302.4) 302.0 [M]+ 300.0 [M–H2]

+ 30

7

Table S5 The overview of validation data including mean recovery (REC, %), repeatability (RSD, %) and MQL

(µg kg-1

) of all target analytes in fish tissue. REC and RSD were calculated from the results of repeated analyses

(n=6) of fish tissue (fat 1%, w/w) spiked at three concentration levels 0.1, 1 and 5 μg kg–1

Analyte 0.1 µg kg-1 1 µg kg-1 5 µg kg-1 MQL

(µg kg-1) REC (%) RSD (%) REC (%) RSD (%) REC (%) RSD (%)

Maj

or

PC

Bs

CB 28 90 6 107 9 101 5 0.005

CB 52 93 9 114 10 114 10 0.005

CB 101 94 7 117 11 116 5 0.005

CB 138 107 10 105 7 119 4 0.005

CB 153 104 9 117 6 105 5 0.005

CB 180 99 8 110 6 118 4 0.01

Dio

xin

-lik

e P

CB

s

CB 77 110 13 105 12 114 10 0.005

CB 81 111 11 118 10 90 9 0.005

CB 126 108 8 103 8 113 3 0.005

CB 169 115 9 102 5 118 4 0.005

CB 105 101 5 116 13 108 6 0.005

CB 114 94 14 117 8 118 5 0.005

CB 118 105 6 102 11 111 9 0.01

CB 123 98 6 118 13 107 2 0.01

CB 156 108 10 117 4 113 5 0.005

CB 157 111 11 105 6 117 4 0.005

CB 167 107 8 119 5 114 3 0.005

CB 189 104 9 119 5 111 3 0.005

OC

Ps

Aldrin 98 12 114 16 102 6 0.1

Chlordan-trans N/A N/A 109 12 109 6 0.5

Chlordan-cis N/A N/A 99 9 107 7 0.5

HCB 79 9 82 13 86 3 0.1

HCH-alfa 103 8 120 11 101 14 0.1

HCH-beta 107 14 73 20 79 9 0.1

HCH-gama 99 10 91 16 101 6 0.1

HEPO-cis N/A N/A 106 11 90 12 0.5

HEPO-trans 90 9 119 3 98 3 0.1

Heptachlor 82 9 113 12 79 8 0.1

O,p‘-DDD 104 3 119 10 116 6 0.1

P,p‘-DDD 114 13 87 8 82 7 0.1

O,p‘-DDE 109 15 82 10 111 7 0.1

P,p‘-DDE 106 6 96 15 118 10 0.1

O,p‘-DDT 95 8 92 8 103 6 0.1

P,p‘-DDT 93 3 103 7 109 11 0.1

BF

Rs

BDE28 98 4 114 6 100 10 0.025

BDE 47 112 10 75 7 73 5 0.005

BDE 49 103 3 107 14 94 6 0.01

BDE 66 92 9 102 14 92 6 0.05

BDE 85 73 7 113 19 101 7 0.05

BDE 99 97 6 106 8 100 5 0.05

BDE 100 102 9 102 8 98 5 0.05

BDE 153 98 16 107 10 101 7 0.05

BDE 154 92 8 99 10 93 6 0.05

BDE 183 N/A N/A N/A N/A 104 10 1

HBB N/A N/A 85 9 100 3 0.5

PBEB N/A N/A 95 12 101 8 0.5

PBT N/A N/A 82 4 95 6 0.5

BTBPE N/A N/A N/A N/A 97 7 1

8

PA

Hs

AN 113 13 119 12 102 8 0.01

BaA 83 4 100 9 104 6 0.025

BaP 95 7 113 12 118 12 0.025

BbFA 90 8 111 10 117 9 0.025

BcF 76 13 104 7 106 6 0.025

BghiP 79 16 113 12 113 3 0.01

BjFA 85 8 110 12 111 9 0.025

BkFA 84 5 106 11 107 6 0.025

CHR 77 9 115 12 118 5 0.025

CPP 90 12 105 5 100 2 0.025

DBaeP 111 13 108 6 103 4 0.025

DBahA 81 7 103 10 100 6 0.005

DBahP 115 18 103 11 108 8 0.05

DBaiP N/A N/A 98 5 99 5 0.25

DBalP 112 19 104 5 104 5 0.025

FL 73 15 95 16 92 10 0.01

FA 78 16 97 12 98 17 0.05

IP 82 4 102 11 104 6 0.01

PY 81 11 102 12 101 14 0.05

1MC 90 7 117 12 117 5 0.01

1MPH 74 12 92 12 90 12 0.005

1MPr 78 17 97 18 109 15 0.01

2MA 72 5 119 17 112 12 0.01

3MC 92 7 119 13 103 3 0.01

5MC 84 8 100 7 93 4 0.005

9

Table S6 The overview of validation data including mean recovery (REC, %), repeatability (RSD, %) and MQL

(µg kg-1

) of all target analytes in fish feed. REC and RSD were calculated from the results of repeated analyses

(n=6) of fish feed (fat 8%, w/w) spiked at three concentration levels 1, 5 and 10 μg kg–1

Analyte 1 µg kg-1 5 µg kg-1 10 µg kg-1

MQL (µg kg-1) REC

(%) RSD (%) REC (%) RSD (%) REC

(%) RSD (%)

Maj

or

PC

Bs

CB 28 107 8 107 6 82 9 0.05

CB 52 110 7 111 4 89 9 0.05

CB 101 115 7 115 4 74 5 0.05

CB 138 118 7 112 4 81 9 0.05

CB 153 103 8 109 5 92 9 0.05

CB 180 117 6 116 4 74 11 0.1

Dio

xin

-lik

e P

CB

s

CB 77 114 7 118 4 92 4 0.05

CB 81 113 5 95 6 83 4 0.05

CB 126 109 9 103 5 84 10 0.05

CB 169 102 7 95 4 76 9 0.05

CB 105 119 8 113 5 104 2 0.05

CB 114 116 7 119 3 74 1 0.05

CB 118 107 9 110 3 98 5 0.1

CB 123 115 9 118 5 78 12 0.1

CB 156 107 9 111 4 77 11 0.05

CB 157 106 9 110 4 76 19 0.05

CB 167 106 7 108 2 76 13 0.05

CB 189 113 11 112 6 79 9 0.05

OC

Ps

Aldrin 79 19 98 5 115 6 1

Chlordan-trans N/A N/A 102 6 109 11 5

Chlordan-cis N/A N/A 107 15 89 6 5

HCB 112 9 120 7 90 9 1

HCH-alfa 102 4 109 9 99 3 1

HCH-beta 105 10 96 3 79 5 1

HCH-gama 94 6 93 9 83 7 1

HEPO-cis N/A N/A 89 15 105 13 5

HEPO-trans 90 9 90 9 89 10 1

Heptachlor 94 10 78 3 114 4 1

O,p‘-DDD 114 12 84 9 98 6 1

P,p‘-DDD 105 9 104 3 92 8 1

O,p‘-DDE 102 7 116 10 78 12 1

P,p‘-DDE 115 10 109 5 119 4 1

O,p‘-DDT 91 11 83 13 104 13 1

P,p‘-DDT 89 15 72 10 93 10 1

BF

Rs

BDE28 102 7 98 17 84 14 0.25

BDE 47 85 16 82 9 90 5 0.05

BDE 49 109 4 89 4 76 5 0.1

BDE 66 103 7 74 6 89 8 0.5

BDE 85 98 11 111 11 74 3 0.5

BDE 99 116 16 108 19 86 7 0.5

BDE 100 115 9 102 16 74 12 0.5

BDE 153 109 14 115 16 81 17 0.5

BDE 154 87 19 98 13 90 13 0.5

BDE 183 N/A N/A N/A N/A 105 6 10

HBB N/A N/A 90 9 93 9 5

PBEB N/A N/A 78 5 90 10 5

PBT N/A N/A 83 8 78 3 5

BTBPE N/A N/A N/A N/A 91 6 10

10

PA

Hs

AN 90 13 87 6 90 12 0.1

BaA 94 7 97 6 75 7 0.25

BaP 100 8 106 6 106 10 0.25

BbFA 89 10 93 8 93 4 0.25

BcF 92 10 89 6 84 7 0.25

BghiP 76 7 75 6 78 10 0.1

BjFA 82 8 82 7 75 9 0.25

BkFA 88 8 91 7 82 7 0.25

CHR 93 11 93 9 73 1 0.25

CPP 107 10 110 7 75 15 0.25

DBaeP 76 20 96 16 109 18 0.25

DBahA 84 9 87 5 73 17 0.05

DBahP 73 17 106 14 86 8 0.5

DBaiP N/A N/A 89 7 78 9 2.5

DBalP 85 9 90 8 114 12 0.25

FL 76 7 95 3 76 5 0.1

FA 73 11 103 9 73 7 0.5

IP 84 8 85 8 73 10 0.1

PY 89 7 77 10 70 13 0.5

1MC 76 10 76 17 98 16 0.1

1MPH 88 12 95 6 81 8 0.05

1MPr 81 9 94 19 85 9 0.1

2MA 94 7 90 14 92 13 0.1

3MC 103 3 103 6 102 17 0.1

5MC 99 14 96 10 74 9 0.05

11

Table S7 Verification of trueness of generated data: Analysis of PCBs, PBDEs and OCPs in standard reference

material – Lake Michigan fish tissue (SRM 1947, NIST, USA)

Analyte Determined value (µg kg-1) 1

Certified/reference value (µg kg-1)

Agreement Yes/No

Maj

or

PC

Bs

CB 28 16.3 ± 2.5 14.1 ± 1.0 Yes

CB 52 37.8 ± 8.0 36.4 ± 4.3 Yes

CB 101 98.7 ± 38.9 90.8 ± 0.3 Yes

CB 138 172 ± 68 162.0 ± 6.9 Yes

CB 153 206 ± 62 201 ± 3 Yes

CB 180 81.5 ± 39.5 80.8 ± 5.0 Yes

Dio

xin

-lik

e

PC

Bs

CB 105 50.4 ± 13.7 50.3 ± 3.7 Yes

CB 118 118 ± 29 112 ± 6 Yes

CB 156 14.8 ± 2.7 13.3 ± 0.9 Yes

CB 157 3.5 ± 0.6 4.08 ± 0.77 Yes

OC

Ps

Chlordan-trans 15.6 ± 4.3 12.8 ± 1.2 Yes

Chlordan-cis 56.9 ± 6.9 49.0 ± 5.5 Yes

HCB 9.4 ± 1.7 7.48 ± 0.66 Yes

HCH-alfa 0.96 ± 0.2 1.06 ± 0.12 Yes

HCH-gama 0.56 ± 0.2 0.355 ± 0.095 Yes

O,p‘-DDD 2.9 ± 1.3 3.31 ± 0.16 Yes

P,p‘-DDD 50.69 ± 1.3 45.9 ± 3.6 Yes

O,p‘-DDE 4.1 ± 0.7 3.39 ± 0.28 Yes

P,p‘-DDE 789 ± 96 720 ± 43 Yes

O,p‘-DDT 17.4 ± 8.4 15.7 ± 0.89 Yes

P,p‘-DDT 65.3 ± 27.7 59.5 ± 6.7 Yes

BF

Rs

BDE28 1.94 ± 1.06 2.26 ± 0.46 2 Yes

BDE 47 79.6 ± 19.3 73.3 ± 2.9 Yes

BDE 49 3.5± 1.6 4.01 ± 0.10 Yes

BDE 66 1.82 ± 0.5 1.85 ± 0.13 Yes

BDE 99 19.6 ± 9.5 19.2 ± 0.8 Yes

BDE 100 14.3 ± 1.3 17.1 ± 0.6 No

BDE 153 3.9 ± 1.54 3.83 ± 0.04 Yes

BDE 154 6.5 ± 1.8 6.88 ± 0.52 Yes

Yes/No – result is/not in agreement with the certified value

n.d. – not detected

1 The uncertainty was estimated by the “top-down” approach combining standard uncertainties calculated for

both precision and trueness. The estimation of the relative standard deviation for reproducibility conditions

was based on the empiric equation between repeatability (RSD) and reproducibility (RSDR)

(RSD = 0.66 * RSDR).

2 Certified value is a sum of BDE 28 and 33

12

Table S8 Verification of trueness of generated data: Analysis of PCBs, OCPs and PAHs in standard reference

material – Mussel Tissue (SRM 1974b, NIST, USA)

Analyte

Determined value (μg/kg)

1 Certified/reference

value (μg/kg) Agreement

Yes/No

Maj

or

PC

Bs CB 28 4.2 ± 0.8 3.43 ± 0.25 Yes

CB 52 7.4 ± 1.6 6.26 ± 0.37 Yes

CB 101 13.6 ± 5.4 10.7 ± 1.1 Yes CB 138 9.0 ± 3.3 9.2 ± 1.4 Yes

CB 153 15.2 ± 4.6 12.3 ± 0.8 Yes

CB 180 1.54 ± 0.51 1.17 ± 0.10 Yes

Dio

xin

-

lik

e P

CB

s CB 105 3.1 ± 0.8 4.00 ± 0.18 Yes

CB 118 10.7 ± 2.6 10.3 ± 0.4 Yes CB 156 0.89 ± 0.51 0.718 ± 0.080 Yes

CB 157 0.25 ± 0.12 0.236 ± 0.024 Yes

OC

Ps

Chlordan-trans 1.21 ± 0.29 1.14 ± 0.17 Yes

Chlordan-cis 1.41 ± 0.51 1.36 ± 0.10 Yes

O,p‘-DDD 0.85 ± 0.11 1.09 ± 0.16 Yes P,p‘-DDD 3.97 ± 1.32 3.34 ± 0.22 Yes

O,p‘-DDE 0.41 ± 0.06 0.336± 0.044 Yes P,p‘-DDE 6.5 ± 3.35 4.15 ± 0.38 Yes

PA

Hs

AN 0.49 ± 0.21 0.527 ± 0.071 Yes

BaA 6.02 ± 1.19 4.74 ± 0.53 Yes BaP 3.12 ± 0.57 2.80 ± 0.38 Yes

BbFA 7.12 ± 1.29 6.46 ± 0.59 Yes BghiP 4.12 ± 1.00 3.12 ± 0.33 Yes

BjFA 3.65 ± 1.03 2.99 ± 0.29 Yes BkFA 4.02 ± 0.81 3.16 ± 0.18 Yes

CHR 6.98 ± 1.90 6.3 ± 1.0 Yes

CPP 0.32 ± 0.14 0.227 ± 0.004 Yes DBahA 0.31 ± 0.06 0.327 ± 0.031 Yes

FL 0.38 ± 0.21 0.494 ± 0.036 Yes FA 14.5 ± 7.91 17.1 ± 0.7 Yes

IP 2.01 ± 0.43 2.14 ± 0.11 Yes

PY 19.6 ± 8.3 18.04 ± 0.6 Yes 1MPH 1.23 ± 0.67 0.98 ± 0.13 Yes

2MA 0.36 ± 0.13 0.232 ± 0.004 Yes

Yes/No – result is/not in agreement with the certified value

n.d. – not detected

1 The uncertainty was estimated by the “top-down” approach combining standard uncertainties calculated for

both precision and trueness. The estimation of the relative standard deviation for reproducibility conditions

was based on the empiric equation between repeatability (RSD) and reproducibility (RSDR)

(RSD = 0.66 * RSDR).

Gas chromatography triple quadrupole tandem mass...

Documents

Transcript of Gas chromatography triple quadrupole tandem mass...