Fractionation of chlorinated and brominated persistent organic pollutants in several food samples by pyrenyl-silica liquid chromatography prior to GC-MS determination

Analy ica Chimica Ac a 565 (2006) 208–213 F ac iona ion o chlo ina ed and b omina ed pe sis en o ganic pollu an s in se e al ood samples by py enyl-silica liquid ch oma og aphy p io o GC–MS de e mina ion Bel´ en G´ oma aa,b, Ca men Ga c´ ıa-Ruiza, Ma ´ ıa Jos´ e Gonz´ alezb, Ma ´ ıa Luisa Ma inaa,∗ aDepa amen o de Qu´ımica Anal´ı ica, Facul ad de Qu´ımica, Uni e sidad de Alcal´a, C a. Mad id-Ba celona Km. 33.600, 28871 Alcal´a de Hena es, Mad id, Spain bDepa amen o de An´alisis Ins umen al y Qu´ımica Ambien al, Ins i u o de Qu´ımica O g´anica Gene al, CSIC, Juan de la Cie a, 3 28006, Mad id, Spain Recei ed 8 No embe 2005; ecei ed in e ised o m 21 Feb ua y 2006; accep ed 22 Feb ua y 2006 Abs ac A high-pe o mance liquid ch oma og aphy (HPLC) me hod using a column o 2-(1-py enyl) e hyldime hylsilyla ed silica was de eloped in his wo k in o de o achie e sa is ac o y and ep oducible ac iona ion o polychlo ina ed biphenyls (PCBs) om b omina ed lame e a dan s (BFRs) (polyb omina ed diphenyle he s, PBDEs; and polyb omina ed biphenyls, PBBs). A e he s udy o di e en ch oma og aphic pa ame e s (mobile phase composi ion and sepa a ion empe a u e we e he mos impo an ) an isoc a ic elu ion wi h isooc ane: oluene (98:2, / ) a a low- a e o 1 mL/min, a empe a u e o 45 ◦C, and UV-de ec ion a 225 nm was selec ed o ac iona ion o PCBs ( ime egion, 4.0–5.8 min) om PBDEs (5.8–9.0 min) and om PBBs (5.8–11.0 min). The applicabili y o his me hod o ood samples was demons a ed by ac iona ing PCBs om PBDEs in h ee ood samples (cheese, milk, and ish). In e e ences om PCBs (p esen in eal samples a much highe concen a ions han PBDEs) we e emo ed in his way. In addi ion, by analysing hese samples by gas ch oma og aphy–mass spec ome y (GC–MS) wi h and wi hou p e ious ac iona ion we we e able o obse e an imp o emen in de ec ion sensi i i y o PBDEs a e HPLC ac iona ion. © 2006 Else ie B.V. All igh s ese ed. Keywo ds: B omina ed and chlo ina ed POPs; F ac iona ion; PYE; HPLC 1. In oduc ion Polyb omina ed diphenyle he s (PBDEs) and polyb omi- na ed biphenyls (PBBs) a e wo g oups o ubiqui ous pollu an s ha a e p esen in en i onmen al samples due o hei use as b omina ed lame e a dan s (BFRs) o dec ease he likelihood and in ensi y o i e in elec onic componen s, plas ics, clo h- ing and a numbe o o he comme cial p oduc s [1]. The e is cu en lyg owing in e es ,pa icula lyin PBDE analysisinen i- onmen al and ood samples due o he ma ked inc ease in he le els o hese compounds de ec ed du ing he las decade [2]. In 2003, he Eu opean Communi y (EC) in oduced a new eg- ula ion o con ol he p esence o PBDEs in he en i onmen [3], wi h he in en ion o in oducing new egula ions simila o hose exis ing o polychlo odibenzo-p-dioxins and poly- ∗Co esponding au ho . Tel.: +34 91 8854935; ax: +34 91 8854971. E-mail add ess: [email p o ec ed] (M.L. Ma ina). chlo odibenzo u ans (PCDD/Fs) and polychlo ina ed biphenyls (PCBs) o oods u s [4] in he nea u u e. These di ec i es in ol e ou ine moni o ing calling o adap a ion o analy ical me hods in o de o p o ide adequa e sensi i i y and selec i i y so as o allow unambiguous de e mina ion o BFR in complex ma ices. In mos cases he analy ical p ocedu e is based on es ablished me hods o simila compounds, such as PCBs and PCDD/Fs. The inal clean-up s ep o hese me hods is designed o ob ain h ee ac ions: (i) PCDD/Fs (ii) plana (dioxin-like) PCBs, and (iii) he bulk o PCBs plus PBDEs and PBBs. While he i s wo ac ions (PCDD/Fs and plana PCBs) a e gene ally anal- ysed by gas ch oma og aphy coupled o high- esolu ion mass spec ome y (GC–HRMS) [5], he las can be analysed ei he by gas ch oma og aphy wi h elec on cap u e de ec ion (GC–ECD) [6] o gas ch oma og aphy coupled o low- esolu ion mass spec- ome y (GC–MS) [7]. The main limi a ion o bo h ins umen al sys ems is GC coelu ions among congene s o he h ee ami- lies (PCBs, PBDEs, and PBBs), which may cause iden i ica ion 0003-2670/$ – see on ma e © 2006 Else ie B.V. All igh s ese ed. doi:10.1016/j.aca.2006.02.053 B. G´oma a e al. / Analy ica Chimica Ac a 565 (2006) 208–213 209 p oblems (in ECD de ec ion) and in e e ence p oblems, inc eas- ing he backg ound noise and de ec ion limi s (in MS de ec ion). This makes i e y di icul o analyse he BFRs, which a e usually ound a much lowe concen a ions han PCBs in eal samples. Some coelu ions de ec ed by ECD can be sol ed using dual GC columns wi h di e en composi ion and/o cha ac e - is ics [9,10] o in mul idimensional GC (MDGC) sys ems (i.e. “hea -cu ing” MDGC) [11]. The ypical coelu ion o PCB 180 and PBDE 47 can be elimina ed using longe GC columns and slowe empe a u e p og ams han usual [10]. These app oaches, howe e , signi ican ly inc ease he analysis ime. Chlo ina ed in e e ences a e elimina ed by nega i e ion chemical ionisa ion (NICI), bu his p ocedu e canno ecognise di e en b omina ed compounds since only b omine ions can be moni o ed. Al hough b omine in e e ences can be esol ed using a GC–MS sys em in elec on impac (EI) ionisa ion in selec i e ion moni o ing (SIM) mode, chlo ina ed o b omina ed compounds wi h equal m/z al- ues canno be dis inguished [12]. HRMS is he mos selec i e and sensi i e de ec ion me hod o hese kinds o analysis bu acquisi ion and main enance cos s a e e y high. The ion ap de ec o in MS/MS ope a ing mode (ITD(MS/MS)) [7] could be a good op ion because o i s good analy ical cha ac e is ics and low cos . Bu , in his case, he analysis o PBDEs in he p esence o PCBs, which a e no mally ound a highe concen a ion han PBDEs in eal samples, could inc ease he backg ound noise and de ec ion limi s. A di e en s a egy o sol e coelu ion p oblems is p io sep- a a ion o PCBs om BFR compounds in an addi ional clean-up s ep. The di e en app oaches ha ha e been de eloped, using open liquid ch oma og aphy columns wi h silica gel [13,14] and alumina [15], a e no comple ely sa is ac o y as sample ma ix and analy e concen a ion can a ec he ac iona ion. In addi ion, he solid phase ex ac ion (SPE) ca idges usually employed a e p epa ed manually and no e y ep oducible. This pape desc ibes he de elopmen o a high-pe o mance liquid ch oma og aphy (HPLC) me hod using a 2-(1-py enyl) e hyldime hylsilyla ed silica column o achie e sa is ac o y, ep oducible ac iona ion o PCBs om BFRs (PBDEs and PBBs). The applicabili y o his me hod o ood analysis was demons a ed by ac iona ing se e al ood samples. 2. Expe imen al 2.1. Appa a us Fo HPLC expe imen s, a HP-1100 Se ies sys em (Hewle - Packa d, Waldb onn, Ge many) equipped wi h an au osample and a iable wa eleng h de ec o was used. Ins umen con- ol and da a acquisi ion we e pe o med wi h he HP-1100 Se ies ChemS a ion so wa e. Sepa a ions we e pe o med on a Cosmosil 5-PYE column, 2-(1-py enyl) e hyldime hylsilyla ed silica gel, 250 ×4.6 mm i.d., pa icle size 5 ␮m om Nacalai Tesque Inc., P omochem GmbH (Kyo o, Japan), and he UV- de ec o was ope a ed a 225 nm. Di e en mobile phases, lows, and column empe a u es we e in es iga ed in o de o ob ain he bes sepa a ion be ween he chlo ina ed and b omina ed compounds s udied. i y mic o li e was injec ed o PCB and PBDE congene s and samples, while 5 ␮L was injec ed o PBB congene s due o hei highe concen a ion in he comme cial s anda d solu ion. Gas ch oma og aphy coupled wi h ion ap mass spec om- e y de ec ion in MS/MS o SCAN mode (GC–ITD(MS/MS) o GC–ITD(SCAN)) analysis was pe o med on a Va ian gas ch oma og aph (CP-3800, CA, USA) equipped wi h an ion ap de ec o (ITD, Sa u n 2000, Va ian). Samples we e injec ed in o a p og ammable empe a u e apo ising injec o (PTV) in ho spli less mode (4 ␮L; injec ion a e: 0.5 ␮L/s; 150 ◦C, hold o 2 min, and hen o 300 ◦Ca 200◦C/min; spli less ime 2.0 min) applying a p essu e pulse o 44 psi du ing 2.10 min. Fo sepa a- ion, a low bleed GC capilla y column VF-5MS (Fac o Fou TM, 55 m, 0.25 mm i.d., 0.25 ␮m ilm hickness) pu chased om Va - ian (CA, USA) was employed. Fo PBDE and PBB iden i ica ion he column empe a u e was p og ammed as ollows: 130 ◦C (1 min) o 200 ◦C (1 min) a a a e o 20 ◦C/min, hen o 280 ◦C (8 min) a 10 ◦C/min and hen o 310 ◦C (15 min) a 10 ◦C/min. PCBs we e iden i ied using he ollowing column empe a u e p og am: 100 ◦C (2 min) o 200 ◦C (3 min) a a a e o 30 ◦C/min, hen o 230 ◦C (15 min) a 3 ◦C/min and hen o 270 ◦C (15 min) a 5 ◦C/min. Helium was used as ca ie gas a a cons an low- a e o 1 mL/min. Fo he h ee amilies, he empe a u e o he ans e line and he ap we e se a 310 ◦C and 220 ◦C, espec i ely. 2.2. Reagen s and s anda ds All sol en s employed o he di e en mobile phases we e o HPLC g ade. N-hexane, n-hep ane and e hanol we e supplied by Scha lau (Ba celona, Spain) and oluene and isooc ane we e om SDS (Peypin, F ance). All sol en s used o sample p epa- a ion and GC–ITD iden i ica ion we e Pes ipu quali y and we e pu chased om SDS, excep n-hexane which was supplied by Me ck (Da ms ad , Ge many). Sulphu ic acid (analysis quali y) and silica gel 60 (0.063–0.200 mm), used o sample p epa a- ion, we e ob ained om Me ck, anhyd ous sodium sulpha e om J.T. Bake (De e e , The Ne he lands) and SupelcleanTM ENVITM-Ca b SPE ca idges om Supelco (Palo Al o, USA). Ou o 209 possible PCB congene s, o his s udy we selec ed PCB #28, 52, 101, 138, 153, 170, 180, and 194, acco ding o IUPAC numbe s [16], since hey a e ela i ely abundan in eal- li e samples [17]. A wo king s ock solu ion was p epa ed in n-hexane om he comme cial PCB s anda ds (D . Eh ens o e , Augsbu g, Ge many) con aining 1 mg/L o each na i e congene . The en na i e PBDE congene s s udied (PBDEs #17, 28, 47, 66, 85, 99, 100, 153, 154, and 183) we e selec ed on he basis o hei ela i e abundance in comme cial mix u es [1]. A wo king s ock solu ion was p epa ed in n-hexane om he indi idual PBDE s anda ds (Welling on Labo a o ies, On a io, Canada) con aining 1 mg/L o each o en na i e compounds. Comme cial mix u e Fi emas e BP-6 (P omochem, Ba celona, Spain) was used as PBB s anda d. The majo componen s o his mix u e a e PBBs #101, 118, 138, 149, 153, 156, 157, 167, and 180 [18]. Fi emas e BP-6 was dissol ed in n-hexane, and a wo king solu ion con aining 500 mg/L o he o al PBB congene s p esen in he mix u e was p epa ed. 210 B. G´oma a e al. / Analy ica Chimica Ac a 565 (2006) 208–213 Table 1 IUPAC numbe s and s uc u es o he PBDE, PCB, and PBB congene s s udied in his wo k IUPAC numbe Congene s uc u e Polyb omina ed diphenyl e he s (PBDEs) 17 2,2,4-T ib omodiphenyl e he 28 2,4,4-T ib omodiphenyl e he 47 2,2,4,4-Te ab omodiphenyl e he 66 2,3,4,4-Te ab omodiphenyl e he 85 2,2,3,4,4-Pen ab omodiphenyl e he 99 2,2,4,4,5-Pen ab omodiphenyl e he 100 2,2,4,4,6-Pen ab omodiphenyl e he 153 2,2,4,4,5,5-Hexab omodiphenyl e he 154 2,2,4,4,5,6-Hexab omodiphnyl e he 183 2,2,3,4,4,5,6-Hep ab omodiphenyl e he Polychlo ina ed/b omina ed biphenyls (PCBs/PBBs) 28 2,4,4-T ichlo obiphenyl 52 2,2,5,5-Te achlo obiphenyl 101 2,2,4,5,5-Pen achlo o/b omobiphenyl 118 2,3,4,4,5-Pen ab omobiphenyl 138 2,2,3,4,4,5-Hexachlo o/b omobiphenyl 149 2,2,3,4,5,6-Hexab omobiphenyl 153 2,2,4,45,5-Hexachlo o/b omobiphenyl 156 2,3,3,4,4,5-Hexab omobiphenyl 157 2,3,3,4,4,5-Hexab omobiphenyl 167 2,3,4,4,5,5-Hexab omobiphenyl 170 2,2,3,3,4,4,5-Hep achlo obiphenyl 180 2,2,3,4,4,5,5-Hep achlo obiphenyl 194 2,2,3,3,4,4,5,5-Oc achlo obiphenyl Table 1 g oups he IUPAC numbe s and s uc u es o he PBDE, PCB and PBB congene s s udied in his wo k so as o show he deg ee and loca ion o halogena ion in hem. 2.3. Sample p epa a ion Di e en ood samples (cheese, milk, and ho se macke el) pu chased in se e al Spanish ma ke s we e chosen o applica- ion o he agmen a ion me hod in es iga ed. The non-edible pa o he ood p oduc s was emo ed, and he edible pa was s o ed in s able condi ions ( eeze-d ied) un il analysis. Ex ac- ion o samples was done by ma ix solid phase dispe sion (MSPD) as desc ibed in de ail elsewhe e [19]. Be ween 5 and 10 g o sample was homogenised wi h 20 g o 1:1 (w/w) silica gel: anhyd ous sodium sulpha e. The mix u e was g ound o a ine powde , loaded in o a column and ex ac ed wi h 400 mL o 1:1 ( / ) ace one:hexane mix u e. Clean-up s eps consis ed in a mul ilaye column illed wi h neu al silica, silica imp eg- na ed wi h sulphu ic acid and silica modi ied wi h KOH. Bo h columns we e elu ed wi h 125 and 100 mL o n-hexane, espec- i ely. F ac iona ion be ween PCDD/Fs and non-o ho PCBs om o ho PCBs plus PBDEs and PBBs was achie ed using SupelcleanTM ENVITM-Ca b SPE ca idges as desc ibed else- whe e [6]. The ac ion con aining o ho PCBs + PBDEs + PBBs was hen e apo a ed unde a gen le s eam o ni ogen, e- dissol ed wi h 60 ␮Lo n-hexane and injec ed (50 ␮L) in o a HPLC o ac iona ion, using he HPLC me hod de eloped o sepa a ion o chlo ina ed and b omina ed compounds. Bo h ac- ions we e collec ed a he end o he HPLC sys em, e apo a ed, e-dissol ed wi h 50 ␮L o nonane and injec ed (4 ␮L) in o he GC–ITD(MS/MS) sys em as desc ibed elsewhe e [7,8]. 2.4. GC-ITD peak iden ifica ion The GC–ITD(MS/MS) p ocess consis s in ionisa ion o GC e luen by elec on impac (EI) a 70 eV ollowed by isola ion o selec ed pa en ions co esponding o each congene and sub- sequen ac iona ion o he pa en ions by collision-induced dissocia ion (CID). The mos abundan ion was selec ed om he molecula clus e o each PCB and PBDE as he pa en ion o each congene excep PBDE 183; in ha case he pa en ion came om he i s EI agmen a ion o he molecula ion, B. G´oma a e al. / Analy ica Chimica Ac a 565 (2006) 208–213 211 because i s molecula ion exceeded he maximum MS ange. Iden i ica ion by GC–ITD(MS/MS) was based on de ec ion, a he app op ia e ch oma og aphic e en ion ime, o he wo mos abundan ions o he daugh e clus e o each congene and he main enance o he heo e ical a io be ween hem wi hin an app op ia e ange. The daugh e clus e o each congene was he p oduc o he loss o wo Cl agmen s by PCBs, and wo B agmen s by PBDEs (excep o PBDE 183, in which no agmen a ion was obse ed unde he expe imen al condi- ions). The di e en pa ame e s a ec ing MS/MS de ec ion o PCBs and PBDEs ha e been published elsewhe e [7]. To iden i y PBB congene s, he GC–ITD sys em was used in he SCAN mode, moni o ing ions in he ange 500–650 m/z. Iden i ica ion was based on he EI spec um o each congene a 70 eV, which has a cha ac e is ic iso ope clus e o each b omi- na ion deg ee. In addi ion, i was assumed ha he elu ion o de o PBBs on a VF-5MS column is equal o he elu ion o de o PCBs in he same column, and his ac was con i med o wo indi idual PBB congene s (PBB #101 and 153). Also, a ios o signal a eas o he di e en PBBs peaks we e used o iden i- ica ion aking in o accoun he Fi emas e BP-6 composi ion [18]. In his way he iden i ica ion o all 9 PBB congene s by GC–EI(MS) was con i med. 3. Resul s and discussion 3.1. De elopmen o a HPLC me hod o ac iona ion o PCBs om BFRs (PBDEs and PBBs) In a p e ious pape [20], a 2-(1-py enyl) e hyldime hylsily- la ed silica gel column was used o simul aneous sepa a ion o coplana and chi al PCBs based on he numbe o chlo- ine a oms in o ho posi ions, using n-hexane a 0.5 mL/min and 25 ◦C. Since hese condi ions we e no app op ia e o he sepa a ion o PCBs om PBDEs and PBBs, all o which a e conside ed “non-pola ” compounds, di e en ch oma og aphic condi ions we e assayed. Fi s ly, inc easing pe cen ages (0, 1, 2, and 3%) o oluene we e added o n-hexane o inc ease he elu ion powe o he mobile phase and o dec ease he e en ion ime o he mos e ained congene s o each amily. Fig. 1 shows he ch oma og ams co esponding o PCBs, PBDEs and PBBs o a mobile phase o n-hexane (Fig. 1a) and n-hexane: oluene (98:2, / ) (Fig. 1b) a 0.5 mL/min and 25 ◦C. No e ha he addi ion o oluene dec eased he e en ion imes o he com- pounds, especially o he las elu ing congene s o each amily. Al hough highe pe cen ages han 2% o oluene in he mobile phase would ha e been necessa y o sho en he e en ion o he las elu ing congene s, he e was an impo an dec ease in sen- si i i y (abou six- old when oluene was inc eased om 2 o 3%) owing o UV-abso p ion o oluene in he UV egion (max- imum abso p ion a 280 nm). E hanol (a 5 and 10%) was also es ed as a modi ie in a mobile phase o n-hexane, bu his pola sol en did no imp o e on he bes sepa a ions achie ed wi h oluene. A e ha , 2% ( / ) oluene in n-hep ane and isooc ane was also es ed. Wi h isooc ane: oluene (98:2, / ), e en ion imes o he PBDE and PBB congene s we e longe han o he PCB congene s, a ou ing sepa a ion be ween he las elu ing PCB congene s and he i s elu ing PBDE and PBB congene s. Using an isooc ane: oluene (98:2, / ) mobile phase, a low- a e o 1 mL/min and a column empe a u e o 45 ◦C ( he maximum ope a ing empe a u e o he column was 50 ◦C) we e applied o educe he analysis ime. The las men ioned ch oma og aphic condi ions we e selec ed o he ac iona ion o selec ed PCB congene s om PBDE and PBB congene s in wo mix u es o s anda d solu- ions con aining espec i ely PCBs plus PBDEs and PCBs plus PBBs. The ch oma og ams o bo h mix u es a e shown in Fig. 2, whe e he ime anges selec ed o ac iona ion o he PCBs om he BFR (PBDEs and PBBs) a e ma ked. In he case o PCB/PBDE ac iona ion, he i s elu ed ac ion (F-I) was col- lec ed om 4.0 o 5.8 min and co esponded o almos all PCBs (excep PCBs #170 + 194), and he second elu ed ac ion (F-II), om 5.8 o 9.0 min, included majo i y PBDEs excep PBDE 17 and 28. I is impo an o no e ha hese wo PBDEs (PBDEs #17 and 28) a e he smalles componen s o comme cial mix u es o PBDEs and a e p esen in eal-li e samples in low concen a- ions [21]. Likewise, one o he PCB congene s (PCB #194) included in PBDE ac ion (F-II) is also a mino i y congene in Fig. 1. Ch oma og ams co esponding o he sepa a ion o PCB (∼1 mg/L, each), PBDE (∼1 mg/L, each) and PBB (∼500 mg/L, o al) congene s in a n-hexane s anda d solu ion wi h a mobile phase o (a) n-hexane and (b) n-hexane: oluene (98:2, / ). Ch oma og aphic condi ions: empe a u e 25 ◦C; low- a e, 0.5 mL/min and de ec ion a 225 nm. 212 B. G´oma a e al. / Analy ica Chimica Ac a 565 (2006) 208–213 Fig. 2. Ch oma og ams o mix u es o PCBs wi h BFRs s anda d solu ions in n-hexane con aining (a) PCBs (∼1 mg/L, each) plus PBDEs (∼1 mg/L, each) and (b) PCBs (∼1 mg/L, each) plus PBBs (∼500 mg/L, o al). Ch oma og aphic condi ions: mobile phase, isooc ane: oluene (98:2, / ); low- a e, 1.0 mL/min, empe a u e, 45 ◦C and de ec ion a 225 nm. eal-li e samples [17]. The ac ions ob ained o PCBs/PBBs imp o ed on PCBs/PBDEs in ha only PCBs #170 + 194 we e collec ed in he PBB ac ion and no PBB congene s we e p esen in he PCB ac ion. The ac iona ion achie ed sepa- a es he bulk o he PCBs o de e mina ion o PBDEs (and PBBs), which, as no ed ea lie , can be a ec ed by he p es- ence o PCB congene s, especially o hose congene s p esen in much highe concen a ions han PBDEs (and PBBs) in eal samples. I is impo an o no e ha he ac iona ion powe o he p o- posed me hod is simila o o e en be e han hose documen ed in he li e a u e. Plana PCBs we e sepa a ed om PBDEs when ac i a ed ca bon and alumina we e used in he sepa a ion col- umn; ne e heless, non-plana PCBs and PBDEs s ill appea ed in he same ac ion,andHRMS wasneeded o de e minePBDEs [15]. The ac iona ion achie ed wi h silica columns [14,22] was sa is ac o y o speci ic PCB and PBDE congene s. Fo exam- ple, Lind e al. [14] s udied he sepa a ion o he majo pa o he PCB congene s om PBDEs 47, 99, 100, 153 and 154, bu hey did no include lowe b omina ed PBDEs, which ou me hod was no able o sepa a e. And again, in he pape pub- lished by Ma ´ ınez e al. [22], quan i a i e sepa a ion be ween PCBs 28, 52, 118, 138, 153, 180, and 209 and PBDEs 47, 85, 99, 100, 153, and 154 was achie ed, bu PCBs 170 and 194, which we e included in ou s udy ( he las wo elu ing PCB con- gene s unde he selec ed ch oma og aphic condi ions) we e no conside ed. 3.2. Applicabili y o he HPLC ac iona ion me hod de eloped o eal samples A e es ing o ac iona ion wi h s anda d mix u es, he same condi ions we e applied o di e en comme cially a ail- able ood samples. Fig. 3 shows he ch oma og am o a s anda d mix u e o PCBs and PBDEs oge he wi h he ch oma og ams o h ee ood samples: cheese, milk and ho se macke el. No e he di e ence in concen a ion be ween PCBs (F-I) and PBDEs (F-II) in eal samples. As was no ed in he in oduc ion, PCB concen a ions we e much highe han PBDE concen a ions, so ha ac iona ion p oduces a PBDE ex ac ee o in e e e - ence om he main PCBs. Mo eo e , he agmen a ion achie ed unde he expe imen al condi ions enabled PCB 180 o be sep- a a ed om PBDE 47 (see Fig. 3 o peak iden i ica ion). Bo h congene s coelu e when a GC sys em is used o hei inal de e - mina ion, causing o e es ima ion (mainly o PBDE 47, which is less abundan han PCB 180) o he concen a ion wi h common non-selec ed de ec o s such as ECD. Finally, i is impo an o conside ha when eal samples we e ac iona ed wi h he p oposed HPLC me hod, he e was an obse able imp o emen in he de ec ion sensi i i y achie ed by GC–MS. Thus, he limi s o de ec ion (LODs, calcula ed as he concen a ion co esponding o h ee imes he noise signal) o he h ee samples s udied, wi h and wi hou he HPLC ac- Fig. 3. Ch oma og am o cheese, milk, and ish samples including a s anda d solu ion in n-hexane con aining PCBs plus PBDEs (∼1 mg/L, each). Ch oma o- g aphic condi ions as in Fig. 2. B. G´oma a e al. / Analy ica Chimica Ac a 565 (2006) 208–213 213 Table 2 LODs, exp essed in pg/␮L injec ed, o GC–MS analysis wi h and wi hou HPLC ac iona ion PBDE Cheese Milk Ho se macke el No ac iona ion HPLC ac iona ion No ac iona ion HPLC ac iona ion No ac iona ion HPLC ac iona ion 17 1.03 0.29 0.41 0.29 0.62 0.33 28 1.05 0.35 0.53 0.32 0.53 0.25 47 0.58 0.19 0.48 0.08 0.58 0.19 66 0.94 0.25 0.31 0.19 0.39 0.19 100 0.72 0.36 0.54 0.29 0.21 0.25 99 1.64 0.33 1.36 0.33 1.09 0.68 85 2.91 1.09 1.82 1.09 2.18 1.45 154 0.99 0.30 0.99 0.34 0.49 0.30 153 1.67 0.60 1.33 0.40 1.00 0.83 183 1144 204 817 245 654 245 iona ion s ep p io o GC–MS analysis, a e shown in Table 2. No e ha LODs o PBDE congene s de ec ed in samples gen- e ally dec eased when HPLC ac iona ion was pe o med p io o GC–MS analysis. These da a, hen, show he use ulness o he p oposed HPLC ac iona ion me hod in e ms o sensi i - i y. In addi ion, his HPLC ac iona ion me hod was highly ep oducible, gi en ha he R.S.D. (in %) o elu ion imes o h ee di e en injec ions o e 3 mon hs using wo di e - en ins umen s was lowe han 0.9% o all he congene s in es iga ed. 4. Conclusions This pape has desc ibed a HPLC me hod using a column o 2-(1-py enyl) e hyldime hylsilyla ed silica o he ac iona ion o PCBs om BFRs (PBDEs and PBBs). The me hod was ep o- ducible (RSD o elu ion imes <0.9%) and enhanced GC–MS sensi i i y by making i possible o elimina e in e e ences om he main PCBs p esen a high concen a ions in eal samples. The pape u he demons a es he use ulness o his HPLC ac iona ion me hod o he ea men o se e al ood samples p io o a GC–MS analysis. Acknowledgemen s The au ho s hank he Comunidad Au ´ onoma de Mad id (Spain) o p ojec S-0505/AGR/0312. B. G´ oma a wishes o hank he Uni e si y o Alcal´ a and CSIC o he g an s and C. Zu˜ niga o he kind help. C. Ga c´ ıa-Ruiz hanks he Minis y o Science and Technology o he Ram´ on y Cajal p og am (RYC- 2003-001). Re e ences [1] M. Alaee, P. A ias, A. Sj¨ odin, A. Be gman, En i on. In e . 29 (2003) 683. [2] A. Co aci, S. Voo spoels, J. de Boe , En i on. In e . 29 (2003) 735. [3] Di ec i e 2003/11/CE, 15.2.2003-L42/45. [4] Council Regula ion (EC) 2375/2001, Amending Commission Regula ion (EC) 466/2001, O . J. Eu . Communi ies 2001, L 321/1–5. [5] E. Abad, J.J. Lle ena, J. Saul´ o, J. Caixach, J. Ri e a, Chemosphe e 46 (2002) 1435. [6] M.A. Conceje o, L. Ramos, B. Jim´ enez, B. G´ oma a, E. Abad, J. Ri e a, M.J. Gonz´ alez, J. Ch oma og . A 917 (2001) 227. [7] B. G´ oma a, M.A. Fe n´ andez, M.J. Gonz´ alez, L. Ramos, J. Sep. Sci. 29 (2006) 123. [8] B. G´ oma a, L. He e o, L.R. Bo dajandi, M.J. Gonz´ alez, Rapid Com- mun. Mass Spec om. 20 (2006) 69. [9] M. Alaee, S. Backus, C. Cannon, J. Sep. Sci. 24 (2001) 465. [10] E. Roge s, M. Pe eas, J.-S. Pa k, G. Zhao, M.J. Cha les, J. Ch oma og . B 813 (2004) 269. [11] L.F. de Alencas o, D. G and ean, J. Ta adellas, Chimia 57 (2003) 499. [12] E. Elja a , A. de la Cal, D. Ba cel´ o, J. Ch oma og . A 1008 (2003) 181. [13] J. de Boe , Chemosphe e 18 (1989) 2131. [14] Y. Lind, P.O. Da ne ud, -S A uma, M. Aune, W. Becke , R. Bje selius, S. Cna ingius, A. Glynn, En i on. Res. 93 (2003) 186. [15] K. D’Sil a, A. Fe nandes, S. Whi e, M. Rose, O ganohalogen Compd. 61 (2003) 179. [16] K. Ballschmi e , M. Zell, Z. F esenius, Anal. Chem. 302 (1980) 20. [17] L. Ramos, E. Elja a , L.M. He n´ andez, J. Ri e a, M.J. Gonz´ alez, Anal. Chim. Ac a 402 (1999) 241. [18] L.W. Robe son, S.H. Sa e, A. Pa kinso, E. Pelliza i, C. Pochini, M.D. Mullin, J. Ag ic. Food Chem. 32 (1984) 1107. [19] L.R. Bo dajandi, G. G´ omez, M.A. Fe n´ andez, E. Abad, J. Ri e a, M.J. Gonz´ alez, Chemosphe e 53 (2003) 163. [20] L. Ramos, L.M. He n´ andez, M.J. Gonz´ alez, Anal. Chem. 71 (1999) 70. [21] A. Fe nandes, S. Whi e, K. D’Sil a, M. Rose, Talan a 63 (2004) 1147. [22] A. Ma ´ ınez, M. Ramil, R. Mon es, D. He nanz, E. Rub´ ı, I. Rod ´ ıguez, R. Cela To ijos, J. Ch oma og . A 1072 (2005) 83.