- Email: [email protected]
Miniaturized matrix solid phase dispersion procedure and solid phase microextraction for the analysis of organochlorinated pesticides and polybrominated diphenylethers in biota samples by gas chromatography electron capture detection
Journal of Chromatography A, 1216 (2009) 6741–6745
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Miniaturized matrix solid phase dispersion procedure and solid phase microextraction for the analysis of organochlorinated pesticides and polybrominated diphenylethers in biota samples by gas chromatography electron capture detection Y. Moliner-Martinez a , P. Campíns-Falcó a,∗ , C. Molins-Legua a , L. Segovia-Martínez b , A. Seco-Torrecillas b a b
Department of Analytical Chemistry, Faculty of Chemistry, University of Valencia, C/Dr Moliner, 50, 46100 Burjassot, Valencia, Spain Department of Chemical Engineering, School of Engineering, University of Valencia, C/Dr Moliner, 50, 46100 Burjassot, Valencia, Spain
a r t i c l e
i n f o
Article history: Received 11 June 2009 Received in revised form 29 July 2009 Accepted 10 August 2009 Available online 13 August 2009 Keywords: Organochlorinated pesticides (OCPs) Polybrominated diphenylethers (PBDEs) Biota Matrix solid phase dispersion (MSPD) Solid phase microextraction (SPME) GC
a b s t r a c t This work has developed a miniaturized method based on matrix solid phase dispersion (MSPD) using C18 as dispersant and acetonitrile–water as eluting solvent for the analysis of legislated organochlorinated pesticides (OCPs) and polybrominated diphenylethers (PBDEs) in biota samples by GC with electron capture (GC-ECD). The method has compared Florisil® -acidic Silica and C18 as dispersant for samples as well as different solvents. Recovery studies showed that the combination of C18–Florisil® was better when using low amount of samples (0.1 g) and with low volumes of acetonitrile–water (2.6 mL). The use of SPME for extracting the analytes from the solvent mixture before the injection resulted in detection limits between 0.3 and 7.0 g kg−1 (expressed as wet mass). The miniaturized procedure was easier, faster, less time consuming than the conventional procedure and reduces the amounts of sample, dispersant and solvent volume by approximately 10 times. The proposed procedure was applied to analyse several biota samples from different parts of the Comunidad Valenciana. © 2009 Elsevier B.V. All rights reserved.
1. Introduction The presence of organochlorinated pesticides (OCPs) and polybrominated diphenylethers (PBDEs) has been investigated in different types of environmental samples such as fish tissue [1–3], atmospheric [4,5], sediment [6], soil [7] and water samples [8–10]. There is an increasing interest in studying biota samples containing these compounds because of their bio-accumulation and persistence [11,12]. Simultaneous determination of both families of compounds has been also investigated in fishes and food web [13,14]. Thus, legislation controls their content in different matrices; in fact some of them are included in the list of priority organic pollutants (POPs) [15]. The complexity of biota samples (fatty matrix) and the low concentrations of OCPs and PBDEs require extraction, clean-up, preconcentration [16–18] and chromatographic techniques [19]. In the last few years miniaturized procedures have been developed [20,21] in order to minimize the environmental effect that analytical processes could cause. Conventional techniques such as Soxhlet [1,2,22], or liquid–liquid extraction (LLE) [16,23,24]
∗ Corresponding author. Tel.: +34 96 3543002; fax: +34 96 3544436. E-mail address: [email protected] (P. Campíns-Falcó). 0021-9673/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2009.08.021
require higher sample sizes, volumes of solvent and time than other techniques such as microwave assisted extraction (MAE) [25,26], supercritical fluid extraction (SFE) [27], pressurised liquid extraction (PLE)[28–30] or matrix solid phase dispersion (MSPD) [31–34]. The simplicity of MSPD [35,36] makes this technique one of the most advantageous in the miniaturization field. MSPD is based on a blending process between the solid samples with an abrasive solid support material. After this, the mixture is transferred to an empty cartridge to form a packing column and finally the analytes are eluted by using the appropriate solvent. The major factors affecting MSPD are the solid support material, the solvent and the matrix. In fact, MSPD methodology using acidic silica and C18 has been employed to the isolation of different drugs and pollutants in food and biological matrices [37]. MSPD miniaturized can be achieved by reducing the amount of sample (approximately 0.5 g) [35] and the corresponding reduction of sorbent, solvent (5–20 mL) and time (less than 1 h). This contrasts with the sample amount (i.e. 10 g) and several hundred of millilitres of solvent and several hours usually required for sample preparation by most traditional method (LLE) or Soxhlet extraction. Also this is a simple procedure that does not require special instrumentation. Until now, only a few authors have studied the MSPD miniaturization [34,36,38], and as far as we know, no miniaturized procedure to simultaneous determination of OCPs and PBDEs in
6742
Y. Moliner-Martinez et al. / J. Chromatogr. A 1216 (2009) 6741–6745
biota samples has been developed. Two different solid support materials were investigated, one using Florisil® and silica, another one using C18 and the last one using C18–Florisil® . SPME and GC-ECD was used as preconcentration, separation and detection system. The proposed procedure was applied to several biota samples, particularly mussels and cockles.
mode employed was splitless. When using the GC-FID the column temperature was held at 90 ◦ C for 1 min, then raised to 220 ◦ C (15 ◦ C min−1 ) and finally raised to 300 ◦ C (8 ◦ C min−1 ) and held for 10 min. The temperature program for the GC-ECD was: initial temperature 90 ◦ C for 1 min, then raised to 190 ◦ C (15 ◦ C min−1 ) and finally raised to 300 ◦ C (15 ◦ C min−1 ) and held for 7 min. 5 L was employed as injection volume (unless stated otherwise).
2. Experimental 2.4. Extraction procedure 2.1. Reagents and materials Hexane was obtained from Carlo Erba (Rodano, Italy) and methanol, ethyl acetate and acetonitrile were from J.T. Baker (Deventer, The Netherlands). Sulphuric acid was purchased from Merck (Darmstadt, Germany). The solid support materials Florisil® (60–100 mesh), silica (230–400 mesh, 60 Å) and silica (70–230 mesh, 40 Å) were obtained from Aldrich (Steinheim, Germany) and Bondesil C18 (40 m) was purchased from Varian (Habor City, CA, USA). Polyethylene solid phase extraction (SPE) cartridges (12 and 3 mL), frits and the solid phase microextraction (SPME) assembly with extraction fibers coated with polydimethylsiloxane-divinylbencene (PDMS-DVB, 65 m) were from Supelco (Bellefonte, PA, USA). Silica and Florisil® were activated at 110 ◦ C for 48 h. Acidic silica (44% sulphuric acid) was prepared by mixing neutral silica (70–230 mesh, 40 Å) in the appropriate volume of concentrated sulphuric acid. The corresponding amount of C18 was set in a cartridge and was conditioned by passing it through 2 mL of methanol and 2 mL of Milli-Q water. 2.2. Standard and samples Isomers of hexachlorocyclohexane (lindane (␥-HCH), ␣-HCH, HCH and ␦-HCH) and hexachlorobutadine (HCBu) were obtained from Rielden de Häen (Seelze, Germany). Hexachlorobencene (HCBe) was obtained from Fluka (Steinheim, Germany). Stock standard solutions of OCPs (65 mg L−1 ) were prepared in hexane and methanol when using GC-FID or GC-ECD, respectively. Polybrominated diphenylether mixture was obtained from Dr. Ehrenstorfer GmbH (Augsburg, Germany). The PBDEs determined were 2,4,2 ,4 -tetra-BDE (BDE 47), 2,4,5,2 ,4 -penta-BDE (BDE 99) and 2,4,6,2 ,4 -penta-BDE (BDE 100). Concentrations of the stock standard solution (prepared in hexane or methanol) were 40, 12 and 40 mg L−1 for BDE 47, BDE 100 and BDE 99, respectively. The concentration of the working standard solutions used with FI detection was from 0.5 to 5 mg L−1 for OCPs and PBDEs from 0.4 to 4.0 mg L−1 for BDE 47 and BDE 99 respectively and from 0.12 to 1.2 mg L−1 for BDE 100. These solutions were prepared by diluting the stock solutions in hexane. The concentrations of these solutions when using ECD were between 0.03 and 1.2 g L−1 for all the analytes. Several lyophilised biota samples (mussels and cockles) collected from the Mediterranean sea, all over the Comunidad Valenciana coast were analysed. Samples were stored in polyethylene containers at 4 ◦ C.
2.4.1. Procedure 1: Florisil® and silica as solid support material Based on Ref. [33], samples were processed by following the procedure: between 0.05 and 0.5 g of fortified samples was placed in a glass mortar and dispersed with Florisil® 0.15 g with a pestle for 5 min (homogeneous mixture). The blended mixtures were transferred to 3 mL SPE cartridge containing a coarse frit and 1 g of acidic silica and analytes were eluted with 2 mL of hexane. After, the extract was passed through a second 3 mL SPE cartridge containing a coarse frit and 0.2 g of neutral silica (230–400 mesh, 60 Å) and collected with 2 mL of the eluent (mixtures of ethylacetate:hexane or hexane). Finally, 5 L of this extract (0.5 mg L−1 for OCPs and 4.0, 1.2 and 4.0 mg L−1 for BDE 47, BDE 100 and BDE 99 respectively) was injected into the GC-FID. 2.4.2. Procedure 2: C18 as solid support 0.1 g of dry sample was placed in a glass mortar and blended with 0.4 g of C18 for 5 min. The homogenised samples were transferred to a 3 mL SPE cartridge containing a coarse frit. A clean-up step with Milli-Q water (0.5 mL) was included. After the sorbent being dried with air, hexane was employed as eluting solvent (1.3 mL). Finally, a volume of 5 L was injected into the GC-FID system. The spiked concentrations were the same for OCPs and 10 times lower for PBDEs than those used in Procedure 1. In this case, two preconcentration procedures were studied. In the first procedure, the extracts were concentrated to 150 L with a nitrogen stream in order to improve sensitivity and 5 L was injected in the GC-FID system. In the second case, a SPME step was included. A PDMS-DVB fiber was immersed into the extracts for 45 min under continuous stirring. The analytes were directly injected through the fiber (desortion time 2 min). Additionally, the FI detector was substituted by an ECD detector. 2.4.3. Procedure 3: C18–Florisil® as solid support This procedure was based on a previously developed MSPD procedure by this research group described for polycyclic aromatic hydrocarbons [38]. In this procedure 0.1 g of lyophilised samples were placed in a glass mortar and blended with 0.4 g of the C18 phase (5 min). Then, the mixture was transferred to a 3 mL SPE tube containing 0.1 g of Florisil® phase. The analytes were desorbed by flusing 1.2 mL of acetonitrile. Then, the extracts were diluted with 2.6 mL of water and then, SPME was carried out in the conditions described in Procedure 2 and EC detection was employed. 3. Results and discussion
2.3. Chromatographic equipment and experimental conditions 3.1. Optimization of the miniaturized MSPD procedure and SPME Gas chromatography was carried out with a Focus gas chromatograph equipped with a flame-ionization detector (GC-FID) and an electron capture detector (GC-ECD) from Thermo (Waltham, MA, USA). A capillary column TRB-5 (30 m × 0.25 mm i.d.), film thickness 0.25 m (Teknokroma, Barcelona, Spain) was used for the separation of the analytes. The carrier gas was nitrogen at a flow of 2.2 mL min−1 for the GC-FID and 60 mL min−1 for GC-ECD. The temperature of the injection was kept at 260 ◦ C and the detector temperature (FID and ECD) was 300 ◦ C. The injection
Several dispersant and fat retainer phases were studied: Florisil® –silica, C18 and C18–Florisil® for MSPD. The comparison of the extraction efficiencies was carried out using hexachlorobenzene, lindane, BDE 47, 99 and 100 as target analytes. The eluted compounds were directly injected into the GC or concentrated by evaporating the extracts and redissolving in low volumes of solvents. SPME was also tested as preconcentration technique. FID and ECD detectors were used.
Y. Moliner-Martinez et al. / J. Chromatogr. A 1216 (2009) 6741–6745
6743
Fig. 1. Extraction efficiencies (n = 3) using (A) Florisil® –silica procedure (0.05 g of sample, 0.15 g Florisil® , 5 g acidic silica, 0.2 g neutral silica and 2 mL of ethylacetate:hexane (30:70) as eluent), (B) C18 procedure without and with the concentration step and (C) C18–Florisil® procedure (for more details, see Section 2).
Following Procedure 1, fortified samples with OCPs and PBDEs were processed. HCBe and lindane were used in an attempt to extend the procedure proposed by Martinez et al. [33] to the simultaneous extraction of OCPs and PBDEs with a miniaturized procedure. Different sample sizes (0.05, 0.25, and 0.5 g) were spiked with the target analytes. The recoveries obtained when using 0.05 g of sample and 2 mL of ethylacetate:hexane (30:70) as eluent (n = 3) are shown in Fig. 1A. As can be seen, the values were satisfactory for PBDEs but not for the OCPs. Hexachlorobencene coeluted with a component of the matrix giving recoveries higher than 150% and significative losses on lindane content was observed (recoveries lower than 60%). The increment of the sample amount resulted in lower recoveries (49%) for hexachlorobencene and lindane. Variations in the amount of Florisil® , acidic or neutral silica or elution solvent did not improve those recoveries. Using hexane as eluent (2 mL) PBDEs were extracted independently of the sample size. However, hexachlorobencene and lindane presented low recoveries similar to that shown above. The use of acidic silica not only removed the lipids but also retained the non-polar compounds. This effect can be attributed to the formation of carbon as a sub-product of lipid’s dehydration that retained the analytes [39]. Probably, the use of large amount of solvent would improve the recoveries. Nevertheless this was not a strategy for a miniaturized procedure. Therefore, the use of Florisil® and silica provided good results for PBDEs but not for all the analytes listed in the CE Directive [15]. Procedure 2 evaluated the use of a unique solid support (C18) as dispersant and fat retainer. The extraction efficiencies for OCPs were good for 0.1 and 0.25 g of sample, for a size of 0.5 g either hexachlorobencene or lindane could not be estimated. 0.1 g was selected as the optimum because sample size, the amount of solid support material (0.4 g) and solvent volume (1.3 mL) were minimized. As can be seen in Fig. 1B, PBDEs and OCPs were eluted in a unique hexane fraction with satisfactory extraction efficiencies. The concentration level of the target analytes in biota samples required a preconcentration step. A commonly used way to improve the detection limits is the use of nitrogen streams. In this
particular case, sensitivity improved by a factor of around 7 if the extracts (1.3 mL) were concentrated to dryness and reconstituted with 150 L of hexane. However, the recoveries after this concentration step decreased down 60% (Fig. 1B) for all the tested analytes. In addition, an increase in the baseline noise was also observed (see Fig. 2A and B for a blank and a spiked sample under those conditions, respectively). This fact caused difficulty the identification and quantification of the analytes in biota samples. Thus, an alternative preconcentration system should be studied. The other preconcentration technique evaluated was a SPME procedure in combination with ECD detection. Under these conditions, hexachlorobencene and lindane were practically lost as the recoveries obtained were lower than 22%. Probably, the reason of these recoveries could be the poor efficiency of the SPME fiber to
Fig. 2. Chromatograms obtained for (A) blank and (B) sample of biota (mussel) spiked with the target analytes obtained with Procedure 2 after the concentration of the extracts with GC-FID. 2.5 mg L−1 HCBe, 2.5 mg L−1 lindane, 4 mg L−1 BDE 47, 1.2 mg L−1 BDE 100 and 4.0 mg L−1 BDE 99 (see Section 2.4.2).
6744
Y. Moliner-Martinez et al. / J. Chromatogr. A 1216 (2009) 6741–6745
Table 1 Detection limits (LOD), quantification limit (LOQ) (expressed as wet mass) and reproducibility (%RSD) obtained for the target analytes with the C18–Florisil® MSPD, SPME and GC-ECD (Procedure 3).
HCBu ␣-HCH HCBe -HCH Lindane ␦-HCH BDE 47 BDE 100 BDE 99 a
LOD (g kg−1 )
LOQ (g kg−1 )
%RSDa (n = 3)
0.3 3.7 0.5 4.7 3.0 4.7 3.0 4.6 7.1
1.0 12 1.7 16 12 16 16 15 23
10 5 9 1 1 5 23 19 5
Concentration level: 23 g kg−1 .
adsorb the target analytes from the extracts of hexane as the analytes had low distribution coefficients in hexane extracts (K = Cf /Cs , where Cf is the concentration in the fiber and Cs is the concentration in the sample) [40]. The use of another solvent would result in better extraction efficiencies in the SPME, but the use of a unique solid support limited the solvent. Hexane provided good recoveries of the analytes with low volume and with the minimum elution of undesirable compounds of the matrix. Finally, we studied the efficiency of Procedure 3 based on the use of C18–Florisil® [36]. The combination of C18 as dispersant and Florisil® as fat retainer, elution with acetonitrile, a SPME step and GC-ECD resulted in good recoveries for all the target analytes (Fig. 1C). The combination of C18 and Florisil® allowed the use of low volumes of MeCN as eluent and so good results in the SPME step as high distribution coefficient could be expected for these analytes in the PDMS-DVD fiber. It provided the adequate sensitivity for tested analytes and allowed concentration range between 0.03 and 1.2 g L−1 . Fig. 1C shows the recoveries obtained for these analytes. These values were satisfactory for all the target analytes. Bearing in mind these results, hexachlorobutadiene, ␣-, -, ␦-HCH were included in the study as they are also legislated as highly persistent compounds. Recoveries for hexachlorobutadiene, ␣-, - and ␦-HCH were 55 ± 10, 114 ± 29, 74 ± 28, 130 ± 30, respectively. 3.2. Figures of merit Under the conditions of Procedure 3 good linearity (r2 between 0.9790 and 0.9994) was obtained in the working interval (0.03 and 1.2 g L−1 ). The LODs, LOQs and RSD values (0.6 g L−1 or 23 g kg−1 expressed as wet mass) are given in Table 1. As can be seen, this procedure provided the adequate sensitivity with good reproducibility for all the analytes. Thus, this procedure was selected to analyse biota samples as it was advantageous in terms of detection limits. Additionally, it presented other advantages such as time consuming, reagent consumption and waste generation.
Fig. 3. Chromatograms obtained for (A) blank and (B) spiked sample with all the analytes (30 ng L−1 ) obtained with Section 2.4.3.
3.3. Application to biota samples The proposed procedure was employed to analyse 17 lyophilised biota samples (11 mussels and 6 cockles). The validation of the proposed methodology was done with a recovery study by spiking the samples with a mixture of the target analytes due to the inexistence of standard reference materials for these POPs in biota. The mean recovery values obtained, between 55 ± 10 and 130 ± 30% (n = 17) for all the analytes showing that the measured concentration were in agreement with the added ones. Fig. 3 shows the chromatogram obtained for a blank and a spiked sample (1.15 g kg−1 of each analyte). Only HCBe was found in three samples (mussels) at levels of 2.2 ± 0.2, 2.9 ± 0.3, and 1.8 ± 0.2 g kg−1 (expressed as mean value in wet samples for n = 3). In the other cases the contents of all the studied analytes in all the samples were below the detection limit showing the good health of the coast under evaluation. 4. Conclusions This paper has studied the simultaneous extraction of OCPs and PBDEs from biota samples (mussels and cockles) with a miniaturized MSPD extraction method. The use of C18 as dispersant and Florisil® as fat retainer combined with SPME–GC-ECD yielded to the development of a procedure with numerous advantages com-
Table 2 Comparison of the miniaturized MSPD proposed with other published procedures for OCPs and PBDEs. Parameters
Martinez et al. [33] PBDEs and PCBsa
Long et al. [31,32] OCPsa
This manuscript OCPs and PBDEsa
Sample amount (g) Sorbent (g)
0.5 1.5 g Florisil®
0.5 C18 2 g
0.1 C18 0.4 g Florisil® 0.1 g
Clean-up
5 g acidic silica 2 g neutral silica n-Hexane (20 mL) n-Hexane (20 mL) n-Hexane–diclhoromethane (80:20), 12 mL – >75%
Florisil® 2 g
Solvent (mL)
Preconcentration Recoveries a
Analytes.
Acetonitrile 8 mL
Acetonitrile (1.3 mL)
– 85 ± 3–102 ± 5
SPME 55 ± 10–130 ± 30
Y. Moliner-Martinez et al. / J. Chromatogr. A 1216 (2009) 6741–6745
pared with previously published procedures. Table 2 summarizes some of these advantages. As can be seen, C18–Florisil® demonstrated a reduction in the use of hazardous solvents and analysis time for the analysis of OCPs and PBDEs in the same fraction. Additionally, the use of SPME–GC-ECD provided the enough sensitivity for OCPs and PBDEs analysis in biota samples. The procedure minimized the required sample amount, sorbent material and solvent consumption. Analytical properties such as recoveries or amount detected were also an advantage of the proposed procedure. Acknowledgements The authors would like to thank the Ministerio de Ciencia y Tecnología of Spain for the financial support received for the project CTQ 2008-01329/BQU and the Government of the Region Valenciana (Generalitat Valenciana) for the project “Application of EU water Framework Directive 2000/60/EC on endocrine disrupters and other priority substances in coastal waters in the Region Valenciana”. Y.M.M. expresses her gratitude for her JdC contract (Ministerio de Ciencia e Innovación). References [1] N. Fidalgo-Used, G. Centineo, E. Blanco-González, A.J. Sanz-Mendel, J. Chromatogr. A 1017 (2003) 35. [2] S. Sinkkonen, A. Rantalainen, J. Paasivirta, M. Lahtiperä, Chemosphere 56 (2004) 767. [3] B. Fägström, M. Athanasiadou, I. Athanassiadis, A. Bignert, P. Grandjean, P. Weihe, A. Bergman, Chemosphere 60 (2005) 836. [4] A. Sanusi, M. Millet, P. Mirabel, H. Wortham, Sci. Total Environ. 263 (2000) 263. [5] K. Hyakawa, H. Takatsuki, I. Watanabe, S. Sakai, Chemosphere 57 (2004) 343. [6] J. Boer, C. Allchin, R. Law, B. Zegers, J.P. Boon, Trac-Trend Anal. Chem. 20 (2001) 591. [7] C. Goncalves, M.F. Alpendurada, Talanta 65 (2005) 1179. ˜ [8] V.M. León, B. Alvarez, M.A. Cogollo, S. Munoz, I. Valor, J. Chromatogr. A 999 (2003) 91. [9] H.P. Li, G.C. Li, J.F. Jen, J. Chromatogr. A 1012 (2003) 129. [10] M.G. Ikonomou, M.P. Fernández, Z.L. Hichkman, Environ. Pollut. 140 (2006) 355.
6745
[11] C. Dong, Z. Zeng, M. Yang, Water Res. 39 (2005) 4204. [12] Y.F. Li, R.W. Macdonald, Sci. Total Environ. 342 (2005) 87. [13] S. Corsolini, A. Covaci, N. Ademollo, S. Forcardi, P. Schepens, Environ. Pollut. 140 (2006) 371. [14] L. Guo, Y. Qiu, G. Zhang, G.J. Zheng, P.K.S. Lam, X. Xi, Environ. Pollut. 152 (2008) 604. [15] Directive 2008/105/CE. [16] A. Covaci, A. Gheorghe, O. Hulea, P. Schepens, Environ. Pollut. 140 (2006) 136. [17] M. Alaee, D.B. Sergeant, M.G. Ikonomou, J.M. Luross, Chemosphere 44 (2001) 1489. [18] S. Corsolini, A. Covaci, N. Ademollo, S. Focardi, P. Schepens, Environ. Pollut. 140 (2005) 371. [19] F.J. Santos, M.T. Galceran, Trac-Trend Anal. Chem. 21 (2002) 672. [20] L. Ramos, J.J. Ramos, U.A.Th. Brinkman, Anal. Bioanal. Chem. 381 (2005) 119. [21] P. Deveau, V.M. Mallet, Int. J. Environ. Anal. Chem. 75 (1999) 331. [22] J. Christensen, M. Glaius, M. Pécseli, J. Platz, G. Pritzl, Chemosphere 47 (2002) 631. [23] G. Wittman, T. Huybrechts, H. Van Langenhove, J. Dewulf, H. Nollet, J. Chromatogr. A 993 (2003) 71. [24] R. Kazda, J. Hajslová, J. Poustka, T. Cajka, Anal. Chim. Acta 520 (2004) 237. [25] S. Bayen, H. Kee Lee, J.P. Obbard, J. Chromatogr. A 1035 (2004) 291. [26] W. Vetter, M. Weichbrodt, K. Hummert, D. Glotz, B. Luckas, Chemosphere 37 (1998) 2439. [27] C. Nerín, R. Batlle, M. Sartaguda, C. Pedrocchi, Anal. Chim. Acta 464 (2002) 303. [28] E. Eljarrat, A. de la Cal, D. Raldua, C. Duran, D. Barcelo, Environ. Pollut. 133 (2005) 501. [29] E. Björklund, S. Sporring, K. Wiberg, P. Haglund, C. von Holst, Trac-Trend Anal. Chem. 25 (4) (2006) 318. [30] P. Suchan, J. Pulkrabová, J. Hajslová, V. Kocourek, Anal. Chim. Acta 520 (2004) 193. [31] A.R. Long, M.D. Crouch, S.A. Barker, J. Assoc. Off. Anal. Chem. 74 (1991) 667. [32] A.R. Long, M.M. Solimna, S.A. Barker, J. Assoc. Off. Anal. Chem. 74 (1991) 493. [33] A. Martinez, M. Ramil, R. Montes, D. Hernanz, E. Rubí, I. Rodríguez, R. Cela Torrijos, J. Chromatogr. A 1072 (2005) 83. [34] B. Gómara, L. Herrero, L.R. Bordajandi, M.J. González, Rapid Commun. Mass Spectrom. 20 (2006) 69. [35] S.A. Barker, J. Chromatogr. A 885 (2000) 114. [36] E.M. Kristenson, L. Ramos, U.A.Th. Brinkman, Trac-Trend Anal. Chem. 25 (2006) 96. [37] N. Furusawa, Anal. Bioanal. Chem. 378 (2004) 7. [38] P. Campins-Falcó, J. Verdú-Andres, A. Sevillano-Cabeza, C. Molins-Legua, R. Herráez-Hernandez, J. Chromatogr. A 1211 (2008) 13. [39] L. Pensado, M.C. Casais, M.C. Mejuto, R. Cela, J. Chromatogr. A 1077 (2005) 103. [40] J. Pawliszyn, Solid Phase Microextraction, Wiley–VCH, New York, 1997.
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Miniaturized matrix solid phase dispersion procedure and solid phase microextraction for the analysis of organochlorinated pesticides and polybrominated diphenylethers in biota samples by gas chromatography electron capture detection Y. Moliner-Martinez a , P. Campíns-Falcó a,∗ , C. Molins-Legua a , L. Segovia-Martínez b , A. Seco-Torrecillas b a b
Department of Analytical Chemistry, Faculty of Chemistry, University of Valencia, C/Dr Moliner, 50, 46100 Burjassot, Valencia, Spain Department of Chemical Engineering, School of Engineering, University of Valencia, C/Dr Moliner, 50, 46100 Burjassot, Valencia, Spain
a r t i c l e
i n f o
Article history: Received 11 June 2009 Received in revised form 29 July 2009 Accepted 10 August 2009 Available online 13 August 2009 Keywords: Organochlorinated pesticides (OCPs) Polybrominated diphenylethers (PBDEs) Biota Matrix solid phase dispersion (MSPD) Solid phase microextraction (SPME) GC
a b s t r a c t This work has developed a miniaturized method based on matrix solid phase dispersion (MSPD) using C18 as dispersant and acetonitrile–water as eluting solvent for the analysis of legislated organochlorinated pesticides (OCPs) and polybrominated diphenylethers (PBDEs) in biota samples by GC with electron capture (GC-ECD). The method has compared Florisil® -acidic Silica and C18 as dispersant for samples as well as different solvents. Recovery studies showed that the combination of C18–Florisil® was better when using low amount of samples (0.1 g) and with low volumes of acetonitrile–water (2.6 mL). The use of SPME for extracting the analytes from the solvent mixture before the injection resulted in detection limits between 0.3 and 7.0 g kg−1 (expressed as wet mass). The miniaturized procedure was easier, faster, less time consuming than the conventional procedure and reduces the amounts of sample, dispersant and solvent volume by approximately 10 times. The proposed procedure was applied to analyse several biota samples from different parts of the Comunidad Valenciana. © 2009 Elsevier B.V. All rights reserved.
1. Introduction The presence of organochlorinated pesticides (OCPs) and polybrominated diphenylethers (PBDEs) has been investigated in different types of environmental samples such as fish tissue [1–3], atmospheric [4,5], sediment [6], soil [7] and water samples [8–10]. There is an increasing interest in studying biota samples containing these compounds because of their bio-accumulation and persistence [11,12]. Simultaneous determination of both families of compounds has been also investigated in fishes and food web [13,14]. Thus, legislation controls their content in different matrices; in fact some of them are included in the list of priority organic pollutants (POPs) [15]. The complexity of biota samples (fatty matrix) and the low concentrations of OCPs and PBDEs require extraction, clean-up, preconcentration [16–18] and chromatographic techniques [19]. In the last few years miniaturized procedures have been developed [20,21] in order to minimize the environmental effect that analytical processes could cause. Conventional techniques such as Soxhlet [1,2,22], or liquid–liquid extraction (LLE) [16,23,24]
∗ Corresponding author. Tel.: +34 96 3543002; fax: +34 96 3544436. E-mail address: [email protected] (P. Campíns-Falcó). 0021-9673/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2009.08.021
require higher sample sizes, volumes of solvent and time than other techniques such as microwave assisted extraction (MAE) [25,26], supercritical fluid extraction (SFE) [27], pressurised liquid extraction (PLE)[28–30] or matrix solid phase dispersion (MSPD) [31–34]. The simplicity of MSPD [35,36] makes this technique one of the most advantageous in the miniaturization field. MSPD is based on a blending process between the solid samples with an abrasive solid support material. After this, the mixture is transferred to an empty cartridge to form a packing column and finally the analytes are eluted by using the appropriate solvent. The major factors affecting MSPD are the solid support material, the solvent and the matrix. In fact, MSPD methodology using acidic silica and C18 has been employed to the isolation of different drugs and pollutants in food and biological matrices [37]. MSPD miniaturized can be achieved by reducing the amount of sample (approximately 0.5 g) [35] and the corresponding reduction of sorbent, solvent (5–20 mL) and time (less than 1 h). This contrasts with the sample amount (i.e. 10 g) and several hundred of millilitres of solvent and several hours usually required for sample preparation by most traditional method (LLE) or Soxhlet extraction. Also this is a simple procedure that does not require special instrumentation. Until now, only a few authors have studied the MSPD miniaturization [34,36,38], and as far as we know, no miniaturized procedure to simultaneous determination of OCPs and PBDEs in
6742
Y. Moliner-Martinez et al. / J. Chromatogr. A 1216 (2009) 6741–6745
biota samples has been developed. Two different solid support materials were investigated, one using Florisil® and silica, another one using C18 and the last one using C18–Florisil® . SPME and GC-ECD was used as preconcentration, separation and detection system. The proposed procedure was applied to several biota samples, particularly mussels and cockles.
mode employed was splitless. When using the GC-FID the column temperature was held at 90 ◦ C for 1 min, then raised to 220 ◦ C (15 ◦ C min−1 ) and finally raised to 300 ◦ C (8 ◦ C min−1 ) and held for 10 min. The temperature program for the GC-ECD was: initial temperature 90 ◦ C for 1 min, then raised to 190 ◦ C (15 ◦ C min−1 ) and finally raised to 300 ◦ C (15 ◦ C min−1 ) and held for 7 min. 5 L was employed as injection volume (unless stated otherwise).
2. Experimental 2.4. Extraction procedure 2.1. Reagents and materials Hexane was obtained from Carlo Erba (Rodano, Italy) and methanol, ethyl acetate and acetonitrile were from J.T. Baker (Deventer, The Netherlands). Sulphuric acid was purchased from Merck (Darmstadt, Germany). The solid support materials Florisil® (60–100 mesh), silica (230–400 mesh, 60 Å) and silica (70–230 mesh, 40 Å) were obtained from Aldrich (Steinheim, Germany) and Bondesil C18 (40 m) was purchased from Varian (Habor City, CA, USA). Polyethylene solid phase extraction (SPE) cartridges (12 and 3 mL), frits and the solid phase microextraction (SPME) assembly with extraction fibers coated with polydimethylsiloxane-divinylbencene (PDMS-DVB, 65 m) were from Supelco (Bellefonte, PA, USA). Silica and Florisil® were activated at 110 ◦ C for 48 h. Acidic silica (44% sulphuric acid) was prepared by mixing neutral silica (70–230 mesh, 40 Å) in the appropriate volume of concentrated sulphuric acid. The corresponding amount of C18 was set in a cartridge and was conditioned by passing it through 2 mL of methanol and 2 mL of Milli-Q water. 2.2. Standard and samples Isomers of hexachlorocyclohexane (lindane (␥-HCH), ␣-HCH, HCH and ␦-HCH) and hexachlorobutadine (HCBu) were obtained from Rielden de Häen (Seelze, Germany). Hexachlorobencene (HCBe) was obtained from Fluka (Steinheim, Germany). Stock standard solutions of OCPs (65 mg L−1 ) were prepared in hexane and methanol when using GC-FID or GC-ECD, respectively. Polybrominated diphenylether mixture was obtained from Dr. Ehrenstorfer GmbH (Augsburg, Germany). The PBDEs determined were 2,4,2 ,4 -tetra-BDE (BDE 47), 2,4,5,2 ,4 -penta-BDE (BDE 99) and 2,4,6,2 ,4 -penta-BDE (BDE 100). Concentrations of the stock standard solution (prepared in hexane or methanol) were 40, 12 and 40 mg L−1 for BDE 47, BDE 100 and BDE 99, respectively. The concentration of the working standard solutions used with FI detection was from 0.5 to 5 mg L−1 for OCPs and PBDEs from 0.4 to 4.0 mg L−1 for BDE 47 and BDE 99 respectively and from 0.12 to 1.2 mg L−1 for BDE 100. These solutions were prepared by diluting the stock solutions in hexane. The concentrations of these solutions when using ECD were between 0.03 and 1.2 g L−1 for all the analytes. Several lyophilised biota samples (mussels and cockles) collected from the Mediterranean sea, all over the Comunidad Valenciana coast were analysed. Samples were stored in polyethylene containers at 4 ◦ C.
2.4.1. Procedure 1: Florisil® and silica as solid support material Based on Ref. [33], samples were processed by following the procedure: between 0.05 and 0.5 g of fortified samples was placed in a glass mortar and dispersed with Florisil® 0.15 g with a pestle for 5 min (homogeneous mixture). The blended mixtures were transferred to 3 mL SPE cartridge containing a coarse frit and 1 g of acidic silica and analytes were eluted with 2 mL of hexane. After, the extract was passed through a second 3 mL SPE cartridge containing a coarse frit and 0.2 g of neutral silica (230–400 mesh, 60 Å) and collected with 2 mL of the eluent (mixtures of ethylacetate:hexane or hexane). Finally, 5 L of this extract (0.5 mg L−1 for OCPs and 4.0, 1.2 and 4.0 mg L−1 for BDE 47, BDE 100 and BDE 99 respectively) was injected into the GC-FID. 2.4.2. Procedure 2: C18 as solid support 0.1 g of dry sample was placed in a glass mortar and blended with 0.4 g of C18 for 5 min. The homogenised samples were transferred to a 3 mL SPE cartridge containing a coarse frit. A clean-up step with Milli-Q water (0.5 mL) was included. After the sorbent being dried with air, hexane was employed as eluting solvent (1.3 mL). Finally, a volume of 5 L was injected into the GC-FID system. The spiked concentrations were the same for OCPs and 10 times lower for PBDEs than those used in Procedure 1. In this case, two preconcentration procedures were studied. In the first procedure, the extracts were concentrated to 150 L with a nitrogen stream in order to improve sensitivity and 5 L was injected in the GC-FID system. In the second case, a SPME step was included. A PDMS-DVB fiber was immersed into the extracts for 45 min under continuous stirring. The analytes were directly injected through the fiber (desortion time 2 min). Additionally, the FI detector was substituted by an ECD detector. 2.4.3. Procedure 3: C18–Florisil® as solid support This procedure was based on a previously developed MSPD procedure by this research group described for polycyclic aromatic hydrocarbons [38]. In this procedure 0.1 g of lyophilised samples were placed in a glass mortar and blended with 0.4 g of the C18 phase (5 min). Then, the mixture was transferred to a 3 mL SPE tube containing 0.1 g of Florisil® phase. The analytes were desorbed by flusing 1.2 mL of acetonitrile. Then, the extracts were diluted with 2.6 mL of water and then, SPME was carried out in the conditions described in Procedure 2 and EC detection was employed. 3. Results and discussion
2.3. Chromatographic equipment and experimental conditions 3.1. Optimization of the miniaturized MSPD procedure and SPME Gas chromatography was carried out with a Focus gas chromatograph equipped with a flame-ionization detector (GC-FID) and an electron capture detector (GC-ECD) from Thermo (Waltham, MA, USA). A capillary column TRB-5 (30 m × 0.25 mm i.d.), film thickness 0.25 m (Teknokroma, Barcelona, Spain) was used for the separation of the analytes. The carrier gas was nitrogen at a flow of 2.2 mL min−1 for the GC-FID and 60 mL min−1 for GC-ECD. The temperature of the injection was kept at 260 ◦ C and the detector temperature (FID and ECD) was 300 ◦ C. The injection
Several dispersant and fat retainer phases were studied: Florisil® –silica, C18 and C18–Florisil® for MSPD. The comparison of the extraction efficiencies was carried out using hexachlorobenzene, lindane, BDE 47, 99 and 100 as target analytes. The eluted compounds were directly injected into the GC or concentrated by evaporating the extracts and redissolving in low volumes of solvents. SPME was also tested as preconcentration technique. FID and ECD detectors were used.
Y. Moliner-Martinez et al. / J. Chromatogr. A 1216 (2009) 6741–6745
6743
Fig. 1. Extraction efficiencies (n = 3) using (A) Florisil® –silica procedure (0.05 g of sample, 0.15 g Florisil® , 5 g acidic silica, 0.2 g neutral silica and 2 mL of ethylacetate:hexane (30:70) as eluent), (B) C18 procedure without and with the concentration step and (C) C18–Florisil® procedure (for more details, see Section 2).
Following Procedure 1, fortified samples with OCPs and PBDEs were processed. HCBe and lindane were used in an attempt to extend the procedure proposed by Martinez et al. [33] to the simultaneous extraction of OCPs and PBDEs with a miniaturized procedure. Different sample sizes (0.05, 0.25, and 0.5 g) were spiked with the target analytes. The recoveries obtained when using 0.05 g of sample and 2 mL of ethylacetate:hexane (30:70) as eluent (n = 3) are shown in Fig. 1A. As can be seen, the values were satisfactory for PBDEs but not for the OCPs. Hexachlorobencene coeluted with a component of the matrix giving recoveries higher than 150% and significative losses on lindane content was observed (recoveries lower than 60%). The increment of the sample amount resulted in lower recoveries (49%) for hexachlorobencene and lindane. Variations in the amount of Florisil® , acidic or neutral silica or elution solvent did not improve those recoveries. Using hexane as eluent (2 mL) PBDEs were extracted independently of the sample size. However, hexachlorobencene and lindane presented low recoveries similar to that shown above. The use of acidic silica not only removed the lipids but also retained the non-polar compounds. This effect can be attributed to the formation of carbon as a sub-product of lipid’s dehydration that retained the analytes [39]. Probably, the use of large amount of solvent would improve the recoveries. Nevertheless this was not a strategy for a miniaturized procedure. Therefore, the use of Florisil® and silica provided good results for PBDEs but not for all the analytes listed in the CE Directive [15]. Procedure 2 evaluated the use of a unique solid support (C18) as dispersant and fat retainer. The extraction efficiencies for OCPs were good for 0.1 and 0.25 g of sample, for a size of 0.5 g either hexachlorobencene or lindane could not be estimated. 0.1 g was selected as the optimum because sample size, the amount of solid support material (0.4 g) and solvent volume (1.3 mL) were minimized. As can be seen in Fig. 1B, PBDEs and OCPs were eluted in a unique hexane fraction with satisfactory extraction efficiencies. The concentration level of the target analytes in biota samples required a preconcentration step. A commonly used way to improve the detection limits is the use of nitrogen streams. In this
particular case, sensitivity improved by a factor of around 7 if the extracts (1.3 mL) were concentrated to dryness and reconstituted with 150 L of hexane. However, the recoveries after this concentration step decreased down 60% (Fig. 1B) for all the tested analytes. In addition, an increase in the baseline noise was also observed (see Fig. 2A and B for a blank and a spiked sample under those conditions, respectively). This fact caused difficulty the identification and quantification of the analytes in biota samples. Thus, an alternative preconcentration system should be studied. The other preconcentration technique evaluated was a SPME procedure in combination with ECD detection. Under these conditions, hexachlorobencene and lindane were practically lost as the recoveries obtained were lower than 22%. Probably, the reason of these recoveries could be the poor efficiency of the SPME fiber to
Fig. 2. Chromatograms obtained for (A) blank and (B) sample of biota (mussel) spiked with the target analytes obtained with Procedure 2 after the concentration of the extracts with GC-FID. 2.5 mg L−1 HCBe, 2.5 mg L−1 lindane, 4 mg L−1 BDE 47, 1.2 mg L−1 BDE 100 and 4.0 mg L−1 BDE 99 (see Section 2.4.2).
6744
Y. Moliner-Martinez et al. / J. Chromatogr. A 1216 (2009) 6741–6745
Table 1 Detection limits (LOD), quantification limit (LOQ) (expressed as wet mass) and reproducibility (%RSD) obtained for the target analytes with the C18–Florisil® MSPD, SPME and GC-ECD (Procedure 3).
HCBu ␣-HCH HCBe -HCH Lindane ␦-HCH BDE 47 BDE 100 BDE 99 a
LOD (g kg−1 )
LOQ (g kg−1 )
%RSDa (n = 3)
0.3 3.7 0.5 4.7 3.0 4.7 3.0 4.6 7.1
1.0 12 1.7 16 12 16 16 15 23
10 5 9 1 1 5 23 19 5
Concentration level: 23 g kg−1 .
adsorb the target analytes from the extracts of hexane as the analytes had low distribution coefficients in hexane extracts (K = Cf /Cs , where Cf is the concentration in the fiber and Cs is the concentration in the sample) [40]. The use of another solvent would result in better extraction efficiencies in the SPME, but the use of a unique solid support limited the solvent. Hexane provided good recoveries of the analytes with low volume and with the minimum elution of undesirable compounds of the matrix. Finally, we studied the efficiency of Procedure 3 based on the use of C18–Florisil® [36]. The combination of C18 as dispersant and Florisil® as fat retainer, elution with acetonitrile, a SPME step and GC-ECD resulted in good recoveries for all the target analytes (Fig. 1C). The combination of C18 and Florisil® allowed the use of low volumes of MeCN as eluent and so good results in the SPME step as high distribution coefficient could be expected for these analytes in the PDMS-DVD fiber. It provided the adequate sensitivity for tested analytes and allowed concentration range between 0.03 and 1.2 g L−1 . Fig. 1C shows the recoveries obtained for these analytes. These values were satisfactory for all the target analytes. Bearing in mind these results, hexachlorobutadiene, ␣-, -, ␦-HCH were included in the study as they are also legislated as highly persistent compounds. Recoveries for hexachlorobutadiene, ␣-, - and ␦-HCH were 55 ± 10, 114 ± 29, 74 ± 28, 130 ± 30, respectively. 3.2. Figures of merit Under the conditions of Procedure 3 good linearity (r2 between 0.9790 and 0.9994) was obtained in the working interval (0.03 and 1.2 g L−1 ). The LODs, LOQs and RSD values (0.6 g L−1 or 23 g kg−1 expressed as wet mass) are given in Table 1. As can be seen, this procedure provided the adequate sensitivity with good reproducibility for all the analytes. Thus, this procedure was selected to analyse biota samples as it was advantageous in terms of detection limits. Additionally, it presented other advantages such as time consuming, reagent consumption and waste generation.
Fig. 3. Chromatograms obtained for (A) blank and (B) spiked sample with all the analytes (30 ng L−1 ) obtained with Section 2.4.3.
3.3. Application to biota samples The proposed procedure was employed to analyse 17 lyophilised biota samples (11 mussels and 6 cockles). The validation of the proposed methodology was done with a recovery study by spiking the samples with a mixture of the target analytes due to the inexistence of standard reference materials for these POPs in biota. The mean recovery values obtained, between 55 ± 10 and 130 ± 30% (n = 17) for all the analytes showing that the measured concentration were in agreement with the added ones. Fig. 3 shows the chromatogram obtained for a blank and a spiked sample (1.15 g kg−1 of each analyte). Only HCBe was found in three samples (mussels) at levels of 2.2 ± 0.2, 2.9 ± 0.3, and 1.8 ± 0.2 g kg−1 (expressed as mean value in wet samples for n = 3). In the other cases the contents of all the studied analytes in all the samples were below the detection limit showing the good health of the coast under evaluation. 4. Conclusions This paper has studied the simultaneous extraction of OCPs and PBDEs from biota samples (mussels and cockles) with a miniaturized MSPD extraction method. The use of C18 as dispersant and Florisil® as fat retainer combined with SPME–GC-ECD yielded to the development of a procedure with numerous advantages com-
Table 2 Comparison of the miniaturized MSPD proposed with other published procedures for OCPs and PBDEs. Parameters
Martinez et al. [33] PBDEs and PCBsa
Long et al. [31,32] OCPsa
This manuscript OCPs and PBDEsa
Sample amount (g) Sorbent (g)
0.5 1.5 g Florisil®
0.5 C18 2 g
0.1 C18 0.4 g Florisil® 0.1 g
Clean-up
5 g acidic silica 2 g neutral silica n-Hexane (20 mL) n-Hexane (20 mL) n-Hexane–diclhoromethane (80:20), 12 mL – >75%
Florisil® 2 g
Solvent (mL)
Preconcentration Recoveries a
Analytes.
Acetonitrile 8 mL
Acetonitrile (1.3 mL)
– 85 ± 3–102 ± 5
SPME 55 ± 10–130 ± 30
Y. Moliner-Martinez et al. / J. Chromatogr. A 1216 (2009) 6741–6745
pared with previously published procedures. Table 2 summarizes some of these advantages. As can be seen, C18–Florisil® demonstrated a reduction in the use of hazardous solvents and analysis time for the analysis of OCPs and PBDEs in the same fraction. Additionally, the use of SPME–GC-ECD provided the enough sensitivity for OCPs and PBDEs analysis in biota samples. The procedure minimized the required sample amount, sorbent material and solvent consumption. Analytical properties such as recoveries or amount detected were also an advantage of the proposed procedure. Acknowledgements The authors would like to thank the Ministerio de Ciencia y Tecnología of Spain for the financial support received for the project CTQ 2008-01329/BQU and the Government of the Region Valenciana (Generalitat Valenciana) for the project “Application of EU water Framework Directive 2000/60/EC on endocrine disrupters and other priority substances in coastal waters in the Region Valenciana”. Y.M.M. expresses her gratitude for her JdC contract (Ministerio de Ciencia e Innovación). References [1] N. Fidalgo-Used, G. Centineo, E. Blanco-González, A.J. Sanz-Mendel, J. Chromatogr. A 1017 (2003) 35. [2] S. Sinkkonen, A. Rantalainen, J. Paasivirta, M. Lahtiperä, Chemosphere 56 (2004) 767. [3] B. Fägström, M. Athanasiadou, I. Athanassiadis, A. Bignert, P. Grandjean, P. Weihe, A. Bergman, Chemosphere 60 (2005) 836. [4] A. Sanusi, M. Millet, P. Mirabel, H. Wortham, Sci. Total Environ. 263 (2000) 263. [5] K. Hyakawa, H. Takatsuki, I. Watanabe, S. Sakai, Chemosphere 57 (2004) 343. [6] J. Boer, C. Allchin, R. Law, B. Zegers, J.P. Boon, Trac-Trend Anal. Chem. 20 (2001) 591. [7] C. Goncalves, M.F. Alpendurada, Talanta 65 (2005) 1179. ˜ [8] V.M. León, B. Alvarez, M.A. Cogollo, S. Munoz, I. Valor, J. Chromatogr. A 999 (2003) 91. [9] H.P. Li, G.C. Li, J.F. Jen, J. Chromatogr. A 1012 (2003) 129. [10] M.G. Ikonomou, M.P. Fernández, Z.L. Hichkman, Environ. Pollut. 140 (2006) 355.
6745
[11] C. Dong, Z. Zeng, M. Yang, Water Res. 39 (2005) 4204. [12] Y.F. Li, R.W. Macdonald, Sci. Total Environ. 342 (2005) 87. [13] S. Corsolini, A. Covaci, N. Ademollo, S. Forcardi, P. Schepens, Environ. Pollut. 140 (2006) 371. [14] L. Guo, Y. Qiu, G. Zhang, G.J. Zheng, P.K.S. Lam, X. Xi, Environ. Pollut. 152 (2008) 604. [15] Directive 2008/105/CE. [16] A. Covaci, A. Gheorghe, O. Hulea, P. Schepens, Environ. Pollut. 140 (2006) 136. [17] M. Alaee, D.B. Sergeant, M.G. Ikonomou, J.M. Luross, Chemosphere 44 (2001) 1489. [18] S. Corsolini, A. Covaci, N. Ademollo, S. Focardi, P. Schepens, Environ. Pollut. 140 (2005) 371. [19] F.J. Santos, M.T. Galceran, Trac-Trend Anal. Chem. 21 (2002) 672. [20] L. Ramos, J.J. Ramos, U.A.Th. Brinkman, Anal. Bioanal. Chem. 381 (2005) 119. [21] P. Deveau, V.M. Mallet, Int. J. Environ. Anal. Chem. 75 (1999) 331. [22] J. Christensen, M. Glaius, M. Pécseli, J. Platz, G. Pritzl, Chemosphere 47 (2002) 631. [23] G. Wittman, T. Huybrechts, H. Van Langenhove, J. Dewulf, H. Nollet, J. Chromatogr. A 993 (2003) 71. [24] R. Kazda, J. Hajslová, J. Poustka, T. Cajka, Anal. Chim. Acta 520 (2004) 237. [25] S. Bayen, H. Kee Lee, J.P. Obbard, J. Chromatogr. A 1035 (2004) 291. [26] W. Vetter, M. Weichbrodt, K. Hummert, D. Glotz, B. Luckas, Chemosphere 37 (1998) 2439. [27] C. Nerín, R. Batlle, M. Sartaguda, C. Pedrocchi, Anal. Chim. Acta 464 (2002) 303. [28] E. Eljarrat, A. de la Cal, D. Raldua, C. Duran, D. Barcelo, Environ. Pollut. 133 (2005) 501. [29] E. Björklund, S. Sporring, K. Wiberg, P. Haglund, C. von Holst, Trac-Trend Anal. Chem. 25 (4) (2006) 318. [30] P. Suchan, J. Pulkrabová, J. Hajslová, V. Kocourek, Anal. Chim. Acta 520 (2004) 193. [31] A.R. Long, M.D. Crouch, S.A. Barker, J. Assoc. Off. Anal. Chem. 74 (1991) 667. [32] A.R. Long, M.M. Solimna, S.A. Barker, J. Assoc. Off. Anal. Chem. 74 (1991) 493. [33] A. Martinez, M. Ramil, R. Montes, D. Hernanz, E. Rubí, I. Rodríguez, R. Cela Torrijos, J. Chromatogr. A 1072 (2005) 83. [34] B. Gómara, L. Herrero, L.R. Bordajandi, M.J. González, Rapid Commun. Mass Spectrom. 20 (2006) 69. [35] S.A. Barker, J. Chromatogr. A 885 (2000) 114. [36] E.M. Kristenson, L. Ramos, U.A.Th. Brinkman, Trac-Trend Anal. Chem. 25 (2006) 96. [37] N. Furusawa, Anal. Bioanal. Chem. 378 (2004) 7. [38] P. Campins-Falcó, J. Verdú-Andres, A. Sevillano-Cabeza, C. Molins-Legua, R. Herráez-Hernandez, J. Chromatogr. A 1211 (2008) 13. [39] L. Pensado, M.C. Casais, M.C. Mejuto, R. Cela, J. Chromatogr. A 1077 (2005) 103. [40] J. Pawliszyn, Solid Phase Microextraction, Wiley–VCH, New York, 1997.