Microwave assisted micellar extraction coupled with solid phase microextraction for the determination of organochlorine pesticides in soil samples

Analytica Chimica Acta 571 (2006) 51–57

Microwave assisted micellar extraction coupled with solid phase microextraction for the determination of organochlorine pesticides in soil samples Daura Vega Moreno, Zoraida Sosa Ferrera, Jos´e J. Santana Rodr´ıguez ∗ Department of Chemistry, Faculty of Marine Sciences, University of Las Palmas de Gran Canaria, 35017 Las Palmas de Gran Canaria, Spain Received 21 March 2006; received in revised form 17 April 2006; accepted 18 April 2006 Available online 29 April 2006

Abstract Microwave assisted micellar extraction (MAME) coupled with solid phase microextraction (SPME) and HPLC-UV determination have been used for the determination of five organochlorine pesticides from agricultural soil samples. A non-ionic surfactant, Polyoxyethlylene 10 Lauryl Ether was used, and the different variables for the optimization of MAME and SPME procedures were studied. This method was applied successfully to the determination of these pesticides in several kinds of agricultural soil samples with different characteristics. Most of the compounds studied can be recovered in good yields with R.S.D. lower than 9% and detection limit ranged between 56–96 ng g−1 for the pesticides studied. © 2006 Elsevier B.V. All rights reserved. Keywords: Organochlorine pesticides; Micellar medium; Extraction methods; Microwave; Solid phase microextraction; Soils; High performance liquid chromatography

1. Introduction Organochlorine pesticides have been used during half century due to their effectiveness as insecticides [1]. These compounds are banned in developed countries since 1970s, but in underdeveloped countries are still used specially for the control of vector-borne diseases as malaria [2,3]. Despite not being used by many countries for some years, they are present in the environment; in this way, the long-term persistence of DDT and its metabolites in soil have been reported [4–7]. Organochlorine pesticides are toxic compounds, which can affect to human health [1]. Their lipophilic nature, hydrophobicity and low chemical and biological degradation rates have led their accumulation in biological tissues and subsequent magnification of concentrations in organisms progressing up the food chain [8]; for these reasons are listed as US Environmental Protection Agency (EPA) priority pollutants [9]. Analysis for environmental organic pollutants require complex procedures

∗

Corresponding author. Tel.: +34 928452915; fax: +34 928452922. E-mail address: [email protected] (J.J.S. Rodr´ıguez).

0003-2670/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2006.04.046

involving several steps, such as clean-up extraction and preconcentration prior to their quantification [10]. The most frequently used methods for the extraction of organochlorine pesticides from solid samples are Soxhlet and ultrasonic extraction [10], which required high amounts or organic solvents and long analysis times. Other analysis methods have been developed, including accelerated solvent extraction (ASE) [11] and microwave assisted solvent extraction (MAE) [12]. In the last decade, MAE has become a viable alternative to the conventional techniques and exhibits many substantial improvements in analytical sample preparations, as it requires much lower volume of organic solvent, reduces extraction time and lets prepared multiple samples in one step [13–15]. However, organic solvents are used as extractant in all of them. An alternative to these types of extractants would be the use of micellar systems. The extraction of organic compounds from solid samples using micellar medium, so called microwave assisted micellar extraction (MAME), offers advantages, such as safety, low cost and its low toxicity (Table 1) [16,17]. At the same time solid phase microextraction (SPME) is a new variation of extraction techniques, which has been used mainly for the analysis of pollutants in environmental water

52

D.V. Moreno et al. / Analytica Chimica Acta 571 (2006) 51–57

Table 1 Comparison between extraction methods for organochlorine pesticides Extraction method

Organic solvent

Volume of solvent (ml)

Time

Toxicity

References

Soxhlet Ultrasonic ASEa MAEb MAMEc

Hexane, acetone, dichloromethane Hexane, acetone Hexane, acetone, dichloromethane Hexane, acetone No

30–100 10–50 15–40 10–30 8

6–24 h 1–3 h 20–60 min 5–20 min 2–14 min

High High High Average Low

[10,15] [10,11] [11] [12–15] [16,17]

a b c

Accelarated solvent extraction. Microwave assisted extraction. Microwave assisted micellar extraction.

samples [18], but this methodology can be coupled with other techniques, as the microwave extraction, for the analysis of soil samples [19]. SPME allows the simultaneous extraction and preconcentration of analytes from a sample. The principle of SPME is equilibration of the analytes between the sample matrix and the organic polymeric phase usually coating a fused-silica fibre. The amount of the analyte absorbed by the fibre is proportional to the initial concentration. Normally, the application of SPME has been performed in combination with gas chromatography, however, it is possible coupled with HPLC [20]. The advantage using HPLC is that compounds with low volatility or thermal labile compounds can be analysed. An organic solvent (static desorption) or the mobile phase (dynamic mode) is used to disorb the analytes from the SPME fibre. In this paper, we present the application of microwave assisted micellar extraction coupled with solid phase microextraction and HPLC determination for the analysis of organochlorine pesticides in agricultural soil samples. In the microwave assisted extraction, we used Polyoxyethylene 10 Lauryl Ether (POLE) as extractant and several variables as surfactant concentration, microwave time and power were optimized. The parameters which affect to the absorption and desorption of the SPME procedure were investigated. Finally, the effect of different soil characteristics, such as pH, organic carbon and clay mineral contents were studied. 2. Experimental 2.1. Reagents Organochlorine pesticides were obtained from Cerilliant Corporation (provided by LGC Promochem, Barcelona, Spain) and prepared by dissolving appropriate amounts of the commercial products in methanol. The organochlorine pesticides are listed in Table 2 (numbers and abbreviations identify the compounds in figures). The non-ionic surfactant used in this study, Polyoxyethylene 10 Lauryl Ether, was obtained from Sigma–Aldrich (Madrid, Spain) and prepared in bidistilled water. The 60 ␮m Polydimethylsiloxane/Divinilbenzene (PDMS/ DVB) fibre of the SPME system was provided by SUPELCO (Madrid, Spain). Methanol HPLC-grade was obtained from Panreac Qu´ımica S.A. (Barcelona, Spain).

All solvents and analytes were filtered through a 0.45 ␮m cellulose acetate membrane filter and ultra-high-quality water obtained by a Milli-Q water purification system (Millipore, USA) was used throughout. 2.2. Apparatus The chromatograph system consists of a Varian pump fitted with a Varian Autosampler 410 with a volume selector, a Column Valve Module with an internal oven and a Varian PDA Detector. The system and the data management were controlled by Star software from Varian (Varian Inc., Madrid, Spain). The stationary-phase column was a Varian Microsorb-MV 100 C18, 250 mm × 4.6 mm, 8 ␮m particle diameter. The analytical column was inserted in the column module and thermostated at 30 ± 0.2 ◦ C. The microwave oven used in this study was a Multiwave with a 6 EVAP rotor and 6 MF100 vessels (Anton Paar, Graz, Austria). 2.3. Procedures 2.3.1. Soil characteristics Different soils samples were collected from different agricultural locations of Gran Canaria (Canary Island, Spain); Tafira and Santa Br´ıgida soils from the Northeast of the island and Valleseco soils from the Centre. These soils do not show any signals of the pesticides studied in the chromatogram when a blank of them was done under our chromatographic conditions. To determine the soil pH and conductivity, 5 g of each soil was mixed with 20 ml of bidistilled water; the slurry was stirred and then allowed to separate the supernatant, pH and conductivity was measured potentiometrically [21]. The organic matter (OM) content was determined by the Sauerlandt method (organic Table 2 List of organochlorine pesticides, wavelenghts and retention times Compound

Abbrevation

λ (nm)a

tR (min)b

4,4 -Dichlorodiphenyldichloroethane Dieldrin 4,4 -Dichlorodiphenyltrichloroethane 2,4 -Dichlorodiphenyltrichloroethane 4,4 -Dichlorodiphenyldichloroethylene

4,4 -DDD Dieldrin 4,4 -DDT 2,4 -DDT 4,4 -DDE

238 220 238 238 238

8.3 9.3 13.0 14.5 16.4

a b

Detection wavelenghts. Retention time.

D.V. Moreno et al. / Analytica Chimica Acta 571 (2006) 51–57

53

Table 3 Soil characteristics Soil

Tafira Sta Brigida Valleseco I Valleseco II

Agricultural uses

Garden Pine forest Potatoes Potatoes

Particle size (%) 0.3 mm

0.2 mm

0.15 mm

≤0.1 mm

44.4 56.5 36.33 25.16

17.2 17.8 16.43 13.17

14.6 12.8 15.19 14.38

23.8 12.8 32.06 47.29

matter oxidation by potassium dichromate and sulphuric acid) [22]. The physical–chemical characteristics of the soils samples are given in Table 3. 2.3.2. Preparation of spiked soils The soil samples were air-dried at room temperature for more than 2 weeks and sifted to a particle size of less than 0.3 mm. These samples were spiked as follows: 2 g of each soil was spiked with a solution of 4,4 -DDD, 4,4 -DDT, 2,4 -DDT and 4,4 -DDE to obtain a final concentration of 1.2 ␮g g−1 of each analyte and with a volume of Dieldrin solution to obtain a concentration of 2 ␮g g−1 . The samples were then stored in amber bottles at room temperature for 24 h before analysis, in order to obtain a dry and homogenous sample. 2.3.3. Microwave assisted micellar extraction Once the soil sample was transferred to the vessel, the optimized volume and concentration of surfactant were added and the soil was subjected to the MAME process [16,23]. The vessels were closed for avoiding the loss of extract during the extraction process at microwave power and time optimized; after this, the vessels were allowed to cool first 10 min with the microwave fan and after another 5 min at room temperature outside the microwave before being opened. The extract solution was removed of the vessels and filtered with a 0.45 ␮m syringe-driven filter and transferred to a glass tube. 2.3.4. Solid phase microextraction (SPME) process Two millilitre of the MAME extract and 2 ml of bidistilled water were introduced in a 4 ml vial for the SPME extraction under the optimized conditions [24]. The desorption of the compounds was done in a 200 ␮l glass vial with 100 ␮l of methanol during 4 min. 2.3.5. Liquid chromatography analysis with UV detection The analysis of the extracted samples was carried out using high performance liquid chromatography with UV detection [20,25,26]. The separation and determination of the compounds under study were performed by injecting 50 ␮l of SPME extract into the liquid chromatograph and the absorbance for each analyte, corresponding to the maximum wavelength, was then measured (Table 2). The mobile phase used for the separation of the five organochlorine pesticides mixture was methanol–water (85:15%, v/v) isocratic with a flow-rate of 1 ml min−1 .

pH

Organic matter (%)

Conductivity (␮S/cm)

8.3 5.9 4.84 3.91

4.8 3.9 4.4 6.2

292 483 202 100

A linear relationship was obtained between peak areas and the analyte concentrations in the range of 30–500 ␮g l−1 , with high correlation coefficients (≥0.995). 2.3.6. Statistical analysis The experimental design of the MAME extraction was performed using Statgraphics Plus software, version 5.1 (Manugistic, Rockville, MD, USA). The studies of partial and bivariate correlations were done using SPSS 11.0 (Chicago, IL, USA). 3. Results and discussion 3.1. Optimization of MAME methodology Optimization experiments were performed using a soil sample from Valleseco I with 4.4% organic matter and pH 4.8 with POLE as extractant for the five organochlorine pesticides under study. Parameters that can influence the MAME process are amount of soil, extractant volume and concentration, irradiation time and power. Optimization of microwave assisted extraction in several applications and different studies have used multivariable factorial design [27,28] or central composite design to find optimal conditions [12]. In this study, we have chosen 2 g of sample to carry out the optimization. The variables were studied with a 23 factorial design in order to obtain the influence of each variable on the recovery and the variables correlations each other. The experimental design parameters in the screening design are shown in Table 4. The analysis of the obtained results for each run (Table 5) shows the different correlations between variables and their influence on the recovery, where it can be observed that all parameters show the same sign for all pesticides. Table 4 Design matrix in the screening design (23 ) Run 1 2 3 4 5 6 7 8

Power (W)

Time (min)

Surfactant concentration (%, v/v)

1000 100 1000 1000 100 100 1000 100

14 2 2 14 14 14 2 2

1 1 5 5 5 1 1 5

54

D.V. Moreno et al. / Analytica Chimica Acta 571 (2006) 51–57

Table 5 Relevant variables and correlation between variablesa Correlation Power Time Concentration Power × conc. Time × conc. Power × time a

4,4 -DDD

Dieldrin

4,4 -DDT

2,4 -DDT

4,4 -DDE

0454 0555 0555 −0441 −0340 −0337

0701 0275 0443 −04853 −01409 −02809

0479 0193 0677 −05018 −01814 −01074

0562 0175 0593 −05004 −01312 −0121

0541 0252 0555 −04292 −01738 −01675

The maxima correlations are +1 and −1.

The volume of surfactant wasn’t included in experiment design because it principally depends on the amount of soils, and it is constant. We chose a volume of 8 ml for 2 g of soils, in order to ensure the sample totally covered by the surfactant. The results of the experiment design and the correlation between variables with the recovery, and each other, show that the surfactant concentration is the variable, which more affect to the recovery and is correlated with the microwave power. The surfactant concentration was optimized using a response surface with 32 factorial design with duplicate of central points where is represented surfactant concentration, microwave power and recovery at extraction time of 8 min. The results obtained for the case of 4,4 -DDT using POLE surfactant are shown in Fig. 1, the optimum surfactant concentration is 5% (v/v). For the rest of compounds the results are similar. The power is the other variable, which more affects to the recovery, and is correlated too with the microwave time. These two variables were going to optimized together using another 32 factorial design with duplicate of central points, where the recovery is the dependent variable. The experiment involves 10 run by duplicate and other variables involved in the extraction process were kept constant in their optimized values: surfactant volume (8 ml), surfactant concentration (5%, v/v) and soil amount (2 g). The concentration of pesticides spiked was also kept constant as were indicated. Fig. 2 shows the response surface for 4,4 DDT where we can observe that 1000 W for microwave power

Fig. 1. Response surface for the effect of surfactant concentration and power on the extraction of 4,4 -DDT using POLE surfactant at extraction time of 8 min.

Fig. 2. Response surface for the effect of power and time on the extraction of 4,4 -DDT using POLE surfactant (5%, v/v).

during 14 min are the optimum conditions. Higher times were not used because of decreasing recovery for some compounds, probably due to volatilization or degradation of these analytes. The behaviour was similar for the rest of pesticides that were extracted with this surfactant. 3.2. Optimization of SPME procedure The efficiency of analytes extraction by SPME can vary widely depending upon the matrix nature, the time period of absorption, temperature, salt addition to the sample and pH [29,30]. To optimize the extraction process, one of the most important step is the determination of the time needed until it reaches equilibrium between the sample matrix and the coating of the fibre. Duplicated samples containing all the analytes were extracted with a PDMS/DVB fibre for periods of time ranging from 20 to 60 min. Fig. 3a shows the time profile obtained for the analytes 4,4 -DDT, 4,4 -DDD and 4,4 -DDE as representatives of the behaviour of all analytes; the rest of compounds gave similar results. Most of analytes reach the equilibrium after about 40 min; with this time, it is obtained a good reproducibility and an acceptable analysis time. Desorption was done using static mode by immersion of the PDMS/DVB fibre into the vial. A desorption time interval was studied between 2 and 12 min, checking that from 4 min wasn’t produced a significant increase in the peak area, it shows that from this time, the desorption of the analytes has been produced and higher times do not get to a signal improvement (Fig. 3b). For this, a desorption time of 4 min was chosen for later studies. The temperature of extraction can play an important role on the absorption of analytes, because it influences the mass transfer rates and the partition coefficients of the analytes. Extraction temperature influence was studied in the range 25–60 ◦ C and it is observed that the peak area of analytes decreases with increasing temperature (Fig. 3c). It is due to the absorption is generally an exothermic process, and the amount of analyte absorbed decreases with increasing temperature [31]; therefore

D.V. Moreno et al. / Analytica Chimica Acta 571 (2006) 51–57

55

Fig. 3. Optimized variables for SPME extraction using POLE coupling with PDMS/DVB fibre.

we decided to carry out all the measurements at room temperature (25 ◦ C). Another extraction parameter, which has a well-establish effect in the extraction procedure is the ionic strength of the sample by addition of salt. This effect has been also studied in SPME by addition of NaCl. Fig. 3d shows the effect of ionic strength on the peak area of the analytes, between 0% and 20% (w/v) NaCl concentrations. It can be seen that for these compounds salt addition doesn’t improve the analyte absorption; because of this, we eliminated the addition of salt in later studies. Finally, and as resume, the optimized parameters were: 40 min for absorption time, 4 min for desorption time, room temperature and non-addition of salt.

Fig. 4. Chromatogram of an extract of a spiked Valleseco I soil after MAMESPME procedure. Chromatographic conditions as described in the text. The numbering refers to Table 2.

3.3. Liquid chromatographic analysis The analysis of the extracted samples was carried out using high performance liquid chromatography with UV detection [20,25]. The chromatogram obtained for the mixture of pesticides extracted of a spiked Valleseco I soil sample is shown in Fig. 4. It can be observed that the mobile phase used allows a good separation of analytes and a short analysis time. The corresponding calibration curves were done with a surfactant concentration of 5% (v/v), PDMS/DVB fibre under the optimum conditions and the corresponding analyte concentrations, ranged between 30 and 500 ␮g l−1 . A linear relationship was obtained between peak areas and the analyte concentrations, with high correlation coefficients (≥0.995) in all cases. With the aim to study the precision of the method MAME coupled to SPME-HPLC, it was applied to the analysis of six samples containing the mixture of pesticides under the selected

conditions. The R.S.D. values are listed in Table 6, where can be observed the R.S.D. values are lower than 9% in all cases. Method detection limits, calculated as twice the noise of each pesticide [32], are in the range 56–96 ng g−1 (Table 6) for all compounds studied. Table 6 Analytical parameters for the determination of pesticides under study using MAME-SPME procedurea Compound

R.S.D. (%)

LOD (ng g−1 )

4,4 -DDD Dieldrin 4,4 DDT 2,4 DDT 4,4 -DDE

7.4 5.5 7.3 8.4 7.9

68 96 72 84 56

a

n = 6.

56

D.V. Moreno et al. / Analytica Chimica Acta 571 (2006) 51–57

Table 7 Extraction and determination of pesticides compounds using MAME coupled SPME-HPLC procedure in a certified reference materiala Compound

Reference valueb

R.S.D. (%)

Confidence interval

Concentration found (MAME-SPME)

4,4 -DDD Dieldrin 4,4 -DDT 4,4 -DDE

1.531 1.863 1.060 1.520

0.476 0.655 0.275 0.410

1.294–1.767 1.539–2.186 0.926–1.195 1.325–1.715

1.637 1.741 1.082 1.774

± ± ± ±

0.046 0.038 0.051 0.044

All values are expresed in ␮g g−1 . The pesticides values in the sample were certified by USEPA SW846 (3rd edition). Extraction methods 3540A/3541 (soxhlet), 3550 (sonication) and analysis method 8081. a

b

In order to prove the validity of the optimized method MAME coupled SPME-HPLC, it was applied to the extraction and determination of the mixture of pesticides 4,4 -DDD, 4,4 -DDT, 4,4 DDE and Dieldrin present in a certified soil sample (CRM804050). The results obtained are showed in Table 7, where we can observe that the recoveries obtained falling within the certificated range for all compounds analysed, so that the proposed extraction and determination procedure is suitable. 3.4. Applications 3.4.1. Matrix effect study In order to study the influence of soil characteristics on the optimized method MAME coupled SPME-HPLC, it was applied to the extraction and determination of pesticides mixture in four different types of agricultural soils from different agricultural areas of Gran Canaria island (Canary Island, Spain), with different levels of organic matter, pH, conductivity and texture (Table 3). Soil samples used in this study do not present peaks of organochlorine pesticides or interferences in the chromatograms. These samples were then spiked with the mixture of pesticides to obtain a final concentration of 1.2 ␮g g−1 for 4,4 -DDD, 4,4 -DDT, 2,4 -DDT and 4,4 -DDE and 2 ␮g g−1 for Dieldrin to investigate the matrix effect on this method. The selected concentration level for spiking was the one typical of acute pollution events that may occur in this kind of soil [33]. The soil samples were stored in the dark at room temperature for 24 h before analysis.

Fig. 5 shows the recovery obtained for the analytes under study in these different kinds of soils. In general, the extraction efficiency is satisfactory for all compounds except for dieldrin in Sta. Brigida soil where the extraction of this analyte was hardly affected by the matrix. Therefore, we can say that the proposed method MAME coupled to SPME-HPLC is applicable to determine the target analytes in real soil samples in the range of concentrations studied (120–2000 ng g−1 ). 4. Conclusions MAME coupled with SPME-HPLC was satisfactory applied to the determination of organochlorine pesticides in agricultural soil samples at the level of concentrations studied. The different parameters, which can affect both MAME and SPME extraction procedures were optimized. The combination of microwave assisted extraction using surfactant as extractant with solid phase microextraction provides a viable alternative to other extraction techniques in real soil samples, because the methodology developed reduced amounts of time and extractant required, thus lowering the cost dramatically. Acknowledgements Daura Vega thanks to the Spanish Ministry of Education and Science for her PhD Student Grant (FPU). References

Fig. 5. Recoveries obtained after MAME-SPME procedure for the pesticides for four different soils using POLE as extractant.

[1] M. Maroni, C. Colosio, A. Ferioli, A. Fait, Toxicology 143 (2000) 61. [2] E. Kusvuran, O. Erbatur, J. Hazard. Mater. 106B (2004) 115. [3] Stockholm Convention on Pops: A Global Public Health Treaty, Sweden, 2001. [4] H.M. Shivaramaiah, I.O.A. Odeh, I.V. Kennedy, J.H. Skerritt, J. Agric. Food Chem. 50 (2002) 5360. [5] W. Chen, L. Zhang, L. Xu, X. Wang, L. Hong, H. Hong, Mar. Pollut. Bull. 45 (2002) 385. [6] A. Nawab, A. Aleem, A. Malik, Bioresour. Technol. 88 (2003) 41. [7] M. Megharaj, D. Kantachote, I. Singleton, R. Naidu, Environ. Pollut. 109 (2000) 35. [8] H. Mwevura, O. Othman, G.L. Mhehe, Mar. Pollut. Bull. 45 (2002) 262. [9] U.S. Environmental Protection Agency, Toxic Substance Control Act, US EPA, Washington, DC, 1979. [10] J. Boer, R.J. Law, J. Chromatogr. A 1000 (2003) 223. [11] K. Li, M. Landrialult, M. Fingas, M. Llompart, J. Hazard. Mater. 102 (2003) 93.

D.V. Moreno et al. / Analytica Chimica Acta 571 (2006) 51–57 [12] O. Zuloaga, N. Etxebarria, L.A. Fern´andez, J.M. Madariaga, Fresenius J. Anal. Chem. 367 (2000) 733. ´ [13] V. L´opez-Avila, R. Young, Anal. Chem. 66 (1994) 1097. [14] F.A. Esteve-Turrillas, C.S. Aman, A. Pastor, M. de la Guardia, Anal. Chim. Acta 522 (2004) 73. [15] M. Barriada-Pereira, E. Concha-Gra˜na, M.J. Gonz´alez-Castro, S. Muniategui-Lorenzo, P. L´opez-Mah´ıa, D. Prada-Rodr´ıguez, E. Fern´andez-Fern´andez, J. Chromatogr. A 1008 (2003) 115. [16] Z. Sosa Ferrera, C. Padr´on Sanz, C. Mahugo Santana, J.J. Santana Rodr´ıguez, Trends Anal. Chem. 23 (7) (2004) 469. [17] C. Padr´on Sanz, R. Halko, Z. Sosa Ferrera, J.J. Santana Rodr´ıguez, Anal. Chim. Acta 524 (2004) 265. [18] A. Pe˜nalver, E. Pocurull, F. Borrull, R. Marc´e, Trends Anal. Chem. 18 (8) (1999) 557. [19] W. Ho, S. Hsieh, Anal. Chim. Acta 428 (2001) 111. [20] J. Aulakh, A. Malik, Crit. Rev. Anal. Chem. 35 (2005) 71. [21] AOAC Official Method 994.16, AOAC Official Methods of Analysis, Fertilizers, p. 40, 2000 (chapter 2). [22] W. Sauerlandt, H. Berwecke, Z. Pflanz. D¨ung. Bodenk. 56 (1952) 204. [23] C. Padr´on Sanz, R. Halko, Z. Sosa Ferrera, J.J. Santana Rodr´ıguez, J. Chromatogr. A 1078 (2005) 13.

57

[24] J. Beltr´an, F. L´opez, F. Hern´andez, J. Chromatogr. A 885 (2000) 389. [25] N. Tomiyama, H. Tsuji, M. Watanabe, M. Takeda, T. Harada, H. Kobayashi, J. Chromatogr. B 748 (2000) 361. [26] S. Mill´an, M. Sampedro, N. Unceta, M. Goicolea, E. Rodr´ıguez, R. Barrio, J. Chromatogr. A 995 (2003) 135. [27] V. Pino, J.H. Ayala, A.M. Afonso, V. Gonz´alez, J. Chromatogr. A 869 (2000) 515. [28] V. Pino, J.H. Ayala, A.M. Afonso, V. Gonz´alez, Int. J. Environ. Anal. Chem. 81 (2001) 281. [29] L. Krutz, S. Senseman, A. Sciumbato, J. Chromatogr. A 999 (2003) 103. [30] F. Hern´andez, J. Beltr´an, F. L´opez, J. Gaspar, Anal. Chem. 72 (2000) 2313. [31] C.L. Arthur, L.M. Killiam, K.D. Buchholz, J. Pawliszyn, J. Anal. Chem. 64 (1992) 1960. [32] S. Lindsay, High Performance Liquid Chromatography, Wiley, New York, NY, 1992, p. 71. [33] D. Kantachote, R. Naidu, I. Singleton, N. McClure, B. Harch, Appl. Soil Ecol. 16 (2001) 85.