Journal of Chromatography A, 957 (2002) 201209 www.elsevier.

com / locate / chroma

Determination of pesticide residues in coconut water by liquid liquid extraction and gas chromatography with electron-capture plus thermionic specic detection and solid-phase extraction and highperformance liquid chromatography with ultraviolet detection q
a a a a b a, N.M. Brito , S. Navickiene , L. Polese , E.F.G. Jardim , R.B. Abakerli , M.L. Ribeiro *
a

, Instituto de Qumica , Universidade Estadual Paulista-Unesp, C.P. 355, 14800 -900 Araraquara, Departamento de Qumica Organica SP, Brazil b , SP, Brazil EMBRAPA-Meio Ambiente, Rodovia SP 340, Km 127,5, C.P. 69, 13820 -000 Jaguariuna Received 8 January 2002; received in revised form 5 March 2002; accepted 5 March 2002

Abstract Two simple methods were developed to determine 11 pesticides in coconut water, a natural isotonic drink rich in salts, sugars and vitamins consumed by the people and athletes. The rst procedure involves solid-phase extraction using Sep-Pak Vac C 18 disposable cartridges with methanol for elution. Isocratic analysis was carried out by means of high-performance liquid chromatography with ultraviolet detection at 254 nm to analyse captan, chlorothalonil, carbendazim, lufenuron and diafenthiuron. The other procedure is based on liquidliquid extraction with hexanedichloromethane (1:1, v / v), followed by gas chromatographic analysis with efuent splitting to electron-capture detection for determination of endosulfan, captan, tetradifon and trichlorfon and thermionic specic detection for determination of malathion, parathion-methyl and monocrotophos. The methods were validated with fortied samples at different concentration levels (0.0112.0 mg / kg). Average recoveries ranged from 75 to 104% with relative standard deviations between 1.4 and 11.5%. Each recovery analysis was repeated at least ve times. Limits of detection ranged from 0.002 to 2.0 mg / kg. The analytical procedures were applied to 15 samples and no detectable amounts of the pesticides were found in any samples under the conditions described. 2002 Elsevier Science B.V. All rights reserved. Keywords: Beverages; Food analysis; Coconut water; Pesticides; Organochlorine compounds; Organophosphorus compounds; Benzimidazole; Benzophenylurea; Thiourea compounds

1. Introduction
q

This study was presented at the I Simposio Nacional sobre a Agua de Coco, Aracaju-SE, 2526 October 2001, and at the Anual da Sociedade Brasileira de Qumica, 248 Reuniao Poc os de Caldas-MG, 2831 May 2001. * Corresponding author. E-mail address: mlucia@iq.unesp.br (M.L. Ribeiro).

The expansion of the Brazilian coconut culture is demonstrated by the production increase in the Sao Paulo State, since it is not a traditional producer. Nowadays it reaches 10 million fruit per annum [1]. The northeast of Brazil is responsible for 82.9% of Brazilian coconut production [2]. Besides the com-

0021-9673 / 02 / $ see front matter 2002 Elsevier Science B.V. All rights reserved. PII: S0021-9673( 02 )00351-5

202

N.M. Brito et al. / J. Chromatogr. A 957 (2002) 201209

mercialization of the by-products of the coconut such as coconut oil and coconut milk largely employed in food and cosmetic industries, there has been an increased consumption of coconut water from young or green coconut as a refreshing drink [3]. It is estimated that the Brazilian consumption of coconut water will increase from 1.25% (125?10 6 l) to 4 or 5% in comparison with the consumption of soft drinks (10 billion liters per year) [4,5]. The coconut water is commercialized in natura or after industrialization (packed in TetraPak or plastic bottles). However, the production of coconut is seriously affected by the occurrence of pests and diseases, which can be controlled by pesticides to minimize the economical waste [6]. About 50% of the Brazilian coconut production is reduced by foliage diseases [2]. Studies of the chemical composition [3] and pesticide residues of young coconuts for food use are limited. The literature does not describe chromatographic methods for the determination of pesticide residues in coconut water. Rao and Murthy [7] estimated the residue accumulation in water and copra of coconut by colorimetry and also worked out the safe period of harvest of coconut after application of monocrotophos. Considering the studies shown above and the Brazilian legislation, which did not establish the maximum residue limits (MRL) in this matrix, except for trichlorfon [8], the purpose of this work was to develop methods for the determination of residues from different classes of pesticides, Fig. 1. The methodologies are based on solid-phase extraction (SPE) and high-performance liquid chromatography (HPLC) with UV detection to determine carbendazim, captan, chlorothalonil, diafenthiuron and lufenuron, and on liquidliquid extraction by gas chromatography (GC) with electron-capture detection (ECD) for tetradifon, trichlorfon, captan, endosulfan I, II and sulfate and thermionic specic detection (TSD) for malathion, parathion-methyl and monocrotophos in coconut water.

dichloromethane (Mallinckrodt Baker, Paris, KY, USA) were nanograde. HPLC-grade water was obtained by ltering deionized water through a 0.45 mm lter with a Waters Millipore (Milford, MA, USA) system. Methanol and water were degassed using a Branson 5200 (Branson Ultrasonic, Danbury, CT, USA) ultrasonic bath. Reference standards of all pesticides were purchased from Dr. Ehrenstorfer (Augsburg, Germany). All standards were at least 98% pure. Three stock solutions containing different pesticides were prepared by diluting 1.0 mg of the standards in 10.0 ml of solvent to obtain a concentration of 100 mg / ml: carbendazim, captan, chlorothalonil, lufenuron and diafenthiuron were prepared in methanol; parathionmethyl, malathion, trichlorfon and captan were prepared in ethyl acetate; monocrotophos, endosulfan I, II, sulfate and tetradifon were prepared in isooctane. The working standard solutions were prepared by diluting the stock solutions as required. Sodium chloride and anhydrous sodium sulfate were analytical-grade reagents (Mallinckrodt Baker). Sep-Pak Vac C 18 cartridges were purchased from Waters (Milford, MA, USA). The cartridges were of 3 ml capacity packed with 500 mg of solid phase. The vacuum manifold from Supelco (Bellefonte, PA, USA) was used to perform the SPE.

2.2. Apparatus 2.2.1. GC TSD A Varian 3300 gas chromatograph (Varian, Sunnyvale, CA, USA) equipped with thermionic specic detector and a splitsplitless injector, and connected to a Varian 4290 reporting integrator, which was used to determine parathion-methyl, malathion and monocrotophos. The capillary column was a DB-5 fusedsilica column (30 m30.25 mm I.D., 0.25 mm; J&W Scientic, Folsom, CA, USA). The injector and detector were operated at 240 and 300 8C, respectively. Sample (1 ml) was injected in the splitless mode (40 s), and the oven temperature was programmed as follows: 160 8C for 1 min, increasing to 240 8C (10 8C / min) to 260 8C (5 8C / min) and holding for 10 min. Nitrogen was the carrier and makeup gas at 30 ml / min. Hydrogen and air were used as detector gases at 4 and 175 ml / min, respectively.

2. Experimental

2.1. Chemicals and materials

Methanol, acetonitrile, ethyl acetate, hexane and

N.M. Brito et al. / J. Chromatogr. A 957 (2002) 201209

203

Fig. 1. Molecular structures of the studied pesticides.

2.2.2. GC ECD A Varian 3300 gas chromatograph equipped with an electron-capture detector, a on-column injector, and a connecting Varian 4290 reporting integrator was used to determine endosulfan I, II and sulfate, tetradifon, trichlorfon and captan. The injection

volume for all the samples was 1 ml. The megabore column was a ZB-1701 fused-silica column (30 m3 0.53 mm I.D., 1.25 mm; Zebron-Phenomenex, Torrance, CA, USA). The injector and detector were operated at 240 and 300 8C, respectively. The oven temperature was programmed as follows: 140 8C for

204

N.M. Brito et al. / J. Chromatogr. A 957 (2002) 201209

1 min, increasing to 240 8C (1 8C / min) to 265 8C (10 8C / min) and holding for 10 min. Nitrogen was the carrier and makeup gas at 30 ml / min.

2.2.3. HPLC UV A Waters high-performance liquid chromatograph (Waters, Milford, MA, USA) equipped with two solvent delivery pumps (Model 501), an injector (Model U6K), a UVVis absorbance detector (Model 486) and a reporting integrator (Model 746) was used for the determination of carbendazim, captan, chlorothalonil, lufenuron and diafenthiuron. A stainless steel analytical column LiChrospher 100 RP18 (12534.6 mm I.D., 5 mm; Merck) connected to a 100 RP18 guard column (2034.6 mm I.D., 5 mm; Merck) was used. The mobile phase consisted of methanolwater (70:30, v / v) at 0.7 ml / min. The detection was carried out at 254 nm. 2.3. Sample preparation and fortication
Coconut water samples (plastic bottles, TetraPak and in natura coconut fruit) were obtained from Paulo State. Araraquara and Campinas cities, Sao Before the solid-phase extraction, the samples were ltered in Whatman 40 and 0.45 mm lters. Fortied samples were prepared by adding standard solutions (0.51.0 ml) to 10.0 ml of coconut water followed by homogenization.

2.4.2. Procedure for benzimidazole, benzoylphenylurea and thiourea pesticides A Sep-Pak Vac C 18 cartridge was preconditioned with 1 ml of methanol followed by 2.5 ml of water. It was placed on top of a vacuum block. An analytical aliquot of 10.0 ml coconut water was transferred to the cartridge at a ow-rate of ca. 0.5 ml / min and allowed the solid phase to dry for 15 min. The pesticides were eluted with 4.0 ml of methanol. A 20 ml aliquot of eluate was analysed by HPLCUV.

3. Results and discussion

3.1. General comments

The use of GC and HPLC with conventional detection methods makes possible the direct, reliable, efcient and economical determination of 11 pesticides in coconut water samples listed in Tables 1 and 2. The pesticides studied include compounds of the organochlorine, organophosphorus, benzimidazole, benzoylphenylurea and thiourea classes, and they are used frequently in coconut culture. Few methods have been reported for determining lufenuron residues in different matrices like fruit, vegetables, blood and groundwater. These studies have used liquid chromatography with mass spectrometric, uorescence or diode-array detection, and the procedures are mainly based on solvent partitioning, supercritical uid extraction and solid-phase extraction [916]. Up to now, there is no published methodology for determination of diafenthiuron. The efciency of the proposed methods was evaluated by means of recovery analyses with coconut water samples fortied at different levels. Since the Brazilian legislation did not establish the MRLs for coconut, except for the trichlorfon insecticide (0.5 mg / kg), the fortication levels were selected according to the MRLs established for other fruit [8].

2.4. Methods 2.4.1. Procedure for organochlorine and organophosphorus pesticides An analytical aliquot of 10.0 ml coconut water sample was placed in a glass-stoppered ask. A 2-g amount of sodium chloride and 10.0 ml of hexane dichloromethane (1:1, v / v) were added and the mixture was extracted for 15 min in a mechanical shaker. The residue was reextracted with another 10.0 ml of the solvent mixture and the organic phase was dried with sodium sulfate. The extract was concentrated using a rotary evaporator (32 8C). The residue was dissolved in ethyl acetate and an aliquot (1 ml) was analysed by GCECD and GCTSD.

3.2. Procedure for organochlorine and organophosphorus pesticides

A small scale liquidliquid extraction method was developed to determine seven pesticides in coconut

N.M. Brito et al. / J. Chromatogr. A 957 (2002) 201209 Table 1 Recovery of pesticide from fortied coconut water employing GCECD and GCTSD Pesticide Endosulfan I Detection method ECD Fortication level (mg / kg) 0.01 0.05 0.1 0.01 0.05 0.1 0.1 0.5 1.0 0.5 1.0 2.0 0.5 1.0 2.0 0.5 1.0 2.0 Range of recoveries (%) 8599 7189 80101 91102 7693 86102 82104 89106 95106 85103 102107 90101 8199 79106 91105 88108 96109 88108

205

Average* recovery, % (RSD, %) 95 (4.8) 81 (7.3) 91 (7.3) 99 (4.0) 85 (7.0) 94 (6.2) 91 (8.7) 99 (4.4) 104 (1.8) 93 (9.5) 104 (1.8) 99 (2.3) 90 (9.0) 95 (11.5) 97 (6.3) 97 (8.4) 103 (5.7) 99 (1.4)

Endosulfan II

ECD

Captan

ECD

Malathion

TSD

Monocrotophos

TSD

Parathion-methyl

TSD

Endosulfan sulfate

TSD 0.01 0.05 0.1 83107 7492 84102 95 (8.2) 83 (8.3) 94 (5.9)

Tetradifon

ECD 0.5 1.0 7286 8088 90105 7189 8192 93100 80 (7.7) 84 (4.0) 97 80 85 98 (3.3) (9.6) (5.4) (2.8)

Trichlorfon

ECD

2.0 0.5 1.0 2.0

* n 558 analyses.

water by gas chromatography with electron-capture and thermionic detection. Preliminary analyses were made using hexane for organochlorine pesticide extraction. Recovery values ranged from 78 to 97% for endosulfan I, II and sulfate. In order to develop a simple multiresidue method to determine organochlorine and organophosphorus compounds in coconut water, the extraction was performed using hexanedichloromethane (1:1, v / v). Average recoveries ranged from 80 to 104% for all pesticides.

The detection limits ranged from 0.002 to 0.3 mg / kg and the quantication limits ranged from 0.01 to 0.5 mg / kg [17]. Gas chromatograms concerning a coconut water sample and a standard solution are shown in Figs. 24. The unfortied coconut water sample used in recovery analyses was free of pesticide residues and interfering compounds. The total running time of GCECD analysis was 18 min and that of GCTSD analysis was 12 min.

206

N.M. Brito et al. / J. Chromatogr. A 957 (2002) 201209

Table 2 Recovery of pesticides from fortied coconut water employing HPLCUV Pesticide Carbendazim Fortication level (mg / kg) 0.5 1.0 6.0 12.0 1.0 2.0 0.5 1.0 0.7 1.3 Range of recoveries (%) 7083 84103 7595 7085 85102 8591 7083 84103 70105 82102 Average* recovery, % (RSD, %) 75 (7.3) 93 (8.3) 83 (7.9) 82 (7.9) 96 (6.9) 88 (2.7) 75 (7.3) 93 (8.3) 82 (11.0) 93 (10.0)

Captan

Chlorothalonil

Lufenuron

Diafenthiuron * n 55 analyses.

Fig. 3. GCTSD chromatogram of control coconut water sample. Conditions: DB-5 fused-silica capillary column (30 m30.25 mm I.D., 0.25 mm); injector (240 8C); detector (300 8C); injected volume (1 ml, splitless mode, 40 s); oven temperature: 160 8C for 1 min, increasing to 240 8C (10 8C / min) to 260 8C (5 8C / min) and holding for 10 min; nitrogen (carrier and makeup gas) at 30 ml / min, hydrogen and air (detector gases) at 4 and 175 ml / min, respectively.

Fig. 2. GCECD chromatogram of standard solution: trichlorfon, 2.5 ng / ml (1); captan, 50 pg / ml (2); endosulfan I, 50 pg / ml (3); endosulfan II, 50 pg / ml (4); endosulfan sulfate, 50 pg / ml (5) and tetradifon, 0.1 ng / ml (6). Conditions: megabore column ZB-1701 fused-silica column (30 m30.53 mm I.D., 1.25 mm); injector (240 8C), detector (300 8C); oven temperature: 140 8C for 1 min, increasing to 240 8C (10 8C / min) to 265 8C (10 8C / min) and holding for 10 min. Nitrogen was the carrier and makeup gas at 30 ml / min; injected volume: 1 ml.

The quantication was performed by comparing sample peak areas to those obtained for standard solutions. Electron-capture and thermionic detection provided linear responses for a wide range of amounts injected and good correlation coefcients. Calibration graphs were plotted with ve points. For each calibration point three injections into the GC system were performed. The correlation coefcients were in all cases higher than 0.99 indicating good performance of the chromatographic methods, Table 3. Captan is preferentially determined by GCECD

N.M. Brito et al. / J. Chromatogr. A 957 (2002) 201209

207

showed deviation from linearity at the lowest con centrations as described by Jimenez et al. [18]. In contrast, the calibration curves generated using pesticides in the matrix were linear for all analytes. For this reason, matrix-matched standards were used for quantication of these compounds in coconut water.

3.3. Procedure for benzimidazole, benzoylphenylurea and thiourea pesticides

A fast and efcient method using SPE and HPLC UV was developed for the simultaneous determination of ve pesticides in coconut water. Solidphase extraction was preferred due to the goal to reduce the use of organic solvents in analytical laboratories. To prevent clogging the Sep-Pak Vac C 18 cartridge the coconut water was ltered to remove suspended matter. Excess water remaining in the cartridge was removed by suction of air for 15 min. Methanol was selected for the experiments because it provided higher or similar recoveries for most of the pesticides in comparison with acetonitrile. Matrix interferences were separated from the pesticides through the optimized isocratic mode with UV detection at 254 nm, as presented in Fig. 5. In order to establish the optimum chromatographic analysis and separation conditions of captan, carbendazim, chlorothalonil, lufenuron and diafenthiuron a UV detector was used. To evaluate the mobile phase different ratios of methanolwater were tested with respect to optimal peak sharpness and separation efciency. The composition of methanolwater (70:30, v / v) at a ow-rate of 0.7 ml / min proved to be the most suitable mobile phase with respect to peak form, time of analysis and chromatographic separation of the ve pesticides investigated. Fig. 6 shows the chromatogram of pesticides extracted from fortied coconut water. The total running time of HPLCUV analysis was 38 min. Sep-Pak Vac C 18 cartridges were satisfactory for the extraction of the pesticides in the coconut water samples. The untreated coconut water samples were fortied at two concentration levels. The accuracy was calculated as a percentage of recovery when using the UV detection ranging from 70 to 105% with relative standard deviation (RSD) values of 6.9 to 11.0%, as can be seen in Table 2. Each recovery analysis was

Fig. 4. GCTSD chromatogram of standard solutions: monocrotophos, 0.5 ng / ml (1); malathion, 0.3 ng / ml (2) and parathionmethyl, 0.3 ng / ml (3). Conditions: DB-5 fused-silica capillary column (30 m30.25 mm I.D., 0.25 mm); injector (240 8C); detector (300 8C); injected volume (1 ml, splitless mode, 40 s); oven temperature: 160 8C for 1 min, increasing to 240 8C (10 8C / min) to 260 8C (5 8C / min) and holding for 10 min; nitrogen (carrier and makeup gas) at 30 ml / min, hydrogen and air (detector gases) at 4 and 175 ml / min, respectively.

due to it higher sensitivity in comparison to that obtained by HPLCUV. Quantitative errors from the matrix effect have been reported in the determination of several classes of pesticides by GC in different matrices [18,19]. In this work, the matrix effect was observed for monocrotophos, malathion and parathion-methyl. So, mean recoveries (n 53) of 43, 51 and 39% were obtained for monocrotophos, malathion and parathion-methyl, respectively, when the quantication was performed with standard solutions prepared in ethyl acetate. Therefore, calibration was performed using pesticide standard solutions prepared in coconut water extract. Further, the calibration graphs generated using pesticide standard solution in solvent

208

N.M. Brito et al. / J. Chromatogr. A 957 (2002) 201209

Table 3 Analytical data: equations of the analytical curves, coefcients of regression, retention time, detection and quantication limits for the pesticides Pesticide Carbendazim Chlorothalonil Captan Lufenuron Diafenthiuron Endosulfan I Endosulfan II Endosulfan sulfate Malathion Parathion-methyl Monocrotophos Tetradifon Trichlorfon
a b

Equation of calibration curve y 514.708x 23938 y 51.479.332x 20.0068 y 52.6000x 112.32 a y 5928.432x 159.283 b y 566 465x 2809 y 514.6085x 22354 y 53.9564x 20.52667 y 52.535x 22.6206 y 53.50699x 24.53749 y 519.43x 20.5292 y 513.26x 20.3586 y 50.00587x 20.1396 y 537.09x 21.7511 y 531 968x 24.9393

r2 0.998 0.994 0.995 0.983 0.998 0.999 0.999 0.997 0.999 0.985 0.994 0.996 0.985 0.991

Retention time (min) 7.6 11.9 10.8 13.2 21.9 37.6 11.7 13.1 14.2 10.6 10.4 9.0 17.1 4.40

Detection limit (mg / kg) 0.2 0.3 0.04 2.0 0.2 0.2 0.002 0.002 0.004 0.3 0.2 0.3 0.3 0.3

Quantication limit (mg / kg) 0.5 1.0 0.1 6.0 0.5 0.7 0.01 0.01 0.01 0.5 0.5 0.5 0.5 0.5

GCECD. HPLCUV.

repeated ve times. The precision and accuracy were considered adequate for the validation of the method. Under the chromatographic conditions described, the calibration graphs (external standard mode) were

constructed by plotting peak areas versus concentrations. Good linearity and a good correlation coefcient were achieved for the ve pesticides. The

Fig. 5. HPLCUV chromatogram of control coconut water sample. Conditions: analytical column LiChrospher 100 RP18 (1253 4.6 mm I.D., 5 mm) connected to a 100 RP18 guard column (2034.6 mm I.D., 5 mm); mobile phase: methanolwater (70:30, v / v), ow-rate: 0.7 ml / min, l5254 nm, injection volume: 20 ml.

Fig. 6. HPLCUV chromatogram of spiked extracts of untreated coconut water: carbendazim, 4.8 ng / ml (1); captan, 0.06 mg / ml (2); chlorothalonil, 0.01 ng / ml (3); lufenuron, 0.1 ng / ml (4) and diafenthiuron, 0.1 ng / ml (5). Conditions: analytical column LiChrospher 100 RP18 (12534.6 mm I.D., 5 mm) connected to a 100 RP18 guard column (2034.6 mm I.D., 5 mm); mobile phase: methanolwater (70:30, v / v), ow-rate: 0.7 ml / min, l5254 nm, injection volume: 20 ml.

N.M. Brito et al. / J. Chromatogr. A 957 (2002) 201209

209

equations of the calibration curves, the coefcients of regressions, detection and quantication limits [17] are presented in Table 3.

References
Paulo, 2000, [1] Suplemento Agrofolha, Folha de S. Paulo, Sao p. F3. [2] E.C. Leal, D.R.N. Warwick, M.L.S. Leal, E.A. Tupinamba, Fitopatologia 33 (1998) 220. [3] U. Santoso, K. Kubo, T. Ota, T. Tadokoro, A. Maekawa, Food Chem. 57 (1996) 299. [4] W.M. Aragao, A Importancia do Coqueiro-Anao, EMBRAPA / Tabuleiros Costeiros, Sergipe, 2000. Tema de Curso e Simposio [5] N. Barroso, Coco sera em Sergipe, EMBRAPA / Tabuleiros Costeiros, Sergipe, 2001. [6] J.M.S. Ferreira, D.R.N. Warwick, L.A. Siqueira (Eds.), A Cultura do Coqueiro no Brasil, EMBRAPA-SPI, 1998, p. 189. [7] B.N. Rao, K.S.R.K. Murthy, Ind. Coconut J. 16 (1985) 7. de Substancias [8] Relac ao para Uso Fitossanitario e Domis Paulo, da Saude, sanitario: Portarias do Ministerio ILSI, Sao 1995. [9] A. Anastassiades, E. Scherbaum, W. Schwack, Dtsch. Lebensm.-Rundsch. 97 (2001) 176. [10] M. Martinez-Galera, T. Lopez-Lopez, M.D. Gil-Garcia, J.L.M. Vidal, P.P. Vazquez, J. Chromatogr. A 918 (2001) 79. [11] A.G. Frenich, M.D.G. Garcia, F.J. Arrebola, J.L.M. Vidal, M.M. Galera, T. Lopez, Chromatographia 52 (2000) 569. [12] N.G. Tsiropoulos, P.G. Aplada-Sarlis, G.E. Miliadis, J. AOAC Int. 82 (1999) 213. [13] M. Hiemstra, A. Tooren, A. de Kok, J. AOAC Int. 82 (1999) 1198. [14] G.E. Miliadis, N.G. Tsiropoulos, P.G. Aplada-Sarlis, J. Chromatogr. A 835 (1999) 113. [15] M. Gamon, A. Saez, R. Pelegri, I. Peris, J.G. de la Cuadra, R. Coscolla, J. AOAC Int. 81 (1998) 1037. [16] J.J. Mackichan, W.F. Hink, J. Liq. Chromatogr. 16 (1993) 2595. [17] H.P. Thier, H. Zeumer (Eds.), Manual of Pesticide Residues Analysis, Deutsche Forschungsgemeinschaft, Pesticide Commission, Vol. 1, Verlag Chemie, Weinheim, New York, 1987, p. 37. a a [18] J.J. Jimenez, J.L. Bernal, M ] .J. Del Nozal, L. Toribio, M ] .T. J. Chromatogr. A 823 (1998) 381. Martn, a a [19] J.L. Bernal, M ] .J. Del Nozal, J.J. Jimenez, J.M ] . Rivera, J. Chromatogr. A 778 (1997) 111.

3.4. Real samples

The proposed methods were applied to the analysis of 15 real coconut water samples (TetraPak, plastic bottles and fresh fruit) collected from six commercial markets located in Araraquara and Cam Paulo State, Southeast, Brazil. None pinas cities, Sao of the target analytes were detected in these samples under the experimental conditions described.

4. Conclusions The methods allow efcient and rapid determination of the selected pesticides in coconut water. Since the coconut water consumption is increasing and literature does not describe chromatographic methods for determination of pesticide residues in this matrix, the proposed procedures could be applied for monitoring programs. Further eld studies are necessary to contribute for the establishment of the maximum residue limits for coconut.

Acknowledgements de AperN.M.B. thanks the CAPES (Coordenac ao feic oamento de Pessoal de Nvel Superior) and S.N. (Proc. No. 98 / 15891-6) thanks the FAPESP (Fun de Amparo a ` Pesquisa do Estado de Sao dac ao Paulo) for the fellowships. The authors thank the EMBRAPA (Empresa Brasileira de Pesquisa Ag ropecuaria)Tabuleiros Costeiros, Aracaju-SE, for the technical support and sample collection.