- Email: [email protected]
Ultrasonic-assisted matrix solid-phase dispersion as an improved methodology for the determination of pesticides in fruits
Journal of Chromatography A, 1212 (2008) 145–149
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Short communication
Ultrasonic-assisted matrix solid-phase dispersion as an improved methodology for the determination of pesticides in fruits J.J. Ramos a , R. Rial-Otero b , L. Ramos a,∗ , J.L. Capelo b,∗∗ a b
Department of Instrumental Analysis and Environmental Chemistry, IQOG (CSIC), Juan de la Cierva 3, E-28006 Madrid, Spain REQUIMTE, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, Quinta da Torre, 2829-516 Monte de Caparica, Portugal
a r t i c l e
i n f o
Article history: Received 18 June 2008 Received in revised form 1 October 2008 Accepted 8 October 2008 Available online 14 October 2008 Keywords: Ultrasound MSPD Pesticides Food analysis
a b s t r a c t An ultrasonic-assisted matrix solid-phase dispersion (UA-MSPD) method has been developed for extracting and cleaning-up 15 organophosphorus pesticides (OPPs) and 9 triazines in fruits. Using reversed phase octasilyl-derivatised silica (C8) as dispersant and ethyl acetate as extraction solvent, two sonication devices, an ultrasonic bath and a sonoreactor, were tested for speeding and increasing the efficiency of the MSPD process. A standard MSPD and a heating-assisted MSPD procedure were also done. Gas chromatography–mass spectrometry was used for analyte determination. 1-min sonication with the sonoreactor at 50% amplitude provided the best results, with reproducibilities below 15% for the pesticides studied. The general low detection limits (1–42 g kg−1 ) ensure proper determination at maximum allowed residue levels set in current legislations, except for dimethoate and disulfuton. The analytical performance of the method was evaluated for apples, pear and apricot, showing little or no matrix effect. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Matrix solid-phase dispersion (MSPD) is a simple and fast sample preparation technique for the treatment of liquid, viscous and (semi-)solid samples [1,2]. Its main advantages over other conventional procedures of sample treatment include short extraction times, use of smaller amounts of sorbents and solvents and reduced sample handling because of the possibility of simultaneously perform the extraction and (preliminary) clean-up of the analytes, while still providing similar or higher extraction efficiencies [2]. MSPD efficiency depends on careful optimisation of the experimental parameters affecting competition among matrix, dispersant sorbent and extraction solvent for both, analytes and potential matrix interferences. Thereby, in the case of very sorptive samples, enhanced solvent extraction techniques, such as Soxhlet or Ultrasonic-Assisted Extraction (UAE), can easily result in higher recoveries [3]. For pesticides, ultrasonic energy has been reported to speed and improved the extraction efficiency during solid–liquid extraction (SPE) [4], also when the cartridge was placed inside an ultrasonic bath [5,6]. This study reports the combined use of MSPD and UAE. The new methodology, so-called ultrasonic-assisted matrix solid-phase
∗ Corresponding author. Tel.: +34915622900; fax: +34915644853. ∗∗ Corresponding author. Tel.: +351212948386; fax: +351212948550. E-mail addresses: [email protected] (L. Ramos), [email protected] (J.L. Capelo). 0021-9673/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2008.10.028
dispersion (UA-MSPD), has been applied to the determination of 15 OPPs and 9 triazines in fruits. The obtained results have been compared with those found when analysing the same samples by conventional MSPD and heating-assisted MSPD. 2. Experimental 2.1. Reagents, materials and apparatus Triazines (ametryne, atraton, atrazine, prometon, prometryne, propazine, simetryn, terbuthylazine and terbutryne) were purchased as pesticide mixture-619 (500 ng L−1 each component) from Chem Service (West Chester, PA, USA). OPPs (bromophosmethyl, bromophos-ethyl, chlorfenvinphos, chlorpyrifos-ethyl, diazinon, dichlorvos, dimethoate, disulfoton, fenthion, ethion, malathion, mevinphos, parathion-methyl and -ethyl, and paraoxon-ethyl) were purchased from Dr. Ehrenstorfer (Augsburg, Germany) as Pesticides-Mix 235 (200 ng L−1 each component). Working standard solutions were prepared from these commercial mixtures in isooctane at concentrations ranging between 100 and 1000 pg L−1 and used for further dilution and spiking of the samples. Lindane (Sigma–Aldrich, Steinheim, Germany) in isooctane was used as syringe standard. All solutions were stored at 4 ◦ C. All solvents were of trace analysis grade and purchased from Merck Co. (Darmstadt, Germany) and Scharlau Chemie (Barcelona, Spain). Milli-Q water (Millipore, Bedford, USA) was used in all
146
J.J. Ramos et al. / J. Chromatogr. A 1212 (2008) 145–149
experiments. Octasilyl-derivatised silica (C8, endcapped, 50 m) was purchased from IST (Mid Glamorgan, UK). Four varieties of apple (i.e., Gala, Granny Smith, Golden delicious and Pink lady), an organic apple (Gala variety), pear and apricot samples were purchased from local markets in Costa de Caparica (Portugal). Representative portions of the selected fruit peel, around 5 g, were homogenised in a Sorvall Omni-mixer (DuPont Instruments, Wilmington, DE, USA) and kept at −20 ◦ C until use. For UA-MSPD, an ultrasonic bath (Transsonic TI-H-5, Elma, Singen, Germany) and a sonoreactor (UTR200, Dr. Hielsher, Teltow, Switzerland) were assayed. 2.2. Sample treatment MSPD extraction procedure. An MSPD procedure previously validated [7] was used as reference method. Briefly, 700 mg of the untreated peel were dispersed on 700 mg of C8 and 200 mg of the dried homogeneous resulting mixture packed on an empty 3 mL SPE glass barrel (J.T. Baker, Deventer, The Netherlands). Pesticides were extracted with 700 L of ethyl acetate. Because of the intended comparison between the several assayed sample treatments, the flow through the MSPD column was stopped for 5 min. Complete solvent elution was ensured by positive pressure at the end of the dropwise elution step. The MSPD extract was concentrated under a gentle nitrogen stream, reconstituted in the syringe standard solution and subjected to gas chromatography–mass spectrometry (GC–MS) analysis. Heated assisted-MSPD (HA-MSPD). Sample preparation was as described for the reference procedure but the matrix-C8 mixture was packed in an empty reversible 2 mL SPE tube (Supelco, Bellefonte, USA). Once the extraction solvent was loaded, both ends were closed with polypropylene caps and the column immersed in a thermosthated bath heated at 40 ◦ C for either 15 or 30 min. The extractant was subsequently eluted and treated as previously described. UA-MSPD with ultrasonic bath. Sample treatment was as described for HA-MSPD but an ultrasonic bath operating at 35 kHz and 100% of amplitude was used instead. Sonication times of 15 and 30 min were tested. Ethyl acetate was then eluted and treated as previously described. UA-MSPD with sonoreactor. Sample handling was similar to that used for the ultrasonic bath experiments, but using a sonoreactor device and sonication times of 1 and 3 min. The extraction efficiency of the several sample treatments assayed was evaluated by analysing a pesticide-free Gala apple sample spiked at the 0.5 g g−1 level, which is close to the maximum residual limits (MRLs) set in the European Union for most of the pesticides included in this study in the test fruits [8]. Peel fruit sub-samples were spiked by adding on top of the sample column the corresponding amount of the pesticide standard mixture containing all test analytes [9] and allowing to stand for 30 min before treatment. In studies concerning the reference procedure, four replicates were done. In all other cases, three separate analyses were done. In all sets of experiments, procedural blanks, i.e. non-spiked pesticidefree samples, were regularly analysed to check for contamination throughout the analytical procedure showing no presence of the analytes of interest. 2.3. GC–MS analysis Pesticide determination was performed on a GC 8000 series gas chromatograph equipped with a mass-selective detector MD 800 (Fisons Instruments, Ipswich, UK). Masslab version 1.3 soft-
ware (ThermoQuest, Manchester, UK) was used for data acquisition. Samples were injected in splitless mode (1 L; 280 ◦ C; splitless time, 1 min) in a BPX5 column (30 m × 0.25 mm i.d., 0.25 m film thickness; SGE, Darmstadt, Germany). The oven temperature was programmed from 60 ◦ C (2 min) to 140 ◦ C (1 min) at 12 ◦ C min−1 , and then at 6 ◦ C min−1 to 280 ◦ C (30 min). The carrier gas was helium with a constant pressure of 100 kPa. The source and transfer line temperatures were 200 ◦ C and 280 ◦ C, respectively. MS detection (electron impact) was performed in full scan (m/z 50–550); the ion energy was 70 eV. Confirmation criteria for pesticide identification and quantification included detection of the target compound peak at the same retention time that the corresponding standard, the mutual agreement between their respective mass spectra, and the maintenance of the two selected ions ratio within ±15% of the theoretical value [7].
3. Results and discussion 3.1. Optimisation of the sample preparation procedure Table 1 summarises relevant analytical data concerning the pesticides investigated with the different sample preparation procedures evaluated. Although satisfactory average recoveries were obtained with the MSPD reference method (mean values of 75% and 78% for OPPs and triazines, respectively), individual values lied in a relatively wide range (53–91% for OPPs and 65–89% for triazines). In addition, despite the still acceptable 18% average RSD obtained for both pesticide families, relatively high RSDs were found for some of the analytes (e.g. dimethoate, chlorphenvinphos or propazine). These results agreed with previously reported when a large number of pesticides with varying chemical structures were involved [10], but also stood for further improvement of the method. Therefore, the influence of the ultrasonic energy in the efficiency of the MSPD process was investigated. An ultrasonic bath was firstly evaluated using sonication times of 15 and 30 min [10,11]. Longer extraction times were not assayed because then one of the main advantages of MSPD, speed, would be lost. UA-MSPD for 15 min yielded results similar to those of the reference method, i.e. mean recoveries of 72% and 78% and RSDs of 20% and 16% for OPPs and triazines, respectively. Increasing the sonication time to 30 min resulted in a general increase of the average recoveries for OPPs to 84% (dichlorvos and mevinphos were not considered in this calculation), with the only exception of disulfoton (11% decrease). A similar trend was observed for triazines (average recovery increase of 8% as compared to the reference method), with atrazine, propazine and terbuthylazine showing the most significant improvements (in the 16–38% range). Lower recoveries were only observed for atraton and prometryne. No improvement was observed regarding repeatability (mean RSDs for OPPs and triazines, 19% and 23%, respectively) as compared to the reference method. Sonication promotes an increase in the bath water temperature (up to 32 ◦ C after 30 min sonication). To evaluate the possible effect of this parameter on the final results and isolate it from that of ultrasound, a new set of experiments consisting on HA-MSPD at 40 ◦ C during 30 min were developed. Heating had a variable but, in general, moderated effect on OPP recoveries. The most significant change was observed for dimethoate, for which the recovery increased from 53% with MSPD at room temperature to 89% at 40 ◦ C. However, other OPPs (diazinon, disulfoton, parathion-methyl and bromophos-methyl) showed significant decreases of up to 17%. Results obtained by HA-MSPD were also rather similar to those found after 30 min UA-MSPD, with the clear exceptions of dichlor-
J.J. Ramos et al. / J. Chromatogr. A 1212 (2008) 145–149
147
Table 1 Retention times and m/z ions selected for quantification (bold font) and confirmation of the OPPs and triazines investigated by GC–MS. Comparison of the recoveries (%) and repeatability (as RSD, n = 3) obtained for the target compounds with the different sample preparation methods assayed in this work (spiking level, 0.5 g g−1 ). Pesticide
tR (min)
Dichlorvos Mevinphos Dimethoate Diazinon Disulfoton Parathion-methyl Paraoxon-ethyl Malathion Fenthion Chlorpyrifos-ethyl Parathion-ethyl Bromophos-methyl Chlorfenvinphos Bromophos-ethyl Ethion
14.02 17.73 24.22 24.70 25.42 27.12 27.44 27.90 28.24 28.54 28.64 29.17 29.69 30.39 32.89
Atraton Prometon Atrazine Propazine Terbuthylazine Simetryn Ametryne Prometryne Terbutryne
24.05 24.17 24.44 24.52 24.92 27.29 27.32 27.34 27.55
Peak no.
1 2 3 4 5 6 7 8 9 10 14 11 12 15 16 13 17 18
m/z
MSPD
HA-MSPD
UA-MSPD (ultrasonic bath)
UA-MSPD (sonoreactor)
% Recov. (RSD)
% Recov. (RSD)
15 min % Recov. (RSD)
30 min % Recov. (RSD)
1 min % Recov. (RSD)
3 min % Recov. (RSD)
109/815 109/127 125/143 152/179 125/153 125/263 109/149 125/173 197/314 125/278 139/291 125/331 267/323 303/331 153/231
74 (17) 79 (18) 53 (26) 86 (18) 91 (17) 80 (16) 83 (14) 79 (8) 59 (17) 81 (21) 82 (18) 75 (15) 64 (25) 74 (16) 68 (19)
77 (27) 83 (25) 89 (20) 69 (11) 76 (7) 67 (21) 86 (20) 82 (22) 72 (23) 74 (24) 74 (24) 61 (16) 74 (16) 64 (22) 72 (12)
70 (30) 53 (24) 102 (23) 71 (15) 75 (13) 70 (27) 70 (11) 75 (9) 75 (22) 69 (21) 69 (18) 67 (22) 71 (24) 69 (25) 72 (20)
143 (77) 124 (17) 102 (10) 90 (18) 80 (31) 81 (20) 81 (15) 86 (14) 83 (20) 84 (18) 85 (15) 79 (23) 84 (21) 81 (21) 78 (23)
82 (16) 71 (6) 90 (12) 121 (4) 124 (2) 101 (6) 92 (23) 89 (13) 100 (14) 92 (13) 100 (17) 85 (14) 94 (9) 92 (14) 93 (19)
50 (86) 91 (4) 71 (10) 90 (17) 70 (10) 62 (17) 96 (12) 101 (14) 124 (6) 101 (14) 90 (9) 87 (23) 73 (12) 80 (17) 87 (10)
196/211 210/225 200/215 214/229 214/229 213/198 212/227 226/241 226/241
83 (13) 72 (19) 89 (14) 69 (24) 83 (13) 65 (22) 79 (20) 88 (21) 77 (13)
218 (63) 190 (47) 76 (10) 84 (8) 84 (5) 135 (52) 201 (58) 151(53) 162(54)
99 (16) 74 (19) 75 (10) 77 (8) 81 (12) 81 (16) 77 (22) 74 (22) 65 (18)
66 (32) 74 (26) 106 (23) 112 (8) 101 (18) 73 (34) 94 (13) 74 (23) 77 (25)
68 (7) 76 (8) 120 (3) 139 (6) 126 (7) 72 (8) 80 (6) 88 (6) 77 (7)
116 (15) 132 (7) 75 (16) 79 (16) 75 (7) 183 (12) 182 (17) 139 (19) 133 (13)
vos and mevinphos, for which the unexpected high recoveries obtained with the latter procedure contrasted with the satisfactory recoveries (77% and 83%, respectively) obtained when the ultrasonic energy was avoided. RSDs systematically increased with the extraction temperature for a significant number of analytes (9 out the 15 test OPPs), a trend also observed with UA-MSPD and that could point to a (partial) degradation of some of the analytes with temperature. The situation became much more dramatic for triazines. Unacceptable recoveries above 135% and RSDs over 47% were found for atraton, prometon, simetryn, ametryne, prometryne and terbutryn with HA-MSPD, a result that, irrespective of its cause (e.g., coextraction of other matrix components [12,13] or degradation [11]), detracted from further investigation. The feasibility of UA-MSPD with a sonoreactor [11] was investigated. Results obtained after 3 min UA-MSPD for OPPs with the sonoreactor follows a similar trend to that observed after 30 min UA-MSPD in an ultrasonic bath (Table 1). The only exceptions were observed for fenthion, which showed much higher recovery in the sonoreactor experiments (124% vs 83% with bath); and for dichlorvos, which shows a low recovery of 50%. This last result was due to the degradation of dichlorvos during the treatment with the sonoreactor (Fig. 1A) and explains its high RSD. A slight (but systematic) decrease was observed in the RSDs calculated (12%) as compared to those of previously discussed treatments. Results obtained for triazines in this part of the study were closer to those found after 30 min MSPD at 40 ◦ C than to those obtained by 30 min UA-MSPD in an ultrasonic bath, with still unacceptably high recoveries for most compounds. Reducing the sonication time to 1 min yielded recoveries in the 82–101% range with a mean RSD of 12% for most of the OPPs investigated (exceptions, mevinphos, diazinon and disulfoton). In addition, the degradation previously observed for dichlorvos was avoided (Fig. 1B; recovery, 82%; RSD, 16%). Despite the still rather low extraction efficiency observed for atraton (66%), results for atrazines were more satisfactory (average recovery, 94%; RSD, 6%) than with previously assayed methods, a fact specially evident for
atrazine, propazine and terbuthylazine (recoveries of 120%, 139% and 126%, respectively). The shorter extraction time involved by this treatment also contributes to avoid the (apparent) degradation observed at higher temperatures (i.e., MSPD at 40 ◦ C) or longer sonication times (i.e., 3 min sonoreactor UA-MSPD). 3.2. Analytical performance The developed 1 min sonoreactor UA-MSPD method was finally selected and evaluated in terms of intermediate precision (between-day RSD, Table 2). Mean recovery for OPPs from the repeatability assays was 95% (RSD, 12%; n = 3), while the value for the intermediate precision was 96% (RSD, 10%; n = 7). These results were similar to those reported in the literature for OPP determi-
Fig. 1. Fragmentograms of dichlorvos after (A) 3 min and (B) 1 min UA-MSPD with sonoreactor. Spiking level, 0.5 g g−1 .
148
J.J. Ramos et al. / J. Chromatogr. A 1212 (2008) 145–149
Table 2 Inter-day recoveries and reproducibility (as RSD; n = 7, within a week) of the UAMSPD with sonoreactor method developed for the OPPs and triazines investigated (spiking level, 0.5 g g−1 ). Limits of detection (LODs) as calculated for real-life apples (g kg−1 ). UA-MSPD-sonoreactor (1 min) Compound
% Recovery
RSD
LODs (g kg−1 )
Dichlorvos Mevinphos Dimethoate Diazinon Disulfoton Parathion-methyl Paraoxon-ethyl Malathion Fenthion Chlorpyrifos-ethyl Parathion-ethyl Bromophos-methyl Chlorfenvinphos Bromophos-ethyl Ethion
81 73 106 108 115 97 98 102 99 93 99 88 94 92 96
11 9 12 11 9 8 12 15 11 10 10 10 10 10 9
13 72 205 12 61 1.4 42 6.6 2.8 1.7 2.1 1.0 4.1 4.9 4.2
Atraton Prometon Atrazine Propazine Terbuthylazine Simetryn Ametryne Prometryne Terbutryne
83 83 106 118 108 89 89 91 87
22 15 11 14 12 24 14 8 14
36 21 14 9.5 3.8 27 11 8.5 18
nation in fruits [12] by MSPD, and in fruit juice by SPE assisted by ultrasound [5], and better than those obtained with other miniaturised techniques [13,14]. Triazines showed a more variable, although still satisfactory, behaviour regarding the between-day precision, with recoveries in the 83–118% range and RSDs varying from 8% to 24%. Although examples publishing more satisfactory reproducibility can be found in the literature [15], these studies typically involved
Fig. 3. Fragmentograms obtained for bromophos-ethyl in a Golden delicious apple (A) without and (B) with fortification at the 0.5 g g−1 level.
larger amounts of sample and solvent and more laborious and time consuming treatment protocols. Detection limits (LODs), determined as the concentration yielding a signal-to-noise ratio of 3 for the extracted quantification ion (Table 2), were in the 1–42 g kg−1 range. Higher LODs were obtained for mevinphos (72 g kg−1 ), but only those of disulfoton and dimethoate (61 and 205 g kg−1 ) were above the EU MRLs set for the investigated fruits. These low LODs proved the general feasibility of the proposed analytical procedure for the accurate and sensible determination of the OPPs and triazines studied even if only 100 mg of sample was used (Fig. 2). Finally, the applicability of the proposed method to the determination of the test pesticides in commercial fruits, i.e. four apple varieties, pear and apricot, was evaluated. None of the target compounds was found at detectable levels in pear, apricot and in the Pink lady and Granny Smith apples. Bromophos-ethyl was detected
Fig. 2. Typical reconstructed fragmentograms obtained for an organic Gala apple sample fortified at the 0.5 g g−1 level after 1 min UA-MSPD.
J.J. Ramos et al. / J. Chromatogr. A 1212 (2008) 145–149
at levels below the quantification limit (LOQ; S/N ratio = 10) in the Golden delicious and Gala apples (i.e., concentration < 15 g kg−1 ). Chlorfenvinphos was also found at a level below LOQ in the Gala apple (Fig. 3). 4. Conclusions A method based on UA-MSPD with a sonoreactor has been developed and its suitability for the accurate determination of OPP and triazine residues in a variety of fruits without any additional clean-up step before GC–MS analysis demonstrated. Compared with classic MSPD, the proposed method improves the general extraction efficiency, decreases the RSDs and allows complete sample treatment within a few minutes. Recoveries above 81% were obtained for all investigated pesticides. RSDs lower than 12%, were obtained for OPPs. More variable but still acceptable values (8–24%) were found for triazines. These results, combined with the low LODs, proved that the proposed procedure is suitable for the accurate determination of the target analytes at levels set in current legislations although as a small amount of sample as 100 mg is used.
149
Acknowledgments R.R.-O. acknowledges to FCT for grant SFRH/BPD/23072/2005. J.J.R. thanks MEC for FPI grant and aid EST20070624838. J.J.R. and L.R. thank MEC (CTQ-2006-14993/BQU). References [1] [2] [3] [4] [5] [6] [7] [8]
[9] [10] [11] [12] [13] [14] [15]
S.A. Barker, J. Biochem. Biophys. Methods 70 (2007) 151. E.M. Kristenson, L. Ramos, U.A.Th. Brinkman, Trends Anal. Chem. 25 (2006) 96. H.B. Xiao, M. Krucker, K. Albert, X.M. Liang, J. Chromatogr. A 1032 (2004) 117. S. Navarro, J. Oliva, A. Barba, C. García, J. AOAC Int. 83 (2000) 1239. B. Albero, C. Sanchez-Brunete, J.L. Tadeo, J. Agric. Food. Chem. 51 (2003) 6915. F. Priego-Capote, M.D. Luque de Castro, Trends Anal. Chem. 23 (2004) 370. J.J. Ramos, M.J. González, L. Ramos (in preparation). Regulation (EC) No. 396/2005 on maximum residue levels (MRLs) of pesticides found in or on food and feed of plant and animal origin, OJ L 70, 16.03.2005, p. 1. S. Lundstedt, B. van Bavel, P. Haglund, M. Tysklind, L. Oberg, J. Chromatogr. A 883 (2000) 151. X.C. Chu, Z.X. Hu, H.Y. Yao, J. Chromatogr. A 1063 (2005) 201. H.M. Santos, J.L. Capelo, Talanta 73 (2007) 589. H.S. Dórea, F.M. Lanc¸as, J. Microcol. Sep. 11 (1999) 367. M. Chai, G. Tan, A. Lal, Anal. Sci. 24 (2008) 273. ˜ A. Juan-García, J. Manes, G. Font, Y. Picó, J. Chromatogr. A 1050 (2004) 119. E.J. Avramides, S. Gkatsos, J. Agric. Food Chem. 55 (2007) 561.
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Short communication
Ultrasonic-assisted matrix solid-phase dispersion as an improved methodology for the determination of pesticides in fruits J.J. Ramos a , R. Rial-Otero b , L. Ramos a,∗ , J.L. Capelo b,∗∗ a b
Department of Instrumental Analysis and Environmental Chemistry, IQOG (CSIC), Juan de la Cierva 3, E-28006 Madrid, Spain REQUIMTE, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, Quinta da Torre, 2829-516 Monte de Caparica, Portugal
a r t i c l e
i n f o
Article history: Received 18 June 2008 Received in revised form 1 October 2008 Accepted 8 October 2008 Available online 14 October 2008 Keywords: Ultrasound MSPD Pesticides Food analysis
a b s t r a c t An ultrasonic-assisted matrix solid-phase dispersion (UA-MSPD) method has been developed for extracting and cleaning-up 15 organophosphorus pesticides (OPPs) and 9 triazines in fruits. Using reversed phase octasilyl-derivatised silica (C8) as dispersant and ethyl acetate as extraction solvent, two sonication devices, an ultrasonic bath and a sonoreactor, were tested for speeding and increasing the efficiency of the MSPD process. A standard MSPD and a heating-assisted MSPD procedure were also done. Gas chromatography–mass spectrometry was used for analyte determination. 1-min sonication with the sonoreactor at 50% amplitude provided the best results, with reproducibilities below 15% for the pesticides studied. The general low detection limits (1–42 g kg−1 ) ensure proper determination at maximum allowed residue levels set in current legislations, except for dimethoate and disulfuton. The analytical performance of the method was evaluated for apples, pear and apricot, showing little or no matrix effect. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Matrix solid-phase dispersion (MSPD) is a simple and fast sample preparation technique for the treatment of liquid, viscous and (semi-)solid samples [1,2]. Its main advantages over other conventional procedures of sample treatment include short extraction times, use of smaller amounts of sorbents and solvents and reduced sample handling because of the possibility of simultaneously perform the extraction and (preliminary) clean-up of the analytes, while still providing similar or higher extraction efficiencies [2]. MSPD efficiency depends on careful optimisation of the experimental parameters affecting competition among matrix, dispersant sorbent and extraction solvent for both, analytes and potential matrix interferences. Thereby, in the case of very sorptive samples, enhanced solvent extraction techniques, such as Soxhlet or Ultrasonic-Assisted Extraction (UAE), can easily result in higher recoveries [3]. For pesticides, ultrasonic energy has been reported to speed and improved the extraction efficiency during solid–liquid extraction (SPE) [4], also when the cartridge was placed inside an ultrasonic bath [5,6]. This study reports the combined use of MSPD and UAE. The new methodology, so-called ultrasonic-assisted matrix solid-phase
∗ Corresponding author. Tel.: +34915622900; fax: +34915644853. ∗∗ Corresponding author. Tel.: +351212948386; fax: +351212948550. E-mail addresses: [email protected] (L. Ramos), [email protected] (J.L. Capelo). 0021-9673/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2008.10.028
dispersion (UA-MSPD), has been applied to the determination of 15 OPPs and 9 triazines in fruits. The obtained results have been compared with those found when analysing the same samples by conventional MSPD and heating-assisted MSPD. 2. Experimental 2.1. Reagents, materials and apparatus Triazines (ametryne, atraton, atrazine, prometon, prometryne, propazine, simetryn, terbuthylazine and terbutryne) were purchased as pesticide mixture-619 (500 ng L−1 each component) from Chem Service (West Chester, PA, USA). OPPs (bromophosmethyl, bromophos-ethyl, chlorfenvinphos, chlorpyrifos-ethyl, diazinon, dichlorvos, dimethoate, disulfoton, fenthion, ethion, malathion, mevinphos, parathion-methyl and -ethyl, and paraoxon-ethyl) were purchased from Dr. Ehrenstorfer (Augsburg, Germany) as Pesticides-Mix 235 (200 ng L−1 each component). Working standard solutions were prepared from these commercial mixtures in isooctane at concentrations ranging between 100 and 1000 pg L−1 and used for further dilution and spiking of the samples. Lindane (Sigma–Aldrich, Steinheim, Germany) in isooctane was used as syringe standard. All solutions were stored at 4 ◦ C. All solvents were of trace analysis grade and purchased from Merck Co. (Darmstadt, Germany) and Scharlau Chemie (Barcelona, Spain). Milli-Q water (Millipore, Bedford, USA) was used in all
146
J.J. Ramos et al. / J. Chromatogr. A 1212 (2008) 145–149
experiments. Octasilyl-derivatised silica (C8, endcapped, 50 m) was purchased from IST (Mid Glamorgan, UK). Four varieties of apple (i.e., Gala, Granny Smith, Golden delicious and Pink lady), an organic apple (Gala variety), pear and apricot samples were purchased from local markets in Costa de Caparica (Portugal). Representative portions of the selected fruit peel, around 5 g, were homogenised in a Sorvall Omni-mixer (DuPont Instruments, Wilmington, DE, USA) and kept at −20 ◦ C until use. For UA-MSPD, an ultrasonic bath (Transsonic TI-H-5, Elma, Singen, Germany) and a sonoreactor (UTR200, Dr. Hielsher, Teltow, Switzerland) were assayed. 2.2. Sample treatment MSPD extraction procedure. An MSPD procedure previously validated [7] was used as reference method. Briefly, 700 mg of the untreated peel were dispersed on 700 mg of C8 and 200 mg of the dried homogeneous resulting mixture packed on an empty 3 mL SPE glass barrel (J.T. Baker, Deventer, The Netherlands). Pesticides were extracted with 700 L of ethyl acetate. Because of the intended comparison between the several assayed sample treatments, the flow through the MSPD column was stopped for 5 min. Complete solvent elution was ensured by positive pressure at the end of the dropwise elution step. The MSPD extract was concentrated under a gentle nitrogen stream, reconstituted in the syringe standard solution and subjected to gas chromatography–mass spectrometry (GC–MS) analysis. Heated assisted-MSPD (HA-MSPD). Sample preparation was as described for the reference procedure but the matrix-C8 mixture was packed in an empty reversible 2 mL SPE tube (Supelco, Bellefonte, USA). Once the extraction solvent was loaded, both ends were closed with polypropylene caps and the column immersed in a thermosthated bath heated at 40 ◦ C for either 15 or 30 min. The extractant was subsequently eluted and treated as previously described. UA-MSPD with ultrasonic bath. Sample treatment was as described for HA-MSPD but an ultrasonic bath operating at 35 kHz and 100% of amplitude was used instead. Sonication times of 15 and 30 min were tested. Ethyl acetate was then eluted and treated as previously described. UA-MSPD with sonoreactor. Sample handling was similar to that used for the ultrasonic bath experiments, but using a sonoreactor device and sonication times of 1 and 3 min. The extraction efficiency of the several sample treatments assayed was evaluated by analysing a pesticide-free Gala apple sample spiked at the 0.5 g g−1 level, which is close to the maximum residual limits (MRLs) set in the European Union for most of the pesticides included in this study in the test fruits [8]. Peel fruit sub-samples were spiked by adding on top of the sample column the corresponding amount of the pesticide standard mixture containing all test analytes [9] and allowing to stand for 30 min before treatment. In studies concerning the reference procedure, four replicates were done. In all other cases, three separate analyses were done. In all sets of experiments, procedural blanks, i.e. non-spiked pesticidefree samples, were regularly analysed to check for contamination throughout the analytical procedure showing no presence of the analytes of interest. 2.3. GC–MS analysis Pesticide determination was performed on a GC 8000 series gas chromatograph equipped with a mass-selective detector MD 800 (Fisons Instruments, Ipswich, UK). Masslab version 1.3 soft-
ware (ThermoQuest, Manchester, UK) was used for data acquisition. Samples were injected in splitless mode (1 L; 280 ◦ C; splitless time, 1 min) in a BPX5 column (30 m × 0.25 mm i.d., 0.25 m film thickness; SGE, Darmstadt, Germany). The oven temperature was programmed from 60 ◦ C (2 min) to 140 ◦ C (1 min) at 12 ◦ C min−1 , and then at 6 ◦ C min−1 to 280 ◦ C (30 min). The carrier gas was helium with a constant pressure of 100 kPa. The source and transfer line temperatures were 200 ◦ C and 280 ◦ C, respectively. MS detection (electron impact) was performed in full scan (m/z 50–550); the ion energy was 70 eV. Confirmation criteria for pesticide identification and quantification included detection of the target compound peak at the same retention time that the corresponding standard, the mutual agreement between their respective mass spectra, and the maintenance of the two selected ions ratio within ±15% of the theoretical value [7].
3. Results and discussion 3.1. Optimisation of the sample preparation procedure Table 1 summarises relevant analytical data concerning the pesticides investigated with the different sample preparation procedures evaluated. Although satisfactory average recoveries were obtained with the MSPD reference method (mean values of 75% and 78% for OPPs and triazines, respectively), individual values lied in a relatively wide range (53–91% for OPPs and 65–89% for triazines). In addition, despite the still acceptable 18% average RSD obtained for both pesticide families, relatively high RSDs were found for some of the analytes (e.g. dimethoate, chlorphenvinphos or propazine). These results agreed with previously reported when a large number of pesticides with varying chemical structures were involved [10], but also stood for further improvement of the method. Therefore, the influence of the ultrasonic energy in the efficiency of the MSPD process was investigated. An ultrasonic bath was firstly evaluated using sonication times of 15 and 30 min [10,11]. Longer extraction times were not assayed because then one of the main advantages of MSPD, speed, would be lost. UA-MSPD for 15 min yielded results similar to those of the reference method, i.e. mean recoveries of 72% and 78% and RSDs of 20% and 16% for OPPs and triazines, respectively. Increasing the sonication time to 30 min resulted in a general increase of the average recoveries for OPPs to 84% (dichlorvos and mevinphos were not considered in this calculation), with the only exception of disulfoton (11% decrease). A similar trend was observed for triazines (average recovery increase of 8% as compared to the reference method), with atrazine, propazine and terbuthylazine showing the most significant improvements (in the 16–38% range). Lower recoveries were only observed for atraton and prometryne. No improvement was observed regarding repeatability (mean RSDs for OPPs and triazines, 19% and 23%, respectively) as compared to the reference method. Sonication promotes an increase in the bath water temperature (up to 32 ◦ C after 30 min sonication). To evaluate the possible effect of this parameter on the final results and isolate it from that of ultrasound, a new set of experiments consisting on HA-MSPD at 40 ◦ C during 30 min were developed. Heating had a variable but, in general, moderated effect on OPP recoveries. The most significant change was observed for dimethoate, for which the recovery increased from 53% with MSPD at room temperature to 89% at 40 ◦ C. However, other OPPs (diazinon, disulfoton, parathion-methyl and bromophos-methyl) showed significant decreases of up to 17%. Results obtained by HA-MSPD were also rather similar to those found after 30 min UA-MSPD, with the clear exceptions of dichlor-
J.J. Ramos et al. / J. Chromatogr. A 1212 (2008) 145–149
147
Table 1 Retention times and m/z ions selected for quantification (bold font) and confirmation of the OPPs and triazines investigated by GC–MS. Comparison of the recoveries (%) and repeatability (as RSD, n = 3) obtained for the target compounds with the different sample preparation methods assayed in this work (spiking level, 0.5 g g−1 ). Pesticide
tR (min)
Dichlorvos Mevinphos Dimethoate Diazinon Disulfoton Parathion-methyl Paraoxon-ethyl Malathion Fenthion Chlorpyrifos-ethyl Parathion-ethyl Bromophos-methyl Chlorfenvinphos Bromophos-ethyl Ethion
14.02 17.73 24.22 24.70 25.42 27.12 27.44 27.90 28.24 28.54 28.64 29.17 29.69 30.39 32.89
Atraton Prometon Atrazine Propazine Terbuthylazine Simetryn Ametryne Prometryne Terbutryne
24.05 24.17 24.44 24.52 24.92 27.29 27.32 27.34 27.55
Peak no.
1 2 3 4 5 6 7 8 9 10 14 11 12 15 16 13 17 18
m/z
MSPD
HA-MSPD
UA-MSPD (ultrasonic bath)
UA-MSPD (sonoreactor)
% Recov. (RSD)
% Recov. (RSD)
15 min % Recov. (RSD)
30 min % Recov. (RSD)
1 min % Recov. (RSD)
3 min % Recov. (RSD)
109/815 109/127 125/143 152/179 125/153 125/263 109/149 125/173 197/314 125/278 139/291 125/331 267/323 303/331 153/231
74 (17) 79 (18) 53 (26) 86 (18) 91 (17) 80 (16) 83 (14) 79 (8) 59 (17) 81 (21) 82 (18) 75 (15) 64 (25) 74 (16) 68 (19)
77 (27) 83 (25) 89 (20) 69 (11) 76 (7) 67 (21) 86 (20) 82 (22) 72 (23) 74 (24) 74 (24) 61 (16) 74 (16) 64 (22) 72 (12)
70 (30) 53 (24) 102 (23) 71 (15) 75 (13) 70 (27) 70 (11) 75 (9) 75 (22) 69 (21) 69 (18) 67 (22) 71 (24) 69 (25) 72 (20)
143 (77) 124 (17) 102 (10) 90 (18) 80 (31) 81 (20) 81 (15) 86 (14) 83 (20) 84 (18) 85 (15) 79 (23) 84 (21) 81 (21) 78 (23)
82 (16) 71 (6) 90 (12) 121 (4) 124 (2) 101 (6) 92 (23) 89 (13) 100 (14) 92 (13) 100 (17) 85 (14) 94 (9) 92 (14) 93 (19)
50 (86) 91 (4) 71 (10) 90 (17) 70 (10) 62 (17) 96 (12) 101 (14) 124 (6) 101 (14) 90 (9) 87 (23) 73 (12) 80 (17) 87 (10)
196/211 210/225 200/215 214/229 214/229 213/198 212/227 226/241 226/241
83 (13) 72 (19) 89 (14) 69 (24) 83 (13) 65 (22) 79 (20) 88 (21) 77 (13)
218 (63) 190 (47) 76 (10) 84 (8) 84 (5) 135 (52) 201 (58) 151(53) 162(54)
99 (16) 74 (19) 75 (10) 77 (8) 81 (12) 81 (16) 77 (22) 74 (22) 65 (18)
66 (32) 74 (26) 106 (23) 112 (8) 101 (18) 73 (34) 94 (13) 74 (23) 77 (25)
68 (7) 76 (8) 120 (3) 139 (6) 126 (7) 72 (8) 80 (6) 88 (6) 77 (7)
116 (15) 132 (7) 75 (16) 79 (16) 75 (7) 183 (12) 182 (17) 139 (19) 133 (13)
vos and mevinphos, for which the unexpected high recoveries obtained with the latter procedure contrasted with the satisfactory recoveries (77% and 83%, respectively) obtained when the ultrasonic energy was avoided. RSDs systematically increased with the extraction temperature for a significant number of analytes (9 out the 15 test OPPs), a trend also observed with UA-MSPD and that could point to a (partial) degradation of some of the analytes with temperature. The situation became much more dramatic for triazines. Unacceptable recoveries above 135% and RSDs over 47% were found for atraton, prometon, simetryn, ametryne, prometryne and terbutryn with HA-MSPD, a result that, irrespective of its cause (e.g., coextraction of other matrix components [12,13] or degradation [11]), detracted from further investigation. The feasibility of UA-MSPD with a sonoreactor [11] was investigated. Results obtained after 3 min UA-MSPD for OPPs with the sonoreactor follows a similar trend to that observed after 30 min UA-MSPD in an ultrasonic bath (Table 1). The only exceptions were observed for fenthion, which showed much higher recovery in the sonoreactor experiments (124% vs 83% with bath); and for dichlorvos, which shows a low recovery of 50%. This last result was due to the degradation of dichlorvos during the treatment with the sonoreactor (Fig. 1A) and explains its high RSD. A slight (but systematic) decrease was observed in the RSDs calculated (12%) as compared to those of previously discussed treatments. Results obtained for triazines in this part of the study were closer to those found after 30 min MSPD at 40 ◦ C than to those obtained by 30 min UA-MSPD in an ultrasonic bath, with still unacceptably high recoveries for most compounds. Reducing the sonication time to 1 min yielded recoveries in the 82–101% range with a mean RSD of 12% for most of the OPPs investigated (exceptions, mevinphos, diazinon and disulfoton). In addition, the degradation previously observed for dichlorvos was avoided (Fig. 1B; recovery, 82%; RSD, 16%). Despite the still rather low extraction efficiency observed for atraton (66%), results for atrazines were more satisfactory (average recovery, 94%; RSD, 6%) than with previously assayed methods, a fact specially evident for
atrazine, propazine and terbuthylazine (recoveries of 120%, 139% and 126%, respectively). The shorter extraction time involved by this treatment also contributes to avoid the (apparent) degradation observed at higher temperatures (i.e., MSPD at 40 ◦ C) or longer sonication times (i.e., 3 min sonoreactor UA-MSPD). 3.2. Analytical performance The developed 1 min sonoreactor UA-MSPD method was finally selected and evaluated in terms of intermediate precision (between-day RSD, Table 2). Mean recovery for OPPs from the repeatability assays was 95% (RSD, 12%; n = 3), while the value for the intermediate precision was 96% (RSD, 10%; n = 7). These results were similar to those reported in the literature for OPP determi-
Fig. 1. Fragmentograms of dichlorvos after (A) 3 min and (B) 1 min UA-MSPD with sonoreactor. Spiking level, 0.5 g g−1 .
148
J.J. Ramos et al. / J. Chromatogr. A 1212 (2008) 145–149
Table 2 Inter-day recoveries and reproducibility (as RSD; n = 7, within a week) of the UAMSPD with sonoreactor method developed for the OPPs and triazines investigated (spiking level, 0.5 g g−1 ). Limits of detection (LODs) as calculated for real-life apples (g kg−1 ). UA-MSPD-sonoreactor (1 min) Compound
% Recovery
RSD
LODs (g kg−1 )
Dichlorvos Mevinphos Dimethoate Diazinon Disulfoton Parathion-methyl Paraoxon-ethyl Malathion Fenthion Chlorpyrifos-ethyl Parathion-ethyl Bromophos-methyl Chlorfenvinphos Bromophos-ethyl Ethion
81 73 106 108 115 97 98 102 99 93 99 88 94 92 96
11 9 12 11 9 8 12 15 11 10 10 10 10 10 9
13 72 205 12 61 1.4 42 6.6 2.8 1.7 2.1 1.0 4.1 4.9 4.2
Atraton Prometon Atrazine Propazine Terbuthylazine Simetryn Ametryne Prometryne Terbutryne
83 83 106 118 108 89 89 91 87
22 15 11 14 12 24 14 8 14
36 21 14 9.5 3.8 27 11 8.5 18
nation in fruits [12] by MSPD, and in fruit juice by SPE assisted by ultrasound [5], and better than those obtained with other miniaturised techniques [13,14]. Triazines showed a more variable, although still satisfactory, behaviour regarding the between-day precision, with recoveries in the 83–118% range and RSDs varying from 8% to 24%. Although examples publishing more satisfactory reproducibility can be found in the literature [15], these studies typically involved
Fig. 3. Fragmentograms obtained for bromophos-ethyl in a Golden delicious apple (A) without and (B) with fortification at the 0.5 g g−1 level.
larger amounts of sample and solvent and more laborious and time consuming treatment protocols. Detection limits (LODs), determined as the concentration yielding a signal-to-noise ratio of 3 for the extracted quantification ion (Table 2), were in the 1–42 g kg−1 range. Higher LODs were obtained for mevinphos (72 g kg−1 ), but only those of disulfoton and dimethoate (61 and 205 g kg−1 ) were above the EU MRLs set for the investigated fruits. These low LODs proved the general feasibility of the proposed analytical procedure for the accurate and sensible determination of the OPPs and triazines studied even if only 100 mg of sample was used (Fig. 2). Finally, the applicability of the proposed method to the determination of the test pesticides in commercial fruits, i.e. four apple varieties, pear and apricot, was evaluated. None of the target compounds was found at detectable levels in pear, apricot and in the Pink lady and Granny Smith apples. Bromophos-ethyl was detected
Fig. 2. Typical reconstructed fragmentograms obtained for an organic Gala apple sample fortified at the 0.5 g g−1 level after 1 min UA-MSPD.
J.J. Ramos et al. / J. Chromatogr. A 1212 (2008) 145–149
at levels below the quantification limit (LOQ; S/N ratio = 10) in the Golden delicious and Gala apples (i.e., concentration < 15 g kg−1 ). Chlorfenvinphos was also found at a level below LOQ in the Gala apple (Fig. 3). 4. Conclusions A method based on UA-MSPD with a sonoreactor has been developed and its suitability for the accurate determination of OPP and triazine residues in a variety of fruits without any additional clean-up step before GC–MS analysis demonstrated. Compared with classic MSPD, the proposed method improves the general extraction efficiency, decreases the RSDs and allows complete sample treatment within a few minutes. Recoveries above 81% were obtained for all investigated pesticides. RSDs lower than 12%, were obtained for OPPs. More variable but still acceptable values (8–24%) were found for triazines. These results, combined with the low LODs, proved that the proposed procedure is suitable for the accurate determination of the target analytes at levels set in current legislations although as a small amount of sample as 100 mg is used.
149
Acknowledgments R.R.-O. acknowledges to FCT for grant SFRH/BPD/23072/2005. J.J.R. thanks MEC for FPI grant and aid EST20070624838. J.J.R. and L.R. thank MEC (CTQ-2006-14993/BQU). References [1] [2] [3] [4] [5] [6] [7] [8]
[9] [10] [11] [12] [13] [14] [15]
S.A. Barker, J. Biochem. Biophys. Methods 70 (2007) 151. E.M. Kristenson, L. Ramos, U.A.Th. Brinkman, Trends Anal. Chem. 25 (2006) 96. H.B. Xiao, M. Krucker, K. Albert, X.M. Liang, J. Chromatogr. A 1032 (2004) 117. S. Navarro, J. Oliva, A. Barba, C. García, J. AOAC Int. 83 (2000) 1239. B. Albero, C. Sanchez-Brunete, J.L. Tadeo, J. Agric. Food. Chem. 51 (2003) 6915. F. Priego-Capote, M.D. Luque de Castro, Trends Anal. Chem. 23 (2004) 370. J.J. Ramos, M.J. González, L. Ramos (in preparation). Regulation (EC) No. 396/2005 on maximum residue levels (MRLs) of pesticides found in or on food and feed of plant and animal origin, OJ L 70, 16.03.2005, p. 1. S. Lundstedt, B. van Bavel, P. Haglund, M. Tysklind, L. Oberg, J. Chromatogr. A 883 (2000) 151. X.C. Chu, Z.X. Hu, H.Y. Yao, J. Chromatogr. A 1063 (2005) 201. H.M. Santos, J.L. Capelo, Talanta 73 (2007) 589. H.S. Dórea, F.M. Lanc¸as, J. Microcol. Sep. 11 (1999) 367. M. Chai, G. Tan, A. Lal, Anal. Sci. 24 (2008) 273. ˜ A. Juan-García, J. Manes, G. Font, Y. Picó, J. Chromatogr. A 1050 (2004) 119. E.J. Avramides, S. Gkatsos, J. Agric. Food Chem. 55 (2007) 561.