Article pubs.acs.org/JAFC
Quantitative Screening of Agrochemical Residues in Fruits and Vegetables by Buffered Ethyl Acetate Extraction and LC-MS/MS Analysis Manjusha R. Jadhav, Dasharath P. Oulkar, Ahammed Shabeer T. P., and Kaushik Banerjee* National Referral Laboratory, ICAR−National Research Centre for Grapes, P.O. Manjri Farm, Pune 412307, India S Supporting Information *
ABSTRACT: A buffered ethyl acetate extraction method is proposed for the simultaneous analysis of 296 agrochemicals in a wide range of fruit and vegetable matrices by liquid chromatography−tandem mass spectrometry (LC-MS/MS). The optimized quantity of acetate buffer (1% acetic acid + 0.5 g of sodium acetate per 10 g of sample) adjusted the pH of each test matrix to 5− 6, which in turn significantly improved recoveries of acidic and basic compounds. The role of diethylene glycol (used in the evaporation step) on signal suppression of certain compounds was evaluated, and its quantity was optimized to minimize such an effect. The method was validated in grape, mango, drumstick, bitter gourd, capsicum, curry leaf, and okra as per the DGSANCO/12571/2013 guidelines. Recoveries in the fortification range of 1−40 μg/kg were within 70−120% with associated relative standard deviations below 20% for most of the compounds. The method has potential for regulatory and commercial applications with a generic approach. KEYWORDS: buffered extraction, agrochemical residues, multiresidue analysis, fruits and vegetables, LC-MS/MS, matrix effect, single-laboratory validation
■
years.8 The natural pH of most of the fruits and vegetables ranges between 2 and 7, which might influence the stability and extraction efficiency of certain ionizable (acidic and basic) compounds. The existing method4 was falling short in achieving sufficient recovery of such compounds. The significance of the pH adjustment by means of buffering with phosphate buffer, acetate buffer, citrate buffer, etc., is reported earlier in the literature.9−11 In the buffered QuEChERS method,10 acetate buffer was used for pH adjustment, which helped to improve the extraction and stability of certain compounds such as folpet, dichlofluanid, chrorothalonil, and pymetrozine, and the method eventually emerged as an AOAC Official Method.12,13 In the current study we endeavored to develop a comprehensive residue analysis methodology with a generic approach encompassing matrices of variable pH range for largescale multiresidue analysis including acidic and basic compounds. The existing ethyl acetate based extraction method4 has been modified by including a pH adjustment step with an optimized quantity of acetate buffer (sodium acetate or acetic acid). Furthermore, the effect of buffering on the performance of cleanup and the effect of dilution on matrix-induced signal suppressions were also evaluated. The quantity of diethylene glycol (DEG) as a keeper at the evaporation/solvent exchange step was optimized. The residue analysis method validation was carried out per the DG-SANCO guideline.14
INTRODUCTION This paper engages with the development of a generic approach for large-scale multiresidue analysis of pesticides in a variety of fruits and vegetables. Analysis of pesticide residues is an important quality evaluation step that addresses the food safety concerns of consumers and allows compliance check for a food consignment to national and international food safety regulations. With the establishment of the Food Safety Standards Authority, the scope of residue monitoring of agrochemicals in India has been expanded enormously over the past decade.1−3 Implementation of an effective residue monitoring program envisages a rapid turn-around time in laboratory analysis, and for this a single multiresidue analysis method should incorporate as many chemicals as possible with simultaneous method applicability to a wide variety of matrices. In one of our earlier publications, a multiresidue method was reported for the simultaneous estimation of 82 pesticide residues in grapes using ethyl acetate extraction and liquid chromatography−tandem mass spectrometric (LC-MS/MS) analysis.4 The method was further successfully evaluated for other matrices such as mango,5 pomegranate, orange, and apple6 and also for gas chromatography−mass spectrometry (GC-MS)-amenable compounds.7 This method has been adopted by commercial testing laboratories of India for pesticide residue testing in the specified commodities. However, the existing method is unable to provide satisfactory recovery (accuracy in analysis) for certain acidic and basic pesticides, especially in matrices with extreme pH values. With increasing trade and food safety requirements, it is important to expand the scope of residue monitoring by covering more and more commodities, especially when rapid alert notifications in the European Union (EU) countries are recorded in recent © XXXX American Chemical Society
Special Issue: IUPAC - Analysis of Residues in Food Received: October 29, 2014 Revised: January 17, 2015 Accepted: February 2, 2015
A
DOI: 10.1021/jf505221e J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
■
methanol/20 mM ammonium formate in water (1:1 v/v) or control matrix extracts. Existing Method. Homogenized samples (10 g) were extracted with ethyl acetate (10 mL) per the procedure reported earlier.4 Optimization of Buffer Concentration. Homogenized grape (10 g) spiked at 10 ng/g was extracted with 10 mL of ethyl acetate either using no buffer or in the presence of acetic acid (1% v/v) plus variable amounts of sodium acetate (0, 0.1, 0.5, 1, and 2 g; n = 6) along with 10 g of sodium sulfate. No PSA cleanup was employed to avoid resultant pH-dependent effects on the recovered analytes. The average recoveries were compared against the respective matrix-matched standards employing single-point calibration. Effect of Dispersive SPE Cleanup Using PSA. For this, two sets of homogenized control grape samples (10 g) were extracted with the existing method4 and optimized sodium acetate buffer method, each in six replicates. The resultant extracts were spiked at 10 ng/g level and treated as (1) no cleanup, (2) cleanup with 25 mg of PSA per 5 mL of extract, and (3) cleanup with 125 mg of PSA per 5 mL of extract. The cleanup efficiencies in each case were evaluated by gravimetric comparison of the residual matter. At each variable amount of PSA, the pH of the corresponding ethyl acetate extract was measured after dilution with water and under continuous stirring with a magnetic stirrer. The average concentration of the analytes at different treatments were estimated as accuracy (%) against solvent standards and compared. Optimization of DEG Quantity. Two milliliters of ethyl acetate extract of control grape and pure ethyl acetate solvent were taken in six different sets of 10 mL evaporation tubes and spiked with a mixture of pesticides at 10 ng/mL. DEG (10% in methanol) was then added in each tube at different volumes of 0, 10, 25, 50, 100, and 200 μL, respectively. After evaporation under a gentle stream of nitrogen at 35 °C, the residues were reconstituted in methanol/water (50:50) and injected into LC-MS/MS after passing through a Nylon-6,6 membrane filter (0.2 μm). The area (%) responses of the test analytes were compared. Full-scan analysis of selective samples was done using information dependent acquisition (IDA) functionalities of a QTrap LC-MS/MS system. Proposed Method. The sample (10 g) was extracted with 10 mL of ethyl acetate (+ 1% acetic acid) in the presence of sodium acetate (0.5 g) and anhydrous sodium sulfate (10 g) by means of vortexing (2 min) and subsequent centrifugation at 5000 rpm for 5 min. An aliquot of the upper ethyl acetate extract (5 mL) was drawn and treated with PSA (25 mg) by means of vortexing for 30 s. The cleaned supernatant extract (2 mL) was placed in a 10 mL test tube, and to it 10% DEG (10 μL) was added. This mixture was subsequently evaporated to dryness under a gentle stream of nitrogen (in a low-volume concentrator) at 35 °C. The residue was dissolved in 1:1 methanol plus 20 mM ammonium formate in water (2 mL) by ultrasonication (1 min) followed by vortexing (30 s). This solution was centrifuged at 10000 rpm for 5 min. The extracts were analyzed by LC-MS/MS after passing through a Nylon-6,6 membrane filter (0.2 μm). LC-MS/MS Analyses. Chromatographic separation of the test compounds was achieved on an Ultra AQ C18 column (3 μm, 100 mm length, 2.1 mm i.d.; Restek Corp., Bellefonte, PA, USA). The mobile phase composition was (A) water with 10 mM ammonium formate and (B) methanol with 10 mM ammonium formate, each with 0.1% formic acid, at flow rate of 0.5 mL/min with a gradient profile.15 The column oven temperature was maintained at 40 (±1) °C. The injection volume was 10 μL. Measurements were done with electrospray ionization (ESI) in positive polarity with optimized multiple reaction monitoring (MRM) transitions. The source parameters, namely, ion source voltage (5500 V), nebulizer gas (30 psi), heater gas (60 psi), ion source temperature (550 °C), and curtain gas (40 psi), were maintained throughout the analysis. Residue estimation was performed by retention time dependent “scheduled multiple reaction monitoring” (sMRM) with two mass transitions for each test molecule, one for quantification and the other for confirmation. The MRM detection window (MRM DW) was 90 s with the target scan time of 1 s. The ion ratio for the two mass transitions was used for unambiguous identification of each pesticide
MATERIALS AND METHODS
Chemicals. Certified reference standards of all test compounds were of >97% purity and purchased from Ehrenstorfer GmbH (Augsburg, Germany) and Sigma-Aldrich (Steinheim, Germany). All of the solvents were of analytical reagent (AR) or ultragradient HPLC grade and obtained from J. T. Baker (Center Valley, PA, USA). All other reagents were of AR grade and purchased from Thomas Baker (Mumbai, India) and Merck India Ltd. (Mumbai, India). The dispersive solid phase extraction (dSPE) sorbent, namely, primary− secondary amine (PSA), was received from Agilent Technologies (Santa Clara, CA, USA). Ultipor Nylon-6,6 membrane filters (0.2 μm pore size and 13 mm diameter) were purchased from Pall Life Sciences (Ann Arbor, MI, USA). HPLC grade water was obtained through Sartorius water purification system (Göttingen, Germany). Apparatus. For sample preparation a mixer with grinder (0.5 and 2 L capacity, model GX7, Bajaj India Ltd., Mumbai, India), vortex mixer (Genie 2T, Imperial Biomedicals, Mumbai, India), ultrasonic bath (Oscar Electronics, Mumbai, India), low-volume concentrator (TurboVap LV; Caliper Life Sciences, Russelsheim, Germany), highspeed refrigerated centrifuge (Kubota 6500, Kubota Corp., Tokyo, Japan), and a microcentrifuge (Microfuge Pico, Kendro D-37520, Osterode, Germany) were used. A pH meter (Thermo Orion, Chelmsford, MA, USA) was used after calibration using buffer solutions of pH 4 and 7. Residue analysis was performed using an LCMS/MS [Shimadzu UFLC XR connected to API 5500 QTrap (AB Sciex, Toronto, Canada) mass spectrometer], supplied by Labindia Instruments Pvt. Ltd. (Mumbai, India). Selection of Matrices and Sample Homogenization. Fruit and vegetable matrices were selected considering their importance for export and domestic consumption. The selected matrices were grape (Vitis vinifera L.), mango (Mangifera indica L.), capsicum (Capsicum annuum L.), okra (Abelmoschus esculentus L.), curry leaf (Murraya koenigii L.), drumstick (Moringa oleifera L.), and bitter gourd (Mormordica charantia L.). The method was initially optimized in grape and then extended to other commodities. The control matrices were obtained from the organically grown local farms and used after ensuring absence of any of the target incurred residues. About 0.5−2 kg sample sizes of different matrices were selected (0.5 kg for curry leaf, 1 kg for other vegetables, and 2 kg for grape and mango) to obtain a homogenized representative sample. The samples were chopped (except grape) after removal of stalks, caps in the case of vegetables, and stone in the case of mango, followed by crushing in a mixer− grinder. In the following step, about 0.1−0.4 kg of this crushed material (0.1 kg for curry leaf; 0.2 kg for bitter gourd, drumstick, and okra; 0.4 kg for grape, mango, and capsicum) was again homogenized in a smaller jar (0.5 L capacity) to obtain a finely crushed homogeneous sample with or without addition of water. In the case of bitter gourd, okra, and drumstick, water was added in the proportion of 1:1 w/w, and for curry leaf, the ratio was 1:3 w/w. The proportion of water added was based on the inherent moisture content of the matrices and was optimized through separately conducted experiments to achieve the desired level of homogeneity (RSD < 10% when analyzed in six replicates). Selection of Agrochemicals. The study included 296 multiclass agrochemicals (including selective metabolites) with different biological activities such as insecticides, acaricides, fungicides, herbicides, and plant growth regulators, which were either neutral, acidic, or basic by chemical nature. The compounds were selected considering the MRL database of the European Union2 and Food Safety Standards Authority of India.3 Details of the test analytes are presented in Supplementary Table 1 of the Supporting Information. Preparation of Standard Solutions. The stock solutions of individual certified reference standards (CRS) were prepared by accurately weighing 10−20 mg of each analyte in volumetric flasks (certified A class) and dissolving in 10 mL of methanol (stored at −20 °C). Intermediate standard mixtures of 10 μg/mL and a working standard mixture of 1 μg/mL were prepared by diluting the stock solutions in methanol. From this, the calibration standards within the range of 0.2−40.0 ng/mL were prepared by serial dilution with B
DOI: 10.1021/jf505221e J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
Figure 1. Effect of buffering on recovery of analytes. per the criteria of the SANCO guideline.14 A representative LC-MS/ MS chromatogram is given in Supplementary Figure S1 in the Supporting Information. Method Validation. The performance of the analytical method was assessed per the SANCO guideline.14 Standard calibration was established on the basis of solvent and matrix-matched standards in the concentration range of 0.2−40 ng/mL. Limits of detection (LOD) and quantification (LOQ) were determined by considering signal-to-noise ratios (S/N) of 3 and 10, respectively, for the quantifier MRM ensuring that the qualifier MRM has S/N of ≥3:1 at LOQ. For the recovery experiment, at first the samples were spiked at the desired level from a standard working solution of 1 μg/mL and then thoroughly homogenized. Ten grams of homogenized mass was weighed in a 50 mL polypropylene tube and then extracted per the proposed method. The recovery experiments were carried out in grape at five spiking levels, that is, 1, 2.5, 5, 10, and 40 ng/g (n = 6). For all other matrices three spiking levels were selected, which were equivalent to 1, 5, and 10 ng/g for mango and capsicum; 2, 10, and 20 ng/g for bitter gourd, okra, and drumstick; and 4, 10, and 40 ng/g for curry leaf on an actual weight basis. Recoveries (%) were calculated against the respective matrix-matched calibrations. The precision in terms of repeatability in recovery was estimated as % RSD. The results were summarized in different categories, namley, I, recovery (70− 120%), RSD ≤ 20%; II, recovery (60−140%), RSD ≤ 20%; III, recovery (60−140%), RSD > 20%; IV, recovery (5 due to better partitioning into ethyl acetate. To achieve the best compromise for optimum extraction of acidic and basic analytes, for the final method, 0.5 g of sodium acetate + 1% acetic acid was selected as the optimized quantity for buffering. The pH of all the test commodities was effectively adjusted within the range of 5−6 irrespective of the initial matrix pH (Supporting Inforamtion, Supplementary Figure S2). Effect of Dispersive SPE Cleanup Using PSA. In the existing ethyl acetate method4 25 mg of PSA per 5 mL of extract had been used for dispersive cleanup to remove matrix coextractives such as fatty acids and sugars. In the initial studies, with this unbuffered method, we found that the PSA cleanup resulted in severe losses of certain acidic pesticides in all test matrices. Therefore, we decided to evaluate the effect of PSA D
DOI: 10.1021/jf505221e J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
Figure 3. Signal suppression of (A) cyromazine and (B) methamidophos (10 ng/mL) corresponding to (1) no DEG, (2) 10 μL of DEG, and (3) 200 μL of DEG.
Figure 4. Demonstration of dilution approach for reducing matrix effect in different commodities.
of the matrix examined. Considering its polar nature and early elution, we assumed that DEG might be contributing to matrix effect, and hence its impact was investigated further. During the DEG optimization experiment, it was observed that the earlyeluting compounds such as methamidophos, acephate, atrazine desethyl, atrazine desethyl 2-hydroxy, cyromazine, and dinotefuran showed significant signal suppressions (30−80%) as the quantity of DEG was increased from 0 to 200 μL (Supporting Information, Supplementary Figure S3). The trend in signal suppression of these analytes was similar in the case of pure solvent and matrix, which demonstrated that DEG rather than the matrix coextractives was responsible for signal suppressions. The effect was most prominent in the case of methamidophos, the earliest eluting compound (tR 1.9 min), which had almost 40% signal suppression with 10 μL of DEG and almost 80% suppression after the addition of 200 μL of DEG. Figure 3 shows how signal suppression due to the addition of DEG resulted in lower signal-to-noise ratios (S/N) and, thus, increased LOQs of cyromazine (by 5 times) and methamidophos (by 8 times).
acetate phase, which reduced direct interaction of PSA with the acidic analytes. This was another important and favorable aspect of buffering, and thus there was no need to omit PSA cleanup; good recoveries of acidic compounds could still be achieved. Furthermore, with this optimum quantity, signal suppression was significantly reduced by 20−50% for compounds, namely, dodine, bupirimate, fenpropimorph, imazalil, mexacarbate, methoprotryne, pirimicarb, prometon, prometryne, spiroxamine, and 6-BA. Optimization of DEG Quantity. In the existing ethyl acetate based extraction method,4 200 μL of DEG (10% solution in methanol) for 2 mL of final extract is employed as a keeper during the solvent evaporation step. The significance of the use of DEG solution is explained by Mol et al.21 In brief, addition of this high boiling, viscous liquid in the extract protects against loss of analytes during evaporation and avoids hard drying of the residual matrix, which otherwise may trap analytes and make their solubilization difficult. In our routine analysis, it was observed that the early-eluting compounds (tR 2−5 min) suffer serious signal suppressions (>50%) irrespective E
DOI: 10.1021/jf505221e J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
Figure 5. Summarized data on the performance of the buffered ethyl acetate method for 296 compounds in all test matrices.
To evaluate this further, DEG solution (equivalent to 200 μL of 10% DEG solution/2 mL) along with methamidophos (10 ng/mL) in methanol was injected in the same chromatographic system and analyzed in full-scan mode using QTrap functionalities. Analysis was performed using enhanced mass spectra (EMS) as “survey scan” and enhanced product ion (EPI) as “dependent scan”. The extracted ion chromatogram of DEG (m/z 107.0) and enhanced product ion spectra revealed prolonged elution of DEG within 1−5 min (Supporting Information, Supplementary Figure S4), which could be the main reason for signal suppression of early-eluting compounds. The spectral confirmation was obtained by comparing the same with the earlier reported spectrum of DEG.22 Due to relatively higher viscosity and boiling point (245 °C) DEG might be incompletely evaporated and reduce ionization of early-eluting compounds. Volatile or semivolatile compounds such as dichlorvos, molinate, and isoprocarb suffered higher recovery losses (20−60%) in the case when no DEG was used in the evaporation step, which could be avoided when 10 μL of DEG solution was added. Therefore, considering the role of DEG to avoid hard drying of residues, instead of completely omitting its use, its volume was optimized to 10 μL in place of 200 μL4 Dilution and Matrix Effect. In Figure 4, the results corresponding to “without dilution” (WD) indicate matrix effect of the respective test matrix. In the case of grape, mango, capsicum, okra, and drumstick, 72−97% of the test compounds had moderate to low matrix effect (groups III and IV). In the case of bitter gourd and curry leaf, 40 and 23% of the compounds suffered from moderate to low matrix effects, whereas 46−56 and 15−22% of compounds belonged to the high (group II) and very high (group I) matrix effect categories, respectively, in terms of signal suppressions. Hence, quantification of the residues of these compounds was performed on the basis of matrix-matched calibration standards. Dilution led to cleaner extracts, reducing signal suppressions significantly.9 In recent publications,23,24 a similar approach has been demonstrated with reference to QuEChERS extraction. From the comparative study (Figure 4), it was concluded that 5−10 times dilution of the final extract effectively reduced matrix-induced signal suppression for most of the compounds. After dilution of the bitter gourd extract by 5 and 10 times, the percentage of compounds falling into groups I (very high matrix effect) and II (high matrix effect) came down from 60 to 10 and 3%, respectively. Similarly, for curry leaf, the dilution of extract by 5 and 10 times reduced the proportion of compounds falling into groups I and II from 77 to 10 and
3%, respectively. This study will thus help employ appropriate dilution for quantification against solvent standard and nullify the requirement of control matrix for preparation of matrixmatched standards. Method Validation. Calibration linearity employing linear or quadratic regression equation was established with r2 ≥ 0.995 for all of the test analytes in solvent and matrix-matched standards within the range of 0.2−40 ng/mL with a weighting factor of 1/x. The method LOD and LOQ values of all the analytes were determined in grape matrix based on S/N ratio and specified in Supplementary Table 2 in the Supporting Information. Results showed that, of 296 test compounds, 79.3% of the compounds had LOQs in the range of 0.2−1 ng/g whereas for 16.2 and 3.7% compounds, the LOQs were in the ranges of 1−5 and 5−10 ng/g, respectively. As about 80% of compounds could be easily quantified at the 1 ng/g level, it was selected as the lowest spiking level (LOQ) for recovery study in grape matrix. Per the SANCO guideline,14 recovery study at minimum two levels (at LOQ and another higher level) need to be performed and thus, considering the entire range of LOQs, that is, 1−10 ng/g, the recovery study was performed at five spiking levels, 1, 2.5, 5, 10, and 40 ng/g in grape matrix. Once the method was validated for grape matrix, it was extended to other matrices and corresponding LOQs were estimated. Recoveries at LOQ were estimated for all matrix−chemical combinations, and the lowest level at which recovery is reported is the method LOQ for that particular matrix (Supplementary Table 2). Figure 5 shows summarized method performance data classified into categories I−VI for all of the matrices. Of the 296 test compounds, 61.2−75.7% had recoveries in the range of 70−120 (±20)% at the lowest spiking level across the commodity matrices. At the regulatory default LOQ level of 10 ng/g, the recoveries for 87.1−94.9% of the compounds were within 70−120 (±20)%. Only 1−3.7% of the total compounds had recoveries 20%) and about 0−1% of compounds had lower recoveries (20%) at the 10 ng/g level. Of the 296 test analytes, 15−34% at the lowest spiking level and 0.33−3% at the highest spiking level (in all matrices) fell into category VI, which included the compounds below the LOQ or compounds that suffered matrix-specific degradation. Low recoveries were observed for compounds such as propamocarb and atrazine F
DOI: 10.1021/jf505221e J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
by offering stability and extractability to a wide range of compounds, especially those the stability of which is dependent on matrix pH. Use of a single extraction protocol for multiple matrices with minimum cleanup makes the method suitable for routine analysis in a commercial testing laboratory setup. The LC-MS/MS method enabled quantitative screening of a large number of target analytes at 1 ng/g and lower traces. Even the pesticides with problematic nature could be satisfactorily screened with sufficiently low LOQs (≤10 ng/g). The method has a good potential for implementation as a regulatory generic method for quantitative screening of residues in fruits and vegetables.
desethyl 2-hydroxy due to poor extraction in ethyl acetate as discussed earlier, although their RSD (%) values were acceptable. Lower and variable recoveries were recorded for alanycarb in grape, mango, drumstick, and okra; cycloxydim in grape, bitter gourd, capsicum, and okra; thiophanate methyl in all of the matrices except okra and bitter gourd; bifenazate in all of the matrices except mango, capsicum, and drumstick; benfuracarb and formetanate in all of the matrices except capsicum and bitter gourd; and ethoxyquin in all the matrices except drumstick, okra, and mango, probably due to their degradation in matrix. Some compounds showed matrixspecific recovery problems, which could be due to different reasons, including metaflumizone degradation in bitter gourd, low recoveries (50−69%) of bromadiolone and quinoxyfen, conversion of quizalofop-p-ethyl, haloxyfop, and fenoxaprop-pethyl into their metabolites, namely, quizalofop free acid, haloxyfop free acid, and fenoxaprop-P, respectively, in okra matrix. However, it was possible to achieve sufficiently low LOQs for most of these compounds for effective screening and quantification at or less than the regulatory MRLs. In the case of bitter gourd and curry leaf, the lowest numbers of compounds falling within the acceptable performance criteria (recovery = 70−120 ± 20%, at the lowest spiking level) were mainly due to very high matrix-induced signal suppressions. Comparison of the Proposed Method with the Buffered QuEChERS Method10 and the Existing Ethyl Acetate Method.4 The comparative method performances of the proposed method vis-a-vis the buffered QuEChERS method10 are shown in Supplementary Figure S5 in the Supporting Information for the grape matrix. For the basic compounds such as imazalil, thiabendazole, and cyromazine, the performances of both methods were comparable. For atrazine desethyl 2-hydroxy, formetanate, propamocarb, and pymetrozine, the recoveries for the proposed method were less as compared to acetonitrile extraction because of the comparatively less polar nature of ethyl acetate, but they showed acceptable RSDs (
