Métodos individual e multirresidual para determinação de pesticidas em grãos e derivados por GC-PFPD, GC-ITD-MS e LC-MS/MS

Abstract

Due to the complexity of grains like soybean, corn and rice, and the very low pesticide concentration expected for these samples, it is really necessary to develop efficient and reliable analytical methods to identify and quantify their residues. The possibility to perform the determination of a large number of pesticides in just a single chromatographic analysis is the major advantage of the multi-residue methods. However, there are important compounds that are not possible to be analyzed together, like the dithiocarbamates (DTC). For that, is so important to develop and validate single methods. Due to low acute toxicity, combined with strong action, low-cost production and low environmental persistence, the dithiocarbamates are still using worldwide. The first part of this study reports a sensitive analytical single method suitable for the quantitative analysis of DTCs in soybeans, rice and corn, using GC-PFPD and GCITD- MS. For the method validation studies, the following parameters were assessed: detection limit (LOD), quantification limit (LOQ), precision (repeatability and intermediate precision) and accuracy, via recovery experiments at 0.05, 0.1 and 0.5 mg CS2 kg-1 spiking levels. For soybeans is necessary a clean-up step due the high fat content in that crop, because of it, same sorbents (Aluminum oxide, Florisil®, Silica, PSA, C18 and C18OH) were tested and the best results were obtained applying silica gel. For the extraction procedure, an amount of sample was weighed and water added. The spike solution (thiram) was added to the sample, 25 mL of isoctane and 150 mL of tin (II) chloride solution in hydrochloric acid were added. The bottles were closed immediately and heated for 1 h in a water bath at 80 °C, under shaking. After cooling, 2 mL aliquot of the upper organic layer of each bottle was pipetted into a tube containing 50 mg of silica gel, for the clean-up step. The analytical curves were linear from 0.02 to 1.0 mg CS2 L-1 with determination coefficients (r2) higher than 0.99 for both detection systems and to all crops. Method LOD and LOQ values were 20 and 50 μg CS2 kg-1, respectively, for soybeans and corn. For rice LOQ was just 40 μg CS2 kg-1. Very good precision was obtained with RSD between 1.8 and 7.9% and recoveries from 67 to 103%, for all matrices. The results obtained in the validation step allow us to conclude that both chromatographic detection systems are appropriate to determine residues of DTCs in grains. The second step was to develop and validate a method for the analysis of 75 pesticides in corn grain and breakfast cereals using LC-MS/MS, ESI positive mode. The extraction procedure was the modified QuEChERS method, however, excluding the dispersive SPE clean-up step. To that end, milled and homogenized corn grain and breakfast cereals with water were spiked with 75 pesticides, at 3 spiking levels (10, 20 and 50 μg kg-1, n=6), plus the procedure internal standard (quinalfos). A volume of 10 mL of acetonitrile with 1% acetic acid was added and the tubes were vigorously shaken for 1 minute. The tubes were uncapped, anhydrous magnesium sulfate were added, the shaking procedure was repeated and the extract was centrifuged at 4000 rpm, for 4 min. The upper layer of the extracts was transferred to chromatographic vial, containing 0.5 mL of the instrumental internal standard (propoxur) in methanol. The linear range wasevaluated and validated between 0.1 and 50 μg L-1 of the analytical curves (7 concentrations, n=6). Also LOD, LOQ, matrix effect, as well as precision (as RSD%) and accuracy (as recovery), for 75 pesticides, were studied The LOD were calculated based on practically realized repeatability RSD. To determine the precision, accuracy and LOQ, blank samples were spiked at 10, 20 and 50 μg kg-1. The results showed recoveries between 70 and 120% and RSDs <20% for the majority of the pesticides studied, even at the lowest level. As expected, the matrix effect did not have influence on the results, and the values of LOD and r2 were considered as satisfactory for the purpose of this study.