Métodos individual e multirresidual para determinação de agrotóxicos em camas biológicas por cromatografia gasosa acoplada à espectrometria de massas

Abstract

Besides the possibility to perform determination of a large number of pesticides in only a single chromatographyc analysis using multiresidual methods, sometimes, due to diversity in pesticides chemical composition and particular characteristics, it is still required develop single-residue method, such as for dithiocarbamates (DTC). Therefore, in this study, firstly, was optimized and validated a single method to determine DTC (mancozeb) in biobeds using GC-MS to evaluate biobeds efficiency on DTC degradation. Validation study was performed and assessed the analytical curve linearity (r2), detection and quantification limits (LOD and LOQ), accuracy (recovery%), precision (RSD%) and matrix effect. Recovery experiments were carried out with a standard spiking solution of the DTC pesticide thiram. Blank biomixture samples were spiked at the three levels corresponding to the CS2 concentrations of 1, 3 and 5 mg kg-1, with seven replicates each (n = 7). The method presented recoveries within the range of 89-96% and RSD ≤ 11%. The analytical curves were linear in the concentration range of 0.05 to 10 μg CS2 mL-1 (r2 > 0.9946). LOD and LOQ were 0.2 and 1.0 mg CS2 kg-1, respectively, and the calculated matrix effects were not significant (≤ 20%). The validated method was applied to 80 samples (biomixture), from sixteen different biobeds. Ten percent of samples presented CS2 concentration below the LOD (0.2 mg CS2 kg-1) and 70% of them showed results below the LOQ (1.0 mg CS2 kg-1), which demonstrates the biobeds capability to degrade DTC. The second step was to devepol and validated a method (modification of the New Dutch mini-Luke method) for multiresidue determination of captan and its degradation product (tetrahydrophytalimide), chlorothalonil, chlorpyrifos-ethyl, fenithrothion, methidathion, phosmet and tebuconazole in biobed, using GC-MS/MS. The dichloromethane use and the evaporation step were eliminated. This improvement avoids chlorinated solvents use and allows final extract direct injection into the GC-MS/MS system. After method optimization, validation study was performed based on SANTE Guidance Document. Linearity of analytical curves (r2), residuals, limits of detection (LOD) and quantification (LOQ) and matrix effects (%), were assessed by the analysis of standards solutions prepared in organic solvent and in blank biobed extract at eight concentrations, with seven replicate injections of each standard solution. Accuracy (as recoveries %) precision (as repeatability of recoveries experiments-RSD% and as intermediate precision) were also determinate. Two labeled internal standards were used for quality control. PCB-153 was used as procedure internal standard (P.I.S.), and HCB-C13 was used as instrument internal standard (I.I.S.). Recovery study was performed by the analysis of blank biobed samples spiked at 50, 100 and 200 μg kg-1 (n=7). All pesticides presented average recoveries within the range of 74 and 117%, except captan which was indirectly detected via tetrahydrophytalimide, its degradation product, and also had acceptable recoveries. The RSD values calculated from the 7 replicated recovery samples at 3 concentration levels were below 20% (2 to 14%). The residuals was lower than 20% for all pesticides. The method was linear within the range of 1 to 300 ng mL-1 for 2 pesticides and of 5 to 300 ng mL-1 for 4 pesticides and for tetrahydrophytalimide (the degradation product of captan). The r2 of the analytical curves calculated from matrix-matched standards were > 0.99 for all pesticides except for chlorpyrifos-ethyl, that had an r2 0.98. All studied pesticides presented the LODs of 10 or 20 μg kg-1 and LOQs of 50 μg kg-1. The calculated matrix effects were higher than 25% for all pesticides, excepted for chlorpiripfós-ethyl, that was -20%.