Resíduos de agrotóxicos em tomate: validação de métodos por cromatografia gasosa acoplada à espectrometria de massas e monitoramento

In-lab validation of an analytical method, GC-MS, was performed to determine pesticide residues in tomato samples. Sample preparation was compared using the modified QuEChERS method and modified NL-method. The modified QuEChERS method was validated and presented better recovery results. The analytical parameters of linearity, selectivity, matrix effects, accuracy, precision, detection limit, quantification limit and measurement uncertainty were evaluated. Linearity was demonstrated by proving the premises of simple linear regression. The deviation from normality calculated by the Ryan-Joiner test was not significant (p > 0.10). Regression residuals were independent as calculated by the Durbin-Watson test (p > 0.10). Homoscedasticity, as demonstrated by the Levene test, had variance homogeneity along the curve (p > 0.05). The regressions were significant (p < 0.001) with nonsignificant linear deviations (p > 0.05), as calculated by the ordinary least squares method, in the ranges from 0.01 to 0.06 mg kg-1 for chlorpyrifos methyl, endrin, etrimfos and metolachlor, 0.05 to 0.30 mg kg-1 for atrazine, chlorfenapyr, ethoprophos, fenitrothion and procymidone and 0.10 to 0.60 mg kg-1 for dieldrin. The method was selective and the curves were prepared in tomato matrices for all analytes. Veracity and precision were demonstrated in the analytes quantification. The mean values of the recovery rate varied between 71.69 and 110.15 %. Relative standard deviations of 13.03 % were obtained under repeatability conditions and 19.61 % under conditions of intermediate precision. The measurement uncertainty was estimated considering the uncertainty of the calibration curve and the data of intra-laboratory reproducibility. The validated method was applied to the monitoring of tomato samples collected in the state of Minas Gerais. 20 tomato samples were monitored and no unsatisfactory results were obtained for the studied analytes, considering the detection limits of 0.04 mg kg-1 for atrazine, chlorfenapyr, dieldrin and fenitrothion, 0.01 mg kg-1 for chlorpyriphos methyl, endrin, etrimfos and metolachlor, 0.03 mg kg-1 for ethoprophos and 0.02 mg kg-1 for procymidone.