Desenvolvimento e validação de método multirresidual para determinação de pesticidas em café verde utilizando GC-MS (NCI-SIM)

Abstract

In this study, was developed and validated a multi-residue method for the determination of 40 pesticides in green coffee beans. To obtain a very homogeneous sample, the coffee beans were first milled, followed by a slurry preparation of coffee and water (ratio 1:4, w/w). For the extraction procedure modifications of the QuEChERS method were developed. Acetonitrile containing 1% acetic acid and the procedure internal standard (P.I.S.) was used as extraction solvent. After the acetonitrile addition, 3 g of anhydrous magnesium sulfate were added, to promote phase separation between the organic phase (acetonitrile) and the aqueous phase. For the dispersive solid-phase extraction (dSPE) clean-up step, two sorbents were tested in separate and in different amounts and were also tested mixtures of them. Ergo the best results were obtained using 500 mg C18-bonded silica together with 600 mg of magnesium sulfate, for drying. The pesticides were determined by gas chromatography tandem mass spectrometry (GC-MS). Considering the characteristics of the pesticides and the complexity of the coffee matrix the GC-MS had to be operated in negative chemical ionization mode (NCI) and single ion monitoring mode (SIM) which provided high sensitivity and selectivity to pesticide detection. The validation was performed by analyzing spiked samples at three different concentrations (10, 20 and 50 μg kg-1), with 6 replicates (n=6) at each spike concentration. For each one of the 51 pesticides determined by GC-MS (NCI-SIM), the linearity (r) of calibration curves, accuracy (as recovery percent), instrument and method limits of detection and quantification (LOD and LOQ), precision (as RSD percent) and matrix effects (percent) were determined. From the 51 pesticides studied, approximately 76% showed average recoveries between 70-120% and 75% RSD ≤ 20% for the spike concentration of 10 μg kg-1. For the spike concentrations of 20 and 50 μg kg-1, the recoveries (%) and RSD (%) values were even better. The method LOQ was 10, 20 and 50 μg kg-1, for respectively 33, 3 and 6 pesticides, 4 pesticides could only be detected via their degradation products and 5 compounds did not presented recoveries between 70-120% and RSD ≤ 20% thus the method LOQ could not be determined for those pesticides. The matrix effect was present for all the pesticides even with the use of a clean-up step, requiring therefore matrixmatched calibration standards for application in routine analysis.