Development of liquid–liquid–liquid extraction clean–up in combination with preconcentration by dispersive liquid–liquid microextraction in analysis of seven pesticides from soya milk

Nowadays, the control of pesticide residues in food commodities is a requirement for compliance with the legislation, ensuring the safety of the population and international and national trade [1]. Frequent application of various pesticides simultaneously requires the development of multi–residue analysis methods. The development of such methods is often difficult since the different pesticides from various chemical classes have different physicochemical properties e.g. polarity, solubility, volatility as well as different octanol–water partition coefficients, which make their extraction and analysis difficult. Moreover, the presence of matrix interferences can significantly affect pesticides identification and quantification. Therefore, sample preparation is a crucial step in most analytical procedures. A perfect sample preparation provides not only extraction of the target analytes, but also sample clean–up. It removes the potentially interfering compounds to achieve a sufficient selectivity and sensitivity to guarantee that the detection limits of the method are in agreement with maximum residue limits. Some sample preparation methods have been developed for the analysis of pesticide residues such as liquid–liquid extraction [2] and solid phase extraction [2, 3], which have some drawbacks. In the present study, a new extraction method based on a three–phase system liquid–liquid–liquid extraction followed by dispersive liquid–liquid microextraction has been developed and validated for the extraction and preconcentration of seven pesticides in soya milk sample. The aim of this work was to develop a sample preparation method in which not only most analytes are transferred from the sample to the extraction solvent but the co–extracted components of the matrix are reduced as far as possible. The extraction mechanism is based on the different affinities of the species from soya milk towards each of the involved phases (aqueous solution, acetonitrile, and n–hexane), which provides high selectivity to the process. Most interference co–extracted components such as lipids, free fatty acids, and other hydrophobic compounds are extracted into n–hexane and will not be present in the final extract. Effect of various experimental parameters on the extraction efficiency was studied and optimized. Under the optimized extraction conditions limits of detection and enrichment factors for the analytes were obtained in the ranges of 0.11–0.35 µg L−1 and 562–933, respectively. The obtained extraction recoveries were between 56 and 93% and the calibration curves were linear in wide ranges with correlation coefficients ≥0.9985. Relative standard deviations were less than 7.0% for intra– (n=6) and inter–day (n=5) precisions (at two concentrations of 10 and 50 µg L−1 of each analyte). Finally, the proposed method was applied on different soya milk samples and chlorpyrifos was determined in one sample of soya milk at a concentration of 19 ± 3 ng mL−1 (n=3)