Box Behnken design-based optimized extraction of non-dioxin-like PCBs for GC-ECD and GC-MS analyses in milk samples

Abstract A multivariate optimization process of the sample extraction procedure by Box-Behnken design through a global desirability function is described for the determination of six non-dioxin-like polychlorinated biphenyls (NDL-PCBs # 28, 52, 101, 153, 138 and 180) in milk by GC-ECD and mass spectrometry. Three factors were involved in refining the extraction conditions: the acetone percentage in the extraction mixture, the sample/solvent ratio, and the extraction time. The three-factor design required 26 experiments that were carried out in duplicate and in a randomized order to minimize the bias effects of uncontrolled variables. The optimized factors (acetone percentage: 30%; sample-to-solvent ratio: 0.11 g mL−1; extraction time: 45 min) ensured a low solvent consumption and a reduced extraction time, allowing a rapid and simultaneous preparation of multiple sample extracts. The method was validated according to the European directives (Decision 657/2002/EC, SANTE 2017/11813/EC) through the evaluation of linearity, selectivity, LOD, LOQ, recovery, precision, and ruggedness.