Simultaneous determination of bisphenol A and bisphenol B in beverages and powdered infant formula by dispersive liquid–liquid micro-extraction and heart-cutting multidimensional gas chromatography-mass spectrometry

Abstract

The purpose of this study was to establish a reliable, cost-effective, fast and simple method to quantify simultaneously both bisphenol A (BPA) and bisphenol B (BPB) in liquid food matrixes such as canned beverages (soft drinks and beers) and powdered infant formula using dispersive liquid–liquid micro-extraction (DLLME) with in-situ derivatisation coupled with heart-cutting gas chromatography-mass spectrometry (GC-MS). For the optimisation of the DLLME procedure different amounts of various extractive and dispersive solvents as well as different amounts of the derivative reagent were compared for their effects on extraction efficiency and yields. The optimised procedure consisted of the injection of a mixture containing tetrachloroethylene (extractant), acetonitrile (dispersant) and acetic anhydride (derivatising reagent) directly into an aliquot of beverage samples or into an aqueous extract of powdered milk samples obtained after a pretreatment of the samples. Given the compatibility of the solvents used, and the low volumes involved, the procedure was easily associated with GC-MS end-point determination, which was accomplished by means of an accurate GC dual column (heart-cutting) technique. Careful optimisation of heart-cutting GC-MS conditions, namely pressure of front and auxiliary inlets, have resulted in a good analytical performance. The linearity of the matrix-matched calibration curves was acceptable, with coefficients of determination (r2) always higher than 0.99. Average recoveries of the BPA and BPB spiked at two concentration levels into beverages and powdered infant formula ranged from 68% to 114% and the relative standard deviation (RSD) was <15%. The limits of detection (LOD) in canned beverages were 5.0 and 2.0 ng l–1 for BPA and BPB, respectively, whereas LOD in powdered infant formula were 60.0 and 30.0 ng l–1, respectively. The limits of quantification (LOQ) in canned beverages were 10.0 and 7.0 ng l–1 for BPA and BPB, respectively, whereas LOQ in powdered infant formula were 200.0 and 100.0 ng l–1, respectively. BPA was detected in 21 of 30 canned beverages (ranging from 0.03 to 4.70 µg l–1) and in two of seven powdered infant formula samples (0.23 and 0.40 µg l–1) collected in Portugal. BPB was only detected in canned beverages being positive in 15 of 30 samples analysed (ranging from 0.06 to 0.17 µg l–1). This is the first report about the presence of BPA and BPB in canned beverages and powdered infant formula in the Portuguese market.