NIR Spectroscopy and Partial Least-Squares Regression for Determination of Natural α-Tocopherol in Vegetable Oils

Abstract

Near-infrared (NIR) spectroscopy and partial least-square regression were used for determination of alpha-tocopherol in edible oils after extraction with ethanol. The standard error of calibration and the standard error of prediction were calculated for evaluation of the calibration models. The chemometric calibration model was prepared in spectral region 6500-4500 cm(-1) for standard alpha-tocopherol solutions (0.54-53.54 mg/mL). Obtained mean concentrations of natural alpha-tocopherol in different types of oils varied from 17.53 to 57.10 mg/100 g. Net analyte signal calculation was used to estimate detection limit (DL = 0.12 mg/mL), quantification limit (QL = 0.40 mg/mL), sensitivity (SEN = 0.045 mg/mL), and selectivity (SEL ranged between 0.24 and 0.54% of the measured reflectance signal) of the proposed NIR method. The comparable precision (RSD = 0.68-2.80% and 0.79-3.06%) and accuracy (recovery, 97.2-102.4% and 96.8-103.2%) for the proposed NIR and standard HPLC methods, demonstrate the benefit of the NIR method in the routine analysis of alpha-tocopherol in vegetable oils.