Classification of edible oils using synchronous scanning fluorescence spectroscopy

Total luminescence and synchronous scanning fluorescence spectroscopy techniques were tested as regards their ability to characterize and differentiate edible oils, including soybean, sunflower, rapeseed, peanut, olive, grapeseed, linseed and corn oils. Total luminescence spectra of all oils studied as n-hexane solutions exhibit an intense peak, which appears at 290 nm in excitation and 320 nm in emission, attributed to tocopherols. Some of the oils exhibit a second long-wavelength peak, appearing at 405 nm in excitation and 670 nm in emission, belonging to pigments of the chlorophyll group. Additional bands were present in the intermediate range of excitation and emission wavelengths in some oils, arising from unidentified compounds. Similarly, bands attributed to tocopherols, chlorophylls and unidentified fluorescent components were detected in the synchronous-scanning fluorescence spectra. Classification of oils based on their synchronous fluorescence spectra was performed using a non-parametrical k nearest neighbours method and linear discriminant analysis. Both methods provided very good discrimination between the oil classes with low classification error. The results presented demonstrate the capability of the fluorescence techniques for characterizing and differentiating vegetable oils. (C) 2004 Elsevier Ltd. All rights reserved.