Discriminant analysis of edible oils and fats by FTIR, FT-NIR and FT-Raman spectroscopy

The goals of this study were to use Fourier transform mid-infrared (FTIR), near-infrared (FT-NIR) and Raman (FT-Raman) spectroscopy for discrimination among 10 different edible oils and fats, and to compare the performance of these spectroscopic methods for edible oil/fat study. The spectral features of edible oils and fats were studied and the unsaturation bond (CC) in IR and Raman spectra was identified and used for discriminant analysis. Linear discriminant analysis (LDA) and canonical variate analysis (CVA) were used for the discrimination and classification of different edible oils and fats based on spectral data. FTIR spectroscopy was found to be the most efficient in classification of oils and fats when used with CVA and yielded about 98% classification accuracy, followed by FT-Raman (94%) and FT-NIR (93%) methods; however, the number of factors were much higher for FT-Raman and FT-NIR methods. Overall, results demonstrated that FTIR, FT-NIR and FT-Raman techniques can be used to rapidly and simply determine the authenticity of edible oils and fats with chemometric analysis.