A novel chemometric strategy for the estimation of extra virgin olive oil adulteration with edible oils

A useful procedure for the qualitative and quantitative determination of vegetable oils (canola, hazelnut, pomace and high linoleic/oleic sunflower) as adulterants in commercial samples of extra virgin olive oil, has been developed. Partial least squares (PLS) was employed for the analysis of Fourier transform infrared spectroscopy (FTIR) spectral data of the blend oil samples. Calibration models were constructed for extra virgin olive oil purity, with wavelength selection in the infrared region, according to their predictive ability, with first derivative and mean centering used as data pretreatment. PLS models were internally validated by the leave-one-out procedure. The method developed was very suitable for the determination of modeled adulterants but it may also reveal an adulteration even if it does not derive from the adulterants employed in this study. (C) 2009 Elsevier Ltd. All rights reserved.