The chromatographic similarity profile –An innovative methodology to detect fraudulent blends of virgin olive oils

Abstract

Extra virgin olive oil is a potentially vulnerable foodstuffthat can be mixed with other vegetal edible oils including poorer quality olive oils in order to obtain illicit profits. These unauthorized operations may take place at any stage of the production process and radically affect the chemical composition. In this paper, the analysis of different vir gin olive oil sam ples before and after blending with other lower-grade olive oils in different proportions were performed. The direct analysis of the samples by (NP)HPLC-DAD in a wavelength range between 190 and 700 nm allowed the simultaneous analysis of several compound families responsible of the colour including chlorophylls, pheophytins, carotenes and tocopherols, the first three responsible for the olive oil colour. Unsupervised pattern recognition techniques applied on the chromatography-agnostic fingerprints of unblended virgin olive oil samples clearly showed the occurrence of groupings according to the sample hue (green and yellow). Two strategies, based on revealing changes in the spectrum-chromatographic fingerprints, are tested in order to detect the occurrence of such fraudulent blends: two-input class clas- sification methods (SIMCA) and similarity analysis. The SIMCA strategy was effective only for detecting blends carried out on virgin olive oils with a greenish hue (high chlorophyll/pheophytin content). Further- more, the similarity profile, developed and applied for the first time in this study evidences the blending in all cases irrespective of the original olive oil hue.