Multivariate quantification of olive oil blended with sunflower oil by portable device SORS

Abstract

This study evaluates the ability of a portable analyser based on Spatially Offset Raman Spectroscopy (SORS) to quantify blends of extra virgin olive oil (EVOO), virgin olive oil (VOO) and sunflower oil (SFO) using multivariate regression analysis. Firstly, normalised spatially resolved (NSR) Raman fingerprints of pure EVOO, VOO, and EVOO/SFO and VOO/SFO binary mixtures were obtained. Multivariate regression models (PLS-R) were then developed from these instrumental fingerprints that could adequately predict the percentage of olive oil in the binary mixtures. The results showed that the portable SORS combined with chemometric tools can efficiently detect the presence of adulterant oil, even at concentrations below 2 %, with a prediction error (RMSEP) 4.96 %, a bias close to 0 and suitable SEP (∼5 %) for quality control applications of virgin olive oil. This non-destructive, sustainable approach combines Raman-SORS technique and instrumental fingerprinting methodologies to offer a versatile, reliable tool for monitoring and authenticating EVOO and VOO against SFO adulteration. Furthermore, it shows that it is possible to make direct measurements through the EVOO bottle, as there are no significant differences in the SORS spectra obtained from measurements made through the instrumental vial or the original sales bottle. The bottle under study was supplied by a reputable olive oil packaging company, which provided us with the relevant technical specifications. This methodology enables quality control to be conducted at various stages of distribution and sale, in order to detect adulteration of these oils with SFO, as they can be monitored both in bulk and in bottles.