Espectrometria de massas com ionização electrospray (ESI-MS) e métodos quimiométricos: caracterização de azeite de oliva (extra virgem e puro) e outros óleos vegetais e quantificação de óleos adulterantes em azeite de oliva extra virgem

Abstract

In this work, electrospray ionization mass spectrometry (ESI-MS), withchemometric approaches, were used to typify olive oils (extra virgin and pure) and other vegetable oils (soybean, corn, canola, sunflower). Furthermore, the detection and determination of the adulterant content in counterfeit samples of extra virgin olive oil, adulterated by the addition of such vegetable oils (including pure olive oil), were also performed.Hence, exploratory analysis techniques (such as principal component analysis, PCA, and hierarchical clustering analysis, HCA), applied to set of the ESI(-)-MS and ESI(+)-MS (mass spectra in the negative and positive ion modes, respectively) data allowed an unequivocal distinction between such vegetable and olive oils. Furthermore, the differentiation between extra virgin and pure olive oils was also successfully achieved after treatment of the ESI(+)-MS data by PCA and HCA approaches. Extra virgin olive oils were adulterated by addition of each vegetable oils (soybean, corn, canola, sunflower) or pure olive oil, in proportions varying from 1 to20% w/w with increments of 1% w/w (hence, a total of 20 samples were prepared for each adulterant oil). The ESI(-)-MS and ESI(+)-MS of the counterfeit and authentic extra virgin olive oil samples were treated by the PCA and HCA methodologies and an unambiguous fraud detection (even at levels as low as 1% w/w) was achieved. Finally, one sample of a commercial extra virgin olive oil was adulterated with increasing proportions of a given soybean oil (from 0.5% m/m to 20% m/m withincrements of 0.5% m/m, a total of 40 samples). The application of multivariate calibration (via the partial least square regression, PLS) to the ESI(-)-MS e ESI(+)-MS data allowed the construction of models that presented excellent performance, with RMSEC (root mean square error of calibration) and RMSEP (root mean square error of prediction) values lower than the ones described in similar works reported in literature.