Desarrollo de metodologías analíticas basadas en la utilización de datos de tres y cuatro vías combinadas con calibración multivariante, para la evaluación de la calidad de alimentos de origen vegetal

Abstract

The present Thesis deepens about the development of new analytical methods using molecular fluorescence in combination with multivariate calibration chemometric algorithms for the determination and identification of main polyphenolic compounds in much appreciated foods. Four food matrices from the economy of Extremadura were selected in this Thesis: grapes, wines, plums and olives. A few references about phenolic compounds quantification fluorescence methods in these food matrices without pretreatments are presented in the bibliography. For this, front-face fluorescence methods in combination with multivariate calibration algorithms, such as PARAFAC and U-PLS, have been developed for wine polyphenolic compounds quantification without pretreatments in addition to aqueous-organic extracts of grapes, plums and olives. The obtained results have been validated with those obtained by liquid chromatography, obtaining high correspondences of resveratrol in grapes and wines, catechins in grapes, wines, plums and olives and chlorogenic acids and procyanidin B2 in plums. The obtained information has been analysed by classification chemometric algorithms such as LDA, LDA-PARAFAC and U-PLS-DA and its application has allowed the discrimination between grapes, plums and olives harvested in different ripening stages. These algorithms have also been used for irrigated and non-irrigated olives discrimination, using two solvent four-way data in case of irrigated and non-irrigated grapes. The methodologies developed have certain noticeable improvements in comparison with the existent ones.