In the field of olive oil, quality and authenticity are among the criteria of consumer purchase. Indeed, consumers are increasingly looking for food products of good and healthy quality where origin must be certified. In this context, the purpose of the present thesis project was to develop new analytical and methodological approaches to determine the quality and authenticity of virgin olive oils (VOO). To better understand the concept of quality, theoretical work on the olive oil sector, the olive oil extraction process, biochemical and chemical reactions taking place during the olive oil extraction, storage, as well as the methods of analysis applied to determine the quality and authenticity of olive oil were performed.
The 1st part of this thesis was devoted to study the potentiality of classical and spectral methods to authenticate 41 VOO samples collected during two crop seasons (2015/2016 and 2016/2017) according to their geographic origin (Beni-Mellal/Khenifra, Fès/Meknès, Marrakech/Safi, Northern and Eastern regions) and variety (Arbéquine, Arbozana, Moroccan Picholine, and Languedoc Picholine). The chemical analyses (free acidity, peroxide value, k232, k270, and the chlorophyll level) did not allow the authentication of VOO samples according to their geographic origins and varieties. The fluorescence and mid infrared (MIR) spectra acquired on the VOO samples allowed to discriminate VOO samples according to their geographic origins (96.72 and 91.87%, respectively) and varieties (95.12 and 91.87%, respectively). This trend was confirmed following the application of partial least square regression (PLSR) to the fluorescence spectra since an excellent prediction of free acidity (R2 = 0.98), and peroxide value (R2 = 0.96), and good prediction of k232 (R2 = 0.88), k270 (R2 = 0.88), and chlorophyll level (R2 = 0.89) were observed, suggesting that fluorescence and MIR spectra could be considered as a fingerprint of olive oil to evaluate their quality and authenticity. The 2nd part of the thesis aimed to determine the potentiality of front-face fluorescence (FFFS) and MIR spectroscopies to monitor the aging and to predict the chemical parameters of 14 VOO samples collected during 2015/2016 crop season. The results showed clear discrimination of VOO samples according to their storage time since 96.67% of correct classification was obtained for the 5 groups aged 7 days, and 3, 6, 12, and 18 months, respectively from the fluorescence and infrared spectra. These results were confirmed by the excellent prediction of the storage time following the application of the PLSR, support vector machine regression (SVMR), principal components regression (PCR) and multiple linear regression (MLR), to the emission spectra acquired after excitation set at 430, 290 and 270 nm since R2 = 0.98, ratio of prediction deviation (RPD) = 7.9, and root mean square error of prediction (RMSEP) = 24 days was noted. Similar results were obtained concerning the prediction of chemical properties of VOO since the validation models allowed to obtain R² ranging between 0.98 and 0.99 for free acidity, peroxide value, k232, k270, and the chlorophyll level. Finally, in the 3rd part of this work, spectral methods demonstrated their ability to detect the adulteration of Moroccan extra VOO with other types of olive oils of low quality (VOO, ordinary VOO, lampante VOO, pomace, and refined olive oils) at different levels varying between 5 and 50%. Indeed, an excellent prediction of the adulteration level was obtained since R2 = 0.99 and RMSEP = 1.28% were obtained by applying PLSR and PCR on the fluorescence spectra. Similar results were obtained by using MIR spectroscopy. The results obtained from this thesis demonstrated the potential of spectral methods combined with chemometric tools to determine the quality and authenticity of VOO, which can build a solid basis for the development of rapid and effective methods.