Controle de qualidade de azeite de oliva extra virgem e misturas diesel/biodiesel utilizando espectrometria de massas e validação multivariada

• Ano de defesa: 2014
• Autor(a) principal: Junia de Oliveira Alves
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Adulteração; Controle de qualidade; Azeite de Oliva extra virgem; PLS; Diesel b; ESI-MS; PLS-DA; EASI-MS; Validação Multivariada; Química analítica; Analise multivariada; Espectrometria de massa; Quimiometria; Azeite
• Link de acesso: http://hdl.handle.net/1843/SFSA-9RMGC8

Aiming at merging modern mass spectrometry techniques (electrospray ionization mass spectrometry - ESI-MS and easy ambient sonic-spray ionization mass spectrometry -EASI-MS) with chemometric methods (partial least squares -PLS and partial least squares discriminant analysis PLS-DA) the present work was developed forquality control of extra virgin olive oil and diesel b (blends diesel/biodiesel). The chapters were organized as following indicated:Chapter 4 was designed for quality control of extra virgin olive oil. A PLS2-DA model was developed basead on ESI-MS data, for classification of seven classes of olive oil (ordinary olive oil, extra virgin olive oil and adulterated with five adulterants oils). The best model was built with eight latent variables and showing good sensitivity (1.000) and specificity (0.967 1.000) values for the training and test sets. PLS models were also built, with seven models built with ESI-MS data, two models with data from a mass spectrometer for high resolution ESI-HRMS and a model constructed from data EASI (+)-MS. The 10 models were constructed for the quantification of adulterants oils( soybean, corn, sunflower and canola) in extra virgin olive oil. The models were validated by means of some figures of merit, was evaluated in models linearity, bias, accuracy, precision, selectivity, sensitivity and analytical sensitivity, limits of detection and quantification and Residual Prediction Deviation (RPD). Chapter 5 was intended for diesel b quality control ESI-MS data was used to construct a model for quantification of biodiesel in diesel. This model was also validated similarly as above mentioned. The proposed methods are promising because they are simple and fast. All models showed high efficiency and can be used in quality control of samples of extra virgin olive oil and biesel b.