Detecção de fraudes em néctares de frutas: análises espectroscópicas aliadas a métodos de calibração eclassificação multivariada

Due to the increasing advancement in the consumption of fruit drinks, fruit nectars stand out because their final price is much lower than pasteurized whole juices. Orange nectar is the most consumed worldwide, and the most consumed nectars in Brazil are grape, peach, orange and passion fruit, in this order. There are many known types of adulteration in juices and nectars that occur during the manufacturing process and some of them are frequent, such as: dilution with water, addition of sugar syrup and undeclared addition of different species. Spectroscopic techniques can be used to evaluate the structure and composition of food and beverages. The combination of these techniques with chemometric methods is widely used to determine authenticity in food. The main objectives of this work were: a) analysis andcomparison of the spectral profile of Gala and Fuji apple varieties involved in the production of nectars in Brazil; b) characterization of commercial nectars (grape, peach, orange and passion fruit) in relation to their Identity and Quality Standards; c) development and validation of multivariate calibration models for quantification of the main fruit content in nectars (grape, peach, orange and passion fruit) unadulterated and adulterated with syrup, cashew juice and apple juice; d) development and validation of supervised classification models for discrimination and classification of cashew and apple adulterants, isolated, in grape nectars using attenuated total reflectance Fourier-transform infrared (ATR-FTIR) spectroscopy; e) development and validation of supervised classification models for discrimination and classification ofcashew and apple adulterants, as a mixture, in grape nectars using ATR-FTIR; f) development and validation of supervised classification models for discrimination and classification of cashew and apple adulterants, isolated and as a mixture, in grape nectars using time-domain nuclear magnetic resonance (TD-NMR). The methods and models developed and validated were adequate for their purposes of use, with potential contribution for fraud detection in fruit nectars.