Potencial da espectrometria no infravermelho médio para análise de lactose em leite

Abstract

Considering that the lactose free dairy sector has been growing considerably and that the legislation for these products is becoming more and more rigid, there is a need to develop more efficient analytical methods capable not only of identifying lactose but also of quantifying it, quickly and accurately. The present work aimed to develop methodology based on the use of infrared spectrometer associated with chemometric methods to detect and predict the lactose content present in milk. In the first it was demonstrated that the possibility of discriminating milk samples with lactose from those without lactose using the medium infrared spectra and Principal Component Analysis (PCA). In the second part, samples were prepared from milk mixtures with and without lactose and analyzed in the medium infrared spectrophotometer. The absorbances in the 935 to 1200 cm-1 spectral range, which concentrates information on functional groups present in the carbohydrates, were used as descriptors to evaluate the efficiencies of the models obtained by Regression by Partial Least Squares (PLS), Multiple Linear Regression (MLR) and Least Squares Regression (MQ). The three methods proved to be efficient in this determination, with the smallest errors being obtained by PLS. In the third part, these models were used to monitor the lactose content during a lactose hydrolysis process using enzyme lactase, demonstrating the applicability and efficiency of the methods in the monitoring of this process. Thus, it was concluded that the Medium Infrared Spectrometry associated with Chemometrics methods constitutes a rapid method that requires sample preparation capable of detecting and determining the lactose content in bovine milk.