Identification and Detection Adulterated of Butter by Methods of Colorimetry and Near-IR-Spectroscopy

Abstract

An express and simple method for identification of oil and fatty products of plant origin by their own fluorescence and diffuse reflection of IR radiation using colorimetry and near-IR spectroscopy is proposed. To record the analytical signal, we used 3D printed devices with built-in UV and IR LED matrices (390 and 850 nm), and a smartphone with the PhotoMetrix PRO® application installed, as well as FT-IR spectroscopy in the near-IR region (10000–4000 cm^–1) with the NIRA attachment, used for the analysis of solid samples. Processing of diffuse reflectance spectra was carried out using the TQ Analyst and The Unscrambler X applications. Identification and differentiation of the studied objects was carried out using chemometric algorithms – principal component analysis (PCA) and hierarchical cluster analysis (HCA). The determination of the mass fraction of fat in the declared products was carried out using univariate and multivariate (PLS algorithm) analyzes. It has been established that on the PCA and HCA graphs, adulterated butter is located separately from natural products and does not intersect with each other on the dendrogram. To construct a calibration relationship and determine the milk fat concentration using the PLS method and one-dimensional analysis, we took samples of butter with different milk fat mass fractions: 61.5, 72.5, 82.5, and 99.0%. At the same time, the calibration error (RMSEC) did not exceed 1.31% and the predictive properties (RMSEP) – 4.45%. The methods under consideration were tested with samples of butter and vegetable oil products from different manufacturers. When multivariate analysis was used, the RMSEP values for dairy products did not exceed 4.97%, and for margarine it was more than 10%. When using univariate analysis, the relative deviation of the results from the values of the mass fraction of fat presented on the packaging did not exceed 4.8%. As a result of the study of margarine, this indicator was in the range of 96.3–96.5%. The data obtained were correlated with the results of Fourier transform infrared spectroscopy.