Independent component analysis and near infrared spectroscopy to evaluate thermal degradation of antioxidants from coconut oil

Methods that promote separation of mathematical signals without need for physical separation are considered as being part of the chemometrics research field. In this field, independent component analysis (ICA) perform the decomposition of signals of a mixture into more statistically independent components. The construction of independent components (ICs) are based on linear combinations of the original variables, and the ICs are assumed to correspond to the signals of the pure source present in the analyzed mixtures. ICA promotes the minimization of the statistical dependence of the signals, and solves the problems as the need of the presence of pure variables and local minimum instead of global minimum that can occur with multivariate curve resolution with alternating least squares (MCR-ALS). In this work, ICA was used to separate the antioxidant signal from coconut oil submitted to thermal degradation. The NIR spectra (900 – 1600nm) were acquired in a JDSU MicroNIR ® equipment. Then, using Matlab ®, the spectra was cut in order to use only the most informative part to this study (1100 – 1600nm). The ICA was performed with JADE (Joint Approximate Diagonalization of Eigenmatrices) algorithm after spectra first derivative to perform baseline correction. Coconut oil is reported to possess antioxidant action due to its phenolic composition. By applying ICA in the near infrared spectra of coconut oil heated from 23oC to 180oC is possible observe that the scores profile change and evolved accordingly with heating. The signal profile correspondent to the phenolic constituents was assigned due to the region of 1150 – 1210nm, which refers to characteristic absorption regions of second overtones of C-H stretching. Analyzing the correspondent scores for this signal, it was possible observe that coconut oil initially contained high quantities of antioxidants and these quantities linearly decrease due to thermal degradation. This result was validated by DPPH (2,2-diphenyl-1-picrylhydrazyl) radical scavenging method. Based on the data, near infrared spectroscopy coupled with independent component analysis is a feasible tool to monitor antioxidants in oils. Moreover, this is a low cost, fast and sample preparation free methodology to evaluate antioxidants in oils.