Chemometric interpretation of the results of chromato-mass spectrometry analysis of mineral and synthetic oils for customs purposes

Abstract

Application of the principal component method (PCA) using the example of distinguishing mineral, synthetic and semi-synthetic motor oils from the results of chromatography-mass spectrometric analysis is considered in the present work. Samples of commercial oils of domestic and foreign production constituted the objects investigated. A generalized algorithm for constructing a classification model, including the stages of preliminary processing of spectral information, as well as conditions for setting up the experiment, is proposed. The principal component analysis (PCA) and the principal component regression (PCR) methods were applied to evaluate the results obtained. Chromatograms of the studied samples acted as the primary data for chemometric processing. Analysis of the graphs of scores in the model constructed by PCA for the first and second principal components revealed that the points corresponding to different groups of objects did not overlap and were located in different areas in relation to the first and second principal components. The loadings plot reflects the characteristic retention times, which are responsible for separating studied samples into mineral and synthetic oils. The PCR method was chosen for modeling chromatography-mass spectrometry data taking into account the task of quantitative identification of mixed compositions of mineral and synthetic oils. Based on the results obtained it was found that chromatography-mass spectrometry in combination with chemometric methods of analyzing multifactorial dependencies can be applied for solving problems, related to determining qualitative and quantitative composition of motor oils. The obtained models demonstrate sufficient accuracy for qualitative and quantitative determination of the composition of mixed motor oils based on mineral and synthetic base components.