FTIR fingerprints associated to a PLS-DA model for rapid detection of smuggled non-compliant diesel marketed in Morocco

Abstract Fourier Transform Infrared Spectroscopy was selected as a rapid, non-destructive, cheap and a high sensitive technique combined with chemometric tools in order to detect and discriminate smuggled diesel samples from authentic products. A total of nineteen authentic diesels and eleven smuggled diesel samples were collected, their quality was checked by determining their chemical composition using the GC MS technique in order to avoid the use of adulterated products. Secondly, infrared fingerprints of all samples were recorded, and the resulting spectra processed by Principal Components Analysis (PCA) to visualize classes and to detect outliers. Partial Least Square Discriminate Analysis (PLS-DA) was carried out to classify samples in two classes (authentic or smuggled). The PCA results were able to indicate the dissimilarities between the two diesel classes, and the PCA scores-plot observed two groups. On the other hand, spectral data associated with PLS-DA models allowed a better classification between the two classes of diesel with a high specificity and sensitivity.