Vibrational Spectroscopy and Chemometrics Applied to the Forensic Analysis of Automotive Paints and Edible Oils

Abstract

Profiling of complex materials (e.g., automotive paint and cooking oil) with infrared and Raman spectroscopy is an active area of research with a large and growing literature. The object of profile analysis is to correlate a characteristic fingerprint pattern in a spectrum with the properties of the sample. Objective analysis of these profiles depends upon the use of multivariate curve statistical methods. In this regard, pattern recognition techniques have been found to be of enormous utility. In this dissertation, several projects were undertaken to demonstrate the advantages of chemical fingerprinting using spectroscopic techniques to solve problems in the areas of food chemistry and forensic science. In one study, Raman spectra of 15 varieties of edible oils obtained from 53 samples purchased over a 3 year period representing different production years and vendors were analyzed by pattern recognition methods using a hierarchical classification procedure. Supplier to supplier variability and seasonal variability within a supplier were the major sources of variation with the Raman spectral data. Edible oils assigned to one group could be readily differentiated from those assigned to other groups, whereas Raman spectra within the same group more closely resemble each other and therefore were more difficult to classify by type. In another study, IR microscopic imaging and a prototype pattern recognition library search system were applied to the forensic examination of automotive paint using a new methodology for cross sectioning paint samples and decatenating infrared spectral images. Successful methods developed in test experiments such as the studies described in this dissertation will become part of the routine analytical practices of chemists in the very near future.