Analysis of Edible Oils Using Fourier Transform Infrared Spectroscopy and Chemometrics

Abstract

The application of infrared spectroscopy and pattern recognition to the problem of discriminating edible oils by type is the subject of this research project. Previously published studies on this topic were generally limited to a small number of samples spanning five or six varieties of edible oils obtained from a single manufacturer over a limited production year range. In this study, infrared spectra obtained from 95 samples spanning 20 edible oils collected from supermarkets in the greater Newark, DE metropolitan area over a three year period were investigated using the three major types of pattern recognition methodology: mapping and display, clustering, and discriminant development. The edible oils could be partitioned into four distinct groups based on their degree of saturation and the ratio of polyunsaturated fatty acids to monounsaturated fatty acids. Edible oils assigned to one group could be readily differentiated from those assigned to other groups, whereas infrared spectra within the same group more closely resembled each other and therefore were difficult to accurately classify by type. These analyses highlight the need for adulteration studies that span multiple sources for each type of edible oil. The supplier to supplier variation for edible oils (and possibly the seasonal variation within a supplier) is greater than within supplier variation. This work also demonstrates that previous studies (which rely on only one source for each type of edible oil) provide an overly optimistic estimate of the ability to classify edible oils or to detect low levels of adulterants by infrared or Raman spectroscopy.