Adulteration of edible oils by blending with less expensive oils is a problem that is of concern to government and regulatory officials as olive oil is the food most subject to adulteration. The successful classification of edible oils by variety is a crucial first-step in solving this analysis problem.  To date, efforts to demonstrate the efficacy of infrared and Raman spectroscopy to classify edible oils by type and detect adulterants have presented ‘best-case’ scenarios with oils from a single batch. While edible oils from any particular batch within a class are well clustered and can be differentiated from other varieties of edible oils from a single source, incorporating uncontrolled variability from sources and seasons presents a more difficult classification problem. The between-source and seasonal variability within a class of edible oils may be comparable to differences between the average spectra of the different varieties of edible oils, thereby preventing a reliable classification of the edible oils and the detection of adulterants if a single model, spanning all sources and years of oils is to be constructed.  Using Raman and infrared spectroscopy, we have demonstrated that edible oils can be partitioned into 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 can be readily differentiated from those assigned to other groups, whereas Raman and IR spectra within the same group more closely resemble each other and therefore are more difficult to classify by variety. Using a hierarchical approach to the classification of edible oils, we will show that variable selection implemented using a genetic algorithm can address a wide range of problems in the detection of adulterants in edible oils.  For example, variable selection can make a discriminant model more robust to unexpected interferences or aid in the transfer of calibrations across instruments.   <br/>

Authentication of edible oils using vibrational spectroscopy and pattern recognition techniques

2019 Fall National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.