Development of Multisensory Gas Chromatography Experimental Apparatus for Scent Analysis

In this presentation we discuss the development of an experimental multisensory gas chromatography (MGC) apparatus in order to analyze chemicals to provide a verbal description of the scent. Scents are categorized into families, determined by common verbal descriptions. For example, notes of lemon and orange correspond to a citrus scent, in the citrus family, whereas notes of cedar and oak correspond to a woody scent, in the woody family. In ``The Atlas of Odor Character Profiles'' (1985), Andrew Dravnieks released a collection of the applicability of a number of verbal descriptors for various chemicals. The scent descriptor applicability dictates each chemical's family placement. Analyzing chemicals with known verbal descriptors using an MGC-based apparatus quantifies their scents providing a chromatogram for each test odor. A Convolutional Neural Network is used to build corresponding function between a set of verbal descriptors and a set of collected chromatograms. This project is funded by the KY NSF EPSCoR URE program.