Record of mint flavor using mass spectrometry

Abstract

Although GC/MS is a standard technique to analyze gases, a combination of mass spectrometry with a pattern recognition technique enables us to evaluate smells rapidly because lengthy measurement of gas chromatography is eliminated. Although several researchers have reported the classification of smells only using mass spectrometry, the estimation of the sensory-test result has not been studied. A neural network enables us to map the mass spectrum into the human sensory space. In this study, mint flavors were focused on. Four spearmint flavors and their mixtures were successfully separated and quantified only using the mass spectrum. Then, the mixtures of those flavors and peppermint flavor were evaluated. The scores of the sensory test such as QDA (quantitative descriptive analysis) method were estimated using the mass spectrometry and SOM (self-organized map). It was found that the accuracy of the estimated value using the SOM was better than that obtained according to the linear superposition theorem