Development of Facile Quantitative Evaluation Method for Citrus Peel Aroma Using Electronic Nose

Abstract

The citrus aroma, which affects the consumer's preference, is an important parameter for breeding. However, the quantitative evaluation of citrus aroma was difficult because the aroma accumulates in the oil glands of the peel. In this study, we improved the aroma evaluation method using a portable electronic nose and the odor-emitting device (OED) that we had invented previously. By standing after shaking the sample and OED, repeatability was improved, and it enabled the quantitative evaluation and comparison of aromas emitted from different citrus cultivars peel. Furthermore, nine parameters calculated from four sequential data were used for PCA and k-means clustering, and they show the classification of citrus aromas. As a result, Lisbon Lemon and Rinoka were clustered in Cluster 1; Encore, Yoshida Ponkan, and Sweet Orange were clustered in Cluster 2; and Ueno Wase and Miyagawa Wase were clustered in Cluster 3. Based on the result of clustering, decision tree analysis was conducted, and the factors contributing to the clustering of the aromas were shown. In this study, we have developed aroma evaluation methods consisting of e-nose, OED, and an algorithm optimized for aroma classification of citrus peels, and this result would contribute to accelerating breeding of valuable citrus cultivars.