Characterization of Differences in Flavor in Virgin Rapeseed Oils by Using Gas Chromatography–Mass Spectrometry, Electronic Nose, and Sensory Analysis

Abstract Flavor profiles of virgin rapeseed oils (VROs) are comparatively studied for the discrimination and authentication of various flavor types by using gas chromatography–mass spectrometry (GC‐MS), electronic nose (E‐nose), and sensory analysis. The relative odor activity values (ROAVs) of key odorants detected by GC‐MS and E‐nose data are visualized to discriminate VROs with various flavors by using multivariate analysis. The correlation of sensory data, ROAVs of key odorants, and E‐nose data is analyzed. Results show that 47 key odorants comprising aldehydes, isothiocyanates, nitriles, pyrazines, ketones, acids, and alcohols are identified based on their ROAVs. The VROs are divided into three groups and the flavor type of each group is determined, namely, green‐pungent, nutty, and oxidized oil flavor by the three techniques. The above flavor types are mainly attributed to the contributions of isothiocyanates, pyrazines, and aldehydes, respectively, according to the compounds' ROAVs. The correlation of the three techniques is investigated by partial least squares regression: GC‐MS data are confirmed to have good correlation with sensory data. Practical Applications : VROs are characterized by their intense flavor and consequently receiving growing interest among consumers. In this study, different flavor profiles of VROs were comparatively studied using GC‐MS, E‐nose, and sensory analysis. It is found that 47 volatile compounds including aldehydes, isothiocyanates, nitriles, pyrazines, ketones, acids, and alcohols were the key odorants to VRO flavors. The VROs were divided into three types by the three techniques, that is, green‐pungent, nutty, and oxidized oil flavor. The three flavor types were attributed to the contributions of isothiocyanates, pyrazines, and aldehydes, respectively, based on the compounds' ROAVs and GC‐MS was confirmed to have good correlation with sensory data. Although E‐nose cannot correlate all flavor types in the oils, this technique is rapid and low‐cost for VRO discrimination according to various aroma characteristics. GC‐MS and E‐nose can be regarded as potential techniques for flavor discrimination in VROs.