Characterization of flavor profiles in chive leaves and stems at different moisture transfer points during combined drying: An integrated approach with e-nose, GC-IMS, GC-MS, and machine learning.

This study investigated the effect of moisture transfer points on volatile organic compounds (VOCs) in chive leaves and stems during the vacuum freeze drying combined with hot air drying process. The VOCs of chive leaves and stems were comprehensively analyzed using electronic nose, HS-GC-IMS, and HS-SPME-GC-MS techniques, with 8 key VOCs identified through HS-SPME-GC-MS combined with the random forest algorithm and odor activity values. Multivariate analysis (principal component analysis and orthogonal projections to latent structures discriminant analysis) validated a clear distinction between high (L1/L2, S1/S2) and low (L3/L4, S3/S4) moisture transfer point groups. Additionally, potential metabolic pathway analysis indicated that dipropyl disulfide and dimethyl disulfide originate from the enzymatic reaction of S-alkylcysteine sulfoxides; terpenoids are associated with carotenoid cleavage; and aldehydes and ketones are closely related to amino acid metabolism and sugar cleavage. This study provides theoretical support for precision processing and flavor enhancement in chives.