| Quality identification and classification for tobacco was great significance in keeping stability of blend threshing and building of tobacco brand. Taking different levels and different modules of flue-cured tobacco (FCT) from12tobacco production areas of Yunnan during2009to2010as the research materials, this study established analytical procedure of GC-MS fingerprints of FCT and its database based on the analysis of the aroma components. The main objectives were (1) to illuminate the aroma components of main FCT from Yunnan;(2) to establish the GC-MS fingerprint of tobacco’s aroma components;(3) to apply GC-MS fingerprint in quality classification of FCT;(4)to identify the structures of the peaks on the HPLC fingerprint spectrum of FCT from different production areas of Yunnan. The main results were as follows:1. Some volatile and semi volatile aroma components in FCT leaves from Yunan were identified by using simultaneous distillation and extraction (SDE) method and GC-MS analysis. The results showed that alcohols, aldehydes and ketones were the main components, followed by acids, phenols and esters. Among them, neophytadiene was the highest one, with its Area%of reaching up to50%on average. In addition, those with the Area%of greater than0.5, included Furfural, Nicotine, Solanone, Damascenone, Megastigmatrienone, Farnesylacetone and Duvatriendiol.2. The standard procedure for fingerprint analysis of tobacco aroma components was established by the technology of GC-MS with dichloromethane as extractant and phenylethyl propionate as internal standard. The precision analysis showed that RSD of relative retention time(RT) of peaks was less than0.1%, and RSD of relative peak area for the biggest peak was6.68%and the rest peaks’ were less than5%respectively. Analysis of the procedure stability showed that the RT of internal standard was between25.52min and25.55min, and its Area%was between10.59and12.03. The RSD of relative RT was less than0.09%.The RSD of relative peak area for the biggest peak was9.13%and the rest peaks’ were less than5.3%respectively. The results for repeatability analysis showed that relative RT’s RSD of each peak was less than0.09%, the angle cosine similarity of all samples were greater than0.9992.3. The common fingerprint peaks database of typical Yunnan FCT were established. Common fingerprint peaks of C3F,B2F,X2F level FCT were55,52and57repectively, for sampling in2009, and were53,53and60respectively, for sampling in2010. In the same time, time-sharing database of fingerprint by a time-sharing software was also constructed, with21time-sharing peaks, i.e.,3min~6min,6min~9min...63min~66min respectively.4. Classification of FCT samples were carried out by principal component analysis and cluster analysis. The results showed, C3F,B2F and X2F level FCT were classed into6,3and4groups for samples in2010, and4,4, and5groups for samples in2009. The blend threshing strips of FCT were divided into3groups.5. The chemical components of HPLC fingerprint peaks were identified by means of±ESI-MS/MS spectrometry experiment. The five characteristic peaks were determined as Neochlorogenic acid, Chlorogenic acid, Cryptochlorogenic acid, Rutin, and Kaempferol-3-O-rutinoside. Furthermore, ESI-MS fragmention pathways of the five components were proposed. |