基于HS-GC-IMS技术对不同程度霉变烟叶的判别  

Discrimination of Different Degrees of Mildewed Tobacco Based on HS-GC-IMS Technique

作　　者：刘显军 张亮 孙海峰 张振 何澎 杨鹏飞 吴薇 尚紫博 LIU Xian-jun;ZHANG Liang;SUN Hai-feng;ZHANG Zhen;HE Peng;YANG Peng-fei;WU Wei;SHANG Zi-bo(Technology R&D Center,Shenzhen Tobacco Industrial Co.,Ltd.,Shenzhen 518000,China;College of Tobacco Science and Engineering,Zhengzhou University of Light Industry,Zhengzhou 450000,China)

机构地区：[1]深圳烟草工业有限责任公司技术研发中心,广东深圳518000 [2]郑州轻工业大学烟草科学与工程学院,河南郑州450000

出　　处：《分析测试学报》2024年第9期1433-1441,共9页Journal of Instrumental Analysis

基　　金：中国烟草总公司重大科技项目(110202201053(SJ-03));中国烟草实业发展中心科技项目(ZYSYQ-2022-17)。

摘　　要：采用顶空-气相色谱-离子迁移谱(HS-GC-IMS)技术测定广东韶关南雄不同程度霉变烟叶中的挥发性化合物,并进行化学模式识别及相对气味活度值(ROAV)分析,构建了基于差异挥发性成分的分类判别模型。结果表明:HS-GC-IMS差异图、挥发性化合物指纹图谱及相似度相关性热图显示不同程度霉变烟叶中的挥发性化合物存在差异;不同烟叶样品共定性出50种挥发性化合物,其中醇类13种,醛类6种,酮类8种,酯类13种,酸类1种,萜烯类1种,其他类8种;主成分分析(PCA)、层次聚类分析(HCA)和偏最小二乘-判别分析(PLS-DA)均可有效区分不同程度的霉变烟叶,以变量重要性投影值(VIP)>1为依据筛选出20种差异挥发性化合物;进一步通过ROAV分析选取了17个差异挥发性化合物用于构建支持向量机(SVM)分类判别模型,所建模型的分类判别率为100%。该方法能有效区分不同程度霉变烟叶,具有高效、准确等优点,可为烟叶霉变的识别及预测提供技术和数据支撑。To explore the differences of volatile compounds in different degrees of mildewed tobacco and achieve rapid discrimination,headspace-gas chromatography-ion mobility spectrometry(HSGC-IMS)was employed to analyze the volatile compounds in different mildewed tobacco from Nanx⁃iong,Shaoguan,Guangdong,and a classification model was constructed based on distinctive vola⁃tile components which was screened by the chemometric pattern recognition and relative odor activity value(ROAV)analysis.The results revealed discernible patterns of volatile compounds among tobac⁃co with varying mold severity by the HS-GC-IMS differential maps,volatile compound fingerprints,and similarity correlation heatmaps.A total of 50 volatile compounds were qualitatively identified in different leaf samples,including 13 alcohols,6 aldehydes,8 ketones,13 esters,1 acid,1 ter⁃pene,and 8 other compounds.Principal component analysis(PCA),hierarchical clustering(HCA),and partial least squares discriminant analysis(PLS-DA)effectively differentiated tobacco samples with different mold severity.20 distinct volatile compounds were screened out based on variable im⁃portance projection(VIP)values greater than 1.Furthermore,17 differential volatile compounds were chosed,using ROAV analysis,to construct a support vector machine(SVM)classification mod⁃el,achieving a 100%classification rate.The method effectively discriminated tobacco with varying mold severity,providing technical and data support for moldew identification and prediction.

关 键 词：烟叶 霉变 顶空-气相色谱-离子迁移谱 挥发性化合物 化学模式识别 相对气味活度值 

分 类 号：O657.7[理学—分析化学] TS45[理学—化学]