The Study On Determining Types And Qualities Of Tea With GC-MS Combining Chemometrics

Tea aroma is an important factor to determine the quality of tea. In recent years, a great deal of research is being carried out for the identification of aroma substances in different tea species, origins and varieties, even different grades of tea with the development of GC-MS technology. Tea aroma has characteristcs in only a small quantity and many kinds of scent, which brought great challenges to analysis and identificationof fragrance.In this study, one-factor test was involved to establish the optimal conditions for tea aroma extraction with headspace solid phase micro extraction about Green tea of Korea, Anxi Tieguanyin and Jasmin tea. The types and relative contents of tea aroma under different conditions including extraction time, extraction temperature and the amount of tea were analysed by GC-MS. The results showed that the optimal conditions for aroma extraction were as following:Green tea of Korea:50℃-70℃,70min,3.0g-5.0g; Anxi Tieguanyin:60℃,50min-60min,3.0g; Jasmin tea:40℃-60℃,50min,3.0g; Water consumption were30mL.Further tea aroma discriminant analysis for the different grades of tea combining sensory evaluation were carried out to predict the feasibility of different levels of tea quality with different aroma through partial least squares regression. Results showed that as below:(1) Discriminant analysis combined principal component was satisfactory well to distinguish among the species of tea leaves, got the discriminant functions about three kinds of tea leaves, they fitted three categories. Green tea:Y1(X)=-7.138-4.094X,-7.876X2; Jasmin tea: Y2(X)=-7.291-2.059X,+8.157X2; Black tea:Y3(X)=-3.060+4.022X1+0.746X2. X, stands for the first principal component, X2stands for the second principal component, formulating discriminant function through cross-validation reached correct rate100%, as discriminating factors also achieved100%correct rate.(2) PLSR analysis was proceeded in different grades of tea leaves and the signal of GC-MS, adopting the signal of GC-MS predicts at the grades or prices of Dafo Longjing, Jasmin tea, Dianhong and Qihong, the correlation of fitted value and actual value is followed by0.9056,0.8855,0.9527,0.9710, except the correlation of Jasmin tea was slightly lower, the rest of them are all above0.9, the results show that using PLSR analysis combining GC-MS signal was good effect to predict quality the different grades of tea leaves.