近红外光谱法快速测定晒青毛茶三种儿茶素组分含量

Rapid Determination of Three Catechin Components in Crude Pu-erh Tea by Near Infrared Spectroscopy

摘要:

为了实现晒青毛茶儿茶素含量的快速检测，该研究利用高效液相色谱技术（High Performance Liquid Chromatography，HPLC）测定了50份晒青毛茶样品中的表没食子茶素没食子酸酯（Epigallocatechin Gallate，EGCG）、表没食子儿茶素（Epigallocatechin，EGC）、没食子酸（Gallic Acid，GA）三种儿茶素单体含量，结合样品的近红外光谱，分别建立了晒青毛茶三种儿茶素单体含量的偏最小二乘法（Partial Least Squares，PLS）模型，并对模型进行验证。实验结果表明，基于EGCG、EGC、GA含量所建立近红外模型的决定系数（Coefficient of Determination，R2）分别为99.99%、99.99%、99.92%；校正标准差（Root Mean Square Error of Calibration，RMSEC）分别为0.17、0.15、0.10；相对标准差（Relative Standard Deviation，RSD）分别为0.19%、0.27%、0.56%；外部验证的结果显示，三种儿茶素单体含量模型的预测值与真实值的平均绝对误差分别为0.13、0.12、0.07；平均相对误差分别为0.17、0.25、0.45。实验结果表明，该研究建立的近红外分析模型具有较高的预测准确度和稳定性，在快速检测晒青毛茶儿茶素方面具有潜在应用价值。

Abstract:

In order to achieve rapid detection of the catechin content in crude Pu-erh tea, high performance liquid chromatography (HPLC) was used to determine the contents of three catechin monomers, namely, epigallocatechin gallate (EGCG), epigallocatechin (EGC), and gallic acid (GA), in 50 crude Pu-erh tea samples, and the partial least squares models (PLS) were also established to predict the contents of these catechins by combining the near infrared spectra (NIRS) of the samples. The established models were then verified. The experimental results showed that the coefficient of determination (R2) values of the models for EGCG, EGC, and GA models were 99.99%, 99.99%, and 99.92%, respectively. The values of root mean square error of calibration (RMSEC) for EGCG, EGC, and GA were 0.17, 0.15, and 0.10, respectively, and the corresponding relative standard deviation (RSD) values were 0.19%, 0.27%, and 0.56%, respectively. The results of external validation indicated that the mean absolute errors between the predicted values and actual values of the three models for catechin monomer contents were 0.13, 0.12, and 0.07, respectively, and the mean relative errors were 0.17, 0.25, and 0.45, respectively. The experimental results show that the near infrared analysis model established in this study has high prediction accuracy and stability, and has potential application value in rapid detection of catechins in crude Pu-erh tea.