Taste Quality Detection Of Longjing Tea Based On Electronic Tongue Techniques

Longjing tea is well known for its value and quality in the world. In recent years, Longjing tea, especially Xihulongjing tea, is subjeted to frauds:the addition of cheap teas and/or the longjing teas from other regions. These acts hurt the benefits of the consumers greatly and disturb the market order. Hence, it is necessary to find a rapid, simple and valid method to evaluate the quality of Longjing tea. An electronic tongue (e-tongue) is a biomimetic sensing device that detects liquid information with a set of sensors array and has advantages such as objectivity, easy-to-handle measurement set-up and rapidness. This techniques will bring good prospect in the field of analyzing and recognizing component and taste in the liquid.In this work, Longjing tea was used as the main material object and the aim was to study the reliability and techniques of the application of the electronic tongue on testing tea quality. A series of investigations about evaluating the quality of Longjing tea were carried out by means of the electronic tongue technique and pattern recogniton methods, including discrimination of tea grade and producing region, identification of the authenticity and estimation of storage time. The chemical analysis and sensory evaluation of the tea samples were also performed. The relationships between the response signals of the electronic tongue and the chemical components or sensory scores of the tea samples were analyzed. The results and conclusions were listed as followed:(1) The response signals of the e-tongue to the tea samples were analyzed. The best testing conditions were determined and it was observed that the sensors'sensitivities were different to the tea samples with different grade, producing region, producing year and storage time.(2) The classification and identification of the tea samples with different grade, producing region, producing year and storage time were investigated by inducing diffenrent pattern recognition methods. It was found that Principal Component Analysis (PCA) could provide general classification informations of the various tea samples and the reducing dimension data provided enough information of the samples raw data in the research of region identification of Xihulongjing tea by the analysis of PCA. The corret predictions of each group tea samples were achieved by Discriminant Analysis (DA). The Soft Independent Modeling Class Analogy(SIMCA)models for identifying the samples from the Xihu region or not, and the freshness of Xihulongjing tea were build up. Hierarchical Cluster Analysis (HCA) was able to correctly discriminate the samples with different storage time.(3) It was found that there were differences among the samples with different grade, region, authenticity or storage time by the chemical analysis and sensory evaluation. The sensory results could explain and demonstrate the classification of the samples by the e-tongue.(4)The every taste chemical component was regarded as the dependent variable, the esponse singals of 7 sensors were regarded as the independent variables. Step linear regression nodel was established. The relation of the chemical component and response signal was inalysized. The Step linear regression model between the sensory scores and the response singals were also established. The result showed all models had statistic significance and good fitting. It was also observed that different sensors played different role in the models. The sensor CA and BB had a great effect on all the taste chemical components models. The response of sensor CA was evident for the taste components except for tea polyphenols.It was indicated that relationship between the sensors and taste chemical components of the tea samples with different grades/geographical origins. The sensor of HA, BB and GA had high correlations with the content of tea polyphenols. The sensor of CA and HA were sensitive to dissociative amino acid of tea sample and were highly related with the taste of tea. The sensor of BA and JB had high correlations with the concentration of tea liquor. The sensor of CA, JB and BB were sensitive to theine of the tea sample with different grade or geographical origin. The sensor of CA and JB were sensitive to the taste chemical components except for tea polyphenols for the tea sample with with different grade or geographical origin. The sensor of ZZ and BA were sensitive to tea polyphenols of the tea sample with with different grade or geographical origin.The sensor of HA and JB were highly correlated with the sensory score in the grade model. The sensor of CA and JB were highly correlated with the sensory score in the region model. The sensor of BB and CA were highly correlated with the sensory score in the authenticity model.