A Customized Electronic Tongue by Developing an Array of Molecular Imprinted Polymer-Based Voltammetric Electrodes for Tea Quality Evaluation

In this facile approach, a well-developed voltammetric electronic tongue (e-tongue) consisting of an array of molecular imprinted polymer (MIP) electrodes has been designed for the detection of theophylline (TH), tannic acid (TAN) and total theaflavins (TF) in real samples of black tea. These compounds are responsible for having a high impact on the taste of tea in terms of astringency, thus contributing to the taste of tea infusion. The developed e-tongue has been found to detect the signatures of these analytes from different tea samples under the same conditions. Thus, this e-tongue, in conjunction with a switching circuit, has been able to extract current responses from the working electrodes, giving an overall signature of the tea quality. The differential pulse voltammograms (DPV) responses were obtained to assess the performance on introducing MIP electrodes into the black tea samples. The obtained dataset was subject to the gradient boosting method. The prediction model was designed based on the responses of the e-tongue and yielded a mean-squared error (MSE) of 0.00003 and a coefficient of determination (R²) of 0.967 for the fused data responses. From the explained variance score and R-square value, it is clear that the overall quality of tea has been predicted better for the sensor array when considering the individual dataset from each sensor. On application of t-SNE (t-Distributed Stochastic Neighbour Embedding) on the fused data set, successful discrimination among the different black tea samples was obtained. The results point towards the possibility of an e-tongue based on an array of MIP electrodes to detect the quality of black tea liquor.