Abstract: The proposed approach reports the combined advantages of biosensors made of molecularly imprinted polymers (MIPs) and the modelling capabilities of Artificial Neural Networks (ANN) in a bio-electronic tongue (BioET) analysis system for the very first time. Molecularly imprinted polymers taylor-made for 4-ethylphenol (4-EP) and 4-ethylguaiacol (4-EG) and their control polymers, non-imprinted polymers (NIPs), were succesfully synthesized with similar morphologies and integrated onto an electrochemical sensor surface, as the recognition element, via sol-gel immobilization. The resulting MIP-functionalized electrodes were employed to arrange an array of different biosensor electrodes to quantify by means of ANN the binary mixtures of 4-EP and 4-EG yielding an obtained vs. expected correlation coefficient >0.98 and a normalized root mean square error (NRMSE) <0.076 (external test subset)
Template and target information: 4-ethylphenol, 4-EP, 4-ethylguaiacol, 4-EG
Author keywords: molecularly imprinted polymer, Electronic tongue, Artificial neural networks, differential pulse voltammetry, Volatile phenols