This work reports the application of a BioElectronic Tongue (BioET) in the estimation of polyphenol content in wine. The approach used an array of enzyme biosensors capable of giving a wide and complete response of the analyzed species, plus a chemometric processing tool able to interpret the chemical signals and extract meaningful data from the complex readings. In our case, the proposed BioET was formed by an array of four voltammetric enzymatic biosensors based on epoxy-graphite composites, one blank electrode and the other three bulk-modified with tyrosinase and laccase on one side, and copper nanoparticles on the other; these modifiers were used in order to incorporate differentiated or catalytic response to different polyphenols present in wine and aimed to the determination of its total polyphenol content value. The obtained voltammetric responses were pre-processed employing the Fast Fourier Transform (FFT); this was used to compress the relevant information whereas the obtained coefficients fed an Artificial Neural Network (ANN) model that accomplished the quantification of total polyphenol content. For comparison purposes, obtained polyphenol content was compared against the one assessed by two different reference methods: Folin-Ciocalteu and UV polyphenol index (I280); good prediction ability was attained with correlation coefficients higher than 0.949 when comparing against reference methods. Qualitative discrimination of individual polyphenols found in wine was also assessed by means of Principal Component Analysis which allowed the discrimination of the individual polyphenols under study. © 2012 Elsevier B.V.
- Artificial neural network
- Electronic Tongue