The new research, published in Electroanalysis, aimed to copy the human taste system due to a combination of “sensor systems and advanced mathematical procedures”.
The device can also detect defects in the wines, said the researchers at the Autonomous University of Barcelona who have been researching and developing electric tongues over several years.
The electronic tongue they have developed can currently identify three types of cava: Brut, Brut Nature and Medium-Dry. However, with proper training the device will be able to identify all wine varieties, claim the researchers.
As the type of cava is determined by the amount of sugar added, it was important to find out the sugar content to distinguish between different wine varieties, said the scientists.
The resulting classifications are: Brut Nature (<3 g/L, no sugar added), Extra Brut (<6 g/L), Brut (<12 g/L), Extra Dry (12-17 g/L), Dry (17-35 g/L), Medium-Dry (33-50 g/L) and Sweet (>50 g/L).
In order to design the electronic tongue, the researcher group identified 21 different cava samples using voltammetric measurements.
Using a combination of principal component analysis (PCA), discrete wavelet transform (DWT), and artificial neural network (ANN), the researchers said they managed to replicate the human taste system and distinguish between different types of cava, obtaining a classification similar to a sommelier or wine professional.
A second order standard addition method (SOSAM) was used to correlate the sought analytical information.
This made it possible to quantify the amount of sugar added in the cava production process, said the researchers.
The scientists said they are currently working on perfecting the device through the incorporation of biosensors.
Electronic tongues contain a sensor matrix to obtain chemical information from samples in the same way that they are obtained by human senses, said the scientists.
The perception of taste is based on the generation of sensory patterns of the nerves activated by the brain and nerve print recognition, this is achieved with the use of computerised systems which interpret data obtained by the sensor matrix, they explained.
As in biological mechanisms, a learning and training process is required to enable the electronic tongue to recognise properties that need to be identified, the scientists added.
“Voltammetric Electronic Tongue in the Analysis of Cava Wines”
Volume 23, Issue 1, Pages 72-78
Authors : X. Cetó, J. Manuel Gutiérrez, L. Moreno-Barón, S.Alegret, M. del Valle