Rapid assessment of red wine compositional parameters by means of a new Waveguide Vector Spectrometer

A new method for rapid assessment of selected wine parameters based on a 1.6e2.7 GHz Waveguide Vector Spectrometer (WVS) is proposed for the first time. The Gain and Phase spectra of forty-two samples of Sangiovese red wines were acquired in the best frequency range (1.6e2.1 GHz) and modelled using Partial Least Square (PLS) regression to predict alcohol, dry extract and glycerol contents. The prediction of alcohol content showed test set validation R2 values of 0.961 (RMSE = 0.11 mL/dL) and of 0.955 (RMSE = 0.13 mL/dL) for Gain and Phase spectra, respectively. Glycerol and dry extract were predicted at the best using Gain spectra with R2 values of 0.834 (RMSE = 0.31 g/L) and 0.861 (RMSE = 0.51 g/L), respectively. A test was also conducted to assess the technique ability in the detection of alcohol and glycerol contents by preparing and using samples of model wines with known added content of these substances. R2 values up to 0.992 (RMSE = 0.16 mL/dL) for alcohol and 0.999 (RMSE = 0.04 g/L) for glycerol contents were obtained using Gain spectra. The suggested technique has shown to be promising for the rapid assessment of the control process of these important attributes of wines. © 2017 Elsevier Ltd. All rights reserved.