The application of near infrared spectroscopy to wine analysis: An innovative approach using lyophilization to remove...

The implementation of near-infrared spectroscopy as an analytical method for the quantification of major wine parameters is limited due to the aqueous nature of wines. Water molecules contribute to a poor signal-to-noise ratio and to suppress important groups' vibrations frequencies, preventing the quantification of most chemical compounds present. This paper proposes an alternative approach for the quantification of major wine indicators based on near infrared spectroscopy using lyophilized wine samples. A diversity of wine samples, including red, white and rosé, were lyophilized and analyzed by NIR spectroscopy. The parameters quantified were: alcoholic degree, volumic mass, total dry extract, total sugars, total acidity, volatile acidity, pH, free sulfur dioxide and total sulfur dioxide. Calibrations using partial least squares (PLS) regression were performed against the results obtained by reference methods. Spectra collected within 10,000 to 4000 cm^-1 range were randomly divided in two sets: one for the optimization of the PLS models and the remaining for external testing. The PLS models obtained were able to accurately quantify total sugars, pH, volumic mass and total dry extract with a range-error-ratio above 10. The quantification of the remaining parameters yielded unsatisfactory results. This methodology proved to be an interesting alternative for the quantification of major wine quality descriptors by circumventing the interference of water bands. Further studies exploring different lyophilization conditions and additional wine chemical compounds present at low concentrations are needed.