Marsala is a dessert wine exclusively produced in the province of Trapani (Sicily, Italy). Twenty-nine different categories of Marsala are present on the market sort by the grape variety, production technology and aging. This research aims to develop a fast and easy method to characterize the different categories using attenuated total reflectance Fourier transform infrared (FTIR-ATR) spectroscopy combined with multivariate analysis. Principal Component Analysis (PCA), applied to spectral data, allowed separating wine samples of different sugar content and distinguishing the tanned samples (Fine, Superiore and Superior reserve) from the most valuable Virgin ones. Moreover, Linear Discriminant Analysis (LDA) was applied to the spectral data with a CV higher than 20% to discriminate among Marsala wines of different aging times. The results showed a complete discrimination of 100%. The confusion matrix of cross validation was equal to 87.76% indicating a high percentage of correct classification also in prediction. The proposed method is promising as it is simple and rapid and no sample pre-treatment steps are required. Moreover, it is environmentally friendly since no organic solvents are used. It could be of great interest to verify the conformity of the Marsala wines to the declaration labeled; moreover, it could be used for wine aging monitoring and/or verifying the effects of innovation in the production process.
Characterization and ageing monitoring of Marsala dessert wines by a rapid FTIR-ATR method coupled with multivariate analysis
Tripodi, Gianluca;
2018-01-01
Abstract
Marsala is a dessert wine exclusively produced in the province of Trapani (Sicily, Italy). Twenty-nine different categories of Marsala are present on the market sort by the grape variety, production technology and aging. This research aims to develop a fast and easy method to characterize the different categories using attenuated total reflectance Fourier transform infrared (FTIR-ATR) spectroscopy combined with multivariate analysis. Principal Component Analysis (PCA), applied to spectral data, allowed separating wine samples of different sugar content and distinguishing the tanned samples (Fine, Superiore and Superior reserve) from the most valuable Virgin ones. Moreover, Linear Discriminant Analysis (LDA) was applied to the spectral data with a CV higher than 20% to discriminate among Marsala wines of different aging times. The results showed a complete discrimination of 100%. The confusion matrix of cross validation was equal to 87.76% indicating a high percentage of correct classification also in prediction. The proposed method is promising as it is simple and rapid and no sample pre-treatment steps are required. Moreover, it is environmentally friendly since no organic solvents are used. It could be of great interest to verify the conformity of the Marsala wines to the declaration labeled; moreover, it could be used for wine aging monitoring and/or verifying the effects of innovation in the production process.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
