*5.2.3 Discrimination based on spectroscopy methods*

In the last decade, vibrational spectroscopy, namely infrared-based techniques (IR), supported by chemometric methods, was found to be a powerful technique because of its widespread use in analytical laboratories, its versatility and low economic impact. Moreover, IR technology requires minimal sample processing prior to analysis [5].

Certainly, IR spectroscopy coupled with multivariate data analysis has been used for the determination of oak volatile compounds [101–103] and for classifying barrels [104]. Subsequently, in recent years, this method has been proposed for discriminating wines aged in different types of wood containers and for different time periods [96, 105, 106], taking advantage on PLS-based calibration between GC-MS and near infrared spectroscopy (NIR).

The first publication appeared in 2012 [107], which investigated two different levels of information on fusion of NIR spectra and midinfrared (MIR) spectra from red wines aged in different ways. A total of 96 red wines, including wines aged in oak barrels, wines aged in stainless steel tanks with oak chips and without, were analyzed. Discriminant models of the three different aged wines were established, and the FDA method was applied to build the classification models of three different aged wines using the NIR, MIR and the merged spectra, reaching up to 98% correct classification with the latter. The results suggest that the spectral fusion of NIR and MIR is a promising technology for discriminating different aged wines.

A recent paper aimed to identify if spectroscopic techniques allow discriminating wines aged with alternative oak products (chips and staves) from different oak woods (American, French and Spanish) and floating micro-oxygenation (20 μg/L), compared with those aged in barrels, after 10 years of bottling [108]. The spectral information and analysis were performed in an FTIR-ATR (Fourier-transform IR-Attenuated Total Reflection). The results indicated that with this technique it is possible to clearly separate the wines aged by the three systems (chips, staves, and barrels) in the case of American oak. In the case of French oak, wines aged with chips were clearly differentiated for wines from a single grape variety and with similar oenological features.

The most recent paper on this topic [96] analyzed approximately 90 red wines issued from the same wine appellation in Italy, including commercial barrel-aged wines and wines aged using different types of commercial oak chips. Wines were analyzed in transmittance using NIR. In order to test if combined explanatory variables made it possible to discriminate treatments, an orthogonal partial least squares discriminant analysis (OPLS-DA) was carried out. Several factors were considered, including the aging process, the type of oak used for aging (wood *barriques*, big barrels or chips) and the wine typologies (differing for some oenological parameters). OPLS-DA application reached >96.5% success in classifying wines depending on the aging process (no wood, barrel/*barrique* and chips) in the internal validation check and > 90% in an external test. Certainly, further studies are needed to validate the potential of the technique in order to improve it for wine authentication, enlarging both the array of wine types (e.g., testing different appellations and grape varieties), and the range of oak alternatives (staves, cubes, blocks, and other sizes of product) to be tested, possibly in the presence of micro-oxygenation.

It is worth noting that IR spectroscopy has been used in other studies on wine aging, although without direct comparison between wood alternatives and barrel aging. For instance, FT-IR and UV-visible (UV-Vis) spectroscopic techniques combined with multivariate analysis were used to obtain regression models to study the aging level of high-quality Sherry wines [106]. Moreover, [105] wines aged in barrels made from different wood species and in stainless steel tanks, were analyzed. A complete differentiation of the samples was achieved according to grape variety, the container type and the aging time based on two spectral regions of their FT-IR spectra. The overall significance of these studies is still provisional because of the restricted data sets and the inhomogeneity of spectral ranges analyzed in the different papers. However, the results show the potential of IR spectroscopy and chemometric analysis for discriminating wines issued from different aging processes. In this regard, identifying discriminating algorithms and creating robust databases is a necessary condition to obtain a method that can be used routinely in a laboratory for quickly acquiring specific information about the type of wood used for wine aging.

Finally, nuclear magnetic resonance (NMR) is a powerful tool for analysis, quality control and authentication of wines. The main advantage of non-targeted wine analysis by <sup>1</sup> H NMR spectroscopy is the ability to collect, relatively simply and fastly, a huge amount of compositional information relating to a single sample [109]. Unlike other analytical techniques, however, it requires complex instrumentation and specialized personnel to be realized, but thanks to its versatility and reproducibility, this analytical technique seems to be very promising for the purposes described in this chapter.
