Puesta a punto de un nuevo método de análisis instrumental por GC-MS y desarrollo de nuevas metodologías basadas en técnicas espectrales para la estimación de la composición aromática varietal de la uva

Abstract

Grape volatile composition is one of the most important parameters for determining must and wine quality. Grape aromas, also called primary aromas, are divided into varietal and pre-fermentative aromas. These compounds are found in very low concentrations in the grape, so their identification and quantification require an efficient extraction method for subsequent chromatographic analysis. The analysis of volatile compounds in grapes by chromatographic methods has many drawbacks, such as time consumption, loss of sample and sample preparation, costs of reagents and instrumentation, and the need for trained personnel, so this equipment is only available in some research centers and major wineries. Currently, there is no method that allows routine real-time analysis of the aromatic composition of grapes throughout the ripening process, which would allow the winegrower to make decisions regarding viticultural practices, the choice of harvest date, or the classification of grapes according to their aromatic quality. This has become more important in recent years due to the unbalanced industrial maturity (sugar/acid content) and phenolic and aromatic maturity in grapes, caused by the effect of climate change. Consequently, the berries reach the appropriate sugar content (measured as Total Soluble Solids (TSS) (ºBrix)) more quickly, but not phenolic and aromatic compounds (compounds that are synthesized later in the berry). In recent years, rapid and non-destructive methods have been developed to relate spectral data to the concentration of specific chemical components (TSS, phenolic compounds, aromatic compounds, nitrogen compounds, etc.). Two of these technologies are Near-Infrared Spectroscopy (NIRS) and Hyperspectral Imaging (HSI). Both have been used to estimate general parameters in grapes, as well as amino acids and phenolic compounds. NIRS has been used to measure the aromatic composition of wines, but has been used very few times to estimate the volatile composition of berries. As for HSI, it has been used to measure the aromatic composition of other matrices such as coffee, eggs, or pork, but only one study has been found in which it has been used to estimate the aromatic composition of grapes, and none to monitor it throughout the ripening process. For all these reasons, this doctoral thesis has developed new methodologies, based on spectral techniques, which allow the estimation and monitoring of the aromatic composition of grapes throughout the ripening process, in a fast and non-invasive way. For this purpose, first of all, three techniques for the extraction of volatile compounds in must were optimized: Stir Bar Sorptive Extraction (SBSE), its variant multi-SBSE (mSBSE), in which polydimethylsiloxane (PDMS), and ethylene glycol (EG) stir bars were used simultaneously, and Thin Film Solid Phase Microextraction (TF-SPME), in which PDMS/carboxen (PDMS/CAR) and PDMS/divinylbenzene (PDMS/DVB) thin films (TFs) were tested. The optimized factors were the extraction mode, speed, time, and temperature, and the addition of NaCl. In turn, the analysis of these compounds by Gas Chromatography - Mass Spectrometry (GC-MS) was developed. Once the three techniques were optimized, the one that provided the best results was chosen, which was TF-SPME-GC-MS, with the PDMS/CAR TF and under the conditions of direct immersion (DI), at 500 rpm, for 6 hours at 20 ⁰C. This technique was subsequently used as the reference method for the determination of volatile compounds in the must samples. In the second place, spectral measurements were taken, under laboratory conditions, of 240 samples of Vitis vinifera L. Tempranillo and 240 of Tempranillo Blanco, collected during the 2019 and 2020 vintage, by HSI in the 400-1,000 nm range and by NIRS in the 1,100 - 2,100 nm range. In the third place, the TSS were measured and the volatile compounds of these same samples were analyzed with the optimized TF-SPME-GC-MS method, the chromatograms were integrated and, in this way, the reference data were obtained. Finally, multivariate calibration models were constructed by Modified Partial Least Squares (MPLS) regression using the acquired spectral data and the reference data. The results obtained showed that HSI and NIRS techniques allow to differentiate between high, medium, and low values of each volatile compound, and even to quantify the concentration of many of them, as well as of the TSS, in the Tempranillo Blanco variety throughout its maturation. On the other hand, the results obtained for the Tempranillo variety showed that HSI and NIRS spectral techniques allow differentiating between high and low values, and sometimes differentiating between high, medium, and low concentration values of volatile compounds and TSS. In this doctoral thesis, for the first time, a tool has been developed to predict technological maturity and aromatic maturity simultaneously, in a fast, contactless, and non-invasive way, throughout the ripening of Tempranillo Blanco and Tempranillo berries.