E. Brenon¹, S Panigai², D Brosseau², C Pascal¹, H Mahé^2
1 Vinventions – Oenology team, Nîmes, France
² Bureau Interprofessionnel des Vins de Bourgogne, Beaune, France

Wine is the result of a complex balance between compounds that is influenced by winemaking practices. In 2020, the Bureau Interprofessionnel des Vins de Bourgogne launched a project aiming to understand how the longevity of white Burgundy wines is built, and at which stages of winemaking.

The longevity of a wine is linked to its intrinsic sensitivity to oxidation and to the more or less oxidative conditions it undergoes during storage. The bibliography describes in detail the impact of oxygen on the evolution of wines in bottle, whether the oxygen intake is linked to bottling or to transfer via the obturator ([1]-[11]). Beyond the contribution of oxygen, polyphenols have been shown to be involved in wine oxidation mechanisms ([12]-[14]), which can lead to a deterioration in the organoleptic properties of the final product, whether red, white or rosé.

One of the aims of this project was to study the extraction of polyphenols during the pressing of white Chardonnay grapes from Burgundy, and to assess the concentration of these compounds in the settling tank prior to alcoholic fermentation.

This work was carried out as part of the VOLTA project, in partnership with some forty partner wineries spread across the whole of Burgundy (from Mâcon to Chablis) and representative of the region’s different types of winery (négociants, caves coopératives, domaines). The measurements were carried out in the cellar, in order to better explore the reality of Burgundy practices, using a technology that allows measurement without prior sample preparation (electrochemistry) and on a sufficiently large number of juices to allow statistical processing of the data. The results presented below correspond to four vintages of data collection (2020 to 2023). They highlight 3 types of polyphenol extraction during pressing in Burgundy, as well as the impact of pre-pressing practices such as sulfiting or mechanical harvesting on polyphenol concentrations in free-run juice and then in settling tanks.

Material and method

Polyphenol concentrations were measured using a potentiostat, the Polyscan (WQS Vinventions), on printed electrodes (carbon working electrode, Vinventions) by linear sweep voltametry (0-1200 mV, 100 mV/s, 10 mV steps). Analysis was performed immediately after sampling on a drop of juice deposited on the electrode. The PhenOx concentration indicator (total polyphenols, expressed as mg/L gallic acid equivalent) was calculated from the raw data as described by Ugliano et al. ([15]).

Measurements of must in settling tanks were taken immediately after sampling at the tasting tap, or at the top of the tank, at least 50 cm below the surface of the liquid. A total of 153 measurements were taken in 2020, 105 in 2021, 142 in 2022 and 25 in 2023.

For press monitoring, around ten measurements were taken over the course of the pressing period. The first analysis was carried out on juice collected when the press was launched. Subsequent analyses were carried out on juices obtained within one minute after reaching the target pressure for a given stage of the press cycle. This allows  collecting juices at a high flow rate at the press outlet, thus obtaining a sample representative of the stage and limiting oxidation of the sample by juice/air contact. A total of 70 pressing cycles were monitored in 2020, 31 in 2021, 15 in 2022 and 7 in 2023. Musts from the monitored presses were also sampled in settling tanks. The start of pressing was defined as the phase between the first pressurization and 25% of the pressing time. The end of pressing was defined as the phase beyond 75% of pressing time. The average PhenOx values for each of these phases (beginning, middle, end) of pressing were calculated for each press monitored. PhenOx variations were thus calculated. The calculation of these synthetic variables made it possible to compare different pressing cycles, both in terms of duration and pressure levels (number, pressure, number of re-drafts), but also with different monitoring (number and distribution of samplings).

A database was compiled from these measurements and supplemented by data qualifying the oenological practices from which the samples were taken: sulfiting, timing and dose of sulfiting, type of pressing cycle, type of press (open cage, closed cage, pneumatic, plate press), type of harvest (manual, machine), cooling of the harvest, press filling time.

Mathematical and statistical data processing was carried out using Excel (Microsoft), R and RStudio.

Results

Overview of polyphenol concentrations in settling tanks

For each one of 2020 to 2023 vintages, measurements were grouped into classes of 100 PhenOx units. The distribution of measurements for 2020 is very close to the distribution obtained for all Polyscan measurements on Chardonnay juice, all wine-growing regions combined, between 2015 and 2020 (over 3,000 measurements). Conversely, the distributions of measurements in 2021, 2022 and 2023 are shifted towards the lower PhenOx classes. The 2020 vintage was a hot one in Burgundy. The vines suffered from water stress (according to data from the BIVB’s Observatoire du Millésime), a condition known to lead to a higher concentration of polyphenols in the juices ([16]). Conversely, the 2021 vintage was cooler, consistent with lower polyphenol levels. Finally, 2022 was a warm vintage, but in which little water stress was observed (Communication 2023 Observatoire du Millésime du BIVB), and a high grape load was present on the vines, possibly contributing to a dilution of these compounds. In 3 of the 4 vintages measured, Burgundy Chardonnay musts had lower polyphenol loads than those observed for the same grape variety, all production areas combined.

Polyphenol extraction dynamics during pressing

Observation of the various press cycles monitored revealed a variety of PhenOx levels and PhenOx extraction dynamics during the cycle.

Each of the 70 presses monitored in 2020 was characterized by 3 variables, namely the PhenOx of the free-run juice, the variation in PhenOx between the beginning and middle, then between the middle and end of pressing, as indicated in the materials and methods.
In order to identify possible pressing categories in terms of PhenOx content and extraction dynamics, a multivariate statistical method was used for clustering:

• Principal Component Analysis on 3 variables per press monitored: principal components are new, independent variables with maximum variance.
• Hierarchical Clustering on the axes of Principal Component Analysis. Cluster analysis is used to divide a dataset into homogeneous subgroups.

Principal component analysis, whose first 2 dimensions explain around 78% of the variance between individuals, and hierarchical ascending classification were used to group the presses into 3 clusters (figure 2).

Figure 3 shows individuals that are representative of each cluster. Cluster (a) comprises 28% of presses, and is characterized by high PhenOx in free-run juice, with no increase in PhenOx during pressing. Cluster (b) accounts for 60% of presses; individuals in this cluster have low PhenOx at the start of pressing, then show an increase in PhenOx during pressing. Even so, until the end of pressing, the PhenOx values measured remain below the median value calculated for Chardonnay musts, all production zones combined, from 2015 to 2020. Finally, cluster (c) covers 12% of presses, which are characterized by low to medium PhenOx values at the start of pressing, and show an increase during pressing to reach high PhenOx levels (above 600).

Press monitoring for the 2021, 2022 and 2023 vintages were subjected to the same statistical treatments, revealing the same 3 types of polyphenol extraction dynamics in Burgundy presses. The percentage of presses in each cluster is also equivalent across all vintages, give or take a few percent.

As a result of these polyphenol extraction dynamics, and as validated by measurements on the settling tanks of juices from the presses of each cluster (figure 4), musts from the presses of cluster (a), which PhenOx is high from the beginning to the end of pressing, are systematically rich in polyphenols in the settling tank, whether or not a separation of free-run and press juices is carried out. Musts from cluster (b), with low PhenOx levels throughout pressing, are systematically low in polyphenol concentration in the settling tank. Juices from cluster (c) presses show greater variability in PhenOx levels in the settling tank, which seems consistent with low to medium levels at the start of pressing and a sharp increase during pressing, depending on whether or not separation was carried out during pressing. It is interesting to note that, in practice, according to the results of a BIVB survey carried out in 2021 among its members (52 respondents), 23% of Burgundy winemakers surveyed said they systematically separate free-run and press juices, 35% never separate them, and 42% sometimes separate them.

None of the pre-pressing or pressing oenological practices observed (type of harvest, harvest cooling, sulfiting on grapes, crushing) nor any type of press or pressing cycle (open cage, closed cage, pneumatic, tray, sequential cycle, automatic cycle…) is statistically significantly linked (Kruskal-Wallis test) to any of the clusters described above. Pre-pressing and pressing technical itineraries are made up of a succession of unitary operations (harvesting, transport, reception, transfer, sulfiting, inertizing, pressing, etc.), each of which may correspond to a variety of oenological practices. The result of one of these operations on polyphenol extraction can influence the result of the next, making it potentially difficult to demonstrate the predominance of one oenological practice on polyphenol extraction dynamics.

Impact of pre-pressing practices on polyphenol concentration in free-run juice

With no demonstrated link between oenological practices and polyphenol extraction dynamics during pressing, the focus was on one of the differences between the clusters, their richness in free-run juice polyphenols, to assess the impact of unitary operations on this parameter, which seems more directly linked to pre-pressing practices.

A statistically significant link (Kruskal Wallis test, on one vintage’s data: p.value=0.01 on 2020 data, p.value = 0.02 on 2021 data) between free-run juice PhenOx levels and harvesting method did indeed emerge. Machine harvesting leads to higher PhenOx levels than manual harvesting (figure 5 a). It has been hypothesized that this higher extraction of polyphenols is linked to the grapes being more thoroughly crushed and macerated with free-run juice before pressing, particularly during transport.

Similarly, long filling times and non-cooling of the harvest lead significantly to higher levels of PhenOx in the free-run juice, a priori favoring diffusion of the compounds by maceration.

On the other hand, crushing (figure 5 b) does not induce a higher level of PhenOx in the drained juice (Kruskal Wallis test, on data from one vintage, p.value=0.81 on 2020 data, p.value=0.73 on 2021 data), so it’s possible that the juice extracted during crushing is not in contact with the harvest long enough to promote extraction of polyphenols by diffusion.

Special case of sulfiting practices on grapes and must

Pre-fermentation sulfiting can be carried out on the grapes (before transport, during reception in the cellar, in the press) or on the juice after pressing (in the press hopper or in the juice reception tank). Its aim is to limit enzymatic oxidation and the development of undesirable flora. It is also known to act as a solvent for polyphenols.

Sulfiting grapes before pressing (before transport, during reception in the cellar or in the press) results in a higher PhenOx content in free-run juice (Kruskal Wallis test, p.value=0.006 on 2020 data, p.value=0.02 on 2021 data). It is probable here that the addition of sulfites to grapes favors the extraction of polyphenols from the juices and maintains their concentration by limiting enzymatic oxidation through inhibition of the PPO (polyphenol oxidase enzyme).

Moreover, even if it has no link with the dynamics of polyphenol extraction at pressing, it is the only pre-fermentation oenological practice that has a significant impact on polyphenol content in the settling tank, i.e. on the result of all pre-pressing, pressing and juice separation operations. However, the data collected in this project have shown that the timing of sulfite application (on grapes or after pressing) does not have the same effect.

If sulfiting is carried out prior to transport, during reception in the cellar or in the press, it results in a significantly higher level of PhenOx (figure 6 a) compared to the level observed during the same vintage on tanks of juice from a process not sulfited during these stages (Kruskal Wallis test, on data from one vintage: p.value=0.006 on data from 2020, p.value = 0.007 on data from 2021). As described for its impact on free-run juice, it seems that sulfites applied to grapes favor the extraction of polyphenols from free-run juice, protecting the polyphenolic load from oxidation and therefore its concentration. On the other hand, if sulfiting is carried out in the press bed or in the settling tank, it has no significant impact on PhenOx levels (figure 6 b), compared with the level of juice in the settling tank that has not received sulfiting (Kruskal Wallis test, on data from one vintage: p.value=0.11 on 2020 data, p.value = 0.21 on 2021 data). In this case, the SO₂ is added after pressing, i.e. after extraction of the polyphenols. It therefore only has a protective role against oxidation during transfer and/or during the settling time in vats, during which juice/air contact is limited compared to that during the previous unitary stages

Conclusion

The VOLTA project presented in this article is original in its approach to large-scale data collection, in real winemaking situations, in structures representative of Burgundy players in terms of geography, size or type (producers, wineries, cooperative wineries). Data was collected over 4 years, enabling conclusions to be confirmed for vintages of varying temperatures and yields.

This study shows that Chardonnay musts from Burgundy are generally less rich in polyphenols than Chardonnay musts from other wine-growing areas. The only exception to this conclusion was observed in 2020, a particularly hot vintage during which the vines in Burgundy underwent significant stress.

The pressing study carried out as part of this project highlighted 3 main categories of polyphenol extraction dynamics in Burgundy. It was shown that nearly 70% of pressings, all  vintages combined, lead to low polyphenol concentrations in the settling tanks.

No statistically significant link could be found between the pre-pressing practices studied (type of harvest, sulfiting of grapes, crushing, cooling of harvested grapes, etc.) and the type of polyphenol extraction during pressing, nor the settling tank content. At best, these practices have a significant impact on the polyphenol content of free-run juice from presses. The only practice that significantly increases the polyphenol content of the settling tanks is sulfiting of the grapes, whether this is carried out in the skip, in the hopper or in the press. On the other hand, post-pressing sulfiting (at the press bed or in the settling tank) does not lead to an increase in polyphenol concentration.

The remainder of this project is focused on gaining a better understanding of how the longevity of Burgundy white wines is constructed, with 1^er aim of establishing the link between polyphenol content in the settling tank and the oxidation sensitivity of wines at the end of alcoholic fermentation. Ageing phases will then be studied.

Acknowledgements: the authors would like to thank all the partners in this project for their hospitality in carrying out the measurements and for sharing their experience on this subject.