Julie Chanut a, c, Aurélie Lagorce a, Régis D. Gougeon b, Jean-Pierre Bellat c, Thomas Karbowiak a
a Univ. Bourgogne Franche-Comté, Institut Agro Dijon, PAM UMR 02 102, 1 Esplanade Erasme, 21000 Dijon, France
b Univ. Bourgogne Franche-Comté, Institut Universitaire de la Vigne et du Vin, 1 rue Claude Ladrey, 21000 Dijon, France
c Univ. Bourgogne Franche-Comté, Laboratoire Interdisciplinaire Carnot de Bourgogne, UMR 6303 CNRS, 9 Avenue Alain Savary, 21000 Dijon, France
Introduction and context
Oxidative stability is one of the main issues related to the shelf life of wine. Thus, the control of the oxygen transfer from the outside environment to the wine inside the bottle is a key parameter for preserving the wine quality. Obviously, many other factors can also influence the evolution of wine during post-bottling aging, such as the composition of the wine itself, the temperature, the relative humidity, the storage position, as well as the amount of oxygen initially present in the bottle [1-3]. However, the oxygen transfer remains the most critical factor. For this reason, the choice of the stopper is crucial in providing the best chance for wine aging in bottle [4, 5].
In a recent case study, oxidation was noticed in four bottles of white wine coming from the same vintage and production lot, i.e., visual examination showed obvious colour evolution after ten years of storage (Figure 1) [6]. To investigate this phenomenon, a multidisciplinary approach was designed combining chemical analyses, and physical investigation with both the wine and the system composed of the stopper and the bottleneck.
First, both the sensory evaluation and the chemical analyses of classical enological parameters unambiguously revealed the different oxidative states of the four bottles, with, for each vintage, one bottle being oxidized (Ox wine) compared to the other (NoOx wine). Further, a metabolomics analysis was performed by FT-ICR-MS. A total of 532 masses (chemical compounds) were significantly more intense in Ox or NoOx wines, of which 175 m/z values were distinct for Ox wines and 357 m/z values for NoOx wines. These molecular markers of non-oxidized wines are predominantly nitrogen-sulphur compounds (sulfonated polyphenols, amino acids/peptides, and glycosylated compounds), as revealed by the corresponding CHOS/CHO and CHONS/CHO ratios [7] (Figure 1). The molecular markers of oxidized wines are characterized by a significantly reduced contribution of nitrogen-sulphur compounds. These observations, show that CHOS and CHONS compounds, naturally involved in the oxidative stability of wines, were consumed in molecular mechanisms following high oxygenation in oxidized bottles.
Second, the oxygen transfer rate was determined through the whole system composed of the glass bottleneck containing the cork stopper, and then on the cork stopper alone with the interface glued (after uncorking). The objective was to characterize the contribution of stoppers on bottle aging of white wines in real condition, with particular emphasis on the bottleneck / stopper interface. The oxygen transfer measured for the cork stopper once extracted from the bottleneck is approximately the same for all four stoppers, with a mean value of 4 x 10-10 m2.s– 1, similar to those measured in cork in previous work for natural cork stopper [8, 9]. However, the transfer of oxygen through the cork stopper / glass bottleneck system was higher than through the cork alone in both cases (18 x 10-10 m2.s-1 for NoOx 2005 wine and 28 x 10-10 m2.s-1 for NoOx 2006 wine, respectively), and much higher for bottles containing the Ox wines (6938 x 10-10 m2.s-1 for Ox 2005 wine and 498 x 10-10 m2.s-1 for Ox 2006 wine, respectively) (Figure 1). The oxidation of these wines is therefore probably due to excessive oxygen transfer in the bottle at the interface between the stopper and the glass bottleneck.
In the following of this work, a more in-depth study was conducted to investigate the role played by the glass / cork interface, paying particular attention to the compression and to the surface treatment of the stopper. This study was performed starting from the diffusion through the stopper alone and ending with a more complex system comprising the stopper covered by a surface treatment and compressed in the glass bottleneck [8, 10, 11]. This comprehensive approach aimed to quantify and differentiate between the two oxygen flows passing through the stopper and at the interface between the stopper and the glass bottleneck.
Material and method
A homemade manometric method was used to determine the diffusion coefficient of oxygen through cork stoppers [8, 9, 12, 13].
The stoppers used for the experiments were microagglomerated stoppers with a diameter of 24.2 mm and a length of 48 mm. The first experiments on cork stopper alone, uncompressed (Figure 2, ①), were carried out on 3-mm thick cork wafers to reduce the time of measurement. In the case of compressed stopper (Figure 2, ②), cork samples were 6-mm thick to prevent buckling during compression. In both cases, cork wafer (compressed or uncompressed), was placed in a metal ring and the interface between the cork and the metal was glued to avoid any gas transfer at the interface. For the second experiments, on cork compressed in a bottleneck, non-coated full-length cork stoppers were inserted in empty glass bottles (Figure 2, ③). The bottleneck was cut and inserted in a metal ring with the part between the ring and the bottleneck glued. The same procedure was applied for cork stoppers coated with a surface treatment product (Figure 2, ④). They were coated with a monolayer surface treatment product composed of an emulsion of paraffin and silicone. For these stoppers, only 6-mm wafers were inserted in the bottleneck to reduce the time of permeation experiments.
The oxygen flow was measured through the sample separating two chambers. First, after a purge in the measuring chamber C1, the initial oxygen pressure was set to an initial at 900 hPa (± 0.1 hPa) value while the other chamber C2 was kept under dynamic vacuum (0.1 hPa). The temperature was kept constant at 20 °C (±1 °C) by a thermostated water circulation surrounding the measuring chamber. The decrease in oxygen pressure in the measuring chamber, caused by the transfer of oxygen through the sample, was monitored over time. From the slope of the pressure decrease over time, the oxygen diffusion coefficient of the cork wafer Dwafer is determined from Fick’s first law once the steady state is reached (Figure 2).
Then, the diffusion coefficient of the full-length stopper was calculated using an extrapolation model (Figure 3) [9, 12]. As the cork stopper is not a homogeneous material, the diffusion coefficient measured through a cork wafer of thickness lwafer is not representative of the diffusion coefficient through a full-length cork stopper having a length of Lstopper. To that end, the stopper was considered as a serial stack of n wafers where each slice represents a local resistance to gas transfer, Rwafer. The sum of these resistances permitted to calculate a global resistance Rstopper of the stopper and to determine Dstopper.
Finally, the oxygen flow through the cork stopper / glass interface, Jinterface (mol.m−2.s−1), is determined by subtracting from the total flow, going through the system comprising the cork stopper inserted in the glass bottleneck, Jtotal, the oxygen flow going through the compressed cork stopper alone, Jstopper.
J interface=J total-J stopper (Equation 1)
For practical convenience, OTR were calculated considering a stopper of 48 mm length and 18.5 mm diameter which is the diameter of a stopper compressed in a bottleneck, with an oxygen pressure differential of 200 hPa. It is given in mg of oxygen going through the stopper per year (Table 1).
Results and discussion
In a first step, the determination of the diffusion coefficient of oxygen Dstopper was carried out on uncompressed and compressed cork wafers, with in both cases the interface glued, in order to determine the effect of compression on the oxygen transfer through the cork stopper. The oxygen diffusion coefficient value from the extrapolated distribution is 1.4 x 10-11 m2.s– 1 and 9.2 x 10-12 m2.s-1 for uncompressed and compressed cork stoppers, respectively. The corresponding OTR are also displayed in Table 1.
Thus, when a compression is applied to the stopper corresponding to a reduction of 40 % in volume in the case of still wines, this leads to a slight decrease of the oxygen transfer through the stopper with a reduction by a factor of 1.5 of the oxygen diffusion coefficient. Such decrease in the oxygen transfer through an agglomerated cork stopper could be attributed to a reduction of the initial porosity between cork particles. This could vary according to the formulation of the agglomerated cork stopper (in particular the adhesive / cork ratio).
In a second step, the diffusion coefficient of oxygen through the stopper / bottleneck system was determined. First, this was done for stoppers without surface treatment and, then for stoppers coated with an emulsion of paraffin and silicon. In the case of the stopper without coating, the diffusion coefficient was tremendously higher, with a mean value of 1.3 x 10-7 m2.s– 1. Such stopper / bottleneck system is thus 10 000 times more permeable than the compressed stopper alone (with the interface glued). This poor resistance to gas transfer could come from the stopper / glass interface. Therefore, this addresses the likely role of defects (macropores, scratches, roughness irregularities, etc.) formed on the cork surface during its processing or at bottling, or of defects present on the surface of the glass bottleneck. Such defects might create preferential pathways for the gas transfer.
Table 1: Mean oxygen diffusion coefficient and corresponding oxygen transmission rate (OTR) determined for uncompressed or compressed cork stoppers alone (interface glued) and for cork stoppers inserted in a bottleneck with or without surface treatment agent. Data in brackets corresponds to the minimum and maximum values obtained from the standard deviation of the statistical distribution.
In addition to permeability measurements, Scanning Electron Microscopy (SEM) analysis was performed for agglomerated stoppers with and without coating (Figure 5). Without surface treatment, the cork particles are entrapped within the adhesive network that maintains the cohesion of the stopper (Figure 5a). With a surface treatment, the deposited coating layer fully covers the surface of the stopper and does not tend to form aggregates (Figure 5c). It is also distributed in a very thin layer over the cork cells and does not fill the volume of the outer empty cells. At a lower scale, the coating thickness could be estimated from cross-sectional images at around 0.5 μm.
Conclusion
The transfer of oxygen in a cork / bottleneck system was assessed through a comprehensive study starting from the diffusion through the stopper alone and ending with a more complex system comprising the stopper covered by a surface treatment agent and compressed in the glass bottleneck. First, the effect of compression on agglomerated cork (with the interface glued) showed that a 40% reduction in the volume of the stopper (compression applied for still wines) only slightly reduced the oxygen diffusion of the stopper. Then, the oxygen transfer through a corked bottleneck was examined. Without coating, the oxygen diffusion coefficient through an agglomerated cork stopper compressed in a bottleneck was significantly higher than through the stopper alone (10-7 m2.s– 1 Vs 10-11 m2.s-1), by a factor 10 000. However, when the stopper was coated with a silicone-paraffin emulsion, the oxygen diffusion coefficient through the cork / bottleneck gave a similar value to that of the compressed stopper (10-11 m2.s– 1). Altogether, the results highlighted the key role of the glass bottleneck / cork interface in the oxygen transfer. A coating with a thickness of less than 0.5 µm therefore provide an efficient barrier for the stopper / bottleneck interface system to act against gas transfer at the interface. In addition to its initial role of ensuring easier uncorking, the surface coating therefore confers a supplementary and unexpected barrier efficiency to the wine sealing system.
P. Arapitsas, G. Speri, A. Angeli, D. Perenzoni and F. Mattivi, The influence of storage on the ‘‘chemical age’’ of red wines, Metabolomics 10 (2014) 816-832 https://doi.org/10.1007/s11306-014-0638-x
P. Lopes, C. Saucier, P.-L. Teissedre and Y. Glories, Impact of storage position on oxygen ingress through different closures into wine bottles, Journal of Agricultural and Food Chemistry 54 (2006) 6741-6746 https://doi.org/10.1021/jf0614239
P. Lopes, M. A. Silva, A. Pons, T. Tominaga, V. Lavigne, C. Saucier, P. Darriet, M. Cabral, P.-L. Teissedre and D. Dubourdieu, Impact of the Oxygen Exposure during Bottling and Oxygen Barrier Properties of Different Closures on Wine Quality during Post-Bottling, 1104 (2012) 167-187 https://doi.org/10.1021/bk-2012-1104.ch011
K. Crouvisier-Urion, J.-P. Bellat, R. D. Gougeon and T. Karbowiak, Gas transfer through wine closures : A critical review, Trends in Food Science & Technology 78 (2018) 255-269 https://doi.org/10.1016/j.tifs.2018.05.021
T. Karbowiak, R. D. Gougeon, J.-B. Alinc, L. Brachais, F. Debeaufort, A. Voilley and D. Chassagne, Wine oxidation and the role of cork, Critical Reviews in Food Science and Nutrition 50 (2010) 20-52 https://doi.org/10.1080/10408390802248585
T. Karbowiak, K. Crouvisier-Urion, A. Lagorce, J. Ballester, A. Geoffroy, C. Roullier-Gall, J. Chanut, R. D. Gougeon, P. Schmitt-Kopplin and J.-P. Bellat, Wine aging: a bottleneck story, Npj Science of Food 14 (2019) https://doi.org/10.1038/s41538-019-0045-9
R. Romanet, F. Bahut, M. Nikolantonaki and R. D. Gougeon, Molecular Characterization of White Wines Antioxidant Metabolome by Ultra High Performance Liquid Chromatography High-Resolution Mass Spectrometry, 9 (2020) 10.3390/antiox9020115
A. Lagorce-Tachon, T. Karbowiak, C. Paulin, J.-M. Simon, R. D. Gougeon and J.-P. Bellat, About the Role of the Bottleneck/Cork Interface on Oxygen Transfer, Journal of Agricultural and Food Chemistry 64 (35) (2016) 6672-6675 https://doi.org/10.1021/acs.jafc.6b02465
S. Lequin, D. Chassagne, T. Karbowiak, J.-M. Simon, C. Paulin and J.-P. Bellat, Diffusion of oxygen in cork, Journal of Agricultural and Food Chemistry 60 (2012) 3348-3356 https://doi.org/10.1021/jf204655c
J. Chanut, J.-P. Bellat, R. D. Gougeon and T. Karbowiak, Controlled Diffusion by Thin Layer Coating: The Intricate Case of the Glass-Stopper Interface, Food Control 120 (2021) 107446 https://doi.org/10.1016/j.foodcont.2020.107446
J. Chanut, Y. Wang, R. D. Gougeon, J.-P. Bellat and T. Karbowiak, Unraveling the Complex Interfacial Properties of Cork-Based Materials in Their Use as Wine Stoppers, ACS Applied Materials & Interfaces 14 (37) (2022) 42602-42612 https://doi.org/10.1021/acsami.2c07299
J. Chanut, A. Lagorce, S. Lequin, R. D. Gougeon, J.-M. Simon, J.-P. Bellat and T. Karbowiak, Fast manometric method for determining the effective oxygen diffusion coefficient through wine stopper, Polymer Testing 93 (2021) 106924 https://doi.org/10.1016/j.polymertesting.2020.106924
A. Lagorce-Tachon, T. Karbowiak, J.-M. Simon, R. D. Gougeon and J.-P. Bellat, Diffusion of oxygen through cork stopper: Is it a Knudsen or a Fickian mechanism?, Journal of Agricultural and Food Chemistry 62 (37) (2014) 9180-9185 https://doi.org/10.1021/jf501918n
