Experimental wine bottle tracks oxygen moving through the cork

To bypass this issue, the team designed a custom experimental rig they called the miniature bottle system. “The idea was to see what mechanisms are at work in this system,” Chanut said. The setup consisted of small glass vials designed to mimic the standard cylindrical geometry of a commercial wine bottleneck. Each vial was sealed using scaled-down cork stoppers ranging in length from 6 to 42 millimeters; the interior could be precisely loaded with either gas or a specific volume of model wine. The reduction in the total volume of both the liquid and gas phases artificially amplified any oxygen concentration changes that occurred. The system acted as a chemical magnifying glass that enabled the scientists to precisely measure extremely subtle physical and chemical mechanisms like outgassing through the cork or the reactions at the interface between the cork and the wine. Armed with their miniature bottle setup, the team loaded half of the vials with wine, left the other half empty, sealed them with the selection of different length corks, filled them with sensors, and left them for 18 months to age. It turned out the oxygen dynamics in the vials was way more complex than a simple, steady leak through the cork. During the experiment, the researchers learned there are four stages of oxygen transfer through the cork, which starts the moment a cork is rammed into a bottleneck. The first phase lasted for the initial 15 days after the vials were corked. “It was an equilibration between the liquid phase of the model wine and the gas phase,” Chanut said. There are differences in the gas content between wine that had been aged in sealed containers and the small bit of air that gets trapped and pressurized by the insertion of the cork. In the experiment, the oxygen dissolved in the liquid phase in the vials escaped back to the gas phase. The second phase that followed, though, was where things got a little more surprising. Chanut’s team observed that, during the first six months, the majority of oxygen that was getting into the wine wasn’t coming from the outside environment. The oxygen, it turned out, was coming from the cork itself, diffusing out of the microscopic spaces in the cork’s cellular structure. The cork was basically outgassing into the bottle. “Because we used model wine in the experiment and focused on oxygen transfer, we didn’t do any tasting tests,” Karbowiak said. But oxygenation does matter for taste and, Karbowiak claims, the team is already getting a lot of interest from both wine makers and cork manufacturers. “Wine is a very particular case of a product without a shelf life. So, the question is, ‘When should I drink my wine?’” Karbowiak said. “And actually, we are not able to answer this question.” His team hopes that obtaining detailed data on how specific types and dimensions of stoppers manage wine’s oxygenation after bottling may one day enable wineries and cork manufacturers to solve this problem. But there’s much more we need to learn before we get there. In the future, Karbowiak’s lab wants to focus on quantifying the exact balance and interplay between the four oxygen transfer mechanisms they discovered. While the team isolated the individual phases, determining how they work with different types of corks and varying environmental aging conditions remains unknown. Also, because cork is an inherently variable biological material, scientists want to examine how its properties change over several years of storage. For Karbowiak and his team, the goal is to develop methods to evaluate a wine’s initial oxidative potential so that winemakers can pair a specific vintage with a stopper that guarantees desired taste at a precise future date. “We need to know how much oxygen the wine should contain when it is optimal for tasting,” Karbowiak said. “If you have that information, you can select the stopper you need for the right preservation over a specific period of time to pinpoint the moment your wine is at its best.” Science Advances, 2026. DOI: 10.1126/sciadv.aed3023