August 2009 Issue of Wines & Vines


Research gets specific about oxygen, closures, bottling and aging

by Jamie Goode

  • New technology that enables scientists to measure oxygen in wine has fueled the recent interest in the wine-O2 relationship.
  • The goal of a new nonprofit association is "the promotion of scientifically based solutions for oxygen-management challenges in the wine industry."
  • The level of oxygen in wine can vary widely--even in wines bottled on the same bottling line.

A few years ago the hot topic in wine science was the closures debate, and more recently it has been reduction. Now it seems that everyone is curious about the complicated interaction between oxygen and wine--a theme that incorporates both of the earlier hot topics, but which puts them in a broader context.

Understanding the way that wine is affected by oxygen at various stages in its development is critical for winemakers who want to produce high-quality wines. But while it's been more than 100 years since Louis Pasteur first pointed out the love-hate relationship wine has with oxygen, it's only recently that wine scientists have focused their attention on this issue in earnest.

Part of the reason for this is that new methods have emerged to measure oxygen levels accurately and easily. Another contributing factor is that one of the world's leading closure companies has initiated a large collaborative project across several research centers, aiming to reveal more about the way different levels of oxygen ingress post-bottling affect various wine styles.

But perhaps the main reason there hasn't been a huge emphasis on understanding and controlling wine's interaction with oxygen is because wine is pretty robust. For centuries winemakers have simply filled bottles by gravity and then popped in a cork. And by and large, they have gotten away with it. There's a surprising variation in the levels of oxygen pickup among modern wineries--and even within wines along the same bottling line. In many cases, the actual level of oxygen pickup isn't monitored.

"Luckily for the wine industry, its product is quite robust compared to other beverages like beer," says George Crochiere, a technical consultant specializing in drinks packaging. "The rule of thumb for beer is that 1ppm of oxygen from filling, headspace and ingress will destroy a beer and determine the end of its shelf-life. This has forced the beer industry to spend a large amount of time and money to minimize oxygen. With wine, the amount of oxygen is higher and varies between wines, and some inconsistency in oxygen levels from filling, sealing and ingress does not have the same devastating effect."

Oxygen sensor
A probe collects information from an in-bottle sensor that monitors oxygen levels.

The big oxygen project
The goal of synthetic closure manufacturer Nomacorc's research initiative is to gain a better understanding of how oxygen, in combination with the closure, influences winemaking. There are two levels to the project. The first objective is to gain a greater understanding of ways to control oxygen during winemaking in terms of quality control.

Nomacorc CEO Malcolm Thompson describes this as the "low hanging fruit"--a relatively easy way to improve the consistency of wine development. "We have identified significant improvements in how to manage oxygen, particularly at bottling," he claims.

The second objective is described as "winemaker intention": to use knowledge of oxygen's effects on wine to put winemakers in a position where they can integrate closure design into winemaking. What sensory attributes are winemakers looking to achieve?

The project brings together several of the leading wine research institutions from four continents, each of which is studying the effects of post-bottling oxygen exposure on specific grape varieties. This research is conducted under an umbrella organization called Oxygen in Wines (O2W), an international nonprofit association consisting of wine industry suppliers and service providers, with the stated objective of "the promotion of scientifically based solutions for oxygen-management challenges in the wine industry."

Participating members include G3, Lallemand, Perrier Bottling Machines, Inter Rhône, Appe and Nomacorc. The project has three phases. The first is to measure oxygen and its impact by developing oxygen-measurement tools, new analytical approaches and rapid wine-aging tools.

The second is to understand the impact of oxygen on sensory attributes, looking at its effect on grape variety, winemaking style, bottling conditions, closure selection and storage and transport. The third is to control the wine-aging process by being able to translate the desired sensory attributes into closure requirements, matching the right closure to the right wine.

Oxygen pickup varies with choice of closure

Closures vary quite widely in their oxygen transmission levels, and this influences how wines develop post-bottling in significant ways.

Screwcaps come with two different liners for wine--tin/saran and Saranex only. It is easy to tell them apart: The tin/saran liner has a metal layer inside and is silver in color, while the Saranex-only liner appears white. The tin/saran liner allows very little oxygen transmission, while the Saranex-only liner allows a little bit more.

Technical corks such as Twin Top and Diam are more consistent in their performance and offer steady levels of oxygen transmission. Typically, they offer the same oxygen transmission properties as good natural corks, allowing just a little more oxygen transmission than tin/saran-lined screwcaps. Diam offers specific products that differ in their oxygen-transmission levels.

Natural cork varies in its oxygen transmission properties. Good corks allow relatively little oxygen transmission, while lower quality corks allow quite a bit more. It isn't possible to predict with any degree of certainty how individual corks will perform, although buying more expensive, higher grade corks will increase your chances of getting a good one.

Paulo Lopes' work has shown that the oxygen transmission rate of cork starts off relatively high, but then reduces significantly after some months in bottle--a fact not taken into account in many measurements. This dynamic transfer rate could be important, and it would be difficult to replicate with an alternative closure.

Synthetic corks allow quite a bit of oxygen transmission because they are made of plastic, which allows oxygen diffusion: The oxygen dissolves into the body of the closure on the air side and diffuses through to the wine side. However, not all synthetics are alike. Since they first appeared, much work has gone into improving their oxygen barrier properties, and many synthetic manufacturers will now offer closures with a range of oxygen transmission levels.


Taking measure

One of the key facets of this collaborative venture is PreSens, the technology being adopted to measure oxygen. Nomacorc has formed an alliance with PreSens Precision Sensing, a German company, to market this technology to the wine industry. The method, which involves the use of luminescence, also is used by manufacturers such as Oxysense.

In the presence of oxygen, a photodiode detects blue light when a red light is shone on an oxygen-sensitive sensor. The attraction of this system is that not only can the sensor be built into a probe, it also can be incorporated into a reusable dot that can be applied inside a bottle to follow oxygen over time in the same bottle of wine. These dot sensors make in-line measurement of oxygen possible, and because the measurement process is non-destructive, they can be used to follow the kinetics on the same bottle filled with wine.

Dr. Olav Agaard of Nomacorc says the cost for a PreSens machine is around 10,000 euros ($14,000), and that many laboratories, bottlers and consultants are interested in using it. The reusable sensors cost 20-25 euros each ($28-$35), and the system comes pre-calibrated. Two sensors are available: One for 0-4% oxygen calculates to an accuracy level of 1ppb, and the 0.5%-50% model has an accuracy rate of 15ppb.

A word on measurement units: Typically, dissolved oxygen concentrations are quoted in parts per million (ppm), which is the same unit as milligrams per liter (mg/liter). The maximum amount of oxygen a wine can contain is around 8 mg/liter, which is known as the saturation level. Typically, tank-to-tank movements will result in oxygen pickup of 0.1ppm-0.2ppm. The normal barrel aging process will add perhaps 25ppm oxygen to the wine at a rate of a few ppms per month. Nomacorc data show that fermenting must consumes as much as 5.5ppm oxygen in just 15 minutes.

Of course, there already exist a number of ways to measure dissolved oxygen. The best established of these is the electrochemical method, commonly known by the proprietary name of Orbisphere, one of the leading manufacturers of this equipment. This is an electrolysis system in which the amount of generated current is proportional to the oxygen partial pressure. This is a destructive method (that is, the wine sample can only be tested once), and is not as convenient as luminescence methods, but because it is a reliable, established technique, it remains the benchmark.

At the Australian Wine Research Institute (AWRI), George Skourounomounis and Elizabeth Waters have developed a dye technique for measuring oxygen ingress. This involves two reagents: methylene blue and BPAA (bis-9,10-anthracene-[4-trimethylphenylammonium] dichloride). It's a complicated chemical system, but its power is that it mimics what occurs in wine when oxygen is encountered: The reaction consumes oxygen and thus maintains a concentration gradient to drive ongoing oxygen diffusion through a closure. Light is used to convert oxygen from the triplet (unreactive) state to the singlet (reactive state) in the presence of methylene blue, which then reacts with the BPAA to cause a color change that can easily be measured using a spectrophotometer.

Paulo Lopez, now working for leading cork manufacturer Amorim, has used another dye method to look at oxygen ingress under a range of closures in work conducted with the University of Bordeaux II. This takes advantage of the yellow-indigo color change that occurs when reduced indigo carmine is exposed to oxygen. The advantage of both these dye techniques is that repeated measurements can be made on the same bottle because they are non-destructive. The disadvantage is that they can only be used with model wine solutions, not real wines.

Finally, MOCON is the industry standard method for measuring closure oxygen transmission, but it is expensive, time-consuming and difficult to adopt as a routine technique. It is also a measurement that can only be made on dry packages.

Oxygen's effects
So just what does oxygen do to wine? Many white wines are made reductively from the start, protecting them as much as possible from oxygen after crushing, but for some white styles and most red wines, oxygen exposure at some stages of fermentation is an important tool in winemaking. For all wines, some oxygen is needed for healthy yeast growth during primary fermentation, and a deficit will result in struggling ferments that are liable to produce sulfides, causing reduction problems.

Once fermentation is complete, the requirement for oxygen is much reduced--especially for unoaked white wine styles. Winemakers will look to protect wines during storage and movement by the use of inert gases and stainless steel tanks. However, the use of oak barrels is a deliberate attempt to make positive use of small levels of oxygen exposure during winemaking to achieve stylistic goals.

This is particularly important for red wines with substantial tannic structure. Increasingly, winemakers are using controlled oxygen delivery during winemaking (known as microoxygenation) to assist in developing structure, color and mouthfeel in red wines--although this is still pretty much an empirical process that involves a good degree of guesswork and tasting rather than exact measurement.

When a winemaker has decided his wine is ready for bottling (which can be as little as a few months and as long as several years after vintage) he needs to decide how that bottling is to take place--and also how the bottle is to be sealed. These decisions will have important implications for the shelf-life of the wine, and also how the wine will appear to consumers at the point of consumption.

One important choice is the level of free sulfur dioxide (SO2) used at bottling, because this is a molecule that has an important interaction with wine. Its reaction with oxygen is not a direct one, because it occurs only slowly. Instead, SO2 mops up initial reaction products of oxygen and prevents them going on to further oxidize wine components such as aroma and flavor molecules.

In wine, SO2 dissociates into free and bound forms, with the former being active in protecting wine from the adverse effects of oxygen. Typically, a wine will be bottled with a free SO2 level of 25-40 mg/liter. This free SO2 level shows a sharp initial decline after bottling because of the presence of oxygen dissolved in the wine, plus the headspace oxygen, plus, in some cases, release of oxygen from the closure body.

After this time, the free SO2 declines at a slower, steadier rate through closure oxygen transmission. For white wines, results from the AWRI have shown that oxidation characters usually begin to appear when the free SO2 level dips below 10 mg/liter. For red wines, this will be somewhat different because of the phenolic content of red wines, which itself is protective against oxidation.

Australian wine scientist Richard Gibson (whose consultancy is called Scorpex) has prepared a set of calculations that can provide an estimate of shelf-life in wines. He begins with the observation that 1mg of oxygen reacts with 4mg of sulfur dioxide. A typical headspace of 5.95ml, if it consists of air, will contain 1.24ml oxygen, which is equivalent to 1.78mg, which in turn equates to 2.37 mg/liter of oxygen, once it is dissolved in the wine. This can react with 9.5 mg/liter of SO2.

Then the amount of oxygen that enters the bottle through the closure over time can be calculated from the closure oxygen transfer rate. A rate of 0.01 cc/day equates to 0.019 mg/liter of oxygen entering the bottle each day, which can react with 27.7 mg/liter of SO2 during a year. From these figures, expected shelf-life of a wine can be calculated. For example, a wine with 35 mg/ml free SO2 at filling, 2 mg/liter of dissolved oxygen, 0.5ml oxygen in the headspace and a closure OTR of 0.008 cc/day, will have a shelf life of 217 days. These sorts of calculations make it clear that both the total pack oxygen (TPO) at bottling, and also the closure oxygen transmission levels are vital factors in determining how the wine will develop over time.

George Crochiere gives some examples of the sorts of levels of oxygen that might be introduced to wine bottles during different filling procedures. In the worst-case scenario, a gravity-filled bottle sealed with a cork without any vacuum will pick up 2.6ppm oxygen during filling and have a further 1.8ppm oxygen in the headspace, giving a total of 4.4ppm.

If a vacuum filler is used and the headspace is evacuated, this figure will fall to just under 1ppm. If a screwcap is used and the bottle is filled using a vacuum filler, then pick-up in bottle filling is 0.6ppm. If liquid nitrogen dosing is used, headspace pickup is 0.7ppm; without this it is as high as 4.75ppm, giving an initial TPO of 5.36ppm. Crochiere points out that in the beer industry, the best-run bottling lines give oxygen pick-ups of between 0.05 and 0.15ppm, while average lines are 0.2-0.4ppm.

These sorts of calculations are relevant to the first preliminary results to emerge from the Nomacorc-sponsored study, which come from Geisenheim in Germany, where Dr. Rainer Jung has been looking at the effect of bottling parameters on Riesling wines. Using PreSens, Jung has shown that the headspace represents a significant source of oxygen contributing to wine evolution.

The project evaluated the evolution of a Riesling wine under different bottling and post-bottling conditions; 375ml bottles equipped with sensor spots were filled using two headspace volumes, each containing three different concentrations of oxygen, emulating conditions typically encountered during actual bottling runs.

Dissolved oxygen in the tank was 0.3ppm, and after bottling the dissolved oxygen levels ranged from 0.9-1.3ppm (0.3-0.5 mg/bottle). Headspace volume was either 5ml-6ml or 17ml-19 ml, and two or three different oxygen concentrations were used in each case. Two closures were employed, a Nomacorc and a screwcap, with the former being stored in either 21% oxygen (air) or zero oxygen. The levels of total pack oxygen (TPO) measured during bottling ranged from 0.2mg per bottle to a high of 6.0mg--a dramatic range.

TPO measurements showed that the wines started with very different levels, but all oxygen was absorbed by 300 days. As well as the oxygen present at bottling, the Nomacorc contributed 2.5 mg/bottle oxygen during the 300 days of the trial, while the Nomacorc in an atmosphere of zero oxygen contributed a bit more than 1 mg/bottle oxygen, which must have come from the body of the closure itself.

The screwcap contributed around 0.3 mg/bottle. The TPO level correlated highly with decline in free SO2 levels and also the change in color of the wine. "Our results show that headspace oxygen, which has largely been ignored by the industry, is a critical factor impacting wine development and, more specifically, a wine's oxidation resistance influencing shelf-life performance," Jung reports.

Nomacorc also has results looking at what happens in real-life bottling runs. In one set of data from several different bottlings, total oxygen pick-up ranged from a low of under 1ppm, to a high of more than 13ppm, with a wide spread around a mean of just under 2.5ppm.

Another study looked at a single bottling line and found a spread of values, with a range of 0.9ppm, depending on the filling head, although in this case all were less than 3ppm. Interestingly, the contribution here from headspace was more significant than the dissolved oxygen level.

Taking it home
The take-home message seems to be that there is an urgent need for the wine industry to come to grips with oxygen pick-up at bottling--and then oxygen transmission by closures--because these important variables have a large effect on wine development and shelf-life.

"By optimizing certain processes on the bottling line and diligently monitoring and controlling headspace and dissolved oxygen levels, winemakers and retailers can extend the shelf-life of a wine and improve its consistency," says Dr. Stéphane Vidal, Nomacorc's chief enologist. Currently, the variation that exists in oxygen pick-up at bottling is liable to dwarf the effects of variations in closure performance, and thus Nomacorc's desired outcome of being able to advise winemakers on closure choice to match wine style will be thwarted by the real world situation of bottling line inconsistency. "Before winemakers can shape wine evolution through closure oxygen transmission, oxygen ingress at bottling must be under control," Vidal says.

It will be interesting to see the results from the various collaborative studies emerge during the next few years. One thing is sure: Enhanced knowledge about the interaction between oxygen and wine is likely to result in more wines reaching consumers in optimal condition, which has to be a good thing. Whether we'll ever get to the situation where winemakers use "designer closures" as part of their stylistic choice, however, is less certain.

London-based writer Jamie Goode is the publisher of and specializes in wine science issues. His first book was published in 2005 by Mitchell Beazley. Contact him through Wines & Vines at

Print this page   PRINTER-FRIENDLY VERSION   »
E-mail this article   E-MAIL THIS ARTICLE   »
Posted on 08.14.2009 - 22:34:50 PST
This is all pretty naive stuff. There is no mention of the work of J Ribereau-Gayon and colleagues, changes in Redox Potential, implications of the Nernst Equation, wine-making as a micro-aerobic process, the desirability of anaerobiosis in bottled wine or enclosed oxygen as a major factor in shelf-life equations published some twenty years ago.
It would be tedious to list all the inaccuracies and misapprehensions. Overall, it leaves me with the impression that it is just an elaborate marketing gimmick for synthetic stoppers.
North Ryde, NE Australia

Posted on 08.19.2009 - 08:31:39 PST
Kronos - We would, nonetheless, like to know precisely what inaccuracies and misapprehensions are contained in the article.

Your litany of citations, unfortunately, doesn't contribute to our understanding or the body of knowledge about this issue.

Rather than just disparage Mr. Goode, why not write and submit your own piece? And I suggest that you read his book "The Science of Wine" in which he deals with this subject in an extended fashion.

Thank you.
Metuchen, NJ USA

Posted on 08.25.2009 - 18:27:05 PST
Kronos what blather, redox potentials, the Nernst equation. Wine making isn't electrochemistry. You're looking at responses to autoxidation processes in the wine. Are you an engineer? The independent variables of chemistry are: time, temperature, pressure, and concentration.
Chicago, IL USA

Posted on 08.20.2009 - 02:02:11 PST
In 200 words or less? The subject of oxygen in winemaking and bottling has been covered extensively by J Ribereau-Gayon and colleagues and is described in the various editions of Traite d'OEnologie I have published a number of technical articles on and around this subject in the last few decades (plus one next month) in the
Australian Grapegrower&Winemaker, in particular "Oxidation serendipity and SO2" (1996) & "Sporadic oxidation of packaged wine"(1998). Over the last few years I have had a number of friendly exchanges of e-mails with Jamie Goode about these matters.
INACCURACIES SO2 does not "dissociate" into free and bound forms - oxygen is not transmitted through corks or Stelvin wads, but it can diffuse between the closure and the glass -there are more wad and gasket formulations than tin/Saran and Saranex - Mocon is not an industry standard for corks, (plastic films, yes) - Mocon has no relevance to the ability to act as a barrier to atmospheric oxygen - Lopes' work shows an e
North Ryde, NE Australia

Posted on 08.20.2009 - 17:05:34 PST
(continued) Lopes'work shows an initialeffusion of some of the oxygen compressed in the cork's cells into the wine - the solubility figure of 8 mg of O2/L is in air at 20C - the "robustness" of addition compounds formed with phenolics - the figures used in the shelf-life estimates are unrealistic - "reduction" is used as a euphemism for SLO (sulfide-like odours) -Twin Top and Diam are denser than natural corks and make a tighter seal, and like natural corks there is no penetration of atmospheric O2 while they are compressed in the neck of the bottle - in AWRI trials,Diam's three levels of OTR gave identical performance (another marketing gimmick) -lower grade corks (those with higher lenticel volumes) exclude atmospheric oxygen and seal just as effectively as higher grades. They don't look as nice and maybe they decline in performance sooner (years). I could go on a bit more, but I don't want to monopolise the site.
North Ryde, NE Australia