The premature oxidative ageing of wine
Dr Valérie Lavigne and Prof Denis Dubourdieu
Faculty of Oenology
Bordeaux Institute of Vineyard and Wine Sciences



Ageing potential, i.e. a wine's ability to develop its unique character over time, is an essential factor in determining a wine's intrinsic quality. But what does "ageing well" actually mean? In my opinion, it is comparable to human beings, and means seeming "reasonably" younger than one's age and, above all, being more interesting than several years previous!

Wine is undoubtedly the most cultural and symbolic of beverages. People identify easily with what they eat, but not yet what with what they drink... In 1998, the sociologist Claude Fischler wrote,"We are completely and somewhat mystically convinced that we 'are what we eat', and that there is a kind of identity that links us to what we ingest".  When we enjoy a wine with a taste that is typical of its origin, we can be said, in a way, to drink in the culture and history of the place it comes from, as well as the philosophy and personality of the people who made it. When we savour an old vintage that has aged well, we magically "absorb" and appropriate its amazing ability to defy the passage of time and delay its "irreparable outrage". This is surely one of the main reasons why expensive wines are necessarily ageworthy wines. This is also why the disappointment is all the more acute when wine enthusiasts discover that certain wines in their cellar have aged badly. In such instances, they feel as though they have been cheated somehow. 

Defective ageing, usually associated with premature oxidation (frequently shortened to "premox"), describes both a wine that

as aged before its time as well as the whole issue of prematurely aged wines in general. When people open a bottle of wine these days, the question seems to be: "Premox or not premox?". This subject raises so many questions and feeds so many rumours and controversies that Google has 15,000 pages of hits for the word "premox"!

The purpose of this talk is to present our work on premature ageing and our conclusions regarding the characteristic odorous molecules in the aroma profile of prematurely old wines – as well as how they are formed, the factors from the vineyard to the bottle that cause this aberration and, obviously, the best means to prevent it.

Premox in white wines

A white wine with ageing potential has the ability, over time, to maintain the aromas of the grape variety (or varieties) from which it has been made and to develop specific aromatic nuances (burnt, mineral and truffle aromas) that constitute the "bouquet" of great wines.

Unfortunately, such ideal ageing is far from common among white wines... Most lose their fruity aromas quickly and develop heavier ones reminiscent of honey, beeswax, moth balls and pine resin. This aromatic development is always accompanied by an increased orangey-yellow color and an impression of bitterness on the aftertaste.

Such "oxidative ageing" takes away a great deal from the wine's personality. It negates the terroir as surely as reductive ageing highlights it.

Obviously, premox is most detrimental when it affects white wines purported to have fine ageing potential, such as great white Burgundy. However, it would be wrong to think that this problem affects white Burgundy more than other wines because of some mysterious influences found only in that region. In fact, premox affects all white wines, still and sparkling, dry and sweet, and all grape varieties and origins. The issue is identical for white wines everywhere, with the same causes and the same effects.
Our studies concerning premox began in the early 2000s. They are now practically finished. Here are the main results and conclusions.

Sotolon, a universal marker of premox in white wines, is formed from acetaldehyde.

Based on the similarity between the odour of honey and white wines affected by premox, extracts of premoxed wines and honey were both subjected to gas chromatographic analysis as well as olfactory analysis (by the human nose) and analysis with a mass spectrometer.

The aromagrams show two shared aroma sectors (ZO1 and ZO2). The first is reminiscent of honey and the second of beeswax. So, describing the odour of premoxed wines as resembling that of honey and beeswax is certainly justified. It is now clear that certain molecules present in honey and beeswax are also to be found in the aroma of wines suffering from premox.

Prior to our studies, 3 compounds had already been identified in prematurely aged dry white wines: methional, phenylacetaldehyde and o-aminoacetophenone.

Methional, which has an odour of boiled potatoes, is formed by the breakdown of methionine.

Phenylacetaldehyde, with an aroma of wilted roses and honey, comes from pheylalanine, and is also due to oxidation.

2-amino acetophenone, with an odour of moth balls and wax polish, was first identified in German white wines that had fallen victim to irregular ageing odeur or Untypische alterung note (UTA). This is formed by the oxidative degradation of indolacetic acid.

However, these molecules cannot alone account for defective aromatic ageing because many premoxed wines do not contain them, or their content is less than the odour detection threshold.

On the other hand, we have clearly shown the role of sotolon in the aroma of all premoxed white wines.

Sotolon is a volatile heterocyclic compound with an intense curry aroma.  Other than its role in the taste of Sherry and vins jaunes from the Jura, this compound also contributes in an important way to the dried fig and rancio aromas of vins doux naturels (VDN) and Port.

We have established that the odour detection threshold in dry white wines is 7 µg/L.

In fact, the sotolon molecule has an asymmetrical carbon atom. Sotolon thus has two enantiomers (R and S). Each of these has different olfactory characteristics. The odour detection threshold of the S form in hydroalcoholic solutions is 0.8 µg/L, and that of the R form is 89 µg/L. The S form is thus far more odoriferous than the R form. The powerful aroma of (S)-sotolon is reminiscent of curry and walnuts, and that of (R)-sotolon smells only of rancid walnuts. Therefore, the S form of sotolon is the one that contributes to the odour of dry white wines affected by premox. The distribution of the enantiomers of sotolon varies from one wine to another. In certain wines, the R and S forms are in equal proportions (racemic mixture), whereas others display a strong concentration of one or the other. Sotolon's contribution to the premox characteristics of a wine depend on its level of S-sotolon.

Sotolon is formed by an aldocondensation reaction between α-ketobutyric acid, present in all wines, and acetaldehyde, whose concentration is increased by oxidation. It is thus easy to see why is it is absolutely essential for winemakers to avoid acetaldehyde at all stages of white wine production.

Glutathione, which is naturally found in grapes, provides a natural defence mechanism against the premature ageing of white wines.
What compounds are likely to protect wines from premature oxidation? For red wines, the answer is polyphenols (tannins and anthocyanins), which are strongly reductive and generally amply present in wines with good ageing potential. What are the naturally reductive compounds that protect white wines, with a low level of phenolic compounds, from oxidation? We have shown that certain peptides or sulphur amino acids and, in particular, glutathione play a key role.

Glutathione is a tripeptide, and a natural constituent of many plants and foods. Its role in protecting against the oxidation as well as detoxifying cells is well known. Like many plants, grapes also have a large concentration of glutathione, several hundred mg/L.
A strong link exists between the glutathione level in musts and their level of available nitrogen. Vines that are too weak, and with a poor nitrogen intake, produce grapes low in glutathione. Summer drought conditions and/or competition from grass left to grow between the vine rows also worsens this deficit.

Glutathione is a very reactive compound. When the juice is extracted, a large part of the glutathione in the must is transformed via oxidation into disulphide of glutathione and/or combined with quinones, produced by the enzymatic oxidation of phenolic compounds in the must, to form GRP. However, glutathione remains in the must before fermentation begins – from a few milligrammes to approximately 20 mg per litre.

The glutathione level decreases at the beginning of alcoholic fermentation, but increases again at the end and during the first month of barrel ageing on the lees. The yeast uses the available glutathione in the must during its growth phase, then releases it at the end of alcoholic fermentation and the early stages of autolysis.

There is a strong correlation between the must's initial glutathione level and that found in the wine at the beginning of barrel ageing. Furthermore, the concentration of glutathione in the must and the wine at the end of AF are also similar.

These results have been confirmed in the laboratory, using a model medium. The initial concentration of glutathione in the medium is very similar to that found at the end of AF, provided that alcoholic fermentation took place normally, and especially thanks to a sufficient level of nitrogen in the must.

In our tests, the glutathione level was established at 20 mg/L and available nitrogen varied from 48 to 190 mg/L. When the medium lacked nitrogen (48 and 96 mg/L), fermentation was sluggish. The quantity of glutathione released by the yeast at the end of AF was negligible. On the other hand, when fermentation took place normally thanks to a suitable amount of nitrogen, the initial and final levels of glutathione in the medium were comparable. In other words, at the end of AF, the yeast "returns" glutathione to the medium that it "borrowed" at the beginning.

The premature ageing of dry white wines may appear either during barrel ageing or later, when the wine is cellared in bottle.
Keeping wine on the lees during barrel ageing retains fruity aromas and mainatins them in an oxidation-reduction state that encourages a reductive bouquet as well as avoiding, or delaying, the appearance of defective "honey-type" aromas.

We followed the simultaneous development of fruity aromas (volatile thiols), markers of premature aromatic ageing (sotolon) and glutathione during the ageing of a Sauvignon Blanc wine kept in new and used barrels, either on its lees or without them. The most oxidative ageing conditions, i.e., new barrels without lees, tended to decrease the fruity aroma, in this instance grapefruit, in the wine (3-MH). The concentration of sotolon is greater when the wine is kept separate from its lees, especially if aged in new oak barrels.

Also, in instances when the lees are removed from the wine, the glutathione level quickly decreases during barrel ageing. This phenomenon is accentuated by new barrels, where oxidative influences are greater. Therefore, the ageing conditions most favourable to retaining the aromatic qualities of dry white wines are ones that limit the decrease in the concentration of glutathione.
The role the lees have in protecting the wine from defective ageing is due to two reasons: their ability to preserve glutathione and their capacity to consume oxygen. If one compares a white wine on its lees and the same wine after filtration, the importance of lees in oxygen consumption is clear. A wine on its lees consumes oxygen 60,000 times more quickly than one without lees. The reductive capacity, and thus the protective capacity of the lees is considerable.

Of course, the defective ageing of white wines is most often discovered once the wine has been bottled. The most disconcerting aspect of premox in bottled wine is its very random nature. If a tasting jury is asked to evaluate the oxidative character of twelve bottles of the same dry white wine with a cork closure (same vintage, bottled the same day, with the same lot of corks, etc.), certain bottles may be found to be very aged, whereas others may show no sign of premature ageing whatsoever.

There is a very close correlation between the quantity of oxygen dissolved in the bottle and a wine's propensity to suffer from premox. Furthermore, the more a wine's aromas are judged to be oxidised by tasters, the more intense the wine's orangey-yellow nuances (higher DO 420). Premoxed bottles have lower levels of free SO2. Naturally, the presence of sotolon is also closely correlated with the intensity of the oxidative character detected during tasting.

These results raise the very tricky issue of the best closure to use, which is too involved a subject to go into here. Significant variations in the dissolved oxygen level in wine have been observed only six months after bottling according to the type of closure used (natural cork, synthetic cork, technical corks, or screwcaps). The level of free SO2 is also very different depending on the type of closure. These results illustrate the extreme heterogeneity of permeability to gas of various closures.

Preventing the oxidative ageing of white wines requires maintaining their level of glutathione.

Doing so calls for 8 relatively simple rules once one has understood the principle:

1- Ensuring that vines are sufficiently vigorous thanks to a nitrogen intake in keeping with their needs – Grapes with plenty of available nitrogen (> 200mg/L) always have a higher concentration of aroma precursors and glutathione, and a lower concentration of phenolic compounds. Excessive yields, drought conditions, competition from grass left to grow between the vine rows and superficial rooting are all factors in weakening vine growth and therefore in promoting premox in white wines.

2- Limiting the extraction of phenolic compounds during pressing to preserve glutathione – The number of times the cake is crumbled must be limited and the selection of juice at the end of pressing very strict.

3- Protecting must and wines efficiently by using inert gas and sulphur dioxide

4- Making sure that alcoholic fermentation takes place relatively quickly and is completely finished – This implies a level of must clarification adapted to the grape variety, a sufficient level of available nitrogen and just the right amount of oxygen during the phase when the yeast multiply.

5- Reducing the time lag before malolactic fermentation – At this stage of winemaking, the wine is no yet protected from oxidation by sulphur dioxide. Stirring lees or inoculating with malolactic bacteria can prove to be effective.

6- Ageing wines in as reductive an environment as possible by maintaining an efficient dose of free SO2, stirring the lees and the moderate use of new oak.

7- Limiting the dissolution of oxygen when preparing the wine for bottling After racking, fining and filtration, the wine is in a fragile state and must be transferred in an inert gas environment.

8- Choosing a closure that is suited to the wine...