Wine faults during production and ageing of red wines; Brett and other issues



Brettonamyces / Dekkera

Volatile phenols responsible for the 'barnyard' aroma in wines are:
vinyl-4-gaiacol and vinyl-4-phenol (white wines)
Ethyl-4-phenol and ethyl-4-gaiacol (red wines)

Precursors for brett can be acides cinnamiques (no smell in the grapes)

Acide p-coumarique for ethyl+vinyl 4-phenol
Acide ferulique for ethyl+vinyl 4-gaiacol

Perception threshold whites 130-180 micrograms/litre
Perception threshold reds around 425 micrograms/litre

Study in late 1990s of classified Bordeaux reds found 36% had threshold quantities or higher. Dubourdieu believes it would be at least the same now, probably higher.

Danger can come from:
- leaving rotten grapes around at harvest, so potential for cross-infection
- leaving wine on the lees for too long without sufficient protection
- insufficiently careful sulphur regime
- incomplete fermentation
- waiting for malolactic fermentation to begin (without sufficient protection)
- barrel ageing (particularly used barrels)
-  warm temperatures encourage proliferation, so summer is a risk period

The yeast brett is the only element responsible for this phenolé character in wines (barnyard etc). Infection from lactobacillus can be inhibited by tannins, but brett is not inhibited by tannins.

Around 300mg/l of residual sugar left in wine after Af and MLF is enough for 1000 cells of brett per millilitre to develop - which is sufficient for the threshold perception of (approx) 450 micrograms/litre.

Bdx wines (in fact almost all dry wines following EU regs) are dry if have under 400mg/l of residual sugar = plenty are at risk (particularly if high pH)

Used barrels have higher incidence, but all barrels more at risk than vats.

Plus higher temperatures more risk, so ageing in barrels over the summer be careful!

Average bdx red wines have around 425mg/l of sugar left after fermentation - not perceptibly sweet, but a threat for brett development.

- barrel hygiene - long washing of barrels with very hot water and then steam cleaning, plus SO2 dose of 18-22mg/l (under 18mg/l SO2 dose = poor protection against brett)
- ensure alcoholic fermentation is complete and that malo starts rapidly
- respecting SO2 rules (according to pH of wine, looking for around 18mg of free SO2) and temperature control.

2) over-ripe grapes and premox in red wines
Wine faults can manifest themselves in must and wine from grapes of over-ripe wines.

Researchers at the university of Bordeaux have recently identified two markers that indicate potential issues:

Y-Nonalactone (Z03)

Giving smell of cooked/stewed fruit

Seuil de perception 60 micrograms per litre

Can come from overripe grapes, exacerbated by toast of oak (plus new oak contains higher levels of the molecule) and oxygen exchange from barrel ageing.

3-methyl-2-4 nonanedione (ZOI) - MND

Giving smell of prunes (pruneux)

Seuil de perception 16 ng/l

These two molecules are markers of potential ageing issues in red wines.

Premature ageing of red wines aromas; new markers found

From ISVV Dubourdieu and Lavigne, and Seguin Moreau researchers

'Dubourdieu: I believe one day this will be a massive scandal, we wont learn about it fully for another decade, by which time 15 vintages will have been affected. Right Bank worst affected because merlot is so vulnerable to it. If Bordeaux is giving away its natural benefits of high acidity and ability to age, it is giving away it's raison d'etre, because this is the only place in the world that can produce this style of wine. 'Why try to be cuit et phenolé?'says Dubourdieu.


The quality of great red wines is intrinsically linked to their ageing ability; their aromatic personality that develops over through bottle ageing and reveals their place of origin.

Red wines are of course less sensitive to oxygen then whites but they are not without risk of premature oxidation. The premox of red wines comes through the appearance of certain aroma markers such as prunes, stewed fruits and dried figs, and are oftrn linked also to a rapid evolution in colour. The issue arises in must and wines made from over-ripe grapes, or in wines that are still undergoing ageing, or that are already in bottle.

The tram at ISVV has already been behind a groundbreaking study into premature oxidation in red wines, and applied the same methodology here - gas chromatography to isolate first aromas and then individual molecules, then tests to determine their thresholds of perception to understand their contribution to overall markers or premox.

They found two specifc odour characteristics - ZO1 close to prune, and ZO2 close to stewed fruit. These two aromas have been identified in red wines, and the same molecules are also present in prunes, so it is reasonable to assume that the same molecules are responsible for the smell of prematurely aged wines and that of prunes.

At the same time, the aromagram of wines made from passerillissage grapes also present the same aroma zones as those of prematurely oxidised reds. On tasting, wines made from dried grapes have strong (and desired) notes of prune. So it seems that the state of ripeness of grapes have an influence on the appearance of this fault in dry reds. Wines made from dried out grapes or overly ripe grapes can show a rapid evolution in their aromas.

Certain winemaking practises will worsen the likelihood of premox. The use of too much new oak barrels for example, low doses of sulphur dioxide particularly when coupled with high pH (low acidity) - both will encourage oxidative reactions within the wine, and therefore the development of stewed fruit and prune smells. 

The effect of oxygen:
The team at the ISVV were able to test and verify in the laboratory the role oxygen plays on the formation of molecules responsible for premox aromas through triangular tests. Ine red wine totally unaffected by oxidation was split into two samples A and B. Sample A was immediately put through gas chromotography while sample B was left in the presence of oxygen (in controlled conditions). After 15 days, lot B was extracted as sample A had been and was found to be significantly more marked by prune/stewed fruit aromas.

Both ZO1 and ZO2 appeared in the presence of oxygen.

ZO2 (ZO3) was identified as y-nonalactone - a molecule found in peach, apricot, prune, depending on intensity (varied between 2 micrograms-40 micrograms. Important aroma marker in old vins doux naturel, and in older beers. Perception threshold 60 micrograms/l

Establishing the correct ripeness of a grape is tough - clearly there are degrees of maturity, but no definitive physiological limit for each degree (effect grape, terroir, wine desired etc). However, the influence of y-nonolactone on wines made from extremely ripe grapes seemed important to establish.

So, took two plots of merlot grapes from saint emilion and bordeaux, separated them into extremely ripe and non-ripe grapes.

From both origins, the over ripe grapes had three times higher levels of y-nonolactone (ZI2/3), but below the threshold of perception (27 mivpcrograms/l). So could not be contributing to oxidative character. However, was not impossible to imagine that it could contribute to this aroma in certain circumstances - specifically use of new barrels - and at least the study had proved that there was a correllation between ripeness and y-nanolactone.

Samples were taken out of wine aged in new oak, and older oak barrels (1 year and 2 years) for six months. Found definite correlation, but again just below threshold of perception. Levels of y-nanolactone in used barrels was lower.

Molecule associated with MND / ZO1 -3-methyl-2,4-nonanedione.

This molecule was identified for the 1st time in wine in this reasarch, but had already been identified in soya, parsley and spinach, as well as green tea. Certain acides gras furanique had been identified as precursors in soya oil. The origin on MND un red wines is not yet known.

Aroma varies with concentration - aniseed and hay at strong concentration and prune at lower concentrations. Seuil 63 nanograms/litre

2011 study looked into effect of levels of this molecule in red wines.

Same wine, different vintages, checking levels -found older and/or warmer vintages had higher levels. Found higher levels corresponded to prune aroma. And found no wines without premox had levels over seuil.

Also found that levels were similar for first few years in 1st and 2nd wines of the same property, but that levels went up far more rapidly in second wines than first. So, 1st wine ages more slowly, is more resistant to oxidation (clearly 1st wine lots are chosen specifically for their ability to age- older vines, better terroir). But checking for levels of MND allowed researchers to verify this analytically.

They had shown that exposure to oxygen led to the appearance of this molecule, found that after 8 days exposure to oxygen, levels of MND were 3x higher than the temoin/control, and surpassed threshold of perception.

So, found both y-nonolactone and MND contributed to red wine premox. Levels of y-nonolactone rarely above threshold, but MND was fairly frequently above this level, and had a direct impact on the pruney character of wines. Kevels influenced by over ripeness and oxidation. Now working on how these two parameters are affecting the appearance of premox in reds.

(nb for premox in white wines, excellent paper by Dubourdieu and Lavigne – specifically on the role of glutathions):