The Australian Wine Research Institute

A wide range of wine faults can be caused by the growth of microorganisms in packaged wine. These include:

• Gas/spritz
• Haze/deposit
• Volatility
• Oxidation
• Bitterness
• Metallic taste
• Medicinal taste (volatile phenols)
• Geranium character
• Reduced acidity

The control of unwanted microorganisms at the time of packaging is therefore very important, so it’s prudent to determine the microbiological status of a wine before bottling. Along with chemical analysis results, the results of microbiological analysis will help guide winemakers about any adjustments or treatments that might be required just before, or at the time of packaging. Typically, microorganisms are controlled by a combination of pH/SO₂ and filtration. However, in the case of sweet wines, an additional treatment may be required due to the higher risk of refermentation if an infection occurs post-sterile filtration. This is especially the case when only a relatively low concentration of free SO₂ can be obtained.

Apart from SO₂, the other main chemical treatments available to the winemaker include dimethyl dicarbonate (DMDC) and sorbic acid. The combination of SO₂ and one of these compounds usually results in effective control of microorganisms. DMDC is classified as a processing aid, whereas sorbic acid is classified as an additive, in the Australia New Zealand Food Standards Code Standard 4.5.1 (Anon 2019).

Dimethyl dicarbonate (DMDC)

Key points:

• Used at bottling to kill wine yeasts
• Maximum legal addition in Australia and New Zealand is 200 mg/L
• Not as effective against bacteria
• Does not have a sensory effect
• Special dosing machine needed due to safety and reactivity issues.

DMDC is an antimicrobial agent which can legally be used in wine in Australia and New Zealand at rates up to 200 mg/L (Anon. 2019). It works by binding with particular amino acid residues located at the catalytic centre of key cellular enzymes, particularly alcohol dehydrogenase and glyceraldehyde-3-phosphate dehydrogenase, deactivating them (Porter and Ough 1982, Portugal et al. 2014). With these enzymes deactivated, glycolysis, which is required for energy production, cannot continue, so the cell dies.

The action of DMDC against microorganisms in wine is rapid (Daudt and Ough 1980); however, it quickly hydrolyses to mainly methanol and carbon dioxide (within 1 hour at 30°C and within 5 hours at 10°C), so should be added to wine in as pure a form as possible (Delfini et al 2002). Consequently, to be most effective at ensuring wine sterility, DMDC should be added just before the filler bowl in the bottling sequence (Renouf et al. 2007). In addition, DMDC must be thoroughly mixed to ensure homogenisation through the wine (Fugelsang and Edwards 2007).

DMDC is an irritant, acutely toxic, corrosive and flammable (Anon. 2020a). Because it is corrosive to skin and eyes and toxic by inhalation and ingestion, certain safety precautions must be taken when handling concentrated DMDC. This includes hand protection, eye protection, skin protection and a respirator if ventilation is inadequate. Detailed precautions for safe handling, conditions for safe storage and first aid measures are given in the product Safety Data Sheet available from suppliers.

Due to its rapid hydrolysis, the need for accurate and homogeneous mixing through the wine, combined with its safety issues, DMDC needs to be added to wine using a metered dosing system (Calisto 1990, Fugelsang and Edwards 2007). Such a dosing system should ensure that the addition of DMDC is well controlled in a way that is consistent with good manufacturing practice (Anon. 2010).

Note that because DMDC is listed as a processing aid in the Australia New Zealand Food Standards Code, there is no requirement for DMDC to be listed on wine labels.

Effectiveness against yeasts
There are several factors that influence the effectiveness of DMDC, including yeast species and strain, initial yeast population, pH, SO₂ concentration, alcoholic strength, temperature and method of addition and homogenisation. As might be expected, DMDC is more effective when initial cell numbers are low, pH is low, alcoholic strength is high and the temperature is just above 20°C (Ough et al. 1978, Porter and Ough 1982, Ough et al 1988, Delfini et al. 2002, Costa et al. 2008). Table 1 summarises results obtained by several researchers on the amount of DMDC needed to kill different strains of yeast at particular levels of contamination/inoculation. The results are mainly for wine, apart from the results reproduced from Renouf et al. (2007), which are for red grape juice.

Table 1. Amount of DMDC needed to kill several different strains of yeasts, as reported by various researchers

Yeast	Contamination/inoculation rate (CFU1/mL)	Rate of DMDC required to kill the yeast (mg/L)	Comments on conditions	Authors
Saccharomyces cerevisiae</td>	<100 (average ~50)	60	Five white wines: pH 3.14−3.4, 10.9−12.9% v/v ethanol, 10 g/L sugar; Five red wines2 : pH 3.03−3.51, 10.2−11.7% v/v ethanol, 10 g/L sugar	Ough (1975)
Saccharomyces cerevisiae Saccharomyces pombe Brettanomyces spp. Dekkera bruxellensis Rhodotorula rubra	17−104	50−100	Wine, 10% v/v ethanol and 20 g/l sugar	Daudt and Ough (1980)
Saccharomyces cerevisiae	100−500	100	Wine, 10% ethanol, 20 g/L sugar, temperature slightly above 20°C	Porter and Ough (1982)
Saccharomyces cerevisiae	330−460	50−75	White wines: 25 mg/L free SO2, pH 3,0−3.6, 10.9% v/v ethanol, 20 g/L sugar	Ough et al. (1988)
Saccharomyces bayanus	580–758	100	White wines, pH 3.0−3.6, 11% v/v ethanol, 12 g/L sugar	Threlfall and Morris (2002)
Brettanomyces bruxellensis (ten strains) Saccharomyces cerevisiae (three strains) Candida cantarelli Candida ethanolica Hanseniaspora uvarum Pichia anomala (two strains) Zygosaccharomyces bailii	1,000	150 100−200 50 100 50 400 250	Red grape juice, pH 3.7, 190 g/L sugar	Renouf et al. (2007)
Saccharomyces cerevisiae (two strains) Dekkera bruxellensis Pichia guilliermondii (three strains) Schizosaccharomyces pombe Zygosaccharomyces bailii Saccharomycodes ludwigii	500	100 100 100 100 25 25	Red wine, pH 3.5, 12% v/v/ ethanol, 2 g/L added sugar	Costa et al. (2008)

Notes:
1. CFU = colony forming units
2. Some viable cells were observed at 60 mg/L DMDC for one of the red wine replicates

Based on the data presented in Table 1, an addition of 100 mg/L of DMDC appears to be sufficient to kill yeast in wine if there are less than 500 viable cells of the yeast per mL. Given the AWRI recommends that sweet wines be sterile filtered (i.e. 0.45 μm membrane filtered) at bottling, the population of any yeast contaminant picked up post-filtration should be less than 500 cells/mL. However, as indicated above, the effectiveness of DMDC is dependent on several factors and these should be considered when deciding how much DMDC to add. Note that DMDC product suppliers indicate the usual dosage for wine is between 125 mg/L and 200 mg/L (Anon. 2020b).

Effectiveness against bacteria
There is consensus in the literature that DMDC is not as effective against wine bacteria as it is against wine yeasts. Table 2 summarises results obtained by some researchers on the amount of DMDC needed to kill wine bacteria at particular levels of contamination/inoculation.

Table 2. Amount of DMDC needed to kill different bacteria reported by various researchers

Bacteria	Contamination/inoculation rate (CFU1/mL)	Rate of DMDC required to kill the bacteria (mg/L)	Comments on conditions	Authors
Acetobacter aceti Lactobacillus sp.	5×106	>200 >200	Synthetic nutrient media	Delfini et al. (2002)
Lactobacillus brevis Lactobacillus buchneri Lactobacillus pasteurianus Pediococcus cerevisiae	500	200 30 300 300	Artificial medium, pH 2.8–4.7, 28°C	Golden et al. (2005)
Acetobacter pasteurianus	250	190–250	Artificial medium	Fugelsang and Edwards (2007)
Oenococcus oeni Acetic acid bacterium	500	>300 >300	Red wine, pH 3.5, 12% v/v/ ethanol, 2 g/L added sugar	Costa et al. (2008)

Notes:
1. CFU = colony forming units

Based on the data presented in Table 2, DMDC at the maximum legally permitted dose of 200 mg/L cannot be solely relied upon to control post-bottling growth of lactic acid and acetic acid bacteria. Control of lactic acid bacteria can usually be achieved by judicious use of SO₂ (in combination with pH), while acetic acid bacteria can be controlled by minimising dissolved oxygen at bottling and using a closure with a low oxygen transmission rate, as acetic acid bacteria are aerobic organisms.

Sensory aspects
Ough (1975) investigated the effect of DMDC on samples of a neutral white wine and established that additions of DMDC up to 200 mg/L were undetectable by a trained tasting panel.

Methanol
As indicated above, DMDC hydrolyses to mainly methanol and carbon dioxide. Wines naturally contain some methanol due to the hydrolysis of methylated pectin, which is formed in grapes by endogenous pectin methylesterase (PME). Pectinases added to juice can also increase the concentration of methanol in wine. Although methanol is a natural component in wine, regulatory authorities set limits for its presence due to its toxicity. While methanol is a toxic substance, Stafford et al. (1976) calculated that an adult weighing 70 kg would need to consume 70 L of wine with a relatively high methanol concentration (340 mg/L) in order to consume a lethal dose of methanol, suggesting that at the typical concentrations found in wine (see below), methanol toxicity due to wine consumption is of insignificant risk to humans (Hodsen et al. 2017).

The Australia New Zealand Food Standards Code (Standard 1.4.1) prescribes limits for methanol in Schedule 19 (Maximum levels of contaminants and natural toxicants), which are summarised in Table 3 below (Anon. 2017).

Table 3. Summary of limits for methanol in wine from the Australia New Zealand Food Standards Code

Product	Maximum level of methanol
Red wine, white wine and fortified wine	3 g of methanol per litre of ethanol
Whisky, rum, gin and vodka	0.4 g of methanol per litre of ethanol
Other spirits, fruit wine, vegetable wine and mead	8 g of methanol per litre of ethanol

Given the maximum concentration of methanol permitted is specified per litre of ethanol, the maximum permitted methanol concentration in a wine will vary depending on its alcoholic strength. Some examples of the maximum permitted concentration of methanol in terms of mg of methanol per L of wine, based on the alcoholic strength of the wine, are provided in Table 4. Note that the AWRI’s winemaking calculator for methanol expressed as proportion of ethanol can be used to determine whether a wine’s methanol level is within the legal limits.

Table 4. Examples of maximum permitted concentration of methanol in wines of different alcoholic strength

Alcoholic strength of the wine (%v/v)	Maximum permitted methanol concentration allowed in the wine (mg/L)
6	180
8	240
10	300
12	360
14	420
16	480
18	540

Hodson et al. (2017) surveyed 60 white and 90 red Australian wines from multiple varieties and vintages and found the mean concentrations of methanol to be 58 mg/L and 170 mg/L, respectively. The theoretical maximum concentration of methanol that could be formed in a wine after the maximum allowed addition of 200 mg/L DMDC is 96 mg/L (Stafford and Ough 1976). Based on the average concentrations of methanol reported by Hodson et al. (2017) and the maximum concentration of methanol that could form from the maximum allowed addition of DMDC, it is highly unlikely that a DMDC addition to a typical table wine would lead to the wine exceeding legal methanol limits. However, it should be noted that wines produced from grapes affected by Botrytis cinerea can have higher concentrations of methanol than other wines (up to 364 mg/L) due to PME produced by this invasive organism. It is therefore recommended that such wines be tested for their methanol concentration before any DMDC is added. This will allow the winemaker to calculate the maximum methanol concentration expected after DMDC addition, so that the legal limits for methanol can be adhered to.

Sorbic acid

Key points

• Used just prior to packaging sweet white wines to prevent refermentation by Saccharomyces cerevisiae
• Maximum legal addition in Australia and New Zealand is 200 mg/L as sorbic acid
• More effective at lower pH, higher alcohol levels and low yeast populations
• Must avoid growth of lactic acid bacteria to avoid the development of devastating geranium off-odour

Sorbic acid, or 2,4-hexadienoic acid, is an unsaturated, short-chain fatty acid which can be used to help prevent post-packaging refermentation by yeast, mainly Saccharomyces sp. (Fugelsang and Edwards 2007). Sorbic acid works by suppressing the oxidative phosphorylation enzyme system (Tromp and Agenbach 1981), which is used to produce energy for the cell. Unlike DMDC, which is toxic to yeast (especially Saccharomyces sp.), sorbic acid is considered a fungistat rather than a fungicide, as it inactivates yeasts but does not kill them. However, sorbic acid is persistent in wine, so its activity does not decrease over time (Boulton et al. 1996). Sorbic acid is listed as a wine additive in the Australia New Zealand Food Standards Code and can be added at rates up to a maximum of 200 mg/L, expressed as sorbic acid (Anon. 2019). Unlike DMDC, which is considered a processing aid, there is a requirement for sorbic acid to be listed on the labelled product.

Effectiveness
The effectiveness of sorbic acid depends on wine composition (mainly pH, alcoholic strength and SO₂ concentration), the type of microorganism and the strain. Sorbic acid is more effective with increasing concentration of SO₂, increasing concentration of ethanol and at lower pH values (Ough and Ingraham 1960, Ribéreau-Gayon et al. 2006, Zoecklein et al., 1995). With regards to microorganisms, sorbic acid is mainly used to inhibit refermentation by Saccharomyces cerevisiae in sweet wines (Fugelsang and Edwards 2007, Zoecklein et al., 1995), as it is not a very effective at inhibiting yeasts such as Brettanomyces, Saccharomycodes and Zygosaccharomyces, lactic acid or acetic acid bacteria (du Toit and Pretorius 2000, Edinger and Splittstoesser 1986). In general, the recommended rate of addition is between 100 and 200 mg/L (Auerbach 1959, Fugelsang and Edwards 2007, Ough and Ingraham 1960, Zoecklein et al. 1995); however, other authors recommend the maximum rate (200 mg/L) as the effective dose (Boulton et al. 1996, Tromp and Agenbach 1981). These authors recommend that 200 mg/L be used regardless of the yeast cell numbers and that sorbic acid should be used in combination with SO₂ and sterile filtration (Boulton et al. 1996, Tromp and Agenbach 1981).

Addition to wine
Sorbic acid has relatively low solubility in water (1.6 g/L at 20°C), so the potassium salt (potassium sorbate) is used instead, which is very soluble in water (Ribéreau-Gayon et al. 2006). When calculating how much potassium sorbate to add to a wine, the value for the desired concentration of sorbic acid must be multiplied by the ratio of the molecular weight of potassium sorbate to that of sorbic acid, which is 1.34. For example, if a sorbic acid concentration of 200 mg/L is required, 1.34 x 200 mg/L = 268 mg/L of potassium sorbate must be added to achieve this. Note that you can use the AWRI winemaking calculator for sorbic acid addition to calculate the amount of potassium sorbate to add. The potassium sorbate should be dissolved in water prior to addition to the bulk wine. Because the pKa of sorbic acid is 4.7, when potassium sorbate is added to wine (pH typically <3.8) it changes from the dissociated form to the undissociated form (sorbic acid) soon after addition (Zoecklein et al. 1995). However, because sorbic acid has low solubility, it can precipitate if the potassium sorbate solution is added too quickly, so the solution of potassium sorbate must be added slowly while the wine is mixed (Ribéreau-Gayon et al. 2006). Due to this, Ribéreau-Gayon et al. (2006) recommend using a dosing pump. It is also recommended that sorbic acid be added just prior to bottling (Fugelsang and Edwards 2007, Ribéreau-Gayon et al. 2006).

Sensory aspects
It has been reported that some tasters can detect sorbic acid in wine (Ough and Ingraham 1960), while Tromp and Agenbach (1981) reported that the threshold for sorbic acid is much higher than the legal limit of 200 mg/L. However, Ribéreau-Gayon et al. (2006) indicate that freshly prepared solutions of sorbic acid have no odour and that sorbic acid does not affect wine aroma when correctly prepared and added to wine which is then stored properly. These authors attribute unpleasant odours (not including geranium off-odour, which is discussed below) to the oxidation of sorbic acid when exposed to air. Because sorbic acid is an unsaturated fatty acid (i.e. it contains reactive double bonds), it can be oxidised in air to give molecules containing aldehyde functional groups, which can impart off-odours. Ribéreau-Gayon et al. (2006) indicate that concentrated aqueous solutions of sorbic acid turn yellow after exposure to air and exhibit pungent odours, which is evidence of the instability of sorbic acid under aerobic conditions.

Geranium off-odour
As indicated above, sorbic acid has very little activity against lactic acid bacteria. In fact, certain lactic acid bacteria can metabolise sorbic acid, ultimately forming an unpleasant odour similar to that of crushed geranium leaves (Sponholz 1993), which is practically impossible to remove from wine. The odour remains after carbon treatment, is present in spirits after distillation and is still perceived after significant dilution (Ribéreau-Gayon et al. 2006). Consequently, it is imperative that wine conditions (the combination of pH, SO₂ and ethanol concentrations) are such that growth of lactic acid bacteria is strongly inhibited and that sorbic acid is not used to treat red wines. More information about geranium off-odour can be found on the faults, taints and flavours page.

References and further reading

Anonymous. 2010. Food Standards Australia New Zealand. Application A1025 – Classification of Dimethyl Dicarbonate. Available from: https://www.foodstandards.gov.au/code/applications/documents/A1025%20DMDC%20AppR%20FINAL.pdf

Anonymous. 2017. Food Standards Australia New Zealand. Australia New Zealand Food Standards Code – Standard 1.4.1 – Contaminants and Natural Toxicants, Schedule 19. Canberra: FSANZ. Available from: https://www.legislation.gov.au/Series/F2015L00454

Anonymous. 2019. Food Standards Australia New Zealand. Australia New Zealand Food Standards Code – Standard 4.5.1 – Wine production requirements. Canberra: FSANZ. Available from: https://www.legislation.gov.au/Series/F2008B00809

Anonymous. 2020a. National Center for Biotechnology Information. PubChem Database. Dimethyl dicarbonate, CID=3086. [monograph online] Accessed June 18, 2020, from https://pubchem.ncbi.nlm.nih.gov/compound/Dimethyl-dicarbonate

Anonymous. 2020b. Lanxess Emerging Chemistry. Velcorin® Spectrum of use, The best choice for wines. [monograph online] Retrieved 15 June 2020, from https://velcorin.com/velcorin/spectrum-of-use/velcorin-wine/

Auerbach, R.C., 1959. Sorbic acid as a preservative agent in wine. Wines Vines 40(8): 26-28.

Australia New Zealand Food Standards Code – Standard 1.4.1 – Contaminants and Natural Toxicants, Schedule 19. Canberra: FSANZ. Available from: https://www.legislation.gov.au/Series/F2015L00454

Boulton, R.B., Singleton, V.L., Bisson, L.F., Kunkee, R.E. 1996. Principles and practices of winemaking. New York, USA: Chapman & Hall: 362.

Calisto, M.C. 1990. DMDC’s role in bottle stability. Wines Vines 71 (10), 18–21.

Chisholm, M.G., Samuels, J.M. 1992. Determination of the impact of the metabolites of sorbic acid on the odor of a spoiled red wine. J. Agric. Food Chem. 40(4): 630-633.

Costa, A., Barata, A., Malfeito-Ferreira, M., Loureiro, V. 2008. Evaluation of the inhibitory effect of dimethyl dicarbonate (DMDC) against wine microorganisms. Food Microbiol. 25(2): 422-427.

Crowell, E.A., Guymon, J.F. 1975. Wine constituents arising from sorbic acid addition, and identification of 2-ethoxyhexa-3,5-diene as source of geranium-like off-odor. Am. J. Enol. Vitic. 26(2): 97–102.

Daudt, C. E., Ough, C. S. 1980. Action of dimethyldicarbonate on various yeasts. Am. J. Enol. Vitic. 31(1), 21-23.

Delfini, C., Gaia, P., Schellino, R., Strano, M., Pagliara, A. and Ambrò, S. 2002. Fermentability of grape must after inhibition with dimethyl dicarbonate (DMDC). J. Agric. Food Chem. 50(20): 5605−5611.

Du Toit, M. and Pretorius, I.S. 2000. Microbial spoilage and preservation of wine: using weapons for nature’s own arsenal. S. Afr. J. Enol. Vitic. Vol. 21: 74−96.

Edinger, W.D., Splittstoesser, D.F. 1986. Production by lactic acid bacteria of sorbic alcohol, the precursor of the geranium odor compound. Am. J. Enol. Vitic. 37(1): 34−38.

Fugelsang, K., Edwards, C. 2007. Wine Microbiology, Practical Applications and Procedures. Springer Science & Business Media: 70−74.

Golden, D. A., Worobo, R. W., Ough, C. S. 2005. Dimethyl Dicarbonate and Diethyl Dicarbonate. Davidson, P. M., Sofos, J. N. and Branen, A. L. (eds) Antimicrobials in food. 3rd Edition CRC Press, Boca Raton, Florida, USA: 305–326.

Hodson, G., Wilkes, E., Azevedo, S. and Battaglene, T. 2017. Methanol in wine. BIO Web Conf. 9: 02028.

Ough, C.S. 1975. Dimethyldicarbonate as a wine sterilant. Am. J. Enol. Vitic. 26(3): 130−133.

Ough, C.S., Ingraham, J.L. 1960. Use of sorbic acid and sulfur dioxide in sweet table wines. Am. J. Enol. Vitic. 11(3): 117−122.

Ough, C.S., Langbehn, L.L. and Stafford, P.A. 1978. Influence of pH and ethanol on the effectiveness of dimethyl dicarbonate in controlling yeast growth in model wine systems. Am. J. Enol.Vitic. 29(1): 60−62.

Ough, C. S., Kunkee, R. E., Vilas, M. R., Bordeu, E., Huang, M. C. 1988. The interaction of sulfur dioxide, pH, and dimethyl dicarbonate on the growth of Saccharomyces cerevisiae Montrachet and Leuconostoc oenos MCW. Am. J. Enol.Vitic. 39(4): 279−282.

Porter, L.J., Ough, C.S. 1982. The effects of ethanol, temperature, and dimethyl dicarbonate on viability of Saccharomyces cerevisiae Montrachet No. 522 in wine. Am. J. Enol. Vitic. 33(4): 222-225.

Portugal, C., Sáenz, Y., Rojo-Bezares, B., Zarazaga, M., Torres, C., Cacho, J., Ruiz-Larrea, F. 2014. Brettanomyces susceptibility to antimicrobial agents used in winemaking: in vitro and practical approaches. Eur. Food Res. Technol. 238(4): 641-652.

Ribéreau-Gayon, P., Dubourdieu, D., Donèche, B., Lonvaud, A. (eds.) 2006. Handbook of Enology, Volume 1: The microbiology of wine and vinifications. John Wiley & Sons: 224−229.

Renouf, V., Strehaiano, P., Lonvaud-Funel, A. 2007. Effectiveness of dimethlydicarbonate to prevent Brettanomyces bruxellensis growth in wine. Food Contr. 19(2): 208-216.

Stafford, P. A., Ough, C. S. 1976. Formation of methanol and ethyl methyl carbonate by dimethyl dicarbonate in wine and model solutions. Am. J. Enol. Vitic. 27(1): 7-11.

Zoecklein, B.W., Fugelsang, K.C., Gump, B.H., Nury, F.S. 1995. Wine analysis and production. New York, NY, USA: Chapman & Hall: 209-212.