Revisiting sulfur dioxide use

The AWRI’s problem-solving service has recently seen a number of samples submitted with problems apparently related to insufficient use of the preservative sulfur dioxide (SO2). These wine samples have presented with a range of ‘symptoms’, including sensory deterioration, premature oxidation, microbial spoilage and high concentrations of volatile acidity (VA). All have been traced back to reveal insufficient SO2 at bottling as a major contributing factor. So, while SO2 usage is a fundamental aspect of winemaking, and one to which chapters of textbooks have been devoted, it seems appropriate to revisit this topic with a focus on practices that can drastically reduce the incidence of the types of problems described above.

The properties, functions and uses of SO2 wine are outlined below:

  • SO2 has both antioxidant and antimicrobial properties, making it an extremely effective preservative for wine.
  • SO2 exists in free and bound forms within wine – but it is the free SO2 that has the important antioxidant and antimicrobial properties. Therefore, when measuring the SO2 concentration of a wine, it is the free SO2 concentration that is generally regarded as the most important (Rankine, 1989).
  • The amount of SO2 in the free form, and the effectiveness of the free SO2, depends on the pH of the wine. The higher the pH, the less SO2 will be in the useful free form AND the less effective this free SO2 will be. (This second factor relates to the concentration of ‘molecular’ SO2 that is present in wine at various free SO2 concentrations and pH levels, and will not be further discussed in this article). Therefore, all other things being equal, at a higher pH not only will more SO2 need to be added to achieve the desired level of free SO2 (see Table 1 below), but the concentration of free SO2 required to have the desired antioxidant and antimicrobial properties will in itself also be greater.
  • When SO2 is added to wine, as a guide only about 35 to 40% is yielded in the free form (depending on pH and other factors). Multiple small additions of SO2 to a wine rather than one equivalent large addition may result in the concentration of free SO2 never reaching a level where it confers a substantive antioxidant or anti microbial effect. Larger less frequent SO2 additions, yielding greater concentrations of free SO2, are therefore considered much more effective in achieving the desired antioxidant and antimicrobial effects. It is apparent that many wines submitted for problem solving investigation are those to which several small SO2 additions have been made, and many of these wines have a comparatively high ratio of bound to free SO2, i.e. the free SO2 concentration being 25% or less of the concentration of total SO2. In most cases this phenomenon is apparently due to products of oxidation and microbiological activity resulting from the low free SO2 concentrations, acting to bind a large proportion of the SO2 that is present.
  • A corollary of this is that the ratio of the free and bound forms of SO2 can provide important information about the state of both white and red wines. SO2 binds very strongly to acetaldehyde and microorganisms present in wine. Thus if a wine contains a very high proportion of bound SO2 compared to free SO2, it can be an indicator that oxidation or microbial spoilage has occurred, and if a change in this ratio during the maturation of a wine is identified, then the causes for the change should be investigated immediately.

Bearing these properties in mind, what factors can be taken into account when determining how much SO2 a particular wine needs? The levels recommended by Rankine (1989) for both white and red table wines provide a good starting point (Table 1). Based on the AWRI’s problem solving experience, these values should be considered as the minimum concentrations required to prevent oxidation and to inhibit microbial growth.

Table 1. Free SO2 levels recommended by Rankine


Wine type pH Concentration mg/L
White table wines pH 3.00 to 3.20 0 to 20 mg/L free SO2
pH 3.21 to 3.40 20 to 30 mg/L free SO2
pH 3.41 to 3.50 30 to 50 mg/L free SO2
pH > 3.50 50 plus mg/L free SO2
Red table wines pH 3.40 to 3.60 10 to 20 mg/L ‘apparent1′ free SO2
or 50 to 150 mg/L total SO2
pH > 3.60 > 20 mg/L ‘apparent’ free SO2 or
> 150 mg/L total SO2

1. Rankine (1989) points out that the ‘apparent’ free SO2 in red wines is predominantly loosely bound to anthocyanin pigments, and so does not have true chemical significance, but still can be considered a useful measure.

As a general rule, the style of wine being produced also needs to be considered when SO2 concentrations at packaging are determined. All other things being equal, if a wine is intended for medium to long term storage it will require a greater concentration of SO2 than a wine intended for consumption over a shorter time frame. However, it should be remembered that wine packed in bladder packs, for instance, may be intended for early consumption, but may also lose SO2 rapidly after packing due to the oxygen permeability of the bladder material.

An additional consideration is the clarity of the wine. If a wine is hazy, free SO2 in the wine is likely to be rapidly bound up and therefore rendered ineffective at preventing oxidation and microbial growth. Thus, all other things being equal, adding SO2 to a more clarified wine is likely to result in a greater proportion of the SO2 addition being yielded in the useful free form. However, this should not be taken to imply that if a wine is hazy then SO2 should not be added, rather, a high degree of clarity should be maintained throughout the maturation period. Clarification (usually filtration or centrifugation) of hazy wine before bottling is therefore strongly recommended in order to achieve the highest possible concentration of free SO2 at bottling. At all stages of winemaking it is very important to ascertain what a haze in wine consists of, and if a haze proves to contain viable microorganisms, filtration should be considered essential.

Finally, it should be remembered that the concentration of dissolved oxygen (DO) in a wine increases during the bottling process. In the AWRI’s experience, the concentration of DO in wine increases by about 0.5 mg/L during bottling under ideal, industry best practice conditions, and will commonly be in the range of 1.0 to 1.5 mg/L when conditions are less than ideal. Boulton et al. (1996) indicate that, theoretically, 1 mg/L of oxygen can consume 4 mg/L of SO2, so this immediate loss of free SO2 due to DO at bottling should also be taken into account when levels of SO2 at bottling are determined. Additionally, it is considered good practice to measure the SO2 concentration in a wine both pre- and at bottling, and perhaps two weeks post-bottling, preferably on multiple bottles, as a written record can be extremely useful should a problem arise later. Opportunities for DO pickup at other stages in the winemaking process should also be identified and managed, and in the experience of Industry services staff, which has been confirmed during many conversations on the subject with winemakers and wine bottlers, at least 5 mg/L of free SO2 can be lost during even a careful wine transfer. It should also be remembered that the lower the temperature of the wine when it is transferred, the higher its capacity to absorb oxygen, and therefore wine may be more prone to loss of SO2 and subsequent oxidation when it is colder, although it should also be noted that oxidation reactions progress more slowly at lower temperatures.

The AWRI’s closure trial (AWRI publication #666) provides some interesting data correlating SO2 concentration with the sensorial attributes of a Semillon wine. In this trial it was found that samples with a concentration of free SO2 below approximately 10 mg/L were rated high in the attribute oxidised. Conversely, samples with a concentration of free SO2 above approximately 10 mg/L were rated considerably lower in the attribute oxidised. Godden (Technical Review #139) later indicated that browning and deterioration of the Semillon wine’s sensorial attributes accelerated when the concentration of free SO2 fell below this apparently ‘critical’ level of 10 mg/L. The Semillon wine used in this study had a pH of 3.1. Analysing the data from these trials, we might expect other wines to show similar deterioration if the concentration of free SO2 falls below a particular ‘critical’ level, specific for each particular wine. We might also expect wines with higher pH values than the Semillon wine used in the closure trial study to deteriorate at higher ‘critical’ levels of free SO2. Supporting this hypothesis, results from recent analyses of multiple bottles of a 2000 Chardonnay with a pH of 3.57 suggest that the ‘critical’ free SO2 concentration for this wine, below which browning and sensorial deterioration of the wine accelerated, was approximately 15 mg/L.

The consequences of insufficient SO2 seen recently in samples submitted to the Institute have been quite varied, but have always had what could be considered severe ramifications in terms of wine quality and potential longevity of the wines. Recent examples include a white wine demonstrating a rapid sensory deterioration post-bottling (probably due to DO pick-up), and a Pinot Noir wine bottled with free SO2 of 8 mg/L and pH of 3.97, which developed extremely high turbidity and a high concentration of VA, apparently due to microbial activity. In some cases wines have been recalled from overseas markets due to spoilage that can be linked back to insufficient SO2. The corresponding financial losses, as well as losses of good name and of potential markets are very high and could have ramifications for the Australian industry as a whole.

Importantly, the Institute is not advocating that all Australian winemakers increase their use of SO2, although this would have been recommended in many of the cases referred to above. Rather, the Institute seeks to promote an understanding of the interactions between SO2 and other wine components in order to maximise the effectiveness of the concentrations of SO2 currently being used in many wineries. It should also be noted that the upper legal limits of SO2 concentrations in Australian wines are: 250 mg/L for wines containing less than 35 g/L of residual sugar, and 300 mg/L for wine containing in excess of 35 g/L of residual sugar. However, the reluctance of winemakers to increase SO2 usage because of possible adverse reactions by consumers, or a perceived consumer demand for lower SO2 concentrations, is understood, as is the notion that large additions of SO2 to red wine can have a negative sensory effect. However, in many cases, negative effects such as the bleaching of colour are transitory, and are far outweighed by the negative sensory effect of subsequent wine spoilage.

Understanding the properties of sulfur dioxide is an essential element of good winemaking and the relationship between SO2 and pH is important when determining the correct concentration of SO2 for each wine. By taking into account the factors discussed above, and by optimising wine conditions in order to maximise the effectiveness of the SO2 additions that had been made, the wide range of problems recently observed by the problem-solving service could most likely have been prevented. While wine instability resulting from insufficient SO2 is obviously not a new phenomenon, the Industry Services team is concerned that the nature of recent investigations might be indicative of a growing trend, and we would like to ‘nip this in the bud’. Team members are happy to discuss SO2 use with winemakers, particularly with regard to wine bottling. More detailed information about the chemistry of SO2 can be found in the references below, and in many other winemaking textbooks.


Boulton, R.B., Singleton, V.L., Bisson, L.F., Kunkee, R.E. 1996. Principles and practices of winemaking. New York: Chapman & Hall: 459.

Godden, P., Francis, L., Field, J., Gishen, M., Coulter, A., Valente, P., Høj, P., Robinson, E. 2001. Wine bottle closures: physical characteristics and effect on composition and sensory properties of a Semillon wine. 1. Performance up to 20 months post-bottling. Aust. J. Grape Wine Res. 7: 64-105.

Godden, P.W. 2002. Update on the AWRI trial of the technical performance of various types of wine bottle closure. Tech. Rev. (139): 6-10.

Rankine, B.C. 1986. Making good wine: a manual of winemaking practice for Australia and New Zealand. Melbourne: Sun Books: 108, 197.