Sulfur dioxide is a traditional, inexpensive additive used widely at many different stages of modern wine making to combat oxidation and microbial spoilage. It can be used pre-fermentation to control microbial populations that come in on grapes and the oxidation of juice, post-fermentation to protect wine from oxidation and spoilage during aging, for storage of barrels to prevent microbial spoilage, and even as a general cleaning agent in the winery (sprays on surfaces, etc…).

When used in the winery, SO₂ is most often in a liquid form. When SO₂ dissolves in water, it doesn’t remain only as SO₂. Instead, it interacts with the water in a way that takes three forms: molecular sulfur, bisulfite and sulfite. Figure 1 gives a graphical representation of the relationship between pH and the forms of sulfur dioxide. Between pH of 3 and 4, which describes most wine, bisulfite is the predominant form of sulfur dioxide while molecular sulfur and sulfite are scarce¹. The balance of forms is important because not all forms of sulfur dioxide have the same activity. 

Figure 1: Forms of sulfur dioxide in solution are pH dependent. Image from Rotter n.d².

 

Molecular sulfur (SO₂) is prized because of its antimicrobial activities. As the only form of sulfur dioxide that is not charged, molecular sulfur can penetrate the cell membranes of microbes and cause cellular damage and death.^1,3,4 Molecular SO₂ is also volatile, which means this is the form that can cause negative sensory aromas in the headspace of wine and can be lost during aging as it diffuses into the headspace of barrels and dissipates¹. The volatility of molecular SO₂ is also the basis for separation from the wine matrix during aeration oxidation testing, and is the form detected in the headspace gas detection tube method of analysis.

The bisulfite form of sulfur dioxide (HSO₃^-) dominates at wine pH. Bisulfite is a potent inhibitor of enzymes that causes enzymatic browning in juice and wine⁴. However, bisulfite binds to many constituents in the wine including acetaldehyde, anthocyanins, and sugars^1,3,4. Once bound, it is no longer active as an antioxidant. This binding is not irreversible, however. Under very acidic conditions and high heat, bisulfite is released. This is the basis of total SO₂ testing, but can be erroneously counted as free when acid is used during free SO₂ testing such as aeration oxidation and Ripper titration.

SO₂ testing is one of the most common analytical tests performed in wineries, with addition rates relying on the results of these tests. Most wineries test for both free and total sulfur dioxide. Free SO₂ refers to any form of SO₂ in wine that is not bound to another molecule and gives an indication of the antimicrobial and antioxidant protection conveyed to the wine. Total sulfur includes both the free and bound fractions. The legal limit for total SO₂ is 350 ppm, however most wineries limit total SO₂ to 200 ppm for sensory reasons^4,5. Free SO₂ targets depend on the antimicrobial and antioxidant goals. A general rule of thumb is to maintain a molecular SO₂ between 0.5 and 0.8 mg/L for antimicrobial properties^6,7. It is also recommended to keep free SO₂ above 10 ppm in red wines and 20 ppm in white wines to avoid oxidation³. Because the bisulfite form of SO₂ binds readily to many common constituents of wine, free SO₂ cannot be calculated from SO₂ additions but rather must be measured. There are several methods currently in use for measuring free SO₂. The principles behind these tests and potential problems with each test are the topic of the remainder of this newsletter. For more detail into the chemistry of SO₂, target SO₂ levels and management strategies, see the July 2020 Newsletter on SO₂ Management.

Figure 2: Forms of free and bound sulfur dioxide in wine. From Zoecklein⁸

---

References

(1) Boulton, R.; Singleton, V. L.; Bisson, L. F.; Kunkee, R. E. Principles and Practices in Winemaking; Chapman and Hall, Inc: New York, 1996.

(2) Rotter, B. Sulfur Dioxide. Improved Winemaking: advanced theory, practical solutions and opinions.

(3) Ribereau-Gayon, P.; Dubourdieu, D.; Doneche, B.; Lonvaud, A. Handbook of Enology Volume 1: The Microbiology of Wine and Vinifications, 2nd ed.; John Wiley & Sons: West Sussex, England, 2006.

(4) Zoecklein, D. B. Sulfur Dioxide (SO2). Enology Notes Downloads, 16.

(5) Stamp, C. How Much SO2 to Add and When. Wines and Vines 2011.

(6)        Margalit, Y. Concepts in Wine Chemistry, 3rd ed.; The Wine Appreciation Guild LTD: San Francisco, California, 2012.

(7)        Cojocaru, G.; Antoce, O. Chemical and Biochemical Mechanisms of Preservatives Used in Wine: A Review. Sci. Pap. Ser. B: Hortic. 2015, 56, 457–466.

(8)        Zoecklein, B. W. Sulfur Dioxide: Science behind This Antimicrobial, Antioxidant, Wine Additive. Practical Winery and Vineyard Journal 2009.

Free SO₂ testing is likely the most common laboratory test performed in wineries regardless of size, with addition rates determined based on the result. In a 2019 survey of winemakers attending a WRE seminar on SO₂ management, most Virginia wineries (90% of respondents) are measuring SO₂ in-house using aeration oxidation (50%), Ripper titration (30%), or autotitration (20%), which is a modification of the Ripper method. Unfortunately free SO₂ tests do not come with standard solutions because we cannot fix SO₂ in the free form, eventually it will bind up and no longer be free. So how reliable are our results? Which of the available methods is best, and can a winemaker be sure he/she is working with a reliable number? In this newsletter we will examine the different methods for measuring free SO₂ in a small to medium sized winery, discuss quality control measures and pros and cons of each approach. Specific protocols may differ among wineries, but these basic principles will apply broadly. For specific protocols to use in your own winery, we recommend those found in “Chemical Analysis of Grapes and Wine” (Iland et al 2004) and Wine Analysis and Production (Zoecklein et al 1995), though several textbooks and lab manuals will contain good examples.

 

Regardless of which method you use, it is important to use the appropriate measurement device. Any lab method has a limit to how robust it is to changes to the procedure¹. Some steps require very careful measurement while others do not. If the protocol calls for dilution with a volumetric flask, this tells you that there is little room for error in this step. Volumetric pipettes are more accurate than transfer pipettes, and transfer pipettes are far better than graduated cylinders. It is worth the limited expense to purchase the appropriate measurement device in order to ensure a more accurate and precise result.

 

Ripper Titration 

The Ripper method relies on a redox reaction between sulfur dioxide and iodine. The wine is first acidulated with sulfuric acid to reduce cross reaction of iodine with wine polyphenols². Starch is added as an indicator that will turn color when it complexes with iodine. When Iodine is initially titrated into the wine/acid/starch mixture, sulfur dioxide from the wine reacts with iodine before it can complex with the starch, so there is no color change. When all of the sulfur dioxide has been complexed with iodine, excess iodine reacts with the starch to form an inky black end product, indicating the end to the titration. Based on the volume of iodine used, the amount of sulfur dioxide can be calculated with a simple conversion factor based on the concentration of iodine^2–4.

There are several elements to the Ripper titration that make it attractive for the small to medium sized winery. The entire procedure if very fast, it requires very little equipment (a burette, a flask, a pipette and a desk light, Figure 1), and it is very inexpensive (Table 1). However, there are several known inaccuracies with this approach. Once acidulated, SO₂ is volatile and can be lost from the liquid/air interface during swirling⁴. Determination of the endpoint by a color change can be difficult in deeply colored red wines. Most importantly, there are other compounds in the wine matrix that also bind to iodine, most notably phenols, sugars, aldehydes, and ascorbic acid. This leads to an overall overestimation of SO₂ present^2,3.

If you are using the Ripper titration, there are some steps to take to make it as accurate as possible. A pinch of bicarbonate can be added to blanket the liquid/air interface with CO₂, making volatilization of SO₂ and dissolution of O₂ less likely^2,3. A small desk lamp can be used to provide backlight to the titration, making endpoint determination easier. Due to the instability of iodine, the concentration of this solution should be checked weekly, or potassium iodate could be used in its place to create iodine in situ⁵.

General QC recommendations also include 

1. Store iodine in a dark cabinet in dark glass bottle since it is light sensitive or the refrigerator to avoid microbial contamination². 
2. Standardize the iodine solution as it will degrade over time (with an estimated shelf life of 3 months when stored properly). Iland et al 2004 has a good protocol for this.
3. Though some kits supply a syringe for titration of iodine, it is better to use a microburette with a narrow bore opening to fine tune the titration.
4. The endpoint is reached when the color changes to a dark blue and persists for 30 seconds.
5. Work quickly (within 2 minutes) to avoid volatilization of SO₂ at low pH and dissociation of SO₂ bound to anthocyanins^2,5

ORP detection methods

Several manufacturers, including Vinmetrica and Hanna Instruments, have developed oxidation/reduction sensing probes that detect a change in the current of the wine in the presence of excess iondine⁶. Sensing redox potential as opposed to a color change allows a better determination of the reaction endpoint in darkly colored wines². The Hanna Instruments autotitrator also reduces human error by automating the titration step. However, since these methods still rely on the Ripper method, they are still subject to interferences from other wine components and iodine should be used promptly or standardized.

Aeration Oxidation

The aeration oxidation method seeks to avoid interferences by other components of the wine matrix by separating SO₂ from the wine prior to measurement. During AO, acid is added to wine which is then subjected to either negative pressure (vacuum) or positive pressure (bubbling air) that separates molecular SO₂ from the liquid. This gaseous SO₂ passes through tubing to a second receptacle where it is trapped by a hydrogen peroxide solution, forming sulfuric acid. Because the wine is initially acidulated, most of the SO₂ is in the molecular form and therefore measurable. After a measured amount of time (dependent on flow rate), the amount of sulfuric acid in the peroxide trap is determined by titration with NaOH. Though this methodology still relies on a color change to determine the end of titration, the peroxide solution is clear and therefore the endpoint is easier to read than in the Ripper titration^2,3. However, several other difficulties remain

Aeration oxidation requires a more elaborate setup than the Ripper titration (Figure 1) with higher initial cost and longer reaction time (15 minutes per test). It still relies on a trained technician to accurately determine the endpoint, and care must be taken to keep solutions up to date. Because this method relies on separation by air movement, the flow rate of air through the wine must be calibrated, and the reaction must be carefully timed as too much or too little will affect the resulting value. When setting up the reaction, all tubing must be secure as any leaks will cause SO₂ to escape, leading to aberrantly low values^2–4.

General QC recommendations also include:

1. Standardize NaOH often. Some recommend daily³, some say weekly⁴. Iland et al (204) recommend purchasing 0.1 M NaOH, which is more stable, and diluting to 0.01 N NaOH as needed, however this dilution must be done volumetrically to avoid error. NaOH can also be made from dry pellets at less cost and then standardized^2–4.
2. Hydrogen peroxide also degrades quickly in air, so a stock solution of 30% peroxide should be purchased from a trusted chemical supplier and stored in a dark bottle away from sunlight. A working solution of 0.3% should be made up at the time of use^2,3.
3. Store NaOH and Hydrogen peroxide in the refrigerator to slow degradation, but make sure working solutions are used at room temperature³. 
4. All testing should occur at cellar temperature³.
5. Do not use NaOH that has been sitting in the burette overnight or for several hours. Good practice is to drain the burette at the end of each testing session. When refilling, make sure to flush any remaining solution completely, refill, and make sure to remove any air bubbles before titrating.
6. The endpoint is reached when the magenta solution shows the first appearance of olive green. If you get to bright green, you have gone too far³.
7. The flow rate of air should be 1L/minute.

Spectrophotometer based methods

In recent years, a third category of free SO₂ determination has become available to the medium sized winery. This methodology relies on a quantitative reaction of bisulfite in the wine with fucsin and formaldehyde to produce a colored compound that can be measured with a spectrophotometer*^5,7. A standard curve is built using standard solutions allowing for calibration within the protocol. Many medium sized wineries use spectrophotometers for measurement of volatile acidity, malic acid, residual sugar and juice nutrients, so this methodology adds one more function to this instrument. The reaction itself has been condensed into kit form by several manufacturers (Unitech, Vinessential, Megazyme and others) and can be run with an automated protocol on instruments such as the Chemwell. Several QC protocols are included in the kit procedure including sample blanking to avoid absorbance by polyphenols and pigments and use of a second reading to determine true SO₂ vs. other oxidizable compounds that have reacted with reagents. 

This methodology has been shown in internal documentation to be within 7ppm of traditional measures (Ripper and AO), with average differences of 1.5ppm for Chardonnay and 5ppm for red wines, with lower variation than either traditional metod⁵. However, this study was done on an automated spectrophotometer, which comes at considerable cost ($11,000-25,000 at the time of this writing). Package inserts from Megazyme and Vintessential also taut close correlation (r²=0.999) with results from AO measures of the same wine, leading to average differences of 1 mg/L (SD 4 mg/L). These kits can be used with a manual spectrophotometer, allowing several reactions to be processed in one run (decreasing time per test), however, manual operation includes the need for a trained technician as well as introducing potential error from pipetting, timing, cuvette cleanliness, etc… In a review of laboratory proficiency over 13 years and over 70 laboratories, Howe et al (2015) found spec based methods for free SO₂ determination had very high values for variation. These authors hypothesize high values are due to a combination of factors including poor calibration of the spectrophotometer, pipette calibration, inappropriate cell path length, and dilution errors.

*The Vintessential kit relies on a different reaction than the Unitech and Megazyme kits. This approach includes acidification of the sample, leading to issues of bisulfite dissociation that will be discussed in the “Headspace” portion of this newsletter. Other aspects of the protocol (potential for automation, pipetting error, etc…) will apply to this kit.

Other methodologies

Several other methodologies for detection of sulfur dioxide in wine have been investigated^8–11, however, most are either not yet ready for the production laboratory or they require expensive instrumentation and skilled operators that place them out of reach for small to medium sized wineries. These approaches are exciting developments and will be reported on as they become accessible for Virginia winemakers.

Figure 1: Comparative glassware needs and setup for the Ripper titration and Aeration Oxidation testing ($395) (www.carolinawinesupply.com)(gillypad.weebly.com)

 

Table 1: Comparative cost of materials and supplies for various methods of SO2 determination

---

References

(1) Butzke, C. E.; Ebeler, S. E. Survey of Analytical Methods and Winery Laboratory Proficiency. American Journal of Enology and Viticulture 1999, 50 (4), 461–465.

(2) Zoecklein, D. B. Sulfur Dioxide (SO2). Enology Notes Downloads, 16.

(3) Iland, P.; Bruer, N.; Edwards, G.; Weeks, S.; Wilkes, E. Chemical Analysis of Grapes and Wine; Patrick Iland Wine Promotions PTY LTD: Campbelltown, Australia, 2004.

(4) Buechsenstein, J. W.; Ough, C. S. SO2 Determination by Aeration-Oxidation: A Comparison with Ripper. 1978, 29 (3), 4.

(5) Anderson, G. Free & Total Sulfur Dioxide Measurement. 9.

(6) Vinmetrica SC-100A User Manual Version 3.1c. Vinmetrica.

(7) Megazyme Free Sulfite Assay Procedure Manual. Megazyme 2017.

(8) Luque de Castro, M. D.; González-Rodríguez, J.; Pérez-Juan, P. Analytical Methods in Wineries: Is It Time to Change? Food Reviews International 2005, 21 (2), 231–265. 

(9) Monro, T. M.; Moore, R. L.; Nguyen, M.-C.; Ebendorff-Heidepriem, H.; Skouroumounis, G. K.; Elsey, G. M.; Taylor, D. K. Sensing Free Sulfur Dioxide in Wine. Sensors 2012, 12 (8), 10759–10773. 

(10) Coelho, J. M.; Howe, P. A.; Sacks, G. L. A Headspace Gas Detection Tube Method to Measure SO2 in Wine without Disrupting SO2 Equilibria. American Journal of Enology and Viticulture 2015, 66 (3), 257–265. 

(11) Jenkins, T. W.; Howe, P. A.; Sacks, G. L.; Waterhouse, A. L. Determination of Molecular and “Truly” Free Sulfur Dioxide in Wine: A Comparison of Headspace and Conventional Methods. Am J Enol Vitic. 2020, 71 (3), 222–230.

The most common question asked about SO₂ methodology is some variation of “which is the best one” or “which one should I be using”? In a perfect world, we all want a method that is accurate, precise, robust, and inexpensive. However, the real answer depends on how much SO₂ testing you are doing, how much time and resources you have to devote to the pursuit, and how accurate you need the results to be to make a good decision. Whichever method you choose to implement in your winery, you should always validate the method both at the beginning of use and periodically throughout its usage to ensure you have set things up correctly and nothing has changed after implementation. Even slight modifications to the way a method is applied in a winery can lead to significant changes in outcomes¹.

Which method is the most precise?

As winemakers we are most concerned with the accuracy (the difference between the measured value and the true value) and precision (how well our methodology will produce the same result when performed multiple times)¹ of a testing method. For free SO₂ analysis, it is difficult to determine accuracy because many variables in the wine matrix affect how much SO₂ is bound vs. free, making it very difficult to prepare standards. Instead, comparison to a mean is often used as a benchmark. But beware, if systematic errors exist, like poorly calibrated equipment or inherent biases in the test, the mean can be far from accurate (see the next section on headspace detection methods for molecular SO₂ for a good example). Precision is easier to determine. One often-used measure of precision is the confidence interval (CV), which indicates the range within which the true value of the analysis is thought to fall given the imprecision in repeated testing. CV is calculated by dividing the standard deviation of a group of readings by the mean of that same group². Each of the testing strategies outlined in the previous section (Ripper, aeration oxidation, spectrophotometric) have, to one degree or another, have been developed with the goals of accuracy and precision in mind. However, when tested in research laboratories and in wineries, determination of free SO₂ has been shown to have some challenges.

There are a few reports of side by side testing of free SO₂ testing methods. Buechsenstein and Ough (1978)³compared Ripper and aeration oxidation values for four wines (two reds, two whites) that had been spiked with 50 mg/L, 100 mg/L and 150 mg/L SO₂. They report higher variation in Ripper (CV of 9.5%) than aeration oxidation (CV of 2.6%). In one Pinot Noir, the range of values found for AO was 61.8 – 64.5 mg/L while the range for Ripper was 51.2 – 60.8 mg/L. In an internal Unitech study of their spectrophotometric method, Geoff Anderson (n.d) tested 13 Chardonnays and 12 red wines and found Ripper accuracy to be +/- 5ppm while aeration oxidation accuracy was +/- 2ppm with AO results typically 7ppm lower than Ripper⁴. All spectrophotometric analyses yielded values within 7ppm of traditional values with differences depending on the wine and method. All Chardonnay samples were with 1.5 ppm of Ripper while spectrophotometric methods averaged 5-7 ppm higher in red wines. It is important to note these tests were done on an automated spectrophotometer, minimizing human error in pipetting. It is expected these rates would be higher using the same kits on a manual spectrophotometer.

Results of precision testing from research laboratories may differ from the precision reached in production wineries due to differences in quality control protocols. In 1988, the ASEV Technical Projects Committee undertook an effort to encourage implementation of quality standards in the American wine laboratory testing through proficiency testing¹ that eventually grew into a program run through the Collaborative Testing Services (CTS)⁵(https://collaborative-testing.com). Participating wineries received two bottled sample wines every 4 months to analyze in-house with as many replications as they chose. Resulting values were reported to the testing agency and compared within and between labs. Participants fell into three categories: formally accredited labs with quality control systems in place, laboratories with few quality control systems in place, and those with neither training, experience, nor knowledge of laboratory quality systems. For free SO₂, accuracy is difficult to measure because the amount of free vs. bound SO₂ is often changing,  so results were compared to the grand mean of all reported results^1,5,6. In a review of progress from 2002, Butzke (2002) reported variation in free SO₂ values ranging from 11.8 – 29.4% depending on the wine tested, which could mean ranges as narrow as 16.6 – 23.2 but as wide as 11.8 – 22.1 mg/L (Table 1). In a report on 13 years of collaborative data from this program, including participation from 30-77 participants per cycle, free SO₂ testing in general was shown to have a within-lab variation (CV) of 5.4% but across lab variation of 19.4%. These data included all methodologies, with 44% of participants reporting using aeration oxidation, 27% reporting using Ripper, and 6.4% reporting a colormetric method (the remainder were using segmented flow, flow injection, and enzymatic methods). In general, Ripper values were 2.7 mg/L higher than those reported from aeration oxidation methods (likely due to binding of non-SO₂ components of the wine matrix). At the outset of this program, Butzke and Ebeler (1998) set a goal of 1% CV for each test. At each reporting, variation for free SO₂ values was considerably higher^5,6, and additional measures of quality control were suggested. 

How to validate a method and test proficiency in your winery

No matter which method you choose, it is important to understand the limitations of the approach and validate that method when it is first implemented as well as periodically check back to ensure it is being performed properly. The process of validation will allow you to identify and rectify sources of error as well as better understand how to interpret the values you generate. Unfortunately, many of the usual methods of validation are not applicable to free SO₂ testing. Standard solutions are difficult to make and degrade quickly, and spiking a wine with a known amount of SO₂ will not result in a commensurate increase in free SO₂, as some (unknown) portion will be bound up^1,2. Following are some approaches to determine if the free SO₂determination method you choose is being performed properly:

1. SO₂ standard solutions can be carefully prepared from potassium metabisulfite². Iland et al (2004) recommend preparing a stock solution of 10 g/L SO₂ (using 75 g of reagent grade potassium metabisulfite dissolved in NaOH and diluted in water). The stock solution can then be diluted to working solutions of 20 mg/L, 50 mg/L and 100 mg/L. Testing replicates of 20 mg/L solution should recover no more than 10% variation (with results from 18-22 mg/L). Testing of each will produce a standardization curve, which should be linear.
2. Analyze a number of samples in tandem with a trusted service lab¹. Send topped, sealed bottles in glass (plastic allows oxygen to permeate which may change free SO₂ values) to a trusted service lab, then run the same samples in-house on the day the lab receives them. Do at least three, and up to seven. From the results, you can calculate your accuracy relative to the service lab.
3. To determine precision, run the same sample seven times². Calculate the mean and standard deviation (Excel has a function for this). Divide the standard deviation by the mean to determine the %CV. A value of 10% or less is generally acceptable for winery decision making². (For a value of 20 mg/L, this would equal a range of 18-22 mg/L.) Keep in mind precision depends on the type of wine, so repeat this procedure with red, white, and sweet wines.
4. Repeat #2 on a different day to get an idea of how robust your method is to routine differences. If multiple people run free SO₂ in your winery, test proficiency in each to identify differences and potential errors.

If the resulting values are unacceptably imprecise or not close enough to the service lab values, there are several questions to ask. First, are your values always higher or always lower? This could point to a systematic error, such as improperly calibrated equipment (for example, the flow rate on the AO is too high or too low) or poorly standardized solutions (0.01 M NaOH is really 0.0093 M), as well as consistent errors in endpoint determination (over-titrating will lead to consistently higher numbers). If variation is too high, perhaps more practice is needed with pipetting or titrating. 

Table 1: freeSO₂ values reported from ASEV Laboratory Proficiency Testing from 1998-2001 (reproduced from Butzke 2002)

---

References

(1) Butzke, C. E.; Ebeler, S. E. Survey of Analytical Methods and Winery Laboratory Proficiency. American Journal of Enology and Viticulture 1999, 50 (4), 461–465.

(2) Iland, P.; Bruer, N.; Edwards, G.; Weeks, S.; Wilkes, E. Chemical Analysis of Grapes and Wine; Patrick Iland Wine Promotions PTY LTD: Campbelltown, Australia, 2004.

(3) Buechsenstein, J. W.; Ough, C. S. SO2 Determination by Aeration-Oxidation: A Comparison with Ripper. 1978, 29 (3), 4.

(4) Anderson, G. Free & Total Sulfur Dioxide Measurement. 9.

(5) Howe, P. A.; Ebeler, S. E.; Sacks, G. L. Review of Thirteen Years of CTS Winery Laboratory Collaborative Data. American Journal of Enology and Viticulture 2015, 66 (3), 321–339. 

(6) Butzke, C. E. 2000/2001 Survey of Winery Laboratory Proficiency. 2002, 7.

Recent development of protocols for free SO₂ determination using gas detection tubes has highlighted significant problems with the accuracy of traditional methods. The acidulation of wine in both Ripper and AO testing releases SO₂ that is bound to anthocyanins and measures it as free. This information was reported as early at 1975 by Bourroughs¹, however, corrective steps required several spectrophotometric measurements using solutions not commonly found in wine labs, and therefore, corrections are almost never made. Unfortunately, the amount of overestimation is significant (up to 66%) but depends on the concentration of each form of anthocyanins in the wine, so cannot be directly correlated to values routinely measured in the winery lab.

Experiments with gas detection tubes began with the goal of measuring free SO₂ without interference from the wine matrix (as if found in Ripper) but also without complex and expensive glassware (as is the case with AO). Pegram et al (2013)² developed a method using gas detection tubes to register sulfur in the gas phase. To move SO₂ from the wine to the tube, wine was acidulated to convert SO₂ into the gas phase, then alka-seltzer tablets were used to create a current that delivered SO₂ to the tubes. However, the alka-seltzer tabs also buffered the acidulated solution, leading to a limit in the range of the test. To solve this problem, Coelho et al (2015) retained the use of the gas detection tubes but, instead of acidulation and airflow, they loaded wine into a closed syringe with known headspace, allowed molecular SO₂ in the wine to equilibrate in the headspace, then depressed the plunger to move headspace air through the syringe³. This methodology measures molecular (gaseous) SO₂ directly, from which free SO₂ can be calculated. 

The headspace gas detection tube method correlated well with AO measurements for white and Rosé wines (R² = 0.97) but the correlation with red wines was poor (R² = 0.72), with headspace methods 2-3 times lower than those reported with AO. After investigating sources of systemic error, the researchers found that in traditional methods the acidulation of wine led to release of SO₂ from anthocyanins while in the HS-GDT method this SO₂ remained bound. They confirmed this was the cause of the difference in free SO₂ values by measuring anthocyanins directly, calculating SO₂ binding to various forms of anthocyanins present at that pH, then correlating back to the difference in free SO₂ between AO and HS-GST. They found a very strong correlation (R² = 0.94)³. Other non-perturbing methods also show this decrease in free SO₂ measures.

Further investigation by Howe et al (2018)⁴ has shown that anthocyanin-bisulfite complexes do not appear to have any antimicrobial properties⁴. This group used sweet wines inoculated with EC1118 yeast to determine effective antimicrobial levels of SO₂ using both AO and HS-GDT methods. They tested both white wine and white wine with added anthocyanin (to mimic red wine, but still compare to the white). In white wines, molecular SO₂ values needed to prevent EC1118 growth were similar using either SO₂ determination method, indicating no change is needed in our practice. In the pseudo-red wine however, a 2-log reduction was found with molecular SO₂ of 0.5-2.0 mg/L when measured by HS-GDT while values greater than 2.0 mg/L when measured by AO were still not sufficient to affect cell counts. (In this study, HS-GDT measures of molecular SO₂ were only 37% those found for AO³.) This means that if SO₂ additions are made based on AO test results in red wines, it is unlikely molecular targets are being met and wine is left susceptible to microbial spoilage.

Unfortunately there is not an easy answer to this problem. The HS-GDT methodology is still very new and has yet to be optimized for the commercial winery. The gas detection tubes used in this method were manufactured for the mining industry, not for wine, and only one of three brands tested by the initial researchers worked^2,3, and even these were discolored by the ethanol found in wine. Several sources of error have been identified, including the timing of syringe depression and leaks in the tubing attachments⁵. The manufacturer’s markings must be converted to linear scale, and calculations must be done based on temperature, pH and ethanol to determine free SO₂⁶. To aid in these calculations, researchers at Cornell have developed a downloadable Excel file (HSGDT calculator). Variation in readings is still large enough for the developers to recommend 5 replicate readings per wine, with each tube costing $7.50 and accommodating 2-3 readings each⁵. So far, this method is not automated, though work done by Jenkins et al (2020)⁷ converts the concept of headspace SO₂ detection to an automatable form. If you would like to try this method in your own winery, Dlubac and Sacks (n.d.)⁵ provide sourcing information and step by step instructions for setup. Make sure you validate the method (by comparing values with white wines) with a known reliable reference method to ensure you are getting reliable numbers.

---

References

(1) Burroughs, L. F. DETERMINING FREE SULFUR DIOXIDE IN RED WINE. Am. J. Enol. Viticult. 1975, 26 (1), 5.

(2) Pegram, Z.; Kwasniewski, M. T.; Sacks, G. L. Simplified Method for Free SO ₂ Measurement Using Gas Detection Tubes. Am. J. Enol. Vitic. 2013, 64 (3), 405–410.  

(3) Coelho, J. M.; Howe, P. A.; Sacks, G. L. A Headspace Gas Detection Tube Method to Measure SO2 in Wine without Disrupting SO2 Equilibria. American Journal of Enology and Viticulture 2015, 66 (3), 257–265.  .

(4) Howe, P. A.; Worobo, R.; Sacks, G. L. Conventional Measurements of Sulfur Dioxide (SO2) in Red Wine Overestimate SO2 Antimicrobial Activity. Am J Enol Vitic. 2018, 69 (3), 210–220.  

(5) Dlubac, G.; Sacks, G. L. Measuring Sulfur Dioxide (SO2) in Wine Using a Headspace Gas Detection Tube Method.

(6) Sacks, G. L.; Howe, P. A. “Free” Doesn’t Always Mean Free: Rethinking SO2 Measurements in the Winery. 6.

(7) Jenkins, T. W.; Howe, P. A.; Sacks, G. L.; Waterhouse, A. L. Determination of Molecular and “Truly” Free Sulfur Dioxide in Wine: A Comparison of Headspace and Conventional Methods. Am J Enol Vitic. 2020, 71 (3), 222–230. 

Winemakers from around Virginia gathered (virtually) to discuss results of two experiments. In the first, precision and accuracy of two commonly used SO₂ detection techniques (aeration oxidation and the Hanna titrator) were tested and compared by Phil Fassieu (Whitehall Vineyards), AJ Greely (Hark Vineyards) and Rachel Stinson Vrooman (Stinson Vineyards). In the second, Kirsty Harmon (Blenheim Vineyards) compared the chemical, microbiological, and sensory effects of different SO₂ addition rates after completion of fermentation in Chardonnay.

Watch Video Here