When it comes to sulfite management, winemakers need reliable analysis of Free SO₂ concentrations to make informed decisions that maximize wine quality. Selecting the right SO2 measurement solution requires consideration of not only the device and method itself, but also the entire measurement and data management process. This includes maintaining a reliable calibration, understanding the limitations of the analysis method, sample collection and processing, data entry, and risks of operator errors. Here are some tips from us on assessing the myriad of options for Free SO2 analysis. These concepts can also be applied to considering and comparing other wine analysis technologies.

1. Repeatability, Repeatability, Repeatability

Also referred to as precision, repeatability is the ability of a device and method to produce consistent and reliable results under the same testing conditions. This should always be the first thing to look at when evaluating a chemical analysis technology. The analogy of throwing darts at a dartboard can help illustrate the difference between accuracy and precision.

Precision (Repeatability) vs. Accuracy

A highly precise method, with high repeatability, will lead to the same decision in the same circumstances. This provides a stable platform for fine tuning the accuracy of a measurement through proper calibration. You can calibrate out an accuracy issue but no amount of calibration fixes  poor precision.

The precision of a measurement method can be quantified a few different ways. This isn’t the time or place for a statistics lecture, but we’ve tried to provide some helpful working definitions here, and you can download our Excel template to calculate these parameters automatically and help with your evaluation process.

  • Maximum deviation: the difference between the minimum and maximum result for repeats of the same sample. This can help give an intuitive understanding of the method’s repeatability, in a useful working way.
  • Standard deviation: This is a measure of how much the results vary from the average value of the repeats.

Comparing a measurement method goes beyond a couple of statistics and there is more to consider. We cover a few of the biggest ones in the rest of this article.

2. Establish an accuracy baseline

Establishing an accurate baseline to compare with the new technology is essential. Your on-site lab or local third-party labs can do the trick. However, it's crucial to approach these comparisons with caution –  how do you know the baseline you've chosen is sufficiently accurate? Consider the known systematic differences between measurement methods. For instance, when measuring Free SO₂, the Ripper Method and Aeration Oxidation would be expected to yield different Free SO2 values for the same sample.

No measurement is perfect, and it is important to remember that you’re comparing two different methods to one another, not comparing the new method to the “truth”. You can use our Excel template to make a comparison plot like the one above. These plots are super helpful in seeing the behavior of two instruments side by side. Does Instrument B systematically over-predict relative to Instrument A (lots of measurement points above the line)? Is the behavior the same in both the high and low ranges of the instruments?  

3. Test in a variety of conditions

Wine matrix effects can vary significantly by variety, alcohol content, extraction, wine age, oxidation state etc. Different methods behave in different ways. For example, some optical methods are very sensitive to color and turbidity, Ripper-based methods can be sensitive to acid and certain sugars. The plot below shows an example of comparing two measurement methods for a selection of very different wines that span the gamut of color intensity, sugar, alcohol, and turbidity.

Make sure to test on samples that are representative and informative of how you’ll be using the instrument for production decisions!

4. Try to use a measurement standard

If there is any reason to question the accuracy of your baseline, create a few standard solutions of known quantities of SO₂ in water (for example, you could make a 10 ppm, 20 ppm, and 40 ppm standard) and measure it with both the baseline and new technology. This can help flush out any systematic variances between measurement methods. Measurement standards have been used for thousands of years to improve accuracy and agreement, you should use one too!

5. Be conscious of sample collection and handling

If you are comparing two methods of analysis, you need to be sure any differences are because of the analysis and not differences in the wine samples being measured. Therefore, it is also essential to consider how the sample was collected and the time between sampling and analysis. For example, if the sample is too small or is kept in a container with a headspace, Free SO2 can be consumed quickly. This can affect the results, particularly if the analysis occurs at a later point in time. Ensure that the sample used for comparisons is drawn from the exact same source. If collecting samples from a barrel, the depth from which the sample was drawn can be a source of measurement variability due to SO2 stratification inside the barrel. When comparing a new instrument with your baseline, do your best to ensure that the exact same sample is being analyzed and, to the best of your ability, at the same time.

6. Evaluate under practical conditions at your wineries

Another important factor in evaluating data quality is to think about the technology in the context of your workflow, and identify potential steps in the process that leave room for errors. This goes well beyond the idealized performance figures that are provided by a manufacturer or that come from a quick lab test.

  • What steps are not automated - sample handling, data handling, or labeling?
  • What steps require specialized skills or undivided attention?
  • How frequently does calibration need to be done by trained technicians?
  • Does the equipment have built-in safeguards to protect the measurement integrity from human and environmental errors? For example, routine self-checks, vendor-side monitoring, etc.
  • Are any reagents or supplies used in the process perishable?

Additionally, issues are bound to arise with any new technology rollout or workflow change - what is your plan to address them?

  • What support systems does the vendor provide (and will the vendor support be fast enough to matter)? Anything that saves you and your team from self-troubleshooting and waiting around will go a long way, especially during busy times.
  • Is the data tracked in a way that can help with troubleshooting? Having a single source of truth is a must, and any extra detail adding context to the measurements is a plus (for example, if the measurement is taken as a spot check after sulfite addition). Generally, a software platform is faster and has fewer problems than paper and spreadsheets.

7. Give it a try!

Ultimately, the best way to evaluate new technology is to test it out for yourself. Many vendors provide at-winery demonstrations or even risk-free trial periods. These offer the opportunity to involve staff from multiple teams, including winemaking, lab, and cellar, to provide a comprehensive assessment of the system's pros and cons. You can also get a sense of the intangible aspects of a new system, such as how intuitive it is for staff to learn and how difficult the system is to maintain.

Closing Notes

New technologies are emerging to help winemakers make better decisions and improve wine quality. While assessing them can be a daunting task, it's essential to take steps to ensure that the data produced is reliable and accurate. When done right, investing in new technology can lead to more consistent wine quality and more efficient winery processes, making it a worthwhile endeavor for wineries that want to stay ahead of the curve.