Mount Vernon Brew Club

UV Exposure, Glass, and the "Skunky" Smell

Sejin Kim, Jason Bennett, Craig Bickle

Tuesday, April 30, 2024

When malt liquors are exposed to ultraviolet, or UV, light, it can change some of the chemical compounds in the beer and cause it to smell "skunky" or like wet cardboard. We wanted to explore this effect in more depth.

Background

Isohumulones, a type of iso-α-acid, are light sensitive. When they are exposed to UV light, they are broken apart into reactive intermediates known as free radicals, which forms 3-methylbut-2-ene-1-thiol.  Chemists will recognize thiol as a sulfur-containing compound, explaining the skunky or wet cardboard smell that is often associated with lightstruck beer. A very informative article in Chemistry — A European Journal Vol. 7 Issue 21 p. 4553-4561 provides a lot more detail on the exact reactions that are occurring.

Some beers are known for their skunking, such as Corona and Heineken imported into the United States. Corona is typically packaged in clear bottles, and Heineken is typically packaged in green bottles. Corona acknowledges (and even leans into) the skunky aroma, suggesting that you serve with a lime wedge to mask some of the odors.

Testing Questions

We wanted to test the differences that glass color had on the detectability of these thiols when beer was exposed to UV light. To test these differences, we posited the following questions.

  • Is it true that green glass lets through as much UV light as clear glass?
  • Is the difference between green glass and brown glass or between green glass and clear glass detectable?
  • What are the qualitative differences? Are there predominant aromas, tastes, or textures that change from skunked beer versus non-skunked beer beyond just the thiols?

Testing Procedure

For the test, we used the Dortmunder Gold Lager from Great Lakes Brewing Company, Cleveland, Ohio. Dortmunder Gold was chosen as a reliable-tasting beer that most participants were familiar with. Great Lakes Brewing Company (based on the fact sheet) uses Mt. Hood (4.0-7.0% AA) and Cascade (4.5-8.9% AA) hops, along with 2-row base malt and caramel 60, yielding 30 IBU and 5.8% ABV.

To prepare the test, Dortmunder Gold Lager (batch number 38741A, expiration 06/18/24) originally packaged in cans was repackaged in a sanitized Corny keg for dispensing into bottles in a light-controlled room. It was then dispensed using a counter-pressure bottle filler into three identical style bottles: one clear, one green, and one brown. The bottles were then stored refrigerated at 42 degrees Fahrenheit (5.5 degrees Celsius) for one day (sunrise 6:35 AM, sunset 8:41 PM) exposed to daylight through a vacuum-sealed, clear, non-low-E, glass window.

Between 9 attendees, two sets of triangle tests were performed using identical plastic tasting cups: one testing whether differences were detectable between green and brown glass, and one testing whether differences were detectable between green and clear glass. Testers were blindfolded, but they were allowed to talk with each other. Testers also qualitatively identified differences between the two samples, and their notes are included.

Triangle Test Results

In all statistical hypothesis tests, a confidence level of 0.10 was used.

In the first triangle test on whether differences could be found between clear and brown glass, four out of nine testers were able to successfully identify the "odd one out". We ran a one-sided exact binomial test, where the alternative hypothesis is that the true probability of correctly identifying the odd one out (from random chance) is greater than 50%. In this test, we got a p-value of 0.7461 and a 95% confidence interval of 0.1687505 to 1.00. 

In the second triangle test on whether differences could be found between green and brown glass, four out of nine testers were able to successfully identify the "odd one out". Again, we ran a one-sided exact binomial test, where the alternative hypothesis is that the true probability of correctly identifying the odd one out is greater than 50%. In this test, we got a p-value of 0.7461 and a 95% confidence interval of 0.1687505 to 1.00.

In the third triangle test on whether differences could be found between green and clear glass, four out of 11 testers were able to successfully identify the "odd one out". Again, we ran a one-sided exact binomial test, where the alternative hypothesis is that the true probability of correctly identifying the odd one out is greater than 50%. In this test, we got a p-value of 0.8281 and a 95% confidence interval of 0.1500282 to 1.00.

Qualitative Observations

In the test between the clear and brown glass, testers observed that the clear glass beer seemed to be "softer", while the brown glass beer seemed to have a more defined hop flavor and was perceive as smoother. However, participants also found it very difficult to find any differences.

In the test between the green and brown glass, testers observed that there was no marked difference, and testers felt like they were guessing randomly. Similarly, in the test between the green and clear glass, testers observed that there was no marked difference. In all tests, no testers were able to identify a clear "skunky" smell when blindfolded.

Results

In our quantitative triangle tests, we found that there is no statistically significant difference at a significance level of 0.10 in any of the trials. In the specific beer that was tested, testers were unable to find any differences between any of the glass types on the "skunky" aroma that is typically associated with lightstruck beer.

Conclusions

In the tests, it became abundantly obvious that the effects of skunking or lightstriking may take longer than one day to take effect. If this were a hoppier beer, the skunking may be more clear, since the UV radiation specifically acts on the hop oils, but in this beer, the other dominant malt flavors made it very difficult to detect skunking.

In our test, we are less worried about skunking in a less hoppy beer as long as the time when the beer is exposed to UV light is relatively short. If you're packaging a beer without too much hoppiness in the morning for consumption that afternoon, we do not believe that skunking is a big issue, even if the bottles are exposed to UV light. However, we are not ready to extend these results to other beer styles.

Citation
Kim, S., Bennett, J., Bickle, C. et al. (2024). UV Exposure, Glass, and the "Skunky" Smell. Mount Vernon Brew Club. https://mountvernonbrewclub.org/site/index.php/ontap/uv-exposure-glass-and-the-skunky-smell/.

This work is licensed under Creative Commons Attribution 4.0 International.