The detectable threshold of detecting diacetyl in beer is commonly understood to be 200 parts per billion (ppb). However, this threshold may need to reconsidered for different styles of beers.
Lallemand Brewing, in partnership with the Siebel Institute of Technology, held a webinar putting laboratory analysis up against sensory analysis. Eymard Freire provided the sensory analysis while Katie Strain provided the lab analysis.
First, let’s discuss the flavors of diacetyl. At low levels, diacetyl gives beer a butterscotch, or buttery, flavor to many lager beers. At higher levels, this off-flavor can overwhelm the intended flavor profile with an artificial butter flavor.
There also is often a mouthfeel difference in high-diacetyl beers. These beers can be more highly carbonated and perceived as more viscous. Some describe it as coating the mouth like movie-theater butter.
During brewing, diacetyl can be controlled by yeast selection. There will be more diacetyl left in the beer if you have a strain of yeast that his highly flocculant. British strains naturally have strains more likely to create diacetyl. Water chemistry also plays a part in how much the yeast will flocculate.
Other factors that influence diacetyl production include: temperature, aeration level, and bacterial contamination.
Diacetyl production may continue even after packaging if oxygen is left. Always perform shelf life stability testing to ensure off-flavors like diacetyl are picked up.
During service, it’s important to maintain adequate sanitation. A draft line contaminated with bacteria can easily lead to cross contamination. For example, it’s common to detect diacetyl in lines that are resting overnight — or longer, in the case of sports arenas that only serve draft beer a few nights a week.
For our head-to-head test, Katie tested for quantification of VDK (2,3-dione and pentadione) and compared it to aroma analysis using GC-FID and spectrophotometry of the headspace. The data analysis tested for compounds most frequently identified.
For the sensory tasting, the panelists used commercial samples and then spiked a second sample with an amount of diacetyl that would bring it to the “detectable” threshold of 200 parts per billion.
Diacetyl is more detectable in the American lager. The lab tests for this beer showed 0.10 parts per million (ppm) in gas chromatography with flame ionization detection (GCFID) analysis. Scaled for normal sensory thresholds, this scaled to 0.5 in a sensory threshold ration.
In an un-spiked sample, sensory analysis concurred with the laboratory assessment: No diacetyl was detectable. In the spiked sample, the sensory test was still fairly subdued. The spiked sample showed the diacetyl levels were increased seven times stronger than the 200 parts per billion threshold.
The typical flavor profile of lagers allows diacetyl to be more easily detected in lagers as compared to other beers.
The flavor profile of a sweet stout makes it difficult to detect diacetyl, even at levels greater than 200 parts per billion. Sensory analysis found the stout to be clean. The spiked sample brought the sample to three times the “detectable” threshold and still was not identifiable.
The spiked sample of the IPA allowed diacetyl to be more easily detectable at three times the threshold. Sensory analysis showed the perceived diacetyl to be greater than the stout and about the same as the American lager. The hop characters did not mask the typical buttery notes of diacetyl.
Training your palette
Typically, sensory tests can include three to six times the 200 parts per billion threshold to easily identify diacetyl. Palettes can become more sensitive through training. The Siebel Institute of Technology offers sensory test kits that include common off-flavors in beer like diacetyl, isovaleric acid, and hydrogen sulfide.
About the panelists
Eymard Freire is the recruitment and product manager at the Siebel Institute of Technology. He is a certified Diploma Bier Sommelier from the Doemens Academy.
Katie Strain, M.S., is a laboratory services manager of Alcoholic Beverage Quality Assurance/ Quality Control and a lecturer in the School of Hospitality at Metropolitan State University of Denver.