Category Archives: Yeast

Sources of DMS in Beer

Quite often, when I run into an issue or a question related to homebrewing that is not answered in the usual homebrewing literature or forums, I turn towards scientific literature. One great example is the Journal of the Institute of Brewing, which makes its issues freely available, with literally more than a hundred years of back issues available.

When earlier today, Brülosophy posted lab results about their DMS exbeeriment, their closing statement irritated me a bit:

But, it’s just as possible our understanding of the relationship between DMS and boil length is simply lacking, that our access to modern technology, higher quality ingredients, and better knowledge about brewing processes has reduced the likelihood of problems brewers of yore had to worry about.

That sentence really made it sound like brewing science is only in its infancy. But that’s definitely not the case. Modern brewing science is lots and lots of organic chemistry, microbiology, and even genetics nowadays, so it would be hardly believable that we knew very little about DMS. I vaguely remember seeing an article about DMS on the mentioned journal, so I dug it up, and first posted a link to it in the comment section of that Brülosophy article. But then I thought, why not write about it? Because the paper itself is great.

So, the interesting that I took out of that article is that there are actually two types of DMS precursor. The paper distinguishes them as “active” and “inactive”. Both have different properties, and come from different sources.

“Inactive” DMS precursor is coming directly from the green malt. If you were to product wort from green malt, the precursor would be decomposed, and the wort would end up with large amounts of DMS, and still a certain amount of DMS precursor. During fermentation, parts of the DMS would dissipate through the gases being given off, while the DMS precursor would be metabolized by the yeast. The yeast wouldn’t make DMS out of it, though, hence why that precursor is called “inactive”.

“Active” DMS precursor is the type of precursor that is created out the inactive precursor during the kilning process. After wort production, there would only be a small amount of DMS in the wort, but still a relevant amount of “active” DMS precursor. During fermentation, that DMS precursor is metabolized by the yeast, which makes DMS out of it. Hence the “active”.

This has some interesting consequences. If the kilning can be done in such a way that no “active” precursor is found in the wort, the yeast will not produce any more DMS during fermentation, and the total amount of DMS in the beer is limited by the amount of DMS in the wort right before pitching. It can even be assumed that some of that DMS will be lost during fermentation. The authors suggest that the amounts of DMS formation during wort production need to be controlled, though. In their experience, it is easier to control DMS levels in beer if it’s derived from DMS precursor in the wort, as the final DMS level can be controlled by using a suitable yeast that keeps DMS production low.

And I think the last bit is the crucial point in the Brülosophy exbeeriment: it’s not that a 30 minute boil vs a 60 minute boil doesn’t have any impact for DMS levels, it’s just that both the specific qualities of the malt and the specific metabolism of the employed yeast do matter, and can have a large impact. My guess is: in the exbeeriment, exactly the “right” malt was used (Bestmalz Pilsner malt, apparently), combined with a yeast strain that only produces low levels of DMS (WLP029, a Kölsch yeast strain). And that perfect combination gave a result that made a 30 minute boil indistinguishable from a 60 minute boil in terms of DMS levels. That said, I would really like to see the same experiment done with a different malt (maybe a less modified floor-malted Bohemian Pilsner malt?) and a yeast strain known for greater DMS level, like a lager yeast. W34/70 comes to mind, for example.