WD Brewer Home Brew Log
← Back to Beer

Diacetyl in Homebrew: The Butter Note and the Warm Rest That Clears It

Beer By Wes Drummond 8 MIN READ OFF-FLAVOR DIFFICULTY: INTERMEDIATE

A friend handed my Vienna lager back to me at a cookout and said it tasted like the inside of a movie theater. He was being kind about it. I had spent six weeks on that beer, kept it cold and clean the whole way, and it came out of the keg slick and buttery in a way I had somehow tasted right past during every gravity check. What I had actually done was pull the beer off the yeast the moment the hydrometer stopped moving, which is the exact mistake that leaves this flavor in a glass. The frustrating part is that the fix costs nothing but two or three days of doing nothing at all.

What the butter note actually is

The compound is diacetyl, and the Beer Judge Certification Program describes it in its beer characteristics reference as an artificial butter, butterscotch, or toffee-like aroma and taste that may also leave a slickness on the palate. That last part is the tell I trust most now. Butterscotch can be argued about; the greasy film it leaves behind on the roof of your mouth is harder to talk yourself out of.

Two warnings from the same reference are worth internalizing before you go hunting for it. A significant number of tasters cannot perceive diacetyl at any concentration at all, which means your palate may simply be blind to it and a friend's palate may not be. And low levels of caramel are sometimes mistaken for it, so a crystal-heavy amber can read as buttery to someone reaching for the word. If you have leaned on steeped specialty grains for toffee character, be careful not to diagnose your own recipe as a fault.

Yeast makes it, then yeast takes it back

Here is the part that reframed the whole problem for me: diacetyl is not contamination and not a mistake. It is a normal fermentation by-product that yeast produce on the way to building amino acids, and which the same yeast normally reabsorb and reduce into far more innocuous compounds called diols. Nearly every batch you have ever brewed passed through a buttery phase. You only taste it in the ones where the yeast never got to finish the second half of the job.

That is why the BJCP lists the two main causes the way it does: prematurely separating the beer from the yeast, or exposing the wort to oxygen during fermentation. Low free amino nitrogen or a mutated culture can also blunt the yeast's ability to clean up after itself. Rack too early, crash too early, or filter too early, and you have removed the only crew capable of doing the cleanup. My Vienna was case one, textbook.

The half-hour test that settles the argument

You do not have to guess whether a beer is done. The forced diacetyl test exploits the fact that the precursor compound is odorless until heat converts it, so warming a sample fast-forwards your beer a couple of weeks. White Labs lays out the professional version in its forced diacetyl testing write-up: pull two samples, hold one at 140–160°F (60–71°C) for ten to thirty minutes, leave the other at room temperature, then cool the heated one back down so both are at the same temperature before you taste them side by side.

My kitchen version is two small jars, a saucepan of water on the lowest burner setting, a thermometer in the water rather than the beer, and twenty minutes. Then the hot jar goes in an ice bath until it matches the other. Smell both, taste both, and pay attention to the finish rather than the first sip.

Reading it is simple. Butter in the unheated sample means the beer already has diacetyl in it. Butter in the heated sample only means the precursor is still sitting there waiting, and packaging now would hand your bottles a fault that shows up three weeks later. Either result points at the same prescription: more time, a little more warmth, or a rousing of the yeast before you go anywhere near a cold crash.

The forced diacetyl test: two samples of the same beer, one held at 140 to 160 degrees Fahrenheit for ten to thirty minutes and then chilled, the other kept at room temperature, tasted side by side. Butter in the unheated sample means diacetyl is present; butter only in the heated sample means the precursor remains and the beer needs more time on the yeast. two jars, one burner, one verdict jar A — room temperature jar B — 140–160°F, 10–30 min the control what the beer tastes like today butter here = diacetyl present chill it back before tasting heat converts the odorless precursor butter here = precursor waiting butter in either jar means the yeast is not finished — do not crash, do not package
Fig. 1 — Same beer, same glassware, same temperature at the moment of tasting. The only variable is twenty minutes of heat, and it buys you a look two weeks into the future.

Warmth is the lever, not the enemy

Homebrewers spend a lot of effort keeping fermentation cool, so the counterintuitive line in the BJCP reference is the one to hang onto: high fermentation temperatures promote both the formation and the elimination of diacetyl, and the elimination is the more effective of the two. Heat makes more of it and clears more of it, and clearing wins. That is the entire justification for the diacetyl rest that lager breweries use, holding the beer in the 60–65°F (15–18°C) range for a few days after racking to the conditioning tank.

For a homebrewer without a glycol setup, the practical version is even easier, because most of us are already fighting to keep a fermenter down rather than up. When the gravity readings flatten out, I simply stop fighting. The swamp cooler comes apart, the ice bottles stop getting swapped, and the fermenter drifts up to whatever my basement is doing over two or three days. Everything I do to hold a temperature in the first place is covered in controlling fermentation temperature without a chamber; the rest is just letting go of it on purpose at the right moment.

Brewer's Note Timing matters more than the exact temperature. Start the warm-up while there is still visible activity — a slow bubble every twenty seconds, or gravity a couple of points above where you expect it to land — because you want an active, suspended population doing the cleanup, not a compacted cake sitting under cold beer. Waiting until the airlock has been silent for a week means warming up a fermenter whose yeast has already flocculated out and gone quiet, which is a much slower way to get the same result. And do not add oxygen at this stage looking to wake anything up. Air after the first day of fermentation buys you staling and, per the same BJCP reference, more diacetyl to deal with rather than less.

When the yeast is not the one to blame

There is a second source, and it does not respond to patience. Diacetyl is produced during the early stages of contamination by lactic acid bacteria, notably Pediococcus damnosus. The pattern looks different: instead of a beer that arrives buttery and cleans up, you get a beer that tasted fine at packaging and turns butterscotch in the bottle over weeks, often with a growing haze, a slick body, and eventually a sourness underneath. A warm rest will not save it, and a keg that keeps producing this across unrelated batches is telling you something about your lines and gaskets rather than your yeast. That is a sanitation problem, and the tap hardware is the usual culprit.

Worth knowing too: a low level is permissible in some styles and should never be obtrusive in any of them. The BJCP lists Czech lagers, English bitters, brown and dark British beers, and Irish red ale as styles where a touch is allowed. It is out of place in German lagers, with Kellerbier the exception, and equally out of place in Irish stout, American ales, and Belgian ales. So the same trace of butterscotch is a feature in a Czech pale lager and a flaw in a hoppy American pale — one more reason the style you think you are brewing decides what counts as a fault.

The end-of-fermentation routine I use now

Four steps, added to the back end of every batch, and I have not shipped a buttery beer since:

  1. When gravity is within two or three points of target, stop all cooling and let the fermenter rise into the low sixties on its own.
  2. Give it three full days there, with the beer still on the yeast and the airlock undisturbed.
  3. Run the two-jar forced test on day three, comparing a heated sample against an unheated one.
  4. Only if both jars come back clean do I cold crash, keg, or bottle. If either one is buttery, it gets two more days and another test.

That is three days of shelf space and about fifteen minutes of actual work, and it is the cheapest quality upgrade left in my process — no gear, no new ingredients, nothing to buy. If you have ever had a beer described back to you as popcorn, buttered toast, or slick, run the two-jar test on your next batch before you crash it. The verdict comes back in half an hour, and it is almost always the same verdict: give the yeast a few more days and a few more degrees.

Some links on WD Brewer may be affiliate links, meaning we may earn a small commission at no extra cost to you. See our affiliate disclosure for details.