A brief and incomplete overview of diacetylIn a typical fermentation, diacetyl begins life as acetolactic acid, which is naturally produced during fermentation.It is unavoidable to find acetolactic acid in fermenting beer. It is just something created by the yeast. During fermentation and the clean up period, the yeast will break down the acetolactic acid and the precursors to acetolactic acid and prevent diacetyl from forming, at least at a level that we can taste. It is commonly a failure to perform a diacetyl rest or prematurely removing beer from the yeast that will leave excessive acetolactic acid in the beer. Acetolactic acid then oxidizes into diacetyl, primarily when oxygen is exposed to the beer during transfer or packaging. Once diacetyl is there, it is hard to get it out of the beer. Yeast can break down diacetyl but it is a slow process and in most packaged beer there is not enough active fermentation to get rid of all of it.
However, fermentation conditions can produce more acetolactic acid than the yeast can chew up, such as mutant yeast cells, underpitching, oxygen exposure during anaerobic fermentation, excessive fermentation temperatures, premature flocculation, poor nutrients in the wort, overpitching and pitching at temperatures over 70F. (See http://www.draymans.com/articles/arts/03.html) Even in fermentations where excessive acetolactic acid is created, sufficient contact time with active yeast can often break down enough of the acid to drop diacetyl below the flavor threshold or leave it at low enough levels the beer is still drinkable.
Diacetyl can also be created in beer through infection. Lactic acid bacteria (LAB) like lactobacillus and pediococcus also create diacetyl through their fermentations (think of the buttery taste in yogurt or sourdough bread) well above what saccharomyces can prevent or break down. You can often tell an infection-caused diacetyl problem because other sensory cues will identify the problem, such as acidity, thinness in the mouthfeel, cloudy appearance and other off flavors.
What (likely) happened in this beerI am fairly confident that this beer did not suffer from an infection. There are no other off flavors in the beer, no unusual cloudy character or excessive carbonation. Additionally, the level of diacetyl has been reducing in the beer. I would expect an infection to continue to chew on the beer and produce more diacetyl. That is clearly not happening.
So that leaves a fermentation problem. I know the beer was properly pitched and aerated like every other beer I brew, so I feel confident scratching out an anomalous fermentation condition as a culprit. What I did with this fermentation is ferment it at 60F with S04, which is very cold for an ale yeast but within this strain's ability to ferment. I chose a cold fermentation to get a clean yeast character and keep as much of the hop aroma from blowing out of the airlock. I fermented the beer for about seven days, cold crashed and then racked off the yeast to dry hop at room temperature for five days. Fermentation was definitely over before racking but I believe the error was not performing a diacetyl rest (by warming the beer) for a few days before racking. Once racked over, oxidation of the acetolactic acid could begin. Why didn't I detect diacetyl in the beer while dry hopping? Oxidation of acetolactic acid into diacetyl doesn't happen that fast. That's why bottle conditioned beer that shows diacetyl almost always goes into the bottle without diacetyl but opens after a few weeks of conditioning with diacetyl.
I would have fixed the problem by warming the beer in a diacetyl rest for a couple of days at the end of fermentation before cold crashing the beer. Easy problem but I got too caught up thinking about preserving the hop character that I forgot about the needs of the yeast. A couple extra days would have been the difference between a great black IPA and a butter bomb.