Monday, 8 December 2014

Failure: Spontaneous Fermentation

A little over a month ago I tried my first foray into brewing a spontaneously fermented beer.  I decided at the start that along as it started fermenting within a week or two, I'd let the beer age out no matter how bad it tasted at first.  Sadly the beer failed to cross even this low hurdle!

I started to write up a post about my process anyway, and in doing it I found myself thinking about why a home brewer would be interested in fully spontaneous beers at all.  In the end I wound up with a rather long and rather rambling discussion of that topic.  I've decided to post the whole thing, in case it was helpful or of interest to other homebrewers: the first half discusses my process, and the second half contains some general layman's thoughts on spontaneous fermentation.


First, the background stuff that is largely irrelevant now that the beer's been ditched. The grist was a blend of wheat malt and pilsner, with about 100g of flaked oats steeped in the wort after mashout to add some starch (in place of a more traditional turbid mash).  I boiled for only ninety minutes, using aged pellet hops at a rate of 4-5 grams per litre.

After flameout, I transferred the kettle to a spot adjacent to an open window and wrapped it up in the insulation I use while mashing (a cut-up exercise mat and a sleeping bag).  Here I was trying to follow advice that Cantillon's Jean van Roy gave to homebrewers in this interview with Basic Brewing Radio (direct link to audio).  He mentions the cooling rate of the wort as a potentially significant difference between beers brewed at Cantillon and beers brewed on a smaller scale: their larger volume of wort takes considerably longer to cool, and this might determine the kind of organisms that are dominant in the wort over the following days.  As evidence of this he mentions a beer brewed on a smaller scale at a local commune that didn't turn out so well---so anecdotal evidence at best, but I thought it was still worth implementing.

After wrapping the kettle in the sleeping bag I removed the lid and covered the opening with a thin mesh bag, then left the kettle by the open window to cool overnight.  The forecast was for temperatures just above freezing, but there had been some colder nights in the preceding week.  Next morning the beer was still in the low 80s, significantly warmer than the temperatures mentioned by Jean van Roy (I'd have to listen again, but I think he mentions temperature that are 10-20 degrees lower).  I transferred the beer to a sealed carboy to let it cool further, but in hindsight I wonder if I shouldn't have left it out for a longer inoculation.  This is one of the points at which I will change my process if I do this again: perhaps longer exposure at lower temperatures would have encouraged inoculation by wild yeast?

Lambic brewers rely as much on organisms inhabiting their barrels as they do on the spontaneous inoculation during cooling, so to imitate this aspect of their process I soaked an oak spiral in my gallon jug of "house culture" for several weeks before brew day, then added it to the cooled wort hoping that it would introduce further diversity into the fermentation.  This is the second point at which I will change my process if I do this again.  I was essentially relying on the spiral for the main inoculation, and I worry that by only leaving it in the house culture for a few weeks I didn't give the brettanomyces and other organisms enough time to establish themselves in the wood. This might also explain the fact that fermentation didn't start as I'd expected.

After inoculation, I waited patiently for signs of fermentation.  A few days in I saw a thin white pellicle form on top of the wort (blurry photo below), which made me optimistic that things were going to plan.  But then after that nothing happened.  Five, six, seven days past and the wort looked exactly the same: a thin white pellicle on top, but no activity in the airlock and no other signs of fermentation.

Eventually, after almost two weeks, I decided to open the fermenter up and take some measurements.  To my disappointment the beer had only dropped a few gravity points, and its pH was still up at around 4.9.  That was high enough that I didn't want to risk tasting it, so I transferred the wort to a smaller carboy to see if the oxygen pick-up helped start a fermentation.  But everything seemed the same the next day, and I decided to give it up.  It wasn't the first beer I've poured down the drain, and it probably won't be the last!

Thoughts on Spontaneous Fermentations

While going through all of this I was thinking a lot about why a homebrewer might be interested in spontaneous fermentation in the first place.  Obviously it has to do with the fact that Belgian lambic brewers use these processes to make such wonderful beers---but can we be more specific about what it is we're trying to imitate, in a way that might help with translating these practices to a smaller scale?  What follows are my own half-baked reflections on why I am interested in spontaneous fermentation, and why I'll probably give it another shot next year.

I'm sure that for a lot of people the allure of spontaneous fermentation comes from the idea of capturing local yeasts and bacteria that might provide a unique range of flavours by the end of fermentation.  I suppose there are two distinct points here: first, the desire for "sense of place", which is often expressed in terms of the notion of "terroir" (there's been a lot of talk about "place-based beer" recently); and second, a desire for the unique range of flavours produced by mixed fermentations.

I'm certainly attracted by the idea of a house character arising from mixed fermentation that would create a distinctive and recognizable flavour profile in my beers; but I don't know that I care if this comes from "local" organisms, bottle dregs, commercial strains, or a mix of all three.  (I guess I'm a little skeptical about the way ideas like "terroir" and "sense of place" are used in marketing beer.  I've been making my way through Authentic Wine in the past few weeks, and one thing that became apparent very early on was how complicated and contested a notion terroir is in the wine world.)

But anyway, if what you want is local organisms there's no need to go through the risk of trying a spontaneously fermented batch; you can try to capture them on a smaller scale in leftover wort, grow them up selectively based on sensory impressions, and pitch as you would a regular yeast.  This is how Jester King cultivated their house culture.  That seems like a much more reliable option than the techniques used by lambic brewers, and its one that's been adapted successfully on a home brew scale.

On the other hand, what if you just want to a mixed fermentation, and don't really care if the organisms are "local"?  The obvious answer is that you can pitch a variety of dregs and cultures at the start of fermentation or as it progresses, but this won't get you anything like the diversity of organisms present in traditional spontaneously fermented lambic (see e.g. this paper).  General consensus seems to be that by the time they are bottled there are only a few organisms active in these beers, and commercial blends are usually limited to a few strains of yeast and LAB.  But why might such diversity be desirable?

This brings me to another reason for imitating the practices of lambic brewers: to encourage a similar fermentation cycle as the one found in their beers.  As lambic ages, different organisms become dominant, fermenting out the remaining sugars and modifying byproducts left by activity earlier in the fermentation.  One thought you might have here is that the sugars and other compounds available at each point play an important role in shaping the flavour profile of the final beer.

Let me give you an example of what I mean, based on my layman's understanding of Wild Brews and the paper I mentioned above about the organisms present in lambic wort. We know that the dominant organisms in the first few days after fermentation are enterobacter.  That might seem to be something undesirable, since these bacteria produce fatty acids and create various unpleasant off-flavours.  In fact, Jeff Sparrow mentions that some lambic brewers have taken steps to prevent this stage of the fermentation (though, based on the paper, that is clearly not true of Cantillon).  However the fatty acids produced by the activity of enterobacter can act as precursors for ester formation by brettanomyces.  In other words, these wild yeasts take an unpleasant compound like butyric acid, which is produced by the enterobacter and smells rancid; and convert it into another compound, ethyl butyrate, which smells like juicy fruit, pineapple, and cognac!  That being the case, some small amount of initial activity by enterobacter might actually be desirable, as it would produce compounds that could be transformed at a later stage in the fermentation.

What about sugars?  Reading through Wild Brews again I was also struck by the claim that the activity of enterobacter in the early days and weeks of fermentation might shape the rest of the cycle, since the enterobacter will consume amino acids in the wort before the saccharomyces get a chance to use them. This will slow down the growth of the yeast and limit their ability to ferment the wort---Sparrow gives O.G.s of 1.022 after 3-4 months, and 1.012 after about 8-12 months. That means that there is lots of fermentable sugar left when Lactic Acid Bacteria become the dominant organisms and take over fermentation. It also means that brettanomyces doesn't become active until relatively late in fermentation, since it grows even slower than saccharomyces, especially in this nutrient poor environment.

So here we have one early stage of fermentation, distinctive of spontaneous beers, that shapes the rest of the cycle by determining the sugars available as each organism becomes dominant and by ensuring there are various compounds present in the beer.  I find the numbers that Sparrow provides quite striking: I don't think I've ever had a beer with an O.G. in between 1.040-50 be as high as 1.022 after three months of secondary fermentation, and I've certainly never had one be as high as 1.012 after a year. What's more, based on what these books say I suspect that the fermentation cycle of a typically inoculated home brew looks pretty different from a traditional lambic; for instance, I'd guess that brettanomyces becomes dominant much earlier in the fermentation, even if its population grows slower than saccharomyces.

Does any of this matter?  I'm not sure. Its certainly true that you can make great sour beer without a spontaneous fermentation.  Its also true that you might be able to achieve some of the same effects by other means.  For instance, Vinnie Cilurzo at Russian River imitates the cycle of lambics by staggering the inoculation of the wort with different organisms (though homebrewers don't seem to have had as much success with this process---perhaps because Vinnie can halt fermentation by filtering the saccharomyces out before he transfers the beer, leaving plenty of fermentables for the LAB and brett).  And perhaps most importantly of all, a fully spontaneous inoculation is a risky affair, especially if you're relying on a single batch rather than the wide range of barrels a real lambic blender can pull from.  There's no guarantee that the brettanomyces will clean up the off-flavours produced by the enterobacter, and there are other risks here too (e.g. the enterobacter will produce acetic acid, which is undesireable in anything other than small amounts and might encourage the formation of an excess of ethyl acetate, making the beer smell like nail polish).

As I finish writing this, my thoughts keep coming back to the long slow fermentation cycle of lambic beers.  One reason that Dave Marliave gave for adding starch to the wort (in imitation of a turbid mash) was that it would extend the life-cycle of the beer: whereas a beer that was mashed high or had maltodextrin in it would be fully fermented after a year or so, in a beer with a starchy wort the slow breakdown of complex starches would result in more residual sugar in a year-old beer, leaving a more balanced flavour profile and creating an extended period of development.  The slower fermentation resulting from the various organisms involved in lambics would also have the same result.

At any rate, I'd certainly like to give spontaneous fermentation another shot, if only so that I can see first hand how such a beer develops over time.


  1. Fantastic write up, Amos. A lot of your thought process has been similar to mine. I recently did my first "spontaneous" fermentation by leaving the kettle out over night with the lid off, but I also inoculated with some trub from a young sour. The Sacch was still alive and started fermenting about 30 hours in. I don't know if the local flora thing will produce much of a difference in the beer... we'll see. Cheers!

    1. Thanks Dan! I think when I do this again I'll make sure I have a backup in case fermentation doesn't start within the first week, it was a shame to ditch the whole batch.