Friday, April 15, 2011

Brettanomyces Experiment - Initial pH, Attenuation, and Secondary Metabolites

Over the years, I’ve found that new brewers typically fall into one of two categories:  those that like to jump in with both feet and learn from experimentation or those that like to research as much as possible before making the leap.  I certainly fall into the latter category and so as expected, when I started venturing into the world of wild fermentations, I scoured the seemingly limited, publicly available information.  

When I came across Chad Yakobson’s dissertation on brettanomyces, I was really intrigued.  In his “Brettanomyces Project”, one of the things that he examined was the effect that varying amounts of lactic acid concentrations have on attenuation as well as secondary metabolites.  Depending on the strain that he was studying, he basically found that, “higher initial concentrations of lactic acid had a significant effect, increasing the level of attenuation observed in each strain while generally decreasing the secondary metabolites produced” (Yakobson).  As a career analyst, the myriad of data fascinated me but as an intrigued homebrewer, I was left without a subjective view.  If the amount of ethyl lactate production is correlated with the initial lactic acid concentrations yet ethyl caproate is inversely correlated, what’s the initial lactic acid concentration that would give me the appropriate balance of flavors that I want in my beer?  What about attenuation?  Is the reduction in ethyl caprylate worth the increased attenuation?  As much as I value Yakobson’s numerical findings, these are the questions that I, as a homebrewer, want to find the answers to.

Even though I don’t have a home filled with sophisticated lab equipment, I wanted to try and replicate Yakobson’s experiment as best I could in order to determine my own subjective opinions.  The Brettanomyces Project dissertation examined the results of eight different strain profiles, but in order to keep things relatively simple, I decided to only work with two commercially available strains of brettanomyces lambicus:  Wyeast 5526 and White Labs 653.   

Even though I could have kept the grain bill of the base beer extremely simple in order to showcase the subtle nuances of the brett fermentation, I opted not to since most likely my future 100% brett beers will not have such a simple wort.  Since the point of this experiment is to find out what flavors I’ll want in future beers and the lactic acid concentrations necessary to arrive at such levels, I decided that using a base beer similar to what I’ll be brewing in the future will provide me with the most realistic results.  There’s no point in using a simple pilsner malt base if later I’ll be brewing with other specialty grains that could mask the very subtle flavors only apparent with a pilsner base.  In the end, I went with a grist of Maris Otter, Munich, Carahell, and Crystal 40.

Once the mash and boil were complete, I filled 12 one-quart mason jars with 28oz. each of clear wort.  These were then split into two groups of six, one soon to be inoculated with Wyeast 5526 and the other with White Labs 653.   In Yakobson’s study, he examined the fermentation at 5 different lactic acid concentrations.  Since I had six jars for each brett strain, I decided to replicate the lactic acid concentrations that Yakobson used plus one additional point.  After determining the g/L of lactic acid in the 88% solution that I had on hand, I was able to calculate the amount of acid needed to dose each 28oz mason jar to reach 0, 100, 500, 1000, 2000, and 3000 mg/L concentrations.

Concentration (mg/L)
ml of Lactic Acid Dosed
Initial pH

Unfortunately I don’t have access to equipment necessary to count the number of yeast cells per ml of wort, so when it came time to inoculate the jars, I tried to be as consistent with each as possible.  Four days before my actual brew date, I made a 600ml starter for each strain.  Since I only have one stir plate, I rotated between the two starters every 24 hours.  When it came time to pitch the yeast, I used a sterile syringe to suck up the yeast from each starter while it was still spinning on the stir plate so as to get a homogenous mixture for each test batch.  By dosing 15ml of homogenous starter to each jar, hopefully I ended up with roughly the same cell count in each batch. 

After fermentation resides in a few months, I plan to measure the final gravities and pH.  I should be able to be get two 12 oz. bottles out of each batch and soon after bottle conditioning, I’d like to hold one blind tasting with one bottle from each of the 12 experimental batches.  Hopefully the results will be obvious and conclusive, but if not, a second blind tasting after another year or so should provide another opportunity for critique.

Batch Specifics
Batch Size (Gal): 5.0
Total Grain (lbs): 9
Anticipated OG: 1.050
Anticipated SRM: 5
Anticipated IBU: 17
Wort Boil Time: 60 minutes
Anticipated ABV: Varies by initial pH / Lactic Acid concentration

89.2 % - 8.0 lbs Maris Otter
7.4% - 0.66 lb Munich Malt
1.7% - 2.5 oz. Carahell
1.7% - 2.5 oz. Crystal 40

60 minutes – 20 grams Cluster (Pellets, 5.0% AA)

Wyeast 5526 Brettanomyces Lambicus
White Labs 653 Brettanomyces Lambicus

Water Profile and Additions
Charcoal filtered Seattle water
Mash Additions: 0.65 g/g Calcium Chloride, 0.35 g/g Epsom Salt (based on 3 gallons)
Boil Additions: 0.65 g/g Calcium Chloride, 0.35 g/g Epsom Salt (based on 3 gallons), and 0.2g/g Salt (based on 6 gallons)
Lactic Acid – dosed into individual batches at rates mentioned above

Mash Schedule
60 minute rest at 152°
15 minute mash out rest at 168°
Sparged with 170° water

4/6/2011 – Started Wyeast 5523 starter (600ml) and placed on stirplate

4/7/2011 – Started White Labs 653 starter (600ml)…unfortunately it didn’t arrive the same day as the Wyeast.  Swapped Wyeast for White Labs on the stirplate.

4/8/2011 – Switched starters on the stirplate.

4/9/2011 – Switched Starters on the stirplate.

4/10/2011 – Brewed solo.

Hit strike 152 mash temp dead on.  Added mash mineral additions.

Collected 5 gallons of 1.060 wort.  Topped off to 7 gallons with filtered water, boiled for 60 minutes, and ended with 6 gallons of 1.050 wort.

Whirlflock and yeast nutrient were added at 15 minutes remaining.

Whirlpool chilled to 70° degrees and then let sit for about an hour and a half covered, waiting for the wort to completely settle.

Used kettle valve to fill 12 quart jars with 28 oz. of clear wort.

Remaining wort was used to fill another mason jar and a 1 gallon jug, both to be placed in various spots in my brewery to test out spontaneous fermentation.

For the 12 test batches, a 1ml syringe was used to dose each of the jars to the appropriate lactic acid concentrations (see grid above).  After lactic acid concentrations/pH levels were correct, 15 ml of brett were added…Wyeast to 6 jars, White Labs to 6.

Test batches were placed in closet where ambient temp is about 70°.

8/17/2011 - Bottled each sample

9/15/2011 - Blind Tasting Results


  1. Hi,

    How did this experiment turn out?


  2. I just bottled the beers about a week and a half ago. Once they're fully carbonated, I'm going to hold an initial blind tasting with some friends and afterwards, I'll post a full writeup. I'm really curious myself and looking forward to the results!

  3. This is a great experiment. I have wanted to do something similar for a while. Can you give some more details on your mason jar setup? It looks like you drilled a hole then used irrigation nipples to aquarium hoses. But what is it connected to?

    You also did not add oxygen, correct? I have found this is one of the biggest factors on Brett fermentations.

    One aspect that I want to look into that is not in Chad's paper is the temperature factor. If Sacc yeast is so sensitive to it then why not Brett yeast. And maybe different strains are more or less sensitive.

  4. Yup, you're dead on for the mason jars. I connected the outlet hoses to a 12-port drip system manifold that I found at home depot (similar to this From there, I just added a standpipe and modified a cap so that a bubble lock could be inserted.

    I didn't intentionally add oxygen, however when I filled up the jars, the wort did splash around a little in the process. Since I've been happy with a few brett beers brewed in the presence of oxygen and without, my verdict is still out. To me it seems like you can still develop a lot of those traditional brett characteristics in either situation. What have you found?

    I'd love to hear your results on the temperature experiment too if you decide to do one!


Related Posts Plugin for WordPress, Blogger...