Birectifier Analysis of a Butyric Acid Bacteria Infected High pH Rum

Another interesting rum crossed my desk for birectifier analysis from the always exciting Cory Widmayer. A series of ferments comprising two different fission yeasts were spontaneously infected with butyric acid bacteria. This was a stinky one, but I am no stranger to the sensory experiences of this phenomenon because I had already explored a series of butyric infections in Jamaica style ferments.

The vector was likely the molasses and this rare event (in this context) was facilitated by fermentation at high pH, despite preparing the molasses with Arroyo’s thick mash process which involves treating with lime and heating to 80°C to kill bacteria as well as centrifuging or decanting precipitated sediment. In this case, spores were able to survive the process. At the same time, another possibility is that spores were introduced to his work environment when dumping fission yeast isolation media. On top of that all, Cory was experimenting with antibiotic regimens which did not prove too effective.

All in all, it represents a fascinating case study of what can happen even if closely adhering to Arroyo’s heavy rum production methods. When considering the results, Cory also said he stoked the fire by adding lime to raise the pH during fermentation. The pH bounced around from 4 to 5.5 for several weeks. The ferments finished at a likely 8% ABV despite the infection.

One of the goals of our projects have been not only to have successes, but also to witness and develop first hand observations with the full spectrum of rum production experiences. In this case, we are observing many phenomena described by Kervegant in Chapter III Distillery Fermentations:

Distillery raw materials normally contain butyric ferments, in the state of bacteria or, more often, spores (molasses). Activity of these microorganisms is annihilated, when the fermentation is carried out with pure yeasts or in the presence of antiseptics. On the other hand, they can play an important role in spontaneous fermentations, especially when they have a long duration. Their intervention results in a decrease in alcohol yield, and also in the production of higher alcohols (butyl, propyl, etc.) and volatile acids (butyric, propionic, formic, etc.), which contribute to bouquet formation in the eaux-de-vie.

Allan was able to isolate in Jamaica’s rhummeries, and more particularly liquids from the “muck hole”, various butyric bacteria. He found that if they do not get along well with cane juice alone, they develop vigorously in the addition of sugar solutions made of albuminoidal substances, as well as in vinasse with the addition of yeast extract. This last medium corresponds substantially, from the point of view of composition, to the must obtained by mixing fresh can juice or molasses with vinasse. The bacteria grow well from 26° C, but have the optimum temperature 35°. The quantities of acids formed however remain rather low (0.3 – 0.4%), if the must is not neutralized, by means of lime for example.

From this, we get some hints that possibly the aroma mixed with ethyl acetate that I had not experienced before in fraction 1 may be appreciable ethyl formate. We also see that the acid may have been limited if lime was not added. In my own experiences with Jamaica style ferments that feature butyric acid, the ferment should smell like cheetos but nothing more or the distillate is at risk of needing to be distilled at higher ABV thus compromising other high value aroma. The distillate analyzed was not cut to be consumed, but merely to survey assets & liabilities. It was a bit of a butyric acid train wreck. One of the older meanings of the tufo term, translated as the stink, was not just the aroma of scorched yeast, but also free butyric acid which can quickly taint a distillate never becoming an ester.

Another paragraph from Kervegant is worth reproducing:

According to the same author [Allan], the bacteria predominate in the tanks towards the end of long-term fermentation in distilleries producing grand arôme rum, to the point where the yeasts of the fermented liquid disappear completely. Intervention of butyric bacteria explains that the fusel-oil of Jamaican rum is formed essentially by normal butyl alcohol.

This eludes to the idea that early Jamaican distillers of grand arôme rum were not afraid to have a ferment get stuck. I would wager that they were distilling ferments that only achieved an ABV of 3-4% which would be considered non-viable to many today.

Next, Kervegant describes Arroyo’s work with butyric acid bacteria:

Finally, Arroyo, recently in Puerto Rico, by fermenting musts of cane molasses with Pombé yeast in symbiosis with Clostridium saccharobutyricum, was able to obtain a rum having all the characteristics of Jamaican rums, as regards bouquet and chemical composition. This author has observed that by cultivating the bacterium in symbiosis with yeast, its multiplication and formation of alcohol were considerably accelerated. The duration of the fermentation of the yeast, from 70 to 96 hours in the case of yeast alone, was reduced to 28-48 hours. The author attributes this fact to the action of radiation emitted by the bacterium and comparable to the mitogenic rays of Gurwitsch (1).

This may be the ultimate ideal but it does not seem like Cory made observations that matched mitogenic radiation due to his very lengthy fermentation duration. Arroyo’s strategy for managing butyric acid bacteria relied upon understanding his particular organism very well and creating narrow windows for it to operate. His organism was particularly sensitive to sugar concentration, pH, and ABV. The ferment had to be monitored to correctly find the window to pitch the pure bacteria culture. The activity of the bacteria was arrested by either adding more sugar, allowing the pH to drop, or the progress of the ABV climbing.

One of the promises of alcoholic fermentation continuing either simultaneously or after the activity of the bacteria is that esterification may take place as a bio-process rather than solely taking place by brute force in the still. However, we have no good observations to explain the magnitude of what is possible. It is up in the air if more esters are created in the ferment versus in the still. Using laboratory scale equipment, there is no notable level of free butyric acid esterified during distillation so it is hard to say how things behave with a full scale double retort. No doubt the process benefits from significant time under heat. One thing that seems to impact rum production mythology is that it is easier to explain esterification taking place in the still than as a bio-transformation.

It is worth noting that fission yeasts are far more resilient to butyric acid that budding yeasts. Any time butyric acid bacteria is present, it is safe to say a fission yeast is required for alcoholic fermentation.

The distillate presented was merely stripped and had a very low ABV of 42%. How would we go about cleaning it up for consumption? One strategy may be to start collecting data on the titrateable acidity of unaged heavy rum role models. This could be followed by collecting data on what percentage of volatile acid is left behind upon redistillation. Another set of data, to inform cut points, may be recording the distribution of volatile acidity during the distillation run at selected intervals of absolute alcohol. This way if you know your stripping run is a bit of a free VA train wreck, you may still have an a really good educated guess on what it will take to tame it and what percentage of the acid will be left behind. Data as a guideline may be increasingly important as volatile acid accumulates in your recycled fractions. See the data may also create reliable sensory queues for operators.

If you have no role models, another way to explore the acceptable amount of free butyric acid in a distillate may be to create systematic dilutions and assemble a tasting panel. Distillates can be observed over time to develop stronger rules of thumb for how much free acid will become ester. With simple methods like acid titration, anyone can likely avoid putting a product in a barrel that will still retain unsellable levels of free butyric acid when it is pulled for bottling.

One other idea to consider is that we still do not know the production rhythms of heavy rums in the Jamaica process. If a single ferment can develop gross surpluses of butyric acid, can that carry multiple distillation runs if it continues to live in the retorts and only slow be drawn down? Would a seemingly non-viable ferment of 3% ABV have different economics if you expect other non-ethanol assets to carry over? Will that dunder be stretched across multiple ferments? Do run amok butyric ferments exist in ratio to others that are more common and clean? Even with all the interest in heavy rum, we just cannot get the producers talking enough to say.


Fraction 1: There is some unique tones in here. Ethyl acetate is easily detectable but there are non-culinary aromas that aren’t glue-like but are more like stuff I have smelt in red wines. Almost a licorice candy character. I think I detect free butyric acid, but maybe it just synergistically modifies the ethyl acetate? This is more concentrated than I expected. Nothing is aroma-negative. Could the extra aroma be related to formic acid?

Fraction 2: Less concentrated version of fraction 1. The extra character is present but does not become any clearer.

Fraction 3: Very neutral as expected, leading me to believe there is no free butyric acid in fraction 1, but instead something else.

Fraction 4: Detectable fusel oil but at the typical fission yeast concentration. Fraction 5 aroma has bled into this fraction. Very slight radiant quality.

Fraction 5: Not overly cloudy and no significant louche. Definite estery character but no obvious pineapple/butyric ester. No real radiance or obvious rum oil character. Faintly detectable free butyric acid but not at an aroma negative level. Significant gustatory acidity, but no acrid character from overly concentrated esters. This fraction isn’t terribly concentrated, but something about it is very attractive. Something to note is that there is no significant rum oil despite there being yeasts that have produced it in the past. Low pH may be the cause.

Fraction 6: No radiant character and possibly less concentrated than fraction 7. Easily detectable free butyric acid. Significant gustatory acidity.

Fraction 7: Possibly less concentrated than fraction 8. Easily detectable free butyric acid. Significant gustatory acidity.

Fraction 8: Easily detectable free butyric acid. Significant gustatory acidity. Cheetos character but easily recognizable as a flaw.

Stillage: [still need to process this]


My tasting session only used 5 ml of each fraction which meant that 20 ml remained. I blended fractions 1-5 and added water to see if I got a result that had no excessive free acidity. Butyric acid was only faintly detectable. My intention was to titrate the spirit and compare it to the original to see what reduction I was experiencing. Sadly, I am not yet set up again to titrate because I am still recovering from flooding in my lab. I need to re-calibrate the machine and find a surface for it to live.

4 thoughts on “Birectifier Analysis of a Butyric Acid Bacteria Infected High pH Rum

  1. Using laboratory scale equipment, there is no notable level of free butyric acid esterified during distillation so it is hard to say how things behave with a full scale double retort.

    This has not been my experience at all, it’s trivially easy to create and break esters in distillation. Keep in mind, a small amount of sulfuric acid is critical in catalyzing the reaction, and it’s something Arroyo commonly did as part of processing. It’s also interesting to completely reverse those reactions, by adding baking soda to the distillate and redistilling (1-2g/l of low wines).

    What was most interesting to me, was to start with pure acids, alcohols, and water, and run numerous test distillations while varying concentrations. Propionic, Butyric, and Lactic (don’t waste your time with Acetic). Go in with a hypothesis about free acid levels, and give it a shot. Start simple, 1 alcohol, 1 acid. Incredibly insightful to understand how these come over in distillation. Probably even more interesting to then run it through the birectifier and plot out your cuts.

    Really just working backwards to eliminate the complexity, starting with simple mixtures. Gets more interesting when you start adding in additional fusel alcohols (1-propanol, isobutanol, isoamyl alcohol, etc), and you can experience how you go from one-note carboxylic esters to something more complex.

  2. Bio-pathway for butyric acid esterification isn’t at all obvious. Saccharomyces is not known to produce ethyl butyrate, Brett can, but only in very small amounts. If that’s the pathway, there is another friend that needs to be identified..

    https://www.researchgate.net/publication/320352577_Ethyl_butanoate_is_synthesised_both_by_alcoholysis_and_esterification_by_dairy_lactobacilli_and_propionibacteria

  3. “One thing that seems to impact rum production mythology is that it is easier to explain esterification taking place in the still than as a bio-transformation.”

    Yes, this is not only a funny sentence, but what you want to say has some truth I think. I also think the esterification power of the still is overrated. At least if you don’t use sulfuric acid as a catalyst (but then it is not the esterification power of the still anyway).
    And the addition of sulfuric acid has disadvantages not easy to bypass (I am thinking of destroying of some good flavors and of making acids more volatile).

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

search previous next tag category expand menu location phone mail time cart zoom edit close