Investigating Lost Spirits Investigations Part I

Sponsor my distilling work simply by sharing the artisan workshop of the Bostonapothecary on social media. Copy, Paste, Support!

I aspire to eventually examine Lost Spirits new aging technique but I thought I should start by looking at their first white paper which I read a while ago. Here is a link to their paper: Trace Carboxylic Acid and Ester Origin in Mature Spirits.

This was my conclusion but I’ll move it to the beginning: I think this paper is really cool, but sort of naive. I’ve wanted to see new distilleries start doing investigations for a while now and I hope to do more myself. The sad thing here is that it isn’t that sophisticated yet made big rounds around the internet and that shows that the spirits community just hasn’t gotten very far. I saw no intelligent comments on the paper from industry peers. Lot of cheer leaders and then lots eye rolling, but nothing constructive.

The biggest disappointment in the paper is the bibliography. It kind of shows I haven’t achieved much. They cite one source when I’ve read & wrote in this territory for years and made countless papers available and annotated & commented on all of them. I’ve tried to create a culture of openness and constructive comment that I found in so many of the giants of distillation that I’ve read. It started with Amerine then it was Valaer, and Guymon, and Willkie and now Piggot. Openess and a high tide lifts all boats is the true culture of the industry and how all the research got lost and forgotten, I don’t know.

Oak matured distilled spirits are one of the least well-understood consumer products in the world.

I would say this statement is less true than people think. This blog hosts and uncovers unending mountains of scholarly research done by the industry. Its true that few in the industry, especially the new arm of the industry, are aware of the research body and I’ve talked to distillers that have been in the business thirty years and they’ve never had the benefit of any of the papers I dig up. They just don’t understand where I got them all from. The library? Inter library loan? I simply go to the library. And then I actually read the papers.

Oak maturation, by contrast, is not well understood. Much of the information printed on the topic also contains gross factual errors and flawed assumptions. Perhaps even less well understood is the potentially important interactions of chemicals formed during the fermentation with compounds extracted from the oak.

Oak maturation is far better understood than people realize. Big players in the industry right now even do tons of private research and use very sophisticated data analysis to learn more and more about oak aging. In the spirits industry, there is a tenuous relationship between tradition and innovation and much of the research that is done is down played and sort of hidden. There are some factual errors in the older literature and you will see some researchers point this out and update ideas as methods got betters. A lot of analysis methods used all over the field of chemistry were pioneered through studying alcohol and a lot of giants of spirits chemistry like Peter Valaer, Herman Willkie, Maynard Amerine, James Guymon, and now John Piggot have made massive contributions. Piggot is my absolute favorite. I love his writing style and he seems to have the best command of both chemistry and neuroscience while others are sort of lopsided.

The last comment, about chemical compounds produced during fermentation reacting with chemical compounds extracted from the oak, probably refers to esterification reactions the paper aspires to study. There is an equilibrium amount of esters a spirit can hold. To move towards the equilibrium esters are either forming or breaking apart (into fatty acids and alcohols). Compounds extracted from the oak change the equilibrium, increasing the amount of esters a spirit can hold.

The most comprehensive study on the topic was published in the Journal of the American Chemical Society in 1908 by C. A. Crampton and L. M. Tolman. Unfortunately Crampton and Tolman lacked modern tools such as gas chromatography and mass spectroscopy making their work very incomplete.

Crampton and Tolman is an interesting paper, but its far from the most comprehensive and so much has happened since it came out. One of the my favorite papers that will have gigantic impact on new distillers is: 1968 ANALYTICAL PROFILE OF CISTERN ROOM WHISKIES Schoeneman, Robert L. and Dyer, Randolph H. J. AOAC (1967), Vol. 51, No. 5, pp. 937-987. I keep procrastinating digitizing my copy (nag me and it will happen). At the end, the paper has a great reflection on the investigations of Crampton and Tolman and where the American whiskey industry has come since then.

Maybe I should whip up a brief & incomplete bibliography to give people ideas about what is out there:
Changes in Whiskey stored for Four years (Peter Valaer 1936)
A study of Whiskey stored for four years in Plywood Barrels (1950)
Changes in Whiskey while maturing (1943)
Comparison of Scotch malt whisky maturation in oak miniature casks and american standard barrels (Piggot 1995)
Effect of cask charring on scotch whisky maturation (Piggot 1993)
Flavor components of Whiskey I (2001)
Flavor components of Whiskey II
Flavor components of Whiskey III
Foreign & Domestic Rum (Valaer 1937)
Influence of distillation system, oak wood type, and aging on composition of cider brandy in phenolic and furanic compounds.
Origins of Flavour in Whiskies and a Revised Flavor Wheel: A review (Piggot 2001)
Role of Organic Acids in Maturation of Distilled Spirits in Oak Casks (1999)
Volatile Fatty Acids in Some Brands of Whisky, Cognac and Rum (1968)
Feed stocks, fermentation, and distillation for production of heavy and light rums
Production of Heavy Rums (Arroyo)
Robert Leuté’s 1989 James Guymon lecture

I could go on and on and I’d list some of the more modern complete grad school text books on making spirits which are really impressive. Then we could also list papers on accelerated aging and why they worked or didn’t and that would give us more clues into what we’re ultimately looking for. There is a cool section on accelerated aging in the Technology Winemaking.

So its safe to say there is a lot more than the work of Crampton and Tolman in 1908. One reason we know so much about aging from the IRS chemists such as Peter Valaer is that to detect fraud in spirits, they had to know what legitimate aging looked like to find the outlying fraudsters. If you say it was aged for X years, why doesn’t it have the chemical hallmarks of a X year product? We didn’t yet say anything was an ordinary, sub par, or extraordinary product, we just counted chemicals to test a claim that is symbolic as well as sensory.

Carboxylic esters are the compounds responsible for fruit flavors found in nature. They have long been observed to form during the oak maturation of distilled spirits and are thus of great interest to us as spirits makers. Carboxylic esters are formed when an alcohol chemically bonds to a carboxylic acid.

Keep in mind esters form as well as split apart. The equilibrium of what can be held together changes as the other variable change due to aging. Besides during aging, esters and their precursor caboxylic acids are inherited from the fruit with some fruit having more than others. Esters and carboxylic acids (also often referred to as fatty acids) are formed during fermentation. Esters also form in the still, especially a pot still because of the longer time under heat seen relative to the typical operation of a column still.

So you can track these aroma compounds and their precursors at every stage of the process and research has done that. And don’t forget, some are more noble than others. Some of the fatty acid ester precursors even get removed during chill filtration of superbly aged spirits so it isn’t all that simple. It would be great to learn more about what gets removed and why they didn’t form esters.

While it is well known that esters form during oak maturation, what is not known is the degree to which precursor carboxylic acids originate from the charring/toasting of the barrel vs from bacteria and yeast in the fermentation and which ones originate where.

This is known and has been the subject of a lot of investigation. This is asking, is there carboxylic acids in the wood? I would say not as significantly as the other steps of the process. Keep in mind, we use new oak, second use, and third use. And none is more superior, each has its purpose, especially the latter in rum aging. So by the third use, tannin is reduced, vanilla like compounds are reduced, and the barrel which can be re-charred is mostly a vessel to soak up congeners (in the char) as well as a vessel with special porosity to get just the right effects of the angels share and slow oxidative changes. Equilibrium has so many variables and you don’t want to change one too fast. The slowness of barrel aging means little reactions keep marching around in a circle and we can catch it at its most beautiful point before things run amok.

In order to study them in more detail 5 commercially available rum samples were subjected to direct inject mass spectroscopy and compared. The instrument also picked up peaks of some relevant aldehydes with similar volatility values.

The problem with direct inject mass spectroscopy is that the reading gives tons of biases. Lots of stuff overlaps and it takes serious computer modeling to be able to untangle a reading. When the industry uses inline monitoring of product with spectroscopy (which feeds them massive amount of data), they can only untangle the reading into something meaningful because have done tons of leg work with chromatography to create robust models to apply to the spectroscopy.

I’m not qualified to say much. I know how to read the results but not to operate the equipment and I know a significant amount of their limitations from reading so much. Using advanced analysis techniques for spirits differs from other fields like biology. Spirits often require exotic sample preparation techniques because all the ethanol or sugar biases the results. If spirits are 99% ethanol & water and 1% congeners, you need to extract that 1% to get a better look with any real fidelity. Often you use serious organic solvents like hexane and dichloromethane to pull the congeners out of the ethanol and then you separate those organic solvent with a vacuum still to isolate the congeners.

I have played with hexane a lot. I wanted to see if I could explore sample preparation in a beautiful context. I tried to suck the congeners out of gins and cognacs and was going to isolate them and then add them to fernet or make a double cognac, cramming twice as much aroma inside. Well it didn’t work like I thought and became wildly expensive. I was getting to a point where I needed to explore continuous liquid-liquid extraction which required expensive glassware (and then I pretty much ran out of money).

NOTE: Traditionally ethyl acetate has been the most extensively monitored carboxylic ester, as it is the easiest to detect due to concentration. It almost certainly originates in the oak, because it is known to increase with every year that a spirit ages without stopping. However, ethyl acetate has a very high aroma detection threshold and thus has less impact on flavor than other trace carboxylic esters we are interested in studying in this paper.

Even a hundred years ago they were aware of the differences in esters and their contributing qualities. Ethyl acetate is not exactly the most monitored, but because esters used to be counted with titration, which can count esters, but not differentiate them, the number of esters was expressed as ethyl acetate which is a chemistry counting convention. Ethyl acetate is the most common ester by far and the most basic in its building blocks. Sometimes carboxylic acids are referred to as long chain or short chain. Acetic acid is the shortest chain and most basic. When winemakers count total acids with titration, they do something similar, counting everything as tartaric even though other types of acids are present. Simplifying total acidity is enough to give them useful data to base decisions on. It is not fair to say that ethyl acetate has less impact on flavor because there is so much of it relative to other esters.

There might be ethyl acetate in the oak, but that is not as significant as the other sources. Robert Léauté’s 1989 James Guymon lecture (page 11) gives an easy to understand chart examining the esters found in cognac wines after fermentation. Ethyl acetate is the most common ester out numbering other esters by giant magnitudes except ethyl laurate. Léauté even gives the advice that fermentation temperatures are carried out at a specific temperature so that some of this ethyl acetate evaporates and then eventually much of it will be removed from the hearts fraction with the heads cut. Léauté’s lecture is the greatest concise primer on distillation ever written.

One reason ethyl acetate can form as spirits age is due to the oxidation of ethanol to form acetic acid and ultimately linking up with an ethanol to become ethyl acetate. There is even some acetaldehyde in there as an intermediate step of the oxidation process. This is all governed by shifting equilibriums. Distillation doesn’t produce something that comes out of the still at equilibrium. Its more like all shaken up and therefore rearranges pretty quickly. Besides the porous nature of oak, which facilitates oxidation, compounds extracted from oak which lower the pH can be a catalyst for reactions and influence the various equilibriums.

The fact that the aging and distillation of these two products was so similar appears to suggest that the key difference originates in the fermentation (likely yeast strain choice).

This quote refers to two chromatograms shown in the paper. Yeast strain choice is a thing, but there are also many other variables that define spirits.

It is possible that a variation in charing of the wood could have provided the difference, or perhaps a subtle difference in distillation protocol. The warehouse climates are assumed to be highly similar so that was likely not a factor. Also idiosyncratic barrels could be ruled out as both products are blends of hundreds of casks.

These ideas are just the tip of the iceberg of potential production differences.

The fact that the fermentation and distillation of these two products was so similar yet the sample on the left is nearly twice the age of the product on the right, and is nearly identical in VOC fingerprint appears to suggest that by the 7-8th year of oak aging all of the volatile range carboxylic ester formation is complete. This would strongly suggest that the carboxylic acid precursors for these pungent trace esters originate entirely in the fermentation and are not derived from the oak. If the precursor acids were derived from the oak we would expect to see far higher peaks in the 15 year rum.

This quote refers to another set of chromatograms. I would say based on every paper I’ve ever read, that desirable ester precursors for rum come from the fermentation. And remember, post distillation esterification is a thing, but esters are also born in the still and when you have the right stuff in your fermentation that is why you go to the expense of a pot still distillation with a long time under heat if you want to make a heavy product. And don’t forget, a column still can be operated to achieve a lot of the same objectives.

The fact that the fermentation and distillation of these two products was so similar yet the sample on the left is over 3x the age of the sample on the right appears to further confirm the suspicion that the carboxylic ester formation is complete by 7-8 years of aging. It also appears to soundly confirm that the trace carboxylic ester profile of a mature rum are essentially predetermined prior to aging. Though it may take as many as 7 years to complete the process – further aging cannot form additional trace carboxylic esters beyond the level of precursors available from in the white spirit.

So you can’t put a light rum in a barrel for 25 years and get a heavy rum.

Given the prior observations comparing and contrasting various column distilled rums a final comparison was made against a 33 year aged pot distilled rum. As was expected the pot distilled rum showed significantly higher peaks for every target ester owing to the fact that the pot still provides much less efficient separation and allows far more of the chemical composition of the fermentation to pass into the final spirit. This observation appears to confirm that the trace ester density is not only predetermined prior to the spirit entering the cask but that the distillation cuts and level of rectification has a massive effect on the final character of the aged spirit. Given the conventional wisdom that aging can “fix” certain off notes in spirits, this is not surprising as many off notes are in fact carboxylic acids that have not yet been esterified during the aging process.

So many variables can come into play here, but one of the major ones again is time under heat. The cuts can be similar and you can distill with a column still at a very low proof but the time under heat in the boiler is going to be much shorter creating less time for acid catalyzed esterification in the still.

Trace carboxylic esters (excluding ethyl acetate) in mature distilled spirits are responsible for the fruit flavors often seen in desirable products. While it is true that the spirit must be aged in oak to increase ester density and convert off notes from carboxylic acids to desirable esters, it was found that their peak concentration is limited by precursor carboxylic acids generated in the fermentation.

One this this misses is the fixative role of ethyl acetate described by Robert Léauté. You want ethyl acetate as close to the recognition threshold as possible without going over. When you go over the recognition threshold, ethyl acetate will smell like nail polish remover, but when below (but well above the absolute threshold), ethyl acetate will be a bridge for the other aromas. Without ethyl acetate to bridge aromas, they will be perceived as disparate and possibly dissonant. The fixative term is used in many different ways but here it brings aromas together (spatially in the mind) to create unique and extraordinary percepts. A large part of distilling and blending is managing ethyl acetate.

It was further observed that pot stills are far better at capturing precursor acids from the fermentation than column stills. However, I would expect column stills designed for lower rectification as is common in Armagnac or Martinique produce to results more closely related to those shown for the pot still rum.

Not every distillery owns a pot still, but precursor potential is a big part of choosing to operate one or not and don’t forget about time under heat. One of the reasons California never had a lot of pot distilled brandy was that their wines were too low in acid to produce enough precursors to justify the added expensive of more time under heat that a pot still generates.

To achieve Lost Spirits’s goals of making the most heavy, robust, rich rum possible, it is apparent that a pot still is ideal. The observations also show that special attention must be paid to the bacteria and yeast strain choices in fermentation. Fermentations could be engineered to generate higher concentrations of favorable precursors. This optimized fermentation coupled with a pot distillation could then generate white spirits more suited to gain substantial flavor density through esterification during the aging process.

Awesome. One of my goals when I started exploring distillation was to explain all the nitty gritty operational differences of still operation which was sort of mystified so that producers could have enough clarity to start working backwards into deeper involvement with other aspects of production like fermentation and cultivation of raw materials. Still operation was just getting too much fetishization and I couldn’t find much intelligent written about it.

Attention will have to be paid to yeast of course, but don’t forget pH, fermentation temperature, the recycling of fractions, the use of dunder, and finally the quality of the molasses.

Of course esterification of trace carboxylic acids (excluding ethyl acetate) is only one component of the aging process. Oak extractives and phenolic compound reactions must be addressed with the same vigor to gain a full picture of the maturation process. The ethyl acetate formation must also be studied in the context of these observations as acetic acid extraction from the oak is likely influencing the equilibrium of the aging spirit (as a buffer solution) in an important way.

And luckily lots of papers address all these concerns. I think that acetic acid extraction from oak won’t be found like the authors think. It will come from other places like the ethanol itself or most definitely in the fermentation.

If you want to learn about any of these concept without having to run a large scale rum distillery, don’t forget to explore my distiller’s workbook. Some of the exercises like the cocoa bourbon or the marmite rye help to explain and explore acid catalyzed esterification in the still. I also did some other unpublished experiments like distilling walnut nut oil or a Sauternes that dramatically illustrate post distillation esterification and the march to equilibrium. After distillation, the nut oil distillate does not organoleptically resemble walnut, but then many months later, a dramatic change occurs and it does. Distilling the Sauternes reveals how much acetaldehyde and plain acetic acid it contains (hides!) and immediately upon distillation it smells horrific and undrinkable. Many months later the distillate mellows and comes to a new equilibrium. It does not end up delicious but it does end up different, illustrating relevant concepts.

2 thoughts on “Investigating Lost Spirits Investigations Part I

  1. Perfect timing for this. Thanks.

Leave a Reply

Your email address will not be published.

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