Thursday, April 2, 2015

Of Replicants and Reactions

These days, it is hard to invoke a response from me. Couple the myriad of medications I'm on with my constant fatigue and general laissez-faire attitude about things and getting me to react in an emotional way is pretty difficult. I wouldn't say I'm robotic in nature but things are definitely taken with a grain of salt. Well...salt substitute. High salt doesn't work with some of my medication. I suppose, in some aspects, that I am a simulacrum of a distiller. I act like a distiller, think like a distiller, live and breathe distilling...but I don't distill. If you're familiar with Blade Runner, I guess I could be considered a distiller replicant. Only, in my world, there is significantly less 80's style film noir and far more video games and coffee.


Recently, a paper came across my desk. A white paper by Lost Spirits Distillery. A white paper that has proved to be my Voight-Kampff Test. In case you've never seen Blade Runner (shame on you), the Voight-Kampff Test is an analysis machine used to evoke an emotional response in its test subject to determine whether or not the subject is a replicant or not. Here is an example from the movie of a Voight-Kampff test:

A replicant, in case you still haven't seen it by the time you've watched that video and realized how awesome of a movie it is and how your life isn't complete without watching it, is a synthetically made android designed in every way, shape, and form to mimic a human...except for their emotional response. Why? Because how do you instill love, hate, passion, perseverance, and understanding into something when it is so varied and poorly understood in the human instilling it? But...I digress.

The Lost Spirits Distillery white paper (available here for full reading) is the latest in "scientific" papers put out by those seeking to tame the savage and patient beast that is barrel aging. I have been watching Lost Spirits for a long time now, carefully examining the papers they put out, and, truth be told, enjoying the work they've been doing. It's not all-encompassing but it's an interesting peek under the hood of one car. However, their latest paper makes some bold claims that, frankly, I see fault in. Not that I necessarily DISAGREE, per se. It is more of a situation in which far more research is needed and, once again, fails to encompass the entirety of the situation. If you'll permit me (and I hope you will considering you're reading this), I'd like to critique it. So let's jump in, eh?

The Model 1, as they call it, is their newest introduction into the world of accelerated barrel aging. A skid mounted chemical reactor to be charged with new-make spirit and oak chips, turned on, and the magic worked. It focuses, almost exclusively, on the esterification process (which I've mentioned here) for  fermentation-based acids and decomposition and esterification of wood / phenolic aldehydes. This takes place in a three-stage process catalyzed by "energy in various forms". Phase 1 esterifies fermentation-based acids into short chain esters. Phase 2 is a catalyzed breakdown of lignin/cellulose into constituent phenolic aldehyde precursors. Phase 3 replicates the maturation process by catalyzing reactions of phenolic aldehydes / oak-based aldehydes into medium and long chain esters.

So far, the science is correct...if seemingly implausible. It seems very much like most of the other products out there. If I had to guess what forms of energy its using, I'd guess ultrasonic, heat, and maybe some sort of electricity-based redox (reduction/oxidation) setup in the line of a plating system...although that last one I'm not sure of because A) there's no highly conducive feedstock and B) RUNNING A REDOX REACTION IN BARREL STRENGTH SPIRITS IS A TERRIBLE IDEA. I'll get back to you on that last one. However, the actual hard data is where they lose me completely. While I won't steal their work and post their data here, I urge you to boot up their paper (once again, found here) and follow along as I get to do what I've always done: break out my big red pen and play TA.

Figure 1:

First off, my problem here is they are using three spirits. Two control (the new make and 33 year old control) and the variable "aged" sample. First off, you can't even calculate a standard deviation. Or a confidence interval. Or error bars. This is bad. If you wish to present a white paper to draw in potential investors/clients, you might want to expand further than one sample. The more samples you present, the more it seems like you've tested it. Also, the less it seems like you cherry-picked data.

Also, it states that the acetyl peak is 60% of the 33 year old sample and, thus, is 15-20 years old. Where did this number come from? There is no source for the "chemical marker concentration" analysis. I mean, 60% of 33 is 19.8 so is that what was went with? There's no context to that statement.

Figure 2:

To start, the graph has no units and that bothers me. Are they picograms? Micrograms? Kilograms? What are the units of measurement here? Once again, two samples are used. Once again, two samples are not enough. But I think the BIGGEST problem I have here is the data itself and what it's telling me. In most cases, save ethyl octanoate and (maybe) isovaleraldehyde, Mark 1 is producing more esters in six days than in a barrel for 33 years. Isn't that a tad...excessive?

You have to understand here, esterification is not completely a two way street. The balance between aldehydes and acids is constantly in flux because of the instability of the additional oxygen molecule to the structure. The polarity of that oxygen destabilizes the whole molecule and makes it prone to, well, breaking the bond. But once you esterify takes a lot of work to go back especially in the instance of barrel maturation. There are two reasons for this. Reason one is that the barrel reactions could potentially undergo Fischer esterification. Fischer esterification is an acid catalyzed esterification of carboxylic acid and alcohol. The acids made during both the aldehyde/acid equilbrium, as well as the acids left over from fermentation, could catalyze the reaction. The equilibrium constant of a Fischer esterification favors the products side (usually Keq > 1) but the reaction will not go to completion. The only way to break that equilibrium would be to invoke Le Chatelier's principle. Le Chatelier's principle, while pretty serious sounding, basically states that you can interrupt equilibrium by concentration, temperature, or pressure and the system will change its equilibrium. In simpler terms, imagine that you have a tug of war team that is stuck in a stalemate. No one is able to pull the other across the line so, in order to win, one side calls in a friend to help out. The tug of war stalemate is equilibrium, the extra friend is the Le Chatelier's principle modifier of "concentration". One side of the rope has more people on it so the balance tips in their favor. How is this done for Fischer esters? Well...since one of the products of esterification is water, lowering the concentration of water will do. evaporation. See where I'm going with this?

That being said, all of this is conjecture and there are a lot of "ifs" in there. There is also the fact that all three of the Le Chatelier's principle variables are changing within a barrel. Temperature, pressure, AND concentration are in constant flux. The Mark 1 does not suffer from evaporative losses so a changing water concentration IS out but if it is a sealed container, temperature and pressure are definitely changing.

Figure 3

Two samples. No units. Unsubstantiated age claim. I won't harp on that. What interests me is the level of sinapaldehyde. Sinapaldehyde should be reacting into sinapic acid / ethyl sinapate but I've seen no mention of it in the paper. Odd.


By now I'm sure you're saying to yourself "But Bacchus, we've let you ramble for twenty minutes and you haven't gotten to the point!" And you'd be right.

The Mark 1 is promising. This paper raises more questions than it does to answer them. Despite its flaws, it intrigues me. But from what I can see is that the spirit the Mark 1 will produce isn't a traditional spirit. It is...a replicant. Exact in all the ways that we know...but yet somehow different. It shares a lot of the markers common in aged spirits but it neglects a lot of the other underlying chemistry that we don't understand. What of the evaporative need for sulfur-based compounds? What of the ethanol-water clustering? What of the concentrations of base aldehydes and acids that ALSO contribute to the spirit? The focus on heavily reacted esters is...admirable...but, like a replicant, it is trying to fabricate one aspect of spirit maturation when there are so many facets that we don't understand and cannot replicate. It may look the same. It may constitute chemical similarity. But on the inside, I don't think it will BE the same. There are so many microcosms of reactions and interactions that I don't think can be replicated by the Mark 1. It sacrifices much for the headlong goal of long-chain esters that may be critical to the flavor, aroma, and mouthfeel. But much like Deckard, I will watch. I will wait. I will see what it can do. It's the most promising foray into maturation science I've seen in awhile and the fact that someone is using GC-MS to quantify spirits makes me happy. This paper has been my Voight-Kampff test. But whether I reacted appropriately remains to be seen. I hope I did.

Otherwise I would not be a very well manufactured replicant.


  1. So you're of the opinion that the paper's release on April 1 was merely a coincidence?

    1. I have contacted them and they said it was not an April Fool's joke.

      That being said, even if it is, it is a very, very expensive April Fool's joke. The GC-MS tests for esters, especially phenolic aldehyde esters, are about $150 per chemical per sample. So if they decided to spend that much on an April Fool's joke then I won't feel terrible about falling for it.

  2. "If I had to guess what forms of energy its using" - Bryan has let slip in the past that he has used light on oak for aging, presumably 'Phase 2'.

  3. A couple more thoughts to add:

    First, being able to measure the major esters probably isn't the entire picture. I'm wondering if this is going to be kind of like artificial vs. natural vanilla. We can replicated the major component - vanillin - just fine, but there are scads of minor components that are necessary to create that real vanilla character. Those minor components, which will necessarily be more difficult to measure, may play important roles in giving the finished product a full character. Steve McCarthy of Clear Creek attempted to use GC/MS analysis of his spirits to improve their quality. After a lot of work, he abandoned that method and went back to organoleptic analysis because he felt like it produced a better product. So the claim that having the same GC/MS chromatogram means that two spirits are the same is not one I'm willing to buy just yet.

    Secondly, I'm just plain skeptical that you can cheat time to that degree. If I could speed up a reaction by many orders of magnitude without causing unwanted side reactions then life would be a whole lot easier in my business. Now maybe he's really figured out something that no one else has, but until there is more independent analysis, I will remain skeptical. It'll also help once the patent is filed and becomes public record, so we can see what the actual methodology is.