[By the end of the post, connections start to be made that I didn’t have a good enough memory to make from the beginning. The existence of this paper was a tip from a particularly smart reader. The punchline may be that components of the mythic rum oil may come from at least two channels. The first is the splitting of glycocides by the enzymatic activity of yeast while the second may be from carotenoids present in the cane itself. Tons of work still needs to be done, but these are some good preliminary guesses of where to look. At the very least, they may point to realizing more terroir in rum, molasses based or otherwise.]
Here is a unique paper, Less Volatile alcohols Esters and Hydrocarbons in a Raw Australian Rum, 1975 (Bundaberg!), which may have a follow up if ILL can track it down. [The follow up is A new approach to the identification of flavour components in rum from the Australian Wine, Brewing, and Spirit Review, 1973. This brief paper was in the bibliography of the other and offers a great summary of what will follow.]
I read this after reading two different modern rum-GCMS papers which were kind of useless for the purpose of learning more about rum history or production. I’ve been aiming to highlight a unique thesis I found with some fantastic explanations of the evolution of chromatography, but I’m short on time and I think I may contact the author first to ask some questions.
In the paper, D.A. Allen reads two early (1966, 1970) rum-GCMS papers and wants to play along, but doesn’t have access to the same equipment. The authors used pentanes to extract congeners from very small samples of rums then analyzed them with GCMS to name volatile components. I’ve actually played with pentane extraction to produce artful creations, but that is another story for another day. Allen could not work with such small samples so he comes up with the novel idea of collecting fusel oil from the side stream of the Bundaberg production then toying with it. Allen’s idea is comparative to that studies that inspired it because most of the unique compounds everyone is looking for are less volatile. The paper ends a little bit abruptly, but he ends up finding the notorius reisling congeners TDN.
I’ll try to describe a little bit of the experiment, but what I should first note with disappointment is that Allen never organoleptically describes anything he is working on. Is he working with that peculiar, wonderful, desired rum oil or is this just low volatility junk? We never really find out here, but maybe we will in his other paper. The whole significance of this paper becomes the old fashioned extraction procedures he uses which may help the contemporary small scale fine producer. Another new possibility is that rum oil congeners may have appreciated in value enough (with our new found fine market) that it is now economically viable to harvest them from a formerly discarded fusel oil fraction. Maybe it is already done for the fragrance industry? Who knows.
Distillery oil, removed in litre quantities from the side of the still was shown to contain these compounds and can be considered as a concentrate of the higher boiling point flavour compounds of rum.
Fractional distillation of distillery oil produced « fusel oil » containing the higher alcohols (n-propanol, isobutanol, isoamyl alcohol and active amyl alcohol, BP to 132°C) and a residue termed « rum » oil containing compounds with a higher boiling point than isoamyl alcohol. Only the analysis of the « rum » oil will be discussed.
Allen uses both a Lecky and Ewell still and a Bower and Cooke still to purify the fractions for analysis. He has citations for each still and it may be helpful to dig them up to see what they were like. A lot of this equipment is still very useful.
He has got an entire liter of rum oil and does not say how it smells. There are a lot of esters in the oil and they get hydrolyzed with sodium hydroxide to concentrate the remaining compounds. The hydrolysate gets fractionally distilled and the fractions analysed. Part of hydrolysate is alcohols that were liberated from the esters by the sodium hydroxide. Due to how the sample was separated from fusel oil, some compounds like acetals reported in rum oil by others may not have survived.
Allen goes on to perform continuous liquid-liquid extraction on multiple liter batches of raw rum. Its seems like he does five batches and winds up with 5 liters of pentane to distill from. The non volatile product is an oil and the volatile product is split into two fractions. The ethanol was in the first fraction and the second fraction was an oil-water azeotropic mixture. The oil was separated, dried with anhydrous calcium chloride and added to the water-free residue in the pot. It would be nice to know how they smell before he blended them together!
This oil gets redistilled in the same apparatus and separated into two fractions collected up to 132°C so everything is well over the boiling point of water. What isn’t clear is if pentane is used in this distillation. These days this distillation would be done under vacuum and a teflon coated spinning band distillation column would be used because holdup, or the clinging of liquid to the glass apparatus, starts to become significant. Descriptions of what is explicitly happening by now have become a little disjointed and I’m having trouble following the transitions.
This oil was redistilled in the same apparatus and separated into two fractions. The larger fraction, collected up to 132°C, contained the higher alcohols to isoamyl alcohol and was called fusel oil. The residue (100 mL) in the pot contained compounds with higher boiling point than 132°C and was called « rum » oil. After four more such distillations, the combined residue amounted to 500 mL.
If any of this smelled really good, wouldn’t he be likely to mention it? Wouldn’t he be likely to show it to a distiller and get some gears turning? Wouldn’t Bundaberg rum be less likely to be so lame?
In the next step, the specialized stills get some use which apparently feature vacuum and a series of 10mL sample were collected until the temperature hit a certain point. The pressure was dropped and more 10mL samples collected. This multi stage pressure drop to avoid decomposition may have been because the equipment was a little more primitive than what we commonly use today. All of the collected fractions see some spectroscopy to identify what they are.
The rum oil goes through some more hydrolysis with more sodium hydroxide with the products extracted into more pentane.
The paper seems to get cut short after Allen identifies 1,1,6-trimethyl-1,2-dihydronaphthalene. Allen does not use the modern abbreviation of TDN, but this is a congener that is infamous in aged Reislings and is responsible for the petrol character which at the right levels is often prized. Allen drops a little bit of history on this compound but does not mention wine at all. I linked to this paper on TDN in the beginning, but here it is again if anyone wants a primer.
The entire work seems to be the basis of a masters thesis to which the next paper I’ve requested may add to.
I really don’t know what to make of the TDN discovery. Allen does torture his sample and we should remember that in continuous column distillates, this fraction is mostly discarded. Google searches for 1,1,6-trimethyl-1,2-dihydronaphthalene +rum yield nothing.
But, when you read the AWRI paper, TDN is noted as related to carotenoids and extra smart blog reader Matt Power brought them up recently (Matt actually inspired the tracking down of this paper after mentioning 1,1,6-trimethyl-1,2-dihydronaphthalene, but I did not immediately connect the dots):
Are components of rum oils microbiologically derived in these manners, rather than from the canes themselves? Carotenoid bio-decomposition is known to lead to a spectacular array of interesting chemicals
This comment come from the Arroyo’s Oidium post about ethyl tiglate and relates to my hypothesis that the peculiar character of rum oil comes from the splitting of glycosides by the enzymatic action of alt yeasts like Schizosacharomyces Pombe. Rum oil may be more complex and the product of more mechanisms.
Outlining the mechanisms may even unlock the potential for finding more terroir in rum from molasses. A rum can only tell us a story of a place if we learn to read it.
[It may be possible to take a modern GC-MS look at a heavy rum and try to categorize all the low volatility congeners found. This may give us a distribution of what channels they come from.]