Methanol and isobutanol measured by GC in pot still experiment

[Stibnut]

Before a new experiment happens please contact me and tell me exactly what yo will do.

I just look at your xlsx. I hope I can compare your data with our data. It will need a few days.

Questions:
- You measured the alcohol strength by density, right? "SG" in your xlsx really means SG or does it mean density (kg/L)?
- "EtOH g/L" is measured with GC or calculated from the SG?
- "Cum." means cumulated?
- You have no measurement data of the remaining liquid after distillation?
- You write: "Approx. wash makeup". How far away from your numbers can it be?

- The SG actually means SG at 60 F/15.56 C. I believe its density is 0.99901 g/mL at that temp, so some of my numbers are off by 0.1%.
- Calculated from the SG.
- Yes
- That's right. For the first run, it should be very close to pure water given that I took everything down to near 0, although a few mg of methanol may still have been in the still. For the second one I definitely should have run the remaining liquid but didn't.
- I measured the methanol and isobutanol using a syringe, so it was sort of imprecise. I'd say +/- 0.02 mL probably. Couldn't find any working micropipettes; we use a different lab at the same site for the microbiology. The ethanol was measured by graduated cylinder so probably +/1 1 mL out of 75 or 125. The final volume was in a volumetric flask so that should be nearly exact.

One other thing, the GC isn't the best so there are error bars of up to +/- 2% for that. I recorded all the digits it gave me rather than rounding to the 3rd sig fig, but the precision isn't anywhere near as good as the number of digits I entered would imply.

[Hügelwilli]

So you measured the abv with a hydrometer calibrated for 60F, right? And did you correct the reading to the measuring temperature? I could do it for you, if you remember the measuring temperature.

I only checked your starting abv number, and it seems to be wrong:
250ml 96%abv made up to 500ml of course is 48vol%. This is the definition of abv%. You wrote 46.12%abv. Probably a measurement value?
Also 75ml 96%abv made up to 500ml is 96 x 75 / 500 = 14.4%abv, not 14.5.
The problem is, if those numbers are proven wrong, what value do the other density measurements have?

I think I can calculate at least the volatility at the start of each distillation. Later it gets more and more speculation.

[Tummydoc]

Hugelwilli, simple volume dilution calculations are not correct. If you take 1 liter of azeotrope ethanol, and add it to 1 liter H2O, you will end up with a volume that is slightly less than 2 liters! And when you top up the volume to 2 liters, your final ABV will be less than simple dilution predicted.

I think if it like pouring a liter of sand into a liter container of 12mm gravel. Final volume isn't 2 liters

[Hügelwilli]

Of course I know the effect of contraction. But here you are wrong. 250ml 96%abv made up to 500ml of course is 48vol%. And of course it's 500ml. But it needs a bit more than 250ml water to reach 500ml. Simple calculations like that work perfectly as long you don't try to calculate the added water.

[Sulaiman]

Thank you for reporting your information and links,
please continue revealing gems,
irrespective of any criticism.

[Stibnut]

So you measured the abv with a hydrometer calibrated for 60F, right? And did you correct the reading to the measuring temperature? I could do it for you, if you remember the measuring temperature.

I only checked your starting abv number, and it seems to be wrong:
250ml 96%abv made up to 500ml of course is 48vol%. This is the definition of abv%. You wrote 46.12%abv. Probably a measurement value?
Also 75ml 96%abv made up to 500ml is 96 x 75 / 500 = 14.4%abv, not 14.5.
The problem is, if those numbers are proven wrong, what value do the other density measurements have?

I think I can calculate at least the volatility at the start of each distillation. Later it gets more and more speculation.

The ABV reported was a measurement value, which was actually measured very precisely, with a temperature-controlled oscillating U-tube density meter. The makeup of the wash was far less precise, as you can tell.

The 96% alcohol that went in was really around 96.1-96.3% although I do not remember it more precisely. I know it was slightly higher than 96% because the beverage-grade product at the facility I work at has a minimum proof spec of 192 proof/96%, while we essentially never go above 193/96.5%; typically the proof is between 192.2 and 192.6.

I don't know what happened with the second, "low wines" wash ABV. Perhaps I misrecorded it and I had actually measured out 120 mL, and/or perhaps I did not bother to let the flask dry at all before putting in the second wash. I didn't really care about hitting any particular ABV value - mostly I just wanted to sample somewhere at the high end of the "low wines" range, a little above the point where methanol becomes more volatile than ethanol.

Thanks for doing this analysis of my data, by the way. I'd have done this more carefully if I'd realized that someone was going to look at it in this amount of detail.

edit to add: Here is a GC/MS study on a large number of congeners in brandy. Perhaps it would be of some interest for some of the more obscure compounds.

[Hügelwilli]

The pdf has this huge list of data. You can see there many different behaviours. But it is limited: It is a slow single run distillation and this is not the typical way of making brandy. So there is only data of something like 10%abv in the boiler and less. So there is nothing to learn about the behaviour at higher abv. And mash means, that there is always high acidity in the boiler, what changes the volatility at least of the acids and the esters. And it promotes chemical developments like esterification.
Main problem is, that they haven't measured what was in the mash. They measured only the distillate. They even don't write the abv of the wine... No chance to see for example, which part of the ethyl acetate simply distilled over into the distillate and which part was built during distillation. And even no chance to see, what portion of acids distilled over and what portion remained in the boiler.

Interesting is the acetaldehyd data. Acetaldehyde is spread over the whole run. Not "apple smell before the first distillate comes" like it is described often here. Here the question is: Is it new formed or not? If yes, does this happen also at higher abv and more neutral pH? Probably not. This would mean, that there is no way to reduce acetaldehyde with a fores cut at a stripping run.

And personally for me I like the ethy lactate data.

[Stibnut]

That acetaldehyde data surprised me too. I would have to guess that the acetaldehyde that came over later was formed by ethanol oxidation during distillation. I don't know why that would happen though - there would have to be some oxidizing agent present to account for it.

I'm pretty sure that most of the acetaldehyde does come over in the heads provided it's not being formed during the distillation. All the other papers I've seen show it concentrating in the heads, as does an experiment with acetaldehyde I did shortly after the methanol/isobutanol one (but it was worse - among other problems I don't know how much acetaldehyde I lost to evaporation as I added it).

[Hügelwilli]

Right, assuming the numbers are right, you have at the 14.5%abv distillation at the start 0.019568%mol MeOH and 0.008565%mol iBuOH.

And Fraction 0 has 0.15167%mol MeOH and 0.28441%mol iBuOH.

So at 14.5%abv methanols volatility is 7.75 and isobutanols is 33.2.

Those numbers are higher than what I see in publications. Methanol should have 4.91 and isobutanol 14.8.
I think you had rectification at the start of the distillation. But rectification would rise only methanol volatility, not isobutanols. I don't know.

[Delecto]

I don't even play the role of scientist on TV, but I'm OK with logic. With that in mind, I do think your experiments are insightful and instructive. However there are many unknown variables in such a simple sample. I would love to see these tests done on an actual distillation rather than a recombined wash. It is very hard to predict what kind of interactions might occur vis. the numerous other components which make up the slurry that is a typical wash.

[Stibnut]

Right, assuming the numbers are right, you have at the 14.5%abv distillation at the start 0.019568%mol MeOH and 0.008565%mol iBuOH.

And Fraction 0 has 0.15167%mol MeOH and 0.28441%mol iBuOH.

So at 14.5%abv methanols volatility is 7.75 and isobutanols is 33.2.

Those numbers are higher than what I see in publications. Methanol should have 4.91 and isobutanol 14.8.
I think you had rectification at the start of the distillation. But rectification would rise only methanol volatility, not isobutanols. I don't know.

I agree that there was some rectification at the start of the distillation. The flasks i used have an arm that slants slightly upward for perhaps 20 cm before sharply bending ~105 degrees to straight down to accomodate the Graham condenser. I can see reflux moving down the arm as it heats up, before recovering any distillate. I don't know of anything to explain isobutanol's high volatility either.

[Hügelwilli]

With such a still you get practical true results. But only for this specific situation. To predict the result for other circumstances is impossible this way. At least if you want more detailed information than only a rough understanding like to show, that methanol is not mainly in the heads.
This is why there are special stills for VLE-experiments, the "Othmer still" for example. With them you can get more universal data. But on the other hand, those results are less practical than yours, because they don't take into account the passive reflux at the start of the distillation every normal still provides.
If you want to predict something, not only to watch at it, you need data from a Othmer still or similar. But then it gets theoretical. It doesn't picture exactly practice anymore.