science question here about drop in proof

[LWTCS]

I think your conclusions are completely plausible.

I don't k ow how to copy the YouTube link on this phone.
But a simple Google search with the key words "vacuum, barrel collaps" should pull up the video.

It's a super basic experiment but really demonstrates the power of a vacuum when hot vapor is super cooled.

[Fivey]

I think the most obvious explanation here is based on the reminder that an alcometer may be marked in ABV but it is not measuring ABV. It is measuring specific gravity relative to water, and assuming an ethanol/water mix at a specific temperature, implying an ABV scale. While for most of a run this is a reasonable approximation, it is not the case in foreshots, where the liquid is a crazy mix of many substances (that’s the reason we toss fores, right?) the mixture of which the alcometer is not calibrated for. So when your alcometer is reading low for fore’s than it does for the first jar of heads, that should not be a surprise. Or, at least, you should realise that it isn’t telling you anything useful.

[Kareltje]

You could certainly have a valid point.
There is only one article I know that gives amounts of different substances in both stripping and spirit run:
http://www.scielo.br/pdf/sa/v67n3/a05v67n3.pdf" onclick="window.open(this.href);return false;" rel="nofollow

In the stripping run the amount of heigher alcohol and summed congeners are both about 1 gr per 100 mL absolute alcohol: more than 1 % by weight. And diminishing.
In the spirit run on the other hand both are less than 0.4 gr per 100 mL, so less than 0.5 % by weight. And augmenting.

Study made with cachaca, so for other ingredients it may be very different.

[Kareltje]

We do not know how much of every substance is present in our fores, heads, hearts or tails, but we can try to make an educated guess.
We can find the density of several substances easily on internet, nowadays.
water: 1.000
ethanol: 0.789
methanol: 0.792
ethylacetate: 0.902
methylacetate: 0.932
acetaldehyde: 0.788
aceton: 0.784
2-methyl-1-butanol: 0.815
pentanol: 0.814
2-propanol: 0.786
1-propanol: 0.803
2-butanol: 0.806
1-butanol: 0.810

So we see: these substances have specific weights very close to that of ethanol. Maybe they do not contract with water as ethanol does, but apart from that they are interchangable with ethanol.
Which means, that these nasties will lower the specific gravity of the mixture the same way as ethanol does.
So I think that a considerable amount of these stuff will cause the alcometer to sink a bit lower, thereby suggesting a higher % rather than a lower %.

I guess the higher esters and other congeners may have a higher density, but they will be seen in the tails, so they are not relevant for the subject at hand.

Be my guest to falsify my reasoning!

[Fivey]

That is pretty much the case, but again we are making use of supposes. - I am simply pointing out that it is wrong to think an alcometer tells you definitively how much alcohol is in the test sample, which is the starting point for much of the rest of the discussion. The reality is we have insufficient information for each cut to formulate any real hypotheses, though we might make some wild ‘reckons’. I’m not saying speculating is a bad idea, but one must not be dogmatic when we are working from incomplete information.

[Kareltje]

And as we always have to work with incomplete information, we must never ever be dogmatic. I fully agree with you on that.

But then, this reasoning brought to its consequence, we can not solve the problem at all.
But there is no fun in that.

[Fivey]

And as we always have to work with incomplete information, we must never ever be dogmatic. I fully agree with you on that.

But then, this reasoning brought to its consequence, we can not solve the problem at all.
But there is no fun in that.

Well, only if you give up. If someone was really interested in fully understanding what is going on, you’d try and collect much more data, more accurately an completely. Lots of small samples, run through a GC. Lots of temperature readings at different points in the still, plus in the distillate, all logged against time. Accurate power input readings, logged against time.

Without more data, you can only speculate as to what us going on. With more data, you can develop a reasonable hypothesis and suggest a model.

They always say the best discoveries are made, not with a “eureka!” but with an: “uh... huh, that’s odd”. So I fully support anyone who is curious about oddities they notice, and the dedicated individuals who doggedly chase the rabbit down the rabbit hole to find out what is going on.

[kiwi Bruce]

What your seeing is very typical of a pot still or a column that's run "wide open" ie without packing. lets say you have a wash that's 10% ABV, it's in your pot and ready to go...as the temp in the liquid goes up and starts to get close to the 170-175 deg F range the vapor that starts to come over (what we call " foreshots") this is a mix of sti# like methanol, acetone etc some ethanol...AND WATER, important this bit, we tend to forget the water part, but most of this is water...then the heads start, depending on your still, say around 175-180 deg F...again water, shi#, and very high proof ethanol...this we keep to redistill...now we get to the heart of the matter...around 180ish the spirit is high in proof with water ---now the water will be at a high % too, around 80 - 75 % of the vapor. In a pot the temp will appear to stall, but what's happening is all the heat that's being added is being used to turn the wash to vapor (the commercial distillers lower the heat input at this point (run it like you were grilling a steak, high heat to start and seal in the juices, then turn it down and cook the inside slowly) the spirit will come over in a higher proportion at a lower temp than if you "push" it...water uses more heat than ethanol to go over...as this happens the total amount of ethanol to water concentration changes and as this concentration drops into the low percentage points the temp will start to rise again as the water now needs more heat to vaporize and the wash will eventually reach 212 deg F waters boiling point---the whole time this is happening the spirit concentration is diminishing to zero at 212 F while the water concentration is increasing to 100%, it's distillation physics. Water boils at 212 F and ethanol at 173 F...but the mixture with start to evaporate (what we call distill) at a different temp range and concentration between these two points.
There are many reasons why the most distillers don't "see" this, the main one being "in my opinion" that many people up the heat at the first signs of foreshots and again "in my opinion" rush or push, the distillation.
Who cares...it dosn't appear to affect the product in any way.
So you wanted to know WHY you were seeing the spirit % drop though out the course of a run...I hope this helps---Kiwi

[Fivey]

What your seeing is very typical of a pot still or a column that's run "wide open" ie without packing. lets say you have a wash that's 10% ABV, it's in your pot and ready to go...as the temp in the liquid goes up and starts to get close to the 170-175 deg F range the vapor that starts to come over (what we call " foreshots") this is a mix of sti# like methanol, acetone etc some ethanol...AND WATER, important this bit, we tend to forget the water part, but most of this is water...then the heads start, depending on your still, say around 175-180 deg F...again water, shi#, and very high proof ethanol...this we keep to redistill...now we get to the heart of the matter...around 180ish the spirit is high in proof with water ---now the water will be at a high % too, around 80 - 75 % of the vapor. In a pot the temp will appear to stall, but what's happening is all the heat that's being added is being used to turn the wash to vapor (the commercial distillers lower the heat input at this point (run it like you were grilling a steak, high heat to start and seal in the juices, then turn it down and cook the inside slowly) the spirit will come over in a higher proportion at a lower temp than if you "push" it...water uses more heat than ethanol to go over...as this happens the total amount of ethanol to water concentration changes and as this concentration drops into the low percentage points the temp will start to rise again as the water now needs more heat to vaporize and the wash will eventually reach 212 deg F waters boiling point---the whole time this is happening the spirit concentration is diminishing to zero at 212 F while the water concentration is increasing to 100%, it's distillation physics. Water boils at 212 F and ethanol at 173 F...but the mixture with start to evaporate (what we call distill) at a different temp range and concentration between these two points.
There are many reasons why the most distillers don't "see" this, the main one being "in my opinion" that many people up the heat at the first signs of foreshots and again "in my opinion" rush or push, the distillation.
Who cares...it dosn't appear to affect the product in any way.
So you wanted to know WHY you were seeing the spirit % drop though out the course of a run...I hope this helps---Kiwi

I think the OP question was not actually why the proof drops during the run - I think that is well-established. Using the word "drops" in the thread title is a bit misleading, as it implies a reduction at a later point in time. My understanding was the OP was asking why their initial samples of distillate showed a lower ABV reading on an alcometer than subsequent ones, it does seem a bit counterintuitive. We expect to follow the phase diagram to the left as the alcohol is removed from the wash (as you describe above), continually getting a lower ETOH concentration in the output. So my understanding of the question, to rephrase it, would be: what is the mechanism that allows the proof of the output to rise during the run (very specifically, near the start)?

[Kareltje]

That is how I understood the OP too.
I have never seen it mentioned, not even in accurate scientific experiments, but I noticed it myself rather often. About half of my runs show this effect.

One of the causes might be the amount of non-water/non-ethanol compounds, but I think we can rule that out.
Another might be traces of tails of the previous run that are left in the condensor, but it is easy to rule this out too.

So there must be a mechanism that causes the first bit of a run to be of lower % than the following bits.

I am chewing on this subject and several observations that seem connected since jon1163 started this thread.
We are not the only distillers in the world, so I guess someone else in some other field has already solved it, but until he or she comes forward to teach us, we will have to search for ourselves.

[still_stirrin]

So there must be a mechanism that causes the first bit of a run to be of lower % than the following bits. I am chewing on this subject....

Remember, your hydrometer (Proof & Traille) only measures density with a calibrated scale that relates ethanol to water ONLY. It doesn’t consider other constituents in the test jar/parrot. So, constituents with a higher volatility (lower boilng point) but have a higher density than alcohol, will accumulate in the test jar early in the run...like the very front end. I gave this explanation earlier in this thread:

...At the start of your reflux run, you’ll get some foreshots, components with a high volatility (higher than ethanol). Those constituents include acetone (784 kg/m3), acetaldehyde (788 kg/m3), and ethyl acetate (902 kg/m3), among others.

And with ethyl alcohol (789 kg/m3), you can see that the “potential” exists for the heads to actually have a slightly higher density than your “pure” ethanol hearts. And in certain washes, the proportion of ethyl acetate (fingernail polish remover) is high, making the hydrometer (Proof & Traille) rise slightly at the start when the heads are coming out, and then sink as you get to the lighter (less dense, not less volatile) constituents.

So, you can theorize that ethyl acetate (a component of the early heads because of its volatility) will show up in your test jar at the beginning of condensate production. As the run progresses further, the composition of the run clears up to the early hearts and your density will reduce slightly as you collect clean(er) alcohol in the water. As a result, the “measured %ABV” will rise slightly. Nothing magic or mysterious really unless you’re a skeptic. It’s the “nature of things”.

Potstills would exhibit this same phenomenon except that the potstill is much, much less effective at separation of the fractions, or compressing the heads and early hearts like a reflux column will. The hydrometer readings are not near as sensitive to minute density changes as the condensate is a greater collection of a wider mixture of constituents.

Again, what your boiler produces first are the constituents with the highest volatility and that will be proportional to the amount of those constituents in the mixture, be-it a wash or low wines. So, if your wash or low wines smell excessively fruity, then you can expect more of the foreshots and early heads will contain ethyl acetate.
ss

[Fivey]

So there must be a mechanism that causes the first bit of a run to be of lower % than the following bits. I am chewing on this subject....

Remember, your hydrometer (Proof & Traille) only measures density with a calibrated scale that relates ethanol to water ONLY. It doesn’t consider other constituents in the test jar/parrot.

I agree SS - this was my own argument about 9 posts ago, that we can’t assume a tool that reads ABV actually tells us what the ABV is.

So, constituents with a higher volatility (lower boilng point) but have a higher density than alcohol, will accumulate in the test jar early in the run...like the very front end. I gave this explanation earlier in this thread:

...At the start of your reflux run, you’ll get some foreshots, components with a high volatility (higher than ethanol). Those constituents include acetone (784 kg/m3), acetaldehyde (788 kg/m3), and ethyl acetate (902 kg/m3), among others.

And with ethyl alcohol (789 kg/m3), you can see that the “potential” exists for the heads to actually have a slightly higher density than your “pure” ethanol hearts.

This was my thinking too, but Kareltje believes that the more volatile molecules will be less dense and the alcometer will sink, not rise, with more of these chemicals. I think I will put the volatility numbers aside and look for actual density numbers. [edit: ignore this last bit, of course the numbers have already been provided above, I was mistakenly thinking they were some measure of volatility]

[Kareltje]

The article I found is about pot stills and indeed shows high amounts of aldehydes and esters in the first samples, especially in the stripping runs. And indeed, both methyl- and ethylacetate have densities of 0.932 and 0.902, quite a bit higher than the range of 0.78 to 0.82 of alcohols. And what is more: the amounts of higher alcohols and congeners, with probably higher densities, were in the order of grams per 100 ml.

We can not end this speculative discussion without accurate measurements, but I see the effect on both stripping runs and spirit runs.
(Found a spirit run of grain with my 2 m long spiral riser: first sample 86 %ABV, second 91.5 %ABV (well, speaking according to your remark: with the same sg's of these alcohol/water mixtures: 0.83722 and 0.81819 ). )

I will collect data of my coming runs, to see if I can find a pattern. It would be nice if others would do the same, to exclude any bias or personal pecularities.

[Kareltje]

One thing I already observed seems to favour the argument of much cogeners: I was redistilling a high wine and observed the effect: first 88, second 92 %.
I diluted both with the same amount of water and lo and behold: the first one got very milky and the second one stayed cristal clear. After a few days the first one even showed some sediment and a thin fleece on top.
An other run of another wine showed the same effect, but less pronounced.