the distillation process

[flyingdutchman]

Ok,
I thought it might be helpful to have an open discussion about the science behind how distillation works (no math). I know there are a lot of really good people who have made great whiskey, azeotope vodka, brandy rum etc with no need to know the hows and whys of distillation. I think thats is great and if that is all they wanted i say more power to them that in and of itself is a great achievement.
However alot of people here are gear heads, trades people etc who want to build a better mouse trap and it doesnt hurt to know how it works b-4 you try that.

So here is a link that does a better job than I ever will job of describing the hows and whys distillation/separations.
http://www.chemguide.co.uk/physical/pha ... d.html#top
Note: page 1 describes Rauolts law and Ideal mixtures you need to really understand these to then go to the second page (link at bottom) to understand the process of how distillation/separations work
Once you read that ask any questions you like and I will absolutely do my best to answer them. I encourage other to read along. hell I dont know everything not even close so correct me if I am wrong.

if you want more here is Zubricks guide to survivng organic chem, he discusses the nuts and bolts of simple, fractional and vacuum distillations starting on page 152
http://survival-training.info/Library/C ... ubrick.pdf
But dont get sidetracked first make sure you understand exactly how and why distillations work.

[LWTCS]

If the molecules are escaping easily from the surface, it must mean that the intermolecular forces are relatively weak. That means that you won't have to supply so much heat to break them completely and boil the liquid.

The liquid with the higher vapour pressure at a particular temperature is the one with the lower boiling point.


Really the foundation of what I have spent most of my time thinking about as I shift and hold bulk amounts of alcohol to the upper most region within my apparatus.
That and also how each system behaves as it absorbs and releases heat as the course of the run progresses.

[flyingdutchman]

Yep,
Thats the thing its just putting into words what we already understand as distillers. The cool thing about his website is he says it in plain english and fills in the tiny gaps from what I know and what I can put into words. Sometimes when I read about things it lets me see it from a slightly different angle and gives me new ideas (or makes for more questions) as well.

[Frosteecat]

Excellent! As a former cook who focused on sauces, complex and unorthodox flavor marriages, etc, I am really interested in the science behind flavor generation, aging and combination impacts, etc. I just need to learn what it is we CAN & CANNOT do to the end result with our stills and aging & flavoring. There is quite a psychological aspect which can be looked at as well...maybe by combining all aspects we can make the "perfect drop"? Here's a link to a scientific examination of flavor perception and process & how to make changes to our advantage. MR posted a link in his Boka thread as well.
http://www.scientificsocieties.com/jib/ ... _5_287.pdf" onclick="window.open(this.href);return false;" rel="nofollow

Here is a picture of a revised flavor wheel enclosed in that document...they draw some interesting conclusions about the correlations between "expectation and perception" of flavors based on appearance, etc. Fascinating adjectives used here....hope you can see it ok...



*edit*
fixed broken link & added flavor wheel

[flyingdutchman]

For anyone reading this make sure you go to the second page after reading the first (there is a link at the bottom)
or just use this link
http://www.chemguide.co.uk/physical/pha ... t.html#top

But for me a big part of what they are saying on the first page is this...

If you boil a liquid mixture, you would expect to find that the more volatile substance escapes to form a vapour more easily than the less volatile one.

That means that in the case we've been talking about, you would expect to find a higher proportion of B (the more volatile component) in the vapour than in the liquid (it came from). You can discover this composition by condensing the vapour and analysing it.

So when you boil it both substances escape B and A (not just one) BUT the more volatile substance escapes in a greater percentage. How much more is a function of just how much more volatile B is than A

Its not about who is heavy and who is light its about who is more volatile.

[flyingdutchman]

Now think about what that implies...

If we have two substance in the liquid A and B, and B is WAY more volatile than A then...
Every time we boil it the resulting vapor will have a whole lot more of B than A (so we can easily separate B from A in just a few plates or even 1 plate)

But if B is just barely more volatile than A then....
Everytime we boil it we get just the tinest bit more of B than A, in the vapor (so it takes a huge number of plates to ever hope of separating B from A)

[Frosteecat]

So how do we make B more volatile than A so we have clearer separation?

[flyingdutchman]

Thats the problem,
IDK of any (reasonable) way to make one substance more volatile than the other, vapor pressure = volatility is a physical property of the substance. You have to play the cards your dealt. So then your options are to
1. Increase the number of plates to get a better separation (running it twice is like doubling your plates)
2. If you have very little of B (more volatile) and alot of A (close but less volatile) you just bleed off a mix of A& B until B is all gone.

Its like what we do with Ethyl Acetate and Ethanol. Look at the differences in BP boiling points just ~1C. Since BP is just the temp where vapor pressure = atmospheric pressure (14.7 psi) what were really saying is look how close their vapor pressures are (or volatility).
There will be a tiny amount more of EA in the vapor at every plate BUT no way do we really isolate EA in the still. At best we just have a slightly more enriched vapor at the top. What we really do is just sacrifice a fair amount of ethanol to bleed off the EA. Luckily we just need to get the EA below our perception level for tasting (I have no idea what that level is srry)

here is another way of looking at it, the difference in BP between ethanol and water is ~ 22C. It takes idk something like 10 plates to get a 10% ethanol in water solution (wash) to separate the ethanol to 95% purity from the water. So think about how many plates you would need to even try and separate something with just a 1C difference in BP temp to 95%.

ps You could hydrolyze the EA with a strong base (like sodium hydroxide) to create ethanol and acetic acid (vinegar) but thats is a synthetic procedure and not changing the physical properties of EA just creating 2 new chenical entities. (there is a thread on this)

Here is a list of common substance that probably occur in your wash
TYPICAL
Acetone 56.5C
Methanol 64C
Ethyl acetate 77.1C
Ethanol 78C
2-Propanol 82C
1-Propanol 97C
Water 100C
Isobutanol 107.9 C
Butanol 116C
2-methyl-1-butanol 128.7C
Isopentanol 131.1 C
Amyl alcohol 137.8 C
Furfural 161C (322F

[Frosteecat]

Methanol 64C

Ethyl acetate 77.1C
Ethanol 78C

2-Propanol 82C


Hmmmm....that's a fairly narrow band sandwiched in a bit wider temp point. How about we just try to get those two out and then do a secondary, fine temp separation of E Acetate and Ethanol?

[LWTCS]

How about we just try to get those two out and then do a secondary, fine temp separation of E Acetate and Ethanol?

That's what the column guys are doing Frosty. In fact, before the "Flute" hit the scene,,,thats the stuff that dominated much of the dialog here....
The group that is most sensitive spends the most amount of time on running, diluting, reruning etc. And also many do practice techniques that exploit optimal heads compression.....

Big R has also talked about rerunning till your spent boiler charge is clean like water....A plain jane low wines charge stinks after the alcohol is liberated....so he runs and reruns his till that funk is worked out.....

[Frosteecat]

So we have already explored inline reboilers and so forth? I don't like the idea of introduction of tertiary compounds to break azeotrope, etc. unless there is a simple and effective way to do it. How much of this optimum separation theorizing is predicated on column height? For instance, with "optimal" packing, how high does a 2" column need to be to reach the 96% range?

[flyingdutchman]

Frost,
LWTC is right (plus add to that if they triple distill then they have effectively have3xs as many plates).
Its a great question. Id love to say i have a definite answer for you but I dont. We know for sure every time we rerun the column the process is not 100% effecient so were taking a hit yield wise (if thats the price for ultra clean your willing to pay cool).
Others might say use high reflux ratios an ultra slow take off and more plates. (thats gonna cost you time and electricity/propane)
see nothing is free
We know that we get a separation based on vapor pressure differences (of the different substances). I dont think the number of components changes that but I could be wrong.
Here's the questionable part. Every general statement has exceptions sometimes big ones. We are assuming that the vapor pressure of each component is exactly that of the substance in pure form (If its an ideal mixture). But hold on what if the interactions between methanol and isopropanol effects how easily EA (ethyl acetate) now escapes and now it cant esacpe the solution as easily? Thats a non-ideal mixture! Now the vapor pressure of EA goes down (so its BP goes up). In this case its better off removing methanol and isopropanol (does this truly happen Idk). See the whole discussion for simplicity sake has been centered around ideal solutions but this isnt truly an ideal solution.

Ps
I personnally have no interest adding a known carcinogen like benzene or cyclohexane to break azeotrope (or molecular sieves) just so i can dilute it afterwards.

[LWTCS]

So we have already explored inline reboilers and so forth?Well that is a work in progress. There is a limited amount of experience but the data is growing. PP's build should help move things forward
I don't like the idea of introduction of tertiary compounds to break azeotrope, None of the nutral guys do. Thats why you gotta let you equipment sufficiently equalize and draw heads by using your sences etc. unless there is a simple and effective way to do it. How much of this optimum separation theorizing is predicated on column height? For instance, with "optimal" packing, how high does a 2" column need to be to reach the 96% range?
This is a great place to have Rad step in and offer up an explanation.

[flyingdutchman]

I believe Nixon and the parent site both claim that mesh/scrubbers make the best packing based on surface area. In the lab mesh is the 1st choice everytime for me.I think the parent website says..
1) ~4 inches of mesh = 1 theoretical plate
2) About 10 plates get you from a 10% wash to ~95%

But more input would be good!

[Prairiepiss]

So we have already explored inline reboilers and so forth?Well that is a work in progress. There is a limited amount of experience but the data is growing. PP's build should help move things forward
I don't like the idea of introduction of tertiary compounds to break azeotrope, None of the nutral guys do. Thats why you gotta let you equipment sufficiently equalize and draw heads by using your sences etc. unless there is a simple and effective way to do it. How much of this optimum separation theorizing is predicated on column height? For instance, with "optimal" packing, how high does a 2" column need to be to reach the 96% range?
With my findings on my new build I think I can hit it with 12" of packing. If I fix my unforeseen screwup. But I know that doesn't really answer the question.
This is a great place to have Rad step in and offer up an explanation.

I would agree. I would love to see Rad throw some stuff on this table.

Edited same color hard to read.

[Frosteecat]

Cool. I appreciate the rehash of old discussions you guys are willing to do for me. I still think there is new ground to break in craft distillation, and with the right blend of linear and abstract thinkers bouncing ideas and builds around, we are going to find some innovative approaches. There should be a patent lawyer in the mix because I truly think we are going to find some methods that take us off in new territories. The amount of knowledge here is pretty stunning...

My goal personally is to make good tasting liquor, in all of its various forms. But building and having at my disposal an efficient, cost effective and easily controlled fractioning device seems intuitive to the overall pursuit. I don't think I'll be striving for a lot of neutral product, but if and when I do, I plan on doing it well!

I'll achieve that thanks to the wide spectrum of input here...I'm sure of it!
My high school alma mater's motto was "Age Quod Agis"... "Do well, whatever you do!" words to live by!

Ok...where's Rad!!??

[rad14701]

So we have already explored inline reboilers and so forth?Well that is a work in progress. There is a limited amount of experience but the data is growing. PP's build should help move things forward
I don't like the idea of introduction of tertiary compounds to break azeotrope, None of the nutral guys do. Thats why you gotta let you equipment sufficiently equalize and draw heads by using your sences etc. unless there is a simple and effective way to do it. How much of this optimum separation theorizing is predicated on column height? For instance, with "optimal" packing, how high does a 2" column need to be to reach the 96% range?
This is a great place to have Rad step in and offer up an explanation.

The answer to the optimal packing question is "it depends"... It's relatively easy to hit 94% - 95% with only several HETP's, even with a short column and iffy packing... But getting from 95% to 96.5% (the highest %ABV without chemicals or vacuum) can be a whole lot harder... Most here don't concern themselves with shooting for anything higher than 95% and consider it luck if they get there...

Here are some figures based on the HETP's Calculator on the parent site... A 2" (~50mm) column 48" (~1.25m) tall with a reflux ratio of 1 and using stainless steel scrubbers and 1500W of heat can produce 94.9% ABV spirits... Increasing the reflux ration to 10 only raises the spirits to 95.1% ABV... However, increasing the column height to 60" (1.5m) with a reflux ratio of 1 can produce 95.1% ABV and a reflux ratio of 3 the results would be 95.2% ABV... Unfortunately, the calculator ceases to function beyond there... I have to check the underlying code to find out why it fails at some point... One thing we do know, however, is that going taller than ~30:1 height:diameter starts showing diminished returns due to the amount of heat required and the amount of heat lost which induces flooding... And a 36" (.9m) packed column is capable of producing 94.2% ABV with a reflux ratio of 1... Heck, even an 18" (.5m) column with a reflux ratio of 1 should be able to produce 92% ABV... But these are all theoretical numbers and your mileage may vary... And remember, this is just the packed column section only, without the condenser head above it...

Clear as mud... I really need to validate the HETP's Calculator because it seems to indicate that height has a greater impact than reflux ratio which seems counter intuitive from my point of view... We know that additional reflux should increase overall purity but this is not proven out by the calculator as well as expected...

Now that we're all confused...

[Frosteecat]

Can you explain "reflux ratio" a little bit and what 1 represents on that scale?

[flyingdutchman]

truth be told I think RAD or PP can answer that a whole lot better than me. I have never used a reflux type still in the lab or even seen one before coming to this site. We have always use what zubrick describes as fractionating columns (run slow as heck)! Even then with proper application of heat I can put those in what we call total reflux, nothing exits the column. But the difference is that my scale is typically 50 - 200 mL

[flyingdutchman]

So,
While I am hoping rad or PP will see this and get to your question Frost, I would like to just make a quick list of important points from page 1 (for any one reading along) so that we can go on to discuss page 2, the actual distillation part.
Just in general terms...
Liquids
1) A liquid has forces that help to keep it together, a big one is intermolecular attraction: the amount of how much the molecules want to "stick together" (hydrogen bonding can play a huge part in this)
2) Molecules are in motion in the liquid (kinetics) and the faster a molecule is moving the more likely it can escape .
3) When a molecule escapes it becomes what we call a gas.
4) increased heat means the molecules move faster are are even more likely to escape.
5) vapor pressure is a measure of the pressure exerted by the molecules escaping the liquid, so think of it as a measure of how easily the molecules escape.
6) vapor pressure is approx = to volatility
7) when the vapor pressure equals the pressure exerted by atmospheric pressure ( ~14.7 psi) the liquid will boil.
8.) So boiling point is the temp at which vapor pressure equals atmospheric pressure (see # 4)
9) Finally increased vapor pressure means increased volatility and as a result a DECREASED boiling point.

Mixtures and Raoults law
in general terms (for ideal mixture)...
1) In ideal mixture of liquids each individual component will behave the way it does as a pure substance (same amount of intermolecular attraction)
2) When we mix 2 substances (say A and B) with different boiling points (BP) the new liquid will have a new boiling point (BP) that is somewhere between the lowest (say B) and the highest (A).
3) What determines where that BP lies exactly depends on the ratio of how much of A we have versus how much of B we have (like a fraction).
4) If we have way more of B than A then the BP will still be between the two but will be way closer to the BP of B. If its way more of A then its still in the middle but way closer to the BP of A.
This makes sense with what we see as distillers at a the beginning of a run when we are taking off mostly ethanol isnt the temp really close to the BP of ethanol (BP ~78C). And later at the end of the run isnt the BP going up and up as we have more and more water (BP ~100C)!
7) Here is the biggie which I will repeat because its really important. Just because we mixed them (in an ideal mixture) we did not change thier individual physical properties. So when we boil a mixture of A and B, and B is more volatile (lower boiling point), then both will escape the liquid (not just B). BUT more of B escapes than A exactly because B is more volatile. We are separating them based on thier vapor pressure! NOT because one is lighter than the other. (water at 18 amu weighs less than ethanol 46 amu but ethanol is more volatile)
8.) A phase diagram does an excellent job of showing you visually what we are talking about in #7
Rad has one here in the forums I have seen it and there is an interactive ones on the parent site here http://homedistiller.org/theory/refluxdesign/steps having read this I encourage you to play with it and think about what we are talkng about.

Ps If i misstated stuff here pls correct me I dont want to give out false info
ps ps--on #7 please dont go down the VM road that last step is a 2nd (and ingenious) separation based on a different physical property than vapor pressure, namely the density of a gas.

[LWTCS]

I believe Harry talks about this RR at several places here some where and about..

I could be wrong but as I recall a 10/1 ratio is optimal for azeotrope for a typical column..Sorry, too lazy to find the thread..
Myles can prolly firm that RR up.

[rad14701]

Can you explain "reflux ratio" a little bit and what 1 represents on that scale?

Reflux ratio is the ratio between how much spirits are returned into the structured packing for reboiling in relation to how much is collected... So 1:1 means 1 part returned and 1 part collected... 10:1 means 10 parts returned and 1 part collected... Virtually all reflux columns work by this principle... We call them reflux columns because we don't collect at different plate levels when running structured packing... It would require a plated column to collect true fractions...

[Frosteecat]

Thank you! So is there also a term for plate takeoff ratios?

[LWTCS]

Say again Frosty?

[Frosteecat]

I guess I'm thinking slant plates (or even perf or bubble?) and if there is also a ratio/scale for how much is taken off based on heat input, reflux etc.
The reason I'm asking is my current design is a bit unorthodox and I'm trying to digest ALL the different Methods of column reflux/fractioning and want to make sure I'm up to speed on the different takeoff and control methods, terminology and rates/ratios.

[Prairiepiss]

A plated still as far as hobby size goes. Is still a reflux still. Plated meaning a flute style CM still. The plates could be perforated, bubblecap, or valved.

A slant plate would be in a boka packed LM still. And used for collecting the distillate.

Basically if it has a way to force reflux. It's a reflux still. Otherwise it would be a pot still. You won't see many fractioning stills in the hobby sector. That's what I think MR is trying to come up with?

[Frosteecat]

I'll just follow along in the textbook. I realize that is like asking how an engine works in every kind of car...

[flyingdutchman]

Ok,
I got a great anonymous question about the mercury tube part on page 1. It basically asked how it was setup. I also included what it was trying to prove. For anyone interested I will add my answer here.

how are ya? Some great questions I like that. Alot of times I read stuff and it answers some questions and creates new ones. Physicists and physical chemists (is there really a difference?) just love the mercury tube thing. I think because its visual so you can see whats happening compared to how the usually do everything with equations. I believe he is making a couple of points here.
1) The atmosphere above us actually has weight. It that pushes down on us to the tune of 14.7 psi. Its counter intuitive because we dont feel it. If you look at a compressor or tire gauge it reads 0 unless the compressor is on or its connected to a tire but thats because the gauges are calibrated to read 0 at 14.7 psi. Think of like were standing in a giant column with a giant stack of air pushing down us all the way up to outer space. O2 has weight and N2 has weight so it exerts pressure on us.
One way to check this visually is the mm of mercury (Hg). Turn the sealed tube upside down and fill to the top with mercury, put your finger over the opening then put it into the pool of mercury and take your finger off (dont really do this). The mercury falls (creating a vacuum in the head space above it )but why doesnt it fall all the way out instead of just a little bit? Because the weight of the atmosphere is actually pushing it up! So now we can measure atmospheric pressure and if you took it to Denver Colo. it would fall a little becuase they are ~ 1 mi high so less air is over their heads. (yes it should work in a 2 inch tube the same way and if anything capillary action is like experimental error because it throws off the results)
2) His second point is that we can now measure precisely the saturated vapor pressure of different substance!

In physical chemistry, saturation is the point at which a solution of a substance can dissolve no more of that substance and additional amounts of it will appear as a separate phase (as a precipitate[2] if solid or as effervescence or inclusion if gaseous). This point of maximum concentration, the saturation point, depends on the temperature and pressure of solution as well as the chemical nature of the substances involved.

so we add his unknown substance and it floats to the top. Why? Because mercury is extremely dense and anything less dense will float to the top. Boats are less dense than water thats why they float (until you put a hole in them and they fill with water). Above the mercury is a vacuum so once the liquid reaches the top it boils! Why because the liquids vapor pressure exceeds the pressure at the top of the tube (~ 0 sort of..). Remember boiling points. Once no more of the liquid can evaporate we see a thin film of it on top of the mercury. Why because now so much of the liquid has evaporated that the pressure of its vapor is now equal to the vapor pressure at which it boils and so we call that saturated (for that temperature). This causes the mercury to be pushed downwards. How much it is pushed down in a measurable quantity (it can be quantified). We now know the saturated vapor pressure of that gas.

Hope this helps

[Frosteecat]

How can surface tension be manipulated to help prevent weeping, promote vapor through liquid filtration, etc? I.e. is surface tension stronger or maybe the effects of this pressure lessened by using non tradiotional plate or even column shapes, like convex or concave....?

[rad14701]

How can surface tension be manipulated to help prevent weeping, promote vapor through liquid filtration, etc? I.e. is surface tension stronger or maybe the effects of this pressure lessened by using non tradiotional plate or even column shapes, like convex or concave....?

You kinda sorta almost answered your own question about ratios with plated columns right there... Because of the inherent differences between plates and structured packing there is no way to calculate a reflux ratio with a plated column... The liquid on each plate performs much of the purification at each successive level and any return to lower plates does some reboiling as well as being returned to the next lower and hotter plate or back into the boiler... I'm just giving you a short and generalized version of what flyingdutchman has been discussing in relation to packed columns...

Keep it up, flyingdutchman... This is going to make for an interesting read when members get a chance to reread it from start to finish...