Sciencemadness Discussion Board - Diluting Ethanol to 95% (Azeotropic)

Diluting Ethanol to 95% (Azeotropic)

nimgoldman - 7-2-2019 at 10:12

Since ethanol I am able to obtain by distillation is only up to 92% ABV, I dry the alcohol to >95% using molecular sieves. Then I filter and distill the alcohol (to remove dust) and measure its density using tared volumetric flask. Then I look up its w/w concentration from table and calculate how much water to add to reach azeotropic concentration, which I believe is 95.5% w/w according to Wiki, though sources vary.

I assume the azeotropic concentration is the most stable one and the concentration most procedures call for (if not specified otherwise).

Unfortunately, my measurements vary from theory and I am not sure why. The ABV alcohol meter (0-100% range with 2% precision) shows the alcohol is 96% ABV and density measurement shows 0.810 kg/L which should be something like 93% w/w.

Since 95% ABV corresponds to 95.5% w/w, this means either the alcohol meter shows higher concentration than actual, or the density measurement shows lower concentration than actual.

I tested the alcohol meter on pure ethanol from pharmacy and it shoved 96% ABV. It also correctly shows 40% concentration on store-bought vodka so I don't think the alcohol meter is off.

I also don't think there is significant amount of admixtures in my ethanol, that would skew the results. I filtered the ethanol with activated charcoal and Celite and fractionally distilled 3 times, always taking the middle fraction running between 78-82 °C.

What also puzzles me is that even excess amount of lab grade molecular sieves (Carl-Roth brand) aren't able to produce 99% ethanol. I use over 20 g of sieves per gram of water (way over the specified absorption capacity). I really have to use crazy excess (like half the volume of alcohol) to produce reasonably anhydrous ethanol for certain syntheses. This then leads to a significant loss of ethanol. I usually let the sieves adsorb water for 1-2 days.

How would you achieve azeotropic ethanol from almost-azeotropic or over-azeotropic one?

Tsjerk - 7-2-2019 at 10:16

What is your source of ethanol?

Edit: can you do a Karl Fischer?

[Edited on 7-2-2019 by Tsjerk]

SWIM - 7-2-2019 at 10:59

If you want azeotropic ethanol from ethanol that's over 95%, how about just leaving the cap off for a while?

Tsjerk - 7-2-2019 at 12:21

Quote: Originally posted by SWIM

If you want azeotropic ethanol from ethanol that's over 95%, how about just leaving the cap off for a while?

You will lose a significant amount of ethanol before reaching azeotropic.

streety - 7-2-2019 at 15:04

What type of alcohol meter do you have? Is the 2% precision over the entire range or is that at a specific concentration?

What volume did you weigh? If a small volume, could differences in the shape of the meniscus throw off the volume? What about the temperature?

Leaving the cap off may be wasteful but you could partly distill and then use what doesn't distill over.

Heptylene - 7-2-2019 at 15:12

Have you considered using a large volumetric flask and a scale? Volumetric flasks measure volume very precisely, scales measure mass very precisely too (if properly calibrated). I'd assume the result would be better than an alcohol meter.

nimgoldman - 7-2-2019 at 15:26

I make the ethanol from sugar wash (white sugar with Alcotec Turbo Yeast 23%). After fermentation, I fractionally distill it using 60 cm Vigreux column, then twice with Vigreux and packed columns with activated charcoal filtering between distillations.

I used 1L volumetric flask and a scale with 1 g precision.

I measured same density with 250 mL flask and scale with 0.01 g precision and got same result.

The amount of liquid under and over meniscus have negligible effect on the result being the concentration measured by density is very different from that given by ABV meter.

The alcohol meter is has 2% steps over whole range, here is one image of its use:

The problem is the ABV meter says my alcohol is 96% ABV which should be more than 95% w/w, but the density table says it's only 93% wt% ... I don't know how to explain the discrepancy. I even tried adjusting temperature of the alcohol to 20 °C but with same result.

Heptylene - 7-2-2019 at 15:32

Maybe your scale is off? If it has not been calibrated, there can be an error of about 0.5 % just due to earth gravitational field differences between where the scale was made and where it is being used. And that's not counting other sources of errors that might contribute.

streety - 7-2-2019 at 16:57

It sounds like you have been very thorough. I would expect 1L and 250mL volumes to be accurate. It might be worthwhile weighing a liter of water just as a sanity check but I think I would trust your density measurements.

For the hydrometer I'm not sure I would assume the accuracy is +/-2% just because the gradations are 2%. The hydrometer you have also has a very wide range. I would expect a more limited range for a high precision hydrometer.

S.C. Wack - 7-2-2019 at 18:34

If accuracy is proven one may perhaps seek a different sort of proof, like the different kinds of chromatography or color tests. AFAIK carbon does little for small molecules...BTW I've seen fractionated samples from the same yeast turn yellow on reflux with KOH, from condensations.

There's probably a good reason why you have no references for the drying of 95% alcohol with sieves.

unionised - 8-2-2019 at 05:40

Azeotropic concentration depends on atmospheric pressure.
I'd check the balance and the flasks by weighing 250 and 1000 ml of water.
The density of alcohol/ water mixtures varies with temperature too so you need to measure that.
Hydrometers are usually calibrated for aqueous solutions. There's a small effect from surface tension if you use them for other materials.

If you use the same balance and the same flask to weigh 250 ml of water and 250 ml of ethanol then calculate the density from tehratio (so it doesn't matter much if your flasks wrong or the balance is miscalibrated) you should get a figure that's more reliable than any meter.

Tsjerk - 8-2-2019 at 09:45

I have the feeling your ABV meter is off. I would trust density measurements over this meter as I have a vague recollection of these meters being imprecise in the very low or very high range.

I also think the way to go is checking your balances with water. If they are accurate density is the way to go as I don't see turbo yeast with white sugar producing much else than ethanol.

happyfooddance - 8-2-2019 at 18:12

Quote: Originally posted by streety

It sounds like you have been very thorough.

No it doesn't. Not once is temperature mentioned anywhere, never is anything calibrated, checked against a standard... I wouldn't expect any accurate anything from the techniques and suggestions in this thread.

nimgoldman - 8-2-2019 at 22:20

Yes temperature has been checked and one of the scales (the 0.01g one) has been recently calibrated using two point calibration. Note that both scales give the same result.

Also when I load the ethanol with sieves to make it practically anhydrous, the hydrometer reads 100% - since both the vodka reading and the anh. ethanol readings seem accurate, I might assume the hydrometer is also.

Well the problem is that I just want 95% ethanol since this seems to be the standard concentration. I don't want to overshoot or undershoot the amount of ethanol in a synthesis that calls for 95% and I also don't like changing the percentage label every time a distill a new batch of ethanol :-)

One problem might be that I measure inside a somewhat slim 100 mL graduated cylinder. Maybe a wider container would be better... Another source of skew might be atmospheric pressure and temperature (likely pressure).

Okay I will try weighing de-ionised water first just for control.

streety - 9-2-2019 at 06:07

happyfooddance, we are all going to have different views on what constitutes thoroughness. I would certainly appreciate reading how you would approach this problem.

A concern I had was the potential use of a small volumetric flask coupled with an imperfect reading of the meniscus could have been one source of error. I was pleased to read that a 1L flask had been used. The use of a second flask and balance exceeded my expectations.

Although not mentioned in the initial post, both temperature and calibration have been mentioned.

nimgoldman, I've been assuming the sieves are 3A. Is this correct?

happyfooddance - 9-2-2019 at 12:37

Thorough would be weighing pure H2O first, not later. Always maintaining 20 degrees Celsius. Using a high precision analytical balance.

A slim volumetric will give you better results than a wider one.

How I tackle this problem (I use lots of ethanol in various grades) is I take etOH and dilute it to roughly 92-94%, and I distill it slowly through a 400mm vigreux, collecting the fraction that distills at 78.1°C. I usually cut the distillation short when the pot gets above 79. Bottom of column never goes beyond 78.1, and since I live by the beach I stay close enough to 1atm for my calculations.

I have never used an alcohol abv meter but I would expect that it would be not very accurate, and if I wanted a more accurate measurement of alcohol concentration I would do a titration against dichromate (again, standardized solutions using de-ionized, de-gassed water at 20°C...)

Sulaiman - 10-2-2019 at 02:16

I have one of these cheap hydrometer kits
https://www.ebay.co.uk/itm/3pcs-Hydrometer-Accurate-Alcohol-...
Three scales and a thermometer with correction tables.
So far I can't find a fault with this kit.

Density changes with temperature cause significant errors if not compensated for.
e.g. "probably about 20^oC" ... is not accurate enough.

Tsjerk - 10-2-2019 at 06:34

Quote: Originally posted by nimgoldman

Then I look up its w/w concentration from table and calculate how much water to add to reach azeotropic concentration, which I believe is 95.5% w/w according to Wiki, though sources vary.

Can you tell us how you do this calculation?

Quote: Originally posted by Sulaiman

e.g. "probably about 20^oC" ... is not accurate enough.

The difference in density of liquids at "roughly 20 degrees" is negligible, at least it is when trying to determine between 95% and 96% ethanol.

[Edited on 10-2-2019 by Tsjerk]

Sulaiman - 10-2-2019 at 08:12

From the temperature compensation tables provided with the kit that I pointed to above;

a reading of 95 indicates 95%ABV at 20^oC and 96%ABV at 15.5^oC
a reading of 96 indicates 96% at 20^oC and 95% at 24.5^oC
so a 4.5^oC difference is equivalent to a 1%ABV change in reading.
(6^oC change for 99/100%AVB)

S.C. Wack - 10-2-2019 at 11:24

Thorough is proof by challenge...an easy and informing challenge would be comparison with gradually diluted undenatured ACS ethanol, AKA calibration. Follow the directions and use lime like everyone else if you want to dry...no you can't just dump in MgSO4 or something and it'll all be cool...

[Edited on 10-2-2019 by S.C. Wack]

happyfooddance - 10-2-2019 at 12:46

Quote: Originally posted by S.C. Wack

you can't just dump in MgSO4 or something and it'll all be cool...

[Edited on 10-2-2019 by S.C. Wack]

While I have to agree with this, I will say that ball-milled anhydrous MgSO4 has been one of my favorite drying agents as of late. It is quick, has a high capacity, and yanks water out of solutions that are already reasonably dry.

The only thing is having to ball mill it, and also storage requires a little planning and care.

macckone - 10-2-2019 at 18:19

The meter shown in the picture is not a high quality, high proof meter.
The recommended range of the meter and temperature matter, a lot.
I suspect that it is pretty accurate though.

Another thing that may throw off the measurements is ethyl acetate and associated impurities.
If you can't get azeotropic by distillation then you probably aren't getting rid of those either.
It takes 5 or more passes through simple distillation to get azeotropic.
With a packed column, you should be able to do it in 2 passes with reflux.
But you stated you were doing a 4C range, and that is way too wide for final distillation.
I would work with a slower distillation and better reflux with a packed column.

nimgoldman - 12-2-2019 at 19:34

I see.

I should have stated the question differently: How to get "standard" 95% ethanol, provided that you have ethanol/water mixture of arbitrary ratio?

The impurities in my ethanol could be one cause of the discrepancy. I am thinkin about removing ethyl acetate, aldehydes and possibly other impurities chemically before proceeding to second distillation.

I run the fermentation in a closet with controlled temperature (24 °C) so I doubt there will be typical impurities such as fusel oils, but ethyl acetate is dfinitely a possibility. The yeast also come with plenty of yeast nutrient (DAP) of which some components might possibly come over.

Of course I don't need ultrapure ethanol, just the "standard" 95% ethanol usable for most amateur lab syntheses (making diethyl ether, alkyl halides, washing and drying products etc.)

Sulaiman - 13-2-2019 at 02:11

Concentrating or diluting ethanol to azeotropic concentration may be a useful exercise,
but I would try for anhydrous ethanol because;
. some reactions need to be anhydrous
. adding H₂O is easy

DavidJR - 13-2-2019 at 03:06

Quote: Originally posted by Sulaiman

Concentrating or diluting ethanol to azeotropic concentration may be a useful exercise,
but I would try for anhydrous ethanol because;
. some reactions need to be anhydrous
. adding H₂O is easy

I agree with this. I use molecular sieves 3A to get the last few percent of the water out - they are slow but they are also highly effective, providing you use a sufficient amount. Store over more fresh sieves to keep the ethanol anhydrous, as absolute ethanol is quite hygroscopic.

nimgoldman - 13-2-2019 at 04:07

Yes but drying all the way to anhydrous requires more sieves (over 20 grams of sieves per gram of water) and this leads to significant loss of ethanol which adheres to the sieves' surface.

Maybe distilling the ethanol off the sieves afterwards might save some alcohol but it is extra work. Using only minimum amount of sieves (high quality zeolites are expensive and regenerating them in a home lab setup is also a lengthy process).

From Wikipedia:

Quote:

Rectified spirit is typically distilled in continuous multi-column stills at 96–96.5% ABV and diluted as necessary.

I guess they simply ensure the ethanol is distilled to the point of being practically azeotropic and then the amount of water added is always the same.

I don't know if there is any filtration or chemical purification in the pharmaceutical-grade ethanol production process (besides distillation), but I've the common impurities (esters and fusels) can be removed by activated charcoal filtration below 50% ABV and by treatment with mild alkali (e.g. baking soda) to hydrolyze esters, namely ethyl acetate into ethanol and acetic acid, which can be more easily separated by fractional distillation. I found table salt is also effective in aiding the separation process.

[Edited on 13-2-2019 by nimgoldman]

DavidJR - 13-2-2019 at 04:13

Quote: Originally posted by nimgoldman

Yes but drying all the way to anhydrous requires more sieves (over 20 grams of sieves per gram of water) and leads to significant loss of ethanol which adheres to the sieves.

I don't think that's correct - I've seen figures of around 20% by weight water capacity for molecular sieves 3A, which would correspond to 5 grams of sieves per gram of water.

You do lose some amount of product but I don't think it's a whole lot.

Alternatively you might want to start with another drying agent to get it to >= 99% then add sieves to get the last little bit.

[Edited on 13-2-2019 by DavidJR]

nimgoldman - 13-2-2019 at 04:54

I have a different experience in my practice.

When I start with say 1 000 mL of 92% ABV ethanol and add just 5 g of 3A sieves per gram of water (Carl-Roth brand), that is 400 g of sieves, the final concentration of ethanol after 2 days is about 97-98%. I have to add huge excess to get near anhydrous in reasonable time, sacrificing significant portion of ethanol by adherence to the sieves surface. The drying process is statistical and I think the theoretical absorption capacity can be approached only asymptotically.

I was thinking about recovering some ethanol by distilling it off the sieves.

My 3A sieves are tested for water absorbtion capacity of at least 15.5% (weight of water per weight of sieves) - that roughly corresponds to the above figures.

[Edited on 13-2-2019 by nimgoldman]

DavidJR - 13-2-2019 at 05:57

According to my calculations, 1 litre of 92% v/v ethanol contains 99g of water, so you'd actually need 500g. Also I'm not sure how long one has to wait as I have a tendency to add sieves and forget about it for a week.

S.C. Wack - 13-2-2019 at 18:55

If Alcotec turbo yeast and sugar is so clean, why did they immediately get a reputation as unclean, why do they sell carbon, yeast mixed with carbon, and special vodka yeast they say actually gives a clean product...they seem eager to give the impression that they either sell a good product or it can be cleaned up.

Could your confidence in yeast, carbon, the number of theoretical plates possible in your setup and the number actually attained in practice, and sieves (particularly in grossly wet alcohol) be over-rated? Ethanol becomes exponentially more difficult to purify by distillation as the water content decreases; I suspect most members would not be using a column with too many plates and too high a reflux ratio.

There are tables that say how much water to add to x amount if your ethanol tests y and you want z.

nimgoldman - 14-2-2019 at 11:24

Quote: Originally posted by DavidJR

According to my calculations, 1 litre of 92% v/v ethanol contains 99g of water, so you'd actually need 500g. Also I'm not sure how long one has to wait as I have a tendency to add sieves and forget about it for a week.

Yes I made a mistake in my calculation. I just subtracted the volume of alcohol, not realizing there is more water due to packing of molecules. Now it makes sense.

Yes I will look up the tables. I was more interested in the theory behind the mixing ratios or whether they are just measured empirically.

As for the purity, I don't know what other significant impurities could the yeast produce.

I used Lalvin EC 1118 yeast earlier (usually sold as wine or champagne yeast), but these were more expensive, plus I had to purchase the extra nutrient - the nutrient is not necessary and indeed adds impurities, but it makes the yesat multiply and thus speed up the process consierably.

Note that I need ethanol pure just enough for simple syntheses and washing glassware.

I think I will filter the alcohol with AC when it is diluted below 50% v/v as the carbon is said to be much more efficient when water will present. Too strong alcohol is said to be able to actually pull the adsorbed impurities back.

I will also check pH and neutralize any acids and try to hydrolyze esters (namely ethyl acetate). I am thinking about treatment with KOH and sulfuric acid to remove possible aldehyde impurities.

The still does not have much plates nor the reflux ratio can be controlled. I can control the pot temperature and also thinking about adding a makeshift dephlegmator to improve the reflux.

Tsjerk - 14-2-2019 at 12:22

First try to measure properly diluted ethanol. And stop using molecular sieves to dry very wet ethanol. Start with some anhydrous salt, and only dry the last percent with sieves. Then properly calculate the amount of water you need. I'm almost sure that will fix your problem (two problems; losing much ethanol to the sieves and the density/percentage thing).

Quote: Originally posted by nimgoldman

I will also check pH and neutralize any acids

Don't try to measure pH of anything not being at least 95% water, or at least don't try to deduct anything out of it. It won't work as any equipment that will give you a value you can do any calculations with won't be designed or won't be calibrated to do so. Anything you can measure pH with will give you a hint of it being acidic or not, not how acidic it is. Or maybe it will tell you it is acidic while it is not. Any pH meter swings between pH 4 and 11 while being in perfectly neutral and pure 3 molar KCl storing buffer for example.

Quote: Originally posted by Tsjerk

Quote: Originally posted by nimgoldman

Then I look up its w/w concentration from table and calculate how much water to add to reach azeotropic concentration, which I believe is 95.5% w/w according to Wiki, though sources vary.

Can you tell us how you do this calculation?

I asked you didn't I? A lovely demonstration experiment is to load 50 ml 99.9% ethanol on top of 50 ml of water in a measuring cylinder and make 4 ml "disappear" buy stirring it.

[Edited on 14-2-2019 by Tsjerk]

nimgoldman - 19-2-2019 at 13:52

As for the drying, I browsed many threads to realize no common drying agent works properly with ethanol. For example, magnesium sulfate is said to not working at all, calcium chloride complexes with ethanol etc.

Most sources talk about rigorous drying, though, not about removing just some water from dilute alcohol.

I found some people use cornmeal or cellulose. I will investigate this further.

So far I will distill up to 92%, then pre-dry with - say - MgSO4 and then finally dry with sieves.

Quote: Originally posted by Tsjerk

Can you tell us how you do this calculation?

I calculate density using volumetric flask and a calibrated scale.

Then I look up the wt% concentration from density using table found on the internet.

The target w/w concentration of an azeotrope is 95.5% so the target concentration can be expressed as

Code:

0.955 = w_EtOH / (w_total + w_H2O)

where w_EtOH is the weight of pure ethanol in the solution and w_total is the weight of the solution.

Solving for w_H2O gives us the weight of water to add to reach the desired concentration.

[Edited on 19-2-2019 by nimgoldman]