Sciencemadness Discussion Board

Some Time Saving Notes on Ethanol Distillation for Non-Experts

aga - 28-8-2014 at 13:33

During a recent run of several distillations and titrations of Ethanol, it became apparent that there are several things to consider when undertaking such work.

The Objective was to distill a 'feedstock' of dilute ethanol, with and without various Columns, and then to titrate the distillates in order to assess the effectiveness of the Columns.

The Titration process required 10ml of each distillate.

First, in order to tolerate later errors, begin with a sufficiently Large volume of feedstock.

Lack of enough feedstock (and brain) was the main reason that the entire procedure had to be repeated 3 times.

Next, it would be a great idea to run the process using a Volume of feedstock, and then calculate if you will end up with enough distillate with that Volume to accomplish at least 2 titrations (to allow for error).

Clearly that reflects back on how much raw feedstock you will need.

This Volume will dictate what size boiling flask to use (e.g. 250ml RBF or 500ml etc).

This affects your Heat Source : a bigger volume in the flask will require more heating.
For amateur setups, it can be as simple as adding extra tea-light candles to compensate.

If a column is used, such as a Vigreux, then the heat source will be heating the Column as well, which in turn implies either a better heat source, and/or insulation of the column, such as scrunched tinfoil, or sliced hot water pipe insulation.

If the glassware joints are not sealing well, you will find that the distillation proceeds slowly, vapour temperature never rises high enough, or you get drips from the joints.
Use vaseline, keck clips and pressure (judicious clamping can help).

On the other hand, ensure that the entire setup is not Totally sealed.
In an unrelated series of ethanol distillations, i inadvertently left a tap closed, which effectively sealed the entire aparatus from the atmosphere.

A 30 minute distillation reached the 30 minutes before i noticed (i use a kitchen timer which beeps).
In all it took 56 minutes, simply because of the closed valve.
Luckily the pressure build-up did not destroy any of the glassware.

In Summary :-

Think through an entire distillation, especially if the Volumes of Distillate and Feedstock are Important.

Calculate as much as you can beforehand. Experiment if required to get Data for the calculations.

Check all joints, seals and the Path for fluids.

Make sure the Heat Source will work. Distillation is a Heating/Cooling game.

Avoid Ethanol when distilling Ethanol.

Oscilllator - 28-8-2014 at 17:41

How did you titrate the ethanol? I would have though density measurements would be perfectly adequate for this kind of work.

aga - 29-8-2014 at 01:55

Comparing the Densities would probably work, but at the time, it was unknown how large or small the differences would be.

The titration was done by oxidising the Ethanol with a Potassium Dichromate solution, then Potassium Iodide was added to use up the remaining dichromate, then the resulting Iodine was titrated with Sodium Thiosulphate solution.

blogfast25 - 30-8-2014 at 02:54

Quote: Originally posted by Oscilllator  
How did you titrate the ethanol? I would have though density measurements would be perfectly adequate for this kind of work.


Density would work fine but you do need a decent pycnometer and 1 mg scales to do it accurately, because the difference in density between EtOH and water is relatively small. You would also need to construct a density-composition calibration curve.

The iodometric determination of EtOH in EtOH/water mixtures works just fine but requires a good level of organisation to be able to process multiple samples 'at once', as well as careful planning.

This project aga is reporting here was in fact a joint venture between him and myself. Aga did all the experimental work, largely based on my experiment design. Three different distillations were carried out on the same EtOH/H<sub>2</sub>O feed stock, one using no column whatsoever, another using an Allihn condenser (w/o water cooling) and the third using a typical Vigreux column without insulation.

I just need to find a bit of time to further process the raw data and draw some conclusions.

[Edited on 30-8-2014 by blogfast25]

aga - 30-8-2014 at 09:16

Quote: Originally posted by blogfast25  
This project aga is reporting here was in fact a joint venture between him and myself. Aga did all the experimental work, largely based on my experiment design.

bloggers did the hard stuff, like the experiment design and the maths.

My part was merely to execute the processes as specified in blogfast25's documentation.

I must take credit for the Quality and Quantity of all of the noob errors though.

DJF90 - 30-8-2014 at 10:11

Is aga not capable of interpreting the results themself? This also does not account for any methanol that may or may not be present, and relies on starting with a pure, aqueous ethanol. Density would be a reasonable measure of purity if this is the case, otherwise refractive index may be more informative. A 10mL sample for dichromate oxidation seems a little excessive (and wasteful of dichromate...). Ideally you could use a ca 1 g sample accurately weighed to 0.01/0.001g.

The practical side of things could be much improved upon as well. Ideally for fractionation you'd have a period of total reflux in order to establish equilibrium in your column. I appreciate this may not be possible without a dividing head, but the efficiency of your separation is going to suffer as a result. You're expecting azeotropic distillate, but it would be more advantageous perhaps to conduct the distillation under reduced pressure in order to break the azeotrope.

aga - 30-8-2014 at 11:28

Yes, i could probably interpret the process entirely wrong for myself, however with blogfast25's help, it all turned out rather well in the end.

Us noobs tend not to have analytical balances capable of accurately weighing milli- or micro-grammes.

The original ethanol solution was OTC 40% Vodka, then distilled 3 times.
Anything 'coming over' not at 78 C Vapour Temperature being discarded, along with the first 10 seconds of distillate that came over when the VP reached 78 C.
This distillate was diluted with distilled water to produce the feedstock.

You may not be aware that Chemistry is New to me (started Feb 2014), which is why i tend to loiter in Beginnings a lot.

Overall, blogfast25's designed processes and suggestions fitted the Requirements and Limitations of the project perfectly.

Certainly the equipment, analytical resolution could be improved, however i doubt that i would have learned more.

It would be wonderful DFJ90 if you would carry out a similar analysis, and post your processes and results.

Being more experienced, i expect it would take you a Lot less time than it took me.

blogfast25 - 31-8-2014 at 05:16

Quote: Originally posted by DJF90  
Is aga not capable of interpreting the results themself? This also does not account for any methanol that may or may not be present, and relies on starting with a pure, aqueous ethanol. Density would be a reasonable measure of purity if this is the case, otherwise refractive index may be more informative. A 10mL sample for dichromate oxidation seems a little excessive (and wasteful of dichromate...). Ideally you could use a ca 1 g sample accurately weighed to 0.01/0.001g.

The practical side of things could be much improved upon as well. Ideally for fractionation you'd have a period of total reflux in order to establish equilibrium in your column. I appreciate this may not be possible without a dividing head, but the efficiency of your separation is going to suffer as a result. You're expecting azeotropic distillate, but it would be more advantageous perhaps to conduct the distillation under reduced pressure in order to break the azeotrope.


As aga explained: no methanol. The vodka was redistilled enough to remove any other dichromate oxidisables (like sugar). The feed stock can thus honestly be considered a pure EtOH water solution.

Using a 50 w% EtOH water standard (prepared from anhydrous EtOH reagent grade) the back titration procedure was also checked and found to find this reference value accurately.

Refraction index would require a reliable refractometer (we don't have one) and a calibration curve.

Only about 1 g of feed stock and distillate was used in the titrations (using 0.1 N solutions), not 10 ml.

The Vigreux was not allowed to come to equilibrium for practical reasons: like you said a reflux splitter would be best for this but we didn't have one (who has? ;) )

We weren't 'expecting azeotrope ethanol', we just wanted to test the column for its separating capability. Using the Vigreux withoust splitter/refluxer we went from 17 w% EtOH (feed stock) to 84 w% EtOH in the distillate. This is about 3 theoretical plates (more about that later). Without column the EtOH content in the distillate was about 58 w% (using the same feed stock of 17 %), or just about 1 theoretical plate.

Removal of the water in azeotropic ethanol is usually done chemically, e.g. drying with CaO.


[Edited on 31-8-2014 by blogfast25]

Magpie - 31-8-2014 at 09:51

This is an interesting study. Your results using the Vigreux, ie, a separation consistent with 3 plates is what I remember from the distillation literature.

In another thread I stated that I had achieved azeotropic, ie, 95%, ethanol using a ss scrub pad packed 20cm Hempel column. I really can't substantiate this claim. I was speculating merely on the basis of the distillate temperature of 78.3-78.5°C. But when I examine the VLE data (composition vs T) for this system it is clear to me that within the limitations of my equipment my distillate composition could have been anywhere from 80-90% mole fraction ethanol. See:

http://commons.wikimedia.org/wiki/File:Vapor-Liquid_Equilibr...

I would be curious to see what distillate composition you find when using a packed column and your more accurate means of composition analysis.

[Edited on 31-8-2014 by Magpie]

[Edited on 31-8-2014 by Magpie]

aga - 31-8-2014 at 10:05

Quote: Originally posted by blogfast25  
Only about 1 g of feed stock and distillate was used in the titrations (using 0.1 N solutions), not 10 ml.

Yes, sorry.

It's my 10m fixation playing up again.

blogfast25 - 1-9-2014 at 08:53

Quote: Originally posted by Magpie  
This is an interesting study. Your results using the Vigreux, ie, a separation consistent with 3 plates is what I remember from the distillation literature.



Thanks Magpie.

I don't know about "a separation consistent with 3 plates is what I remember from the distillation literature". Obviously the number of plates achieved will depend on column height and surface area/volume ratio. The reflux ratio itself will affect it (at 'no reflux' the column would really act as a single plate!)

It would be very interesting to compare a simple Vigreux with a with a comparably sized packed column. But from my studying days (chemical engineering) I remember that finding comparative studies between various column designs (plates, various types of packing/redistribution of flow etc) should not be a big problem, provided one has access to a number of highly specialised publications.

I think a well packed column (not too tight, not loo loose) could increase the number of plates very significantly.

[Edited on 1-9-2014 by blogfast25]