Sciencemadness Discussion Board

Super-dry ethanol: test for.

FrankMartin - 31-5-2010 at 21:20

I am seeking a simple test to check if super-dry ethanol is not water contaminated.

Perhaps some added reagent, which will indicate a certain percentage of water (> 0.05%) is present?

Please help, Frank

not_important - 31-5-2010 at 21:50

The Karl-Fischer test is the standard method, but it takes expensive equipment. There are lower cost instrumental methods, search for "moisture analysis" or "water activity analyser". These are based on humidity sensors, conductivity measurements, or conductivity-capacitance measurements; you might be able to build your own such device for not too much.

A simple way is to add a bit of the liquid to anhydrous CoCl2 or CuSO4, but I'm not remembering the eye-ball detection limits with these. A slightly more complex related method is in the attached PDF, you could build a simple instrument using several LEDs and a visible light photosensor to measure transmission at several wavelengths. It looks like it would work at the limit you give, but how you'll keep ethanol that dry without a gas blanket would seem to be a problem.




Attachment: Spectrophotometric Determination of Water Content in Alcohol Organic Solvents.pdf (106kB)
This file has been downloaded 1170 times

FrankMartin - 31-5-2010 at 22:46

Thank you. Just now I have discovered a method in a very old Organic Reagent book, and this suggests to add to the suspect sample of ethanol a small quantity of Aluminium Ethoxide in benzene. If the ethanol has more than 0.05% water a "fluffy and voluminous" precipitate of Aluminium hydroxide is formed.

I will now buy some Al(OEt)3 and try this. I will advise if it works.

Regards, Frank

Vogelzang - 1-6-2010 at 13:47

How effective would a refractometer be?

UnintentionalChaos - 1-6-2010 at 14:55

Not very. Consider how small a mole fraction say 0.5% water in ethanol would be. Then pretend the relationship between the refractive index and composition is linear with respect to mole fraction (which it wont be). Water and ethanol have refractive indices of 1.333 and 1.361 respectively according to wikipedia. In fact, I'd go so far as to say it's about 'durned useless.

not_important - 1-6-2010 at 16:51

Yes, RI measurements just don't change fast enough with many common liquids to give high resolution of concentration. At 15 C going from 0% to 10% EtOH changes the RI by less than 0.007, a 1% change would be a tenth of that.


Even some recent designs don't look as if they would give the needed resolution

Attachment: Micro_thermophy_eng.pdf (273kB)
This file has been downloaded 993 times


Note that aluminium alkoxides in solution a fairly difficult to keep from reacting with atmospheric moisture. If you plan on using the reagent repeatedly, you may need to invest in at least some fine grade fritted filters to clean it up before use.

In fact a possible source of error is that the partial hydrolysis products are more soluble than pure (EtO)3Al until the amount of HO- groups gets high enough. This can both mask low concentrations of water in the tested sample, and cause false positives if the reagent has reacted to a sufficient degree to be near the point where solubility starts dropping again, so that only a small amount of extra water pushes it over the edge. You may find it necessary to run calibration checks on the reagent.

UnintentionalChaos - 1-6-2010 at 17:28

Quote: Originally posted by FrankMartin  
Thank you. Just now I have discovered a method in a very old Organic Reagent book, and this suggests to add to the suspect sample of ethanol a small quantity of Aluminium Ethoxide in benzene. If the ethanol has more than 0.05% water a "fluffy and voluminous" precipitate of Aluminium hydroxide is formed.

I will now buy some Al(OEt)3 and try this. I will advise if it works.

Regards, Frank


Just a thought; we have a a thread on making your own aluminum isopropoxide. If you mean to simply test a sample of your reagent, this would probably work fine. I imagine that alkoxide exchange is somewhat rapid since the aluminum is lewis acidic and can coordinate an ethanol oxygen, undergo a (solvent mediated?) proton transfer and expel isopropanol.

If you have access to mostly dry ethanol as it is, you can probably just make the ethoxide by the same procedure.