Sciencemadness Discussion Board

Drying alcohols with Sodium Glycoxide

unome - 3-1-2010 at 17:24

Attached please find a paper on the drying of alcohols using sodium glycoxide (the sodium alkoxide of ethylene glycol), which can apparently be formed by thermal dehydration of a solution of sodium hydroxide in ethylene glycol.

Now, what I'm wondering is, if an equilibrium mixture of NaOH + EtOH <==> NaOEt + H2O, what happens when ethanol (99% dry ethanol, ie. virtually absolute) is added to the sodium glycoxide in glycol/ethanol? What would be the pKa of the glycoxide?

The pKa of ethanol is 15.9 & that of ethylene glycol = 14.2, so that would make ethanol a weaker acid, so I'm guessing that EtOH would not displace the glycol, but if that is the case (and the basicity of an alkoxide apparently increases with size - I'm not at all sure about this shit but;)), what would be the pKa of the glycoxide? Would it be a stronger base than sodium methoxide?

Attachment: Dehydration of Alcohols with Alkali Metal Alcoholates.pdf (392kB)
This file has been downloaded 885 times

Attachment: php8AsQpX (542kB)
This file has been downloaded 796 times

[Edited on 4-1-2010 by unome]

Nicodem - 4-1-2010 at 03:54

I do not know the pKa of ethylene glycol in ethanol, but I would estimate that by dissolving NaOCH2CH2OH in ethanol you would achieve a similar concentration of NaOEt as when dissolving an equivalent amount of NaOH. I doubt it would give a higher concentration.
What is your goal anyway? If you want to make a solution of NaOEt in ethanol then it is obviously much more practical to dissolve NaOH in absolute ethanol. This would gives a solution containing mostly ethoxide with some hydroxide (the ratio depending on the amount of NaOH dissolved and can be calculated from the pKa's of H2O and EtOH in ethanol - the result being accurate only for the more diluted solutions). You can make it a solution of pure sodium ethoxide in ethanol by using one of the methods for removing water from ethanol (azeotropic distillation or molecular sieves) or by isolating pure NaOEt as a solid and then redissolving it in another batch of absolute ethanol. The other method is to dry a saturated solution of NaOH in EtOH by a few batches of solid NaOH. Or you can just react sodium with excess absolute ethanol. All these methods are described in several threads in the forum. The easiest and safest way is to just buy solid NaOEt, which is very cheap, and then dissolve it in absolute ethanol.

unome - 4-1-2010 at 18:48

The idea was that if one dissolved the glycoxide in ethanol, then the ethoxide would be generated without the concomitant formation of water, thus serving to fuck up the alkoxide again. The glycoxide (wouldn't it be NaOCH2CH2ONa (or Na2C2H4O2)?) appears to be an 'easily' generated solid (although unstable), which can be thermally dehydrated (cf ethoxide). It just seemed that if this were so, then the generation of it thermally would beat the azeotropic route...

As to the concentration - the following is from the paper (above, p.2/3):


Material charged - 496 g of glycol, 333 g of 48.1% NaOH (Solution presumably) and 19,775g of 92% (by weight) ethanol

Material obtained - 541 g of sodium ethylene glycoxide residue, 1,591g water fraction, 849g middle alcohol fraction, 17, 494 g strong alcohol fraction 97% (by weight).

That suggests, especially considering they acidify the glycoxide residue to get their glycol back, that the concentration of glycoxide post distillation is pretty much the entirety of the residue. Thus the extractive distillation of the water, alcohol (and some glycol) leaves only the glycoxide behind.

Dissolving that in EtOH should give a higher concentration of NaOEt for mine, there is no water generated... The equilibrium would be entirely glycoxide/ethoxide (providing the alcohol was dry), wouldn't it?

rrkss - 5-1-2010 at 11:47

I just use a 3A molecular sieve. Its very cheap and can easily be regenerated in your oven. Leaves pretty much absolute alcohol without a fuss.

unome - 6-1-2010 at 01:00

Attaching 2 papers, both studies on the H_ acidity / pKa of sodium glycoxide/glycol solution... Interestingly enough, they put the pKa of a 3M solution of the glycoxide in glycol ~17.2

Given that is about the range of other better known alkoxides, (potassium t-butoxide for one) why couldn't it be used 'as is' for condensation reactions?

Attachment: Aiyar.Datta.Kundu.Determination.H.Acidity.Na.Glycoxide.Glycol.pdf (605kB)
This file has been downloaded 929 times

Attachment: Kundu.Aiyar.H.Acidity.Fx.Li.Na.K.Glycoxides.pdf (473kB)
This file has been downloaded 826 times

[EDIT] This glycoxide (the sodium one) was implicated (more to the point, its runaway exothermic decomposition) in the Seveso accident/disaster.:o

[Edited on 6-1-2010 by unome]

Nicodem - 7-1-2010 at 01:06

You can not compare the pKa of glycol in glycol (~17.2) with the pKa of glycol in water (~14). pKa values are solvent dependent, so that such a difference in values is completely normal. You can only compare trends of pKa's of a series of compounds from one solvent to the other. And even here the order might not remain the same. For example, I would expect glycols to be more acidic in nonpolar solvent that normal alcohols because the anion is stabilized by the intramolecular H-bonding while in normal alcohols this is not possible. In water this effect is more or less irrelevant because water molecules compete for H-bonding so that ethylene glycol is just somewhat more acidic than ethanol (about 10 times more acidic).
Quote:
The idea was that if one dissolved the glycoxide in ethanol, then the ethoxide would be generated without the concomitant formation of water, thus serving to fuck up the alkoxide again. The glycoxide (wouldn't it be NaOCH2CH2ONa (or Na2C2H4O2)?) appears to be an 'easily' generated solid (although unstable), which can be thermally dehydrated (cf ethoxide). It just seemed that if this were so, then the generation of it thermally would beat the azeotropic route...

Even though no equilibrium containing water is obtained if you dissolve the glycoxide in ethanol, you nevertheless get the equilibrium with ethylene glycol which in ethanol is probably more or less just as acidic as water (probably more). That is why I said that I think you would obtain a similar concentration of NaOEt as if you would dissolve an equivalent of NaOH. Since the ratio of NaOEt/NaOH obtained by dissolving NaOH in absolute ethanol are up to 9:1 and more (depending on the concentration), this is not really important. It is however important for the reactions where it is important to have anhydrous conditions, like for the Claisen condensation for example. But reactions like this one proceed very poorly in ethanolic NaOEt. They give best results when solid NaOEt or some other alcoxide is used in a nonpolar aprotic solvent. So it would be pointless to waste a glycoxide by dissolving it in ethanol when it would work better as such. Besides, I think you are wrong. Preparing solid NaOEt is very much simpler than the process you describe for this sodium glycoxide. It is even simpler to prepare sodium isopropoxide.

sonogashira - 7-1-2010 at 02:17

Quote: Originally posted by Nicodem  
It is even simpler to prepare sodium isopropoxide.


If you have the time would you briefly describe what you did? I recall reading the patent on this method and I believe heat was sometimes used? Is it necessary, or did you just add molecular sieve... or excess of hydroxide? Thanks :)

Nicodem - 7-1-2010 at 14:08

I never prepared sodium isopropoxide (these alkoxides are too cheap to waste time preparing unless for amateur reasons). But as you must have noticed from the now many reports that you can use molecular sieves to remove water from the solutions of NaOH in alcohols.

For other methods look in Vogel under the methods for drying alcohols. It states that alkoxides can not be used as drying agents because alcohols/water azeotropes still distil first leaving the alkoxide in dry alcohol behind. The ethanol/water azeotrope has a too low water content so that there is a need to either add toluene or to use a very efficient distillation column. But the isopropanol/water azeotrope (88:12) is just perfect as it has a larger water content so that it is not even that necessary to use a distillation column.

So all it takes is to prepare a saturated solution of NaOH in isopropanol, dry it over NaOH (which removes most water anyway) and then rotavap the supernatant to dryness to remove the last traces of water. This should leave a fairly pure sodium isopropoxide (which should be treated with due care as alkoxides can go into flames if heated in air).

Slightly more problematic simple alkoxide is the methoxide. As far as I know the only way to prepare NaOMe from NaOH is using molecular sieves (as already described in some other threads). Though the azeotropic removal of water with added toluene might also work with methanol. Solid NaOEt is simple to make as all it takes is to prepare a saturated NaOH/EtOH solution, precipitate out the NaOEt with acetone, isolate the solids by filtration, wash with toluene and dry in a desiccator (there is a patent describing the process, certainly already cited in some other thread). I tried this years ago and it indeed works well. Or maybe even simpler is the use of NaOH as drying agent to dry NaOH/EtOH into NaOEt/EtOH.

sonogashira - 7-1-2010 at 15:52

Thanks very much!

hinz - 7-1-2010 at 17:59

Does the molecular sieve thing really work? Its well known that a hydroxide solution in water or alcohol attacks borosilicate glass if you leave it around at room temperature for some days.
So I wonder how can those incredible fine pores (average size in 3A is 0,3 nm, or 3 Ångström, hence it´s name) can survive the hydroxide. I would expect that those molecular sieve prills get quickly eaten by the alkoxide solution even at room temperature because they have such a big surface area.

Probably another nice nice way of preparing sodium isopropoxide is adding NaOH to a solution of aluminium isopropoxide, precipitating Al(OH)3 and leaving a solution of Na isopropoxide behind.