Sciencemadness Discussion Board

Synthesis of anhydrous potassium alkoxides?

blogfast25 - 17-4-2013 at 05:37

Has anyone here have experience with/knowledge of the synthesis of anhydrous potassium/sodium alkoxides from their hydroxides (or other common K or Na salts)? The usual reaction of the metal with the alcohol is no interest here.

At a glance and w/o any further research, maybe refluxing of anhydrous IPA (e.g.) with KOH, followed by removal of the water by distillation with an azeotropic forming solvent could bring solace? Toluene perhaps as an azeotrope former? This may lead to a product that is a relatively pure potassium alkoxide.


[Edited on 17-4-2013 by blogfast25]

Dr.Bob - 17-4-2013 at 06:58

The main way I have ever made them is from sodium and the alcohol. That works OK, but even then, I normally use them as the solution in the alcohol, as drying them all of the way is tough, but likely doable, if not a humid day.

I have also made small amounts of solutions of sodium methoxide by simply dissolving NaOH in methanol, this will make a mixture of the hydroxide and methoxide in equilibrium, but getting rid of the traces of water will be tough unless the alcohol boils at a temp below that of water, which is not trivial, as methanol, ethanol, and IPA all boil well below water, although in theory you could azetrope off ethanol and water, but I have found azetropes to rarely be that good at fighting equilibrium.

I believe that Nicodem has mentioned ppting NaOEt from acetone as a way to get the solid, you could search for that, look here also:

^ US patent 1978647, Olson, E. & Twining, R. H., "Method for Making Alkali Metal Alcoholates", issued 1934-10-30

sbbspartan - 17-4-2013 at 07:14

I'm not sure if this will help, but in example 2 and 3 of this patent (US patent number 1712830), they talk about making potassium butylate from potassium hydroxide, butyl alcohol, and in example 2, toluene.

http://www.google.com/patents?id=xcdWAAAAEBAJ&printsec=a...

Quote --> "This invention relates to improvements in the production of anhydrous alkali metal alcoholates, and more particularly to the production of anhydrous alkali metal alcoholates."

edit -- The thread Dr. Bob mentioned is here: http://www.sciencemadness.org/talk/viewthread.php?tid=2656&a...

[Edited on 17-4-2013 by sbbspartan]

BromicAcid - 17-4-2013 at 07:43

Of Interest

http://www.sciencemadness.org/talk/viewthread.php?tid=2656

Nicodem - 17-4-2013 at 09:22

Making anhydrous alcoholic sodium or potassium alkoxides from the respective hydroxides is relatively simple for the lower primary and secondary alcohols, but very difficult for tertiary alcohols (I have not seen any reliable reference for these).

This is a recurrent topic and you can find plenty of examples and references posted by users, if you take some time to UTFSE. For example, Ullmann gives a method of preparing anhydrous methanolic sodium methoxide by using molecular sieves on methanolic sodium hydroxide (should work for ethanol and some other alcohols as well). Organikum's thread that BromicAcid linked gives a simple method suitable to simple primary alcohols higher than methanol. Azeotropic removal of water is supposed to work for the propanols and perhaps some other alcohols as well, but requires a good distillation column as the alcohol/water azeotropes have bp close to the bp of the pure alcohols (for ethanol, the difference is too small for practical use - it would take too much of the anhydrous ethanol). Continuous removal of water with a Dean-Stark trap and toluene as the carrier solvent might work for some less volatile primary alcohols (n-pentanol and higher), but in my experience it does not work on t-butanol and t-pentanol (the reaction does go forward, but stops at a certain equilibrium point). In fact, if you develop a method that allows the preparation of potassium t-alkoxides from KOH, it would be quite something. Molecular sieves perhaps do work in this case as well, but I have never seen any examples (I would tend to think that only partial conversion could be achieved).

What is a bit more involved is to obtain the solid alkali alkoxides. You can rotavap the alcoholic solutions and purge out the residual alcohol by adding toluene, triturating and rotavaping it away (residual toluene usually does not harm reactions that require solid alkali alkoxides). Another method is to precipitate them from the alcoholic alkali hydroxides where the equilibrium favors the alkoxides due the the pKa(alcohol) of alcohols being lower than that of water (works well for sodium ethoxide with acetone as the antisolvent - the reference is posted somewhere on the forum).

Beware, that dry alkali alkoxides are pyrophoric and can catch fire if heated on air or in intimate contact with some materials (some can light up a paper towel used to wipe them). Though most are quite tame when compared to the classical pyrophoric stuff, some should be treated with care (e.g., alkali alkoxides of secondary alcohols).

iAmOK

blogfast25 - 18-4-2013 at 09:56

Today I tried to synthesise potassium 3-methyl-1-butoxide (IUPAC?) from isoamyl alcohol and KOH, according to US patent 1,978,647, with mixed results.

46.4 g of dry isoamyl alcohol was mixed with 14.8 g of technical KOH flakes, that’s a molar ratio of 2/1 (based on 100 % KOH).

Refluxing at about 122 C for less than 30 minutes dissolved the KOH effortlessly, so far so good, and resulted in a two phased layer: a smaller clear, colourless layer at the bottom and a larger clear, light yellow layer at the top. I presumed that the bottom layer was mainly water + some KOH. I was planning to separate the two with a separation funnel but on cooling first the bottom solidified and then the top followed. So I reheated it a bit and the top melted again to a quite syrupy yellow liquid which I decanted off.

Here’s the solid bottom, which is clearly crystalline:



I then added slowly about 50 ml of technical acetone to the warm top and the colour changed to wine red. A white crystalline material also dropped out but the amount was less than a quarter of the solids in the bottom fraction, so probably not worth working up.



I’m presuming that due to the relatively low molar ratio of alcohol to KOH, the solubility limit of the alkoxide may have been reached. Perhaps it would have been advisable to add some inert (in this context) solvent at the start or to the hot top.

I’m now going to see if there may be any alkoxide in the solid bottom.

Any suggestions would be welcome.

Edit: I dissolved quite a bit of the bottom white solid in water (in which it dissolved easily), then acidified the solution. I didn't see any alcohol floating to the top (isoamyl alcohol is only sparingly soluble in water). Looks like that solid is mainly KOH... :(

[Edited on 18-4-2013 by blogfast25]

Nicodem - 18-4-2013 at 10:47

Quote: Originally posted by blogfast25  
Refluxing at about 122 C for less than 30 minutes dissolved the KOH effortlessly, so far so good, and resulted in a two phased layer: a smaller clear, colourless layer at the bottom and a larger clear, light yellow layer at the top.

Rather avoid such high temperatures when dealing with primary alkoxides or hydroxides in primary alcohols. They tend to decompose and the resulting aldehyde immediately condenses to colored crap (the initial yellow color can get quite dark, if you let it reflux longer). I suggest you to rather try dissolving the hydroxide at lower temperatures, as you need a saturated solution at room temperature anyway, given that you try to precipitate the alkoxide with acetone. To get proper yields of the alkoxide this way, I suggest you to first prepare a saturated anhydrous potasium isoamylate in isoamyl alcohol first, by using KOH flakes as the drying agent (as described in the patent for sodium ethoxide cited elsewhere on this forum). Thus you need to stir KOH in isoamyl alcohol (better do this overnight, as this is a very slow process at room temperature). You can warm it up (e.g. to 60 °C) for some time, but then cool it to room temperature before the phase separation (grinding the KOH flakes will also speed up things). Remove the excess KOH and the adhering saturated aqueous phase away by decanting the upper alcoholic phase. Add more KOH flakes to it and repeat the process. After two or three repetitions, you should have a fairly anhydrous solution of the alkoxide. Then cool the solution and try the precipitation with acetone and filter rapidly under a dry atmosphere, wash immediately with plenty toluene and dry in a desiccator.

However, I believe it would be better to just rotavap the solvent away. The yields with the acetone process certainly cannot be good, because acetone is nearly not a good antisolvent for potassium alkoxides of such chain length (potassium isoamylate is certainly way more soluble in acetone than sodium ethoxide).
Quote:
I presumed that the bottom layer was mainly water + some KOH. I was planning to separate the two with a separation funnel but on cooling first the bottom solidified and then the top followed. So I reheated it a bit and the top melted again to a quite syrupy yellow liquid which I decanted off.

That's quite a normal behavior for such a ratio of KOH vs. the alcohol. Consider that there is not enough water formed to dissolve such a huge amount of KOH. That's why doing it by successive dissolving&drying with several portions of KOH should be way more effective in removing water and leaving you with prevalently anhydrous alkoxide solution already before you selectively precipitate it out.
Quote:
I then added slowly about 50 ml of technical acetone to the warm top and the colour changed to wine red. A white crystalline material also dropped out but the amount was less than a quarter of the solids in the bottom fraction, so probably not worth working up.

You should not do the precipitation in a warm solution. Acetone self-condenses to crap when exposed to strong bases or acids, especially upon heating. Rather follow the exact patent instructions. I don't remember, if there is an example for potassium isoamylate or any related alkoxide in the patent, but you should read it carefully nevertheless as there are pertaining issues.
Quote:
I’m now going to see if there may be any alkoxide in the solid bottom.

Don't bother with that fraction.

By the way, what does this have to do with this thread topic? You just recently created a thread about alkali alkoxides. (edit: moved the off topic posts to the on topic thread)

[Edited on 20/4/2013 by Nicodem]

blogfast25 - 18-4-2013 at 11:58

Nicodem:

Starting with your last point, I got your U2U but the https thread won’t open for me. And I can’t find the thread anywhere in the ‘Organic section’. That’s why I parked it here. If you want to transplant it, feel free and let me know here it is.

It’s related to this thread in the sense that I’d like to try the reduction of a K alkoxide with Mg in sat. hydrocarbons, an alkoxide other than K t-butoxide, as per the sticky K thread. K 3-methyl-1-butoxide ‘felt’ like the right chain length to me.

It looks like I’ve bitten off more than I can chew though, going by your illuminating remarks. I’ve also ran out of isoamyl alcohol but could prepare some more from the isoamyl acetate that I bought for that purpose. Stinks like a megaton ‘banana bomb’, that stuff!

I think I’ll try with isopropyl alcohol over the week end: the longer alkoxides are definitely more problematic than the methoxides/ethoxides, using the patent’s method. The patent is good on broad lines, not great on the detail. It also makes you wonder how the inventors of the KOH/Mg/ROH method prepared their K t-butoxide, if not by reaction of the metal and the alcohol.

Thanks.

Nicodem - 18-4-2013 at 14:27

Quote: Originally posted by blogfast25  
Starting with your last point, I got your U2U but the https thread won’t open for me. And I can’t find the thread anywhere in the ‘Organic section’. That’s why I parked it here. If you want to transplant it, feel free and let me know here it is.

OK, so I transplanted the posts to keep it together, but you have been around long enough to know that I automatically move all the threads from the Organic section that started by using the beginner's discourse to the Beginners section. There is a good reason for that. Read the forum guidelines for more info on how to avoid this issue.
To access the HTTPS protocol you need to accept the certificate when your browser offers it (I thought most members use HTTPS anyway, as there is no good reason to use HTTP). Otherwise, when you look for your own posts, there is a nice function in your profile page that says: "Search for all posts by this user" (or just click the Find button bellow any of your own posts).
Quote:
It’s related to this thread in the sense that I’d like to try the reduction of a K alkoxide with Mg in sat. hydrocarbons, an alkoxide other than K t-butoxide, as per the sticky K thread. K 3-methyl-1-butoxide ‘felt’ like the right chain length to me.

But iAmOK and iPrOK are not tertiary alkoxides, they are primary and secondary respectively. The "make potassium" reaction requires temperatures of 200 °C. How are you going to prevent the beta-hydride eliminations in these alkoxides when this reaction occurs already at way lower temperatures?

blogfast25 - 20-4-2013 at 04:32

Thanks all for your responses, I’ve only just been able to access this thread.

Nicodem, to address your last point first, it’s an ‘Aha!’ moment for me: thick as two very short planks on organic chemistry here never realised why tertiary alcohols are needed for the potassium reaction. So now I know!

Yesterday and still in the absence of that knowledge, I tried the same procedure (that I tried on isoamyl alcohol) on 99 % isopropanol, because it’s quick and easy, with a similar result.

After about 20 minutes of refluxing at 82.5 C of a mixture of 50 ml IPA and 12.2 g of KOH flakes (a molar ratio of alcohol to KOH of 3:1) a perfectly clear and colourless liquid was obtained. On cooling though a bottom phase developed, which a bit later solidified, into an opaque, colourless crystalline substance. The amount, judging by volume, was about the same as the starting volume of KOH.

The supernatant liquid, which this time showed no signs of solidifying, was decanted off into a beaker and put on an ice bath. As said, this was clear and colourless. While this cooled, the crystalline bottom solid was investigated.

Firstly I washed it with small amounts of water to get rid of superficial IPA on its and the RBF’s surface, replaced the IPA vapour inside the RBF with fresh air, then dissolved the solid in water. Only a small amount of water was needed, it dissolved quickly with some heat generated and the solution was slightly turbid. Then it was neutralised carefully with HCl, much heat evolved but I couldn’t detect the smell of IPA. So this solid in all likelihood contained no alcoholate. Crystals of KCl formed on cooling.

At 5 C, 50 ml of acetone was added to the cooled IPA supernatant phase. At first some turbidity formed but it disappeared again. Adding another 50 ml of acetone caused permanent turbidity and discolouration to light orange. I’m leaving this to chill further in the fridge but hopes of obtaining worthwhile quantities of alcoholate seem slim.

I guess that to establish a ‘baseline’ I’ll now have to try the patent’s procedure using methanol. Refluxing should not be necessary with methanol.

It would appear that for isoamyl alcohol and isopropyl alcohol the equilibrium:

KOH(s) + ROH(l) ↔ K+(diss.) + RO-(diss.) + H2O(diss.) [with R = isopropyl, isoamyl]

… shifts to the right on heating and back to the left on cooling. That would suggest (entropic effects aside, the change in entropy must be positive here) that the reaction enthalpy (left to right) is endothermic. It also suggests that cold solubility of the alcoholate is poor (but that’s saying the same thing in different words).

To put this to a quick and oversimplified test, about 50 ml of IPA and 20 g of KOH were loaded into a clean 100 ml RBF, stoppered and with a thermocouple sticking into the solvent. That was allowed to acclimatise without stirring, to 14.3 C. Then the flask was swirled vigorously for a few minutes, during which temperature dropped to 14.0 C. That’s hardly significant but then very little KOH dissolved during that time and at least there wasn’t any exotherm.

I’m beginning to seriously doubt if this patent works for anything but methoxides and ethoxides.

I think I need to go back to your suggestion on another thread of making isoamyl chloride from banana oil, then use that with acetone to convert it to a t-alcohol via Grignard...


[Edited on 20-4-2013 by blogfast25]

Nicodem - 20-4-2013 at 08:00

Quote: Originally posted by blogfast25  
The supernatant liquid, which this time showed no signs of solidifying, was decanted off into a beaker and put on an ice bath. As said, this was clear and colourless. While this cooled, the crystalline bottom solid was investigated.

I already gave you a hint in some previous reply that you are looking for the product in the wrong phase. Potassium isopropoxide is highly soluble in isopropanol. For this reason, you will not find it in the KOH phase.
Quote:
At 5 C, 50 ml of acetone was added to the cooled IPA supernatant phase. At first some turbidity formed but it disappeared again. Adding another 50 ml of acetone caused permanent turbidity and discolouration to light orange. I’m leaving this to chill further in the fridge but hopes of obtaining worthwhile quantities of alcoholate seem slim.

As I said previously for your isoamyl alcohol attempt, the chances that acetone works as an efficient antisolvent for alkoxides other than sodium ethoxide and methoxide are slim (so much less for the potassium salts and even much less so for the secondary alkoxides - e.g., iPrOK). Perhaps it might cause the precipitation of some product, but the yield will likely be poor. I know I keep on repeating myself, but solvent removal is a much better option for alcohols other than methanol or ethanol, especially if you have an efficient distillation column to take away also the residual water.
Quote:
It would appear that for isoamyl alcohol and isopropyl alcohol the equilibrium:

KOH(s) + ROH(l) ↔ K+(diss.) + RO-(diss.) + H2O(diss.) [with R = isopropyl, isoamyl]

… shifts to the right on heating and back to the left on cooling.

I think you are mixing up two different equilibriums here. One is the proton transfer equilibrium which is in favor of the alkoxide side of the equation (pKa values in alcohols demonstrate this). The other is the solubility equilibrium which will force the hydrated KOH to precipitate out of a saturated solution made by dissolving large amounts of KOH in refluxing alcohols.

unionised - 20-4-2013 at 08:08

Have you allowed for the fact that commercial KOH is not very pure?
85% is typical

blogfast25 - 20-4-2013 at 08:44

No, Nicodem, I’m not looking for the alkoxide in the wrong phase, rather I was corroborating that it wasn’t in the bottom phase. Not the same thing. Also, the IPA experiment was carried out before I could access this thread.

As regards giving preference to evaporative removal of solvent, agreed but I don’t have a rotovap or access to decent vacuum, only to about 250 mm Hg. And w/o effective removal of water the product will only hydrolyse.

Perhaps for more aliphatic alkoxides another antisolvent could work better?

Unionised, I’m certainly aware of it (as I hinted to it) but how does one allow or account for it?

All this has become a bit surplus to requirements because the only alkoxides I’m after would be those of t-alcohols. I have to assume these need to be prepared by refluxing the alcohol with the metal and subsequent evaporative removal of the excess alcohol.


[Edited on 20-4-2013 by blogfast25]