Sciencemadness Discussion Board

Help with base-catalyzed claisen rearrangement of catechol monoallyl ether

bleckster - 20-11-2013 at 15:22

I'm attempting to synthesize vanitrope (3-hydroxy-4-ethoxy-1-propenylbenzene) by producing the monoallyl ether of catechol using allyl chloride and sodium iodide in acetone, doing a base-catalyzed Claisen rearrangement in ethanol with sodium ethoxide to produce 4-allylcatechol, ethylating that to 3-hydroxy-4-ethoxy-1-allylbenzene, and finally isomerizing that to vanitrope. This prodecure is according to:

Laskina; Chemical Abstracts 61 11919c 1964
https://www.dropbox.com/s/7m61kwi6t6rvg7t/Chemical%20Abstrac...

Which states 100g (0.9 mol) catechol, 100g (1.3 mol) allyl chloride in anhydrous acetone (200ml) with 15g (0.1 mol) sodium iodide and 106g (1 mol) Na2CO3 yields the mono allyl ether in 75% yield.

The problem I am having is I can only get 38% yields from this procedure. Furthermore, I am failing to produce 4-allylcatechol by rearranging the catechol monoallyl ether through either a base-catalyzed method or a normal claisen. I only get a little 3-allylcatechol and the rest is not distillable. I could really use some advice.

I figure I should start by describing the first synthesis just in case the problem lies there.

As far as I understand it, the above reaction is a williamson ether synthesis combined with a finkelstein reaction, substituting the iodide for the chloride in the allyl chloride, leaving the chloride to combine with the sodium and precipitate out since sodium chloride is insoluable in acetone, whereas sodium iodide IS soluable. Then the allyl iodide swaps the allyl group with a hydrogen on the catechol, thus freeing up the iodide to combine with more allyl chloride. I know that some allyl chloride still reacts with the catechol directly, so the sodium iodide acts as a PTC and it doesn't need to be equimolar with the allyl chloride. I believe the sodium carbonate is there to neutralize the left over chloride and iodide ions.

The problem is, I'm only getting 38% yields. I've scalled this reaction down to 5% of the above amounts in my experiments. No reaction time is specified, but I chose 8 hours based on a German reference I found for a similar reaction:

Mauthner; J. prakt. Chem. 1937 148 95
https://www.dropbox.com/s/0cvw92w009jnkz9/Mauthner%20J.%20pr...
https://www.dropbox.com/s/yfed7itumwjlrad/Mauthner%20J.%20pr...

60 g guaiacol, 50 g allyl chloride, 50 g of sodium iodide dissolved in 200 cc of acetone with 55 g of potassium carbonate for 8 hours gave 54.5 g Allylguaiacol ether.

You'll notice that he uses 3 times as much sodium iodide as Laskina. I made that change but there was no change in yield. Mauthner's yields were 69% of theoretical. My product distills between 80-83C at 4mmHg. According to the references I've found, catechol monoallyl ether distills at 103-104.5/8mm, 107.5-109/15mm and 108-113/18mmHg, so my BP seems to track.

I should mention that I dried my acetone over 3A mole sieves for 2 days and that the catechol and allyl chloride were purchased from a reputable supplier, however the sodium iodide and anhydrous sodium carbonate came from eBay. The iodide was in a merck bottle (not sealed), and the carbonate repackaged in a plastic bag. I suspected both of the eBay chems, so I assayed the iodide according to http://www.youtube.com/watch?v=ZFP7fYGrFRI and it's definitely sodium iodide.

I also confirmed the carbonate through a pH test, solubility, and by adding HCl to a saturated solution and observing the carbon dioxide and sodium chloride precipitate. I further suspected it might not be anhydrous Na2CO3, so I dried 25g in the oven and got 24.5g (there was some residue left on the parchment paper). If it was the monohydrate, there should have been a 3 gram loss.

I've UTFSE and noticed two instances when Nicoderm has recommended not using sodium carbonate in this type of reaction but potassium carbonate instead. I've got some on order now and I plan to try that next, but other than that I can't think of anything else to try. Any other help on yields would be appreciated.

I am aware that the mono allyl ether is normally made using allyl bromide, however I already had the sodium iodide and the allyl chloride was one third the cost. Believe me, I'm already regretting not springing for the bromide in the first place but these are the reagents I've got and I'd really like to see this through if it's possible.

Oh, I should mention the workup I am using and the various workups I've seen. I am generally following this reference:

Perkin & Trikojus; Journal of the Chemical Society 1927 1663
https://www.dropbox.com/s/bw6c35h1m87ayih/Journal%20of%20the...

My workup (from a larger 0.5 mol batch): "The acetone is recovered and the remainder taken up in 100 mL water, acidified with 10M HCl, and extracted with 2 x 50 mL DCM. The mono allyl ether and unchanged catechol were extracted from the DCM with 2 x 100 mL 5% NaOH. The alkali extracts were acidified with 25 mL 10M HCl and the bottom oil layer taken up in 50 mL DCM, separated, and the water extracted with another 50 mL DCM. The DCM was removed from the combined extracts and the oil fractionally distilled in vacuo."

I don't wash the final non polar with water anymore like they do because I've found that a considerable amount of oil is dissolved in it (I had to saturate with NaCl to recover it).

Rao & Rao in Ind. J. Chem. 2, 323 (1964) use this:

"(after removing acetone) the filtrate was dissolved in ether, acidified and washed free of acid. It was dried, concentrated and distilled"

In other words, they skip the step of separating the diallyl ether and they don't report distilling any either, claiming a modification to the synthesis avoids formation of it. As far as I can tell this modification is to add the allyl bromide dropwise.

Anyway, the problem I'm REALLY having is rearranging the ether. A normal claisen rearrangement reportedly yields 4-allylcatechol in 35-40% yield, but according to:

Ollis, JCS Chemical Communications 1974 494
https://www.dropbox.com/s/amlli3cmwapzlxk/JCS%20Chemical%20C...

The ether can be rearranged in an 80% yield if reacted with an equimolar amount of sodium ethoxide in ethanol. Once again, no reaction time is given. In fact, no experiments are reported at all in that paper, no workup or anything. On my first run through, I ran it for 4 hours but I dropped the sep funnel during the workup (grr) and lost half the product and when I distilled the remainder nothing came over. On the second run through I ran it for 2 hours and got mostly unreacted ether and some 3-allylcatechol. At this point I decided to make a bigger batch of the ether, poor yields and all, so I could at least run a number of rearrangements. I also decided to try a regular old claisen just so I could get the final product and confirm that the ether was what I thought it was. And here's where I've gotten so much more confused.

All reports say to heat the ether in an inert atmosphere to around 160-170 and a vigorous reaction will ensue, with momentary boiling, with the temperature shooting up to 260-280C and the color turning red. The Raos held their reaction in the oil for 5 mins before cooling. Perkin & Trikojus don't make any mention of reaction length, so I can only assume they stopped once the temperature began to drop. Laskina mentions heating the oil for 30-40 mins at 160 before a temperature jump takes place. That is very confusing because I can tell you from experience the reaction kicks in the minute you hit the 160-170 target range. Perhaps it's because they were heating 100g of the ether and it just took that long to heat up. In all of these reports, considerable amounts of ether were used (30g - 100g). They all used oil baths.

I've run several claisen rearrangements on 2.5 g of my product under argon and all I've ever managed to get is a fraction boiling between 116-118/4mm which would not solidify at normal pressure (3-allylcatechol) and nothing distilling over past that. The first time, I heated it directly on the hotplate, observed the boiling and the change in color to red with the temp rising but it did not reach 260 (I think it was around 230ish). I think it didn't get up that hot because there was only 2.5g instead of 30-90g. I also thought it had to hit 260 so I let it get up there and then cooled and the whole dark red mass solidfied and nothing came over. Next I used an oil bath at 215C and brought the mixture just up to 160-170 wherein the reaction kicked in and the temp jumped to ~230 and started dropping after a few seconds, removed it from the bath and cooled it (it was orangish), distilled and got unreacted ether and 3-allyl. I took the unreacted ether up to temp very slowly on the hotplate and removed it at 245 (it was dark red but it never did boil). It solidified when cooled and did not distill. Another reaction brought to 232 produced a tiny bit of 3-allyl and the rest did not distill.

In one reference I did notice this paragraph (I'll have to reproduce the text as I ran out of change for the copy machine while at the library and took a picture of the page instead):

J. Am. Chem. Soc. 52 1700 (1930)
Concerning the mono allyl ether of resorcinol: "In another experiment 40 g. of the ether was taken. At an outer temperature of 235, the inner temperature ascended to 310 and the contents of the flask were boiling vigorously. Only 5 g. of distillate could be obtained from this. It was a very viscous, red liquid which hardened on standing. The residue in the flask solidified to an amber resin. Quantities less than 20-25 g. should be taken to avoid this vigorous exothermic reaction which leads to tars. In addition to this precaution the heat should be removed from underneath the bath once the exothermic reaction starts."

So obviously there is an upper temperature limit involved and the more ether being reacted, the hotter the reaction gets.

Next I decided to investigate whether the temperature resulted in the color change or if the reaction duration did:

Catechol mono-allyl ether (2.53 g, 62 mmol) was heated in an argon atmosphere very slowly on a hotplate until reaching 165-170 upon which a vigorous reaction ensued with boiling and a sharp temperature rise to 208 and the ether turning yellow. A timer was started and the temperature slowly dropped before settling at 198. At T+4 minutes the temperature was a deeper yellow. At T+6 the temp was 200C and color had hints of orange. At T+9 the temp was 202 and color dark orange. At T+13 temp was 202 and color was reddish orange. Heat was discontinued at this point. A T+19 the temp was 157C and color a cherry red.

Once again only unreacted catechol and 3-allyl. I neglected to weigh the amounts, so I apologize for that. That would have possibly provided some hints whether I was getting close or not.

I'm hard pressed to draw any conclusions from this. Perhaps when larger amounts of the ether are involved, the temp range--that sweet spot--where the rearrangement happens is larger and the self-sustained reaction is enough to get it there, whereas with smaller amounts of ether that range is smaller? Or maybe the temperature needs to shoot up above a certain limit right when the reaction first kicks in and only then is the 4-allyl produced?

Perhaps I need to react a larger amount of ether, such as 30g, and see if I can get some results? Or does the problem lie in the ether? Perhaps it isn't what I think it is--although I don't see how. The boiling point matches and it DOES rearrange...

Anyway, finally this brings me to the base-catalyed rearrangement. The big issue I'm having is with the workup. I took it from a sketchy reference and had a feeling the author who came up with it had never tried it:

Catechol mono-allyl ether (2.53 g, 17 mmol) was refluxed with sodium ethoxide (1.15 g, 17 mmol) and anhydrous ethanol (6.91 g, 150 mmol) for 8 hours. The solvent was evaporated on a water bath and replaced with 20 mL 2M NaOH. The alkali solution was washed with 10 mL DCM, acidified to <3 with ~8 mL concentrated HCl, and extracted thrice with 20 mL portions of DCM. The DCM was dried with MgSO4, concentrated, and vacuum distilled with 1.32 g coming over between 110-120 at 4-5mmHg. The flask was dry with black crud.

I know what you're asking--what happened to the other 1.2g? I attempted to distill the alkali water but apparently something steam distilled over with it, as it was cloudy and smelled a bit like an allyl compound and something else and also had some small droplets floating on the surface. I added some NaOH to it and it turned black, and from the smell I realized there was ethanol in there and I had turned it back into ethoxide! Oops. I extracted it with a bit of DCM, which was mostly clear. And there I have left it. I also forgot to stopper the DCM and it evaporated overnight leaving a few drops of a colored oily substance. I'm unsure where to go from here--so I decided to wait and ask for help. Regardless, given the fact that it's only 1.2 g, that's definitely not 80% of 2.53 g. I've decided to order some TLC plates so I can properly track the progress of this reaction so I can at least know when the ether has been changed.

Can anyone suggest a better workup for next time?

I appreciate everybody's help.

Help on a Williamson Ether Synthesis / Finkelstein

bleckster - 22-11-2013 at 07:04

I posted part of this in another thread but I think I went into far too much detail, so I'll try to be more brief. I'm attempting to synthesize vanitrope following this procedure:

Laskina; Chemical Abstracts 61 11919c 1964

The first synthesis states 100g (0.9 mol) catechol, 100g (1.3 mol) allyl chloride in anhydrous acetone (200ml) with 15g (0.1 mol) sodium iodide and 106g (1 mol) Na2CO3 yields the mono allyl ether in 75% yield.

What amount of acetone would you use for this?

How long would you run the reaction for?

Would you add all the reactants together and bring them to reflux? I was thinking that perhaps bringing everything but the allyl chloride to reflux first and adding that dropwise would help the finkelstein move along better, since there would always be an excess of sodium iodide?

I've also read in Organic Synthesis Vol 2 pg 22 that methyl ethyl ketone can be used instead of acetone to improve reaction time due to the higher BP.

One thing I don't understand is that the williamson is supposed to react a phenoxide with an alkyl halide, but Na2CO3 is insoluble in acetone. If it doesn't dissociate into 2 Na+ and CO32-, how can the phenol anion be formed?


[Edited on 22-11-2013 by bleckster]

Nicodem - 22-11-2013 at 09:35

Don't cross-post! And link threads when you refer to them. Read the forum guidelines for more information.
Quote: Originally posted by bleckster  
I know that some allyl chloride still reacts with the catechol directly, so the sodium iodide acts as a PTC and it doesn't need to be equimolar with the allyl chloride. I believe the sodium carbonate is there to neutralize the left over chloride and iodide ions.

No, the NaI does not act as a phase transfer catalyst. It is only used for the nucleophilic catalysis. The sodium carbonate is there to act as a base for the deprotonation of the catechol (the phenolate anions are way more nucleophilic than phenols). The chloride and the iodide ions don't need any neutralization, they are not acidic, utmost they could be basic, but that is irrelevant.
Quote:
The problem is, I'm only getting 38% yields.

You don't mention whether the problem is in the poor catechol conversion or poor selectivity.
If the problem is in the poor conversion, the obvious physical solution is to prolong the reaction time. However, given that the reaction time for that was 8 h, you would need to prolong it to at least 24 h to potentially get to about 85% (but since the reaction is heterogeneous, it is unlikely you would get above 50% even at 24 h).
The most obvious chemical solution would be to use a more suitable method. Sodium carbonate in acetone is crap for alkylations of phenols. I'm surprised you even got that much product. You should be using potassium carbonate.
Quote:
You'll notice that he uses 3 times as much sodium iodide as Laskina. I made that change but there was no change in yield.

I'm not surprised. Nucleophilic catalysis can't do miracles in a reaction where the obstacle is not so much the reactivity of the electrophile.
Quote:
I should mention that I dried my acetone over 3A mole sieves for 2 days

That is unnecessary.
Quote:
I've UTFSE and noticed two instances when Nicoderm has recommended not using sodium carbonate in this type of reaction but potassium carbonate instead.

This will then be the third instance. If you insist on sodium carbonate, then you will have to change the solvent to methanol or PEG, or to do it solventless with a PTC. Sodium carbonate might also work in DMF, though not necessarily. Obviously, all such changes in the allylation method will also change the reaction selectivity. (and it is "Nicodem", not "Nicoderm")
NaOH or KOH in methanol, ethanol or 2-propanol are very efficient for the alkylations of phenols, but you might loose on the monoallylation selectivity. Also, in protic solvents you can get a higher amount of C-allylation of phenols in general due to H-bonding shielding of the phenolate oxygen (5-10% of the C-allylated and C,O-diallylated product can be normal under such conditions). Some C-allylation can also occur in aprotic solvents as well (it is hard to avoid it, as the allyl halides are soft electrophiles, so they have a higher tendency to attack the softer positions of the ambident nucleophiles).
Quote:
I don't wash the final non polar with water anymore like they do because I've found that a considerable amount of oil is dissolved in it (I had to saturate with NaCl to recover it).

With the calculated logP of 2.2, the monoallyl catechol ether can not partition significantly from dichloromethane into water at pH < 8. Whatever you salted out must have been something else.

I seem to remember that there exists an article claiming that direct C-allylation of catechol is possible with some allyl halide. The trick was to allylate the Cu-catechol complex, where the O-akylation is blocked. You might check the literature for the direct allylations and avoid all the problems with the allylation and the rearrangement. I might sacrifice some time to find you the reference, but only after you convince me that you were unable to find it yourself. But then again I might have confused it with something else (perhaps it was the C-allylation of guiacol to give eugenol?).

bleckster - 22-11-2013 at 13:34

Thank you very much in taking the time to respond.

Quote: Originally posted by Nicodem  
No, the NaI does not act as a phase transfer catalyst. It is only used for the nucleophilic catalysis. The sodium carbonate is there to act as a base for the deprotonation of the catechol (the phenolate anions are way more nucleophilic than phenols). The chloride and the iodide ions don't need any neutralization, they are not acidic, utmost they could be basic, but that is irrelevant.


Ah, I misspoke twice, it seems. I must apologize. I should have said the NaI acted as a catalyst, not a PTC. I only meant to imply that it did not need to exist in an equimolar (or excess of) ratio to the allyl chloride, as in a normal Finkelstein. Also, I should not have said the carbonate was there to "neutralize" the halogens. What I meant was, the halogen anions would bond with one of the sodium cations from Na2CO3, leaving NaHCO3 (the H coming from the protonated phenol) and NaCl and NaI. At least, I think.

Quote: Originally posted by Nicodem  
You don't mention whether the problem is in the poor catechol conversion or poor selectivity.


Very good point, I should be recovering and quantifying the unreacted catechol, which I haven't been doing. That would be the obvious way to infer the reaction rate when experimenting without TLC. I will do that in all future experiments.

In my last run from 49.55 g (450mmol) catechol I got 28.84 g mono and 15.55 g diallyl, so the conversion was 65.68%. Hmm, I never thought about the fact that the original chemical abstracts never mentioned Laskina making any diallyl. I thought they were just being brief, but perhaps they never did. Maybe my problem is in the method. Remember I mentioned that Rao & Rao modified the allyl bromide method, adding it dropwise, and thus produced no diallyl? I should try the same thing with the chloride. What do you think?

The KCO3 is here and I will use that next time as well.

Quote: Originally posted by Nicodem  
NaOH or KOH in methanol, ethanol or 2-propanol are very efficient for the alkylations of phenols, but you might loose on the monoallylation selectivity. Also, in protic solvents you can get a higher amount of C-allylation of phenols in general due to H-bonding shielding of the phenolate oxygen (5-10% of the C-allylated and C,O-diallylated product can be normal under such conditions). Some C-allylation can also occur in aprotic solvents as well (it is hard to avoid it, as the allyl halides are soft electrophiles, so they have a higher tendency to attack the softer positions of the ambident nucleophiles).

I seem to remember that there exists an article claiming that direct C-allylation of catechol is possible with some allyl halide. The trick was to allylate the Cu-catechol complex, where the O-akylation is blocked. You might check the literature for the direct allylations and avoid all the problems with the allylation and the rearrangement. I might sacrifice some time to find you the reference, but only after you convince me that you were unable to find it yourself. But then again I might have confused it with something else (perhaps it was the C-allylation of guiacol to give eugenol?).


Very interesting. I agree, let me sacrifice *my* time researching the direct allylation methods first. I will only impose on you if I'm at a loss.