eugenol demethylation

[Cadenza]

The intermediate 4-allylcatechol has been obtained in good yield (75%+) from eugenol.  The process is easy, simple, and scalable; however, there are some issues that I would like to address before writing further on the topic.

1.  Melting point analysis indicates a range between 38-41 Celsius.
2.  IR spec is inconclusive- I can make out twin catechol peaks in the ~3400-3600 range but it is very messy- lots of static everywhere, especially in the upper range (absorbance).  There appears to be no upper limit to the IR spec- this is also a problem with eugenol (but at least eugenol has a clearly defined range > 2000).

In conclusion, does anyone know what might be causing this interference?  Do we even need an IR spec analysis? (Previous reports in literature have been analyzed only through M.P.- gotta love the old days when that was good enough...)

Note: Subsequent methylenation of 4-allylcatechol synthesized by this method has been shown to produce a water-white liquid with a strong root-beer smell.  So I know that there is at least some 4-allylcatechol in there, but the IR is an anomaly, to say the least...

[embezzler]

spectrum that you can post?

[Cadenza]

It's still at the lab- I could it retrieve it, but I would be very embarrased to post it, as there are so many artifacts in it.     It sure isn't readable to me, but then again I am a very poor analytic chemist (actually not a chemist at all) and perhaps others might be able to shed some light on it.

Incidentally, I am not familiar with IR technique, but generally I don't encounter this sort of spectrum.  Even with the other demethylations that produced tar, severe dilution of the product with IR solvent would get rid of some of noise.  I have diluted the original IR mix at least 100^6 fold, and it is still pretty much producing the same spec.  Perhaps there is some sort of solvent contamination going on here.

Anyway, I will grab a copy of the spec soon and post it if there is interest.

[embezzler]

but running the solvent as background  should remove any solvent signal, what solvent did you use?

it will certainly be interesting to see......

[_mu_]

You (or somebody else) didn't make the IR machine dirty somehow, did you?

[embezzler]

especially in college where accountability is nil. the ionic disks are often in need of serious replacement.

[homeslice]

JW, but what demethylation method was chosen?

The process is easy, simple, and scalable; however, there are some issues that I would like to address before writing further on the topic.

Im sure more than a few bees, including myself, are very interested in your future writings on the topic. A writeup of how it all went would be fantastic.  

[Cadenza]

embezzler and __mu__:

The background solvent used was CCl₄.  I am not fond of working with such toxic reagents, but the catechol and eugenol IR specs from library references used CCl₄, so CCl₄ was used for the sake of consistency (background should be automagically filtered out, but CCl₄ has an advantage in that it doesn't have much of a signature in the OH range).

The IR machine is/has been clean, and for the most part is well-behaved.  Perhaps I will try a different solvent (maybe toluene?)  Or maybe the product is inherently dirty with all sorts of polymerization (it looks OK to me, and the melting point range isn't that bad, but you never know).

With regards to the writeup, I will give more details soon, pending resolution of the IR issue.  For all I know, I've synthesized some crazy polymerized crap (although the subsequent methylenation into safrole appears to demonstrate that at least something is there).  Does anyone know how polymerization or rearrangement into the corresponding quinone will affect melting point range? (i.e. if the quinone is more stable/polar, I would expect the melting point to rise).

homeslice:

This was an amine-base method that I worked out a while ago.  For a while, based one extensive experimentation, I had concluded that triethylamine*HCl caused excessive decomposition of the product under microwave radiation.  More recently I discovered that using a biphasic aqueous layer w/conc. HCl and triethylamine, under long microwave irradiation, would result in demethylation without significant thermal degradation.  The yield of the reaction was largely unknown until Psychokitty gave me some pointers about purifying the allylcatechol (I have not yet found a way to crystallize the product).  It is unknown whether demethylation will proceed under traditional reflux conditions, but I doubt it because the reaction proceeds at the thermal boundary layer between the two phases.  The microwave method relies on the fact that eugenol heats up much faster than water/triethylamine/HCl- I am working out some rough heat transfer calculations right now.

The result is that finally I have a product that is can be ground up (i.e. not viscous at room temp).  I suspect that literature reports about 4-allylcatechol being a "red-yellow viscous solid" are due to eugenol contamination, as this is the type of product that was obtained pre-purification.

In addition, 4-allylcatechol (if this is indeed what I have) has some very interesting optical qualities that I will discuss later.

[Nicodem]

If the product is a solid why didn't you simply make the KBr-pellets for the IR? If it is somewhat oily than you can just disolve it in some DCM and put a drop on a NaCl window, wait the DCM to evaporate an cover with the other NaCl monocrystalic window for the IR. That is the usual procedure for as much as I know. If the spectrum still isn't nice reduce the amount (even 0.5mg can sometimes still give you a nice spectrum).

background should be automagically filtered out, but CCl4 has an advantage in that it doesn't have much of a signature in the OH range

What do you mean by this? How can a bacground bee filtered off without runing it? And where did you hear CCl4 is a solvent for IR? I thought only nujol oil is a useful solvent, though avoided whenever possible. But then again I dont know anything more on making IR spectra than the rutine procedures.

Is the product completely colorless? No 4-allyl-orthoquinone present?

[Cadenza]

nicodem:

KBr/Nujol mull disks are the most common methods taught in organic chemistry labs (I have done my fair share) but the IR spec setup that I have uses a cuvette (it also does solid-phase scans but setting it up for that is more involved).

Of course the background was run before scanning the sample- you can use any solvent that you want (obviously some are better than others).  However I do not have Nujol on hand (neither do I have KBr discs), so CCl₄ it was.  In my experience, liquid phase IR samples are even easier to prepare than KBr-- admittingly, it has been a long time since I have prepared a KBr disc.  If you look at IR databases, you'll see that CCl₄ is used quite commonly (although not as common as Nujol/KBr).

The product is far from colorless- it is a characteristic tan color (think wet sand).  Why do you think that 4-allylcatechol is colorless?  Do you have a reference for this?- it has been described as everything from brilliant white to red-yellow to tan colored.  When I was running experiments with pyridine hydrochloride demethylations, the closest I got to a lighter colored substance was a yellow viscous solid (might have been eugenol contamination).

I personally believe that there might be some ortho-quinone present, but I cannot prove this.  If you have any information on the color/MP/properties of o-quinone-allyl, I would be very interested in seeing it, as this would help greatly in identifying contaminants.  In addition, if there is a way to reduce the quinone back to the o-catechol species, I would be interested in that too  

[Nicodem]

In addition, if there is a way to reduce the quinone back to the o-catechol species, I would be interested in that too

That's the reason I ask about color.
The allylcatechol shouldn't bee coloured but the quinone counterpart surely is (probably from green, yelow, brown to red - who knows? ). There are many ways to reduce orthoquinones and a suitable reducent could bee added at the work up stage too purify/recycle/discolour the product.

[embezzler]

was wet ?? water would throw some oh peaks in there...

swim doesnt like to suggest it but too large a particle size will scatter the beam, this problem is common in liquid samples?

[Cadenza]

Nicodem:

In general quinones are darker than their phenolic counterparts, right?  I think that your assumption that pure allylcatechol is either colorless or white (like phenol) is probably correct, which is unfortunate considering that I have never come close to obtaining a colorless product sample.  I do wonder if the catechol oxidation is due to lack of inert atmosphere (easily fixed, even in the microwave reactor).

With regards to reducing the quinione contaminants back to o-catechol, do you have any suggestions off the top of your head?  I had always thought that this would be difficult but I have never looked into it.  It would be nice to clear up the color- brown is never a good sign in my opinion!  

embezzler:

The CCl₄ is (or at least is supposed to be) of very high purity, I doubt that much water gets in the bottle since carbon tetrachloride is so volatile (high vapor pressure).  Since the liquid phase IR always worked on eugenol (some crazy artifacts under 2000 but very clean above that range) I assume that it is an impurity problem with the product rather than the solvent system.

[lugh]

According to the Dictionary of Organic Compouns, the easiest method for purification of 4-allylcatechol is crystallizaton:



Someone should retrieve and post Schopf's article, he synthesized 4-allylcatechol from eugenol using methylmagnesium iodide  

[Cadenza]

Lugh:

Thank you for the reference, I will reattempt crystallization tomorrow.  I am not familiar with arcane chemistry notation-- does C₆H₆-petrol refer to benzene, a mixture of benzene and something another solvent, or petroleum ether, etc?

[lugh]

The Dictionary of Organic Compounds is a British publication, the definition from a chemical dictionary

Petrol: A name used in the British Empire (except Canada) to designate gasoline.

IMHO since so many non polar organic solvents are fractions of crude petroleum, others may serve as well as gasoline without additives    The naptha used by various bees for pill extraction is such a fraction, if that's available it would be a possible option  

[Cadenza]

Lugh:

Thank you for the clarification (I was aware of the British name for petrol, but I didn't know that it applied here   ).  Since I am accustomed to using laboratory reagents, I might try some sort of equivalent mix-- maybe hexanes or pentane.

Incidentally I think there might be some water in the mix (as embezzler noted a while ago) contributing to melting point depression.  This might also be the reason that the IR spec is going crazy-- hopefully slow recrystallization will clear this up.

Do you have more information on the Schopf reference that was alluded to earlier?  I can grab it from the library and scan it in if there is interest.

[Rhodium]

Here is the article containing the grignard demethylation of eugenol - The actual procedure is detailed on pages 49-50 and they say that their yield of pure 3,4-dihydroxyallylbenzene from methyleugenol after treatment with methylmagnesium iodide is 33%, and that both eugenol and safrole reacts just the same, but gives a slightly less pure product.

Die Synthese des 3-Oxy-4-methoxy-phenylacetaldehyds (Homoisovanillins) und des 3,4-Dioxyphenylacetaldehyds (Homoprotocatechualdehyds)
von Clemens Schöpf, Eva Brass, Ernst Jacobi, Walter Jorde, Walter Mocnik, Ludwig Neuroth und Walter Salzer

Ann. Chem. 544, 30-62 (1940)

(https://www.thevespiary.org/rhodium/Rhodium/djvu/schopf-ann.chem.544.30-62.1940.djvu)

[L42L]

Wow cadenza cool beans.

As others, can’t wait to hear about your method.

Would LOVE to hear about your experiments with pyridine-HCl as well.

“under long microwave irradiation”
Long as in time? Are you using a mono-modal or domestic-ish system.

Have you toyed with the methanesulfonic acid demethylation?
(post # No 379200)
Would love to see how that one works in a non mono-modal reflux.

You sound busy, thanks for your time and experimentation!

[Cadenza]

Rhodium:

Thanks for the ref-- interesting info! (where do you find this stuff?  I thought that I had pretty much read everything in the literature concerning eugenol->4-allylcatechol)

L42L:

The microwave was a standard commercial model (i.e. multi-mode) with a hole drilled in the top to allow for a passage to a reflux condensor.  In fact, I got the idea for the microwave setup from one of Rhodium's posts.

Microwave reactors are very easy to construct from domestic microwaves.  I do not recall if anyone on this board has actually made one, but it took less than 2 hours to build, equip, and make the necessary modifications in order for the oven to work.  The entire cost of the project was on the order of $40 ($10 for the actual modification, $30 for the microwave detector). 

In my experience, the microwave leakage is almost negligible.  Of course, it is always wise to check irradiation levels, but I noticed that within a five foot radius, there was essentially no radiation-- even without the pipe shielding the condensor.

The main reason that I like microwave experiments is that they are very fast and easy to conduct- far more convenient than traditional reflux, which is about as easy as you can get in terms of orthodox methods.  The irradiation time used in this particular instance was 90 minutes, but I think it was total overkill; I hypothesize that it could reduced to < 30 minutes without change in yield, and perhaps this would help reduce polymerization etc.

The two most important outcome of the experiments (thus far) were:

1)  The substitution of triethylamine for pyridine in an aqueous reaction matrix.  Hitherto, the type of amine-base used in microwave cleavage had been severly restricted by the melting point of the respective hydrochloride salt.  But since the reflux point is lowered with the addition of water, product composition does not appear to occur (it is true that the reaction probably also occurs much slower, but experimentation suggests that it doesn't impact the yield).

I would like to point out that there are a lot of amine-bases available, including variations of pyridine (i.e. picoline etc.) as well as other aromatic/non-aromatics.  Research into these could yield useful results not only for eugenol demethylation, but for ether cleavage in general (who really wants to mess with BBr₃ or pyridine when deprotecting aryl ethers?)

2)  The mole ratio was bumped down from 5:1 all the way to 1.5:1 (triethylamine:eugenol) and triethylamine is still present in excess.

I haven't tried the methanesulfonic acid method-- it might be interesting for a future project, but I chose to stick with the amine-bases this time, mainly because there are so many of them.

I would suggest that if anyone is interested in the conversion of eugenol to safrole (or in methylenation reactions in general) that the second synthetic step (catechol -> benzodioxole) be first established.  I have tried many of the methods, including PTC w/TBAB/wide variety of surfactants as well as DMSO, with poor results.

The reason I say this is that the result from the first (and only) reaction that I ran from 4-allylcatechol -> safrole resulted in a yield of 20% safrole (overall).  This reaction was carried out with some no-name PTC that I had never heard of, but subsequent runs with established PTCs failed to improve the situation.  So until a reliable method is found to methylenate m-sterically hindered catechols (and I mean by experimentation, literature can fudge a lot    ) there will be a lot of guesswork involved.

As a side-note, neither pyridine*HCl nor triethylamine*HCl/triethylamine*aqueous HCl demethylated vanillin at all under microwave irradiation.  But then again, if you want to synthesize protocatechualdehyde (isn't this available commercially?) you could always use amine-base/AlCl₃ in ethyl acetate, etc.