Prototocatechualdehyde methylenation. Photo-essay.

What follows is a slight variation of the commonly referenced catechol methylenation procedure, easily found copied and pasted all over the internet. This variation would theoretically produce heliotropin (used in perfumery).
I couldn't find any experimental data to confirm whether this synthesis actually works or not. So lets find out.

Chemicals used:
138g protocatechualdehyde
120 ml 50% NaOH aqueous solution (91g NaOH, 91g dH2O)
500 ml dimethylsulfoxide (DMSO)
120ml (160g) dichloromethane (DCM).

Equipment used:
1L RBF & condenser (for methylenation), stirrer hotplate with oil bath
500 ml RBF (for addition mixture), stirrer mantle
500 ml mixing beaker
3L RBF & heater for steam distillation
Suitable large container for collecting distillate (I used 2L beaker)
Converted steam cooker & aquarium tubing for steam distillation (with short glass tube & cork)
Funnel/cork/glass tubing for reactant addition (though not essential)
1 L separating funnel (for dichloromethane extractions of steam distillate)

No inert atmosphere was used for this experiment.

300 ml of DMSO is added to the 1L RBF, then 120ml dichloromethane is added. Reflux condenser added with cooling water flowing, oil bath heated to 125-130 deg C.

Whilst the bath is heating, the addition mixture is prepared. 120 ml 50% NaOH, 138 g protocatechualdehyde and 200 ml DMSO are combined in a 500 ml beaker (with stir bar), there is some heat evolution and the mixture is stirred on the hot plate for a few minutes to fully dissolve the protocatechualdehyde and any residual NaOH. Once dissolved, the mixture is poured into the 500 ml RBF (rinsed down the funnel with a little extra DMSO), and then transferred to the 500 ml stirrer mantle for heating and stirring. Mixture was kept hot but without boiling. No temperature measurement was used for the addition mixture.

Methylenation solution on left, addition mixture on right.


Once the DMSO/DCM mixture was nicely refluxing (oil bath at round 125 deg C), 15 more minutes were allowed for the addition mixture to stir before addition was started.

Orange addition funnel was connected to a small glass tube, this makes addition much easier, much easier than addition via the condenser.
The drop of addition liquid held in position by capillary force, prevents DCM vapor from exiting the funnel - this works quite well and I can contend this is far easier than using an eyedropper (pipettes tend to melt at this temp), which is very impractical on this scale.


Addition commenced (with vigorous magnetic stirring)


Wearing a rubber glove for heat protection, the hot 500 ml RBF was literally poured into the small addition funnel. A few minutes were allowed in between funnel additions.
This waiting time is necessary because the reaction causes DCM to boil vigorously, it could easily overwhelm the condenser if the additions are not spaced apart (also, according to literature sources, reactant dilution is necessary to improve yields).
In total about 40 minutes were required to complete the addition. The reaction (with vigorous stirring and good reflux) was continued for another 90 minutes, before turning the heat off and allowing to cool (with stirring).

Once cool, the mixture was transferred to a larger container suitable for steam distillation. I had a 3L RBF at hand, so used that. About half a litre of water was added prior to steam distilling.

3L RBF has been set up for distillation (using double surface condenser for good rate of steam condensation), and brought to boiling temperature, and now steam is being added via a small glass tube that is submerged nearly to the bottom of the RBF (though only partly visible in the photo below).



Steam distillate, distilling over quite rich in reaction product:


About 2 gallons of water were put through the reaction mixture (using a pressure cooker as steam source) and condensed.
Once cooled, much of the product crystallises in the water, and is easily removed by filtration. The heliotropin laden water is kept for solvent extraction later. It still contains a few grams per litre.

Glass sintered filter funnel (after filtering a small amount of cooled steam distillate).


2L beaker


Filtering the product


The total dried crystals yield was around 34 grams.
The filtered water was extracted with about 30 ml DCM per litre.

The combined DCM extracts were combined, and added to a 500 ml RBF for distillation (using the large double surface condenser).
Once the DCM was mostly distilled off, the condenser was changed for a small liebig condenser, and set up for vacuum distillation.
I was expecting multiple fractions, however the still head (under full vacuum) rapidly rose to 140 deg C and slowly rose to 143 by the end of the distillation. So I only ended up collecting one single fraction,







Product rapidly crystallising in receiver






Product crystallising in the condenser


Vacuum distillation yielded 26 grams,far more than i was expecting. There was some product still in the steam distilled water when it was discarded (as it still had a strong smell of vanilla/piperonal). If I'd known it held up so much product, I'd have carried out at least 3 x 30 ml DCM extractions per litre.
Also, I stopped the steam distillation when the distillate stopped appearing milky/turbid. However, the last litre, although not milky in appearance, still produced lots of crystals, so I feel I stopped the steam distillation too soon.

The total combined yield was 60 grams, of a product which appears indiscernible (in terms of color, taste, smell, texture and melting point) to another small sample of heliotropin that I've compared it with.
Assuming this product is pure heliotropin/piperonal, then this equates to an overall molar yield of 40 %. Not too bad, but clearly there is still room for improvement.

Attachment: methylenation of catechols.pdf (348kB)
This file has been downloaded 2450 times

[Edited on 30-1-2014 by CycloKnight]

Excellent write-up cycloknight, love the pics. I'll bet your lab reeked of sunshine and rainbows after that!

Apparantly aromatherapists believe heliotropin has anti anxiolytic properties. Did you find your concerns and consternations all blissfully washed away in the wake of your experiment!?

Edit: How in gods name is protocatechualdehyde pronounced?

[Edited on 31-1-2014 by zig]

Yes, sunshine and rainbows indeed, interesting medicine this.

A note regarding this methylenation I should add, the reaction is rather sensitive to reagent temperatures and concentrations.
A previous run wouldn't reach the 125 deg C bath temperature without the refluxing DCM overwhelming the condenser, so I left it at that. Turns out I'd added a little more DCM than needed (I was trying to compensate for some DCM vapor that had boiled away by adding a little extra). The reaction was a complete failure, virtually nothing steam distilled from it, so worth keeping in mind that the stated reaction temperature must be attained.

Quote: Originally posted by stoichiometric_steve

Using CH2I2 and doing this reaction with hydroxyde and a phase transfer catalyst in a two-phase system will probably even yield much higher.

Quote: Originally posted by stoichiometric_steve

Using CH2I2 and doing this reaction with hydroxyde and a phase transfer catalyst in a two-phase system will probably even yield much higher.

Quote: Originally posted by Electra

Not only would the yield be higher, but the reaction would occur much more rapidly

Quote: Originally posted by zig

Quote: Originally posted by stoichiometric_steve   Using CH2I2 and doing this reaction with hydroxyde and a phase transfer catalyst in a two-phase system will probably even yield much higher. Quote: Originally posted by Electra   Not only would the yield be higher, but the reaction would occur much more rapidly Why is that? Got any good links on the subject? As far as choosing a PTC, would one be looking for something like a crown ether (15-5 for Na+, if memory serves) or a water soluble salt? I guess I'm unsure of which reactant needs to be ferried where.

Thanks for the peer-review.

Do you think this will work on 4-allylcatechol (obtained by demethylation of eugenol) ?

I am currently leaning more towards KF-catalysed methylenation in DMF merely because the method in newer, but the DMSO approach might have advantages.

I will try both methods but now I am working on eugenol demethylation via pyridine.HCl to obtain the catechol and all I got is awful sticky sludge, but at least it looks demethylated as the smell of eugenol is gone.

Attaching the picture of demethylated product, crashed from toluene, supposedly 4-allylcatechol loaded with impurities. I am afraid this has to be purified before attempting methylenation.



[Edited on 6-9-2018 by nimgoldman]

Quote: Originally posted by CycloKnight

Now that I have some time put aside for this, I'll be revisiting this process very shortly, and for purification the plan is to try crystallising in petroleum ether before taking it any further.

I am now in the process of experimenting with eugenol, also invested huge bulk of time so far - please don't take my motivation away

The Py.HCl demethylation looks promising though I will try direct heat input instead of MW. I read another patent on how to optimize the process. None of the authors mention inert atmosphere. I am new to this kind of reaction so it's just try and fail...

My first attempt was very crude, the eugenol could be impure (should have used freshly distilled) and the ether extract should have been washed properly before evaporation. But polymerisation is most probably the cause of resulting gunk and avoiding air and moisture is always a good idea. Are there other possible factors for polymerization? Time? Heat?

I have read about polymerization risks only regarding to methylenation, not regarding to o-demethylation of methyl aryl ethers, such as catechol, allylcatechol etc.

Quote: Originally posted by CycloKnight

and for purification the plan is to try crystallising in petroleum ether before taking it any further. I've prepared a quantity of elemental bromine, and am keen to carry out a few millimole scale trial runs using an aluminium bromide complex that is effective at demethylating vanillin to protocatechualdehyde. It'll likely be a waste of time on eugenol, but maybe worth a try anyway. If its a total failure, even at higher dilution, then ho hum, at least I can cross it off the to-do list.

Piperine in acetone sounds like a nifty approach to piperonal, I have a number of tesla coils I've built, 2 of which (solid state dual resonant tesla coils, computer controlled) generate half a million volts - I'd imagine any of them could be adapted to power one heck of an ozone generator.

Can't go wrong using pure eugenol, I vac distilled mine and currently have about 1.5L vac distilled and a few hundred ml of sodium eugenolate purified oil. Obviously, that was made by reacting vac distilled eugenol with NaOH, solvent washed before reverting it back to the oil. The aim was to eliminate impurities that might be causing problems, but actually I never noticed any difference between that and the vacuum distilled oil, so never used much of it.

As for the polymerisation, naturally the problem is worse with increased concentration. I recall higher reaction temperatures generating tar faster but there was too much eugenol being used, so it was more concentrated that it should've been. On one occasion, after removing the oil after an attempted 48% HBr demethylation run, I observed the whole lot appearing to polymerise into a gooey mass in a matter of second, before I could get it to the methylenation apparatus.

The final product I made did have a strong smoky odour (as its often described) and appeared to quickly degrade when not kept in the freezer. I didn't isolate the pure product, but the mixture was very viscous at room temperature. I recall having problems separating the product from the impurities, so later on I acquired the petroleum ether since its described as a recrystallisation solvent. Something new to try soon.
I tried separating the final product by vacuum distilling, and collected the higher boiling point fraction and assumed that to be the product (mixed with a proportion of eugenol still) but it wasn't pure enough to crystallise.

------------------------------------------------------------------------------
Physical properties of 4-allylcatechol
Alternate names: 3,4-dihydroxyallylbenzene, 4-allylpyrocatechol, 2-Hydroxychavicol.
Melting point: 46-48.5°C
Boiling point: 123-124°C/1mmHg, 141-144°C/7mmHg, 147-149°C/10mmHg, 155-157°C/13mmHg, 156-158°C/16mmHg. 289.2/760 mmHg.
Solubility (H2O): 25 g/l at 26°C.
Recrystallization solvents: Benzene, petroleum ether, diethyl ether.
Density: 1.148 g/cc
Flash point: 141.7 deg C.
Molecular weight: 150.177 amu.
------------------------------------------------------------------------------


These are images from those experiments back in 2014.

48% HBr/eugenol after a several hour demethylation attempt.


Final product mixture


Same sample the next day if left at ambient temperature

Quote: Originally posted by zed

I have read that de-methylation of Eugenol by HX, in whatever form, is not easy. But, NaOH may be more effective. Excepting that, NaOH will probably isomer-ise the Allyl side chain to Propenyl. The demethylation of Vanillin by HX is generally more successful, but easier still is that of Ethyl Vanillin by H2SO4. Sadly, for this synthetic pathway, the world provides an abundance of Eugenol, whilst I have never heard of an Ethyl vs Methyl, version of Eugenol. But, I am going to check on it. OK, I checked around. If it exists, it is not an easy find. [Edited on 14-9-2018 by zed] [Edited on 14-9-2018 by zed]

Quote: Originally posted by Loptr

Quote: Originally posted by zed   I have read that de-methylation of Eugenol by HX, in whatever form, is not easy. But, NaOH may be more effective. Excepting that, NaOH will probably isomer-ise the Allyl side chain to Propenyl. The demethylation of Vanillin by HX is generally more successful, but easier still is that of Ethyl Vanillin by H2SO4. Sadly, for this synthetic pathway, the world provides an abundance of Eugenol, whilst I have never heard of an Ethyl vs Methyl, version of Eugenol. But, I am going to check on it. OK, I checked around. If it exists, it is not an easy find. [Edited on 14-9-2018 by zed] [Edited on 14-9-2018 by zed] Some might be interested in that isomerized compound.