Sciencemadness Discussion Board

Demethylation of Vanillin and Eugenol

CycloKnight - 18-9-2018 at 03:25

Further to recent discussions in the Protocatechualdehyde Methylenation thread, I thought it pertinent to start a separate thread specifically for demethylation discussion. All feedback and contributions welcome.

I'll start with yesterday's vanillin demethylation to protocatechualdehyde attempt. Based on US patent 2975214, see attached file.
This uses bromine and aluminum to form a complex with the aromatic solvent (toluene or xylene work well, among certain others). Said complex is used to to demethylate the aryl ether (in this case vanillin), in situ.
Because elemental bromine is being used, overhead forced ventilation (or equivalent) is essential. I'm using a diy fume hood.

The reagent ratios used here are the same as per the patent, only rescaled to 0.0083. So instead of 600g bromine, this will use 5g bromine. All other reagents scaled accordingly.

General setup:

0.56 g Aluminum

5 g Bromine

Ice bath under RBF. Bromine addition to 25 ml toluene with 0.56 g aluminum already added, with good stirring. RBF glass stopper is loosened for pressure equalization during addition.
Temperature maintained at around 20 deg C. Whilst the reaction is exothermic, it was still very easy to maintain the temperature on this scale.
Addition was complete in about 2 minutes, then stirring was continued for another 50 minutes (until heat ceased to be evolved) to form the Al Br ArH complex with the toluene (complex is detailed in the patent). Xylene works fine also if toluene isn't available.
Patent doesn't mention reaction times, but I've found through experiment that its important to get the Al reacted before continuing with the next step. The finer the aluminum the better. Reaction rate is controlled by Br addition rate.

1.26 g vanillin crystals

Vanillin addition. Vanillin would not dissolve in 10 ml toluene, so ended up having to wash with extra toluene. In hindsight, I should have just added the vanillin in powder form, one scoop at a time to the RBF. Fumes aren't a problem at this stage, just a little HBr fuming. During addition temperature is maintained between 15 and 20 deg C. Some heat is released during this step, but it was easy to to control the temperature.

As per the patent, this mixture was maintained between 15 and 20 deg C, stirred for 1 hour. This is the demethylation reaction, methyl bromide is evolved.

Then afterwards heating applied and gradually brought up to about 90 deg C (patent calls for 100 deg C, but this seems to work). Then was stirred for another 20 minutes or so at 90 C, then allowed to cool.

Reaction complete and cooled to ambient temperature. Ready for workup.

Reaction mix is poured into a beaker with 83g ice and 2ml of 35% HCl (aq), with good stirring.

The purple mix soon changed to light amber when combined with the water. This due to the AlBr2 (attached to where the methyl used to be) decomposing in contact with the water, and then forming the protocatechualdehyde.

I'm not clear on why the patent calls for so much ice, I had to gently apply heat to melt the ice, not particularly exothermic.

Separate the two phases, naturally the toluene layer is on top.

Aqueous phase is extracted with dichloromethane (though the patent calls for diethyl ether, and then to evaporate the ether to yield oily protocatchualdehye)

The organic phase (toluene) is then extracted with 3 portions of 9 ml 5% NaOH solution. This pulls the majority of product.

Then those 3 extracts are combined, and acidified by adding several ml of 35% HCl, the solution color changes to a light brown solution as the protocatchualdehyde precipitates out. This was then covered with cling film and put in the freezer overnight.

Next morning.

0.55 grams crude protocatechualdehyde. This does not include the DCM extractions, since they haven't been worked up yet.

In conclusion, compared with the eugenol variation of this process (which I've already attempted nearly a dozen times on this same scale), this is a walk in the park. The patent appears to work, as claimed.
No emulsions, the phases separate immediately without any fuss, and altogether quite friendly. I'm sure this could easily be scaled up to any size within reason, the main downside obviously is it involves working with bromine.

Additionally, according to the patent most of the inorganic bromine can be recovered from the aqueous phase by acidifying and either oxidizing or gassing with chlorine. The patent claims that between the methyl bromide and inorganic bromide, the total recoverable bromine is 88-92% per cycle.

Attachment: Process for manufacture of protocatechuic aldehyde US2975214.pdf (394kB)
This file has been downloaded 868 times

[Edited on 18-9-2018 by CycloKnight]

DJF90 - 18-9-2018 at 05:03

Nice to see you working on a sensible scale. You say the reaction was complete but provide no details as to how this was determined (TLC?)? You also provide no data to support your claim that the isolated solids were indeed the desired product, and that it was clean (melting point, TLC)?

I don't mean to demotivate, but merely to stimulate the use of good scientific principles.

CycloKnight - 18-9-2018 at 05:15

As for the eugenol variation, the reaction proceeds rather similarly until the workup, except the eugenol addition is much easier without the solids to contend with.
However, the aqueous/organic phase separation is very slow, there is some tar & aluminum sludge that sticks to the separating funnel walls which is a nuisance.

For the product extraction from the organic phase, when using 5% NaOH as for protocatechualdehyde, the product instantly turns to dark red sticky goop. Irreversibly so. Using K2CO3 instead, avoids the red goop, but final yields are nowhere near as high as with vanillin (protocatechualdehyde) I still haven't yet established whether its the product that's polymerizing with the NaOH, or whether its an impurity that's complicating this step. Steam distillation perhaps?

The extracted product might be the target hyroxychavicol (4-allylcatechol) but haven't positively identified it, and may not be able to isolate enough without first scaling up somewhat beyond 5 g bromine. One could only imagine an allyl protecting group for the eugenol making life a whole lot easier.

[Edited on 18-9-2018 by CycloKnight]

morganbw - 18-9-2018 at 06:47

Nice write-up.
Thanks for posting this up.

CycloKnight - 18-9-2018 at 09:32

Quote: Originally posted by DJF90  
You say the reaction was complete but provide no details as to how this was determined (TLC?)? You also provide no data to support your claim that the isolated solids were indeed the desired product, and that it was clean (melting point, TLC)?

No definitive claims at this stage, this is just a trial run for the scaled up runs later. I can't say whether the product is pure or not. No doubt there remains an element of faith still, however as the thread progresses we will find out whether the product is the target compound or not. All a work in progress.

JJay - 18-9-2018 at 13:21

Pretty cool. I've been thinking about giving this a try for making dopamine. I wonder about the effect of increasing the amount of aluminum without increasing the amount of bromine.

myr - 18-9-2018 at 15:02

Have you seen the paper for AlI3/tetrabutylammonium? Xylene or toluene should substitute nicely for benzene.

Synthesis 1985; 1985(4): 437-439
DOI: 10.1055/s-1985-31235

clearly_not_atara - 18-9-2018 at 17:37


For the product extraction from the organic phase, when using 5% NaOH as for protocatechualdehyde, the product instantly turns to dark red sticky goop. Irreversibly so. Using K2CO3 instead, avoids the red goop, but final yields are nowhere near as high as with vanillin (protocatechualdehyde) I still haven't yet established whether its the product that's polymerizing with the NaOH, or whether its an impurity that's complicating this step. Steam distillation perhaps?

The extracted product might be the target hyroxychavicol (4-allylcatechol) but haven't positively identified it, and may not be able to isolate enough without first scaling up somewhat beyond 5 g bromine. One could only imagine an allyl protecting group for the eugenol making life a whole lot easier.

Polymerization seems likely, the eugenol/AlCl3 reaction is notoriously difficult to work-up or perform, with tars being the major product of most attempts. A different strategy must be chosen for the extraction of the final product.

Protocatechualdehyde should polymerize less because the aldehyde draws electron density from the ring and adds a quinone-like enolized resonance structure. However this same resonance structure makes methylenation of protocatechualdehyde more difficult than other catechols.

Using K2CO3 instead, avoids the red goop, but final yields are nowhere near as high as with vanillin (protocatechualdehyde)

For catechol, pKa1 = 9.45; pKa2 = 12.8. If we assume that the dianion of catechol is what causes the trouble -- because it is easily oxidized to a semiquinone radical -- then our goal should be to deprotonate the catechol one time, so it becomes water soluble, but not two times, to prevent oxidation and polymerization. Potassium carbonate does exactly that: carbonic acid's pKa2 is 10.3, which is enough to deprotonate catechol once but not twice.

The problem with K2CO3 is that it tends to reduce the solubility of organic compounds in water:

One possibility is instead to use milk of magnesia, which also has a pH of around 10.5, but which does not tend to salt-out organic compounds. The obvious disadvantage is that Mg(OH)2 is a heterogeneous base, so you cannot "wash" a solid residue with a magnesium hydroxide suspension.

Another possibility is to use an amine base, such as pyrrolidine (pKaH+ = 11.3), isopropylamine (pKaH+ = 10.7), diisopropylamine (pKaH+ = 11), methylamine (pKaH+ = 10.6), with ammonia being not preferable due to the negative effect of NH4+ on solubility. Which amine I leave up to you, but most small aliphatic amines qualify; aromatic amines tend to be weaker and less soluble, so they are not suitable.

Still another possibility is to use lithium or sodium carbonate, which have a slightly weaker salting-out effect than K2CO3.

Propenyl Guaethol - 19-9-2018 at 07:57

Check out the workup in this patent.

clearly_not_atara - 19-9-2018 at 08:07

"The addition of aluminum iodide (2.242g, 5.5mmol) to a IOOml eggplant-shaped flask, acetonitrile (40ml), potassium tert-butoxide (1.237g, ll.Ommol) and eugenol (0.819g, 5.0mmol) and heated to 80 ° C, stirring was stopped after 18 hours, cool to room temperature acidified by addition of 2mol / L dilute hydrochloric acid (IOml) eggplant type flask, and extracted with ethyl acetate (50ml X 3), the combined organic phases, washed first with saturated aqueous sodium thiosulfate (IOml), then saturated brine (IOml), dried over anhydrous magnesium sulfate, filtered, and the filtrate evaporated on a rotary evaporator, the residue was purified by flash column chromatography (eluent ethyl acetate / petroleum ether = 1: 4, volume ratio) to give 0.733 g crude 4-allyl-catechol, 4-allyl-catechol take crude (0.709 g) under reduced pressure sublimation oil Li to afford 4-allyl catechol 〇.681g (white waxy solid, 93% yield)."
There are methods on the forum for synthesizing acetonitrile and metal alkoxides, but hopefully you can buy them.

[Edited on 19-9-2018 by clearly_not_atara]

CycloKnight - 19-9-2018 at 11:40

clearly_not_atara All very helpful, thanks. I hadn't properly considered that the K2CO3 might have been hindering the extraction.
Whilst there was a large amount of red tar with the NaOH extraction(s), it was far more than the total recovered product when using K2CO3 instead. The numbers didn't add up, that was why I was suspicious that possibly another impurity was involved with forming the tar.

I might revisit the eugenol variation in the next few days, run on a larger scale, sampling the product mixture and applying different work up strategies. Can't do any harm.

Yesterday I ran another vanillin demethylation attempt based on 50 g bromine, 12.6 g vanillin. I ran this one hotter towards the end and then accidentally overheated to about 120 deg C for a short while, but during the workup today I didn't note any obvious signs of degradation/tar. Solution is sparkling with crystals now, should find out the yield tomorrow.

Myr - I haven't read it yet but will check it out later when i get a chance, thanks for the suggestion.
Propenyl Guaethol, good find, the workup procedure may come in handy later.

Magpie - 19-9-2018 at 13:02

Might the OrgSyn preparation of catechol from guaiacol be useful here?

JJay - 19-9-2018 at 13:33

Quote: Originally posted by myr  
Have you seen the paper for AlI3/tetrabutylammonium? Xylene or toluene should substitute nicely for benzene.

Synthesis 1985; 1985(4): 437-439
DOI: 10.1055/s-1985-31235

They used almost double the stoichiometric amount of aluminum. Using some extra aluminum sure beats halogenating the solvent....

clearly_not_atara - 19-9-2018 at 18:57

Quote: Originally posted by Magpie  
Might the OrgSyn preparation of catechol from guaiacol be useful here?
I think you might get 4-(2-bromopropyl)-catechol and/or a polymer. The bifunctional phenol/bromoalkane might polymerize upon basic workup.

For vanillin, HBr is known to work.

CycloKnight - 20-9-2018 at 02:00

One of my acidified aqueous solution extracts from the eugenol trials (5g bromine/0.56 g Al/1.36 g eugenol) was put in the corner of my work area several days ago with no lid and forgotten about.

Quite distinctive needles.
I'll put this aside for any future testing.

Magpie - 20-9-2018 at 09:10

Quote: Originally posted by clearly_not_atara  

For vanillin, HBr is known to work.

The guaiacol to catechol synthesis uses HBr, not Br2.

clearly_not_atara - 20-9-2018 at 13:57

I never said otherwise. 4-(2-bromopropyl)-catechol is the product of adding HBr to the double bond in eugenol. Whether the intermediate carbocation (if formed) would undergo other reactions is anyone's guess.

JJay - 20-9-2018 at 14:42

Quote: Originally posted by clearly_not_atara  
I never said otherwise. 4-(2-bromopropyl)-catechol is the product of adding HBr to the double bond in eugenol. Whether the intermediate carbocation (if formed) would undergo other reactions is anyone's guess.

Oh yeah... you could bring back a double bond with KOH I guess, but it might be in a different spot.

S.C. Wack - 20-9-2018 at 16:55

I wonder what Pearl and Beyer of the JACS article thought about the patent they're mentioned by...maybe their AlBr3 was made by them that same way, just didn't explicitly say so...

A solution of 15.2 g. (0.1 mole) of vanillin in 45 ml. of nitrobenzene at 15° was treated with a solution of 53.4 g. (0.2 mole) of anhydrous aluminum bromide in 60 ml. of nitrobenzene at 10°. The gel which formed was stirred with 125 ml. of nitrobenzene, warmed on the steam-bath to 95°, and allowed to stand at room temperature for 30 minutes. The dark mixture was cooled and poured into 1 L of water containing a little hydrochloric acid. This mixture was extracted with ether, and the ether was extracted with 5% NaOH solution. The alkaline solution was washed with ether, acidified with dilute sulfuric acid, and extracted with ether. The ether was dried and distilled, leaving 12.8 g. (93%) of oily protocatechualdehyde which crystallized on cooling. Recrystallization from toluene gave colorless needles melting at 151-152° and not depressing a mixed melting point with authentic protocatechualdehyde.

JACS 75, 2630 (1953)

nimgoldman - 20-9-2018 at 19:24

I am trying eugenol demethylation using pyridine hydrochloride (Py.HCl).

This article studies demethylation rate at constant temperature, depending on time:

Schmid, Christopher R., et al. "Demethylation of 4-methoxyphenylbutyric acid using molten pyridinium hydrochloride on multikilogram scale." Organic process research & development 8.4 (2004): 670-673.

One problem with Py.HCl demethylation is large excess of it is needed and possibly a PTC. The above paper, however, shows that almost complete demethylation is possible in 5 hours at 190 °C with just 1.5 molar excess of Py.HCl and 1-2 hours with 4 molar excess.

They measured 4-Methoxyphenylbutyric Acid but I guess any methyl aryl ether will work.

I am waiting for an argon pressure regulator and then prepare Py.HCl, distill my eugenol and conduct the demethylation under inert atmosphere in a stoppered flask.

The reaction will be quenched with cold water and extracted with ether. The ether extract will be washed several times with water, dried and evaporated to hopefully yield hydroxychavicol (allylcatechol).

chemplayer... - 22-9-2018 at 20:42

Sexy colours! Nice work. We tried lots of ways to do this with vanillin and in the end the best result was using pyridine hydrochloride (still not a great yield, but might be better if we could get the reagent more dry - it's *insanely* hygroscopic). This method looks worth a try though.

Heptylene - 23-9-2018 at 01:51

Chemplayer! I re-watched your attempt at vanillin demethylation with pyridine HCl. I wonder if the pyridine HCl could be dried with a Dean-Stark setup using toluene. Otherwise, I you could bubble HCl gas in pyridine.

Any new videos coming next?

byko3y - 23-9-2018 at 02:42

Reaction and decomposition temperatures are both much higher than boiling temperature of water, so by the time reaction starts most of the water is gone no matter how much there was on the begginning.

Heptylene - 23-9-2018 at 06:27

It doesn't work like that. You still have to use anhydrous DMF (BP: 153 °C) and DMSO (BP: 189 °C) for reactions that fail in the presence of water. The water doesn't just boil off completely.

Same for ionic liquids (IL). Pyridine hydrochloride could be considered an ionic "liquid", melting at only 145 °C which is quite low for an ionic compound. The definition of an IL is arbitrary and varies a lot. A typical procedure for drying ILs is heating for several days at 100 °C under high vacuum.

I know you said "most" of the water is gone, my point is you cannot consider something dry by just heating it in the open above 100 °C.

chemplayer... - 23-9-2018 at 06:42

Pyridine hydrochloride turns from white crystals into slush in front of your eyes in under 10 seconds! It's the most extreme I've seen. Of course, it's just speculation that it's the water content which caused the low yield, but the toluene solvent and Dean-Stark trap is an excellent idea for a re-try.

nimgoldman - 24-9-2018 at 07:48

Today I prepared some Py.HCl and it seems to be impossible to store it properly. Even in dry bottle with desiccant packets, it clumped and turned into droplets on the walls. Maybe I should have flame dried the bottle beforehand...

Prolonged drying under vacuum is not an option as pyridine.HCl volatilizes easily.

I think the only sensible way is to prepare Py.HCl on demand - fortunately the preparation is easy (bubbling HCl into a cold solution of pyridine in ether, filtering and putting the solids in a vacuum desiccator without applying vacuum or maybe only very slight vacuum - we just want to remove the ether).

I plan to keep distilled eugenol reacting with excess Py.HCl at 190 °C for 3 hours under argon.

I think the demethylated product (hydroxychavicol) can be distilled under vacuum - is it possible?

I have no idea how hydroxychavicol (i.e. allylcatechol) looks like - a crystalline solid?

[Edited on 24-9-2018 by nimgoldman]

clearly_not_atara - 24-9-2018 at 18:37

There's another method I forgot but which is worth considering, particularly since it doesn't smell like pyridine. The attached paper reports a demethylation of eugenol (Page 3742, left side) by extended reflux with lithium chloride in DMF. LiCl is much easier to handle than aluminium halides or PyHCl. I speculate that LiBr may be more effective.

Attachment: kraft2003.pdf (217kB)
This file has been downloaded 527 times

[Edited on 25-9-2018 by clearly_not_atara]

nimgoldman - 25-9-2018 at 01:42

Here is that synthesis step for conveniece:

LiCl (292 g, 6.89 mol) was
added to a solution of eugenol (354 mL, 2.30 mmol) in DMF
(3.7 L), and the mixture was refluxed for 44 h, with additional portions of LiCl (292 g, 6.89 mol) being added after 4 h, 22 h and 29 h.
The reaction mixture was allowed to cool down to room temp.,
and diluted with toluene (2 L). The formed precipitate was filtered
off and washed with toluene, the washings were combined with the
organic solution and concentrated in a rotary evaporator. Silicagel FC (Et2O/pentane, 1:1, Rf = 0.37) provided 4-allylpyrocatechol
(173 g, 50%)

50% yield is better than the procedure I found.

I will try the AlCl3/DMF approach after the Py.HCl.

My pyridine hydrochloride sort of caked overnight (see photos) and there are droplets of the liquefied stuff on the glass. The silicagel packet look soaking wet.

I don't think this is all water, the Py.HCl more likely just melted in a presence of minuscule amount of H2O.

It however stays white and fluffy in a desiccator (it seems to be then filled with pyridine vapour) but as long as exposed to the air, it quickly evaporates and liquefies before your eyes.

DSC_0808.JPG - 130kBDSC_0813.JPG - 124kB

CycloKnight - 25-9-2018 at 02:57

The LiCl route is one I've considered, I have some LiCl (made from Li2ClO3 and HCl aq) and about 500ml DMF. LiCl is a little on the expensive side, but I see no reason why the litium couldn't be recovered and recycled. I could give this a try in the near future.

I wonder if the more commonly available DMSO would work instead of DMF.

Disappointing results.

CycloKnight - 27-9-2018 at 03:59

I attempted a 13g eugenol demethylation (slight deficiency compared with the vanillin patent). Reaction seemed to run very well, instead of pyro grade aluminum I used cut up pie tins and then superficially activated with copper. This just involved a short NaOH solution wash then a several minute dilute CuSO4 solution soak, and finally oven dried. Everything was dried well before the reaction.

Reaction was quite vigorous, when the eugenol addition was begun, the ice cooled reaction mix temperature increased from 10 C to 20 C in about 15 seconds, took 55 minutes to complete the addition (maintaining below 20 C).
The reaction was run according to the patent, including an extra hour at 100 C before cooling down.

Interestingly, the final mix turned into milkshake during the complex decomposition with the ice/water. After external heating was applied (I estimate above 50 C or so), the viscosity rapidly dropped off and the mixture settled and the two phases quickly separated.

By morning there was just a clear water phase and organic phase, with a thin emulsion later in the middle.

For the workup, I just steam distilled the lot (could have separated the emulsion layer, but I included that too). The purpose of this was to test for compatibility. There was only a little foaming. All foaming disappeared completely after most of the toluene had been driven off.
A lot of oil came over too, some solidifying in the condenser and receiver (hydroxychavicol mp is ~49 C). It was looking very promising.

The steam distillate (about 2.5 litres) contained about 200ml of toluene in the first litre. I used that to extract the rest of the distillate later on, with another 100ml makeup toluene also.

Working under very non-ideal circumstances, since lacking much analytical data, one has to make some best guess judgements based on limited info available.
That said, whilst I'm confident there was some hydroxychavicol produced the amount "seemed" to diminish later on with the workup, only to be replaced with something else. A fragrant chemical that smells like a lemon/spring cleaning product, a lot like urinal cakes/tabs or floor cleaner. I don't know what it is. It may have just become more noticeable later in the workup as the smoke hydroxychavicol aroma (once dominant) seemed to become less and less. All very subjective.

I'd imagine the o-quinone to have a disagreeable odor, whereas this product isn't disagreeable at all (also reminds me a little of methyl benzoate aroma). Whatever it is, the eugenol appears to have been largely converted to it one way or another.

The aqueous steam distillate (after workup) tested very alkaline on litmus paper, I was expecting it to be acidic (given the post-reaction mix is acidic). Not sure if that's a relevant detail or not.

Before beginning the steam distillation, I took a small sample from the organic phase, and added it to 4% NaOH solution, and it appeared to polymerize immediately.
Organic phase sample is on the right.

A few drops of the organic phase into the alkaline yields this.

Forming a sticky mess that sticks to the sides of the container, but is easily cleaned with isopropyl alcohol.

The final majority oil product after steam distillation does not polymerise with NaOH solution (though a small amount reacts, it reversible with acidification), so it seems logical to assume something has changed.

On the next run, I'll get samples at varying stages. The purpose being to discern if the product is indeed rearranging or degrading, because if it is then that potentially could affect all synthesis/workup attempts not just this approach. Prudent to rule that out if possible.

[Edited on 27-9-2018 by CycloKnight]

CycloKnight - 27-9-2018 at 08:58

After cleaning up all the glassware, work area smells like a freshly cleaned public restroom. Can almost hear the janitor's trolley squeaking in the distance. That lemon/naphthalene fragrance isn't quite the one I was aiming for.
Any suggestions before the next run are most welcome, running out of ideas as far as this particular aluminum bromide approach goes. I was considering using a eugenol excess next time, to perhaps rule out the possibility of a secondary reaction occurring to the demethylated product (assuming it existed). Another idea is to ease up on the reaction conditions a bit, maybe the reaction temp at 100 C for an hour is a little too aggressive for eugenol? I only did the toluene workup (K2CO3) on 100ml of the steam distillation toluene extracts, still have 200ml toluene extracts that I plan to just vacuum distill and collect the individual fraction(s).


CycloKnight - 30-9-2018 at 12:22

Further from the last run, the workup on the remaining 200 ml of the toluene extractions and vacuum distillation has mostly been completed.
Quite unexpected results.

Toluene distilled off using vigreux column.

Vacuum distilled (smaller vigreux column used), vacuum was consistent throughout the entire vacuum distillation.

3 fractions collected.
Residual toluene, came off at 20 C.

Second fraction, 30 - 50 C. Most came over at 49 C. This is the fragrant compound described in my last post. Yield = 28.62g.

I can only assume this has been formed in some way from the toluene during the demethylation. Any guesses what this compound is? Its the main side product so understanding how it's formed might help improve this method later.

Third fraction, 110 - 144 C. Yield = 3.85 g. Plus another few grams solidified in the column and condenser.

Second fraction on the left. Yet to be identified.

Third fraction, with a little carry over from the last.

Third fraction crystallizing in the apparatus.

Isopropyl alcohol solution evaporating in a pyrex dish. Product forms crystalline needles. You can see from the crystal mass on the edge that the solution is fully saturated, a good few grams in there. This is only the column and condenser washings, it does not include any of the 3.85 grams in the final fraction.

The waxy crystalline solid in the pyrex dish appeared to collect some water under the waxy surface (I assumed from the cooling effect from the fan blowing on the alcohol solution pulling humidity from the air), so I applied gentle heat on the hotplate and the crystals melted at around the hyrdoxychavicol melting point (49 C), and color began to slowly change to a browner shade. Not a particularly stable compound. Didn't get a photo unfortunately.

Therefore I was unable to isolate all of the pure solid for a yield measurement.
The product was reacted with potassium carbonate solution to form the salt, to try and prevent any further decomposition. It reacts instantly and the solution turns pink, but when acidified it turns clear again and the solidified oil forms on the surface. I don't have any reason (yet) to conclude any significant decomposition has taken place.

Potassium salt/catecholate solution

This appears to be the target compound, 4-allylcatechol (hydroxychavicol).
No final yield yet, this is only about half to 2/3rds of the actual yield, since I took a large sample of the toluene steam distillation extracts for workup testing (not included in these results). However, considering that the steam distillation wasn't quite completed since it was only desired to have enough product for testing and for product fraction isolation, the yields could've been somewhat higher.

Will be isolating the potassium salt in the next few days or so. I'm hoping this compound is stable in the salt form, will make things a lot easier.
Still early days yet, but this now looks to be headed in the right direction at least.

[Edited on 1-10-2018 by CycloKnight]

S.C. Wack - 30-9-2018 at 14:29

BTW there are claims from the betel-growing areas of over 25% (really) yield of AC/APC/HC from dried and ground betel leaf, with alcoholic extracting...lots of interest lately except for methylenating.

CycloKnight - 30-9-2018 at 18:13

Recovered salt from the first of two extractions.

There's a good tablespoon of crude chunky potassium catecholate there, haven't weighed it yet. Will weigh it all together when the rest of the work up is complete. So far it isn't looking too bad from a mere 13 g eugenol, all considering! Lots of refinements in the pipeline.

Oven dried it at 60 C. Seems stable, no obvious signs of decomposition so far. I don't know what the other extraction contains yet.

As an aside, I've been unable to find much useful data (solubility, solvent compatibility, etc) for this salt nor the hydroxychavicol, and it's not from lack of trying. So I'll determine and provide it. Tests planned, consider this yield expendable for said purpose. No doubt the data will come in useful for formulating an efficient workup.
Any requests, let me know.

I neglected to mention earlier that the pink catecholate solution also contained about 10 ml of isopropyl alcohol. When the alcohol was evaporated off it changed from pink to a light brown/beige color.

[Edited on 1-10-2018 by CycloKnight]

CycloKnight - 1-10-2018 at 09:30

A few words about product purity, I can't offer any guarantees as it hasn't been thoroughly analysed yet. However, I'm certain at least there isn't any discernible quantity of eugenol in the workup. I've completed this reaction about 15 times (in varying renditions) over 20 days, with experience its quite easy to tell what's eugenol and what isn't. The final product reacts like a catechol should, it decomposes rapidly when melted and exposed to air, and it melts at around 49 C. Can do an accurate mp test later. Crystallizes in needles and has a waxy appearance (as hydroxychavicol does). It smells like diesel and tastes like soap, not easily confused with an obvious workup contaminant such as eugenol.

Today I've been busy trying to wrap up the second extraction. Some product appears to be trapped in about 20 ml of toluene. Rather than waste any more time on it, I might just acidify, rinse, and throw it in with the next vacuum distillation (or vac distill on its own). Either way, there's a small amount I've recovered (still evaporating now), but nowhere as much as the first extraction. So the total second extraction yield is as yet still undetermined, and might remain so.

The first extraction potassium catecholate dried yield is just over 4 g. As that is only from as much as, but no more than 2/3rd of the total steam distillation toluene extractions, then the total yield would extrapolate to at least 5.3 g of catecholate. Add on to that any yield from the second extraction if and when it materializes.

The best workup strategy so far seems to be to run the demethylation reaction, then after the water/ice/HCl aluminum bromide decomposition step, save the organic phase, separate the aqueous layer and extract with toluene.
(Save the aqueous layer for later bromide recovery.)
Steam distill the organic phase (also throw in the toluene from the aqueous phase extraction). Product is the last fraction to come over, so put a lot of steam through it.
Use the toluene that steam distilled in the first litre or so, to back extract the rest of the steam condensate, also extract with extra toluene to get the rest of the product out. Can also add salt to the condensate prior to toluene extractions.

Distill off the toluene using a short column and vacuum distill what remains. The last fraction is the final product, it is then combined with water and isopropyl alcohol and basified with K2CO3 solution. Evaporate to dryness, and then you have your potassium catecholate for whatever purpose you had in mind.

HCl acidified sample

CycloKnight - 1-10-2018 at 12:35

This is about a gram of assumed potassium catecholate, several drops of HCl were added directly without any solvent.
There was some effervescence, indicating some residual K2CO3 contamination with my final yields.

The (assumed) catechol is released as a waxy oil immediately and solidifies on cooling.
Afterwards it was all dissolved by adding a few ml of isopropyl alcohol and then K2CO3 solution added to convert it back to the salt, it was simply dripped in until the effervescence ceased. The catechol is extremely soluble in isopropyl alcohol.
The solution will then be evaporated to recover the salt along with KCl contamination, but can always clean it up later.

S.C. Wack - 1-10-2018 at 15:48

BTW Perkin and Trikojus and Schöpf (hive post 507775, UTFSE) note a green color with ferric chloride and allylcatechol.

I suspect one might want to extract such a diol from an aqueous phase with something a bit more polar than toluene.

myr - 1-10-2018 at 16:25

I would recommend extraction with ethyl acetate- worked very nicely for 5-hydroxyvanillin, might work for any substituted catechol.

(The paper I listed earlier calls for extracting the benzene with water, then washing the water with ethyl acetate, fyi)

CycloKnight - 1-10-2018 at 16:27

Thanks for the suggestion, I'll try that test after the next run. I had a mishap a short while ago. Whilst trying to remove the excess K2CO3 all at once, the product became degraded during the workup. Oops, or something like that.

CycloKnight - 9-10-2018 at 05:35

FeCl3 solution test of assumed hydroxychavicol.

Eugenol yields a yellow-orange color.

Potassium carbonate solutions have proved nearly useless for extracting the product from toluene. It extracts, but only a small percentage per separating funnel extraction using 5% K2CO3 aq. solution. Not proving to be a viable workup strategy unfortunately.

I did manage to isolate some of the product from the completed reaction mixture using solvents and vacuum filtration only.

I removed about 4 ml at round 80 C, evaporated most of the toluene on a glass sample dish, then added a >15 ml of water to decompose the Al Br2 aryl complex. While stirring on the hotplate, several ml of isopropyl alcohol was added, then it was basified with K2CO3. The precipitate was vac filtered (with difficulty as the product was very fine). The gray clay-like precipitate (assumed to contain a lot of aluminum hydroxide) was acidified with 36% HCl (aq) and a little water and extracted with ethyl acetate. 3 extractions of a few ml ethyl acetate removed all of the detectable hydroxychavicol product (using FeCl3 indicator test). The golden yellow ethyl acetate combined extracts were then carefully evaporated in a desiccator with full vacuum applied. Yielding a golden yellow viscous/waxy oil.

The filtrate from the previous filtration step also tested green for some (assumed) hydroxychavicol product also. But, at least this is one way to isolate some product (albeit with impurities) from the reaction mix, without having to steam distill nor vacuum distill.

[Edited on 9-10-2018 by CycloKnight]

myr - 9-10-2018 at 07:35

Do you have the capacity to carry out TLC plates?

CycloKnight - 9-10-2018 at 10:43

Not difficult to do but I don't have any TLC plates I'm afraid. Its on my list of things still to acquire.
I'm currently building a small standalone recirculating steam distillation and decanting rig that I think might extract the product with minimal hassle. Aiming to have it finished soon. I'm running a little short on extra clamps and fittings so having to improvise a bit.
Steam distillation is the only extraction method I've tried that's produced an extract free of the tar and discoloration, one which can be reliably vac distilled to yield a pure white product that crystallizes.

Once I have a larger pure quantity without the contamination, I imagine it should be a lot easier to positively identify.

CycloKnight - 21-11-2018 at 09:52

Finally have some results to update on.

The fragrant compound I described as a major side product of my previous demethylation attempts using the Al Br method, appears to be none other than bromotoluene (ortho meta para). Same density and boiling point(s) for the isomers. I measured the boiling point mixture ranging between 178 and 184 C, and the density around 1.37 g/ml, which ties in quite closely with that for bromotoluene.

The crystalline substance I isolated before isn't pure 4-allylcatechol, but I assume mostly the dibenzoyl with some 4-allylcatechol. I measured the melting point as being 72 deg C. The reference I have for the dibenzoyl, mentions the melting point as being 72 deg C.

This would explain why the crystals were turning brown when heated to the melting point as I mentioned previously, but it wasn't the main crystalline product that was degrading. The pure dibenzoyl (assumed based on melting point) I've isolated can be melted without obvious signs of decomposition.

From the last steam distillation.
3L beaker as steam condensate receiver, covered in crystals.


Crystals in FeCl3 solution, an impurity in the crystals is giving green indication, but not the bulk of the crystal mass.

Ethyl acetate containing hyroxychavicol (presumably) extract.

About half the above solution was evaporated in a glass dish at 6 C with no heating applied. It was an ethyl acetate extract of an acidifed, 5% NaOH extract of the final fraction from vacuum distillation of the steam distillation extracts.
Turned black in air, though this may have been accelerated due to impurities, or decomposition with the NaOH. The NaOH extract was dark red, and then a light yellow/beige upon acidification .

This is an ethyl acetate extract of the steam condensate that's been lying open in the air for about 4 days now. Exposed surfaces turning black.

Its mostly the product that melts at 72 C (dibenzoyl or otherwise), but clearly there is another product in there that degrading on exposure to air.

I've learned that the Al Br demethylation reaction can be run far more concentrated for eugenol than for vanillin. Presumably the dilution for the vanillin demethylation is due to the relatively poor solubility of vanillin in the aromatic solvents referenced in the patents. The last demethylation reaction used 100ml solvent, 26g eugenol, 12g aluminum and 100g bromine. Overhead stirring was used.

I've tweaked the reaction conditions somewhat from the original vanillin demethylation patent procedure, getting better results now.

CycloKnight - 22-11-2018 at 18:11

Results from latest vac distillation.

I see we now have a new player in town. This is the fraction after the dibenzoyl fraction, coming over at >160 C at 24 mbar abs (18 mmHg).

Started off white but soon discolored. The solids shown here solidified in the condenser, nearly completely blocked it (even though I stopped the water flow) but came out mostly in one piece. Turning dark in air, FeCl3 test yields green.

Crystalline solid yield is just under 12 g (including final fraction receiver), and there's a bit more in the glassware alcohol rinsings. This from a 26 g eugenol demethylation run. Steam distillation was run for 10 hours, but was incomplete. Will directly vac distill the organic phase next time.

I've wrapped in foil and put it all in the freezer for safe keeping.

[Edited on 23-11-2018 by CycloKnight]

CycloKnight - 22-11-2018 at 19:00

Final fraction receiver is on the right. Still crystallizing when the image was taken, soon after completely solidifying.

CycloKnight - 23-11-2018 at 07:47

Final fraction mp: 38 C
Condenser solids mp: 48-51 C (some dibenzoyl crystals remained solid)

In the final fraction, there was a drip of clear oil that separated from the crystal mass at room temperature that I hadn't noticed before, so this sample will most likely need to be refined for a more accurate mp test.

CycloKnight - 23-11-2018 at 07:55

Just a short update on the procedure I'm now using.

A while back I had the idea of running the eugenol demethylation reaction using the main side product, bromotoluene, in place of toluene. I've been doing that and it appears to be working so much better as an improved substitute for toluene/xylene complex solvent for this eugenol demethylation approach. It fulfill's the same function (complex formation) just as toluene does. The last two runs have been with bromotoluene (still using toluene as eugenol solvent for the eugenol addition step)

The heating step is now avoided altogether, so after the eugenol addition is complete, only stirring up to 20 deg C but no higher.

This has been working so well, that after the reaction completion and then the complex decomposition step, there is very little tar. Vac distillation of the organic phase looks to be worth trying.

Steam distillation is acceptable for small quantities, but for say a 50 g eugenol demethylation run, it would be necessary to steam distill for perhaps several days just to get the product out.

The bromotoluene is simple to prepare, the reaction conditions (Al Br + HBr) catalyze its formation, so that any excess bromine reacts with available toluene in solution and yields the bromotoluene, which remains after the complex decomposition step, and can be extracted with toluene, then vacuum distilled to separate.

CycloKnight - 23-11-2018 at 08:24

Complex decomposition step.
Final reaction mix in the 500 ml RBF. About a third already added to the beaker.

Note the color after decomposition, much cleaner than before.

CycloKnight - 24-11-2018 at 07:05

"The work on betel leaf (Piper betle) consisted of
heating the chopped leaves with groundnut oil at 200°/5»6 mm.
for 5 min. and of obtaining the essential oil by solvent
extraction as well as by steam distillation. The essential
oil was sepeur-ated into phenolic and non-phenolic constituents*
The former on careful distillation under vacuum gave solid
and liquid fractions* The solid fraction on recrystallisation
from benzene-petrol ether melted at 48.5°C. and gave a
dibenzoyl derivative melting at 72°C. On further
examination, it was found to be identical to 4-allyl
catechol (hydroxy-chavicol), the presence of which in Java
betel leaf oil had been reported earlier by Schimmel^and
its presence in Indian betel leaf oil has sincee been
confirmed by Dutt"

(page 261 of attached pdf)

Allyl-pyrocatechol, C9H10O2, exists in betel leaf oil. It is a crystalline
body melting at 48° to 49° and boiling at 139° at 4 mm. It yields a
dibenzoyl derivative melting at 71° to 72°. Its constitution is as follows :—

By methylation with dimethyl sulphate and potash, it yields methyleugenol,
boiling at 248° to 249°;


Could use a little help. I'm looking for information on this dibenzoyl derivative, its probable structure, etc, and if it can be easily reverted to the 4-allylcatechol. Any feedback appreciated, thanks.

This file has been downloaded 832 times

S.C. Wack - 24-11-2018 at 09:20

The alcohol part(s) is being converted to benzoate ester, perhaps with benzoyl chloride. Hydrolysis might be a challenge if AC is a little sensitive to everything, so it's hard to say without some investigation.

Thanks for letting us know what's going on.

S.C. Wack - 25-11-2018 at 14:26

PS For instance Allen's Commercial Organic Analysis (1910) tells of the method of Thoms that determines eugenol in clove oil by making the benzoate quantitatively (mp 70C.) and weighing it. The no-surprise method they give for ester hydrolysis in general (30 min. 0.5N ethanolic KOH reflux) is that sort of thing IDK how well plays with the allyl part, even though this isomerization isn't super sounds like a good place for inert gas in any case...obviously BzCl is not an obstacle except for expenses; homemade from BnH/Cl2 if nothing determine total eugenol (free and combined such as eugenol acetate) Allen's says 5 g. of clove oil is heated for 30 min. with 20 ml. 15% NaOH, 6 g. BzCl is added, etc.

[Edited on 25-11-2018 by S.C. Wack]

CycloKnight - 26-11-2018 at 06:41

Thanks for the feedback. So far I've tried both 1 hr NaOH (30%) aq reflux and dilute HCl reflux, with no discernible change at all. I can try in ethanolic KOH next. The AC solution (ethyl acetate extract of steam distillate, when evaporated the residue turns black when left exposed) I have turns red in alkaline solution. I'd have expected any AC yield from hydrolysis to also change the solution color to red, but there was no color change in previous attempts.
I carefully measured the crystal mp (72 C) before becoming aware of the dibenzoyl derivative (so no possible confirmation bias there), it was measured when heating to melting point, not when crystallising. Now there is also a waxy crystalline product measuring higher, closer to low 80's C. No obvious reaction with acid/alkali. Its a mixture. No TLC yet.

I've now redistilled the eugenol I've been using (using vigreux column), I was concerned there could be some eugenol acetate contamination but didn't see any sign of this. In fact it came over at a slightly lower temperature than expected, whereas eugenol acetate would come over higher. Its now water clear and has lost some of its viscosity, will soon find out if this has made any apparent change to how the demethylation reaction performs.

[Edited on 26-11-2018 by CycloKnight]

Melgar - 26-11-2018 at 22:35

I posted a paper a while back where demethylation was performed with lithium iodide. Although it seemed like an unorthodox reagent to use, it's actually pretty straightforward to prepare, as long as you remember that lithium will react vigorously with nitrogen gas if it gets too hot. IIRC, it works better if the I2 is molten, and Li metal is added in very small portions. The reaction between lithium metal and elemental iodine is the least vigorous reaction between an alkali metal and a halogen, but it's still reasonably energetic. I guess it forms LiI3 first, then LiI.

Now I can't remember if I was successful in the workup or not, so attempt this at your own risk. All the sources I've found say LiH should be reacted with elemental iodine in ether, incidentally.

CycloKnight - 28-11-2018 at 17:24

Update on the latest demethylation experiment, interesting results this evening!

Completed another eugenol demethylation run. Reagents used were 26 g eugenol (in 50 g toluene), 100 g bromine, 12 g Al in 110 ml bromotoluene/toluene.
Reaction and workup were carried out differently this time, but more on that later.

After hydrolysis, vacuum distilled the organic phase and toluene extracts. No steam distillation for this run.

Collected the fraction from 150 to 185 C.
Crystallized in isopropanol.

First vac filtration.
(Sample in FeCl3 hasn't been stirred, so the green granules are visible. When stirred the solution turns very dark green.)

First two.

Small sample in 5% aq NaOH.

Third is still crystallizing. The camera isn't too good, but the first crop is almost fluorescent white. It appears to melt in water at 50 C (forming a light brown oil) and solidifies rapidly when cooled below that temperature. I measured using an old thermometer, not the most accurate, so give or take a degree or so.
Intense dark green in alcoholic FeCl3 solution.
Don't have the final yield yet, still drying under vacuum. The damp weight of the two filter cakes is just over 13 g.

This may well be the product that's been contaminating the samples from previous runs, yielding a (false) green FeCl3 indication.

It has a rather unique (and useful!!) solubility characteristic in isopropanol; its solubility drops off very rapidly when cooled.

The crystals, in very fine needles drop out of solution and the discoloration remains in solution, yielding a snow white precipitate. No assay on purity yet.

[Edited on 29-11-2018 by CycloKnight]

clearly_not_atara - 28-11-2018 at 18:57

Quote: Originally posted by Melgar  
I posted a paper a while back where demethylation was performed with lithium iodide. Although it seemed like an unorthodox reagent to use, it's actually pretty straightforward to prepare, as long as you remember that lithium will react vigorously with nitrogen gas if it gets too hot. IIRC, it works better if the I2 is molten, and Li metal is added in very small portions. The reaction between lithium metal and elemental iodine is the least vigorous reaction between an alkali metal and a halogen, but it's still reasonably energetic. I guess it forms LiI3 first, then LiI.

I also posted a paper where demethylation was performed with LiCl, but it required a long reflux in DMF. Lithium-based reagents may be the key to this reaction because they will not catalyze Friedel-Crafts reactions, unlike Al.

CycloKnight - 29-11-2018 at 06:08

Total dried yield is 12 g. The alcohol solution is still showing a strong green FeCl3 indication, perhaps removing solvent under vacuum and chilling would crash out more yield but time to move on I think.

As for the total yield, it's worth bearing in mind that no inert atmosphere was used. I had initially hoped that the Al Br complex would shield against oxidation/degradation but that doesn't seem to be the case. With each demethylation reaction, the reactants on the edge of the flask most exposed to air turn dark very quickly, particularly during the heating cycle. It creates a lot of tar, which causes problems with the workup later.
Anyhow, it would seem it is indeed necessary to heat to 100 C after the eugenol addition is complete, and hold for an hour. I've got the argon supply sorted now and will run the next reaction under argon, and I suspect it will make a significant improvement.
If there isn't too much tar, then after hydrolysis I'll try simply removing the solvent (bromotoluene & toluene) under full vacuum then add to isopropyl alcohol to try to precipitate the product. If that works then that will make life a lot easier.

That last vac distillation was a pain, too hot and slow. Everything solidified in the condenser, and the tar in the distilling flask will take some days to dissolve. That's the tar from the heating cycle during the demethylation, carried over in the solvent.

As an aside, the last bromine distillation run I did took 2.5 hours to complete (before drying). That was using the 3L RBF and overhead stirring. After drying with H2SO4, total bromine yield was exactly 625 g. Or enough to demethylate 163 g eugenol with the current procedure.

Water 350 ml (minimum quantity to dissolve NaBr)
NaBr 308g (3 mol)
KMno4 94g

After the above are mixed, 500g H2SO4 is slowly dripped into solution to distill the bromine. Expected yield is 228 g bromine or 95% of theoretical.

Last run used 3x scale as above, except I only used 750g H2SO4 instead of 1500 g (hence why my yield was less than 684 g).

[Edited on 29-11-2018 by CycloKnight]

clearly_not_atara - 29-11-2018 at 18:06

Very nice work! Store the product carefully, it oxidizes.

CycloKnight - 29-11-2018 at 18:40

Thanks! Its all sealed, under argon and in a freezer now.
Just completed another eugenol run with continuous argon purge, same scale as previously. Haven't completed the workup yet but there is virtually no tar. The organic phase after hydrolysis had the appearance of off white toothpaste. Quite a difference to the usual dark brown and tar everywhere. It emulsified heavily, but after a little heating the phases have separated fine. If the FeCl3 indication is anything to go by, this looks to have gone very well so far. Will see what tomorrow brings.

clearly_not_atara - 29-11-2018 at 18:59

Oxygen is a jerk. It always climbs into the flask, helps itself to some electrons, and doesn't give a crap what else happens.

CycloKnight - 30-11-2018 at 05:31

After hydrolysis step.
(Previous demethylation step also involved the heating to 100 C. Without inert atmosphere this normally creates a lot of tar.)

It whipped up into a vicious emulsion with the overhead mixer. Not HCl nor adding more toluene made much difference. Heating up to about 55 or 60 C did the trick, and separated into a clear bottom layer and honey/amber organic top layer of about 700 ml of (mostly) toluene. It was left to cool overnight to ambient temperature.

The morning after.....

Isopropyl alcohol & FeCl3 solution.

Looks just like fiberglass, and there's quite a lot of it. The toluene hasn't even been fridge/freezer chilled and its saturated. Will be removing toluene under vacuum and then crystallizing straight from the toluene. Can't get much easier than that. So no vacuum distillation of the product, thank goodness.
Will find out the total yield maybe later today or tomorrow.

CycloKnight - 30-11-2018 at 06:49

Quote: Originally posted by clearly_not_atara  
Oxygen is a jerk. It always climbs into the flask, helps itself to some electrons, and doesn't give a crap what else happens.

That oxygen chap does sound quite lazy with all that climbing into flasks and hanging around, not much of an aerobic I assume.


[Edited on 1-12-2018 by CycloKnight]

CycloKnight - 30-11-2018 at 10:45

Its looking like a final synthesis and workup procedure isn't far off now, with provisions for refinements of course.
There is a lot that needs to be said regarding the complex formation step (bromine addition), by far the most hazardous step. There are some serious issues that need addressing (other than bromine handling). Particularly with reference to thermal runaway, and catastrophic gaseous HBr release due to over zealous bromine addition rate. The reaction can be deceptive, in that it can appear to have started when it hasn't and if bromine addition is continued anyway it can take off and go out of control very quickly. I've worked out how it's done safely so as long as some simple rules are followed then all should be fine. I'll include all this in more detail when I post the final write up.

CycloKnight - 1-12-2018 at 09:58

Total yield for the last experimental run looks to be close to 20 g, just removing the last traces of solvent still.

The exact bromine mass used for the last run was 105 g, and eugenol was 30 g.

The toluene solution was Guinness black by the time it was distilled down to 500 ml, so there has been a fair bit of degradation in the workup. I added a good few hundred ml trying to break the emulsion after the hydrolysis and probably about 100 ml 35% HCl.

For the next trial, will keep it all under argon even for the hydrolysis & phase separation steps and won't be adding solvent nor extra aq HCl to break the emulsion, just heating under argon.

Reagent ratios.
It's worth noting that the demethylating complex reagent to eugenol ratio being used so far is about 2.63 to 1, or very close.

The prototcatechualdehyde demethylation patent (US patent 2975214) claims a vanillin demethylation efficiency of 62% when using a complex to vanillin molar ratio of just 1:1. Compared with 91% when using a ratio of 2.5:1.

For this eugenol variation, if we consider the bromine being the limiting reagent (rather than eugenol), then that would scale to 67 g eugenol per 100 g bromine, 12 g Al + solvent.

Therefore, for the next trial - will be keeping the other reagent ratios the same except for eugenol which will be increased to 60 g (per 100 g bromine) to see how this performs.

Loptr - 17-12-2018 at 15:30

Has anyone seen this paper about demethylation of methyl vallinate, and eugenol, using AlI3, ethyl acetate, and DIC in acetonitrile with yields up to 98%?

There is also an IBX/sodium hyposulfite method. There are a couple of others that I have not seen referenced before.

Attachment: tian2017.pdf (530kB)
This file has been downloaded 774 times

Screenshot_20181217-183243_Adobe Acrobat.jpg - 523kB

[Edited on 17-12-2018 by Loptr]

clearly_not_atara - 17-12-2018 at 21:16

DIC in that article stands for what I expected, diisopropylcarbodiimide.

For reasons I should not have to explain, there is no need to wonder about whether any amateur is ever going to perform this reaction; the answer is no.

IBX is well-known. IBX is already not used because it's inaccessible; diisopropylcarbodiimide makes IBX look like toothpaste.

[Edited on 18-12-2018 by clearly_not_atara]

Loptr - 18-12-2018 at 03:17

Quote: Originally posted by clearly_not_atara  
DIC in that article stands for what I expected, diisopropylcarbodiimide.

For reasons I should not have to explain, there is no need to wonder about whether any amateur is ever going to perform this reaction; the answer is no.

IBX is well-known. IBX is already not used because it's inaccessible; diisopropylcarbodiimide makes IBX look like toothpaste.

[Edited on 18-12-2018 by clearly_not_atara]

Yeah, I know what they are and for have a similar reagent on my shelf. It's a liquid similar to DCC.

The DIC is acting as a acid scavenger in this paper. Could another one with similar properties not be tried?

This was another paper that I hadn't seen before and thought I would share.

clearly_not_atara - 18-12-2018 at 11:26

You could probably do it with all sorts of things. Particularly, the rxn is known to work with no DIC, but in much worse yield.

DIC is a strange choice of "acid scavenger" (base?). I'd guess this was its purpose:

AlI4- + DIC >> iodo-DIC- (iPrN(-)C(I)=NiPr) + AlI3

i.e. it sequesters iodide from tetraiodoaluminate, reactivating it. That requires more than a typical "acid scavenger" like triethylamine. If you have a carbodiimide handy, it should be fine, but to me this seems like a waste compared to demethylations with LiX or PyHX etc.

monolithic - 2-1-2019 at 19:33

Quote: Originally posted by Loptr  
Has anyone seen this paper about demethylation of methyl vallinate, and eugenol, using AlI3, ethyl acetate, and DIC in acetonitrile with yields up to 98%?

There is also an IBX/sodium hyposulfite method. There are a couple of others that I have not seen referenced before.

[Edited on 17-12-2018 by Loptr]

Check out 0.1055/s-0037-1610996 (attached) for a related paper. Still, I think one of their control experiments is a lot more interesting. 1 molar equiv. eugenol, 1.1 molar equiv. aluminum isopropoxide and 1.1 molar equiv. KI in acetonitrile. Heat for 18 hours at 80C to yield 52% 4-allylcatechol.

Yields are terrible compared to their AlI3 procedure, but aluminum isopropoxide is a hell of a lot cheaper to buy or prepare and KI is pretty much OTC.

[Edited on 3-1-2019 by monolithic]

[Edited on 3-1-2019 by monolithic]

Attachment: 10.1055_s-0037-1610996.pdf (479kB)
This file has been downloaded 632 times

CycloKnight - 24-6-2019 at 05:42

Just a quick update.
I've abandoned the previous method I've used here for demethylating eugenol, this is why I haven't posted the procedure. Its incomplete.
Its a lot of work and rather low yielding (for the bromine used) to be of any interest other than strictly academic, and the product is contaminated with what I believe is likely to be 4-propyl catechol, mp 60 C, which is a nuisance to separate. There are also many side products, including various esters and some kind of vanilla analog(s) that I've yet to identify.

Before giving up on this particular demethylating reagent approach, I tried using bromo eugenol instead of eugenol.
Much better results. No decomposition at all, I even let the temperature run to 115 C to push the limits somewhat, and for over 90 minutes, during the final demethylation heating stage. This would have turned eugenol to sludge.

The difficulty I've been having is the dehydrobromination step (necessary to form the 4-allylcatechol). So have been experimenting with bromo eugenol dehydrobromination to dial in the reaction parameters, before running it with the demethylated product (circa 7 grams from the last run).

The last dehydrobromination run was with 1g bromo eugenol, 20 ml anhydrous ethanol, and 4 g KOH (16:1 molar ratio, KOH:bromo eugenol), and refluxed for 5 hours. Still, little conversion to eugenol. Tested by checking for polymerisation in conc. sulfuric (sample diluted using 1 drop in a few ml methanol)
Suspecting the sec-alcohol has been formed instead, I tried heating a sample in a test tube with KHSO4 and then retesting as before, and there was a lot more polymerisation (strong red indication in conc sulfuric), and there was some moisture released indicating possibly that the alcohol has indeed been formed instead of the alkene.

It might be easier to just go down the aqueous alcohol KOH reflux to the sec-alcohol and take it from there, KHSO4 dehydration, etc. That's the next thing I'll be trying.

If I can't reproduce the allyl group in decent yields then there isn't much promise for this route, at least for 4-AC.
Anything else I could try?

clearly_not_atara - 24-6-2019 at 08:36

When I had suggested using bromoeugenol previously, the next transformation was with acetate salts to form the acetate ester, 4-(2-acetoxypropyl)-catechol. This acetoxy group is easier to attach than dehydrohalogenation and it is quite stable. The acetoxy compound can be used directly in further transformations, eg methylenation.

Methanol or GAA are probably good solvents for this rxn (Sn2 Br >> AcO), but an aprotic solvent like propylene carbonate or DMF is ideal.

CycloKnight - 25-6-2019 at 03:55

Thanks for the tip! I'll give it a try.

CycloKnight - 22-9-2019 at 08:45

Quote: Originally posted by clearly_not_atara  
When I had suggested using bromoeugenol previously, the next transformation was with acetate salts to form the acetate ester, 4-(2-acetoxypropyl)-catechol. This acetoxy group is easier to attach than dehydrohalogenation and it is quite stable. The acetoxy compound can be used directly in further transformations, eg methylenation.

Methanol or GAA are probably good solvents for this rxn (Sn2 Br >> AcO), but an aprotic solvent like propylene carbonate or DMF is ideal.

It works!
Yesterday's experiment went as follows:

Ester formation
5 g bromoeugenol (0.0204 mol)
3.8 g sodium acetate trihydrate (0.028 mol) (1 g excess)
50 ml methanol

Refluxed 90 minutes.
Water added, methanol distilled off.
Extracted with ethyl acetate

Ester + 2 g NaOH + 50 ml methanol
Refluxed for 4 hr 25 minutes
Water added, methanol distilled off.
Extracted with ethyl acetate

Ethyl acetate evaporated to yield 3.77 g amber oil

A small sample was placed in a test tube, along with a small amount of KHSO4, and heated to reflux using a hot air gun.
Water vapour came off immediately, condensing on the glass and droplets falling back into the hot oil and popping as water does in hot oil.

The remaining product contains eugenol, with a small amount of tar. No discernible alcohol remaining.
Normally this reaction would be carried out at lower temperature and under vacuum, so some tar is to be expected no doubt.

The alcohol (1-(4-Hydroxy-3-methoxyphenyl)-2-propanol) is quite distinctive as it's quite acrid, like pepper spray or capsaican. Much like its vanillyl methyl ketone sibling.

After I repeat this on a larger scale, I'll be able to work out the yield % but for now I'm pleased with the results!

[Edited on 22-9-2019 by CycloKnight]

CycloKnight - 22-9-2019 at 09:49

Update on the bromoeugenol demethylation experiments.

The bromoeugenol preparation method I'm using is based on the Fester Method for bromo-saf. preparation, which I found somewhere on the internet. It'll come up with a quick internet search.
By volume, 1 part clove oil, 1 part GAA, 2 parts 48% HBr. Place in ice bath, maintain between 10 and 15 C, and gas with HCl until it turns dark purple blue. Mag stir for a few days, aroma changes from clove oil to fruity.
Quench, toluene extract, wash, and vac distill.
I lost a rotary pump due to HX release during the vac distillation(s), so an inline NaOH trap is a must (!!).

Workup is the same, except obviously the NaOH wash is omitted, since it will pull out the bromoeugenol. 120 g clove oil (just over 100 g eugenol) will yield about 100 g bromoeugenol.

For the demethylation experiments, I had a number of spectacular failures, couldn't quite figure out what the problem was. The toluene used for the AlBr2 demethylation complex solvent must be 100% dry, to achieve this I've been using chemically dried toluene and adding a small amount of mercurous chloride to the aluminium/toluene mixture, and allowing to stir an hour or two before beginning the Br2 addition.

If ANY moisture is present at all, then the final product will be heavily polymerised. On a small scale (~10 g bromoeugenol), it works fine. But when scaled up, the result was failure but I mistakenly assumed it to be moisture causing the problem.
It was actually poor mixing, mag stirring will not suffice on the >70 g bromoeugenol scale I had been using. I believe its down to a coating (either Al or Ar. AlBr2 complex?) that forms on the side of the RBF that accumulates, Overhead stirring is a must on that scale. The good news is that its all working now.

Additionally, I've found that it is necessary to perform an alkali hydrolysis (2 hr alkali reflux) with the demethylated product to yield the diol. As per S.C. Wack's suggestion earlier in the thread.
With the demethylated bromoeugenol, there is a lot of tar that's been forming in this step, about half the product being lost to tar. Argon atmosphere is used always.
Base hydrolysis is with KOH/methanol. So next step is to try with NaOH/methanol and reduce the molar excess.
So far I haven't made any 4-allyl catechol using the bromoeugenol variation yet, but shouldn't be too much longer before we find out whether this is a viable approach or not.

[Edited on 22-9-2019 by CycloKnight]


CycloKnight - 14-10-2019 at 12:37

The bromoeugenol demethylation worked on the small scale (10 g), confirmed via methylenation, but the previous scaled up run (105 g) did produce a fair bit of polymerisation, I suspect due to insufficient Hg2Cl2 added, I only added a small amount as I'm running quite low. Haven't completed the workup on it yet, but I suspect the yield to be on the low'ish side.

As an aside, early in the summer I developed a method for isoeugenol synthesis from eugenol distilled from clove oil. Ethylene glycol is used as solvent (antifreeze, straight from the bottle without any refinement), and argon/inert atmosphere.
The ratios were eventually dialed in as follows:

120 ml ethylene glycol
150 g eugenol (vac distilled)
240 g potassium hydroxide

The potassium hydroxide is slowly added to warm ethylene glycol / eugenol in a 500 ml 3 neck flask with argon flow. This takes about 30 minutes, waiting for the KOH to dissolve and without it boiling over. Set up for reflux. The temperature is then brought to 155 - 160 C and held for 5-6 hours.

After which, its allowed to cool, neutralised, extracted with toluene, washed with distilled water and then vacuum distilled. Isoeugenol purity tested by buffered peracid oxidation and subsequent glycol hydrolysis to methyl vanillyl ketone, which it does though the ketone yield was consistently about 50%, somewhat lower than expected.

Anyhow, this method produces a high boiling point fraction that I had assumed was likely to be cis isoeugenol. An isomerisation time of 5 hr 5 min was recorded.
From my notes, the vacuum distillation went as follows.
Fraction 1:
(19:00) 124 C, rapid heating to 132 C. Change to fraction 2.

Fraction 2 (isoeugenol)
(:03) 132.5 C. (:04) 134 C, ~ 4 ml. (:05) 136 C, 137 C. (:06) 138.5 C, 1.5 drips per second. (:07) 139 C, 140 C. (:08) 142 C. (:10) 144.5 C, water clear. (:12) 154 C, 155 C, ~ 30 ml. (:15) 155 C. Change to fraction 3.

Fraction 3
(:17) 160 C. (:19) 161 C, ~ 25 ml, water clear. (:22) 162 C, ~ 40 ml. (:25) 163 C, ~45 ml. (:28) 164 C, HEAT OFF.

Final yields:
F1 = 7 g, F2 = 54 g, F3 = 54 g.

Most of fraction 3 was used for other experiments, only several ml were left in a plastic container, back in June. Labelled as "High BP fraction" and hasn't been touched since.
Here it is now:

It was a water clear oil, now a very viscous semi-solid goop. The partial vacuum which has collapsed the container is likely from the oxygen being absorbed. All of my demethylation waste solutions (including bromide recycle bottles) containing leftover diols do the same, if the air isn't flushed out with argon before sealing.

Is there anything other than catechols, likely to be present, that will absorb oxygen under these circumstances?
Its probably nothing to get excited about, but no doubt worth following up!

[Edited on 15-10-2019 by CycloKnight]

Mush - 31-10-2019 at 14:16

4-allyl catechol

4-allyl catechol really only serves one purpose - as an intermediate to synthetic safrole. It is a white to tan powder with a characteristic smell of diesel/smoke. Quick or easy are two words that should never describe the synthesis of 4-allyl catechol. While it's not particularly hard, the reagents needed have to be made or bought from a special supplier. It is produced by reacting with an aluminum halide to form the phenol adduct, which is then hydrolyzed.

Aluminum Iodide demethylation

This route is perhaps the easiest, because I find it significantly easier to purchase iodine crystals than anything else. It is made by reacting aluminum with iodine in a solvent, then introducing the eugenol to this. I can't stress this enough: everything must be DRY! The only downside to this reaction is it takes an assload of iodine (which is expensive) to do any sizeable amounts.

2Al + 3I2 ----> 2AlI3

C10H12O2 + AlI3 ----> AlI2C9H9O2 + CH3I

AlI2C9H9O2 + 3H2O ----> C9H10O2 + 2HI + Al(OH)3

Aluminum metal (foil, powder, chips)
Benzene or cyclohexane (NOTE A)
Inert Atmosphere (NOTE B)
Your favorite fabric softener

Separatory funnel
Magnetic stirrer recomended
addition funnel (24/40 joints)
Vacuum filtration equipment
Reflux Column w/ 24/40 joints
Reflux Column w/ 24/40 joints
3 neck 1000ml RBF w/ 24/4joints

Prelab: After all of the glassware is washed, it should be dried in an oven to remove all traces of water. Set up the three neck flask with one stopper, one reflux column and the addition funnel. Once put together, attach a calcium chloride drying tube to the top of the condensor. Water elimination is extremely important as even 1ml of water can reduce the yeild to 0.

Into the 1000ml RBF, pour 150ml of benzene or cyclohexane (NOTE A). To this, add 60g of iodine crystals, followed by 4.5g of aluminum (foil, chips, powder). Reflux this until the red/violet iodine color disappears and a grey-tan color is introduced. Once this is done, set up your inert atmosphere (NOTE B) by hooking it to the previously stoppered joint on the flask. Displace any air in the system with the inert atmosphere as you cool it in an ice bath to below 30C. Once below 30C (or roughly). Charge the addition funnel with 30g of eugenol, 3-5ml of your favorite fabric softener (I used snuggle), and a little (~20ml) benzene or cyclohexane to wash it all down. Begin slowly adding this to the mixture while maintaining the cool water through the condenser. If you do not have a 3-neck flask, or an addition funnel, you may drip it through the reflux column.

At this point, methyl iodide is being formed. It boils at 42C, so you can see why reducing the temp below 30C is important. Methyl iodide is quite poisonous and adequate ventilation is necessary. Continue slowly adding, while maintaining the inert atmosphere. It is recommended to include magnetic stirring at this point, but I did without it. Once it has all been added, reflux it for 1-2 hours with the inert atmosphere still running. After this time, you'll have a flask full of anything from dark brown to light tan solids. Let it cool to room temp and add 100-200ml of water. Continue the inert atmosphere the entire time. If there are clumps in there, you will have to break them up in some fashion to fully hydrolyze it. Once it is a fine suspension, transfer it to a beaker of adequate size. Use a wash bottle to get out any of the remaining solid. You may discontinue the inert atmosphere at this time. Let the precipitate settle to the bottom for a couple hours. Once settled, decant the water into a filter, and follow it by the precipiate. Use vacuum and get off as much liquid as possible in the filter. You must let it settle first, otherwise you will be waiting forever for it to filter, and atmospheric exposure to wet 4-allylcatechol can decompose it.

Place 100-150ml of diethyl ether in the 1000ml flask from before to get any residue, then pour it into a 600ml beaker. Add the solid material on the filter to the beaker of ether and stir it well. This will extract the 4-allyl catechol from the aluminum hydroxide and other salts. Cover the beaker and let it settle for a few hours again. Once again, decant the ether into a filter, followed by the solid. Run a small portion of ether through the solid before it clogs up the filter completely. You must let it settle and decant first, otherwise the superfine aluminum salts will clog up the filter and your ether will evaporate faster than it can go through the filter. If there is water in your ether, separate it with a separatory funnel, then just evaporate the ether off with a water bath or with vacuum. You should be left with a white to tan colored solid. This can be recrystallized from petroleum ether. Theoretical yeild is 23.67g but you should obtain about 20g.

NOTE A: Substitute the solvent as you like with a non-polar solvent. Do not use diethyl ether or ethyl acetate. Useable solvents include benzene, cyclohexane, toluene, xylene, chloroform (reduced yeild), nitrobenzene (poor yeild), chlorobenzene, etc.

NOTE B: For the inert atmosphere, nitrogen is recommended, but this requires a cylinder of nitrogen. Propane can be used without any problem. Use a torch and connect a rubber tube to it. Run it through a calcium chloride dryer first, then run it into the flask. At the top of the reflux condenser, place a stopper with a small peice of glass tubing extending from it. Since propane is very flammable and also contributes to global warming, it is best to burn it as it is used. Let propane flow through the flask/condenser for 30 seconds, then light the little glass tube. You only want a very small flame - similar to a cigarette lighter - to come from the tube. You can increase the size if it wont stay lit. A benefit of this, is it will burn off any methyl iodide that leaves the condenser, producing CO2, H2O and I2 in the gas phase. Iodine is toxic, so ventilation is required, but it's less dangerous than methyl iodide and gives an interesting purple smoke as it burns.

CycloKnight - 3-11-2019 at 14:08

One of the diol side products I've been preparing, when methylenated yields an oil with an unusually high density. A pleasant fragrant oil, an aroma similar to sassafras but different. I had initially assumed it to be dihydrosafrole owing to its similarity. It doesn't absorb Br2 as quickly or as much as eugenol does. It will absorb Br2, but only slowly. Suggesting ring bromination rather than allyl.

It has a density of about 1.29 g/cc. I have about 8 ml of it left, the diol compound wasn't prepared from bromoeugenol, but from eugenol.
I suspect it's 3,4 methyledioxy indane, and the diol compound being 4,5-indanediol (predicted density of 1.3 g/cc).
Its only generated when the reaction (demethylation) conditions are hot, iirc about >90 C.
I can't find any data on the methylenated indanediol.
Assuming this is what it is, is there likely a practical way to revert the indane back to the allyl form?
I'm assuming not, so will be aiming to avoid its synth in the first place.
Next demethylation experiment will be with isoeugenol in place of bromoeugenol/eugenol, will be interesting to see the difference.

Cou - 7-11-2019 at 18:07


Mush - 29-3-2020 at 12:25

Quote: Originally posted by CycloKnight

Could use a little help. I'm looking for information on this dibenzoyl derivative, its probable structure, etc, and if it can be easily reverted to the 4-allylcatechol. Any feedback appreciated, thanks.

(ii) Benzoyl derivatives (benzamides)
Suspend 1 g of the amine in 20 mL of 5% aqueous sodium hydroxide in a well-corked flask, and add 2 mL benzoyl chloride (fume hood!), about 0.5 mL at a time, with constant shaking. Shake vigorously for 5 - 10 min until the odour of the benzoyl chloride has disappeared. Ensure that the mixture remains alkaline. Filter off the solid derivative, wash with a little cold water and recrystallise from ethanol.

Experiment 10 Preparation of the benzoate of phenol.


Many phenols yield crystalline benzoyl derivatives with benzoyl chloride in the presence of sodium hydroxide (Schötten-Baumann method).


To the phenol (0.5 g) is added 5% sodium hydroxide (10 mL) in a well-corked boiling tube or a small conical flask.Benzoyl chloride (2 mL, density 1.21 g cm-3) is added in small quantities at a time, and the mixture shaken vigorously with occasional cooling under the tap or in ice water.
After 15 minutes the solid benzoate separates out: the solution should be alkaline at the end of the reaction; if not alkaline, or if oily, add a solid pellet of sodium hydroxide and shake again.Collect the benzoate, wash thoroughly with cold water, and recrystallise from ethanol (NO FLAMES!). Carry out the tests for phenols and esters (p 57).