Author Topic: Demethylation of vanillin.  (Read 7599 times)

0 Members and 1 Guest are viewing this topic.


  • Guest
Demethylation of vanillin.
« on: April 03, 2002, 09:51:00 PM »
Here's a small research :)

As a brief introduction, let me note that demethylation of vanillin to protocatechualdehyde with H2SO4  IS possible. I haven't been able to find an explicit procedure of this preparation, however, there is a

Patent US3367972

which describes a selective demethylation of veratraldehyde: the main product is isovanillin, the rest is vanillin and protocatechuic aldehyde (PCA). The conditions are very harsh: 96% H2SO4 is used and the temp is ~100  C ::)  (Osmium, how about applying this to eugenol? :o ) Luckily, there is a table that shows correlation between time, temp and proportion of products (see the patent, page 3). The most PCA they arrive at (remember, their goal is isovanillin!)  is 17-18%, gotten by either 3 hrs at 103-105 C, or 143-145 C for 4 minutes. The rate of charring is worse (somewhat) in the 2nd case.

Quite reasonably, i think, one could suggest that by cooking the reaction for a longer time, one would arrive at PCA being the major product. The rate of charring doesn't seem to bee high (10% after 3 hrs at 105 C - and will probably get even better if lower temp/longer time is used). Don't forget that in our case, vanillin is already once demethylated. So some research will bee needed still... If only someone could find such a patent for vanillin!

Generally, all literature stresses the fact that vanillin is a bitch to demethylate. HCl at superatmospheric pressure, as well as AlCl3  give "very poor" yields (or a large xcess AlCl3 and some co-catalyst,such as urea, is required). AlBr3 works, but the yield is also not good. HBr also yields less than 50%, giving in addition a ring-brominated by-product, which is difficult to separate. All in all, they say, 50% yield is never exceeded (just don't ask me for the refs - this is a digest of many patents and textbooks, which i've been digging into for not less than 6 months now :) )

But the real MacCoy (could someone, please, explain me, who is McCoy?) is the

Patent US2975214

. Here, vanillin is demethylated in a 91% yield using a catalytic complex of AlBr3 w/an aryl hydrocarbon, and to make things even simpler, AlBr3 is prepared in situ from aluminium and bromine. See the patent for a lot of interesting details, i'll just type in one of the examples:

67,5 g of Al granules (superficially activated w/copper -by that he simply means dipping it into CuSO4 aq for some time, then washing it and drying) are suspended into 3,0 kgs of dry technical xylene, and bromine, volatilized in a stream of air (also possible to add it in liquid form) is passed thru the well agitated mixtr, with cooling to keep the temp at 15-20 C, until a total of 600 g Br2 have been added. To the resulting complex there's added in small portions w/good stirring, 152 g (1 mole) vanillin. The rxn mixtr is agitated for 1 hr at 15-20 C, heated to 95 C in the course of an hour, then heated at 95-10 C for an hour w/continuous agitation. Methyl bromide, evolving during the rxn may bee condensed & recovered.
The rxn is cooled to RT, poured into 250 cm3 of 22 Be (what's this?) HCl acid in 10 kg of cracked ice and agitated until the ice is completely dissolved.
The aq. phase is xtracted w/4x500ml ether, and the solvt is stripped off. The org. phase is xtracted w/3x1000ml of 4% caustic soda, then acidified w/conc HCl, cooloed overnight at 0-5 C, and the precipitated PCA is filtered off. The aq. phase is then xtracted w/the same 4x500 ml ether used bee4 and ether removed (NO further cleaning!!! No vac. distillation, bisulfite adduct formation, no nothing! )

Yield 126 g of protocatechuic aldehyde, 91% of theory, mp 152-153 C (!)

In case you dislike the large amt of solvents/reagents needed for the rxn - the authors state that less catalyst can bee used, jeopardizing some of the yield, namely:

AlBr3/Vanillin ratio              Yield%

1,0                                           62
1,5                                           71
2,0                                           89
2,5                                           91
3,0                                           92

Thus, on the same scale it's possible to get ~two times as much (214g) of PCA from  ~three times as much (380 g) vanillin.

Well... Methylenation of PCA to piperonal is a known process.

So if you happened to like this post, feel free to write in reply something pleasant :)

Empathetically yours,



  • Guest
I happened to run into the Japanese patent ...
« Reply #1 on: April 03, 2002, 11:58:00 PM »
I happened to run into the Japanese patent JP10265428 in which ethyl vanillin was dealkylated to protocatechualdehyde with sulfuric acid in 91% yield according to the abstract. The hive needs more Japanese bees!


  • Guest
22 Be (what's this?)
« Reply #2 on: April 04, 2002, 09:10:00 AM »
Well done once again Antoncho !

An incremental contribution to this thread:

"22 Be (what's this?)"

°Be is a scale for hydrometers when estimating the strength of an acid from its density.

In Europe,   Density, g/ml = 144.3/(144.3 - °Be)
In USA,      Density, g/ml = 145/(145 - °Be)

Therefore I assume that since the patent is from the US the latter applies, so 22 °Be is equivalent to 1.179 g/ml which would mean the HCl is about 36%.


  • Guest
« Reply #3 on: September 01, 2002, 04:19:00 PM »

I've been working on JP10-265428.  The full literal translation isn't done yet,
but I thought I would post a quick paraphrasing of the patent since it might be
of interest and I may not get around to actually finishing the whole translation
anytime soon (since I am a lazy, lazy bastard and my job at the Ministry of
Truth eats up most of my time). 

The actual mechanics of the process in the patent are pretty straightforward
and at the same time quite vague.  Cold 3-alkoxy-4-hydroxybenzaldehyde is added
to cold non-halide acid, the mix warmed up and allowed to react, then some
water dumped in.  They give a wide range of reactants, quantities, and
temperatures however.  There are three concrete examples at the end.


Publication date: 1998-10-06


Applicant(s): UBE IND LTD

Requested Patent: JP10265428

Application Number: JP19970072239 19970325

IPC Classification: C07C47/565; C07C45/64

First comes the abstract; the patent is for an industrially suitable method
of producing protocatechualdehyde in high yield by dealkylation of
3-alkoxy-4-hydroxybenzaldehydes in the presence of a non-halide acid.  The
alkoxy radical of the 3-alkoxy-4-hydroxybenzaldehyde here *must* have at least
two carbon atoms.

Next it describes the scope of the patent application and lists the patent
application clauses.  These are under the headings marked "[blah blah 1]"
through "[blah blah 5]".

After this comes the detailed explanation of the patent, under numeric
headings 0001 to 0017.

0001: A description of the field under which this invention falls.

0002-0005: Descriptions of older techniques for producing
protocatechualdehyde and why they suck.  Vanillin/AlCl3 method (J. Org. Chem.,
27, 2037(1962)) is the chief reference, with further mention of vanillin
processes involving HCl, HBr and acetic acid, and PCl5.

0006-0008: The patent starts listing example categories of suitable
3-alkoxy-4-hydroxybenzaldehydes. 0008 gives some specific examples that are
definitely suitable: 3-ethoxy-4-hydroxybenzaldehyde,
3-n-propoxy-4-hydroxybenzaldehyde, 3-isopropoxy-4-hydroxybenzaldehyde,
3-n-butoxy-4-hydroxybenzaldehyde, 3-isobutoxy-4-hydroxybenzaldehyde,
3-sec-butoxy-4-hydroxybenzaldehyde, 3-tert-butoxy-4-hydroxybenzaldehyde, "etc.".

0009-0011: Examples of suitable non-halide acids are given: sufuric acid,
pyrosulfuric acid (disulfuric acid), non-halide sulfonic acids, methane sulfonic
acid, ethane sulfonic acid, ethane disulfonic acid, benzene sulfonic acid,
benzene disulfonic acid, toluene sulfonic acid, napthalene sulfonic acid,
napthalene disulfonic acid. 

Some guidelines to the process are given in these sections as well. 
For every mole of 3-alkoxy-4-hydroxybenzaldehyde, 5-20 moles of acid are
required, with 5-10 moles being desirable.  The acid should be 70-100% pure
by weight, with 85-100% being desirable.  The reaction temperature should be
25-100 degrees C, with 25-70 degrees C being desirable.  The initial
temperature should be 0-70 degrees C, with 0-40 degrees being desirable.  The
amount of water added should be 0.5-10 liters per mole of 3-a-4-hb, with
0.5-5 liters being desirable.

0012-0016: "Examples of Execution" where they actually give details on
some tiny runs.  These I present fully:

0013 Example of Execution 1
Inside an Argon gas stream, at room temperature, 96% by weight sulfuric acid
(60.20mmol) was placed into a 25ml flask; after reducing the temperature
to 5 degrees C, this temperature of 3-ethoxy-4-hydroxybenzaldehyde
(6.02mmol) was added, all the while stirring.  Next, the temperature was
raised to 65-70 degrees C; after an additional 2 hours of stirring heated at
the same temperature, the reaction liquid was cooled to room temperature.
Then, to the thus obtained reaction liquid was added, with stirring, ice
water (20ml) and the reaction liquid stirred for 1 hour.  After, when the
aqueous solution was analyzed by high speed liquid chromatography, it was
established that the conversion ratio of 3-ethoxy-4-hydroxybenzaldehyde was
97%, yield of protocatechualdehyde was 91%.  Also, the entire procedure
described above was done under normal pressure.

0014 Example of Execution 2
Same, except the amount of 96% by weight sulfuric acid was changed to
116.19mmol, and the reaction time was changed to 1 hour.  Procedure was
performed in the same manner as Example of Execution 1.  The result was
conversion ratio of 3-ethoxy-4-hydroxybenzaldehyde 99%, yield of
protocatechualdehyde 96%.

0015 Example of Execution 3
Same, except the amount of 96% by weight sulfuric acid was changed to
30.15mmol, and the reaction time was changed to 11 hours.  Procedure was
performed in the same manner as Example of Execution 1.  The result was
conversion ratio of 3-ethoxy-4-hydroxybenzaldehyde 91%, yield of
protocatechualdehyde 84%.  Results of the Examples of Execution are shown
in Table 1.

[Table 1]
|Example of Execution  | Sulfuric Acid | Reaction Time | Protocatechualdehyde|
|                      |     mmol      |        hr     |        yield %      |
|Example of Execution 1|     60.20     |        2      |         91          |
|Example of Execution 2|    116.19     |        1      |         96          |
|Example of Execution 3|     30.15     |       11      |         84          |


0017: A summary.


DISCLAIMER: Myne japsaneses aint perfikt super number one, doubleplusgood etc.
etc. blah blah blah (aforementioned).



  • Guest
« Reply #4 on: September 01, 2002, 06:19:00 PM »
Thanx a lot, man!


It's always a great plesure to meet an intelligent person ;)



  • Guest
« Reply #5 on: September 01, 2002, 06:58:00 PM »
Why does the patent state that the alkoxide radical *must* Be ethyl, propyl and so on?
Is there a different consequence if the alkoxide is methyl?


  • Guest
Short answer: dunno
« Reply #6 on: September 02, 2002, 04:04:00 PM »
The patent just interjects in parentheses "excepting that the alkoxy radical possesses at least 2 carbon atoms" in several places without explanation.

There is a bit involving this in 0007 that I haven't fleshed out yet, but it doesn't look like they have much to say on this point. 

I'll post again after I grind that bit out, anything before that would just be speculation on my part.


  • Guest
E1 elimination?
« Reply #7 on: September 03, 2002, 08:09:00 AM »
Under those conditions, the alkyl group is probably cleaved off by acid catalyzed E1 elimination to give a phenol and a alkene. Formation of methyl cations is much more difficult than ethyl or higher alkyl cations, and methylated phenols either are not cleaved under those conditions or they give some nasty side reaction.


  • Guest
Patent sec 0007...
« Reply #8 on: September 03, 2002, 07:45:00 PM »
...doesn't give any reason after all.  It does however throw in a bit that there must be 2-10 carbons, and it's preferable that there are 2-8 carbons.  Again, they don't give a reason, just a list of example radicals (ethyl, n-propyl, isopropyl, n-butyl, iso-butyl, blah blah blah).

So, the obvious candidates for eschewing  10>Nc<2 : 
1) all or some of those don't work (for __ reason, immaterial really)
2) all or some of those are covered by another patent.