Sciencemadness Discussion Board

Preparation of 2-hydroxybenzylamine

Boffis - 21-4-2021 at 05:36

As part of my ongoing investigation of Schiff type condensation compounds that act as ligands I want to prepare a few from 2-hydroxybenzylamine and various aldehydes. The salicylaldehyde Schiff compound is a well known fluorescent reagent for aluminium and gallium. The compounds are easily prepared once you have the aldehyde and the hydroxybenzylamine.

The compound was originally prepared by Goldschmidt (1890; Berichte de Chm.; v23, p2741-2746.) by treating 2-methoxybenzylamine with conc HCl at 150 C. The methoxy compound being prepared from 2-methoxybenzaldehyde oxime (o-Anisaldehyde oxime) by reduction with Na amalgam.

The problem is how to prepare a small amount (say 5-10g) of the 2-hydroxybenzylamine? Looking at the compounds I have with a 2-hydroxyphenyl motif 2-hydroxyacetophenone, salicylaldehyde and salicylaldehyde azine look like promising precursors.

I wondered if it might be possible to reduce the azine, splitting the N-N linkage and generating the desired benzylamine? Nice and simple but is it possible? Anyone any ideas? I have found references to the reduction of benzaldehyde azine but this gives with Zn and acid dibenzylamine with the loss of ammonia or with Na amalgam to the symmetrical dibenzylhydrazine is formed.

Looking at Goldschmidt's method I wondered if it would be possible to reduce salicylaldehyde oxime directly. Are oximes hard to reduce to amines?

I also envisage a more involved route from 2-hydroxyacetophenone via a Willgerodt reaction to 2-hydroxyphenyl acetic acid; esterify; convert to amide and finally a Hoffmann degradation to the desired amine. A problem I foresee here is the possibility of intramolecular esterification and cyclotisation of the intermediate acid. Will there be sufficient free acid in the equilibrium mixture to drive the reaction to completion?

Another route I thought of was via 2-hydroxybenzyl alcohol via the bromide and then a Delepine type reaction to the amine but then how do you get to the alcohol?

Has anyone any other ideas or comment on any of the above routes?

clearly_not_atara - 21-4-2021 at 06:35

Reduction of this?

"Benzisoxazole may be prepared from inexpensive salicylaldehyde, via a base catalyzed room temperature reaction with hydroxylamine-O-sulfonic acid."

Alternatives include O-acetylhydroxylamine. See Kemp & Woodward 1965, attached.

Attachment: kemp1965.pdf (1.4MB)
This file has been downloaded 38 times

Boffis - 21-4-2021 at 13:31

@ clearly_not_atara; Thank you that is a really interesting route. I have just got to figure out a way to those O-substituted hydroxylamines. Since this method is very similar to Goldschmidt's do you think I might be able to reduce salicylaldehyde oxime, which is easily prepared, directly?

njl - 21-4-2021 at 13:41

Unrelated to @clearly_not_atara's idea there are processes based on phenol/formaldehyde resin chemistry to generate phenolic benzyl alcohols. Of the top of my head I can't remember any specific references but I believe a very simple step in one of the pages on Rhodium uses this reaction to make a benzyl alcohol from a phenol, formaldehyde, and base. From there substitution of the benzylic oxygen with chloride from HCl gives benzyl chloride. BzCl alkylates amine (ammonia or equivalent).

Boffis - 22-4-2021 at 01:13

@njl, that's an interesting idea too. I had thought about something similar but felt that the p- position would be preferred or at a significant side reaction giving a mixture that I would have to separate. Also I seem to recall that the reaction is difficult to control to just the monosubstituted derivative. It might work better with 4-substituted phenols as this blocks the p- position and in some case may reduce the reactivity of the phenol.

Edit: well done njl, I followed up your lead and it is indeed possible to prepare 2-hydroxybenzyl alcohol by the phenol + formaldehyde route. BUt while reading up about these compounds they mention that they resinify in mineral acids so converting it into the 2-hydroxybenzyl halide is going to be tricky. Any ideas?

It looks like I am back to the phenylacetic acid amide degradation or the reduction of the isoxazole.

[Edited on 22-4-2021 by Boffis]

clearly_not_atara - 22-4-2021 at 05:32

I must have been mistaken, because O-acetylhydroxylamine is unstable and rearranged to the N:

So instead the method Kemp and Woodward refer to via O-acetylsalicylaldoxime must involve acylating the oxime oxygen of salicylaldoxime! I'm surprised you could even do that. However, this paper appears to say that acetic anhydride attacks the salicylaldehyde oxime oxygen selectively:

And indeed, the attached study from 1914 firmly establishes that the oxime oxygen is preferentially acetylated by methylating the result and hydrolysing to obtain hydroxylamine instead of O-methylhydroxylamine.

So this may provide a way from salicylaldoxime if Ac2O is available.

Attachment: brady1914.pdf (632kB)
This file has been downloaded 38 times

[Edited on 22-4-2021 by clearly_not_atara]

njl - 22-4-2021 at 06:56

Boffis, that is a good point however no polymerization will be possible since after you form the alcohol, workup of the reaction mixture will remove excess formaldehyde that is needed for further crosslinking. The polymerization is both acid and base catalyzed but I believe the process is slowed in water. I am fairly certain this should be possible, likely at lower temperatures. I would try a solution of NaCl/HCl with the benzyl alcohol at perhaps 0-10 C. The benzyl chloride is immiscible with water.

kmno4 - 23-4-2021 at 05:44

Quote: Originally posted by Boffis  
Are oximes hard to reduce to amines?

It depends on oxime.
Universal methods, working always (almost), possible to perform by amateurs:
1) Reduction by Al(Hg) in aqueous THF or dioxane
2) Reduction by NaBH4 with Ni ("NiB") in methanol
3) Reduction by Na in ethanol
4) Reduction by H2 with Pd/C (or alumina) in alcohol

One can find countless "simple and efficient" methods, but most of them are crap, or working only for very limited classes of oximes.
The borohydride method is the best one I know - very fast, giving (very) high yields of amine and very simple workup.

Chloromethylation of simple phenol gives polysubstituted produsts, even when performed at high dillutions.
The last instance preparation for benzyl alcohols from aldehydes is Cannizzaro rection, of course. Unfortunately, it gives 50% loss of an aldehyde at start :(

clearly_not_atara - 23-4-2021 at 06:11

Hm, even better: Salicylaldehyde condenses with formamide in ~80% yields simply by heating the reagents neat on a water bath:

With a trace of pyridine. - Salicylaldehyde 3 g., formamide 1 g. and pyridine 0.285 g. (1 : 1 : 0.15 mol.) were heated on a water-bath for eight hours, left overnight and the salicylidene-formamide extracted as usual. Yield 3.1 g. or 87% of theory. The heating had to be longer than usual, otherwise a viscous mass was formed and the yield was diminished.


Without any reagent. - Several experiments were made with changes in the period of heating. With 18 hours' heating the yield became 2.8 g. or about 81% of theory.

Other amides and aldehydes need catalysts, but salicylaldehyde reacts with formamide without prompting. The product is N-formylsalicylaldimine, which could hopefully be reduced by such reagents as NaBH4. H2/Pd@C may give o-cresol.

kmno4 - 23-4-2021 at 08:47

Formamide..... It sounds very familiar, ha.
Maybe good old Leuckart would make the day ?
Large excess of ammonium formate would prevent secondary amine formation, at least to some extent. Worth trying :D

S.C. Wack - 23-4-2021 at 14:08

Aldehyde, oxime, amine; not complicated enough?

Quote: Originally posted by kmno4  
It depends on oxime.

And further on the appears that some are more easily hydrolyzed than others.

Na/Hg is particularly well suited for those with Hg and without fear and the (not particularly expensive) aldehyde, since it also produces (using p-toluidine) the desired aldehyde from the acid.

[Edited on 23-4-2021 by S.C. Wack]

clearly_not_atara - 23-4-2021 at 15:56

Quote: Originally posted by kmno4  

Maybe good old Leuckart would make the day ?

Instead of the classic Leuckart it might be a better bet to treat the pre-formed N-formylsalicylaldimine with formic acid, which should reduce if not eliminated condensations.

The product may be 4H-benz[e]1,3-oxazine instead of the expected N-formylsalicylamine due to cyclocondensation. But this should hydrolyse easily.

Corrosive Joeseph - 23-4-2021 at 18:53

Zn-Ni-NH3 should reduce that oxime nicely...


Boffis - 23-4-2021 at 23:16

Many thanks for the suggestions guys; they are very helpful. Looking at the various ideas I think that the route via salicylaldehyde oxime looks like the best route. I'll have a go at some of these reductions when I am next at home.

zed - 23-4-2021 at 23:46

Ummm. Salicyl Alcohol + HCl ---> Salicyl Chloride

Salicyl Chloride + NH3 (or Potassium Phthlamide)----->Maybe?

Boffis - 23-4-2021 at 23:54

Hi zed, the reading I have done suggests that the alcohol group condenses with vacant o and p positions on other molecules of hydroxybenzyl alcohol with the elimination of water thus causing polymerisation under acidic conditions.

zed - 24-4-2021 at 00:53

Well, it seems like there might be difficulties. I expect aromatics to behave like aromatics. Pretty stable.

But, when the rings are substituted in odd ways... All bets are off. Strange things happen.

Route via Azine... Yeah, that's doable. Least wise, it is a known route to amines. As I recall, it is a LiAlH4 reduction.

Might be more common via the Benzyl Alcohol. Alcohol + P-Toluene Sulfonyl Chloride ---> Ester

Benzyl Sulfonate Ester + NaN3 -----> Benzyl Azide

I'd try to avoid it. But, lots of folks are more intrepid than I am.

OK, I looked up Salicaldehyde Azine. Not quite the same thing.

It's a dimer, with linked nitrogens.

Proceeding via the Methyl Ether, then cleaving... seems OK.

Though the Ethyl Ether is usually easier to cleave.

The difficulty in the reduction of Oximes, is case by case.

[Edited on 24-4-2021 by zed]

[Edited on 24-4-2021 by zed]

clearly_not_atara - 24-4-2021 at 10:46

The trouble with salicyl alcohols is the tendency to eliminate to the o-quinone methide. This is a reactive polyene which rapidly forms tars.

Some nucleophiles can trap the quinone methide, but ammonia is difficult due to its volatility and the reversibility of its addition; irradiation of salicyl alcohol in liquid ammonia at the right frequency may give some o-hydroxybenzylamine.

AvBaeyer - 24-4-2021 at 20:07

The attached provides relevant information regarding 2-hydroxybenzylamine. I have been following the discussion since it started. I am surprised that nobody has found this yet


This file has been downloaded 37 times

kmno4 - 25-4-2021 at 04:00

I am surprised that nobody has found this yet

Yes, it deserves a lollipop.
However, the preparation looks to me a bit too good.
There is a paper (DOI: 10.1021/ja01651a002), where this procedure was repeated (with modifications), with slightly worse affect. The same paper gives procedure for bis-o-hydroxybenzylamine formation via H2 reduction over Pt/ethanol of the same oxime. Interesting, another paper (DOI: 10.1246/bcsj.43.226) gives procedure for the oxime reduction to desired salicylamine under similar conditions (Pd/C in methanol/water). One important difference - in the latter preparation HCl(aq) is added to the reaction mixture. This prevents secondary amine formation by trapping primary amine, as soon as it is formed, into ammonium salt. Simple but effective trick. The paper gives also reference to Al(Hg) reduction, unfortunalely in Japanese. The reference is available online, but the only thing I understood from it is 0,3-0,4 g of amine from 1 g of oxime (?).

Also interesting, that procedure for reduction of the oxime by Zn/acetic acid is also available (DOI: 10.3390/pharmaceutics2010018) The yield is not very high, but better than 60%. This is very old method of reduction of oximes, but not always working. See also this paper:

I want a lollipop too.

[Edited on 25-4-2021 by kmno4]

zed - 25-4-2021 at 14:10

I like the yield of that Na-Amalgam reduction. Straight forward method.

Form the Oxime. Reduce it!

Still, I might look for another reduction method.

Our heros of yesteryear, seem to have utilized about a kilo of Mercury to get the job done. But, I'm gonna recheck that. Could that be true?

Boffis - 26-4-2021 at 09:03

Hi guys, many thanks for the ideas and references and I will try an find you some digital lollipops!

The references also contain links to the azine reduction that I had original proposed but after reading them I think the direct oxime reduction route is the simpler route with better yields; this is in spite of the fact that I have a lots of salicylaldehyde azine, a by-product of a recent experiment. I have plenty more salicylaldehyde too. I may try several reduction routes. Unfortunately I am away from home at present and for some time so these experiments will have to wait until I return.