We want to try a Schmidt reaction starting with an alkyl carboxylic acid and converting to an intermediate azide then hydrolysis of this via an
intermediate isocyanate to the corresponding primary amine.
This is the 'Schmidt Reaction' if the process commences with the carboxylic acid directly, but there is an alternative 'Curtius Reaction' commencing
with the acid chloride. The 'Curtius Rearrangement' seems to be the middle step from the carboxylic azide intermediate to the isocyanate (common to
both reactions).
Safety-wise we think it's viable provided everything is kept chilled and dilute (<5% theoretical hydrazoic acid). We did a small experiment with a
small amount of dilute (5%) chilled aqueous sodium azide and adding a little dilute HCl to see what the fumes were like. There's no sign of any fumes
using ventilation but on allowing the reaction mixture to come up to room temperature you can detect a slight pungent aroma rather like 'metallic
formic acid' coming from the mixture. It's powerful and even the slightest sniff does generate a slight 'alkyl nitrite esque' lightheadedness. So
chilled and dilute and with good ventilation is clearly the way.
To give the game away the substrate we were looking to work with was actually ibuprofen. It's a isobutyl substituted 2-phenylpropionic acid derivative
and so there are no other functional groups which should interfere. We don't have a use for the primary amine product; it just seems like an
interesting and different reaction to try and something a bit challenging to get right and safely.
Time for some references; OrgSyn is always a good place to look at practical examples and learn:
Org. Synth. 1967, 47, 28 covers a preparation of cyclobutylamine from cyclobutanecarboxylic acid. This seems like a fairly archetypal reaction schema,
but it's a non-aqueous reaction and uses chloroform as a solvent and with concentrated sulfuric acid + sodium azide.
On Google we found a very useful PDF entitled "4.4 Degradation Reactions by Takayuki Shiori" (obviously from a book / collection, but not sure which)
which actually talks about 3-phenylpropionic acids (pretty similar) reacting readily via Schmidt and yielding the primary amine. Again, they suggest
chlorinated solvents.
In fact most references we can find about performing this reaction use not an aqueous solution but a chlorinated organic solvent. There are some which
refer to using aqueous hydrochloric acid however. For ibuprofen, water alone won't work for us since the substrate just isn't soluble.
So the questions:
1. Are chloroform / 1,2-dichloroethane etc. really safe to use with azides both in terms of explosion risk in dilute solution and in terms of
reactivity? We read lots of horrific things about azides and dichloromethane and there are plenty of examples out there about azides reacting readily
with chloro groups in compounds. Any other non-reactive non-aqueous solvents which might be safer?
2. We could try an alcohol/water mixture as a co-solvent if we can get the ibuprofen / sodium azide to dissolve - initial experiments suggest unlikely
but we could keep trying. Or does the presence of lots of water somehow mess up this reaction? Water is after all generated as a byproduct of the
reaction to form the isocyanate, but perhaps someone here knows better?
Any advice appreciated.gdflp - 25-2-2016 at 19:00
I highly recommend reading up on the Schmidt reaction and the Curtius rearrangement in Organic Reactions Volume 3 in the forum library. In
particular, it seems like there might be a method that is more amenable to you. It involves first forming the azide by reacting sodium azide with the
acid chloride, then rearranging the azide in benzene instead of a chlorinated solvent, finally working up with dil. HCl. More details are available
on page 339.BromicAcid - 25-2-2016 at 19:42
What substrate are you working with? Sometimes if your product is volatile you can use that to your advantage by charging your starter into a
pre-heated pot and letting your product volatize out as it reacts thus limiting your reaction mass and the potential for a run-away. I have run a few
Curtius Rearrangement reactions to produce isocyanates and each was tailored to the product.DJF90 - 26-2-2016 at 02:35
I hope you're not asking for answers to a commercial or academic problem on an amateur science forum. Who are you referring to when you say "we"?
As for the transformation, the Schmidt, Lossen and Curtius rearrangements are all similar but use different starting materials/reagents. There's a
fairly recent OPRD paper on a modified Lossen rearrangement demonstrated at kilo scale - find attached. If you have access to the required reagents
this may be one of the safer methods you could use.
I certainly wouldn't employ halogenated solvents with azide. Chloroform is known to undergo nucleophilic attack with ethoxide (see OrgSyn prep of
Triethyl orthoformate), and dichloroethane will behave similarly (probably worse than DCM to be honest due to reduced sterics).
I hope you're not asking for answers to a commercial or academic problem on an amateur science forum. Who are you referring to when you say "we"?
The plural personal pronoun is chemplayer's style -- evident on their YT channel.
I gather from a comment made recently that chemplayer is two people.clearly_not_atara - 26-2-2016 at 03:22
Quote:
I highly recommend reading up on the Schmidt reaction and the Curtius rearrangement in Organic Reactions Volume 3 in the forum library. In particular,
it seems like there might be a method that is more amenable to you. It involves first forming the azide by reacting sodium azide with the acid
chloride, then rearranging the azide in benzene instead of a chlorinated solvent, finally working up with dil. HCl. More details are available on page
339.
Unfortunately he won't be able to make the acyl halide... however, you can use mineral acids to make a phenyl ester with phenol:
Phenyl esters are mild acylating agents; I think they'll attack azide.
[Edited on 26-2-2016 by clearly_not_atara]chemplayer.. - 26-2-2016 at 03:47
Wow it's very honouring that you think we might be doing something commercial with the answer! But no. If I wanted to risk probably thousands of
dollars of funding on a brand new commercial reaction pilot then even if my employer was too stingy to put consultants fees in the budget, I'd at
least find someone I can pay some sort of nominal fee for 'advice' in order to cover my ass when it inevitably screws up.
"We" is me and my alter ego. It's ok don't worry we'll seek counselling at some point when time allows
Thanks very much for the Organic Reactions Volume 3 reference - very helpful. My limited understanding was that this process of converting to the acid
chloride first was known as the 'Curtius Reaction', with the 'Schmidt Reaction' being the direct hydrazoic acid reaction with the carboxylic acid.
I guess an obvious question would be, other than the advantage of using sodium azide (rather than dangerous hydrazoic acid created in-situ) is there a
fundemental difference here? Are these different reactions better suited to different substrates fundamentally, or can either one work for a given
carboxylic acid? The literature we've looked at doesn't say anything in relation to this last point, so the assumption was that they're sort of
interchangeable reactions. But perhaps you know better?clearly_not_atara - 26-2-2016 at 04:04
The Curtius rearrangement is a much safer modification (so is more popular in modern research) because it does not involve HN3. HN3 is not very
explosive but it has toxicity comparable to cyanide or H2S so is not to be taken lightly. However HN3 is stable under acidic conditions and in fact
salts of the H2N3+ ion are known. Notably the acid must be doubly protonated (RCO2H2+) for the Schmidt reaction to occur, so very strongly acidic
conditions are necessary.
Other than that the reactions are essentially the same. Schmidt reaction is incompatible with substrates that will be destroyed by strong acids (e.g.
acetals) or heating. Curtius rearrangement is not applicable if the acyl chloride cannot be formed (e.g. amino acids). The reaction can also be
carried out with this:
Still a bit nervous about using a chlorinated solvent to do this though. Will have a search and see if this is possible using perhaps something like
toluene or hexane, then perhaps try on a really small scale.gdflp - 26-2-2016 at 06:57
Unfortunately he won't be able to make the acyl halide... however, you can use mineral acids to make a phenyl ester with phenol:
Why not? Ibuprofen is essentially a carbon skeleton with an aromatic ring and the carboxylic acid. Chlorination should be easy with a variety of
different chlorination reagents.
Still a bit nervous about using a chlorinated solvent to do this though. Will have a search and see if this is possible using perhaps something like
toluene or hexane, then perhaps try on a really small scale.
Since Organic Reactions mentions benzene, I suspect that toluene will work as well. Of course, try it on a small scale first to see.clearly_not_atara - 26-2-2016 at 16:36
Quote:
Why not? Ibuprofen is essentially a carbon skeleton with an aromatic ring and the carboxylic acid. Chlorination should be easy with a variety of
different chlorination reagents.
Because if he could, he'd've told us! :p I can only assume that if you're trying to make an acyl- anything, the acyl halide is going to be at the top
of your mind -- your ability to get thionyl chloride, however, may leave something to be desired.
Incidentally, the acyl halide forming reagent with the fewest shipping regulations is trichlorotriazine, IIRC. You can also look into
triphenylphosphine/dipyridyl disulfide, the latter two reagents being nearly as safe as chloroform, and I don't think either one requires hazmat
forms...
(I have a friend who's ingested a significant amount of Ph3PO, either because he was curious or just a huge fucking idiot, depending on how you view
the situation... he's alive)
[Edited on 27-2-2016 by clearly_not_atara]Dr.Bob - 26-2-2016 at 17:32
I have not done one of these personally, but have friends who have, even on large scale. Be careful how much of the acyl azide intermediate that you
create at one time, and try to keep it dilute. The smaller the alkyl group on it, the more unstable it is.
My research adviser in college managed to knock himself unconscious with HN3, and nearly made a real mess when he passed out while helping one of his
graduate students do a similar reaction. And he was a very well trained chemist, so I learned to be extremely careful with azides.
That is not to say I don't use them, but I use a lot of care and try to keep them small scale and well ventilated. If you can use a solvent like
toluene, that might be a good choice. And be careful of any metals contacting the azides at any time, use plastic, ceramics, glass, but not metals.
AvBaeyer - 26-2-2016 at 20:15
The paper posted above by DFJ90 (thanks!) tells what you need to know to safely convert a carboxylic acid to an amine - see Scheme 4.
Carbonyldiiimidazole (CDI) is quite a safe activation reagent and is usually available on ebay as is hydroxylamine hydrochloride. I use CDI routinely
for a variety of reactions requiring carboxylic acid activation. The only requirement is anhydrous solvents such as toluene, THF or acetonitrile. The
imidazolide is not the most reactive of acylating agents but it can usually be pushed to react. Several references to reviews on CDI can be found by
searching on Google or Google Scholar.
Also, refering to Scheme 4, once the rearrangement to the isocyanate (eg 5) is complete, t-butanol can be added to obtain the N-Boc derivative which
is often easier to purify than the amine itself. The Boc group can then be cleaved by a strong acid such as easily available 90% formic acid. Just a
thought.
AvBchemplayer.. - 26-2-2016 at 23:00
Actually we have a bit of PCl3 and we could in theory use this to convert to the acid chloride, but the problem then is how to purify this as an
intermediate (SOCl2 would be nice, but outside of our capability/willingness to produce and not available to buy). We could dissolve the ibuprofen in
DCM and then treat with PCl3 (the phosphorus acid forming a separate layer) but in practise you do seem to end up with weird phosphorus containing
compounds in the product unless you can distill it. We don't have the capability to distill ibuprofen chloride!
On the Lossen rearrangement - that's really interesting so thanks for that. A reaction we'd never heard of. Unfortunately hydroxylamine and CDI are
required for the reaction though; the hydroxylamine as a first step, and then the CDI to effect the rearrangement. Hydroxylamine.HCl is no problem.
CDI is a bit more sophisticated and difficult to hunt down.AvBaeyer - 27-2-2016 at 16:46
CDI is currently available from a reputable seller on ebay. There is no need to be worried about buying it.
By the way, CDI is used in the reaction to form the hydroxamic acid and then again in the rearrangement. The hydroxamic acid can also be made via the
carboxylate ester. I have made the hydroxamic acid of ibuprofen from the methyl ester as well as using CDI. I will write these experiments up and post
in a separate thread in a day or two.
AvBDr.Bob - 27-2-2016 at 18:55
AvBaeyer's method sounds very good. I use CDI for amide couplings often, it is great, very clean workup and cheaper than many other coupling agents.
And the idea of making the Boc derivative is a good one, they are great for purification. Removing the Boc can be done with TFA/DCM or better yet
with HCl in any of several solvents. Look forward to seeing your prep.chemplayer.. - 27-2-2016 at 20:32
Yeah our problem is that glassware/equipment we can import but chemicals are damn near impossible due to some very 'complex' customs regulations. Try
to import 2kg of salt and it will get held up pending an import license. Just don't ask.
But thanks everyone for the really helpful advice. We will give Schmidt a go using toluene on a very small scale (0.2g) sometime next week using a
high dilution factor and see. Only worry we have is that if any of the (dangerous) intermediate compounds happen to be insoluble in the solvent then
they might precipitate out and then there's a big risk, so this is really what we want to check.
If this works then a video will hopefully be in the pipeline.Loptr - 15-5-2017 at 08:40
Yeah our problem is that glassware/equipment we can import but chemicals are damn near impossible due to some very 'complex' customs regulations. Try
to import 2kg of salt and it will get held up pending an import license. Just don't ask.
But thanks everyone for the really helpful advice. We will give Schmidt a go using toluene on a very small scale (0.2g) sometime next week using a
high dilution factor and see. Only worry we have is that if any of the (dangerous) intermediate compounds happen to be insoluble in the solvent then
they might precipitate out and then there's a big risk, so this is really what we want to check.
If this works then a video will hopefully be in the pipeline.
Was there any follow up to this? I am also interested in attempting a schmidt reaction, and have just begun my first steps in researching.clearly_not_atara - 15-5-2017 at 12:16
I wonder if trifluoroacetyl bromide can be made from trifluoroacetic acid and acetonitrile with HBr:
MeCNH2Br+ + CF3CO2H >> MeCONH3+ + CF3COBr (g)
which appears justifiable on the basis of bond energies and acidity of the reactants (amide being the strongest base). The electrophile is a n
N-protonated imidoyl chloride RC(Cl)=N+H2. Trifluoroacetyl bromide is preferable because the chloride won't condense.
This is chemplayer's acetyl chloride but I used HBr here because CF3COCl has a very low boiling point. The resulting mixed trifluoroacetic anhydrides
are activated acids afaikchemplayer... - 16-5-2017 at 01:39
In the end we couldn't get it to work using toluene and we never dared to try with DCM. We were probably being far too cautious with our approach
though.
When time allows at some future date we'll try again perhaps using DCM and a robotic lab assistant (to match the voiceover!).Loptr - 27-7-2017 at 06:51
The paper posted above by DFJ90 (thanks!) tells what you need to know to safely convert a carboxylic acid to an amine - see Scheme 4.
Carbonyldiiimidazole (CDI) is quite a safe activation reagent and is usually available on ebay as is hydroxylamine hydrochloride. I use CDI routinely
for a variety of reactions requiring carboxylic acid activation. The only requirement is anhydrous solvents such as toluene, THF or acetonitrile. The
imidazolide is not the most reactive of acylating agents but it can usually be pushed to react. Several references to reviews on CDI can be found by
searching on Google or Google Scholar.
Also, refering to Scheme 4, once the rearrangement to the isocyanate (eg 5) is complete, t-butanol can be added to obtain the N-Boc derivative which
is often easier to purify than the amine itself. The Boc group can then be cleaved by a strong acid such as easily available 90% formic acid. Just a
thought.
AvB
I am assuming that something like n-methylhydroxylamine couldn't be used in this transformation because the intermediate isocyanate wouldn't be able
to form. Is that correct?
I was thinking about how this could be used to arrive at secondary amines, but then realized if an isocyanate was the intermediate, then it could be
possible to reduce them directly. Not sure how that could be accomplished at this time, or if it would even be practical. Maybe someone has some
references on this?
[EDIT]
It looks like diborane will react with isocyanates, such as phenyl isocyanate, to produce N-methylanilinoborane derivatives, which dimerize to higher
order boranes forming cyclical structures. Are these able to be hydrolyzed to their parent compounds? I am not too familiar with boron chemistry, but
this was the first result that popped up on Google. My initial thought was that NaBH4/H2SO4 could be used to effect the reduction.
[EDIT]
I found a JOC article that references the reduction of isocyanates to N-methylated amines using LAH (surprise!), if that matters because I am hesitant
to acquire LAH for fear of the humidity in my area.
Quote:
The Reduction of Isocyanates and Isothiocyanates with Lithium Aluminum Hydride
A. E. Finholt, Charles Dean Anderson, C. L. Agre
J. Org. Chem., 1953, 18 (10), pp 1338–1340 DOI: 10.1021/jo50016a012
Publication Date: October 1953
[Edited on 27-7-2017 by Loptr]clearly_not_atara - 27-7-2017 at 16:23
I'm curious: what is desirable in a solvent? Chlorobenzene is far less electrophilic than chloroalkanes, but much more polar than toluene (and close
to DCM), so I think that might work.
