Sciencemadness Discussion Board

Sodium Diformylamide for primary amines(and secondary)

karlos³ - 24-5-2020 at 00:46

Ok, so I got interested in that gabriel reagent a while ago, as it seemed so easy to work with.
It can be attached relatively mild, and can be removed under really mild and simple conditions.
It even works well for the rather instable aminoketones, as one ends up with the relatively pure HCl salts which are stable in that case.
And the best about this substance is, that the preparation is really simple and cheap, provided one has the reagents or can buy them, otherwise it is not so easy(keep in mind, formamide has to be redistilled 5-7 times until it is pure).
The preparation of it is usually in quantitative yields though.
The Lithium salt of it is known, adding some more OTC-availability to it.

One can react sulfonylate esters, haloalkanes or haloketones with it without a problem.
There is just one downside of it: either DMF or acetonitrile are needed for the SN2 reaction.
But as soon as one was able to make the diformylimido compound, the deprotection is very simple, just a reaction with conc. HCl in ethanol, where the concentration of the acid in the alcohol determines the course of the reaction.
It can be run either very mild with a low concentration at RT for two days, or under warming with a higher concentration in just a few hours.
I prefer the former as it is much milder.
Afterwards one distills the volatiles off, and then adds dry acetone to the residue in the still pot, which will after some cooling, result in a relatively clean precipitate of the HCl salt of the product.

As for the secondary amines that can be possibly made with this: as the attached paper shows, the deprotection will if done with a hydroxide and without water, result in a monoformylamine, a formamide.
This is useful as it can be reduced with NaBH4 combined with certain other reagents, for example NaBH4/NiCl2 resulting in nickel boride, or NaBH4/CH3COOH resulting in sodium triacetoxyborohydride(both are made in-situ), which will result in the reduction of the formyl into a methyl group.

Here is an example of the preparation and use of this gabriel reagent:

Quote:

- Sodium Diformylamide
A mixture of
- 17,1g/15,2ml formamide(0,4mol) and
- 10,26g NaOMe in MeOH(0,19mol), (freshly made with -
4,45g Na(0,19mol) and
- 40ml MeOH) and is stirred at RT for 1 h.
The solvent is stripped off, then 2x30ml more MeOH were added through the add. funnel while distilling it off, which aids in removal of the ammonia. Finally, 40ml toluene are added and used likewise, but only half of that is distilled off, then it was allowed to cool down and the whole solids were filtered(this is where product loss happened for me, not cold enough).
The crystalline solid obtained is now pure enough for the next reaction.
It is however recommended to powder it as fine as possible, to have a good reaction, since it is not really soluble in the reaction solvents.
yield 17,3g(0,182mol)

- Diformylamidopropiophenone
- 7g a-bromopropiophenone(33mmol) was dissolved in 30ml of DMF and
- 3,9g sodium diformylamide(41,2mmol) were added.
The mixture was heated for 4 h.
Post reaction, the mixture was simply added to 80ml of cold water and collected in 30ml DCM, washed with 15ml brine, dried(anhydrous sodium sulfate) and solvent removed to give a yellow oil, the N,N–diformylamidoketone(205,23g/mol) (4,3g, 21mmol, 56%).

- 2-Aminopropiophenone
This N,N–diformylamide derivative was dissolved in 5% ethanolic hydrochloric acid (50mL) and stirred overnight at RT, and left to stand for another day at RT.
The solution was then evaporated to dryness, acetone was added, and the precipitate was collected by filtration.
The intermediate was used directly without the need for further purification.
It had a weight of 2,37g/12,7mmol, 60% from diformylamidoketone.


With this reagent on hand, the gabriel reaction finally becomes an easy and preferable task to make primary amines, no side product, an easy synthesis without any annoying side products(like ammonia in the delepine for example) and overall pretty convenient.
I can really recommend it.

Attachment: sodium diformylamide gabriel reagent.pdf (234kB)
This file has been downloaded 546 times

Sigmatropic - 24-5-2020 at 08:20

I cannot go into details here but, the deprotection can also be done in the same solvent as the initial reaction, acetonitrile, with conc. HCl.
No need for a solvent swap to alcohol and then back to acetone.

Edit: According to the article the same is true while using ethanol, interesting!

Is the ammonium chloride really that annoying in the delepine reaction? It is insoluble in ethanol, whereas the product hydrochloride is not, right?

[Edited on 24-5-2020 by Sigmatropic]

karlos³ - 25-5-2020 at 16:08

Yes thank you, but in the specific case of an aminoketone as I described, the effort is warranted since you can't isolate the product as freebase.

Well, in my own experiments, I found the delepine to be much dirtier and harder to isolate a pure product from.
The ammonia was never purely discarded with the usual workup.
It just seemed to me to be overly complicated, while the gabriel with this certain reagent runs so smooth and easy, it was absolutely surprising how well and clean it has worked.

This is really something people should be aware of.
Besides its use being so clean and easy, it also has this easy and cheap preparation which all speaks for its use instead of any other gabriel reagent.
It doesn't hurt to have some amount of that stuff in the chemical cabinet :)

Methyl.Magic - 10-6-2020 at 15:05

Hey ! You really made my day by pointing out to this pretty awesome reagent.

It resolves the main problem of amino ketone : not going through basic environment to avoid self-dimerisation of substrate. 56% is not so high yield but pretty decent. Delepine can be good on primary substrate but shitty on secondary and does not work well on 'not active' halides. Did you tried delepine on this substrate ? (I mean secondary alpha-bromo ketone)

The name of this reagent is "gabriel" ? Like the phtalimide ?

Your reagent gives me an idea, because I want to take my vengence on the failed phenylpropanolamine via akabori : so a-bromopropiophenone + Na diformylamide > then NaBH4 ketone reduction > then HCl in EtOH, BAM !!!

karlos³ - 10-6-2020 at 16:47

It can be improved with proper work, the described trial was my second one on a bromoketone.
Prolonging the single reactions does improve a lot too.
I did not need to try the delepine on that substrate, because it does not work.
Many have already proven that on substrates like mine, and I didn't intended to reinvent the wheel including to fail, I mean, I already have the real wheel with this gabriel reagent ;)
Also I think the delepine is comparably dirty, in my opinion, to much fudging around and not as straight as this gabriel reagent.

The gabriel reaction can be done with a handful of reagents, of which phthalimide is just the first, and honestly, one of the absolutely worst too.
This here, is probably the best one, I know of only a second similar well working one... but this isn't that simply made, so Na-DFA is overall really the best(probably, I am always open for anything disproving that assumption :D).

Your idea is not a good one.
If you reduce the alcohol before the final amination, the borohydride will attack from the other side and you will end up with norpseudoephedrine in the end.
I was surprised to get mostly that when I reduced the bromoketone and reacted the bromohydrin with Na-DFA, as upon deprotection the purified product was consistent with the melting point of racemic cathine, norpseudoephedrine.
A friend offered that explaination then, and it actually indicates the same in very similar reaction(bromohydrin and methylamine yield pseudoephedrine mostly, bromoketone and methylamine ephedrine).

Methyl.Magic - 10-6-2020 at 17:49

norephedrine is my goal here :)

Ahh ok you are talking about the trans- isomer instead of the cis- ? Ok this is due to the inversion after na-DFA attacking bromo via SN2.

I also thing maybe the Na-DFA is too basic on the bromhydrin and with result with epoxide formation (?)

I would like to attack the bromoketone with Na-DFA, then ketone reduction with NaBH4 and finally go with HCl to precipitate norephedrine, now I have mostly the cis- isomer right ?

But finally I think I'll go with the azide route : bromoketone + NaN3 > azidoketone
azidoketone + LiAlH4 > norephedrine :)
or 2 step with NaBH4 reduction of ketone + NaBH4/Ni (or Cu) nanoparticles to norephedrine , what do you think ?

karlos³ - 10-6-2020 at 18:33

No I just made the norpseudoephedrine to play around with that reaction.
I don't understand the mechanism behind it completely with the inversion, I don't know if there even is one taking place?
I think not, and that it is, as I mentioned, due to the reduction but I can't explain it, only repeat what said friend was sure to have happened.
For a conversion/retention to take place, it must happen at a chiral part of the molecule.
And that clearly isn't happening, or at least clearly to me, maybe I'm wrong still?

Retention/inversion is something I've seen and done with chiral alcohols, which inverted their chirality when they are halogenated, but kept it when sulfonylated.
Azidation, and likewise sodium diformylamide, will then cause a retention on the halo-compound, back to the original chirality of the alcohol started from.
Doing the same on the sulfonyl ester, will result in inversion to the opposite chirality of the starting alcohol.
By the way reduction of the azide/deprotection of the diformylimide will not result in either of them, I mean inversion or retention.
The substitution leads to that, as that is a SN2 mediated effect.
I can't see how this should take place at this molecule, you must have misunderstood something about what I meant?.
The bromine is not a stereocenter, the carbonyl could become one, but this isn't where the action is.

He explained that in strictly logical way and I am convinced he is right... but its at least 2 years ago and I can't say more than, it must have been somehow connected with the reduction, in that it took place sideways and actually this is where I can only make helpless gestures and don't know better how to explain.

It is very funny that you mention the azidoketone!
Because I attempted an one-pot reduction to the amino alcohol, with borohydride and a copper salt... and same! Low yield, but composed of norpseudoephedrine instead of norephedrine.
I hope someone more knowledgeable chimes in an can actually find out what my friend meant when he told me this explaination?
But as said, since the ketone gets reduced first, and then the azide, it was the threo and not erythro mostly.

And straight LiAlH4 gives the racemic norephedrine instead?
Because this reducing agents proceeds otherwise, my friend even used that as an example for what he meant.

The order of the single reactions has quite a lot of influence on that, or rather, the reduction of the carbonyl is what has a lot of influence seemingly, but as said, helpless to explain that in detail, sorry.

And no, Na-DFA is not basic at all, it has nothing to do with epoxide formation.
That is a whole other mechanism responsible here.

As for cis and trans... norephedrine has neither, it has actually threo and erythro enantiomers, where threo means the norpseudo- enantiomer and erythro the norephedrine enantiomer.

But I get it, of course you must talk about that lousy oxazolidine, the one with an almost mythical reputation which in practice it can't keep up with..
I don't like that substance you are hinting it at all, boring, long, and despite being called euphoria, no euphoria... what a bad joke! :D

In the preparation with BrCN, the threo will give racemic trans-, and erythro racemic cis-, thats right.

But my own suggestion, if you are just after norephedrine, make the monoximino ketone from propiophenone and reduce that with the french nickelammine-zinc pseudo-Urushibara reduction from the 40s.
Here is the reaction used for something else: http://www.sciencemadness.org/talk/viewthread.php?tid=20153
And here is the specific reduction I mean described: FR971429
I did that once and the yield was surprisingly good, and all that in just two separate reactions from propiophenone.

[Edited on 11-6-2020 by karlos³]

karlos³ - 11-9-2021 at 14:16

Here is a collection of almost every single documented use of sodium diformylamide(I'm honest with you, there weren't much, so it was not very hard), including a few theses which have only a single reaction with it, but 400 or more pages...

But lots of useful informations contained regarding the yield of that reaction considerably.
Have a look for yourself, its around 40mb, mostly because of the theses, and thus too large for the forum: https://anonfiles.com/1756R1G8u4/kl_tenlimonade_rar