The Hive > Serious Chemistry

Discussing the Mannich reaction

(1/3) > >>

High Bees  ;) ,

I 've been wondering a bit about the general applicability of the Mannich reaction procedure in Patent US6538136.  Me thinks this happens in example 2:

Flask 1:
1 mol Bzl-NH2 + 2,25 eq. CH2O  --> Bzl-NH-CH2-OH  + 1,25 eq. excess CH2O (don't know if imine will be formed without acid, since the optimal pH for this is 4 - 5)

Flash 2:
1,06 eq. CH3-CO-CH-CH(CH3)2 + 1,11 eq. HCl --> CH3-COH-C=C(CH3)2 + 1,11 eq. HCl (acid catalysed keto-enol equilibrium, notice the thermodynamic enol will be generated predominantly)

Drip flask 1 in flask 2:
Bzl-NH-CH2-OH  + 1,25 eq. excess CH2O + CH3-COH-C=C(CH3)2 + 1,11 eq. HCl  --> iminium ion formed in situ Bzl-NH(+).Cl=CH2 + H20 & reacts with enol  --> Bzl-NH-CH2-C(CH3)2-CO-CH3 + HCl regenerated & reacts with just formed secundary amine or protonates another aminoalcohol

Add 1,1 eq. iPr2EtN to freebase the secundary amine  --> Bzl-NH-CH2-C(CH3)2-CO-CH3 + iPr2EtN.HCl + 1,25 eq. + 0.3 eq.(yeah, why not add some more) CH2O --> Bzl-N(CH2-OH)-CH2-C(CH3)2-CO-CH3

Now comes the tricky part:
We have Bzl-N(CH2-OH)-CH2-C(CH3)2-CO-CH3, 1,1 eq. iPr2EtN.HCl (and 0,01 eq. HCl, but in example 1 there is  more base used to neutralize the HCl, so no HCl is present, just Et3N.HCl and 0.05 eq Et3N) with an excess CH20 in a H2O / EtOH solution

Me thinks the aminoalcohol gets protonated first with the HCl liberated from iPr2EtN.HCl  iPr2EtN + HCl.  At the same time iPr2EtN is generated which supposedly abstracts a proton from the CH3-CO-C(CH3)2-R ketone.

Now, the problem with the general applicability of this patent is that a weak base in a protic solvent gives the thermodynamic enolate (the one with the least H 's attatched to it).  To generate the 3,3-dimethyl piperidone this is no problem, since only the kinetic enolate can be formed.  But with 3-methyl piperidone (and even worse with ordinary piperidone), the wrong enolate will be formed.  A sterically hindered base is used, but do you think that will solve the problem of the inherently favored thermodynamic enolate formation?  It makes me rather suspicous, since they don't demonstrate the general applicability of the patent with some other examples ... (mabey the yields sucks)

This is how me thinks the reaction proceeds further:
Bzl-N(CH2-OH)-CH2-C(CH3)2-CO-CH3 + iPr2EtN.HCl  --> Bzl-N(+).Cl(=CH2)-CH2-C(CH3)2-CO-CH2(-) + iPr2EtNH(+)  --> N-benzyl-3,3-dimethyl-4-piperidone + iPr2EtN.HCl regenerated

I think an excess of formaldehyde is used to shift the equilibrium in the favorable way, but mabey this isn't a good idea for other piperidones than the 3,3-dimethyl ones.

What do you bees think about the proposed reaction mechanism?  It is rather amazing that this all takes place in one pot  8) .  Do you also forsee problems for other (less 3 substituted) piperidones.

Another problem with this procedure is that primary amines can bee dialkylated.  I think this problem is minimized by slowly dripping the aminoalcohol / excess formaldehyde in the ketone / HCl solution ... even though I am wondering whether the excess formaldehyde is really necessary.

Mabey the use of a stoichiometric amount of formaldehyde can solve the problems for the less 3-substituted piperidones.  Particulary with the second mannich reaction: freebase the secondary amine, let it react with 1 eq. formaldehyde   --> iminium formation & mabey bring the pH back to 4 - 5.  Only the kinetic enol can react with the iminium & no residual CH2O is left in the mixture.

In some chemical encyclopedia, they say that in the reaction (CH3)2NH + (CH3)2CH-CO-CH3 + formaldehyde, the thermodynamic product is formed in a 3,5 / 1 ratio, relative to kinetic mannich base (oddly, no further references were given ...).  What worries me is: could the  (excess) formaldehyde react with the Bzl-N(+).Cl(=CH2)-CH2-C(CH3)=COH-CH3 or Bzl-N(+).Cl(=CH2)-CH2-CH=COH-CH3 enol (in a 3-methyl-4-piperidone and 4-piperdione synthesis respectively), which will be formed predominantly.

If this reaction could be (made) generally applicable, it would bee the easiest piperidone synthesis up to date.

Are these mechanistic worries justified, or is this bee totally screwing up here?  I guess it is all a matter of equilibrium & pH.

Paraformaldehyde is a polymeric acetal which cleaves under acidic conditions.

As I understand it, the reaction is acid catalysed (so the acid will be regenerated).  I think it will be favorable to do the reaction in absolute ethanol: due to the equilibria in the iminium formation and le Chatelier ...

But, the dropwise addition of the aminoalcohol to the ketone / acid  - ethanol solution is a very good idea, since in this medium:
 1) the enol forms (acid catalysed)
 2) the aminoalcohol, which is dripped in, is converted in situ to the iminium salt (consumes 1 eq. of acid, but it is regenerated by the addition of the enol to the iminium salt, and the liberated acid is captured then by the formed secundary amine.)
 3) the in situ formed imine reacts with an excess of enol.  Since no excess formaldehyde is used, dialkylation is less likely to occur

The Mannich reaction on primary amines has been done with paraformaldehyde in ethanol has been done in
Patent DE379950 example 5: I particulary like that one, since equimolar amounts of methylamine.HCl and paraformaldehyde are used in alcohol.
Patent DE542253 example 5

Another inspiring articles are:

I guess here the amine salt dissociates and the formaldehyde is formed in situ (grace à the liberated HCl) and the amine reacts with it.
There 's another nice iminium ion formation procedure in:
(the Freifelder article).  Here the iminium salt is generated at last (with a 1 eq. addition of acetic acid).  It seems that the amine reacts with the paraformaldehyde even without the addition of acid (but I seem to remember that bases react with the polymeric acetal to a lesser extent than acids).

Would it be a good idea to
1) formation of the aminoalcohol in absolute EtOH, with a bit of acid to catalyze the reaction (trifluoroacetic acid or HCl)
2) add the above solution dropwise to a absolue ethanol solution with (at least 1 eq.) of TFA or HCl

Since the entire process is acid catalysed (please correct me on this if I am wrong), I 'd make sure that the pH is between 4 - 5 when all the aminoalcohol is added.

Don't hesitate to participate in this thread.  All replies (critical remarks, other proposed routes, answers towards uncertainties, ...) will be appreciated  :) .

I guess here the amine salt dissociates and the formaldehyde is formed in situ (grace à the liberated HCl) and the amine reacts with it.

The same methodology is used here (synthesis of mannich bases 6 a - d)
It seems to take some time beefore te acetal gets depolymerisised.

An interesting example of Mannich cyclization is the synthesis of myrtine:
Tetrahedron Letters (1978), p 397 (has to bee checked out).

This sounds like being one of those comprehensive review article, known to give the dedicated chemist an orgasm:

Tramontini,M.  Advances in the chemistry of Mannich bases.  Synthesis (1973) p 703 - 775


[0] Message Index

[#] Next page

Go to full version