Imperial College London

Henry Rzepa

The story [so far](http://www.ch.imperial.ac.uk/rzepa/blog/?p=16844).
Imines react with a peracid to form either a nitrone (σ-nucleophile) or
an oxaziridine
(π-nucleophile).\[cite\]10.1016/S0040-4039(00)98582-4\[/cite\] The
balance between the two is on an experimental knife-edge, being strongly
influenced by substituents on the imine. Modelling these reactions using
the "normal" mechanism for peracid oxidation did not reproduce this
knife-edge, with ΔΔG (π-σ) 16.2 kcal/mol being rather too far from a
fine balance.

There are two general reasons why computational modelling using quantum
mechanics may not match experimental outcome. Until perhaps 10 or so
years ago, the culprits may often have been the approximations necessary
to apply the theory, as bounded by the limitations of the CPU power of
the then available computers to evaluate the associated equations.
Nowadays, an equally likely explanation is that the molecular model for
which these equations are solved is either wrong or maybe just
incomplete. For an organic reaction, these models are initially set out
by "arrow pushing" a possible mechanistic pathway. Such speculations
have been a common feature of most new articles reporting the outcome of
reaction experiments for perhaps 60 years now. It is now more common
(but by no means universal) to augment this with a computational reality
check. So
[previously](http://www.ch.imperial.ac.uk/rzepa/blog/?p=16844), when I
applied a reality check on the "standard" epoxidation mechanism, it did
not pass the test.

So time to revise the mechanism, as per below. The difference is that
the model includes an extra water molecule to facilitate proton
transfers, with the imine now being protonated by the peracid to form a
zwitterion, which collapses to an addition product and it is this
species that rearranges to the final oxaziridine. Free energies relative
to the reactant **1** are shown in red below.^‡^

[![azir](http://www.ch.ic.ac.uk/rzepa/blog/wp-content/uploads/2016/09/azir.svg){.aligncenter
.size-large .wp-image-16915 decoding="async"
width="400"}](http://www.ch.ic.ac.uk/rzepa/blog/wp-content/uploads/2016/09/azir.svg)

The IRC for 2 (TS) is shown below, being a proton transfer mediated by
the transfer agent (water in this case, but it could be also peracid or
eventually the product acid) to form a ion-pair.

![2ts-irc1](http://www.ch.ic.ac.uk/rzepa/blog/wp-content/uploads/2016/09/2TS-IRC1.gif){.aligncenter
.size-full .wp-image-16904 decoding="async" width="450"}

4 (TS) shows the collapse of the ion-pair to form an addition product
across the imine.

![4ts-irc1](http://www.ch.ic.ac.uk/rzepa/blog/wp-content/uploads/2016/09/4TS-IRC1.gif){.aligncenter
.size-full .wp-image-16908 decoding="async" width="450"}

6 (TS, below) is the most interesting and also the high point on the
free-energy pathway (*i.e.* the rate determining step). The addition
product cyclises to an oxaziridine as induced by the nitrogen lone pair
helping to evict the acetate anion. This is followed at IRC \~7 by a
transfer of the N-H proton back to the carboxylic acid, again using
water as a transfer agent with the whole being part of a concerted but
asynchronous mechanistic step.

![6ts-irc1](http://www.ch.ic.ac.uk/rzepa/blog/wp-content/uploads/2016/09/6TS-IRC1.gif){.aligncenter
.size-large .wp-image-16905 loading="lazy" decoding="async" width="450"
height="430"}

[![6g](http://www.ch.ic.ac.uk/rzepa/blog/wp-content/uploads/2016/09/6G.svg){.aligncenter
.size-large .wp-image-16917 decoding="async"
width="450"}](http://www.ch.ic.ac.uk/rzepa/blog/wp-content/uploads/2016/09/6G.svg)

Crucially, 6 (TS) is 23.4 kcal/mol below the oxaziridination transition
state modelled without a prior proton
transfer\[cite\]10.14469/hpc/1698\[/cite\],\[cite\]10.14469/hpc/1701\[/cite\]
and even 7.6 kcal/mol below the transition state for nitrone
formation.\[cite\]10.14469/hpc/1697\[/cite\],\[cite\]10.14469/hpc/1702\[/cite\]

So the original mechanism is now replaced by an alternative, which
really only differs in the timing of how the acidic proton attached to
the peracid responds to the process. By getting actively involved prior
to the crucial reaction with the nitrogen lone pair of the imine, this
proton enables a lower energy route to be established. We are now ready
for the next "reality check" on these mechanisms, which are the effects
of substituents on the imine. If these can be replicated, we can then
really start to claim that computation has put the mechanism of this
reaction onto a firmer footing than that based just on "arrow-pushing".

------------------------------------------------------------------------

^‡^Calculations (ωB97XD/Def2-TZVPP/SCRF=dichloromethane) for the species
above are archived as a collection at DOI:
[10.14469/hpc/1704](http://doi.org/10.14469/hpc/1704)\[cite\]10.14469/hpc/1705\>\[/cite\]
and individually at 1\[cite\]10.14469/hpc/1695\[/cite\], 2
(TS)\[cite\]10.14469/hpc/1703\[/cite\],
3\[cite\]10.14469/hpc/1696\[/cite\], 4
(TS)\[cite\]10.14469/hpc/1692\[/cite\],
\[cite\]10.14469/hpc/1700\[/cite\], 5\[cite\]10.14469/hpc/1690\[/cite\],
6
(TS)\[cite\]10.14469/hpc/1694\[/cite\],\[cite\]10.14469/hpc/1693\[/cite\],
7\[cite\]10.14469/hpc/1691\[/cite\].

::: {.wp_orcid_field}
[https://orcid.org/0000-0002-8635-8390](https://orcid.org/0000-0002-8635-8390){target="_blank"
rel="author"}
:::


The story so far. Imines react with a peracid to form either a nitrone (σ-nucleophile) or an oxaziridine (π-nucleophile).[cite]10.1016/S0040-4039(00)98582-4[/cite] The balance between the two is on an experimental knife-edge, being strongly influenced by substituents on the imine.

Messaggi di Rogue Scholar

σ or π nucleophilic reactivity of imines? A mechanistic twist emerges.