ACS Catalysis
Research Article
The proton transfer from the weakly bound AB to amido-N is
a high energy process. The respective transition state (TS-I)
for proton transfer exhibits an energy barrier of 21.5 kcal
mol−1. After proton transfer, the deprotonated AB binds to
Mn(I) in an η1 fashion, giving intermediate III where the
amido linkage has been fully converted to the amine. This
assertion is fully reflected in the Mn−N bond length (2.10 Å)
in optimized intermediate III, which has been changed from
1.87 Å in I.
To facilitate the β-hydride elimination, the thiomethoxy arm
of the ligand decoordinates and creates room across the
manganese center, generating intermediate IV, where AB is
bound to manganese through the nitrogen end. Intermediate
IV slightly rearranges and stabilizes it further by creating a β-
agostic interaction with the manganese, leaving intermediate
IVA. Notably, this is the preparatory stage of β-hydride
elimination, where a clear agostic interaction of B−H (B−H
bond, closer to Mn, is elongated to 1.32 Å, while the Mn−H
distance 1.71 Å) is noticed. It seems that an actual kite-like
transition state to initiate β-hydride elimination lies on a very
shallow potential energy surface and eluded optimization,
despite multiple trials. However, due to the flat nature of the
potential energy surface during this step, the barrier for such a
TS is not high and will not dictate the significant energy
demand of the reaction.26 So, the inner-sphere pathway poses a
barrier of 21.5 kcal mol−1.
Interestingly, we were able to optimize the TS for β-hydride
elimination (TS-IIA, not shown in PES) when the
thiomethoxy arm is bound to manganese (a seven-coordinate
species), and it poses a large barrier of 27.9 kcal mol−1 so that
this pathway can safely be discarded. After β-hydride
elimination, the thiomethoxy arm rebounds to engender
intermediate V, which is a six-coordinate Mn-hydride. The
alternative, outer-sphere pathway traverses through TS-V with
a barrier of 20.0 kcal mol−1. The TS is concerted in nature but
highly asynchronous in transferring two hydrogens to nitrogen
and manganese, respectively. From the energy values for
dehydrogenation of AB, we conclude that the outer-sphere
pathway is slightly favored over the inner-sphere one, although
the energy difference is not large.
This observation is also supported by the experimental
determination of the kinetic isotope effect when AB was
appropriately labeled with deuterium (Figure 4). The overall
kinetic isotope effect KIE (kNH·BH/kND·BD = 3.2) with the
doubly labeled substrate is the product of KIE1 (kNH·BH/kNH·BD
= 2.4) and KIE2 (kNH·BH/kND·BH = 1.5), while KIE1 reflects the
effect of deuterium incorporation in BH3, and KIE2 addresses
the impact of deuterium incorporation in NH3. The value of
the overall KIE suggests that both B−H and N−H bond-
breaking contribute to the rate-determining step, and that is
likely if the outer-sphere mechanism is operative. Indeed,
multiple previous AB dehydrogenation studies have proved the
validity of the outer-sphere mechanism by carefully labeling the
substrate and showing that both B−H and N−H bond-
breaking influence the total rate.23 Such a competitive scenario
between both inner-sphere and outer-sphere mechanisms have
also been observed previously for a bifunctional iridium
catalyst, comprising a N-heterocyclic carbene ligand with a
tethered alcohol arm.27
In the second phase of the reaction, redelivering the stored
hydrogen in I starts with opening up the thiomethoxy arm to
create a vacant site for substrate binding, resulting in
intermediate VI. This ligand detachment costs a penalty of
11.6 kcal mol−1. The coordination of benzonitrile confers some
stability, and bound intermediate VII is stabilized by 7.0 kcal
mol−1 compared to intermediate VI. Insertion of a nitrile into
the Mn−hydride bond is a relatively facile process and
overcomes a barrier of only 19.0 kcal mol−1 via TS-III. Upon
nitrile insertion, reattachment of the thiomethoxy arm leads to
intermediate VIII, which releases the imine after protonation.
Notably, the imine formation marks the completion of the first
stage of reduction. The proton migration happens from N−H,
and the process is very facile, posing a barrier of only 4.6 kcal
mol−1 through TS-IV (to the energy of intermediate VIII).
Such a shallow energy barrier for protonation is associated with
a low-barrier hydrogen bond,28 which we have previously
observed during alcohol dehydrogenation by Mn1.9 So, the
major barrier during the nitrile hydrogenation is the insertion
step in manganese hydride, posing a barrier of 19.0 kcal mol−1.
A close look at the alternative pathway of the outer-sphere
hydrogenation of the nitrile reveals an asynchronous concerted
TS-VI, posing a barrier of 17.5 kcal mol−1. This transition state
for the hydrogenation maintains a microscopic reversible
feature to the outer-sphere transition state for dehydrogenation
that follows an asynchronous concerted mechanism.29
Although the computed energy demand for these outer-sphere
and inner-sphere hydrogenation processes are competitive, we
feel the nitrile binding via the inner-sphere pathway is
important. Additionally, as credible proof of the importance
of substrate binding to the metal center, we observe the
competence of the thiomethoxy arm to display hemilabile
nature as shown in Table 1 (26% yield of 2a using Mn4 as a
catalyst in comparison to 92% yield using Mn2). In fact, the
addition of strong Lewis bases, such as pyridine, 4-picoline,
and triphenylphosphine, poisons the catalyst easily, drastically
reducing the amount of product (18−27% yield of 2a, Scheme
2), supposedly by blocking the vacant site for nitrile binding to
manganese during the inner-sphere hydrogenation process.
Scheme 2. Control Experiments To Prove the Importance
of the Hemilabile Arm in the Ligand
Figure 4. Comparison of the initial rates of the Mn2-catalyzed
formation of 2a using H3N·H3B and its deuterium isotopologues as
the hydrogen source.
2791
ACS Catal. 2021, 11, 2786−2794