Journal of the American Chemical Society
Article
(5b), pentoxifylline (5d), and fluoxetine (5e) effectively
underwent hydroamination, affording corresponding desired
products. This methodology was also applied to more complex
alkene substrates derived from docosahexaenoic acid (DHA)
(5c) and cholesterol (5f and 5g), and these substrates could
undergo hydroamination with a loratadine-derived amine
transfer reagent and a monoalkyl-substituted amine source to
yield corresponding desired products.
As amines are ubiquitous and the most versatile functional
groups for subsequent transformations and diversification, we
subsequently turned our attention to the scope of amine-derived
picolinamide (PA) alkenes to demonstrate the synthetic utility
of the current method.14e To our delight, the desired aminated
products were obtained with comparable reactivity under the
standard conditions, which enabled the proximal selective
hydroamination of various classes of δ,ε-unsaturated alkenes, as
shown in Table 4. Beyond the δ,ε-unsaturated alkenes, this
method could be expanded to the γ,δ-unsaturated alkenes,
effectively yielding the corresponding product. In addition,
structurally complex loratadine could be effectively installed to
afford the desired product 7k. Therefore, this NiH-catalyzed
system significantly broadens its potential and general utility by
providing a new platform to deal with the challenging and broad
substrates with excellent levels of site selectivity and reactivity
under the mild reaction conditions.
To elucidate the mechanism, control experiments were
conducted, as depicted in Scheme 2. First, to investigate the
features of the hydrometalation step, we performed deuterium
labeling experiments by subjecting δ,δ-dideuterated alkene d2-1a
to the standard reaction conditions (Scheme 2a). The analysis of
1H NMR spectroscopy confirmed that the deuterium atoms at
the δ-position remained intact, and no H/D exchange was
observed at the γ-position. Additionally, we performed a
crossover experiment with 1j and deuterium-labeled olefin d2-
1a as substrates and observed that products 3j and d2-3a were
produced (Scheme 2b). These results indicate that olefin
dissociation may not occur prior to the formation of the desired
product.19 Next, to investigate the origin of the hydrogen atom
source, an isotope labeling experiment was conducted with
Ph2SiD2 under the standard reaction conditions (Scheme 2c).
Importantly, we observed deuterium incorporation at the δ-
position, which supports that the hydrogen is provided by silyl
hydride from the generated NiH species. Next, kinetic isotope
effect (KIE) experiments were performed with Ph2SiH2 and
Ph2SiD2, in which the observed kH/kD value was found to be 1.57
(Scheme 2d). Moreover, an inverse secondary KIE was observed
with 1a and d2-1a, in which the kH/kD ratio was 0.86 (Scheme
2e). These results indicate that the migratory insertion step is
involved in the turnover-limiting step.
To further investigate the reaction mechanism in the catalytic
system, in-depth kinetic studies using reaction progress kinetic
analysis (RPKA) experiments20 were conducted, and reaction
progress data in the NiH-catalyzed hydroamination reaction are
presented in Figure 1. First, we established standard conditions
for RPKA using 4-pentenoic acid derivatives 1a, morpholino
benzoate (2a), and diethoxymethylsilane with dimethyl
terephthalate as an internal standard. The initial concentration
of alkene 1a in the same-excess 1 represents the situation where
alkene 1a reached a 50% conversion at standard conditions. In
the same-excess 2, product 3a was added to the reaction mixture
of the same-excess 1. On the one hand, in our experiments, the
rate profiles of the same-excess case 1 and 2 overlapped, which
indicates that the generated product does not inhibit the
reaction in a significant way. On the other hand, there was no
overlay at the rate profile between a standard run condition and a
same-excess 2 when time-adjusted, indicating that the
concentration of the on-cycle catalyst is decreased in this
catalytic reaction.
Next, different-excess experiments were conducted to
determine the rate orders of the reaction components. On the
one hand, different initial concentrations of morpholino
benzoate (2a) and silane did not change the reaction rate,
indicating zero-order kinetics (see Supporting Information for
detail). On the other hand, there is no overlap between the
different initial concentrations of Ni catalyst and alkene 1a in the
plotted graph over a function of time (Figure 1B). To visualize
the direct comparison of profiles between the reaction rate and
́
concentration of Ni catalyst and alkene 1a, the Bures graphical
rate analysis21 was employed by plotting the concentration of
the product over normalized time·[component]n, as shown in
Figure 2. Notably, there are significant overlaps with different
loading changes of each Ni catalyst and alkene 1a (Ni: n = 0.62
and 1a: n = 0.42), which means that this reaction has a rate order
of 0.62 for Ni catalyst and 0.42 for alkene 1a, respectively. The
observed fractional order of Ni and alkene 1a (near 0.5) showed
the possibility of the presence of dimeric nickel complex Ni-1a as
an inactive off-cycle resting state, and NiH dimer would undergo
reversible dissociation to the corresponding active NiH
monomer prior to turnover-limiting migratory insertion.20b,22
To further support the generation of off-cycle bridged NiH
dimers, we conducted density functional theory (DFT)
calculations, which supported the possibility for the mono-
mer−dimer equilibrium, as illustrated in Figure 3.
We next turned our attention to investigate the electronic
effects of the amine electrophile components on the reaction
rate via Hammett parameter analysis.23 A series of amine
electrophiles bearing different para-substituted benzoates were
prepared and subjected to the standard reaction conditions.
Similar reaction rates were observed, regardless of the electronic
properties of benzoates, indicating that the oxidative addition
step and NiH regeneration step are not significantly involved in
the turnover-limiting step of this transformation (see the
Supporting Information for details). The zero-order depend-
ence on the silane concentration as shown in Figure 1B is also
consistent with these observed results.24
To gain further insight into the mechanism and the regio-
determining step, DFT calculations were performed using
alkene 1a and morpholino benzoate (2a) as the model reactants,
and the calculated reaction energy profiles of the Ni-catalyzed
hydroamination are illustrated in Figure 3. Regarding redox
manifolds of Ni catalysis, Ni(0)/Ni(II) and Ni(I)/Ni(III)
catalytic systems could be considered as possible pathways.
However, the energy barrier of the reductive elimination step in
the Ni(0)/Ni(II) system is too high to be mechanistically
plausible in this current reaction (>50 kcal/mol), ruling out the
reaction pathway involving Ni(0)/Ni(II) catalysis. Therefore,
we performed a DFT calculation based on the Ni(I)/Ni(III)
catalytic system. The catalytic cycle is initiated by the formation
of active Ni(I) hydride species A from the A-dimer, which is
generated from hydrosilane and the bidentate directing group-
tethered alkene 1a.22c,25 The intermediate A then undergoes
migratory insertion to form alkylnickel intermediate B. Our
calculations support that this migratory insertion step is the
regio-determining step, leading to the formation of two possible
nickellacycles. The transition state A-TS that leads to the six-
membered nickellacycle B is found to be 10.0 kcal/mol lower in
G
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX