1. 4
-4.5
(-27.9)
PMe2
Si RhMe
PMe2
H
5
+2.3
(-28.5)
PMe2
Si RhMe
PMe2
H
H
H
PMe2
Si RhMe
PMe2
H
H
H
TS5-6
+13.7
(-18.4)
6
-9.5
(-39.8)
7
-12.1
(-32.1)
PMe2
Si RhMe
PMe2
H
H
PMe2
Si RhMe
PMe2
H
H
H
-18.2
(-36.9)
PMe2
Si RhMe
PMe2
H
H
2'
TS4-8
+7.0
(-16.2)
8
+2.1
(-21.1)
TS8-1
+6.0
(-16.4)
1'
-21.0
(-30.3)
PMe2
Si RhMe
PMe2
H
PMe2
Si RhMe
PMe2
PMe2
Si RhMe
PMe2
H
PMe2
Si RhMe
PMe2
H
nba
nba
H2
4
10.3
(10.9)
1
0.0
(0.0)
17.8
(7.6)
16.4
(4.9)
5
-3.1
(-15.6)
TS5-6
12.8
(-0.7)
6
3.5
(-9.6)
TS6-7
26.7
(14.4)
7
21.8
(10.3)
2
8.8
(-3.6)
30.2
(16.9)
3
-19.0
(-31.0)
RhSi
PMe2
PMe2
OTf
RhSi
PMe2
PMe2
TfO
RhSi
PMe2
PMe2
O C
O
TfO
RhSi
PMe2
PMe2
O C
O
OTf
RhSi
PMe2
PMe2
O C
O
OTf
RhSi
PMe2
PMe2
O C
O
TfO
RhSi
PMe2
PMe2
O C
O
TfO
RhSi
PMe2
PMe2
O C
O
TfO
TS2-3
RhSi
PMe2
PMe2
O C
O
OTf
RhSi
PMe2
PMe2
O C
OTf
O
RhSi
PMe2
PMe2
O
C
O
OTf
RhSi
PMe2
PMe2
O
TfO C
O
TS4-5
TS1-2
Computa(onal
Studies
of
Metal-­‐Catalyzed
Hydrogena(on
and
CO2
Ac(va(on
Michael J. Trenerry, Matthew T. Whited,* Buck L. H. Taylor*
Department of Chemistry, Carleton College, Northfield, MN
Introduc(on
Compe(ng
Hydrogena(on
Mechanisms
A@er
Branching
Point
Experimental
Studies
Financial Support
This work was supported by the National
Science Foundation and donors of the
American Chemical Society Petroleum
Research Fund
Acknowledgements
We thank Teddy Donnell and Kate DeMeulenaere
for work on complimentary experimental studies.
Computer Resources
Computational resources were provided by the
Midwest Undergraduate Computational Chemistry
Consortium cluster, which is supported by the
NSF (CHE-0520704 and CHE-1039925).
Rhodium and silicon form a polar bond where electron density is focused on the rhodium
metal center. The presence of chelating pincer ligands enhances this effect by forcing the
rhodium and silicon to remain in close proximity of one another. Rhodium-silicon pincer
complexes consequently contain a reactive site around the Rh-Si bond and can potentially be
used to initiate small molecule activation. Using computational methods (DFT), we have
studied the mechanism of rhodium-catalyzed hydrogenation of norbornene and the splitting of
CO2 gas across the Rh-Si bond.
After the branching point at 4, hydrogenation may proceed by one of two competing
mechanisms. The proposed σ-CAM mechanism, shown in black, involves the formation of a
second dihydrogen complex intermediate 5 and a four-centered transition state TS5-6 in the
rate-determining step. An alternative mechanism, shown in blue, requires movement of the
norbornane ligand in 4 to be cis to the hydride (TS4-8) in the rate-determining step prior to
reductive elimination (TS8-1). The reductive elimination mechanism requires a lower
activation energy relative to the σ-CAM process and is thus deemed more energetically
favorable, although the difference in activation energies between reductive elimination and
σ-CAM mechanisms is minor (ΔΔG=6.7 kcal/mol) and due to entropic effects.
Whited, Deetz, Donnell, and Jensen, Dalton Trans. 2016, 45, 9760.
Computational studies indicate norbornene hydrogenation through the reductive elimination
mechanism is more energetically favorable than through the σ-CAM mechanism. However,
the small difference in activation energies between these processes suggest that there is
not a strong preference for one particular mechanism versus another. Experimental studies
provide further evidence that both reaction mechanisms are accessible.
A labeling study performed by the Whited group observed a roughly 1:3:1 distribution of
hydrogenated norbornane products via GCMS. The addition of HD as a η2 ligand to the
rhodium center means this result is consistent with the statistical distribution expected from
a combination of the reductive elimination and σ-CAM mechanisms, in which η2 HD ligand
addition is respectively performed once and twice. Competition between unimolecular and
bimolecular pathways is the subject of ongoing computational and experimental studies.
Possible
CO2
Inser(on
Mechanisms
Mechanisms proceeding from the original complex (black) versus a silylene intermediate
(blue) are shown above. In both mechanisms CO2 is added across the rhodium silicon bond
through a [2+2] cycloaddition reaction and yields 4-membered ring intermediates 2 and 5. All
apparent steps leading to the formation of 3 have been identified for the pathway involving
the original complex. The exact process of triflate migration in the silylene pathway remains
unclear and is the subject of ongoing investigation. Elucidating the full range of potential
triflate migration steps involved in the silylene pathway will likely prove crucial in assessing
its viability as a favorable reaction mechanism.
Theoretical methods:
DFT calculations were performed in Gaussian 09. Density fitting was enabled:
B97D / def2-TZVP / W06 / SMD(diethyl ether) // B97D / def-2SVP / W06 (gas-phase)
Rh
Si
P
Cy2
OTf
P
Cy2
Rh
Si
P
Cy2
OTf
P
Cy21 atm CO2
diethyl ether, benzene, DCM
25o C
O
C
O
nba
1 atm H2
d6-Benzene
25o Cnbe
Rh
Si
P
Cy2
Me
P
Cy2
Hydrogena(on
Shared
Reac(on
Pathway
Both mechanisms follow the same reaction pathway from the entry point of the catalytic
cycle at 1 to the branching point at 4. This process involves the formation of a dihydrogen
complex 2 and the subsequent addition of norbornene as a η2 ligand to form 3. Migratory
insertion of norbornene yields 4, where the rhodium center forms sigma bonds to a hydride
and a norbornane ligand.
PMe2
Si RhMe
PMe2
H
H
PMe2
Si RhMe
PMe2
H
HPMe2
Si RhMe
PMe2
0.0
(0.0)
1
+6.9
(-16.7)
TS3-4
4
2
3
-6.4
(-7.8)
+4.9
(-19.6)
-4.5
(-27.9)
PMe2
Si RhMe
PMe2
H
H
PMe2
Si RhMe
PMe2
H
nbe
H2
PCy2
Si Rh
PCy2
OTf
Rh
Si
P
Cy2
OTf
P
Cy2
Rh
Si
P
Cy2
OTf
P
Cy2 O
C
O
CO2
possible intermediate
The triflatosilyl pincer complex reacts
with CO2, but the methylsilyl relative
does not. This suggests the reaction
depends on the lability of the triflate
group and may proceed via a silylene
intermediate. Computational literature
investigating similar reactivity of an
iridium carbene complex also supports
this claim. Yates, Brookes, Ariafard, and Stranger, JACS. 2009, 131, 5804.
Rhodium-silicon pincer complexes can
be used to drive the catalytic
hydrogenation of norbornene. The
catalytic cycle is hypothesized to yield
norbornane via a reductive elimination
or sigma complex-assisted metathesis
(σ-CAM) process.
Our computational studies suggest
that the reductive elimination pathway
is more energetically favorable.
However, the difference in calculated
activation energies between pathways
is sufficiently small such that major
intermediates of both pathways are
potentially accessible. This is
supported by previous laboratory work
performed by the Whited group and is
being further investigated by recent
experiments.
Whited, Deetz, Donnell, and Jensen, Dalton Trans. 2016, 45, 9760.
PMe2
Si RhMe
PMe2
PMe2
Si RhMe
PMe2
H
H
PMe2
Si RhMe
PMe2
H
H
PMe2
Si RhMe
PMe2
H
H2
H2
PMe2
Si RhMe
PMe2
H
H
H
1
5
3
24
nba
nbe
+nbd
-nbd
nba
Rh
Si
P
Me2
Me
P
Me2
Migratory
Inser(on
Reduc(ve
Elimina(on
σ-­‐CAM
Proposed
Cataly(c
Cycle
of
Norbornene
Hydrogena(on
CO2
Inser(on
Reac(on
(CyPSiP)Rh(nbd)(Me)
diethyl ether
98% HD
~20% ~60% ~20%
H
H
D
H
D
D