Half-sandwich ruthenium catalyst bearing an enantiopure primary amine tethered to an N-heterocyclic carbene for ketone hydrogenation

Article abstract of DOI:10.1021/acscatal.7b02346

By using a copper transmetalation reagent [Cu-(Kaibene)₂]I, the NHC ligand (S,S)-MeNC₃H₂NCHPhCHPhNH2 "Kaibene" was transferred to ruthenium to make a precatalyst [RuCp^?(Kaibene)(MeCN)](PF6) (Cp^? = 1,2,3,4,5-pentamethylcyclopentadienyl), 7, in high yield as a mixture of two diastereomers. Under relatively mild conditions (0.02 mol % Ru, 0.16 mol % KOtBu, iPrOH, 50 °C, 25 bar of H₂), this compound catalyzes the hydrogenation of aryl ketones and one alkyl ketone effectively with excellent activity and productivity (TOF up to 48 s^-1, TON up to 10⁴). At higher hydrogenation pressure (46 bar), the catalytic hydrogenation of N-phenyl-benzylimine to the corresponding amine is efficiently achieved. The hydrogenation of prochiral ketones resulted in low ee (35% for 4-chloroacetophenone). NMR spectroscopy was used to observe diastereomeric hydrides RuCp^?(Kaibene)(H) 13-R/S that were generated by reaction of 7 with H₂ and base in THF-d_8. Complementary DFT studies suggest that either the heterolytic splitting of dihydrogen to form 13-R/S or the hydride transfer to the substrate can be rate-determining depending on the substrate. Experimental and computational results support mechanisms that involve the heterolytic splitting of dihydrogen to the nitrogen of the amide-ligated form of Kaibene in THF or the heterolytic splitting to an outer-sphere alkoxide derived from the product alcohol or 2-PrOH solvent. An unusual feature is the rapid drop in ee of the product alcohol from as high as 60% (R) to 0% in some cases; this might be due to racemization of the Kaibene ligand in THF caused by the strong base or competitive inhibition of one diastereomer of the catalyst by reaction with the product (R)-alcohol.

Full text of DOI:10.1021/acscatal.7b02346

Research Article
pubs.acs.org/acscatalysis
Half-Sandwich Ruthenium Catalyst Bearing an Enantiopure Primary
Amine Tethered to an N‑Heterocyclic Carbene for Ketone
Hydrogenation
Kai Y. Wan, Molly M. H. Sung, Alan J. Lough, and Robert H. Morris*
Department of Chemistry, University of Toronto, 80 Saint George Street, Toronto, Ontario M5S 3H6, Canada
*
S Supporting Information
ABSTRACT: By using a copper transmetalation reagent [Cu-
Kaibene) ]I, the NHC ligand (S,S)-MeNC H NCHPhCHPhNH
2
(
2
3
2
“
Kaibene” was transferred to ruthenium to make a precatalyst
[
RuCp*(Kaibene)(MeCN)](PF ) (Cp* = 1,2,3,4,5-pentamethyl-
6
cyclopentadienyl), 7, in high yield as a mixture of two diastereomers.
Under relatively mild conditions (0.02 mol % Ru, 0.16 mol % KOtBu,
iPrOH, 50 °C, 25 bar of H ), this compound catalyzes the
2
hydrogenation of aryl ketones and one alkyl ketone effectively with
−1
4
excellent activity and productivity (TOF up to 48 s , TON up to 10 ).
At higher hydrogenation pressure (46 bar), the catalytic hydrogenation
of N-phenyl-benzylimine to the corresponding amine is efficiently
achieved. The hydrogenation of prochiral ketones resulted in low ee
(
35% for 4-chloroacetophenone). NMR spectroscopy was used to
observe diastereomeric hydrides RuCp*(Kaibene)(H) 13-R/S that
were generated by reaction of 7 with H and base in THF-d . Complementary DFT studies suggest that either the heterolytic
2
8
splitting of dihydrogen to form 13-R/S or the hydride transfer to the substrate can be rate-determining depending on the
substrate. Experimental and computational results support mechanisms that involve the heterolytic splitting of dihydrogen to the
nitrogen of the amide-ligated form of Kaibene in THF or the heterolytic splitting to an outer-sphere alkoxide derived from the
product alcohol or 2-PrOH solvent. An unusual feature is the rapid drop in ee of the product alcohol from as high as 60% (R) to
0
% in some cases; this might be due to racemization of the Kaibene ligand in THF caused by the strong base or competitive
inhibition of one diastereomer of the catalyst by reaction with the product (R)-alcohol.
KEYWORDS: asymmetric hydrogenation, chiral N-heterocyclic carbene, ketone reduction, DFT, bifunctional catalysis,
ruthenium catalysis
INTRODUCTION
Previously our group reported several catalysts (1−4, 6) for
the hydrogenation of ketones and esters using primary amine-
■
Chiral alcohols and amines and their derivatives are important
synthetic intermediates in the production of pharmaceuticals,
fragrances, agrochemicals, and other fine chemicals. One
economical and effective method for the production of
enantiopure alcohols and amines is through the asymmetric
reduction of the corresponding prochiral ketones or imines
using hydrogen gas in combination with enantioselective
transition-metal-based catalysts. As the most versatile ligands
for transition metals to date, phosphines remain the most
popular choice for homogeneous hydrogenation catalysis,
despite their sensitivity to oxygen and the potentially harmful
17−22
tethered NHC as the ligands (Scheme 1).
Under relatively
mild conditions, these ruthenium or iridium complexes can
catalytically reduce ketones to alcohols. Compound 2 can also
reduce imines to amines, or esters to alcohols. Complex 2 is a
superior ketone hydrogenation compared to compounds 1, 3,
4
, and 6. Despite containing the new enantiopure “Kaibene”
1
−9
ligand, compound 6 displayed no enantioselectivity during
ketone hydrogenation.
With our experience in the design of these transition-metal
catalysts based on NHC-R-NH ligands, two essential concepts
2
1
,10−13
in addition to the chiral element are required for the new
generation of highly active asymmetric hydrogenation catalysts.
effects when released during synthesis and catalysis.
Over
the past two decades, N-heterocyclic carbenes (NHC) have
emerged as a class of ligands that are σ-donors of comparable
strength to phosphines and that are readily modified to achieve
23
First, the computational study on compound 2 suggested that
24
the “N−H” effect is essential for high reactivity. The
1
4,15
a wide range of steric and electronic properties.
Moreover,
NHC have proven to be excellent ligands for asymmetric
hydrogenation applications when used in conjunction with
transition metals like ruthenium, iridium, and rhodium.
Received: July 16, 2017
Revised: August 24, 2017
Published: August 28, 2017
16
©
XXXX American Chemical Society
6827
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842

ACS Catalysis
Research Article
Scheme 1. Ketone Hydrogenation Catalysts Containing a
negative charge on the pentamethylcyclopentadiene (Cp*)
ligand, or compared to Ru(p-cymene) compounds that contain
Bidentate NHC-R-NH Ligand
2
27,28
an anionic tosylated amide group.
The neutral hydride
produced when 2 is treated with base and dihydrogen proved
to be much more reactive to ketones than the cationic hydrides
produced from compounds 1, 4 and 6. The last three are
thought to catalyze ketone hydrogenation via an inner sphere
mechanism where the amino group dissociates from the
17,20
metal.
Therefore, although compound 6 carries a chiral
ligand 5 in its structure, no enantioselectivity was observed
21
during catalysis.
Here we report the new ruthenium precatalyst 7 that can
rapidly hydrogenate ketones once activated by base under
dihydrogen pressure. Complex 7 containing a Cp* ligand
produces the desired more nucleophilic neutral hydride
complex. Compound 7 also contains enantiopure “Kaibene”
5
, which provides a primary amine donor for Noyori
bifunctional catalysis. Although many asymmetric NHC ligands
1
4,15,29−41
are known,
NHC are still rare.
enantiopure primary amine-tethered
1,42,43
2
Under mild conditions (50 °C, 25
bar H ), compound 7 hydrogenates ketones with outstanding
2
catalytic reactivity with turnover frequencies (TOF) up to 48
−1
s
and turnover number (TON) up to 10 000. For the first
time, enantioselectivity, although low, is induced in the product
alcohol using a half-sandwich ruthenium system that contains
an NHC-R-NH ligand. Extensive investigation on the catalytic
2
intermediates during the catalysis was performed, and several
hydride species that might be responsible for this highly active
ketone reduction catalytic process were discovered. Related
studies have detected hydrides in other ruthenium ketone
hydrogenation systems.
Computational studies on ketone hydrogenation catalyzed by
formation of an (H)N-M(H) intermediate during catalysis
provides a highly organized transition state that transfers a
hydride from the metal to the polar unsaturated substrate. The
stabilization through extra hydrogen bonds and a highly
ordered six-membered cyclic transition state structure greatly
reduces the energy barrier, thus producing high reactivity.
Many ketone and imine hydrogenation catalysts utilize this
concept, and extensive modeling and experimental work have
been conducted to verify the significance of the involvement of
23,44−49
compound 2 suggested that dihydrogen activation at ruthenium
23
occurs by an outer-sphere, alcohol-assisted mechanism.
A
similar mechanism might be assumed for compound 7.
However, from compound 2 to compound 7, the chelating
ring size is decreased from a seven-membered ring to a six-
membered ring, and our past experience with iron hydro-
genation catalysts has proven that a different ring size can result
in a fundamentally different catalytic mechanism with different
13,19,20,22,25,26
the N−H group for high reactivity.
The second
factor is the overall neutral charge on the intermediate hydride
species. Assuming complete formation of the corresponding
hydride species, cationic para-cymene compounds 1, 4, and 6
have poor activity compared to compound 2, which bears a
50,51
active intermediates.
Therefore, a detailed computational
study focusing on the catalytic mechanism comprises the last
part of this paper.
a
Scheme 2. Synthesis of CuI(Kaibene) , 9 and [RuCp*(Kaibene)(MeCN)](PF ), 7
2
6
a
21
The synthesis of compound 8 was reported elsewhere.
6
828
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842

ACS Catalysis
RESULTS AND DISCUSSION
Synthesis of the Catalyst. The imidazolium salt of
Kaibene, (S,S)-5 was synthesized from 1-methylimidazole and
Research Article
Reacting 9 with RuCp*(cod)Cl in refluxing acetonitrile
■
produced [RuCp*(Kaibene)(NCMe)]PF 7 almost quantita-
6
tively. A driving force of the reaction may be the formation of a
+
stable [Cu(MeCN) ] byproduct. Interestingly, using KPF
an enantiopure sulfamidate through an S 2 reaction as reported
4
6
N
2
1
during reflux in an attempt to drive the reaction by
precipitation of chloride induced decomposition. The presence
of the chloride in the ligand sphere might be crucial during the
transmetalation to stabilize the corresponding Ru-NHC
intermediate, presumably a neutral compound with the formula
RuCp*(Kaibene)(Cl)(NCMe) with an uncoordinated amine.
Indeed, heating a pure sample of compound 7 under the
identical condition caused decomposition of compound 7.
Compound 7 is composed of two diastereomers generated by
the chiral center at the metal. We attempted without success to
replace acetonitrile by other labile coordinating ligands such as
tert-butyl cyanide or DMSO in order to change the ratio of
previously. In order to coordinate the NHC ligand to a
ruthenium Cp* fragment, we first attempted to use the
previously reported silver transmetalation reagents: [Ag-
2
1
(
Kaibene) ](PF ) and [Ag(Kaibene) ](I) (8). For example,
2 6 2
complex 8 is prepared by reacting (S,S)-5 with silver iodide,
potassium iodide and the strong base, potassium hexamethyldi-
silazide (KHMDS, Scheme 2). However, the reactions of these
silver compounds with either [RuCp*Cl] or RuCp*(cod)Cl
4
(
cod = 1,5-cyclooctadiene) at various temperatures and in
several solvents produced a complicated mixture that contained
starting materials, (S,S)-5 and unidentified products. For
example, the conditions of acetonitrile at reflux are needed to
cause the dissociation of the cod ligand from RuCp*Cl(cod) to
initiate a reaction. However, under these conditions, the basic
Kaibene ligand on silver is protonated to reform the
imidazolium salt (S,S)-5 (Scheme 2).
1
diastereomers in favor of one isomer. A VT H NMR
experiment was also conducted, but the 1:1 ratio between the
two diastereomers in CD Cl remained unchanged between
50 °C to +60 °C. Interestingly, when 7 was dissolved in THF-
d , a dominant diastereomer was observed with a Cp* methyl
resonance at 1.70 ppm along with the minor one (ratio 7:1)
with a resonance at 1.71 ppm. Thus, solvation can influence the
relative stability of the isomers, possibly through hydrogen
2
2
−
8
As an alternative to the silver reagent 8, we synthesized the
copper analogue [Cu(Kaibene) ](I), 9 (Scheme 2). The use of
2
copper(I) NHC complexes to transfer NHC ligands is not as
common as that of silver(I) complexes.
more stable and less light-sensitive than 9. Unlike 8, 9 could
be prepared without the use of potassium iodide.
Single crystals of 9 for X-ray crystallography were obtained
from a DCM and hexane solution. As for 8, crystals of 9 have
two independent molecules in the asymmetric unit. Similar to
compound 8, the unit cell for compound 9 is composed of “T-
shaped” cations that each embrace an iodide between two
amino groups (Figure 1). The two molecules of 9 have C−Cu−
5
2−55
Compound 8 is
5
6
bonding of the THF with the NH groups.
2
Single crystals of 7 could not be obtained possibly due to the
presence of the diastereomers. Therefore, we used mass
spectrometry (ESI-HRMS) and advanced NMR techniques to
elucidate its structure. The mass spectrum has two major peaks
at 514.1801 and 546.1705 m/z. The first peak matches the
formula [RuCp*(Kaibene)] , that is, 7 without the acetonitrile
ligand, which is labile under the ESI conditions. The second
peak reveals the presence of the dioxygen adduct [RuCp*-
Kaibene)(O )] , formed by reaction with air during the
+
+
(
2
measurement. Nevertheless, the presence of two acetonitrile
ligands, each corresponding to one diastereomer, can be
1
3
1
confirmed by NMR. The C{ H} NMR spectrum of
compound 7 in CD Cl reveals two NHC signals at 193 and
2
2
1
1
89 ppm for the two diastereomers, in a similar region to the
82 ppm resonance of compound 9. To determine whether the
amine group of Kaibene is coordinated to the ruthenium, 2D
1
15
H− N gHSQC is used (Figure 2) as described previously for
21
15
6
. Using this technique, two NH2 N signals at 0.22 and 6.53
ppm, each corresponding to one diastereomer, are observed for
15
7
. The N signal at 6.53 ppm is correlated to two inequivalent
Figure 1. Molecular structure of 9. Selected bond lengths (Å) and
hydrogen nuclei, proving that the bound NH has lost its
2
angles (deg): Cu1A-C1A- 1.915(1), Cu1A-C19A 1.909(1), Cu1A-I1A
freedom of inversion. Similarly, the signal at 0.22 ppm is also
correlated to two hydrogen atoms. Although only the signal of
2.912(7), C19A-Cu1A-C1A 155.3(4), C19A-Cu1A-I1A 105.2(3),
1
15
C1A-Cu1A-I1A 99.4(3). Thermal ellipsoids are plotted at the 50%
probability level.
one of these hydrogen atoms could be observed in the H− N
gHSQC NMR spectrum, the other hydrogen could be
1
identified using a 2D H COSY, indicating that this
C angles of 151.8(3)° and 155.3(4)°, respectively; these
compare with the corresponding larger angles of 160.3(5)° and
diastereomer also has Kaibene as a bidentate ligand.
Asymmetric Hydrogenation of Acetophenone. Com-
pound 7 was first tested at relatively high concentration against
200 equiv of acetophenone in THF, and complete conversion
1
2
3
65.9(6)° in 8. The shorter Cu−I distances of 2.912(7) and
.849(1) Å in 9, when compared to the Ag−I distances of
.252(1) and 3.216(1) Å in 8, are accommodated by the
was observed within 0.5 h at 50 °C, 25 bar H (Table 1, entry
2
smaller C−M−C angle in 9 (155.3(4)°) compared to 8
165.9(6)°). Although the M−I distance might be too long for
significant covalent bonding, the complete absence of covalent
interaction is also unlikely. The position of the iodide anion
in the solid state can be a combination of electrostatic
interactions, weak covalent bonding, and weak hydrogen
bonding.
1). However, no enantioselectivity was induced when THF was
used as the solvent. We then decreased the pressure to about 5
bar, but there was no conversion (Table 1, entry 2). Thus, the
kinetics has a hydrogen concentration dependence. Increasing
the ketone concentration by a factor of 8 (Table 1, entry 3,
1600 equiv) causes an approximately 7-fold increase in the
(
57
TOF from 0.6 s− to 4.3 s . This indicates that the
1
−1
6
829
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842

ACS Catalysis
Research Article
1
15
15
1
Figure 2. H− N gHSQC spectrum of 7 in CD Cl (f1: N, f2: H, 700 MHz).
2
2
a
Table 1. Testing of Compound 7 for the Hydrogenation of Acetophenone
c
entry
1
[Ru] (mM)
C:B:S
solvent
TOF (s−1)b
TON
ee (%)
0.36
0.36
1:8:200
1:8:200
THF
THF
0.63
0
200
0
0
n/a
d
2
3
4
5
6
0.36
1:8:1600
1:8:1600
1:8:5000
1:8:10000
1:8:10000
1:8:2500
1:8:5000
1:8:150000
1:8:15000
1:8:20000
1:0:5000
THF
4.3
12
8.7
21
0
1600
1600
5000
10 000
0
0
0.36
2-PrOH
2-PrOH
2-PrOH
2-PrOH
2-PrOH
MeOH
10 (11)
5 (9)
16 (44)
n/a
0.057
0.057
0.057
0.057
0.057
0.057
0.057
0.057
0.057
e
7
f
8
0
0
n/a
9
0.65
0
500
0
7 (10)
n/a
g
1
0
acetophenone
2-PrOH
2-PrOH
2-PrOH
h
11
5.9
0
4250
0
4 (11)
n/a
12
i
13
0
0
n/a
a
b
General conditions: concentration of base and acetophenone varied according to the C:B:S. TOF is measured from the slope of the linear region
in the catalytic profile. The catalytic profile was developed via periodic collection of a small catalytic solution aliquot (ca. 0.1 mL) at different time
c
d
e
intervals from the Parr reactor. ee (R) at the end of catalysis (ee at the beginning). This experiment was performed under 5 bar of H . This
2
(g)
f
g
experiment was performed under Ar(g) in the absence of H . This experiment was conducted at 25 °C. This experiment was conducted in neat
ketone without any additional solvent. Only 28.3% conversion was observed. No base was added to this reaction.
2(g)
h
i
concentration of the substrate is also part of the rate law. In
order to examine whether the 2-PrOH-assisted dihydrogen-
splitting mechanism observed for 2 is also observed for 7, the
solvent was changed to 2-PrOH (Table 1, entry 4). Here the
TOF almost tripled, and now some ee in the alcohol was
detected (10%, R).
The catalyst concentration was then decreased to increase
the substrate to catalyst ratio without changing the concen-
tration of substrate significantly (Table 1, entries 5−13). Using
as the reaction progresses. Catalysis by transfer hydrogenation
from the 2-PrOH solvent is a common source of product
23
12,58
racemization over time.
In order to rule out this possibility,
the catalyst was tested under Ar in 2-PrOH, and no reaction
was observed for 30 min (Table 1, entry 7). Thus, there must
be another explanation for the decline in ee of the product as
discussed further below. There was no conversion after 30 min
when the temperature was lowered to 25 °C. MeOH was not as
effective a solvent for catalysis (Table 1, entry 9). The lower
activity might be attributed to the formation of a stable
2
-PrOH, we observed the TOF to double approximately when
the substrate concentration was doubled (Table 1, entries 5 and
), as in the THF experiments. There is a short induction
23,45,59,60
methoxide complex as proposed in other work.
There
6
was no conversion when the substrate was used as the solvent
(Table 1, entry 10). Here an enolate complex (see Figure S1,
Supporting Information) may be sequestering the catalyst as
period while the catalyst is activated (Figure 3). The ee of the
product alcohol is observed to fall rapidly from 44 to 16% (R)
6
830
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842

ACS Catalysis
Research Article
entry 13). The effect of the base is crucial for catalysis as it
facilitates the formation of an active, neutral hydride
intermediate which plays an important role in the catalytic
cycle (see below in the mechanistic and DFT sections).
Hydrogenation of Other Substrates. The scope of
substrate structures (Table 2) was then investigated using the
conditions established for acetophenone (Table 1, entry 5).
The enantioselectivity increased as the size of the alkyl
substituent of the phenyl ketone was increased (Table 2,
entries 1−3) but with some loss in ee during the reaction as
observed for acetophenone. However, a further increase in the
alkyl group size to t-Bu resulted in lower activity and opposite
enantioselectivity (Table 2, entry 5). Changing to a cyclohexyl
group, of intermediate size between i-Pr and t-Bu, unexpectedly
completely shut down catalysis (Table 2, entry 4). Aryl ketones
with electron-withdrawing substituents on the aryl rings were
also tested for catalysis (Table 2, entries 6−9). Notably, the
ketone with chlorine in the para position was reduced most
rapidly and with somewhat higher ee (Table 2, entry 8), while
the ketones with electron-withdrawing groups at the ortho or
meta positions were less reactive compared to acetophenone.
The ortho substitution might result in steric hindrance to the
Figure 3. Reaction progress of the hydrogenation of acetophenone
catalyzed by 7 and base in 2-PrOH (Table 1, entry 6).
23
proposed previously for analogous catalyst 2. This effect was
further observed when the catalysts were tested at 1:15 000 and
1
:20 000 C:S ratio (Table 1, entry 11, 12). A control
hydrogenation while the meta substitution with CF groups
3
experiment without the presence of base was also conducted,
and no conversion of the substrate was observed (Table 1,
might cause side reactions due to the acidity of the product
alcohol. Increasing the size of the aryl side of the ketones did
§
Table 2. Substrate Scope for CO Bond Hydrogenation Catalyzed by 7
§General conditions: [7] = 5.72 × 10−5 M, [KOtBu] = 4.46 × 10 M, [Substrate] = 0.286 M in 6 mL of 2-PrOH. TOF is measured from the slope
−4
a
b
of the linear region in the catalytic profile. ee at the end of catalysis (ee at the beginning); R unless otherwise specified.
6
831
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842

ACS Catalysis
Research Article
not improve the enantioselectivity significantly (Table 2, entry
for hydride transfer with a lower energy barrier. The
introduction of stereogenic phenyl groups on the ligand
scaffold in 7 does induce enantioselectivity, albeit low, in
certain cases. However, the more rigid structure and these
phenyl substituents introduce steric encumbrance near the
active site which appears to slow or suppress the reduction of
bulky ketones (Table 2, entries 4 and 5).
1
0 and 11) but still resulted in excellent reactivity. The higher
reactivity for 4′-methylacetophenone and 2-acetylnaphthalene
over acetophenone during catalysis might be due to a better
3
C(sp )H/π attraction between the Cp* ligand and the more
electron-rich aryl rings on the ketones. Similar effects had been
proposed in the past for the Ru(hexamethylbenzene)(tosyl-
2
4
dpen) system. Ketones that are part of a cyclohexyl ring
displayed no reactivity (Table 2, entries 12 and 13). The
catalytic activity ceased at about 80% in the case of 2-
acetylpyridine, possibly due to catalyst poisoning by the
product (Table 2, entry 14). 2-Acetylfuran (Table 2, entry
Observation of Hydride Complexes by NMR. The
investigation of the mechanism of catalysis is valuable as it
provides useful information for the future direction of catalyst
design in order to improve enantioselectivity while retaining the
high reactivity of our system. A plausible mechanism for ketone
hydrogenation in THF, similar to the one proposed for 2
15) could not be reduced, possibly due to substrate poisoning
23
at the active site of the catalyst. In the case of α,β-unsaturated
ketones (Table 2, 16), the lack of reactivity might be caused by
a Michael addition of the amino or amide group of the catalyst
on to the olefin. Significantly, our catalyst can efficiently reduce
the difficult substrate 3-methylbutan-2-one, which contains a
nonbenzylic positioned carbonyl carbon (Table 2, entry 17).
Benzophenone is hydrogenated very rapidly (Table 2, entry
previously, is shown in Scheme 4. The reaction of precatalyst
Scheme 4. Proposed Catalytic Cycle of Ketone Reduction by
Compound 7 in THF
18). Benzaldehyde (Table 2, entry 19) cannot be reduced by
this catalyst, possibly because the NH group of the catalyst
2
undergoes imine formation with the aldehyde, effectively
preventing the “N−H” effect required for rapid turnover.
In the cases where ee is observed in the product (entries 1−
3
, 5, 8−11 in Table 2), there is a decline in ee over the course
of the reaction as observed for acetophenone reduction.
Possible reasons for this are discussed further below.
Compound 7 and base also catalyzes the hydrogenation of
N-phenylbenzylimine at higher H2 pressures than those
required for ketone reduction (Scheme 3). The rate of amine
Scheme 3. N-Phenylbenzylimine Hydrogenation Catalyzed
by Compound 7
formation is 200 h− under 46 bar of H and only 100 h
1
−1
2
under 35 bar of H , much lower than that for ketone reduction.
2
The substrate imine and product amine might be competing
with dihydrogen for the active site on the catalyst and therefore
a higher hydrogen concentration in the solution is required to
reduce this aldimine efficiently. Since the reaction needs to be
performed under basic conditions, it is highly plausible the
hydrogen transfer is a “hydride first, proton second” sequence
that is similar to ketone hydrogenation, instead of an alternative
“proton-first” mechanism involving an iminium ion intermedi-
ate, which is more common for cationic imine hydrogenation
catalysts in an acidic catalytic environment.
61,62
7 with base and hydrogen is thought to generate the hydride
complex 13-R/S, which may exist as diastereomers due to the
chiral center at ruthenium. The hydride(s) can be generated
through two possible pathways when both base and hydrogen
are present, starting from an unobserved amine alkoxide
intermediate 10 formed after the dissociation of the acetonitrile
ligand 7. The first catalytic pathway is the heterolytic splitting
of dihydrogen through an amide compound 11. The second
In general, 7 has TOF for acetophenone conversion that is
approximately 4 times greater than that of the achiral NHC
catalyst 2 when the conditions of the temperature of operation
23
and catalyst loading are considered. The higher reactivity of
compound 7 can be attributed to two factors. The first is the
higher electron richness at the ruthenium center contributed by
the Cp* ligand and the dialkyl substituted NHC (alkyl-aryl
substituted in compound 2). Second, the more rigid and
organized six-membered chelating ring may improve the
catalyst stability and create a more organized transition state
2
pathway is the deprotonation of a cationic η dihydrogen
compound 12 by an outer-sphere alkoxide. The hydride
compound 13 subsequently delivers a hydride to the substrate
6
832
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842

ACS Catalysis
Research Article
Scheme 5. Hydride Formation Produced by Reacting 7 with Base and Hydrogen
to regenerate an alkoxide compound 10, where the alkoxide
ligand might be from the product or the added base in 2-PrOH.
The DFT study below suggests that both pathways of Scheme 4
are plausible and exist as competing processes in the presence
of hydrogen gas and base.
complex 11 can have two possible conformations where the
two phenyl groups on the backbone are either diaxial or
diequatorial in the five-membered ring. The diaxial conforma-
tion is about 2.78 kcal/mol more stable than the diequatorial
configuration and hence likely to be the true conformation of
compound 11 in solution (Figure 4 and see below). As
In order to detect the hydride species of Scheme 4, NMR
spectra were obtained after the reaction of solutions of 7 in
THF-d with KOtBu (approximately 10 equiv) in a J Young
8
tube under Ar and then with 2 bar H at room temperature
2
(Scheme 5).
Under Ar, a new compound with a Cp* resonance at 1.67
1
ppm appeared in the H NMR, apparently at the expense of the
major diastereomer of 7, since the Cp* of the minor
diastereomer still was present. The color of the solution
changed from bright yellow to brown during the base addition.
Other very small Cp* peaks were also present. The formation
of an amide compound after the addition of base to a
ruthenium-amine complex in the absence of dihydrogen has
been proposed for a number of hydrogenation cata-
2
3,45,59,60,63
lysts.
Therefore, we propose that the new species
is a Cp*Ru(amide) compound 11 (Scheme 5) where the
primary amine on Kaibene is deprotonated. To obtain a clean
spectrum of 11, a separate experiment was conducted where
THF was evaporated and the residue was extracted with
filtration into deuterated benzene (Figure S12). One major
compound was observed with two CHPh resonances found at
Figure 4. DFT-optimized structure of compound 11. The CH was
3
rotated to obtain a minimum NH---HC distance, assuming CH free
3
5.84 and 6.04 ppm. This assignment was confirmed by gCOSY
rotation in solution. DFT settings: M11-L/6-31++G(d,p) (H, C, N,
O); ECP-SDD (Ru)/IEF-PCM (SMD:2-PrOH)/Grid (99, 590).
analysis. In addition, no acetonitrile (free or coordinated) can
be detected in this spectrum, indicating that this ligand was
removed. A new singlet was observed at 4.68 ppm and assigned
as the amido NH peak. The NMR shift of this signal is similar
to that of the amido NH in RuH(S-binap)(app) (app = 2-
illustrated, the NH group has close approaches to the CH3
hydrogens of the Cp* ligand (approximately 2.3 Å) and the
adjacent CHPh (approximately 2.2 Å). 1-D NOESY experi-
ments focused on the 4.68 ppm of NH peak were conducted
and three hydrogen atoms with significant nuclear Overhauser
63
amido-2-(2-pyridyl)propane) at 4.3 ppm. The widths at half
height was 1.23 Hz for the CHPh peaks and 1.91 Hz for the
NH peak. The 0.7 Hz difference represents relaxation
14
contributions from the CH groups of the Cp* and a
N
effect were observed: adjacent CHPh (6.05 ppm), CH from
3
3
dipolar contribution. Based on the DFT calculations, the amide
Cp* (1.75 ppm), and OH from a tert-butanol alcohol (1.48
6
833
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842

ACS Catalysis
Research Article
1
Figure 5. H NMR (THF-d , 500 MHz) spectrum of hydrides formed when compound 7 is reacted with base under hydrogen. The Gibbs Free
8
Energies of R-13 and S-13 were obtained via DFT (see Figure 4 for the DFT settings).
Figure 6. NMR-scale experiment (THF-d , 500 MHz) of catalytic acetophenone reduction at 4% catalyst loading, approximately 2 bar of hydrogen
8
pressure, 25 °C in THF-d8.
ppm) produced during the experiment which hydrogen bonded
to the NH.
−10.40 (s, S-13), −10.64 (d, 15), and −10.80 (d, 15) ppm.
The narrow width of these peaks and the long T (1.5 to 0.6 s
1
Exposure of this amide compound to hydrogen gas
when determined at 700 MHz) are good indicators that all of
immediately generated several hydride species that can be
detected by H NMR spectroscopy (Figure 5). The color of the
these species are ruthenium hydrides. Ruthenium Cp*
1
dihydrogen species like unobserved 14 would have T less
1
6
4−66
solution also brightened from brown to deep yellow. The
reaction presumably goes via an unobserved, short-lived
dihydrogen compound 14. Three of these hydride species
have much higher concentrations than the others; these have
than 100 ms at 700 MHz
and are known to undergo
6
6,67
oxidative addition to produce a trans-dihydride isomer.
The
doublets of a dihydride 15 of unknown structure at −10.64 and
−10.80 ppm with J = 7.1 Hz were verified to be coupled to each
1
1
1
H chemical shifts in the hydride region at −8.22 (s, R-13),
other using a H− H 2D-gCOSY experiment. Several potential
6
834
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842

ACS Catalysis
Research Article
Figure 7. Calculated catalytic cycle for acetophenone reduction in 2-PrOH but in the absence of alcohol assistance; the free energy of each stage is
shown with the enthalpy in brackets. Figure 4 lists the computational settings.
geometries of compound 15 including cis and trans dihydride
were calculated using DFT methods, but none had a reasonable
energy (Figure S3, Supporting Information).
The two hydride resonances of major intensity are assigned
to two diastereomeric ruthenium monohydrides, R-13 and S-
generating new monohydride and dihydride diastereomers.
However, these hydrides that form after several hours might
not be relevant to our catalysis since typical catalytic runs will
achieve 100% conversion or cease reacting within 30 min.
There is a report of the racemization of a chiral NHC derived
from an ester of valine when coordinated to ruthenium in the
13. They likely form by the heterolytic splitting of dihydrogen
41
over the metal-amide bond in 14, a common reaction for
presence of base.
23,49,63,68
ruthenium-amine based hydrogenation catalysts.
DFT
One drop of acetophenone was added to a J Young tube
containing predominantly R-13 under argon (Figure 6, this was
about 4% catalyst loading according to NMR integration).
Immediately we observed a significant reduction in the
concentration of R-13 at −8.22 ppm relative to the other
hydride species including S-13 and the dihydride 15, thus
indicating a stereoselective reaction had occurred (however, the
catalytic hydrogenation in THF produced no ee). We also
observed the formation of the −9.04 ppm hydride immediately
in the system, but at low concentration. A small quantity of the
ketone reduction product, 1-phenylethanol was observed with
the characteristic doublet at 1.34 ppm (methyl group of 1-
phenylethanol). The J Young tube was then refilled with
approximately 2 atm hydrogen, and the reaction was monitored
over time. As illustrated in Figure 6, acetophenone was slowly
calculations (see below) indicate that hydride R-13 is about 1.6
kcal/mol more stable than S-13. This allows us to assign
complex R-13 to the singlet of greater intensity at −8.22 ppm
and S-13 to that of less intensity at −10.40 ppm. Dihydrogen
can be removed from such a mixture by the freeze−pump−
thaw method, and the spectrum in the hydride region remains
the same, showing that dihydrogen addition to the amide
complex is irreversible (Figure 6, bottom spectrum). Solutions
of the hydrides are stable for 0.5 h under an H atmosphere,
2
and no decomposition of existing hydrides or formation of new
hydrides could be observed. However, an attempt to observe
the 13C signals of the carbene carbons of these hydride species
1
13
required a H− C 2D-HMBC experiment of several hours
duration at 700 MHz. Over this time period, new hydride
signals appeared at −9.04 (s), −11.82 (d, 7.7 Hz), and −12.17
ppm (d, 7.7 Hz). This NMR experiment was conducted under
a dihydrogen atmosphere. These peaks were also observed at
converted into 1-phenylethanol under H with a conversion up
2
to 62% on the fourth day. Therefore, the catalyst at this high
loading is active at room temperature in THF-d . Importantly,
8
the beginning after base addition to 7 under H but at
no hydride signals were observed in the 28% conversion
spectrum and other spectra after it. The resting state at this
stage would be the amide complex 11 or an alkoxide complex
formed by its reaction with the product alcohol (see below);
these do not have hydride resonances.
2
significantly lower concentration. The peak at −9.04 is
1
3
correlated to a carbene C resonance at 191 ppm. The
doublets are characteristic of a ruthenium dihydride compound
and are associated with an unusually low field carbene 13
C
chemical shift of 211 ppm. These complexes may arise from the
base-catalyzed racemization of the Kaibene backbone thus
Decline in ee with Conversion. While no ee was detected
in the alcohol product when THF was used as the solvent, there
6
835
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842

ACS Catalysis
Research Article
Figure 8. Selected optimized geometries for intermediates and transition states within the catalytic cycle when 2-PrOH is used as the solvent. CH3
hydrogens, phenyl hydrogens, and phenyl carbons except the ipso carbon on the backbone were removed for clarity. All bond lengths are in Å.
were measurable values for several of the alcohols from the
ketones of Table 2 when 2-PrOH was used. For 1-
phenylethanol produced by the hydrogenation of acetophenone
in 2-PrOH (Figure 3), the ee starts at 44% (R) and drops to
Proposed Catalytic Cycle Supported by Computa-
tional Calculation. With all of the NMR and experimental
data collected, density functional theory (DFT) was then used
to further probe the catalytic cycle from a quantum chemical
perspective. A similar mechanism might be expected between
our system and the previously reported similar achiral analogue
1
6% by the time 100% conversion is attained. We have ruled
out that this decrease is caused by transfer hydrogenation
Table 1, entry 7). The decrease is clearly related to the nature
69
(
2
. M11-L was used as the functional since it has been proven
7
0−72
of the solvent: (a) in THF there is no ee; (b) in 2-PrOH with
certain ketones, there is good ee when the product
concentration is low; (c) in 2-PrOH with product concen-
trations greater than 0.2 M there is low ee and lower catalyst
activity; (d) in neat ketone there is no catalysis. There are
several possible explanations for this decrease. It could be
caused by the evolution of different catalytic species such as
diastereomers of the hydrides R-13 and S-13 and their
racemized forms that produce alcohols with either a lower or
opposite enantiomeric ratio. Another possibility is that the
active species are sequestered as alkoxide complexes and that
to be useful for calculating organometallic systems.
Ruthenium was treated with the Stuttgart effective core
73
potential (SDD-ECP), and all other elements were calculated
74,75
with the 6-31++G** basis set.
The calculation was
performed using 2-PrOH as the modeling solvent at 298.15
K. The amide compound R-11 was set at zero free energy and
the entry point of the catalytic cycle from the precatalyst. The
following notation will be used to label intermediates R(or S)-
number-[R(or S)A] where the first R(or S) refers to the metal
(
priority: NH > NHC > Cp* > H), the second R(or S) refers
2
−
to the alcohol (A) being produced. The results are shown in
Figure 7, and key structures are found in Figure 8.
certain combinations of alkoxide ligand (2-PrO , (R) or (S)-
−
RPhCHO ) and that ruthenium NHC produce catalysts of
We chose to calculate the R isomer as the active catalyst for
several reasons. First, R-13 is about 1.6 kcal/mol more stable
than S-13. One of the monohydride compounds observed in
our NMR experiment (previous section) was in dominant
concentration compared to the other ones. The same study also
revealed that the dominant hydride, presumably R-13, reacted
immediately with acetophenone to generate the corresponding
alcohol. This indicated that R-13 could be the major active
species during catalysis. Moreover, during catalysis, the
enantioselectivity diminished over time. An explanation for
this phenomenon is that the kinetic hydride (R-13) provides
the product alcohol with high enantioselectivity, while later in
the reaction another hydride appears that provides the product
with the opposite chirality or as a racemate.
different activity and enantioselectivity. The competitive
inhibition of one diastereomer of the catalyst by reaction
with the product (R)-alcohol to form a more stable alkoxide
might be the cause (supported by the DFT calculations below).
The high activity of the catalyst suggests that a Noyori-type
mechanism is operational and that the decline in ee might be
due to the base-induced racemization of the Kaibene ligand.
The higher base strength of alkoxides in THF would cause the
62
more rapid racemization of the ligand than in 2-PrOH.
Another explanation for the decline of ee over the course of
catalysis is the presence of a racemization catalyst in the system.
This was ruled out since no racemization occurred when the
reaction was conducted with 0.286 M of enantiopure R-1-
phenylethanol instead of the substrate.
6
836
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842

ACS Catalysis
Research Article
Figure 9. Energies of two diastereotopic structures R-20 required to initiate an inner-sphere transfer of hydride to ketone are already higher than the
energy of the transition state R-TS17/18 for the outer-sphere hydride transfer; the free energies of each stage are shown with the enthalpies in
brackets. Figure 4 lists the computational settings.
Precatalyst 7 reacts with KOtBu in THF to give an active
catalyst solution. The experiments just described and computa-
tional work point to the initial formation of the amide complex
R-11. The optimized structure of R-11 is illustrated in Figure 8.
The sum of the angles around the amide nitrogen is 358.3°,
which is consistent with a trigonal planar structure around the
nitrogen atom. The generated molecular orbitals displayed
obvious overlap (see Figure S2, Supporting Information)
between the filled p-orbital of the amide, d-orbitals of the
metal and orbitals of the Cp* ligand (HOMO−3 and HOMO−
dihydrogen over the Ru−N bond occurs through the transition
1
−
state R-TS14/13 (G° = 18.1 kcal/mol, ν = 748i cm ),
producing the monohydride R-13 (G° = −7.3 kcal/mol), the
lowest thermodynamic point, and the resting state structure of
the catalytic cycle. This is the hydride that we propose was
observed in our NMR experiment. In the calculated structure,
the amine-ruthenium bond is further elongated to 2.16 Å, and
the Ru−H bond is 1.65 Å. These bond lengths are very similar
to several phosphine-based ruthenium hydride compounds that
7
8−83
are active for ketone hydrogenation.
The NHC−
4). The calculated amide-ruthenium bond length is 1.95 Å,
ruthenium bond length continually decreases at each stage of
the activation and splitting of dihydrogen.
slightly longer than that of the observed Ru−N distance of 1.90
Å in the amide complex [Ru(p-cymene)(Tsdpen)], an effective
When acetophenone is introduced to hydride R-13, it first
forms a hydrogen-bonded adduct R-17 (G° = 6.6 kcal/mol
relative to R-11). Then a six-membered ring hydride transfer
transition state R-TS17/18-SA (G° = 17.9 kcal/mol) leads to
an alkoxide ion pair intermediate R-18-SA (G° = 11.0 kcal/
mol), located as a local energy minimum. Here in R-18-SA, the
alkoxide is hydrogen-bonded to the amine of the cationic
ruthenium complex. Such a transition state and ion-pair
intermediate were also identified in calculations for the achiral
76
transfer hydrogenation catalyst.
The first intermediate detected by computation in the
·
··
reaction of R-11 is a weakly interacting H Ru adduct (Figure
2
7
, R-16) with G° = 2.4 kcal/mol. This noncoordinated
hydrogen intermediate was also calculated in our previous
2
3,77
work.
A dihydrogen coordination transition state R-TS16/
1
4 similar to previous work is also found with ΔG° = 8.0 kcal/
mol. At this transition state, there is a slight elongation of the
amide-ruthenium bond and the dihydrogen bond (0.77 Å
23
system 2. Deprotonation of this amine is easily achieved via a
flat potential energy surface to give intermediate R-19-SA (6.0
kcal/mol relative to R-11) where an alcohol was hydrogen-
bonded to the ruthenium-amide. Attempts to locate a transition
state between R-18-SA and R-19-SA failed as the calculation
either converged to R-18-SA or to R-19-SA. We noticed such a
transition state is more difficult to locate for full structure
calculations but is usually found when simplified catalyst
compared to 0.75 Å calculated for free H ) while the carbene−
2
ruthenium bond length decreases. The geometry around the
amide is pyramidal instead of trigonal planar, which indicates
that the amide ligand is no longer forming a dative π bond with
the ruthenium. Over this barrier, lies the dihydrogen compound
R-14 (G° = 7.2 kcal/mol). The H−H distance in the
dihydrogen molecule is 0.88 Å, while the amide-ruthenium
bond is increased to 2.10 Å. The heterolytic splitting of
50
structures are used, or when a concerted dihydrogen transfer
6
837
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842

ACS Catalysis
Research Article
Figure 10. Dihydrogen splitting to an outer-sphere alkoxide; the free energies of each stage are shown with the enthalpies in brackets. Figure 4 lists
the computational settings.
2
3
transition state is located. Therefore, we conclude that with
all the steric factors included in the calculation, such a proton
transfer is smooth and barrierless. The dissociation of alcohol
from the amide produces the amide complex R-11, but now at
G° = −4.0 kcal/mol. This change in energy of R-11 from the
beginning to the end of the cycle corresponds to the free
energy of hydrogenation of acetophenone using our basis set
and functional.
kcal/mol), and hydride transfer transition state S-TS17/18-SA
−1
(G° = 19.7 kcal/mol, ν = 789i cm ). These intermediates and
transition states of the S catalysis path mirror the R catalysis
path but with an overall energy span 26.4 kcal/mol,
approximately 1.0 kcal/mol higher than the corresponding R
energy span of 25.4 kcal/mol. These data are consistent with
the hypothesis that the R diastereomer is operating faster than
the S diastereomer during catalysis.
In the catalytic cycle shown in Figure 7, the heterolytic
splitting of dihydrogen R-TS14/13 and hydride transfer R-
TS17/18-SA are the two most energy-demanding steps, both
with similar energy. This is consistent with our experimental
observation that the rate of catalysis is dependent on hydrogen
pressure, substrate concentration, and the structure of the
substrate. The energy span for the catalytic cycle from the
lowest energy point, the hydride R-13, to the highest energy
point, R-TS14/13, is 25.4 kcal/mol, which is consistent with
the temperature of 50 °C that we observed was needed for
In order to further validate that the mechanism is purely an
outer-sphere one instead of one with a competing inner-sphere
attack of the hydride on coordinated ketone, we calculated the
structures and energies of two intermediates required for the
inner-sphere mechanism. These have the amino group of the
NHC ligand uncoordinated and substituted by acetophenone in
the pro-R and pro-S orientations (R-20-SA and R-20-RA, Figure
9
). These features were present in the inner sphere mechanism
17,20
calculated for a ruthenium cymene NHC−R−NH system.
2
8
4
However, the energies of R-20-SA and R-20-RA are already
higher than the hydride transfer transition state of the outer-
sphere mechanism (Figure 9) even before adding the energy
necessary to reach the transition state for inner sphere hydride
transfer (previous studies show that, in an inner-sphere
mechanism, the TS can be about 12.2−15.6 kcal/mol higher
rapid catalysis.
Although the R version of the diastereomer was used as the
model for full DFT calculation, several important intermediates
and transition states of the S diastereomer were also calculated.
In comparison, the lowest point of the S diastereomer is also
the hydride compound S-13 with G° = −5.7 kcal/mol, 1.6 kcal/
mol less stable than R-13. The highest point in the S energy
profile is again the heterolytic splitting of dihydrogen over the
metal−amide bond at 20.7 kcal/mol (S-TS14/13, ν = 876i
2
0,85
in energy than the ketone-bound intermediate
). A
nondissociative inner-sphere mechanism was also considered.
However, this mechanism produces an extremely crowded
ligand sphere with a ketone adduct of high energy (G° = 35.9
−
1
cm ), 2.6 kcal/mol less stable than the corresponding
transition state of the R diastereomer. Three other local
maxima/minima of the S catalyst were also calculated: amide S-
5
3
kcal/mol) when the Cp* ligand is switched from η to η
coordination mode to allow an extra ligand site for the
11 (G° = −2.8 kcal/mol), dihydrogen adduct S-14 (G° = 7.5
substrate. Therefore, an inner-sphere mechanism is unlikely.
6
838
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842

ACS Catalysis
Research Article
During our catalytic testing, we observed a significant
increase in reactivity when the solvent was switched from
THF to 2-PrOH. Past experience on the achiral analogue 2
suggested that the catalysis of 7 in 2-PrOH might proceed
mol), and to R-1-phenylethanol (ΔG° = 5.3 kcal/mol). Thus,
as the concentration of the product alcohol increases, there will
be more R diastereomer being inhibited by the product, leaving
more of the less enantioselective S diastereomer to produce
almost racemic product.
23
through an alcohol-assisted mechanism. In order to test this
hypothesis, one explicit 2-PrOH molecule was introduced as a
hydrogen shuttle in the hydrogen splitting step according to O
CONCLUSIONS
2
3
■
et al. As illustrated in Figure 10, the relative energy of the
hydrogen splitting to the outer-sphere 2-propoxide transition
state R-TS24/25 (G°=18.3 kcal/mol) is similar to that of the
splitting to amide transition state R-TS14/13 (G° = 18.1 kcal/
mol), both calculated with 2-PrOH solvation. In contrast, the
relative energy of intermediates R-11, R-16, R-14, and R-13 are
all higher with an explicit 2-PrOH molecule positioned via
hydrogen bonding. These new intermediates, R-21, R-22, R-23,
and R-24 all had lower enthalpy compared to their
corresponding alcohol-free intermediates. As a result, the
overall energy span for the dihydrogen activation part of the
catalysis is reduced from 25.4 to 16.6 kcal/mol (from the key
solvated hydride R-25 to the dihydrogen splitting R-TS24/25).
Including an explicit 2-PrOH in the hydride transfer step (R-
TS26-RA in Figure 9, G° = 26.6 kcal/mol) does not provide a
lower energy pathway for this half of the cycle as was also
A new ketone asymmetric hydrogenation catalyst 7 based on
ruthenium with an amino-functionalized enantiopure NHC was
prepared. It exhibited excellent reactivity under relatively mild
conditions (50 °C, 25 bar H ). To our knowledge, the high
2
−1
4
TOF (up to 48 s , TON up to 10 ) of compound 7 makes it
one of the most-active ruthenium NHC catalyst for ketone
hydrogenation. In addition, this is the first time that
enantioselectivity, although modest, has been observed for
this class of catalyst. The precatalyst was synthesized through
transmetalation from a copper reagent and comprised two
diastereomers. Multiple hydride intermediates were observed
by NMR, and one of these appears to be responsible for the
reactivity and some enantioselectivity. The use of the solvent 2-
PrOH was superior to THF; this was explained by an alkoxide-
assisted dihydrogen splitting step as modeled by DFT.
Computational studies suggested that an outer-sphere attack
on the ketone by the hydride was more favorable in energy
where the metal-hydride and amine together delivered a pair of
hydrogen atoms to the ketone substrate. Both the experiments
and computation support the idea that either the dihydrogen
splitting step or the hydride attack on the ketone are rate-
determining. A new generation of this type of chiral ruthenium
catalyst that has higher enantioselectivity is currently under
development in our lab.
23
observed for catalyst 2.
In order to explain why poor enantioselectivity is observed
for this catalyst, producing small enrichment in the R alcohol,
attempts were made to discover R-TS26-SA, the TS to produce
the S-alcohol. Unfortunately, we could only locate a TS through
a bond-restricted calculation (see Supporting Information) The
relative free energy of R-TS26-SA is approximately 27.3 kcal/
−1
mol (ν = 713.22i cm ), 0.7 kcal/mol higher than R-TS26-RA.
The small energy difference between these enantiodetermining
steps is consistent with the low ee measured.
EXPERIMENTAL SECTION
Several key intermediates of the alkoxide-assisted mechanism
were also calculated for the diastereomer that is S at ruthenium
■
General Considerations. All of the syntheses of transition-
metal complexes were performed under an argon or nitrogen
atmosphere using standard Schlenk-line or glovebox techniques
unless stated otherwise. The synthesis of compound (S,S)-5
(
see Supporting Information). Overall, the energies for this
cycle were consistently higher than the energy for the R
diastereomer and hence should be less important in catalysis.
As noted above, the ee of the product alcohol declined over
the course of the reaction for the systems that produced
enantioenriched products (e.g., Figure 3). This might be caused
by the product alcohol increasing in concentration. This could
affect which diastereomer of the catalyst is more active (see
Supporting Information) or which diastereomer is selectively
inhibited by alkoxide formation. The alkoxide adducts 21, 27,
and 28 formed from the amide compound 11 are consistently
lower in energy in the R configuration (Table 3). This energy
difference increases as the alcohol is changed from 2-PrOH
21
was reported previously. All of the other reagents were
purchased from commercial sources and used without further
purifications. All of the solvents were degassed and dried using
standard procedures prior to all manipulations and reactions
unless stated otherwise. The deuterated solvents were
purchased from Cambridge Isotope Laboratories or Sigma-
Aldrich, degassed, and dried over activated molecular sieves
prior to use. NMR spectra were recorded at ambient
temperature and pressure using a Varian Gemini 400 MHz
1
13
spectrometer (400 MHz for H, 100 MHz for C, 376 MHz
19
for F), Agilent DD2−500 MHz spectrometer (500 MHz for
(ΔG° = 3.7 kcal/mol), to S-1-phenylethanol (ΔG° = 4.8 kcal/
1
13
H, 125 MHz for C), or an Agilent DD2−600 MHz
1
13
spectrometer (600 MHz for H, 151 MHz for C, 564 MHz
Table 3. Relative Gibbs Free Energy of Alkoxide Adducts 21,
for 19F) unless stated otherwise. The H− N gHSQC was
1
15
2
7, and 28
1
measured by an Agilent DD2−700 MHz spectrometer. The H
G° (kcal/mol) of alkoxide adduct
13
and C NMR were measured relative to partially deuterated
R-1-phenylethanol
S-1-phenylethanol
adduct 28
2-PrOH
solvent peaks but are reported relative to tetramethylsilane
adduct 27
adduct 21
(
TMS).
R-
1.6
0.0
−1.6
Synthesis of Cu(Kaibene) I, 9. A 40 mL vial was charged
a
a
2
diastereomer
with 400 mg of (S,S)-5 (0.94 mmol), 92 mg of CuI (0.48
mmol), and a stir bar and then 10 mL of THF was added to
form a suspension. With vigorous stirring a KHMDS solution
S-
6.9
4.8
2.1
diastereomer
a
(
240 mg of KHMDS (1.20 mmol) dissolved in 4 mL THF) was
added. Extra THF was added to make up a total volume at
about 20 mL. The solution was then stirred overnight. The
R(or S) in this table refers to the corresponding hydride compound
for consistency, not the alkoxide which would have the reverse
configuration by convention. Figure 4 lists the computational settings.
6
839
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842

ACS Catalysis
Research Article
reaction mixture was then filtered to remove insolubles, and the
residue was washed with DCM (5 mL × 3). All of the washings
were combined and then evaporated, and the residue was
dissolved in DCM. The DCM solution was filtered through a
pad of Celite to remove any insolubles and then treated with an
C H F N PRu·0.4CH Cl : C, 49.8; H, 5.2; N, 7.6. Found:
30
37
6
4
2
2
C, 50.0; H, 5.3; N, 6.7. Unsatisfactory nitrogen analyses were
measured, possibly because of a combustion problem due to
the presence of the hexafluorophosphate anion as previously
86
reported in literature.
excess of hexanes to cause the formation of the product as a
NMR Experiment for Hydride Detection. Initially, 4 mg
of 7 (6 μmol) and 5 mg of KOtBu (45 μmol) were dissolved in
1
white precipitate. Yield: 350 mg (99%). H NMR (DMSO-d ,
6
2
98 K, δ): 7.78−6.89 (m, 4,5−CH of imid and Ph on the
0.5 mL THF-d , and the solution was transferred into a J Young
8
backbone, 12H), 5.71 (d, CH next to imid., 1H), 5.09 (dt, CH
next to NH , 1H), 3.70 (s, CH , 3H). C { H} NMR (DMSO-
tube. Hydrogen gas was introduced into the J Young tube by
first evacuating argon gas using the freeze-pump technique,
then the solution was allowed to thaw under a hydrogen
atmosphere. Acetophenone was introduced into the J Young
1
3
1
2
3
d6, 298 K, δ): 181.65 (CNHC), 141.21 (aromatic C near imid),
1
1
38.85 (aromatic C near NH ), 129.0−126.0 (aromatic CH),
2
1
21.64 (4-CH of imid.), 120.82 (5-CH of imid.), 69.38 (CH
tube under argon inside a glovebox, and the H NMR spectrum
near imid.), 57.64 (CH near NH ), 37.52 (CH ). MS
2
3
was recorded.
unavailable due to water sensitivity. Anal. Calcd for
C H CuIN ·0.5CH Cl : C, 55.66; H, 4.99; N, 10.67.
To obtain a cleaner spectrum for the amide compound 11,
the precatalyst and base was stirred in normal THF for 5 min.
The solvent was then evaporated, and the residue was extracted
with deuterated benzene and submitted to NMR analysis (see
the SI).
3
6
38
6
2
2
Found: C, 56.79; H, 5.61; N, 9.81. Single crystals for the X-
ray diffraction study were obtained by slow evaporation of a
saturated hexane/DCM solution.
Synthesis of [RuCp*(Kaibene)(CH CN)](PF ) 7. A 25 mL
3
6
General Procedure for Ketone Hydrogenation Using
Schlenk flask was charged with 44 mg of (1,2,3,4,5-
pentamethylcyclopentadienyl)Ru(1,5-cyclooctadiene)Cl (0.116
mmol) and a stir bar in 10 mL of acetonitrile. The reaction
mixture was refluxed for 0.5 h before 57 mg of 11 (0.076
mmol) in 2 mL of acetonitrile was added. The resultant mixture
was refluxed overnight, and then all of the solvent was
evaporated. To this residue, 10 mL of THF and 80 mg of KPF6
7
. Submilligram samples were prepared by evaporating the
solvent of stock THF solutions. Under argon, solution A with a
total volume 4 mL was prepared from 0.24 mg of 7 (0.343
μmol), 1.715 mmol of substrate, and 50 mg of 1,4-
ditertbutylbenzene in 2-PrOH. Solution B with a total volume
2
mL was prepared from 0.3 mg of KOtBu in 2-PrOH. The two
solutions were then taken up into two syringes that were then
stoppered at the end of the needles inside a glovebox. Against a
stream of hydrogen gas, solution A was first added into the Parr
reactor, then solution B. The reaction pressure was then
increased to 25 bar of H2(g), and the catalytic starting time was
recorded when the stirrer was switched on. Small aliquots of
the reaction mixture were collected at timed intervals using
stainless steel sampling dip tube installed on the Parr reactor.
The dip tube was 30 cm long and has an internal diameter as
(
0.43 mmol) was added and stirred for another 3 h to allow salt
metathesis to occur. Then, all of the insoluble salts were filtered
off through a pad of Celite, and all of the THF was evaporated
under vacuum. To remove any residual salt, DCM was added to
the solid and the cloudy suspension was again filtered through a
pad of Celite. Evaporating all of the DCM afforded the product
as an orange powder that contains two diastereomers. Yield: 79
1
mg (98%). H NMR (DCM-d , 298 K, δ): 7.50−6.80 (m, Ph
2
on the backbone, 10H), 7.08 (d, 5−CH of imid., 1H, isomer
A), 6.94 (d, 5−CH of imid., 1H, isomer B), 6.54 (d, 4−CH of
imid., 1H, isomer A), 6.31 (d, 4−CH of imid., 1H, isomer B),
0
.005 or 0.010 in. At each sampling, the first two aliquots were
discarded and the third aliquot is diluted with oxygenated THF
to 2 mL. GC-FID equipped with a chiral column was used to
analyze the conversion and enantioselectivity. The details of
GC temperature and chromatographs were identical to those
5
.43 (d, CH next to imid., 1H, isomer A), 5.21 (d, CH next to
imid., 1H, isomer B), 4.53 (t, CH next to NH , 1H, isomer A),
2
4
.00 (dt, CH next to NH , 1H, isomer B), 3.94 (s, CH , 3H,
2
3
87
found in Demmans et al.
isomer A), 3.87 (s, CH , 3H, isomer B), 3.62 (m, NH , 1H,
3
2
Computational Details. Density functional theory calcu-
isomer B), 3.56 (m, NH , 1H, isomer A), 3.41 (t, NH , 1H,
2
2
88
lations were performed using the Gaussian09 package and the
M11-L hybrid functional. All atoms were treated with the 6-
isomer B), 3.30 (m, NH , 1H, isomer B), 3.23(m, NH , 1H,
2
2
69
isomer A), 2.37 (s, CH CN, 3H, isomer B), 2.31 (s, CH CN,
3
3
3
1++G(d,p) basis set while ruthenium was treated with the
3
H, isomer A), 1.70 (s, CH of Cp*, 15H, isomer B), 1.66 (s,
3
13 1
7
3−75
Stuttgart effective core potential.
An ultrafine grid
CH of Cp*, 15H, isomer A). C { H} NMR (DCM-d , 298
3
2
(
pruned: 99,590) and integral equation formalism polarizable
K, δ): 193.40 (CNHC, isomer B), 189.42 (CNHC, isomer A),
42.27 (aromatic C near imid., isomer B), 141.56 (aromatic C
continuum model (IEF-PCM) with radii and nonelectrostatic
terms from the SMD solvation model in isopropanol were used
for all optimizations.
1
near imid., isomer A), 139.45 (aromatic C near NH , isomer
2
89−91
Ground states were connected to
A), 135.45 (aromatic C near NH , isomer B), 130.5−123.8
2
their transition states by performing intrinsic reaction
(
aromatic CH), 123.06 (4-CH of imid., isomer A), 122.89 (5-
92
coordinate (IRC) calculations. The amide compound R-11
was used as the zero-energy point for both the enthalpy and
Gibbs free energy comparisons. Stationary points were
characterized by normal-mode analysis. A frequency analysis
was performed on all geometries and this produced no
imaginary frequencies for the ground state structures and
only one imaginary frequency for transition states. The Gibbs
free energy of hydrogen in the solution was corrected by
0.00301 hartree to account for the change from 1 atm to 1 M
CH of imid., isomer A), 121.11 (5-CH of imid., isomer B),
20.47 (4-CH of imid., isomer B), 81.73 (C Me ), 68.48 (CH
1
5
5
near imid., isomer B), 68.39 (CH near imid., isomer A), 68.30
NCCH , isomer B), 67.60 (NCCH , isomer A), 67.30 (CH
(
3
3
near NH , isomer A), 59.72 (CH near NH , isomer B), 38.46
2
2
(
CH of NHC, isomer A), 37.78 (CH of NHC, isomer B),
3
3
1
4
0.73 (CH of Cp*, isomer B), 10.51 (CH of Cp*, isomer A),
3 3
15 1
.40 (NCCH3). N { H} NMR (DCM-d , 298 K, δ): 6.53
2
1
9
(NH , isomer A), 0.22 (NH , isomer B). F NMR (DMSO-d ,
2
2
6
93
2
5
98 K, δ): −72.69 (d, PF ). MS (ESI, methanol; m/z):
solution. Three-dimensional visualization of optimized geo-
metries were constructed using ChemCraft.
6
+
14.1801 [M − CH CN − PF ] . Anal. Calcd for
3
6
6
840
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842

ACS Catalysis
ASSOCIATED CONTENT
Research Article
(
2
(
4
(
3
(
17) O, W. W. N.; Lough, A. J.; Morris, R. H. Organometallics 2009,
■
* Supporting Information
8, 6755−6761.
18) O, W. W. N.; Lough, A. J.; Morris, R. H. Chem. Commun. 2009,
6, 8240−8242.
19) O, W. W. N.; Lough, A. J.; Morris, R. H. Organometallics 2013,
2, 3808−3818.
20) Ohara, H.; O, W. W. N.; Lough, A. J.; Morris, R. H. Dalton
Trans. 2012, 41, 8797−8808.
21) Wan, K. Y.; Lough, A. J.; Morris, R. H. Organometallics 2016, 35,
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.7b02346.
NMR spectra, supporting figures, full Gaussian 09
reference, Cartesian coordinates, and free energies and
enthalpies for optimized structures (PDF)
Crystallographic data (CIF)
(
1
(
604−1612.
22) O, W. W. N.; Morris, R. H. ACS Catal. 2013, 3, 32−40.
23) O, W. W. N.; Lough, A. J.; Morris, R. H. Organometallics 2012,
(
AUTHOR INFORMATION
■
3
(
1, 2137−2151.
Corresponding Author
E-mail: rmorris@chem.utoronto.ca.
24) Noyori, R.; Yamakawa, M.; Hashiguchi, S. J. Org. Chem. 2001,
*
6
(
6, 7931−7944.
25) Dub, P. A.; Gordon, J. C. Dalton Trans. 2016, 45, 6756−6781.
26) Yoshimura, M.; Tanaka, S.; Kitamura, M. Tetrahedron Lett.
ORCID
(
2
(
Robert H. Morris: 0000-0002-7574-9388
Notes
The authors declare no competing financial interest.
014, 55, 3635−3640.
27) Haack, K. J.; Hashiguchi, S.; Fujii, A.; Ikariya, T.; Noyori, R.
Angew. Chem., Int. Ed. Engl. 1997, 36, 285−288.
28) Sterk, D.; Stephan, M.; Mohar, B. Org. Lett. 2006, 8, 5935−
5938.
29) Jahier-Diallo, C.; Morin, M. S. T.; Queval, P.; Rouen, M.; Artur,
̌
(
ACKNOWLEDGMENTS
■
(
NSERC Canada is thanked for a Discovery Grant to R.H.M.
and two graduate scholarships to K.Y.W. The authors
acknowledge the Canadian Foundation for Innovation, project
number 19119, and the Ontario Research Fund for funding of
the Centre for Spectroscopic Investigation of Complex Organic
Molecules and Polymers, and the facilities of the Shared
Hierarchical Academic Research Computing Network
́ ́
I.; Querard, P.; Toupet, L.; Crevisy, C.; Basle, O.; Mauduit, M. Chem. -
Eur. J. 2015, 21, 993−997.
(
30) Bourissou, D.; Guerret, O.; Gabbaï, F. P.; Bertrand, G. Chem.
Rev. 2000, 100, 39−92.
31) Herrmann, W. A. Angew. Chem., Int. Ed. 2002, 41, 1290−1309.
(32) Corberan, R.; Lillo, V.; Mata, J. A.; Fernandez, E.; Peris, E.
(
́
Organometallics 2007, 26, 4350−4353.
(
SHARCNET: www.sharcnet.ca) and Compute/Calcul Cana-
(33) Lee, K.-s.; Brown, M. K.; Hird, A. W.; Hoveyda, A. H. J. Am.
Chem. Soc. 2006, 128, 7182−7184.
34) Li, S.-J.; Zhong, J.-H.; Wang, Y.-G. Tetrahedron: Asymmetry
da. R.H.M. thanks the Canada Council for the Arts for a Killam
Research Fellowship.
(
2
(
2
(
2
(
5
006, 17, 1650−1654.
35) Cesar, V.; Bellemin-Laponnaz, S.; Gade, L. H. Chem. Soc. Rev.
004, 33, 619−636.
36) Usui, K.; Yoshida, T.; Nakada, M. Tetrahedron: Asymmetry 2016,
7, 107−113.
37) Wysocki, J.; Ortega, N.; Glorius, F. Angew. Chem., Int. Ed. 2014,
3, 8751−8755.
38) Benhamou, L.; Besnard, C.; Ku
3, 260−266.
39) Schumacher, A.; Bernasconi, M.; Pfaltz, A. Angew. Chem., Int. Ed.
REFERENCES
■
(
2
1) Ager, D. J.; de Vries, A. H. M.; de Vries, J. G. Chem. Soc. Rev.
012, 41, 3340−3380.
2) Busacca, C. A.; Fandrick, D. R.; Song, J. J.; Senanayake, C. H.
Adv. Synth. Catal. 2011, 353, 1825−1864.
3) Clapham, S. E.; Hadzovic, A.; Morris, R. H. Coord. Chem. Rev.
004, 248, 2201−2237.
4) Jolley, K. E.; Zanotti-Gerosa, A.; Hancock, F.; Dyke, A.; Grainger,
(
(
2
(
(
3
(
̈
ndig, E. P. Organometallics 2014,
D. M.; Medlock, J. A.; Nedden, H. G.; Le Paih, J. J. M.; Roseblade, S.
J.; Seger, A.; Sivakumar, V.; Prokes, I.; Morris, D. J.; Wills, M. Adv.
Synth. Catal. 2012, 354, 2545−2555.
2
(
013, 52, 7422−7425.
40) Zhu, Y.; Burgess, K. Acc. Chem. Res. 2012, 45, 1623−1636.
41) DePasquale, J.; White, N. J.; Ennis, E. J.; Zeller, M.; Foley, J. P.;
(
(5) Klingler, F. D. Acc. Chem. Res. 2007, 40, 1367−1376.
(6) Malacea, R.; Poli, R.; Manoury, E. Coord. Chem. Rev. 2010, 254,
Papish, E. T. Polyhedron 2013, 58, 162−170.
42) Bonnet, L. G.; Douthwaite, R. E.; Hodgson, R.; Houghton, J.;
Kariuki, B. M.; Simonovic, S. Dalton Trans. 2004, 3528−3535.
43) Wang, W. F.; Zhang, T.; Wang, F. J.; Shi, M. Tetrahedron 2011,
7, 1523−1529.
44) Li, T.; Bergner, I.; Haque, F. N.; Zimmer-De Iuliis, M.; Song,
D.; Morris, R. H. Organometallics 2007, 26, 5940−5949.
45) Hamilton, R. J.; Bergens, S. H. J. Am. Chem. Soc. 2008, 130,
1979−11987.
46) Pannetier, N.; Sortais, J.-B.; Dieng, P. S.; Barloy, L.; Sirlin, C.;
Pfeffer, M. Organometallics 2008, 27, 5852−5859.
47) John, J. M.; Takebayashi, S.; Dabral, N.; Miskolzie, M.; Bergens,
S. H. J. Am. Chem. Soc. 2013, 135, 8578−8584.
48) Sandoval, C. A.; Ohkuma, T.; Muniz, K.; Noyori, R. J. Am.
Chem. Soc. 2003, 125, 13490−13503.
49) Clapham, S. E.; Hadzovic, A.; Morris, R. H. Coord. Chem. Rev.
(
7
(
29−752.
7) Ohkuma, T.; Ooka, H.; Hashiguchi, S.; Ikariya, T.; Noyori, R. J.
(
6
(
Am. Chem. Soc. 1995, 117, 2675−2676.
(
(
8) Xie, J. H.; Bao, D. H.; Zhou, Q. L. Synthesis 2015, 47, 460−471.
9) Vilhanova, B.; Vaclavik, J.; Sot, P.; Pechacek, J.; Zapal, J.; Pazout,
R.; Maixner, J.; Kuzma, M.; Kacer, P. Chem. Commun. 2016, 52, 362−
65.
10) Clarke, M. L.; Roff, G. J. In The Handbook of Homogeneous
(
1
(
3
(
Hydrogenation; de Vries, J. G., Elsevier, C. J., Eds. Wiley-VCH Verlag
GmbH: Weinheim, 2008; pp 413−454.
(
(
11) Lagaditis, P. O.; Sues, P. E.; Sonnenberg, J. F.; Wan, K. Y.;
Lough, A. J.; Morris, R. H. J. Am. Chem. Soc. 2014, 136, 1367−1380.
(
̃
(
12) Zuo, W.; Lough, A. J.; Li, Y. F.; Morris, R. H. Science 2013, 342,
1
(
(
080−1083.
(
13) Morris, R. H. Acc. Chem. Res. 2015, 48, 1494−1502.
14) Hopkinson, M. N.; Richter, C.; Schedler, M.; Glorius, F. Nature
2004, 248, 2201−2237.
2
(
2
(
014, 510, 485−496.
15) Wang, F.; Liu, L. J.; Wang, W.; Li, S.; Shi, M. Coord. Chem. Rev.
012, 256, 804−853.
16) Zhao, D.; Candish, L.; Paul, D.; Glorius, F. ACS Catal. 2016, 6,
978−5988.
(50) Prokopchuk, D. E.; Morris, R. H. Organometallics 2012, 31,
7375−7385.
(51) Prokopchuk, D. E.; Sonnenberg, J. F.; Meyer, N.; Zimmer-De
Iuliis, M.; Lough, A. J.; Morris, R. H. Organometallics 2012, 31, 3056−
3064.
5
6
841
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842

ACS Catalysis
Research Article
(
52) Bidal, Y. D.; Santoro, O.; Melaimi, M.; Cordes, D. B.; Slawin, A.
M. Z.; Bertrand, G.; Cazin, C. S. J. Chem. - Eur. J. 2016, 22, 9404−
409.
53) Bullough, E. K.; Little, M. A.; Willans, C. E. Organometallics
013, 32, 570−577.
54) Santoro, O.; Lazreg, F.; Cordes, D. B.; Slawin, A. M. Z.; Cazin,
C. S. J. Dalton Trans. 2016, 45, 4970−4973.
55) Mormul, J.; Steimann, M.; Nagel, U. Eur. J. Inorg. Chem. 2014,
014, 1389−1393.
56) Boehme, C.; Frenking, G. Organometallics 1998, 17, 5801−
809.
57) Cordero, B.; Gomez, V.; Platero-Prats, A. E.; Reves, M.;
Echeverria, J.; Cremades, E.; Barragan, F.; Alvarez, S. Dalton Trans.
008, 2832−2838.
58) Pamies, O.; Bac
59) Takebayashi, S.; Dabral, N.; Miskolzie, M.; Bergens, S. H. J. Am.
Chem. Soc. 2011, 133, 9666−9669.
60) Bertoli, M.; Choualeb, A.; Lough, A. J.; Moore, B.; Spasyuk, D.;
Gusev, D. G. Organometallics 2011, 30, 3479−3482.
61) Hopmann, K. H.; Bayer, A. Coord. Chem. Rev. 2014, 268, 59−
2.
(88) Frisch, M. J., et al. Gaussian 09 Revision B.01; Gaussian, Inc.:
Wallingford, CT, 2010.
(89) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105,
9
(
2
(
2999−3093.
̀
(90) Tomasi, J.; Mennucci, B.; Cances, E. J. Mol. Struct.:
THEOCHEM 1999, 464, 211−226.
91) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B
(
2
(
(
009, 113, 6378−6396.
(
2
(
5
92) Fukui, K. Acc. Chem. Res. 1981, 14, 363−368.
93) Dub, P. A.; Henson, N. J.; Martin, R. L.; Gordon, J. C. J. Am.
Chem. Soc. 2014, 136, 3505−3521.
(
2
(
(
̀
̈
kvall, J. E. Chem. - Eur. J. 2001, 7, 5052−5058.
(
(
8
(62) Morris, R. H. Chem. Rev. 2016, 116, 8588−8654.
(63) Hadzovic, A.; Song, D.; MacLaughlin, C. M.; Morris, R. H.
Organometallics 2007, 26, 5987−5999.
(
(
(
64) Morris, R. H. Coord. Chem. Rev. 2008, 252, 2381−2394.
65) Morris, R. H. Coord. Chem. Rev. 2009, 253, 1219.
66) Jia, G.; Lough, A. J.; Morris, R. H. Organometallics 1992, 11,
1
(
5
(
2
(
1
61−171.
67) Chinn, M. S.; Heinekey, D. M. J. Am. Chem. Soc. 1990, 112,
166−5175.
68) Abdur-Rashid, K.; Lough, A. J.; Morris, R. H. Organometallics
000, 19, 2655−2657.
69) Peverati, R.; Truhlar, D. G. J. Phys. Chem. Lett. 2012, 3, 117−
24.
(
(
70) Zhu, L.; Qi, X.; Lan, Y. Organometallics 2016, 35, 771−777.
71) Luo, X.; Bai, R.; Liu, S.; Shan, C.; Chen, C.; Lan, Y. J. Org. Chem.
2
016, 81, 2320−2326.
72) Shan, C.; Luo, X.; Qi, X.; Liu, S.; Li, Y.; Lan, Y. Organometallics
016, 35, 1440−1445.
73) Peterson, K. A.; Figgen, D.; Dolg, M.; Stoll, H. J. Chem. Phys.
007, 126, 124101.
74) Frisch, M. J.; Pople, J. A.; Binkley, J. S. J. Chem. Phys. 1984, 80,
265−3269.
75) Lynch, B. J.; Zhao, Y.; Truhlar, D. G. J. Phys. Chem. A 2003, 107,
384−1388.
76) Haack, K.-J.; Hashiguchi, S.; Fujii, A.; Ikariya, T.; Noyori, R.
Angew. Chem., Int. Ed. Engl. 1997, 36, 285−288.
77) O, W. W. N.; Lough, A. J.; Morris, R. H. Organometallics 2012,
1, 2152−2165.
78) Guo, R.; Chen, X.; Elpelt, C.; Song, D.; Morris, R. H. Org. Lett.
005, 7, 1757−1759.
79) Zhang, L.; Han, Z.; Zhao, X.; Wang, Z.; Ding, K. Angew. Chem.,
Int. Ed. 2015, 54, 6186−6189.
80) Clapham, S. E.; Guo, R.; Zimmer-De Iuliis, M.; Rasool, N.;
Lough, A.; Morris, R. H. Organometallics 2006, 25, 5477−5486.
81) Kumar, P.; Singh, A. K.; Yadav, M.; Li, P.-z.; Singh, S. K.; Xu, Q.;
Pandey, D. S. Inorg. Chim. Acta 2011, 368, 124−131.
82) Guo, R.; Lough, A. J.; Morris, R. H.; Song, D. Organometallics
004, 23, 5524−5529.
83) Picot, A.; Dyer, H.; Buchard, A.; Auffrant, A.; Vendier, L.; Le
(
2
(
2
(
3
(
1
(
(
3
(
2
(
(
(
(
2
(
Floch, P.; Sabo-Etienne, S. Inorg. Chem. 2010, 49, 1310−1312.
(
(
84) Kozuch, S.; Shaik, S. Acc. Chem. Res. 2011, 44, 101−110.
85) O, W. W. N.; Lough, A. J.; Morris, R. H. Organometallics 2011,
3
(
2
(
0, 1236−1252.
86) Marco, A.; Compan
003, 142, 13−19.
87) Demmans, K. Z.; Ko, O. W. K.; Morris, R. H. RSC Adv. 2016, 6,
8580−88587.
́
̃
́
o, R.; Rubio, R.; Casals, I. Microchim. Acta
8
6
842
DOI: 10.1021/acscatal.7b02346
ACS Catal. 2017, 7, 6827−6842