Mechanism of Asymmetric Hydrogenation of Aromatic Ketones Catalyzed by a Combined System of Ru(π-CH2C(CH3)CH2)2(cod) and the Chiral sp2N/sp3NH Hybrid Linear N4 Ligand Ph-BINAN-H-Py

Article abstract of DOI:10.1021/jacs.5b02350

The combination of a Goodwin-Lions-type chiral N4 ligand, (R)-Ph-BINAN-H-Py ((R)-3,3′-diphenyl-N²,N^2′-bis((pyridin-2-yl)methyl)-1,1′-binaphthyl-2,2′-diamine; L), with Ru(π-CH₂C(CH₃)CH₂)₂(co

Full text of DOI:10.1021/jacs.5b02350

Article
pubs.acs.org/JACS
Mechanism of Asymmetric Hydrogenation of Aromatic Ketones
Catalyzed by a Combined System of Ru(π-CH2C(CH3)CH2)2(cod) and
the Chiral sp2N/sp3NH Hybrid Linear N4 Ligand Ph-BINAN-H-Py
Hiroshi Nakatsuka,† Tomoya Yamamura,† Yoshihiro Shuto,† Shinji Tanaka,† Masahiro Yoshimura,‡
and Masato Kitamura*,†
†Graduate School of Pharmaceutical Sciences, Graduate School of Science, and Research Center for Materials Science, Nagoya
University, Chikusa, Nagoya 464-8601, Japan
‡Division of Liberal Arts and Sciences, Aichi Gakuin University, Iwasaki, Nisshin 470-0195, Japan
S
* Supporting Information
ABSTRACT: The combination of a Goodwin−Lions-type
chiral N4 ligand, (R)-Ph-BINAN-H-Py ((R)-3,3′-diphenyl-
N2,N2′-bis((pyridin-2-yl)methyl)-1,1′-binaphthyl-2,2′-diamine;
L), with Ru(π-CH2C(CH3)CH2)2(cod) (A) (cod = 1,5-
cyclooctadiene) catalyzes the hydrogenation of acetophenone
(AP) to (R)-1-phenylethanol (PE) with a high enantiomer
ratio (er). Almost no Ru complex forms, with A and L
remaining intact throughout the reaction while generating PE
quantitatively according to [PE] = kobst2. An infinitesimal
amount of reactive and unstable RuH2L (B) with C2-Λ-cis-α
stereochemistry is very slowly and irreversibly generated from
A by the action of H2 and L, which rapidly catalyzes the hydrogenation of AP via Noyori’s donor−acceptor bifunctional
mechanism. A CH-π-stabilized Si-face selective transition state, CSi, gives (R)-PE together with an intermediary Ru amide, D,
which is inhibited predominantly by formation of the Ru enolate of AP. The rate-determining hydrogenolysis of D completes the
cycle. The time-squared term relates both to the preliminary step before the cycle and to the cycle itself, with a highly unusual
eight-order difference in the generation and turnover frequency of B. This mechanism is fully supported by a series of
experiments including a detailed kinetic study, rate law analysis, simulation of t/[PE] curves with fitting to the experimental
observations at the initial reaction stage, X-ray crystallographic analyses of B-related octahedral metal complexes, and Hammett
plot analyses of electronically different substrates and ligands in their enantioselectivities.
intermolecular-type donor−acceptor bifunctional catalyst (In-
termol-DACat) (Figure 1a).1c,7
1. INTRODUCTION
Since the BINAP−Ru method made its debut in 1987,1
homogeneous asymmetric hydrogenation of ketones has
become established as one of the most reliable strategies for
the synthesis of optically active secondary alcohols. The first
report on homogeneous ketone hydrogenation dates back as far
as 1938, when Calvin demonstrated the catalytic activity of
CuOCOCH3.2 In the shadow of immense efforts toward the
development of homogeneous catalysts for olefin hydro-
genation using Rh, Ru, or Pt complexes in combination with
various organic ligands,3 little progress in this area was made
until 1970, when Schrock and Osborn discovered the high
utility of monocationic RhH2(PPhMe2)2(solvent)2 in the
hydrogenation of acetone to 2-propanol.4 Their discovery
triggered an increase in the number of studies using chiral
phosphines,5 eventually leading to the revolutionary catalyst
system consisting of Ru(OCOCH3)2(BINAP) and a strong
Brønsted acid.1 The first generation of BINAP−Ru chemistry
realized the hydrogenation of β-keto esters and a wide range of
functionalized ketones with nearly perfect enantioselectivity,6
the successful development of which relies on the concept of an
A second revolution took place in 1995, when a chiral Ru
8
catalyst consisting of RuCl2(BINAP)(dmf)n and 1,2-diphenyl-
ethane-1,2-diamine (DPEN) was developed.9 This BINAP−
Ru−DPEN ternary system, which was based on the concept of
an intramolecular-type donor−acceptor bifunctional catalyst
(Intramol-DACat) (Figure 1b), shows high efficiency in the
hydrogenation of unfunctionalized ketones. Since the start of
this apparent trend of BINAP−Ru systems, a vast number of
hydrogenation-active Ru complexes with trivalent phosphorus
ligating atoms have been reported.10 Their effectiveness has
been examined and compared to that of the “privileged”11
BINAP chiral catalyst and its derivatives. Nevertheless, no
omnipotent catalyst exists even now, and new frameworks for
designing asymmetric catalysts are required.
From this viewpoint, nitrogen-based ligands have hidden
potential because they have not been extensively researched as
Received: March 5, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b02350
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
tetradentate chiral ligand (R)-L can form five geometrical
isomers in an octahedral metal complex13C2-Λ-cis-α-, C2-Δ-
cis-α-, C1-Δ-cis-β-, C1-Λ-cis-β-, and C2-trans-isomersalthough
the C2 symmetric Δ-cis-α and C1 symmetric Λ-cis-β isomers are
unlikely because of their highly distorted structures. These
isomers are in equilibrium with each other (Figure 3a), and the
Figure 1. Two epoch-making asymmetric hydrogenation reactions of
ketones and the underlying concepts.
yet, although Ohgo et al. reported the first asymmetric
hydrogenation of ketones catalyzed by Co(II)-
(bisdimethylglyoximato)/quinine in 1971,12 around the same
time as Schlock−Osborn’s report.4 After three unproductive
decades, there is now increasing attention on all nitrogen-
ligating systems. The Goodwin−Lions-type tetradentate sp2N/
sp3NH linear N4 ligand R-BINAN-R′-Py (3,3′-R,R-N2,N2′-
bis((6-R′-pyridin-2-yl)methyl)-1,1′-binaphthyl-2,2′-dia-
mine)13,14 is one of the key examples setting a precedent in this
new trend.15 A combination of the nonphosphine ligand Ph-
BINAN-H-Py (L) and the Ru π-allyl complex Ru(π-CH2C-
(CH3)CH2)2(cod) (A)16 quantitatively hydrogenates various
aromatic ketones with a substrate/catalyst (S/C) ratio of
>10 000 to give the secondary alcohol with an enantiomer ratio
(er) of up to >99:1.13 Most interestingly, the reaction proceeds
with almost no formation of any Ru complexes. Here, we would
like to present a mechanism explaining this unusual
phenomenon and the origin of enantioselection in the (R)-L/
A-catalyzed asymmetric hydrogenation of acetophenone (AP)
to (R)-1-phenylethanol ((R)-PE) (Figure 2).
Figure 3. Equilibrium between three stereoisomers of (R)-Ph-BINAN-
H-Py−RuH2 (B) (a) and (R)-H-BINAN-H-Py−RuH2 (b). Red arrows
indicate a steric repulsion.
overall catalyst performance is the average of each isomer’s own
reactivity and enantioselectivity.17 A density functional theory
(DFT) calculation of an imaginary RuH2((R)-L) (B) complex
using the B3LYP/LANL2DZ set showed that the relative Gibbs
free energies of Δ-cis-β-B and trans-B isomers are, respectively,
14.8 and 19.2 kcal/mol higher than that of Λ-cis-α-B. Steric
repulsion (red arrow) between the C(3) phenyl group of the
naphthalene ring and the pyridylmethyl group in the Δ-cis-β
and trans isomers would shift the equilibrium to the Λ-cis-α-B
side. The energy difference indicates that the Δ-cis-β-B and
trans-B isomers exist at, respectively, only one-trillionth and
one hundred-trillionth of the level of Λ-cis-α-B. This property is
advantageous for unifying the catalytic species to enhance
catalyst performance. Regarding the original Goodwin−Lions
ligand,14 which has no 3,3′-phenyl substituents of (R)-L, the C1
symmetric Δ-cis-β isomer is the most stable (Figure 3b)
according to the same DFT calculation. The existence of the
two different possible reaction sites in the Δ-cis-β isomer may
complicate the reaction pathways. In this sense, the C2
symmetric Λ-cis-α-B isomer is simple.
2. RESULTS AND DISCUSSION
2.1. Characteristic Features of (R)-Ph-BINAN-H-Py ((R)-
L). 2.1.1. Λ-cis-α Selectivity. Theoretically, the linear
2.1.2. Metal Capturing Ability. [Ru((R)-L)(CH3CN)2]-
(PF6)2 (1) was quantitatively prepared by mixing (R)-L and
18
[Ru(C6H6)(CH3CN)3](PF6)2 in CH3CN at room temper-
ature (rt) for 48 h, and yellowish prism crystals (melting point
(mp) = 245 °C) were obtained in 68% yield from a CHCl3−
THF (THF = tetrahydrofuran) solvent system. Ru-
(OCOCH3)2-((R)-L) (2) was also quantitatively obtained
19
from Ru2(OCOCH3)4 (CH3OH, rt/48 h then 60 °C/2 h)
and crystallized as reddish-brown platelets (mp = 225 °C) in
60% yield from CH2Cl2 and hexane. In a similar way,
Mn(OTf)2((R)-L) (60 °C, 30 min, CH3CN), Fe(OTf)2((R)-
L) (25 °C, 1 h, THF), and Cu(OTf)2((R)-L) (25 °C, 1 h,
C2H5OH) were prepared.20 The molecular structures in crystal
are shown in Figure 4, reflecting a general tendency of (R)-Ph-
BINAN-H-Py ligand ((R)-L) to form the C2-Λ-cis-α isomer.
Figure 2. Standard conditions used for this mechanistic study of
asymmetric hydrogenation of acetophenone (AP) catalyzed by the
(R)-Ph-BINAN-H-Py/Ru(π-CH2C(CH3)CH2)2(cod) ((R)-L/A)
combined system (PE = 1-phenylethanol).
B
DOI: 10.1021/jacs.5b02350
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(irradiation of HS) and 1.0% (irradiation of C(8)H); (R)-2,
3.5% (irradiation of HS)].
2.2. Proposed Reactive Species: Λ-cis-α-RuH2((R)-L)
(B). 2.2.1. Catalyst Performance. The above preformed Ru
complexes (2 mM), Λ-cis-α-[Ru((R)-L)(CH3CN)2](PF6)2 (1)
and Λ-cis-α-Ru(OCOCH3)2((R)-L) (2), quantitatively hydro-
genate acetophenone (AP) (2 M) in 2-propanol (iPrOH)
containing potassium tert-butoxide (tBuOK) (20 mM) to give
(R)-1-phenylethanol ((R)-PE) and (S)-PE with an er of, at
most, 93:7 for complex 1 (H2, 100 atm, 50 °C, 24 h) and 84:16
for complex 2 (H2, 50 atm, 25 °C, 24 h).20 In contrast to the
present (R)-L/A combined system (Figure 2) and against
expectation, the catalyst performance of that system was low.
The RuX2−L-type complexes require an excess amount of base,
because the balance point lies toward the left side in the
equilibria of both “RuX2−L + H2 ⇄ RuHX−L + HX” and
“RuHX−L + H2 ⇄ RuH2−L + HX”.21a,22c The reaction system
might be complicated by the remaining tBuOK and coproduced
salts such as KPF6 and KOCOCH3, which might disturb the
clean generation of the presumed reactive species, Λ-cis-α-
RuH2((R)-L) (B).
2.2.2. Trial for Detection of B. Related mechanistic studies
on the hydrogenation of ketones using Ru complexes of
phosphine-containing ligands suggest that a RuH2 mechanism
operates.21−24 Furthermore, RuH2 complexes can be prepared
in high yields from Ru(π-CH2C(CH3)CH2)2(diphosphine) and
Ru(π-CH2C(CH3)CH2)2(cod) (A)/PNP ligand.25 On the
basis of this information, we tried to detect either RuH2((R)-
L) (B) or any Ru complexes from A and L with or without H2
and at 25 or 60 °C, but in vain (Table 1).20 Nothing occurred
at 25 °C, but a black precipitate, most probably a Ru(0) cluster
Table 1. Results of the Trial for Detection of Ph-BINAN-H-
a
Py−RuH2 (B)
b
b
% convn
% yield
c
c
entry
solvent
H2, atm temp, °C
A
L
IB IBD
1
2
(CD3)2CDOH
(CD3)2CDOH
(CD3)2CDOH
(CD3)2CDOH
(CD3)2CDOH
C6D6
0
0
25
60
60
25
60
25
60
60
25
60
25
60
60
25
60
25
60
60
25
60
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
8
0
1
2
0
0
0
6
6
0
0
0
3
d
21
0
e
3
0
0
4
5
6
7
50
50
0
0
0
d
d
25
0
0
0
Figure 4. Molecular structures of (a) [Ru((R)-Ph-BINAN-H-Py)-
(CH3CN)2](PF6)2, (b) Ru(OCOCH3)2((R)-Ph-BINAN-H-Py), (c)
Mn(OTf)2((R)-Ph-BINAN-H-Py), (d) Fe(OTf)2((R)-Ph-BINAN-H-
Py), and (e) Cu(OTf)2((R)-Ph-BINAN-H-Py) in the crystalline state.
OTf = OSO2CF3.
C6D6
0
42
0
e
d
8
C6D6
0
5
2
9
C6D6
50
50
0
0
0
d
d
10
11
12
C6D6
32
0
4
THF-d8
0
THF-d8
0
63
3
1
The structure of Λ-cis-α in solution was confirmed by H
e
d
13
THF-d8
0
6
0
NMR analyses of two diamagnetic Ru complexes, [Ru((R)-
L)(CH3CN)2](PF6)2 (1) and Ru(OCOCH3)2((R)-L) (2). For
both complexes, only one set of signals was observed in the 1H
NMR spectra:20 (R)-1 (δ 4.23 (dd, J = 18.6 and 5.5 Hz,
CHRHS), 4.58 (dd, J = 18.6 and 8.9 Hz, CHRHS), 6.35 (dd, J =
8.9 and 5.5 Hz, NH)) and (R)-2 (δ 1.46 (s, CH3COO), 3.97
(dd, J = 17.2 and 6.0 Hz, CHRHS), 4.29 (dd, J = 17.2 and 8.1
Hz, CHRHS), 9.13 (dd, J = 8.1 and 6.0 Hz, NH)). These simple
spectra were consistent with both complexes 1 and 2 having a
symmetric C2-Λ-cis-α structure. In addition, this Λ-cis-α
geometry in solution phase was supported by the observation
of a nuclear Overhauser effect (NOE) between the closely
located naphthalene C(8)H and methylene HS [(R)-1, 1.6%
14
15
THF-d8
50
50
0
0
0
d
THF-d8
55
0
16
f
16
hexane
f
d
d
17
hexane
0
41
46
0
e
,f
18
19
20
hexane
0
f
hexane
50
50
f
d
hexane
41
a
b
Conditions: [A] = [L] = 10 mM; 24 h. Determined by the 1H NMR
c
analysis. IB = isobutene, IBD = isobutene dimer (2,5-dimethylhexa-
d
1,5-diene). Black precipitate was observed. No 1H NMR signal except
e
f
for those of A and L was detected. In the absence of L. 1H NMR was
measured in C6D6 after evaporation.
C
DOI: 10.1021/jacs.5b02350
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
1
(“Ru(0)”) or a RuH cluster (“RuH”), was generated at an
2.2.3. In Situ NMR Study. Figure 6 shows the H NMR
elevated temperature. Figure 5 shows the possible pathways
spectra of the reaction mixture before and after introduction of
Figure 5. Supposed decomposition pathways from A, (R)-L, and H2 to
a black precipitate. L′ = cod, solvent, H2, (R)-L, etc.; n = 1−4.
from A, (R)-L, and H2 to the black precipitate. A′ is an
imaginary bismethallyl species, in which cod of A is fully or
partly replaced with solvent, H2, (R)-L, etc. In route (a),
coupling of the two methallyl ligands in A or A′ gives isobutene
dimer (IBD), 2,5-dimethylhexa-1,5-diene, and “Ru(0)”.26 In
route (b), A or A′ is converted to isobutene (IB) and a
trimethylenemethane Ru(II) species,27 which decomposes to
“Ru(0)” via “RuH”. In route (c), A or A′ directly reacts with H2
to generate IB and “RuH”,25 which decomposes to “Ru(0)”. In
all routes, the decomposition may proceed via RuH2((R)-L)
(B).
Figure 6. In situ 1H NMR monitoring of the hydrogenation of
acetophenone (AP, 75 mM) to 1-phenylethanol (PE) in
(CD3)2CDOD containing (R)-Ph-BINAN-H-Py ((R)-L, 10 mM),
Ru(π-CH2C(CH3)CH2)2(cod) (A, 10 mM), and tBuOK (6.7 mM) at
25 °C. (a) Spectrum obtained 1.5 h after allowing the reaction mixture
to stand without H2. (b) Spectrum obtained 27 h after applying 50 atm
of H2 pressure to the reaction in (a) in a pressure-tight sapphire tube.
The signal/noise ratio in spectrum (b) is calculated to be 177, as
In discussing the results shown in Table 1, care must be
taken because (i) IB and IBD were not quantitatively detected
1
and (ii) the H NMR spectra tended to broaden. The black
precipitate may absorb these alkenes to some extent, and
aggregates of RuH species may be formed.25,28 With such
uncertainty in mind, the results were examined. In
(CD3)2CDOH, which is the solvent of choice in the present
asymmetric hydrogenation, no reaction of A with L occurred at
25 °C even after 24 h at 0 or 50 atm of H2 (entries 1 and 4). An
increase in temperature to 60 °C at 50 atm of H2 decomposed
25% of A to generate a black precipitate together with 8% of IB
(entry 5). Little IBD was detected, negating route (a). Removal
of H2 from the condition of entry 5 also led to a black
precipitate, but, in this case, IB and IBD were generated in ca.
1:3 ratio (entry 2). Routes (a−c) may proceed in parallel.
Interestingly, in the absence of L, nothing occurred (entries 2
and 3). The NH proton of L or association of L with A may
accelerate the decomposition.29 A basically similar tendency
was observed in aprotic solvents such as C6D6 and THF-d8
(entries 6−15) in terms of the black precipitate. Unlike the case
with (CD3)2CDOH, however, IBD was detected at 50 atm of
H2 (entries 10 and 15). Routes (b) and (c) are suppressed in
aprotic solvents for some reason. In hexane, A decomposed at
60 °C regardless of whether H2 or L was present or not (entries
16−20).
Although the decomposition occurred only at an elevated
temperature, the series of results in (CD3)2CDOH are not
inconsistent with the view that an infinitesimal amount of B is
formed via hydrogenolysis of methallyl groups by two hydrogen
molecules under the real hydrogenation conditions at 25 °C
(see sections 2.4 and 2.5). As compared with the phosphorus
atom-ligating RuH2, B, which possesses electron-donating and
non-π-acceptable sp3N atoms trans to two hydrides, would be
significantly destabilized to decompose without AP. Further-
more, replacement of cod in A would be more difficult with L
than with the phosphorus-based ligand.
○
△
shown in the inset. The signals are indicated by ((R)-L), (A), ×
□
(AP), and (PE).
H2 ([(R)-L]0 = [A]0 = 10 mM, [tBuOK] = 6.7 mM, [AP]0 = 75
mM, 50 atm H2, (CD3)2CDOD, 25 °C, 27 h). In the 1H NMR
spectrum obtained 1.5 h after the addition of all components
except for H2 (Figure 6a), sets of signals corresponding to A
(△), (R)-L (○), AP (×), and tBuOK were independently
observed, although the CH3 signal intensity of AP had
decreased owing to hydrogen−deuterium (H/D) exchange
via formation of the enolate of AP. After pressurizing the
system to 50 atm of H2, the AP signals (×) disappeared and the
PE product signals (□) appeared, whereas there was no change
in the signals of A or (R)-L. The R/S ratio of the PE product
formed was 86:14. Within the limits of the signal-to-noise (S/
N) ratio (177), no Ru−(R)-L complex formed. Nonetheless,
we believe that a very tiny amount of a chiral and reactive
species B ([B] < 10/177 mM = 0.0565 mM) is slowly
generated under the hydrogenation conditions and is operating
in this catalysis. B would survive only in the presence of the
substrate AP.
2.3. Time−Conversion (t/[PE]) Curves. The time course
of the IR CO stretching band intensity of AP at 1690 cm−1
was obtained by use of a total internal reflectance measurement
system during the hydrogenation of AP using the (R)-L/A
combined catalyst under the standard conditions ([(R)-L]0 =
[A]0 = [tBuOK] = 2 mM, [AP]0 = 2 M, 50 atm H2, iPrOH, 25
°C).20 The data were converted to the t/[PE] curve shown in
Figure 7a, together with those obtained at [AP]0 = 0.5 and 1 M.
D
DOI: 10.1021/jacs.5b02350
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 7. (a) Intriguing phenomenon observed in the t/[PE] relation
at [AP]0 = 0.5, 1, and 2 M (conditions: [(R)-L] = [A] = [tBuOK] = 2
mM, 50 atm H2, iPrOH, 25 °C). (b) Plot of [PE] versus t2.
The following intriguing observations can be made from the
three t/[PE] curves: (i) the velocity of PE production (d[PE]/
dt) increases exponentially in the early stages but follows a
rather straight relationship in the later stages; (ii) d[PE]/dt
does not decelerate even at the very final stage; and (iii)
d[PE]/dt decreases as [AP]0 increases. Observation (i) is not
inconsistent with the scenario that the reactive species B is
gradually generated during the course of the reaction, but it
argues against the possible existence of an acceleration effect of
product PE. The slight rate deceleration (0.28 M h−1 ([AP]0 =
0.5 M), 0.27 M h−1 (1 M), and 0.26 M h−1 (2 M)) in the
straight line region of 70−100% conversion would instead
indicate very weak product inhibition. Observation (ii) implies
a zeroth-order dependence on AP concentration in the reaction
system, while observation (iii) clearly shows that the reaction is
inhibited by the initial concentration of AP ([AP]0). The
difference between the reaction profile in the early stages and
that in the later stages may be ascribed not only to the time-
dependent change in [AP], [PE], and [B] but also to the
decomposition of B at a lower concentration of AP. The rate is
not enhanced as much as expected in the later stages, where the
concentration of B increases. The balance between B
generation and decomposition may result in observation (i).
To simplify the subsequent discussion, we will focus on the
kinetic scheme of the present asymmetric hydrogenation only in the
early stages (0−30% conversion of AP) of the reaction.
2.4. Supposed Catalytic Cycle and Rate Law Analysis.
On the basis of the above experimental results and the
mechanism reported for other Ru(II)-catalyzed hydrogena-
tions,21−24 we supposed the kinetic model shown in Figure 8, in
which there are two processes occurring in parallel. One is the
preliminary step with the rate constant k0 from Ru(π-
CH2C(CH3)CH2)2(cod) (A)/(R)-Ph-BINAN-H-Py ((R)-L)/
H2 to the reactive and unstable Λ-cis-α-B intermediate, and the
other is the catalytic cycle itself for hydrogenation of AP to PE
based on the Intramol-DACat mechanism (Figure 1b). In the
absence of AP, the short-lived B easily decomposes by the
liberation of L, while in the presence of AP, the polarized Hδ+−
Figure 8. Reaction pathway and enantioface selection in the Intramol-
DACat-based hydrogenation of acetophenone (AP) using (R)-Ph-
BINAN-H-Py ((R)-L) and Ru(π-CH2C(CH3)CH2)2(cod) (A).
Nδ−---Ruδ+−Hδ− moiety of the 18e Ru dihydride B captures the
similarly polarized Cδ+Oδ− of AP to move to transition state
C. In this charge-alternating six-atom system, the red and blue
hydrogen atoms are delivered to the CO double bond to
release PE together with generation of the 16e Ru amide D,
which would have a trigonal bipyramidal structure with a Ru
N double-bond character.21c The coordinatively unsaturated D
interacts with H2 in an η2 manner to form F, which undergoes
heterolytic cleavage to reproduce the chain carrier B, thereby
completing the catalytic cycle. The rate of the cycle would be
determined by the D −> B step with k2 and not by the B −> D
step with k1 (k2 ≪ k1).21c The Ru amide D interacts not only
with H2 but also with AP, PE, and iPrOH to form equilibria
with the corresponding Ru enolate and Ru alkoxide EQH and
the equilibrium constant KQH, thereby causing inhibition.21f,23c
In the present particular case, the catalyst precursor A is
consumed by the action of H2 and L to generate B with an
extraordinarily slow rate on the time scale of hydrogenation,
enabling [A] and [L] to be approximated by [A]0 and [L]0,
respectively ([A] ≈ [A]0 and [L] ≈ [L]0). Hydrogen pressure
([H2]) and tBuOK concentration ([tBuOK]) do not change
during the reaction. Therefore, the total concentration of the
Ru species in the cycle ([Rucycle]) can be expressed as
20
k0[A]0 [L]0 [H2]nt. Furthermore, under the condition of k2
l
m
E
DOI: 10.1021/jacs.5b02350
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 9. Dependence of the rate constant kobs on [Ru(π-CH2C(CH3)CH2)2(cod) (A)]0, [(R)-Ph-BINAN-H-Py ((R)-L)]0, H2 pressure (pH2), and
[acetophenone (AP)]0 in the hydrogenation of AP in iPrOH at 25 °C. (a) Plot of kobs as a function of [A]0 from 0.5 to 4 mM ([(R)-L]0 = [tBuOK]
= 2 mM, [AP]0 = 2 M, 50 atm H2). (b) Plot of kobs as a function of [(R)-L]0 from 0.1 to 4 mM ([A]0 = [tBuOK] = 2 mM, [AP]0 = 2 M, 50 atm H2).
(c) Plot of kobs as a function of pH2 from 25 to 90 atm ([A]0 = [(R)-L]0 = [tBuOK] = 2 mM, [AP]0 = 2 M). (d) Plot of kobs as a function of pH23 of
graph (c). (e) Plot of kobs as a function of [AP]0 from 0.5 to 4 M ([A]0 = [(R)-L]0 = [tBuOK] = 2 mM, 50 atm H2). (f) Plot of kobs as a function of
1/[AP]0 of graph (e).
≪ k1, the Rucycle is distributed mainly to D and EQH ([Rucycle] =
[D] + Σ[EQH] = [D](1 + ΣKQH[QH])), leading to eq 1.
accurate cubic function. Figure 9d shows a third-order plot
affording a straight line from the origin with a correlation
coefficient of 0.978, establishing n + 1 = 3 in eq 2. In contrast to
the acceleration effect of pH2 on the rate, the kobs value
exponentially decreases as [AP]0 increases from 0.5 to 4 M
(Figure 9e). Replotting the kobs values versus the reciprocal of
[AP]0 established a reasonable linearity with a correlation
coefficient of 0.987 (Figure 9f). The −1st-order dependence on
[AP]0 is consistent with eq 2. The rates observed with [AP]0 at
3 and 4 M were rather slower than those expected from the
0.5−2 M data. This deviation is probably due to changes in the
properties of the solvent, which comprises iPrOH and AP in a
2:1 to 1:1 ratio, because the characteristics of AP exert a
significant effect on the polarity and dielectronic constant of the
solvent system.23d In all cases, the high enantioselectivity was
maintained with [A]0, [L]0, pH2, and [AP]0 varied (Figure 9,
red squares).
l
[D] = k0[A]0 [L]0m[H ]nt/(1 +
KQH[QH])
∑
2
(1)
Steady-state assumption for D gives d[PE]/dt = k2[D][H2] =
k0k2[A]0 [L]0 [H2]n+1t /(1 + ΣKQH[QH]). With the definitions
of ΣKQH[QH] = KAP[AP] + K(R)‑PE[(R)-PE] + K(S)‑PE[(S)-PE]
+ KiPrOH[iPrOH] and KPE[PE] = K(R)‑PE[(R)-PE] + K(S)‑PE[(S)-
PE] = ZKAP[PE], integration of the d[PE]/dt equation from 0
to t followed by the second-order Taylor expansion at t = 0
gives eq 2.20
l
m
l
[PE] = (k0k2[A]0 [L]0m[H2]n+1/2(KAP[AP]0 + KiPrOH[iPrOH]
+ 1))t2 = kobst2
(2)
Consistent with eq 2, time-squared plotting of the t/[PE]
relations in Figure 7a gave straight lines, at least in the early
reaction stages (Figure 7b), implying that the preliminary step
from (R)-L/A/H2 to B and the catalytic cycle converting AP to
PE are operating in parallel.
2.5. Reaction Orders for [A]0, [L]0, [H2] (pH2), and [AP]0.
The time-squared equation, [PE] = kobst2, was used to deduce
the reaction orders for A, (R)-L, pH2, and AP. The values of
Equation 2 can therefore be rewritten as eq 3.
[PE] = (k0k2[A]0[H2]3/2(KAP[AP]0 + KiPrOH[iPrOH] + 1))t2
= kobst2
(3)
2.6. Inhibitory Effect. Equation 3 applies only in the early
stages of the reaction; therefore, no [PE] term is included.
Inhibition of the catalytic activity, however, should be related to
all possible QH species, including substrate AP, product PE,
and solvent iPrOH, in the equilibrium D + QH ⇄ EQH with
KQH in Figure 8. Formation of the Ru enolate and alkoxide as
kobs in all of the kinetic experiments were derived by using data
obtained at the initial stages of the reaction, between 0% and
30% conversion, where the linearity is the highest (Figure 7b,
red line).20 The relationship between kobs and each parameter is
shown in Figure 9a−f.
E
QH is the most probable inhibitory process,21f,23c and the order
The plots of kobs versus [A]0 from 0.5 to 4 mM gave a
reasonably linear relationship from the origin with a correlation
coefficient of 0.964, indicating a first-order dependence on [A]0
(l = 1 in eq 2) (Figure 9a). The rate was essentially not affected
by [L]0 (0.1−4 mM) (m = 0 in eq 2) within the range of error
distribution (Figure 9b). The kobs value exponentially increases
as pH2 increases from 25 to 90 atm (Figure 9c) according to an
of the inhibitory effect is thought to be AP > PE > iPrOH
based on the pKa (H2O) values (AP enol (10.34),30 PE
(15.4),31 and iPrOH (16.5)32). The inhibitory factor is not
simple, being affected not only by acid and base principles but
also by the concentration, steric and electronic properties of
E
QH, and the emergence of other reaction pathways specific to
EQH
.
F
DOI: 10.1021/jacs.5b02350
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 10. Comparison of the inhibitory effect of acetophenone (AP), cyclohexyl methyl ketone (CMK), and 2,2-dimethylpropiophenone (DMPP)
under the standard conditions ([(R)-Ph-BINAN-H-Py ((R)-L)]0 = [Ru(π-CH2C(CH3)CH2)2(cod) (A)]0 = [tBuOK] = 2 mM, 50 atm H2, iPrOH,
25 °C). (a) Reciprocal plot of kobs versus [AP]0. (b) Reciprocal plot of kobs versus [CMK]0. (c) Reciprocal plot of kobs versus [DMPP]0. In (a), (b),
and (c), the slope corresponds to 2(KQH[QH]0 + KiPrOH[iPrOH] + 1)/k0k2[A]0[H2]3 (QH = AP, CMK, or DMPP). Because the value of the
denominator is constant under a given condition, a steeper slope means greater inhibition.
To make a qualitative interpretation of the inhibitory effect,
the following experiments were carried out by using AP,
cyclohexyl methyl ketone (CMK), 2,2-dimethylpropiophenone
(DMPP), (R)-PE, and (S)-PE: CMK is enolizable but lacks the
benzene ring of AP, whereas DMPP is not enolizable but has a
benzene ring. (R)-PE is the major product of (R)-L/A-
catalyzed asymmetric hydrogenation, while (S)-PE is the minor
enantiomeric product. In comparing inhibition among AP,
CMK, and DMPP, however, care must be taken because these
ketones are completely different both sterically and electroni-
cally and vary in reactivity and selectivity in the (R)-L/A-
because DMPP causes little inhibition in the present catalysis, it
shows less reactivity than AP (kobs = 1.67 × 10−2 vs 2.50 × 10−2
M h−2). A steric effect of DMPP would be the reason.
This series of experiments indicates that the inhibition
resulting from AP and CMK would be caused mainly by
formation of the Ru O-enolate21f,33 and partly by the phenyl
group of AP. In addition, the reaction of AP at 1 M was not
inhibited by the addition of ethylbenzene in the range of 1−2
M under the standard conditions.20 This result supports the
idea that the major inhibitory factor of AP originates from
formation of the CH-π-stabilized Ru O-enolate EAP and not
from the arene−Ru complex.
2.6.2. Enantiomeric Products: (R)-PE and (S)-PE. To
evaluate the inhibitory effect of PE quantitatively, reaction
rates were determined for the hydrogenation of AP (0.5 M)
under the standard conditions except for the presence of 0.5
and 1 M PE. By defining the externally added (R)-PE and (S)-
PE as (R)-PEex and (S)-PEex, respectively, eq 5 can be
deduced.20
catalyzed asymmetric hydrogenation: namely, 94:6 er and kobs
=
2.50 × 10−2 M h−2 for AP; 42:58 er and kobs = 8.31 × 10−2
M
h−2 for CMK; and 93:7 er and kobs = 1.67 × 10−2 M h−2 for
DMPP at a substrate concentration of 1 M. Bearing this in
mind, and with the assumption that the same mechanism is
operating as for AP, the inhibitory effect was investigated as
discussed below.
2.6.1. Ketone Substrates: AP, CMK, and DMPP. To analyze
the inhibitory effect of ketone substrates and iPrOH, eq 3 was
converted to eq 4.
1/kobs = 2(KAP[AP]0 + K(R)‐PE[(R)‐ PEex]
+ K(S)‐PE[(S)‐ PEex] + KiPrOH[iPrOH] + 1)
1/kobs = 2(KAP[AP]0 + KiPrOH[iPrOH] + 1)
/k0k2[A]0[H2]3
/k0k2[A]0[H2]3
(5)
(4)
The plots of 1/kobs versus [AP]0 gave a straight line with a
small y-intercept value (−4.98) within the range of 0.5−2 M
(Figure 10a). The direct proportion from the origin shows that
(i) AP is the major inhibitor of the formation of D; (ii) the D
⇄ EAP equilibrium point lies toward the far EAP side (1 ≪
Because 1/kobs shows a linear relation to both [(R)-PEex] and
[(S)-PEex] in eq 5, the slope in the plots of 1/kobs against
[PEex] indicates the degree of the inhibitory effect. As shown in
Figure 11a, the product (R)-PE, which is generated from (R)-
L/A catalysis, had little effect on the reactivity, giving a shallow
slope. By contrast, the enantiomeric product (S)-PE, which is
the minor product in the present R catalysis, significantly
inhibited the reaction. The K(S)‑PE/K(R)‑PE ratio approached 15
(slope ratio = 176/12). The 15-fold difference in the
equilibrium constant may be due to a CH-π interaction
between the phenyl group of (S)-PE alkoxide and PyC(6)H of
KAP); and (iii) inhibition by iPrOH is negligible (KiPrOH
≪
KAP). The inhibition profile of CMK also showed the same
tendency as that of AP (Figure 10b), albeit with a 4.5-fold
shallower slope than AP (47.5 vs 10.5). We assume that this
difference may originate from CH-π stabilization of EAP,
whereby the C6H5 π system interacts with the PyC(6)H of the
Ph-BINAN-H-Py ligand. The extra inhibitory factor exists only
for AP and not for CMK, reflecting the lower reactivity of AP
(kobs = 2.50 × 10−2 M h−2 for [AP]0 = 1 M; kobs = 8.31 × 10−2
M h−2 for [CMK]0 = 1 M). By contrast, DMPP, which has no
proton, gave a straight line with a shallow slope (20.4) and a
large y-intercept (+39.7) as shown in Figure 10c. This profile
can be understood by assuming that nonenolizable DMPP
relatively enhances the degree of the iPrOH inhibitory effect
toward the catalytic cycle via a 16e Ru amide D. Nonetheless,
(R)-L, which would stabilize E(S)‑PE.
34 The conformer with such
a CH-π interaction in the corresponding Ru−(R)-PE alkoxide
(E(R)‑PE) would be energetically unfavored for steric reasons.
The methyl group is forced into a location in the crowded
region (see Figure 11, upper structure). These results imply
that chiral amplification may occur in the present asymmetric
hydrogenation via a new mechanism.7b,35
2.7. Simulation. A series of kinetic studies revealed that the
present asymmetric hydrogenation of AP proceeds without
G
DOI: 10.1021/jacs.5b02350
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
the hydrogenation proceeds according to eq 6 with a k0k2/KAP
value of 4.1 × 103 M−2 h−2 in the early stages.
2.8. Effect of [L]0 and [tBuOK]. As shown in Figure 9b, the
initial concentration of (R)-L in the range of 0.1−4 mM has
little effect on the rate. Therefore, even a reaction with a smaller
amount of L than A ([L]0 = 0.5 mM; [A]0 = 2 mM; L/A = 1/
4) showed essentially the same reactivity and selectivity as that
with an L/A ratio of 1:1. Moreover, a further decrease in the L/
A ratio to 1/20 ([L]0 = 0.1 mM; [A]0 = 2 mM) produced PE in
an R/S er of 96:4 without a significant loss of reactivity. As
shown in Figure 13, an even further decrease in the L/A ratio
Figure 11. Quantitative analysis of the inhibitory effect of (R)-1-
phenylethanol ((R)-PE) and (S)-PE, and possible structures of the Ru
alkoxides, E(R)‑PE and E(S)‑PE. (a) Reciprocal plot of kobs versus [(R)-
PEex]. (b) Reciprocal plot of kobs versus [(S)-PEex]. Hydrogenation of
AP (0.5 M) was carried out in the absence and presence of PE (0.5
and 1 M) under the standard conditions ([(R)-Ph-BINAN-H-Py ((R)-
L)]0 = [Ru(π-CH2C(CH3)CH2)2(cod) (A)]0 = [tBuOK] = 2 mM, 50
atm H2, iPrOH, 25 °C). The red arrow indicates a steric repulsion.
Figure 13. Time−product concentration curves obtained with an
extremely low concentration of (R)-L ([L]0 = 0.02 mM and 0.05 mM,
[A]0 = [tBuOK] = 2 mM, [AP]0 = 2 M, 50 atm H2).
to 1/40 ([L]0 = 0.05 mM; [A]0 = 2 mM) led to a straight line
in terms of the t/[PE] relationship to the point of the reaction
completion, giving (R)-PE with a 94:6 er. In the L/A ratio of 1/
100 ([L]0 = 0.02 mM and [A]0 = 2 mM), the straight-line
phenomenon was observed. Although the enantioselectivity
decreased to 87:13, the degree of chiral multiplication
approached 87 000. Under these conditions, the t/[PE] curve
no longer follows the time-squared equation ([PE] = kobst2). At
such an extremely low concentration of L, all Ph-BINAN-H-Py
molecules would be used to generate B. Consistent with this
view, the rate of [L]0 = 0.02 mM was proportionally slowed to
2/5 that of [L]0 = 0.05 mM (0.0271 vs 0.0743 M h−1). The rate
of 0.26 M h−1 in the later stages under the standard conditions
(Figure 7a) determines the approximate concentration of
Rucycle to be 0.1−0.2 mM, supporting the fact that the rate is
essentially not affected by [L]0 > 0.1 mM (Figure 9b).
significant inhibition from the solvent iPrOH and the major
product (R)-PE; therefore, the kinetic profile shown in Figure 8
can be expressed by eq 6 in the early stages of the reaction
under the standard conditions.
[PE] = kobst2 = k0k2[A]0[H2]3t2/2KAP[AP]0
(6)
To confirm its validity, eq 6 was used to simulate the
relationships between time and PE concentration, which were
then compared with the t/[PE] curves observed under different
conditions of [A]0, pressure pH2, or [AP]0. The kobs values
obtained under the conditions of [A]0 = 1−4 mM, pH2 = 25−
90 atm, and [AP]0 = 0.5−2 M (Figure 9a, c, and e) gave k0k2/
K
AP = 4.1 × 103 M−2 h−2 as the average value.36 This value was
Within the range of 1−20 mM, the initial concentration of
tBuOK was also not influential on the rate and enantiose-
lectivity (Figure 14a). A concentration of tBuOK (1 mM) that
was 50% lower than that of the Ru π-allyl precursor A (2 mM)
substituted into eq 6 to simulate the t/[PE] relationship
(Figure 12a−c). Close agreement between the simulated (blue
surfaces) and experimental (red lines) curves confirmed that
Figure 12. Simulation (blue surfaces) and observation (red lines) of the t/[PE] relationship from 0% to 30% conversion in the (R)-L/A-catalyzed
hydrogenation of AP in iPrOH at 25 °C. (a) t/[PE] relationship as a function of [A]0 ([(R)-L]0 = [tBuOK] = 2 mM; [A]0 = 1, 2, and 4 mM; [AP]0
= 2 M; 50 atm H2). (b) t/[PE] relationship as a function of pH2 ([(R)-L]0 = [A]0 = [tBuOK] = 2 mM; [AP]0 = 2 M; 25, 50, 70, and 90 atm H2). (c)
t/[PE] relationship as a function of [AP]0 ([(R)-L]0 = [A]0 = [tBuOK] = 2 mM; [AP]0 = 0.5, 1, and 2 M; 50 atm H2). Simulation conditions: [PE]
= k0k2[A]0[H2]3t2/2KAP[AP]0 with k0k2/KAP = 4.1 × 103 M−2 h−2.
H
DOI: 10.1021/jacs.5b02350
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 14. (a) Plot of kobs as a function of [tBuOK] from 0.5 to 20 mM ([A]0 = [(R)-L]0 = 2 mM, [AP]0 = 2 M, 50 atm H2). (b) Time−product
concentration curves obtained without tBuOK ([tBuOK] = 0 mM; [L]0 = 2 mM; [A]0 = 1, 1.5, 2, and 4 mM; [AP]0 = 2 M; 50 atm H2). (c) Plot of
[PE] as a function of t3 of graph (b).
was enough to maintain reactivity. When [tBuOK] was
decreased to 0.5 mM, the reactivity was halved (kobs = 1.17 ×
10−2 vs 6.89 × 10−3 M h−2). Furthermore, as shown in Figure
14b, even in the absence of tBuOK ([A]0 = 1, 1.5, 2, and 4
mM), the reaction went to completion, although the rate was
extremely slow in the early stages. Replotting the t/[PE]
relation in Figure 14b as a function of t3 exhibited a reasonably
linear relationship (Figure 14c), although the t3/[PE] curve
obtained with [A]0 = 1 mM deviated from the ideal one. The
change in the order of t from second to third in the absence of
[tBuOK] implies that the ever-changing species [Rucycle], [AP],
and [PE] are involved in this cycle in a different way than when
[tBuOK] is present under the standard conditions. The slopes
of the linear lines in the t3/[PE] plotting (blue lines) have a
second-order dependence on [A]0. So, when [A]0 is increased
from 1 to 1.5 mM, the slope is steeper by a factor of (1.5/1)2 =
2.25, which is close to the observed value, (0.80 × 10−4)/(0.40
× 10−4) = 2. The second-order dependence on [A]0 can be also
observed when 1.5 mM is increased to 2 mM ((2/1.5)2 = 1.78
vs (1.52 × 10−4)/(0.80 × 10−4) = 1.9), and when 2 mM => 4
mM ((4/2)2 = 4 vs (6.84 × 10−4)/(1.52 × 10−4) = 4.5). One
possible explanation is that two Rucycle molecules are involved
in the catalytic cycle established in the absence of tBuOK. One
Figure 15. Proposed mechanism of enantioselection in the (R)-Ph-
BINAN-H-Py−RuH2-catalyzed hydrogenation of AP.
Rucycle species may act as a base instead of tBuOK. [Rucycle
]
increases with a first-order time dependence; as a result, the
involvement of another Rucycle in the cycle would alter [PE] =
k
possessing a different ring substituent X at the para position,
including CH3O, CH3, and CF3.20 All of the reactions were
completed within 18 h under the standard conditions. The
presence of an electron-donating substituent tended to increase
the enantioselectivity, which ranged from an R/S er of 90:10
(4-CF3) to an R/S er of 99:1 (4-CH3O). The Hammett plot of
ln(R/S) versus the standard σ constants for the substituent
parameter exhibited a reasonably linear relationship (Figure
16a), indicating that a single reactive mechanism is operating in
the asymmetric hydrogenation with the involvement of factor I.
By contrast, the enantioselectivity with (R)-Ph-BINAN-H-Py
ligands possessing a substituent Y at the para position of the
C(3)-,C(3′)-phenyl group virtually remained unchanged (R/S
er = 96:4−97:3) (Figure 16b). Overall, the transition state CSi
is most likely to be stabilized by factor I: namely, a CH-π
interaction between PyC(6)H and the aromatic π system of
AP. This is the origin of the enantioselectivity in the (R)-L/A-
catalyzed hydrogenation of AP giving (R)-PE.
obst2 to [PE] = kobst3. The inclusion of at least 1 mM tBuOK
might prevent the aggregation of intermediary Ru species
leading to B, or it might enhance the reactivity of A toward H2,
thereby initiating the reaction without an induction period.
2.9. Enantioface Selection. Figure 15 shows the B + AP
−> D + PE step in detail. We believe that the reaction proceeds
via the prior formation of an N−H---OC hydrogen bond
between B and AP. A clockwise rotation37 of the CO axis
forms the transition state CSi, in which the Si face of the
electrophilic CO group is facing the nucleophilic Ru
hydrogen atom. An anticlockwise rotation37 gives CRe, the
structure of which is sterically more favored than that of CSi.
Nevertheless, CSi is more favored than CRe, resulting in the
production of (R)-PE as the major product. This sense of
enantioselectivity holds for various aromatic ketones.13 An
attractive interaction is expected to stabilize the CSi transition
state: possible candidates would be (i) a CH-π interaction
between PyC(6)H and the aromatic π system of AP (factor I)
and (ii) a CH-π interaction between ortho-HPh of AP and the
C(3)-,C(3′)-Ph π system of (R)-L (factor II).
2.9.2. 12C/13C Kinetic Isotope Effect. In this B −> D step,
the hydrogen bond formation between B and AP would have
no energy barrier, and the rate would be determined by the
hydride transfer step via CSi, where the blue-colored hydrogen
atom on Ru and the red-colored hydrogen atom on sp3N are
delivered to the CO bond in a concerted manner (see Figure
2.9.1. Hammett Analysis. To determine which of these
factors is involved, we examined the enantioselectivity of AP
I
DOI: 10.1021/jacs.5b02350
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
hydrogenation but not via transfer hydrogenation from iPrOH
as follows.20,23c,d Under the standard conditions using
(CH3)2CDOH instead of iPrOH, (R)-PE and (S)-PE were
quantitatively obtained in a 94:6 ratio. After separation of the
enantiomers, the degree of protium incorporation into (R)-PE
Figure 16. Hammett plots of ln([(R)-PE]/[(S)-PE]) as a function of
standard σ constants. (a) Effect of substituent X at the para position of
AP on enantioselectivity. (b) Effect of substituent Y at the para
position of the Ph group of (R)-Ph-BINAN-H-Py on enantioselectiv-
ity.
1
and (S)-PE was determined by H NMR analysis. The relative
15). Direct attack of RuH on CO without formation of an
sp3N−H---OC hydrogen bond is the most unlikely
possibility.
H signal intensities toward the Cpara−H signal of the phenyl
group ranged from 0.990 to 1.007 (Figure 18). Within the error
To confirm the occurrence of Intramol-DACat-based hydride
transfer, the 12C/13C isotope effect was measured.20 The
catalytic hydrogenation of AP was conducted on a 50 mmol
scale under the standard conditions, and the reaction was
stopped at 93% conversion (13 h). The reaction mixture was
concentrated and subjected to silica gel column chromatog-
raphy, which gave PE with an R/S er of 97:3 and AP (421 mg)
in 6.9% yield. The recovered AP was subjected to 13C NMR
measurement, whereby the 13C signal intensities were
compared by using the 13C signal at Cpara of the phenyl
group of AP (not involved in the reaction) as an internal
standard. The relative proportion of the 13C isotopic
composition at C(1)O was increased by 8.4% (Figure
17a). No other signals showed significant changes in 13C
Figure 18. Ratio of protium incorporation in (R)-1-phenylethanol
((R)-PE) (a) and (S)-PE (b) obtained under the standard conditions
using H2/(CH3)2CDOH and H2/iPrOH, respectively (acetone-d6 with
45° pulse and 30-s relaxation delay).
of the analysis, no D was introduced at any carbons of (R)-PE
or (S)-PE, clearly indicating that a transfer hydrogenation
mechanism was not involved in the (R)-L/A-catalyzed
reduction. The most likely explanation is that the major and
minor products, (R)-PE and (S)-PE, are both produced by the
same mechanism17 with an energy difference of 2.07 kcal/mol
between CRe and CSi.
3. CONCLUSION
Figure 17. Measurement of the 12C/13C isotope effect in
acetophenone (AP) recovered in 6.9% yield after 93% conversion to
PE under the standard conditions. (a) 13C isotopic composition
determined by 13C NMR analysis. (b) 12C/13C kinetic isotope effect
calculated by using Cpara as a standard.
The present asymmetric hydrogenation of acetophenone (AP)
using a combined system of (R)-Ph-BINAN-H-Py ((R)-L) and
Ru(π-CH2C(CH3)CH2)2(cod) (A) proceeds with virtually no
change in L and A, giving (R)-PE with an R/S ratio of 97:3.13
The series of mechanistic studies described has revealed the
complete picture of the catalysis, which follows the rate law of
[PE] = k0k2[A]0[H2]3t2/2KAP[AP]0 with k0k2/KAP = 4.1 × 103
M−2 h−2. The time-squared term relates to two processes
occurring in parallel: namely, the preliminary step before the
cycle and the cycle itself.
The detailed reaction sequence in the preliminary process is
unclear. Most probably, the introductory step would initiate
with very slow and irreversible hydrogenolysis of the Ru-π-allyl
bonds of A with 2-mol amounts of H2 molecules by generation
of (CH3)2CCH2,39 producing a tiny (<1%) amount of an
ambiguous RuH2 species in first-order for [A]0 and second-
order for [H2]. Immediately after its generation, the RuH2
species is trapped without any energy barrier by the particular
tetradentate sp2N/sp3NH linear N4 ligand (R)-L. The
characteristics of (R)-L, including very strong metal capturing
ability with a high level of cis-α selectivity in the formation of
octahedral metal complexes, enable the generation of such a
delicate Λ-cis-α B intermediate, which works as a very reactive
catalyst in the presence of AP but readily decomposes in its
isotopic composition. Using the equations of Singleton and
Thomas,38 the 12C/13C isotope effect was calculated to be 1.031
(Figure 17b). This observation of the 12C/13C isotope effect at
C(1) is consistent with the Intramol-DACat mechanism. In
addition, the lack of an isotope effect at C(2)H3 of AP argues
against the involvement of hydrogenation of the AP enol.
2.9.3. N−H---OC Hydrogen Bond. The molecular
structure of Λ-cis-α-Ru(OCOCH3)2((R)-L) (2) in crystal
(Figure 4b) suggests the existence of an sp3N−H---OC
hydrogen bond. The H---O length is considered to be 2.067−
2.112 Å, as judged by the theoretical location of H on N (see
top view). In addition, the dihedral angle made by CH3COO−
Ru−N−H is 17.93°. The hydrogen bond, as well as the small
dihedral angle, would also be reflected in the B/AP complex,
effecting Intramol-DACat ability to transfer the two hydrogen
atoms to the CO double bond of AP via CSi.
2.10. Hydrogenation versus Transfer Hydrogenation.
The present asymmetric reduction was proved to proceed via
J
DOI: 10.1021/jacs.5b02350
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
absence. The first- and zeroth-order kinetics observed for [A]0
and [L]0, respectively, as well as the minimal effect of [A]0,
[L]0, and [H2] (pH2) on enantioselectivity, excludes the
possibility of prior replacement of the cod ligand of A with (R)-
L, followed by hydrogenolysis of the Ru-π-allyl bonds. In
parallel to its reluctant generation, B rapidly converts AP to PE
by means of another H2 molecule. In the B −> D step, the NH
in B quickly captures AP to move to a charge-alternating six-
membered transition state CSi. Here, the two hydrogen atoms
of the polarized Hδ+−Nδ−---Ruδ+−Hδ− moiety are delivered in a
concerted manner to the similarly polarized Cδ+Oδ− of AP
via the prior formation of an N−H---OC hydrogen bond, for
which the energy barrier is lower than that for the subsequent
hydride transfer. The Si face of AP is selected because of a CH-
π attractive interaction between the PyC(6)H of L and the
aromatic π system of AP. The electronic factor surpasses the
steric advantage in CRe. The Intramol-DACat mechanism
operates in CSi to give the coordinatively unsaturated Ru amide
D. This B −> D step proceeds with a zeroth-order dependence
on [AP] in the reaction system, but the turnover rate is slowed
mainly by the formation of a Ru enolate of AP, which is also
stabilized by a CH-π attractive interaction. Inhibition from
iPrOH and (R)-PE is negligible, but the degree of inhibition by
the enantiomeric product (S)-PE is 15 times higher than that
by (R)-PE.
Carbon Utilization (ACT-C) from Japan Science and
Technology Agency (JST). We thank Masahiro Tsuzuki for
preliminary kinetic study and Prof. Donna G. Blackmond and
Prof. John F. Hartwig for valuable discussions for the kinetics.
REFERENCES
■
(1) (a) Noyori, R.; Ohkuma, T.; Kitamura, M.; Takaya, H.; Sayo, N.;
Kumobayashi, H.; Akutagawa, S. J. Am. Chem. Soc. 1987, 109, 5856−
5858. (b) Kitamura, M.; Ohkuma, T.; Inoue, S.; Sayo, N.;
Kumobayashi, H.; Akutagawa, S.; Ohta, T.; Takaya, H.; Noyori, R. J.
Am. Chem. Soc. 1988, 110, 629−631. See also: (c) Kitamura, M.;
Nakatsuka, H. Chem. Commun. 2011, 47, 842−846. (d) Noyori, R.;
Kitamura, M.; Ohkuma, T. Proc. Natl. Acad. Sci. U.S.A. 2004, 101,
5356−5362.
(2) (a) Calvin, M. Trans. Faraday Soc. 1938, 34, 1181−1191.
(b) Calvin, M. J. Am. Chem. Soc. 1939, 61, 2230−2234.
(3) (a) James, B. R. Homogeneous Hydrogenation; Wiley: New York,
1974; pp 1−525. (b) Birch, A. J.; Williamson, D. H. Org. React. 1976,
24, 1−186. (c) James, B. R. Addition of Hydrogen and Hydrogen
Cyanide to Carbon−Carbon Double and Triple Bonds. In
Comprehensive Organometallic Chemistry; Wilkinson, G., Stone, F. G.
A., Abel, E. W., Eds.; Pergamon Press: Oxford, U.K., 1982; Vol. 8, pp
285−369. For a recent review, see: (d) Tsukamoto, M.; Kitamura, M.
Reduction by Homogeneous Catalysis or Biocatalysis. In Science of
Synthesis, Houben−Weyl Methods of Molecular Transformations;
Hiemstra, H., Schaumann, E., Eds.; Thieme: Stuttgart, Germany,
2009; Vol. 48, Chapter 48.1.3.6.2, pp 341−357.
The reaction frequency of the induction step is <2.1 × 10−4
h−1, while the turnover frequency of the cycle is >5.7 × 104
h−1.40 The eight-order difference in these frequencies causes the
seemingly unusual phenomenon that PE is quantitatively
produced without any detectable Ru complex formation. The
L/A/2H2 to B process follows a first-order dependence on [A]0
and a zeroth-order dependence on [L]0; as a result, even a 1/
100 ratio of L (0.02 mM) to A (2 mM) maintains high catalytic
performance, attaining a chiral multiplication of 87 000.
There are likely to be many other examples in which a
negligible amount of reactive and unstable species is operating
in catalysis.41 The present study should stimulate our ideas for
the further development of asymmetric molecular catalysis. By
using preformed well-defined metal complexes, opportunities to
find a high performance catalyst may sometimes be missed.
(4) Schrock, R. R.; Osborn, J. A. Chem. Commun. 1970, 567−568.
(5) (a) Bonvicini, P.; Levi, A.; Modena, G.; Scorrano, G. J. Chem. Soc.,
Chem. Commun. 1972, 1188−1189. (b) Tanaka, M.; Watanabe, Y.;
Mitsudo, T.; Iwane, H.; Takegami, Y. Chem. Lett. 1973, 2, 239−240.
(c) Sih, C. J.; Heather, J. B.; Peruzzotti, G. P.; Price, P.; Sood, R.; Lee,
L.-F. H. J. Am. Chem. Soc. 1973, 95, 1676−1677. (d) Solodar, J. Chem.
Technol. 1975, 421−423. (e) Levi, A.; Modena, G.; Scorrano, G. J.
Chem. Soc., Chem. Commun. 1975, 6−7. (f) Heil, B.; Toros, S.; Vastag,
̈
̈
́
S.; Marko, L. J. Organomet. Chem. 1975, 94, C47−C48. (g) Hayashi,
T.; Mise, T.; Kumada, M. Tetrahedron Lett. 1976, 17, 4351−4354.
(h) Fiorini, M.; Marcati, F.; Giongo, G. M. J. Mol. Catal. 1977/1978,
3, 385−387. (i) Ojima, I.; Kogure, T.; Achiwa, K. J. Chem. Soc., Chem.
́ ́
Commun. 1977, 428−430. (j) Toros, S.; Heil, B.; Kollar, L.; Marko, L.
̈
̈
J. Organomet. Chem. 1980, 197, 85−86.
(6) Reviews and books: (a) Noyori, R.; Kitamura, M.
Enantioselective Catalysis with Metal Complexes. An Overview. In
Modern Synthetic Methods; Scheffold, R., Ed.; Springer Verlag: Berlin/
Heidelberg, Germany, 1989; pp 115−198. (b) Noyori, R. Angew.
Chem., Int. Ed. 2002, 41, 2008−2022. (c) Kitamura, M.; Noyori, R.
Hydrogenation and Transfer Hydrogenation. In Ruthenium in Organic
Synthesis; Murahashi, S.-I., Ed.; Wiley-VCH: Weinheim, Germany,
2004; pp 3−52. (d) Ohkuma, T.; Kitamura, M.; Noyori, R. Ligand
Design for Catalytic Asymmetric Reduction. In New Frontiers in
Asymmetric Catalysis; Mikami, K., Lautens, M., Eds.; John Wiley &
Sons: Weinheim, Germany, 2007; pp 1−32.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental procedures for NMR study, 12C/13C kinetic
isotope effect determination, and isotope labeling; experimental
and spectroscopic data for the new ligands and Ru complexes;
X-ray crystallographic data (CIF); detailed rate law deduction;
and kinetic data and graphs. The Supporting Information is
available free of charge on the ACS Publications website at
DOI: 10.1021/jacs.5b02350.
(7) For the original reaction establishing DACat, see: (a) Kitamura,
M.; Suga, S.; Kawai, K.; Noyori, R. J. Am. Chem. Soc. 1986, 108, 6071−
6072. (b) Noyori, R.; Kitamura, M. Angew. Chem., Int. Ed. Engl. 1991,
30, 49−69. Redox-mediated DACat: (c) Kitamura, M.; Miyata, K.;
Seki, T.; Vatmurge, N.; Tanaka, S. Pure Appl. Chem. 2013, 85, 1121−
1132.
(8) Kitamura, M.; Tokunaga, M.; Noyori, R. J. Org. Chem. 1992, 57,
4053−4054.
(9) (a) Ohkuma, T.; Ooka, H.; Hashiguchi, S.; Ikariya, T.; Noyori, R.
J. Am. Chem. Soc. 1995, 117, 2675−2676. (b) Noyori, R.; Ohkuma, T.
Angew. Chem., Int. Ed. 2001, 40, 40−73. For the asymmetric
hydrogenation of tert-alkyl ketones, see: (c) Ohkuma, T.; Sandoval, C.
AUTHOR INFORMATION
Corresponding Author
*kitamura@os.rcms.nagoya-u.ac.jp
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was aided by a Grant-in-Aid for Scientific Research
(No. 25E07B212 and No. 23005914) from the Ministry of
Education, Culture, Sports, Science and Technology (Japan)
and an Advanced Catalytic Transformation Program for
A.; Srinivasan, R.; Lin, Q.; Wei, Y.; Muniz, K.; Noyori, R. J. Am. Chem.
̃
Soc. 2005, 127, 8288−8289.
(10) (a) Shang, G.; Li, W.; Zhang, X. Transition Metal-Catalyzed
Homogeneous Asymmetric Hydrogenation. In Catalytic Asymmetric
K
DOI: 10.1021/jacs.5b02350
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Synthesis, 3rd ed.; Ojima, I., Ed.; John Wiley & Sons, Inc.: Hoboken,
NJ, 2010; pp 343−436. (b) Tang, W.; Zhang, X. Chem. Rev. 2003, 103,
3029−3069.
Cole-Hamilton, D. J.; Klankermayer, J.; Leitner, W. J. Am. Chem. Soc.
2014, 136, 13217−13225.
(28) (a) Bau, R.; Teller, R. G.; Kirtley, S. W.; Koetzle, T. F. Acc.
Chem. Res. 1979, 12, 176−183. (b) Venanzi, L. M. Coord. Chem. Rev.
1982, 43, 251−274.
(11) Yoon, T. P.; Jacobsen, E. N. Science 2003, 299, 1691−1693.
(12) Ohgo, Y.; Takeuchi, S.; Yoshimura, J. Bull. Chem. Soc. Jpn. 1971,
44, 583.
(29) (a) Six, C.; Gabor, B.; Gorls, H.; Mynott, R.; Philipps, P.;
̈
Leitner, W. Organometallics 1999, 18, 3316−3326. (b) Schweda, L.;
(13) Huang, H.; Okuno, T.; Tsuda, K.; Yoshimura, M.; Kitamura, M.
Nader, A.; Menzel, R.; Biletzki, T.; Johne, C.; Gorls, H.; Imhof, W. J.
̈
J. Am. Chem. Soc. 2006, 128, 8716−8717.
Organomet. Chem. 2010, 695, 2076−2082.
(14) H-BINAN-H-Py: Goodwin, H. A.; Lions, F. J. Am. Chem. Soc.
1960, 82, 5013−5023.
(30) Chiang, Y.; Kresge, A. J.; Santaballa, J. A.; Wirz, J. J. Am. Chem.
Soc. 1988, 110, 5506−5510.
(15) Diamine−Ru, −Rh, −Ir, or −Pd: (a) Pinel, C.; Gendreau-Diaz,
(31) Mehta, S. P. S.; Mehrotra, R. N. Transition Met. Chem. 1991, 16,
402−406.
́ ́
N.; Breheret, A.; Lemaire, M. J. Mol. Catal. A: Chem. 1996, 112, L157−
L161. (b) Ito, M.; Hirakawa, M.; Murata, K.; Ikariya, T. Organo-
metallics 2001, 20, 379−381. (c) Hedberg, C.; Kallstrom, K.;
(32) Wade, L. G. Organic Chemistry; Prentice Hall: Upper Saddle
River, NJ, 1999.
̈
̈
Arvidsson, P. I.; Brandt, P.; Andersson, P. G. J. Am. Chem. Soc.
2005, 127, 15083−15090. (d) Utsumi, N.; Tsutsumi, K.; Watanabe,
M.; Murata, K.; Arai, N.; Kurono, N.; Ohkuma, T. Heterocycles 2010,
(33) For Ru O-enolates and Rh O-enolates, see: ref 21b. (a) Hartwig,
J. F.; Bergman, R. G.; Andersen, R. A. J. Am. Chem. Soc. 1990, 112,
3234−3236. (b) Slough, G. A.; Bergman, R. G.; Heathcock, C. H. J.
Am. Chem. Soc. 1989, 111, 938−949. Regarding the possibility that
the Ru C-enolate is the inhibitory factor, see: ref 23c.
(34) There may be an NH-π stabilization. For the related report, see:
Sandoval, C. A.; Shi, Q.; Liu, S.; Noyori, R. Chem.Asian J. 2009, 4,
1221−1224.
́
80, 141−147. Shiff base−Ru: (e) Karame, I.; Jahjah, M.; Messaoudi,
A.; Tommasino, M. L.; Lemaire, M. Tetrahedron: Asymmetry 2004, 15,
1569−1581. Thiourea−Ru: (f) Tommasino, M. L.; Casalta, M.;
Breuzard, J. A. J.; Lemaire, M. Tetrahedron: Asymmetry 2000, 11,
4835−4841. For the importance of N4 ligands, see: (g) He, Y.-M.;
Fan, Q.-H. Org. Biomol. Chem. 2010, 8, 2497−2504. (h) Yoshimura,
M.; Tanaka, S.; Kitamura, M. Tetrahedron Lett. 2014, 55, 3635−3640.
(16) Powell, J.; Shaw, B. L. J. Chem. Soc. A 1968, 159−161.
(35) Soai, K.; Shibata, T.; Morioka, H.; Choji, K. Nature 1995, 378,
767−768.
(36) The concentration of hydrogen was estimated to be 0.18 M at
50 atm of pH2; see Brunner, E. Ber. Bunsenges. Phys. Chem. 1979, 83,
715−721.
(17) Ishibashi, Y.; Bessho, Y.; Yoshimura, M.; Tsukamoto, M.;
Kitamura, M. Angew. Chem., Int. Ed. 2005, 44, 7287−7290.
(18) (a) Stebler-Rothlisberger, M.; Ludi, A. Polyhedron 1986, 5,
̈
(37) Clockwise and anticlockwise mode is changed when the oxygen
lone pair cis to aryl group is selected.
1217−1221. (b) Luginbuhl, W.; Zbinden, P.; Pittet, P. A.; Armbruster,
̈
T.; Burgi, H.-B.; Merbach, A. E.; Ludi, A. Inorg. Chem. 1991, 30,
̈
(38) Singleton, A. D.; Thomas, A. A. J. Am. Chem. Soc. 1995, 117,
2350−2355.
9357−9358.
(19) Lindsay, A. J.; Wilkinson, G.; Motevalli, M.; Hursthouse, M. B. J.
Chem. Soc., Dalton Trans. 1985, 2321−2326.
(39) Detection of (CH3)2CCH2 and cod was difficult because, at
most, B was generated only 10/177 mM under the 1H NMR
monitoring conditions (Figure 6).
(20) For details, see Supporting Information.
(40) The k0 value was assumed to be 6.46 × 10−3 M−2 h−1 at most, as
estimated from the S/N ratio of 177 observed in the 1H NMR analysis
(Figure 6b). By use of the values of k0, [H2] = 0.18 M, [AP]0 = 2 M,
and k0k2/KAP = 4.1 × 103 M−2 h−2, the frequency of the generation of
B from A (v0/[A]0 = k0[H2]2) and the turnover frequency of B-
catalyzed hydrogenation (v2/[Rucycle] = k2[H2]/KAP[AP]0 = (k0k2/
KAP)(1/k0)([H2]/[AP]0)) are calculated to be <2.1 × 10−4 and >5.7 ×
104 h−1, respectively. Here, v0 is defined as −d[A]/dt = k0[A]0[H2]2,
and v2 is defined as d[PE]/dt = k2[Rucycle][H2]/KAP[AP]0.
(41) For some examples, see: (a) Brunel, J. M.; Heumann, A.; Buono,
G. Angew. Chem., Int. Ed. 2000, 39, 1946−1949. (b) Makio, H.;
Kashiwa, N.; Fujita, T. Adv. Synth. Catal. 2002, 344, 477−493.
(21) (a) Abdur-Rashid, K.; Faatz, M.; Lough, A. J.; Morris, R. H. J.
Am. Chem. Soc. 2001, 123, 7473−7474. (b) Abdur-Rashid, K.;
Clapham, S. E.; Hadzovic, A.; Harvey, J. N.; Lough, A. J.; Morris, R.
H. J. Am. Chem. Soc. 2002, 124, 15104−15118. (c) Clapham, S. E.;
Hadzovic, A.; Morris, R. H. Coord. Chem. Rev. 2004, 248, 2201−2237.
(d) Abbel, R.; Abdur-Rashid, K.; Faatz, M.; Hadzovic, A.; Lough, A. J.;
Morris, R. H. J. Am. Chem. Soc. 2005, 127, 1870−1882. (e) Hadzovic,
A.; Song, D.; MacLaughlin, C. M.; Morris, R. H. Organometallics 2007,
26, 5987−5999. (f) O, W. W. N.; Lough, A. J.; Morris, R. H.
Organometallics 2011, 30, 1236−1252.
(22) (a) Daley, C. J. A.; Bergens, S. H. J. Am. Chem. Soc. 2002, 124,
3680−3691. (b) Hamilton, R. J.; Leong, C. G.; Bigam, G.; Miskolzie,
M.; Bergens, S. H. J. Am. Chem. Soc. 2005, 127, 4152−4153.
(c) Hamilton, R. J.; Bergens, S. H. J. Am. Chem. Soc. 2006, 128,
13700−13701. (d) Hamilton, R. J.; Bergens, S. H. J. Am. Chem. Soc.
2008, 130, 11979−11987. (e) Takebayashi, S.; Bergens, S. H.
Organometallics 2009, 28, 2349−2351.
(23) (a) Yamakawa, M.; Yamada, I.; Noyori, R. Angew. Chem., Int. Ed.
2001, 40, 2818−2821. (b) Sandoval, C. A.; Ohkuma, T.; Muniz, K.;
̃
Noyori, R. J. Am. Chem. Soc. 2003, 125, 13490−13503. (c) Sandoval,
C. A.; Ohkuma, T.; Utsumi, N.; Tsutsumi, K.; Murata, K.; Noyori, R.
Chem.Asian J. 2006, 1, 102−110. (d) Ohkuma, T.; Utsumi, N.;
Tsutsumi, K.; Murata, K.; Sandoval, C. A.; Noyori, R. J. Am. Chem. Soc.
2006, 128, 8724−8725.
(24) (a) Zhang, J.; Leitus, G.; Ben-David, Y.; Milstein, D. J. Am.
Chem. Soc. 2005, 127, 10840−10841. (b) Zhang, J.; Leitus, G.; Ben-
David, Y.; Milstein, D. Angew. Chem., Int. Ed. 2006, 45, 1113−1115.
(25) Prechtl, M. H. G.; Ben-David, Y.; Giunta, D.; Busch, S.;
Taniguchi, Y.; Wisniewski, W.; Gorls, H.; Mynott, R. J.; Theyssen, N.;
̈
Milstein, D.; Leitner, W. Chem.Eur. J. 2007, 13, 1539−1546.
(26) Goliaszewski, A.; Schwartz, J. Tetrahedron 1985, 41, 5779−5789.
(27) (a) vom Stein, T.; Weigand, T.; Merkens, C.; Klankermayer, J.;
Leitner, W. ChemCatChem 2013, 5, 439−441. (b) vom Stein, T.;
Meuresch, M.; Limper, D.; Schmitz, M.; Holscher, M.; Coetzee, J.;
̈
L
DOI: 10.1021/jacs.5b02350
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX