Iridium-catalyzed hydrogenation of N-heterocyclic compounds under mild conditions by an outer-sphere pathway

Article abstract of DOI:10.1021/ja2014983

A new homogeneous iridium catalyst gives hydrogenation of quinolines under unprecedentedly mild conditions-as low as 1 atm of H₂ and 25 °C. We report air-and moisture-stable iridium(I) NHC catalyst precursors that are active for reduction of a wide variety of quinolines having functionalities at the 2-, 6-, and 8-positions. A combined experimental and theoretical study has elucidated the mechanism of this reaction. DFT studies on a model Ir complex show that a conventional inner-sphere mechanism is disfavored relative to an unusual stepwise outer-sphere mechanism involving sequential proton and hydride transfer. All intermediates in this proposed mechanism have been isolated or spectroscopically characterized, including two new iridium(III) hydrides and a notable cationic iridium(III) dihydrogen dihydride complex. DFT calculations on full systems establish the coordination geometry of these iridium hydrides, while stoichiometric and catalytic experiments with the isolated complexes provide evidence for the mechanistic proposal. The proposed mechanism explains why the catalytic reaction is slower for unhindered substrates and why small changes in the ligand set drastically alter catalyst activity.

Full text of DOI:10.1021/ja2014983

ARTICLE
pubs.acs.org/JACS
Iridium-Catalyzed Hydrogenation of N-Heterocyclic Compounds
under Mild Conditions by an Outer-Sphere Pathway
Graham E. Dobereiner,† Ainara Nova,‡,§ Nathan D. Schley,† Nilay Hazari,† Scott J. Miller,†
Odile Eisenstein,*,‡ and Robert H. Crabtree*,†
†Department of Chemistry, Yale University, 225 Prospect Street, P.O. Box 208107, New Haven, Connecticut 06520, United States
‡Institut Charles Gerhardt, Universitꢀe Montpellier 2, CNRS UMR 5253, cc 1501, Place E. Bataillon, 34095, Montpellier, France
S Supporting Information
b
ABSTRACT: A new homogeneous iridium catalyst gives hydrogenation of quinolines
under unprecedentedly mild conditions—as low as 1 atm of H2 and 25 °C. We report air-
and moisture-stable iridium(I) NHC catalyst precursors that are active for reduction of a
wide variety of quinolines having functionalities at the 2-, 6-, and 8- positions. A combined
experimental and theoretical study has elucidated the mechanism of this reaction. DFT
studies on a model Ir complex show that a conventional inner-sphere mechanism is
disfavored relative to an unusual stepwise outer-sphere mechanism involving sequential
proton and hydride transfer. All intermediates in this proposed mechanism have been isolated or spectroscopically characterized,
including two new iridium(III) hydrides and a notable cationic iridium(III) dihydrogen dihydride complex. DFT calculations on full
systems establish the coordination geometry of these iridium hydrides, while stoichiometric and catalytic experiments with the
isolated complexes provide evidence for the mechanistic proposal. The proposed mechanism explains why the catalytic reaction is
slower for unhindered substrates and why small changes in the ligand set drastically alter catalyst activity.
’ INTRODUCTION
sphere mechanisms that have been proposed for the hydrogena-
tion reactions or hydrogen transfer reactions of polar bonds,16
almost all current proposals for quinoline hydrogenation reac-
tions (Scheme 1) assume inner-sphere coordination of substrate.
For example, Fish suggested that a [(Cp*)Rh(MeCN)3]2þ
precatalyst formed an imperfectly defined Cp*Rh hydride inter-
mediate which could insert into the bound quinoline CdN
double bond and then reductively eliminate the hydrogenated
product.4 A similar pathway has been proposed for a number of
catalysts, such as [Rh(cod)(PPh3)2]PF6 by Sanchez-Delgado
et al.,5 an iridium triphos system by Rosales et al.,17 and a
rhodium tripodal polyphosphine complex by Bianchini et al.18
Rosales' CNDO2 calculations suggest that CdN bond reduction
in a model of [RuH(CO)(MeCN)2(PR3)2]þ occurs through a
redox-neutral cycle,19with 1,2-hydride insertion into the bound
substrate, followed by heterolytic H2 cleavage to liberate sub-
strate and regenerate the hydride. Zhou et al. also suggested a
redox-neutral cycle involving heterolytic H2 cleavage, but they
proposed a 1,4- rather than 1,2-hydride insertion.13 In contrast to
these mechanisms, Chan recently proposed an outer-sphere
mechanism, which involved proton transfer from a Ru dihydro-
gen complex to the substrate followed by hydride transfer to the
protonated substrate.8 This type of sequential mechanism is
unusual and is rarely invoked for the hydrogenation of organic
substrates.20
The hydrogenation of nitrogen heterocycles, including qui-
nolines, is of current interest.1,2 The tetrahydroquinoline core,
formed from quinoline hydrogenation, is present in many natural
products and pharmaceuticals, and direct hydrogenation is an
efficient route to tetrahydroquinolines and other saturated
N-heterocycles.1 Furthermore, we and others have recently
proposed nitrogen-containing heterocyclic liquids as hydrogen
storage materials that may be safer than metal hydrides, but more
efficient hydrogenation catalysts are needed to achieve this goal.2
Many homogeneous catalysts can reduce isolated CdC,
CdO, and CdN double bonds in olefins, carbonyls, and imines,
but hydrogenation of nitrogen-containing aromatics is more
challenging. Homogeneous N-heterocycle hydrogenation has
proved possible with many Rh, Ru, Ir, and Os catalyst pre-
cursors,3ꢀ6 including asymmetric hydrogenation catalysts for
quinoline conversion to tetrahydroquinolines.7ꢀ14 Unfortu-
nately, almost all quinoline hydrogenation catalysts require either
high pressures or elevated temperatures. For example, homo-
geneous iridium catalysts for the room-temperature hydrogena-
tion of quinolines typically require an excess of 15 atm of H2 for
good conversions, while catalysts that operate at lower hydrogen
pressures require elevated temperatures (∼100 °C or higher).5,6
In addition, many systems require the presence of harsh additives
such as I2.
Rational design of quinoline hydrogenation catalysts is diffi-
cult, because there have only been a few mechanistic studies and
in some cases closely related systems have even given contra-
dictory results.15 Despite the array of inner-sphere and outer-
Received: February 17, 2011
Published: April 21, 2011
r
2011 American Chemical Society
7547
dx.doi.org/10.1021/ja2014983 J. Am. Chem. Soc. 2011, 133, 7547–7562
|

Journal of the American Chemical Society
ARTICLE
Scheme 1. Prior Mechanistic Proposals for the Hydrogenation of Quinolines4,8,19
Scheme 2. Synthesis of the Precursor 1a
We now describe a new air- and moisture-stable, homoge-
neous catalyst for the hydrogenation of aza-heterocycles having
unprecedented activity at ambient temperature and hydrogen
pressures as low as 1 atm of H2. Our mild conditions have
facilitated a detailed mechanistic study. Both experimental and
computational results suggest that the hydrogenation does not
involve direct coordination of the substrate to Ir; instead, we
propose that the reaction proceeds via an unusual stepwise outer-
sphere pathway that includes a key dihydrogen complex. Each of
the intermediates in our stepwise mechanism has been isolated or
spectroscopically characterized.
Chart 1. Iridium Precursor Complexes Screened for Cataly-
tic Activity
synthesis is shown in Scheme 2). As found by Buriak,23 these
convenient precursors are air stable both in the solid state and in
solution and thus require no special handling, and the purification
of the cationic complexes can be performed in air without any loss
in yield.
Six different ionic complexes were synthesized (Chart 1), of
which only 1d,f were previously known.23,27 Benzimidazole-
and imidazole-based NHCs with various wingtip substitutions
were included. The analogue of 1a with the noncoordinat-
ing anion BArF4 (1b; BArF4 = tetrakis[3,5-bis(trifluoromethyl)-
phenyl]borate) was also prepared, as this anion has been
reported to improve the efficiency of olefin hydrogenation
catalysts versus PF6.28 All the new Ir complexes were fully
characterized, and the structure of 1a was confirmed by X-ray
crystallography (Figure 1). As expected, the geometry around Ir
in 1a is square planar and bond lengths and angles are consistent
with those reported for 1f.27
’ RESULTS AND DISCUSSION
Synthesis of Iridium(I) Precursor Complexes. No reports of
N-heterocyclic carbenes (NHCs) as ancillary ligands in quinoline
hydrogenation catalysts have appeared,21 although NHCs do
feature in several olefin hydrogenation catalysts.22,23 Iridium-
based catalysts are some of the most active for hydrogenation,
especially those based on the (cod)IrI fragment (cod = 1,5-
cyclooctadiene).24 We therefore began by synthesizing a series
of Ir(I) NHC phosphine complexes, a class first reported by
Buriak.23
We first prepared neutral iridium NHC species either through
base-assisted NHC metalation25 from an imidazolium salt and
[Ir(cod)Cl]2 or by transmetalation from a silver carbene.26 The
resulting neutral [Ir(cod)(NHC)X] compounds react with neu-
tral ligands in the presence of the noncoordinating anion X to yield
ionic [Ir(cod)(NHC)L]X species such as 1a (a representative
7548
dx.doi.org/10.1021/ja2014983 |J. Am. Chem. Soc. 2011, 133, 7547–7562

Journal of the American Chemical Society
ARTICLE
Catalyst Optimization. Initial screens at high hydrogen
pressures (70 atm) in toluene at room temperature (Table 1)
used 2-methylquinoline as the substrate with catalyst 1a. Com-
plex 1a is indeed capable of generating trace amounts of
2-methyl-1,2,3,4-tetrahydroquinoline, but without appreciable
catalytic turnover. In contrast to the case for some earlier
catalysts,12,14 stoichiometric iodide does not increase yields.
Superstoichiometric quantities of KPF6 and HBF4 also had no
effect. However, the addition of triphenylphosphine greatly
improved activity. In this case, we observed good conversions
even at 1 atm of H2 at reasonable reaction times (<24 h). The
activity was solvent dependent, with high conversions observed
in toluene, in which the precatalyst is sparingly soluble, and
trifluorotoluene, in which the precatalyst is fully soluble. Reac-
tions in other solvents or solvent mixtures failed to generate any
product.
A screen of 1aꢀf (Chart 1) indicated that the ancillary NHC
ligand is crucial (Table 2). We performed hydrogenation reac-
tions in both toluene and trifluorotoluene for each iridium
precursor. In general, trifluorotoluene gave higher conversions
than toluene, presumably due to the greater solubility of the
iridium complexes in this solvent. In toluene poorer product
conversions were obtained when the PF6 salt (1a) was replaced
with the BArF salt (1b); however, both were very active in
4
trifluorotoluene. Complex 1c, which features a benzimidazole
NHC with butyl wingtips, gave low conversion in both solvents.
The imidazole-based NHC complex 1d was inactive in toluene
and somewhat active in trifluorotoluene, while the related
imidazole-based complexes 1e,f gave no conversion to the
2-methyltetrahydroquinoline product in either solvent. This
may be due to the greater steric demand of the isopropyl and
mesityl wingtips. The versatility of 1a allowed us to perform our
substrate screen in toluene, which has a lower cost and greater
availability than trifluorotoluene.
Overall, our screening reactions with NHC Ir(I) precatalysts
resulted in a system that can hydrogenate 2-methylquinoline
Table 2. Screening of Iridium Precursors for Quinoline
Hydrogenation under Optimized Conditionsa
complex
time, h
conversion,b %
1a
1b
1c
1d
1e
1f
18
18
18
18
18
18
>95 (>95)c
59 (>95)
15 (30)
0 (18)
0 (0)
Figure 1. Crystal structure of catalyst precursor 1a (hydrogen atoms
and anion omitted for clarity). Selected bond lengths (Å) and angles
(deg): Ir(1)ꢀP(1) = 2.3215(17), Ir(1)ꢀC(1) = 2.042(7), N(1)ꢀ
C(1) = 1.365(8), N(1)ꢀC(2) = 1.458(10), N(2)ꢀC(1) = 1.347(9),
N(2)ꢀC(3) = 1.447(9); P(1)ꢀIr(1)ꢀC(1) = 89.5(2), C(1)ꢀN(1)ꢀ
C(2) = 126.0(7), C(1)ꢀN(2)ꢀC(3) = 124.4(7), Ir(1)ꢀC(1)ꢀN(1) =
127.1(5), Ir(1)ꢀC(1)ꢀN(2) = 127.15.
0 (0)
a Conditions: 2-methylquinoline (1 equiv), Ir complex (1 mol %), PPh3
(1 mol %), 1 atm of H2, 35 °C, 18 h. b The percent product conversion
was obtained by NMR integration of the product and any starting
material remaining and normalizing to 100%. c Values in parentheses are
conversions in trifluorotoluene.
Table 1. Initial Optimization of the Hydrogenation of 2-Methylquinoline
additive (amt)
KI (5 equiv)
solvent
toluene
pressure of H2, atm
time, h
conversion,a %
70
70
70
70
1
24
24
24
24
18
18
18
18
<5
<5
<5
>95
>95
0
50% HBF4 (5 equiv)b
KPF6 (5 equiv)
PPh3 (1 mol %)
PPh3 (1 mol %)
PPh3 (1 mol %)
PPh3 (1 mol %)
PPh3 (1 mol %)
toluene
toluene
toluene
toluene
DCM
1
1/1 toluene/MeOH
trifluorotoluene
1
0
1
>95
a The percent product conversion was obtained by NMR integration of the product and any starting material remaining and normalizing to 100%.
b Added as a 48% solution in water.
7549
dx.doi.org/10.1021/ja2014983 |J. Am. Chem. Soc. 2011, 133, 7547–7562

Journal of the American Chemical Society
ARTICLE
Table 3. Substrate Scope for the Hydrogenation Precatalyst 1aa
a Catalytic conditions: complex 1a (1 mol %), PPh3 (1 mol %), toluene, 35 °C, 18 h. See the Experimental Section for full procedures. b The percent
product conversion was obtained by NMR integration of product and any starting material remaining and normalizing to 100%. c Values in parentheses
are isolated yields.
under the mildest conditions reported to date. However, it is
remarkable that such a specific ligand set is needed for good
hydrogenation activity and that small changes to the ancillary
NHC make such a large difference to activity.
pressures of hydrogen (5 atm) improved the yields for less
hindered substrates such as quinoline; however, in all cases our
pressures are far lower than those typical for this type of
homogeneous hydrogenation. Furthermore, the reaction is com-
pletely insensitive both to water and to air prior to the addition of
hydrogen and, therefore, no special precautions or purifications
are required. Unconverted starting material was the sole organic
impurity after hydrogenation, and no side products were observed.
The imine N-benzylideneaniline can undergo hydrogenation to
full conversion under these conditions, while acetophenone is
completely unreactive.
Isolation and Observation of Relevant Iridium Hydrides.
We were interested in isolating any potential catalytic intermedi-
ates with the hope of understanding the mechanism. The neutral
meridional iridium trihydride species 2 (Scheme 3) was isolated
from a mixture of 1a, PPh3, and H2 in the presence of the
Substrate Scope. We were interested in exploring the sub-
strate scope (Table 3). Although room temperature is sufficient
for full conversion of the optimization substrate, we found that
temperatures of 35 °C are needed for the full conversion of
several other compounds, possibly due to the increased solubility
of 1a. Several substitutions on the quinoline are permitted,
including functionalities at the 2-, 6-, and 8-positions. Homo-
geneous hydrogenation of quinolines with substitutions at the
8-position is unusual and rarely reported.9 Halo substituents are
also tolerated, and the resulting aryl halide products could
potentially be used as substrates for Pd cross-coupling reactions,
allowing complex molecules to be rapidly constructed. Higher
7550
dx.doi.org/10.1021/ja2014983 |J. Am. Chem. Soc. 2011, 133, 7547–7562

Journal of the American Chemical Society
ARTICLE
Figure 2. 1H NMR spectrum of the hydride region of 2 (500 MHz NMR, CD2Cl2). Signals and coupling constants: HA, δ ꢀ11.08 (td, JPH = 17.6 Hz,
J
HH = 4.4 Hz, 2H, IrꢀH); HB, ꢀ13.86 (tt, JPH = 19.3 Hz, JHH = 4.4 Hz, 1H, IrꢀH).
Scheme 3. Generation of Trihydride 2 from Catalytic Pre-
cursor 1a
noncoordinating base DBU (1,8-diazabicyclo[5.4.0]undec-7-
ene). Although X-ray-quality crystals of 2 could not be obtained,
the 1H and 31P NMR chemical shifts and 1Hꢀ H and 31Pꢀ H
coupling constants unambiguously confirm the proposed mer-
idional structure. Signals at δ ꢀ11.08 and ꢀ13.86 ppm (CD2Cl2)
integrating to two protons and one proton, respectively, are
assigned to two types of hydride (Figure 2). The signal at δ
ꢀ11.08 ppm is a triplet of doublets (JPH = 17.6 Hz, JHH = 4.4 Hz),
while the δ ꢀ13.86 resonance is a triplet of triplets (JPH = 19.3
Hz, JHH = 4.4 Hz). The small JPH coupling constants require that
all the hydrides be cis to the phosphines, which must also be
mutually trans. The coupling assignments were confirmed in the
1H{31P} spectrum, which shows a doublet (δ ꢀ11.08) and a
triplet (δ ꢀ13.86). Only one resonance was observed in the
31P{1H} NMR spectrum, consistent with the PPh3 ligands being
mutually trans. On the basis of these data we assign the δ ꢀ11.08
signal to the hydrides cis to the NHC and the δ ꢀ13.86 signal to
the hydride trans to the NHC. Meridional trihydrides of iridium
are rare—the fac geometry is usually preferred, and only a few
mer examples have been reported, mostly with tridentate pincer
ligands.29,30
1
1
Figure 3. Plot of ln T1 vs inverse temperature for iridium hydride
complexes 2 and 3a,b at 11.7 T (500 MHz).
Scheme 4. Generation of the Dihydrogen Complex 3a
and could only be identified by NMR spectroscopy (Scheme 4).
1
Addition of 1 equiv of the strong acid [H(Et2O)2][BArF ]31 to
A single broad hydride resonance in the H NMR of 3a at
4
complex 2 afforded the single species 3a, which was not isolable
δ ꢀ7.71 ppm and corresponding to 4H per Ir shows no
7551
dx.doi.org/10.1021/ja2014983 |J. Am. Chem. Soc. 2011, 133, 7547–7562

Journal of the American Chemical Society
ARTICLE
Scheme 5. Formation of the Iridium Dihydride 4 and the Dihydrogen Species 3b
discernible coupling, with no apparent change in line shape from
183 K to 293 K in CD2Cl2. The NMR data are consistent with
several possible structures. Leading possibilities include a flux-
ional classical iridium(V) tetrahydride and a fluxional nonclassi-
cal iridium(III) dihydrogen dihydride. Combined experimental
and computational data discussed below permit a reasonable
proposal for the solution structure.32
The minimum spinꢀlattice relaxation time (T1(min)) of
hydride NMR signals with variable temperature is a well-
established diagnostic tool for differentiating between classical
and nonclassical hydrides.33 Figure 3 shows the results of
inversion recovery experiments for the two hydride signals seen
in the 1H NMR spectrum of 2, as well as the single resonance in
the spectrum of 3a.
A nonclassical hydride signal typically exhibits a T1(min) value
no larger than 35 ms at 250 MHz.33 This corresponds to 70 ms at
500 MHz because T1(min) scales linearly with field strength. The
minimum T1 value seen for complex 3a is 30 ms at 233 K at
500 MHz (CD2Cl2). This is in contrast with the minimum value of
350 ms at 253 K for complex 2 (CD2Cl2). These values suggest that
complex 3a must contain at least one H2 ligand, while complex 2
only contains classical hydrides. The broad hydride peak in complex
3a can then be assigned to the coalescence of both classical and
nonclassical hydrides due to rapid fluxional exchange combined with
“anomalous broadening” probably due to ion pairing with the
counterion.34 Although evidence for this fluxional exchange could
not be obtained by a variable-temperature NMR experiment,
computational studies suggest that the nonclassical and pure tetra-
hydride isomers are close in energy (vide infra). In addition, in the
former isomer, the dihydrogen ligand prefers to be trans to a hydride
ligand, as shown in Scheme 4.
A different iridium hydride can be isolated from the reaction
of catalyst precursor 1a, PPh3, and hydrogen in THF, without
the need for any base (Scheme 5). During the reaction the orange-
red solution becomes colorless as a white solid precipitates. The
precipitated white solid readily converts to an orange material on
applying a vacuum, presumably owing to loss of hydrogen. This
color change is completely reversible under 1 atm of H2, and the
orange color fades to colorless within a few seconds.
X-ray crystallography indicates that this orange material is an
ionic iridium hydride, 4 (Figure 4). The PPh3 ligands are mutually
trans, as in the trihydride. The P(1)ꢀIr(1)ꢀP(2) bond angle is
170.78(7)°, comparable to that observed in other iridium(III)
species with trans phosphines. Although the hydrides could not
be located in the difference density map, we can infer their presence
from charge balance and the 1H NMR spectrum of the complex and
the data are consistent with a cis hydride geometry. We have
previously reported the X-ray structures of coordinatively saturated
trans-bis(triphenylphosphine)iridium(III) NHC complexes which
feature cis hydrides.35
Figure 4. ORTEP representation of 4, as determined by X-ray crystal-
lographic analysis. The PF6 anion has been removed for clarity. Selected
bond lengths (Å) and angles (deg): C(1)ꢀIr(1) = 2.083(7), N(1)ꢀC-
(1) = 1.353(10), N(2)ꢀC(1) = 1.361(10), N(1)ꢀC(2) = 1.448(11),
N(2)ꢀC(3) = 1.452(11), Ir(1)ꢀP(2) = 2.315(2), Ir(1)ꢀP(1) =
2.294(2); P(1)ꢀIr(1)ꢀP(2) = 170.78(7), P(1)ꢀIr(1)ꢀC(1) =
95.3(2), P(2)ꢀIr(1)ꢀC(1)
=
93.3(2), C(1)ꢀN(2)ꢀC(3)
=
123.9(7), C(1)ꢀN(1)ꢀC(2) = 125.6(7), Ir(1)ꢀC(1)ꢀN(1) =
129.8(6), Ir(1)ꢀC(1)ꢀN(2) = 125.1(6).
of T1 data were complicated by the behavior of the complex at
variable temperature (Figure 5). At room temperature a single
hydride resonance is present, integrating to two protons. At 188 K,
this resonance is absent; instead, two signals can be seen, in a 1:1
intensity ratio, at ꢀ11.34 and ꢀ18.50 ppm. Both sets of resonances
are present at intermediate temperatures. The spectra are incon-
sistent with normal coalescence, and two species are clearly
present at intermediate temperatures. Possible explanations for
this behavior are coordination of the anion or of solvent at low
temperatures. Treatment of 4 with H2 forms complex 3b, which
shows a broad single hydridic peak at ꢀ7.77 ppm, similar to the
case for the dihydrogen complex 3a. The T1(min) value (27 ms
at 233 K; Figure 3) confirms that this is an H2 complex. We
therefore suggest that 3b is the PF6 analogue of 3a. Complex 4
can also be generated by application of vacuum to a sample of
dihydrogen complex 3b.
Complex 4 reacts with CD2Cl2 over several days to form
[(NHC)(PPh3)2IrHCl]PF6. This is a possible mode of catalyst
deactivation in catalytic reactions in dichloromethane, perhaps
Efforts to confirm that 4 is a classical iridium(III) dihydride
rather than an iridium(I) dihydrogen complex by interpretation
7552
dx.doi.org/10.1021/ja2014983 |J. Am. Chem. Soc. 2011, 133, 7547–7562

Journal of the American Chemical Society
ARTICLE
Figure 5. 1H NMR spectra of the hydride region of 4 at variable temperatures (500 MHz, CD2Cl2).
(170.78(7)° to 173.17(7)°), bringing that geometry closer to
linearity.
Computational Studies of the Iridium Hydrides. To gain
further insight into the structure of the neutral complex 2 and the
cations of 3a,b and 4, we carried out a DFT study on the full
system, the only approximation being that the counteranion
is not explicitly treated. To reduce the computational cost,
QM/MM calculations were performed using the ONIOM
(B3PW91:UFF) methodology, including the Ph in the MM part.
This level of calculation gives accurate geometries, as illustrated
in the present case by comparison with the crystal structures of 4.
Single-point calculations including the solvent effects (toluene)
on the optimized structures were then carried out with the M06
functional. These energies are shown in Figure 7 (see Computa-
tional Details).
From among all the isomers possible in an octahedral complex
with four different ligands, only the four isomers depicted in
Figure 7 were considered. The selection is based on the fact that a
dihydrogen ligand prefers to be trans to a strong σ donor such as
carbene or hydride. Complexes with cis and trans phosphines
(I used as energy reference, II vs III, IV) were considered, in
order to check the preference for the trans phosphines observed
experimentally. In each case, we searched for dihydrogen trans
to either carbene or hydride. The lowest energy structures
Figure 6. ORTEP representation of [(NHC)(PPh3)2IrHCl]PF6, as de-
termined by X-ray crystallographic analysis. The PF6 anion has been
removed for clarity. Selected bond lengths (Å) and angles (deg): C(2)ꢀIr-
(1) = 2.004(8), N(1)ꢀC(1) = 1.468(12), N(2)ꢀC(3) = 1.469(12),
N(1)ꢀC(2) = 1.385(10), N(2)ꢀC(2) = 1.360(11), Ir(1)ꢀP(2) =
2.353(2), Ir(1)ꢀP(1) = 2.348(2), Ir(1)ꢀCl(1) = 2.388(2); P(1)ꢀIr-
(1)ꢀP(2) = 173.17(7), P(1)ꢀIr(1)ꢀC(2) = 91.3(2), P(2)ꢀIr(1)ꢀC(2)
= 91.5(2), C(2)ꢀN(2)ꢀC(3) = 127.5(6), C(2)ꢀN(1)ꢀC(1) =
123.9(7), Ir(1)ꢀC(2)ꢀN(1) = 121.3(6), Ir(1)ꢀC(2)ꢀN(2) = 131.8(6).
correspond to dihydrogen trans to the hydride, II (ΔEsol
=
ꢀ6.0 kcal molꢀ1) and IV (ΔEsol = ꢀ6.4 kcal molꢀ1), which is the
ligand with the largest trans influence. Of these two structures,
the isomer with trans phosphines, IV, is more stable.
The relative trans influence of hydride and NHC also accounts
explaining the poor conversions seen in that solvent (vide supra).
Figure 6 shows that the monohydride complex is square pyr-
amidal with a vacancy trans to H consistent with the high trans
influence of H. There is a small decrease in the carbeneꢀIr bond
length upon substitution of H by Cl (2.004(8) Å in the
monohydride vs 2.083(7) Å in the dihydride) due to the lower
trans influence of chloride versus hydride. The PꢀIr bond
lengths are approximately 0.05 Å longer in the monohydride.
There is a corresponding increase in the PꢀIrꢀP bond angle
for the change in H H distance of the coordinated H2 ligand
3 3 3
depending on the isomer considered. The higher trans influence of
the hydride is responsible for the shorter HꢀH bond (0.86 Å trans
to H in II, vs 0.88 Å trans to NHC in III). The long IrꢀC(NHC)
distance of 2.15 Å in III decreases the trans influence of this ligand.
This, in turn, elongates the H H distance in III, which is best
3 3 3
considered as an Ir(V) tetrahydride complex.
The energetic preference for isomer IV is consistent with the
experimental structure of the cation of 3a. In addition, the broad
7553
dx.doi.org/10.1021/ja2014983 |J. Am. Chem. Soc. 2011, 133, 7547–7562

Journal of the American Chemical Society
ARTICLE
Figure 7. Isomers considered for the cation of 3a or 3b, with selected distances in Å. Energies are in kcal molꢀ1 relative to I.
Figure 8. Equilibrium proposed on the basis of the observed NMR data of 3a,b.
Figure 9. Optimized structures of the cationic dihydrogen dihydride IV, the neutral trihydride V, and the cationic dihydride VI.
hydride peak in the NMR spectrum of 3a could be explained by a
fast exchange between IV and IV0 going through a higher energy
intermediate III (Figure 8).
Scheme 7. Attempted Catalysis Using Isolated Iridium Spe-
cies as Catalytic Precursors
The mer geometry of V (Figure 9) is consistent with the NMR
data obtained for trihydride 2 (vide supra). Hexacoordinated
Ir(III) trihydrides have an intrinsic preference for a fac geometry
to avoid hydrides to be mutually trans. However, they tend to be
mer when encouraged to do so by the accompanying ligands.
This happens with a pincer ligand30 or, as in 2, in the presence of
two bulky phosphines that prefer to be trans.36 Upon proton-
ation to give the nonclassical tetrahydride 3a, we expect the H
ligand trans to another H to be the most basic and thus most
easily protonated. The geometrical features follow the relative
basicity, since the IrꢀH bond is longer when trans to H (1.66 Å)
than to NHC (1.63 Å). The dihydrogen dihydride product IV
retains the structure of V, having H2 trans to H (Figure 9).
The energy to dissociate H2 from complex IV has been
evaluated in the full system as 11.0 kcal molꢀ1. This is consistent
with the facile H2 elimination from complexes 3a,b. Upon loss of
H2, the dihydride complex, VI in Figure 9, has a square-based-
pyramidal structure with an apical hydride trans to the empty
coordination site. The IrꢀH bond trans to the empty coordina-
tion is shorter than that trans to NHC (1.53 Å vs 1.63 Å).
Reactivity of Isolated Iridium Hydrides in Hydrogenation
Reaction. We evaluated the catalytic competence of the isolated
iridium hydride complexes to establish whether they were
potential intermediates in the catalytic cycle (Scheme 7). Trihy-
dride 2 failed to hydrogenate 2-methylquinoline under our
previously optimized conditions, suggesting the neutral complex
is not a competent precatalyst (Scheme 7a). However, catalytic
hydrogenation of 2-methylquinoline to the tetrahydro form
became possible when [H(Et2O)2][BArF ] was added to
4
7554
dx.doi.org/10.1021/ja2014983 |J. Am. Chem. Soc. 2011, 133, 7547–7562

Journal of the American Chemical Society
ARTICLE
Scheme 8
alone for the present system. Thus, even though the 18e
trihydride can only directly participate in a stepwise outer-sphere
mechanism, it could be protonated and deliver H2 to give a 16e
cationic dihydride that could give an inner-sphere reaction.
Consequently, we evaluated the preference between the outer-
and the inner-sphere mechanisms by DFT calculations. These
calculations were carried out modeling the phosphine ligand by
PH3 and the experimental carbene by C2H2(NMe)2C. The
energy profiles using this model were sufficiently distinct to
demonstrate that the inner-sphere mechanism is unlikely for this
catalyst. The energies reported are Gibbs energies with solvent
(toluene) effects, and the units are kcal molꢀ1 (see Computa-
tional Details). All Gibbs energies are given at 25 °C relative to
separated species, the dihydrogen dihydride model for the cation
of 3a and 2-methylquinoline. The activation barrier is defined as
the difference in Gibbs energy between a transition state and its
associated reactants.
complex 2 (1:1 acid:complex, 1 mol % loading) (Scheme 7b).
Most likely, prior protonation of substrate allows for hydride
delivery from 2. Unlike complex 2, dihydride 4 is competent
for catalytic quinoline hydrogenation in the absence of acid
(Scheme 7c). Since we have observed the formation of a
dihydrogen complex upon exposure of 4 to hydrogen, we surmise
that a dihydrogen complex analogous to 3a is formed in situ and
that this acidic species is then capable of stepwise proton delivery
and hydride delivery to substrate.
Our labeling for the model system is as follows: the organo-
metallic complexes are labeled by the nature of the ligand trans to
the apical hydride of the 16eꢀ dihydride complex, labeled as [0].
The trihydride is labeled as [H] and the dihydrogen dihydride
complex as [H2]. The labels AꢀD are used for the organic
species: A is the unprotonated quinoline, B the monoprotonated
species, Cxy for the dihydrogenated form with H atoms in
positions x and y, D for the trihydrogenated product, and E for
If 3a is capable of sequential proton transfer/hydride transfer,
it should also be capable of stoichiometric reduction of CdN
double bonds in the absence of hydrogen. To check this point, 3a
was generated in situ by combining 2 and [H(Et2O)2][BArF ]
4
in a 1:1 mixture, and 1 equiv of 2-methylquinoline was added.
In the absence of hydrogen, no reaction occurred (Scheme 8a),
but upon addition of hydrogen, the reduction product
appeared. In contrast to 2-methylquinoline, acridine can be
reduced in the absence of hydrogen gas by a 1:1 mixture of 2 and
the tetrahydrogenated product. Finally labels such as [H2-H]q
AB
denote transition states (TS); for instance, [H2-H]qAB is the TS
in which [H2] hydrogenates A into B and turns itself into [H].
The inner-sphere mechanism starts by the decoordination of
H2 to generate a vacant site at Ir (Figure 10). From the
dihydrogen complex, [H2], this process has an energy cost of
9.2 kcal molꢀ1. After this dissociation, the quinoline insertion to
the basal IrꢀH bond of the square pyramid has an activation
barrier of 41.2 kcal molꢀ1. In addition to the large energy barrier,
this process is endoergic by 23.3 kcal molꢀ1. This mechanism is
clearly unfavorable even for PH3; thus, there is little need to
consider the same process with the real phosphines since the
steric influence of these ligands should not make this process
more favorable.
[H(Et2O)2][BArF ] (Scheme 8b). 9,10-Dihydroacridine was
4
observed in the 1H NMR spectrum of a 1:1:1 mixture of complex
2, acridine, and [H(Et2O)2][BArF ]. Additionally, the formation
4
1
of complex 4 is indicated in the H NMR spectrum by the
presence of a hydride resonance at δ ꢀ22.19 ppm. Successful
reduction in the absence of hydrogen, coupled with the forma-
tion of complex 4, strongly suggests that 3a has a role in the
catalytic hydrogenation reaction. Acridine may react because
it only requires 2H to achieve a stable final product, while
quinolines require 4H. A 1,4-hydride/proton addition would
not be expected to occur as easily as a 1,2-addition if the reaction
were concerted. The easy reduction of acridine, presumably by a
1,4-addition, is thus consistent with a stepwise process.
Computational Studies on the Hydrogenation Mechan-
ism. Two main classes of mechanism have been proposed for the
hydrogenation of heteroaromatic substrates: an inner-sphere
mechanism (Scheme 9a) and an outer-sphere mechanism where
the proton and the hydride could be transferred in a concerted
manner or in two separate steps (Scheme 9b).16 However, it is
difficult to choose between them on the basis of experiments
A concerted outer-sphere mechanism, where a proton from
the dihydrogen ligand and the hydride are simultaneously
transferred to the NdC bond of the quinoline, could not be
located. Instead we found that the two H transfers occur via
separate transition states, as represented in Figures 11 and 12.
Figure 11 shows the energy profile for the first H2 addition to
quinoline A, while Figure 12 shows the second H2 addition to
dihydroquinoline C34. Note that in both profiles the
Scheme 9. Two Mechanistic Proposals for the Hydrogenation of Quinoline: (a) Inner-Sphere Hydrogenation Mechanism;
(b) Stepwise Outer-Sphere Mechanism
7555
dx.doi.org/10.1021/ja2014983 |J. Am. Chem. Soc. 2011, 133, 7547–7562

Journal of the American Chemical Society
ARTICLE
Figure 10. Gibbs energy profile, ΔG, in kcal molꢀ1, for the inner-sphere
hydrogenation pathway.
Figure 12. Gibbs energy profile, ΔG, in kcal molꢀ1, for the second
outer-sphere hydrogenation of 2-methylquinoline.
[H]-B, the hydride transfer is endoergic by 4.1 kcal molꢀ1
.
Further reaction of [0] with dihydrogen is thus needed to
complete an exergonic process. This could explain why 3a does
not react experimentally with 2-methylquinoline and acid in the
absence of H2.
As shown in Figure 11, the formation of [0] þ C14 is endoergic
by 2.6 kcal molꢀ1, relative to the initial reactants, [H2]þA,
although isomerization of C14 to C34 stabilizes this species by
1.5 kcal molꢀ1 (not shown in Figure 11). The hydrogenations of
C14 and C34 by [H2] have significantly different energy profiles.
In the case of C14, complex [H2] can only protonate the carbon
in the 3-position; this step has a transition state with a Gibbs
energy of 24.5 kcal molꢀ1. For C34, this protonation takes place
at the nitrogen and is energetically more favorable, with a
transition state at 16.2 kcal molꢀ1 (Figure 12). In both cases,
the protonation yields the ion pair [H]-D, with a Gibbs energy of
10.0 kcal molꢀ1. The transition state for the hydride transfer to
the 2 position has a Gibbs energy of 23.5 kcal molꢀ1, which is
very similar to that of the TS for the first hydride transfer (ΔGq =
23.9 kcal molꢀ1). This process leads to the final product, with a
Figure 11. Gibbs energy profile, ΔG, in kcal molꢀ1, for the first outer-
sphere hydrogenation of 2-methylquinoline.
favorable Gibbs energy of 7.7 kcal molꢀ1
.
These computational results suggest that a stepwise outer-
sphere pathway is the most likely mechanism of hydrogenation
(Scheme 10). A cationic iridium H2 complex first gives proton
transfer to substrate, resulting in a neutral species and a proto-
nated substrate. Hydride transfer to the substrate in a later step
results in the formation of a coordinatively unsaturated, cationic
metal center, which then coordinates hydrogen and completes
the catalytic cycle. Isomerization of dihydroquinoline can take
place with classical acid catalysis in the presence of protons
liberated by the acidic iridium dihydrogen complex, thus allowing
for the dihydro form to undergo subsequent hydrogenation to
the tetrahydroquinoline. In our experimental work we have
isolated or spectroscopically characterized all of the proposed
intermediates, which provides strong support for this mechan-
ism. Furthermore, in stoichiometric reactions we have demon-
strated that each of the intermediates can perform the reactions
we propose in our mechanism; the molecular hydrogen species
organometallic complexes are the same ([H2] f [H] f [0]);
therefore, the difference in energy in those cases comes from the
organic fragment.
The first hydrogen transfer from [H2] to A starts with proton
migration from the dihydrogen ligand to the N of the quinoline.
The transition state for this H transfer is 17.2 kcal molꢀ1 above
the isolated reactants: 8.2 kcal molꢀ1 of this energy is used to
form the ion pair [H2]-A and 9.0 kcal molꢀ1 for the proton
transfer. This reaction is slightly endoergic, with the product
[H]-B being 2.5 kcal molꢀ1 above [H2]-A. Once A is protonated,
positions 2 and 4 are activated for the subsequent hydride
transfer. Both possibilities have been considered, and the en-
ergies for the transition states [H-0]q
and [H-0]q
are
BC12
BC14
ΔGq = 26.3 and 23.9 kcal molꢀ1, respectively. Figure 12 shows
the second possibility, which has the lowest energy barrier. From
7556
dx.doi.org/10.1021/ja2014983 |J. Am. Chem. Soc. 2011, 133, 7547–7562

Journal of the American Chemical Society
ARTICLE
Scheme 10. Proposed Stepwise Outer-Sphere Mechanism for the Hydrogenation of Quinolines
Figure 13. Relative energies of complexes with coordinated quinoline and 2-methylquinoline.
can be deprotonated by base, the trihydride can deliver a hydride,
and the resulting coordinatively unsaturated species can bind
dihydrogen.
orbitals of quinoline cannot efficiently stabilize the hydride in the
insertion step, even with the assistance of the metal. In contrast,
in the two-step mechanisms, the protonation of the quinoline
decreases its aromaticity and significantly lowers the energy of its
empty π* orbitals. The hydride addition is thus facilitated. Direct
π-binding of aromatic substrates to the metal may also be
disfavored by steric effects relative to the outer-sphere route.
These points may well apply to other aromatic hydrogenation
catalysts and make the outer-sphere hydrogenation process the
path of choice.
The substrate selectivity may be partially limited by pKa
differences between proton donor (the dihydrogen complex)
and proton acceptor (substrates). Ketones are not as basic as
quinolines, and this decrease in basicity may make substrate
protonation relatively unfavorable. An increase in the acidity of
the H2 complex may allow for the hydrogenation of these classes
of substrates. Chan et al., who invoke a stepwise outer-sphere
mechanism for their quinoline hydrogenation reaction, report
that their system preferentially performs CdN hydrogenation
over CdO hydrogenation.8
Steric effects may also be responsible for defining the substrate
scope. Hydrogenation of nitrogen heterocycles is most efficient
with 2-methylquinoline and derivatives with other substituents
at the same position. Remarkably, catalytic efficiency falls for
quinoline itself and other less hindered substrates such as
ketones. The proposed mechanism absolutely requires H2 co-
ordination to the metal center to allow protonation of the
substrate. This coordination is not possible if the substrate is a
better ligand than H2. This explains the failure of the reaction
with a variety of sterically unhindered nitrogen- and oxygen-
containing substrates. With bulky substrates, coordination is
prevented and the binding of H2 is allowed. To quantify this
Implications of Mechanistic Studies for Catalysis. The
mechanistic results provide a rationale for several observations made
during the original screening. Additional triphenylphosphine is
presumably required to form the intermediate iridium hydride
complexes that are responsible for catalysis, while the comparatively
small N-methyl wingtip groups on the NHC allow for the coordina-
tion of two phosphines. The noncoordinating PF6 anion also allows
the dihydride to remain coordinatively unsaturated so that it may
react with free hydrogen. This may be an advantage over previous
iridium catalysts, which feature coordinating halides that may inhibit
the formation of the active catalyst or slow the rate of the
hydrogenation by competing for a coordination site.
Though proposals for a stepwise outer-sphere reaction mecha-
nism are unusual and are not often considered, it is possible that
prior quinoline hydrogenation catalysts, in particular those with
[Ir(cod)Cl]2 as precursor,9ꢀ14 could function by this type of
mechanism. The considerable steric bulk of the chiral bisphosphines
employed in many asymmetric quinoline hydrogenations, coupled
with the steric demand of the hindered substrates typically used,
could make inner-sphere substrate coordination unfavorable.
Furthermore, acid increases the rate of reaction in several of these
hydrogenations11,18 and the acid may have a role in proton transfer.
Acid presumably prevents coordination of a substrate’s nitrogen
lone pair due to protonation. It is noteworthy that several systems
contain I2 as an additive. Under the reaction conditions I2 could
plausibly react with metal hydrides formed during the hydrogena-
tion reaction to generate hydroiodic acid.
The high energy barrier for the inner-sphere hydrogenation of
quinoline may be due to its aromaticity. The high-lying empty π*
7557
dx.doi.org/10.1021/ja2014983 |J. Am. Chem. Soc. 2011, 133, 7547–7562

Journal of the American Chemical Society
ARTICLE
effect, the energy required to replace H2 with 2-methylquinoline
or quinoline has been computed (Figure 13). This shows that
substitution of H2 by the former substrate is almost thermo-
neutral (IV f VII, ΔEsol = ꢀ0.9 kcal molꢀ1), while the latter
is thermodynamically favorable (IV f VIII, ΔEsol = ꢀ9.1 kcal
molꢀ1), in agreement with our hypothesis.
The inability of hindered substrates to compete with hydrogen
gas for coordination is a result of the considerable steric demand
of the triphenylphosphine ligands. The excellent conversions
obtained in the quinoline hydrogenation reaction with the
appropriate ligand set demonstrate that the optimal ligands are
bulky enough to prevent substrate coordination yet not bulky
enough to prevent two phosphines from maintaining coordina-
tion. This specific steric requirement may explain the narrow
range of ligands, which demonstrate activity for the reaction.
spectra were recorded at room temperature on a 400 or 500 MHz Bruker or
Varian spectrometer and referenced to the residual solvent peak (δ in
ppm and J in Hz). High-pressure hydrogen gas reactions were performed
using a reactor vessel and a glass sleeve from Parr Instrument Co. (Moline,
IL). IR spectra were obtained using a Nicolet 6700 FT-IR spectrometer.
Elemental analyses were performed by Robertson Microlit Laboratories
(Madison, NJ).
Iodo[η4-1,5-cyclooctadiene](N,N-dimethylbenzimidazol-2-
ylidene)iridium.40 To a flame-dried Schlenk was added [Ir(cod)Cl]2
(778.3 mg, 1.158 mmol). The flask was evacuated and back-filled with
nitrogen. A degassed 1:1 mixture of dichloromethane (20 mL) and
tetrahydrofuran (20 mL) was added via syringe. Under a positive flow
of nitrogen gas potassium tert-butoxide (259.4 mg, 2.316 mmol) was
addedto the flask. Thereaction mixture became a red-brownsolutionafter
stirring for 1 h, at which point N,N0-dimethylbenzimidazolium iodide41
(634.7 mg, 2.316 mmol) was added to the flask under positive flow of
nitrogen gas. The reaction mixture was stirred for 4 h. The solvent was
removed by rotary evaporation. The residual brown solid was purified by
silica gel chromatography (CH2Cl2 eluent, under air) to provide a brown
solid (1.26 g, 94% yield). The isolated material was spectroscopically
identical with that reported previously.40
’ CONCLUSIONS
We report an unusually mild homogeneous hydrogenation of
N-heterocycles at low pressures of hydrogen. Our air-stable pre-
cursors and the low pressures required for reactivity allow for the
reaction to be performed without exhaustive purification or special
handling. The catalysts are selective for CdN bonds over CdO
bonds and tolerant of halide functional groups, which could make
them suitable for practical applications. Mechanistic studies were
performed using both computational and experimental techniques.
DFT calculations showed that the classical inner-sphere mechanism
in which the substrate coordinates to the metal and inserts into a
metalꢀhydride bond is not energetically accessible for these com-
plexes. Instead, the full hydrogenation of the N-heterocycle occurs
in a remarkable succession of outer-sphere hydrogenation events,
involving a proton transfer from a coordinated acidic dihydrogen
ligand, followed by hydride transfer from the resulting classical
hydride. The key species was found to be a dihydrogen dihydride
complex. Each of the proposed intermediates in our catalytic cycle
was either isolated or spectroscopically characterized, and a series of
stoichiometric reactions confirmed their role in the catalytic cycle.
We believe that preferential binding of H2 to the metal center over
substrate is a crucial factor for the outer-sphere mechanism, and this
may explain why the catalyst is highly sensitive to the nature of the
ligands in the coordination sphere. It also accounts for the easier
hydrogenation of more hindered substrates. Although we have only
unambiguously established the stepwise mechanism in this case, we
believe that observed results in other hydrogenation systems are
consistent with our proposal. Thus, our results could lead to the
development of asymmetric catalysts that perform under far milder
conditions than existing systems and direct future catalyst design.
Bromo[η4-1,5-cyclooctadiene](N,N-di-n-butylbenzimidazol-
2-ylidene)iridium. This compound was synthesized following the above
procedure using [Ir(cod)Cl]2 (160.3 mg, 0.239 mmol), potassium tert-
butoxide (53.4 mg, 0.477 mmol), and N,N0-di-n-butylbenzimidazolium
bromide42 (157.0 mg, 0.477 mmol). The reaction mixture was stirred for
24 h. The solvent was removed by rotary evaporation and purified by silica
gel chromatography (CH2Cl2 eluent, under air) to afford a yellow solid
(198 mg, 66% yield). 1H NMR (500 MHz, CD2Cl2): δ 7.25 (m, 2H, arene
CH), 7.15 (m, 2H, arene CH), 4.70ꢀ4.42 (m, 6H, ꢀNCH2Cꢀ, COD
CH), 2.92 (m, 2H, COD CH), 2.20ꢀ2.10 (m, 4H), 2.05ꢀ1.95 (m, 2H),
1.83ꢀ1.69 (m, 4H), 1.59ꢀ1.41 (m, 4H), 0.97 (t, J = 7.4, 3H). 13C{1H}
NMR (126 MHz, CD2Cl2): δ 191.93, 135.53, 122.66, 110.67, 85.73, 53.66,
48.62, 33.85, 31.92, 30.22, 20.97, 14.15. Anal. Calcd for C23H34BrIrN2: C,
45.24; H, 5.61; N, 4.59. Found: C, 45.58; H, 5.55; N, 4.51.
[η4-1,5-Cyclooctadiene][triphenylphosphine](N,N-dimethyl-
benzimidazol-2-ylidene)iridium Hexafluorophosphate (1a). In a
round-bottom flask was added iodo[η4-1,5-cyclooctadiene](N,N-dimethyl-
benzimidazol-2-ylidene)iridium (1260 mg, 2.195 mmol), PPh3 (575 mg,
2.195 mmol), and KPF6 (404 mg, 2.195 mmol). Tetrahydrofuran (5 mL)
was added to the flask under air, resulting in a red-orange solution. Additional
PPh3 and KPF6 can be added to shorten the required reaction time. After
16 h the solvent was removed in vacuo and dichloromethane was added into
the flask. The reaction mixture was washed three times with water in a
separatory funnel under air. The organic layer was dried over magnesium
sulfate, filtered, and evaporated, leaving a red residue. Fifteen milliliters of
dichloromethane was added to the residue, and red solid was precipitated by
the addition of 150 mL of diethyl ether. After it stood for 1 h at ꢀ20 °C, the
suspension was filtered. The solid was rinsed with diethyl ether and dried at
40 °C for 16 h under vacuum over P2O5 to provide the anhydrous red solid
product (1477 mg, 76% yield). Crystals suitable for X-ray diffraction were
obtained by recrystallizing the material in dichloromethane/diethyl ether. 1H
NMR (400 MHz, CDCl3): δ 7.34ꢀ7.10 (m, 19H, PPh3, NHC ꢀCHꢀ),
4.41 (m, 2H, cod CH), 3.90 (m, 2H, cod CH), 3.68 (s, 6H, NꢀCH3), 2.37
(m, 4H, cod CH2), 2.06 (m, 4H, cod CH2). 13C{1H} NMR (125 MHz,
CDCl3): δ 186.43, 135.24, 133.57 (d, J = 10.9), 131.21 (d, J = 2.0), 129.95
(d, J = 51.2), 128.86 (d, J = 10.2), 123.35, 109.85, 86.60 (d, J = 11.5), 82.48,
34.71, 31.19, 30.64 (d, J = 2.9). 31P{1H} NMR (162 MHz, CDCl3): δ 17.42
(s), ꢀ143.93 (hept, J = 714.2). Anal. Calcd for C35H37F6IrN2P2: C, 49.23;
H, 4.37; N, 3.28. Found: C, 49.16; H, 4.16; N, 3.00.
’ EXPERIMENTAL SECTION
General Considerations. The syntheses of the ligand precursors
and the metal complexes were conducted under nitrogen using dry degassed
solvents, unless otherwise indicated. Solvents were dried and degassed
prior to use. [Ir(cod)Cl]2,37 NaBArF ,38 and [H(Et2O)2][BArF ]31 were
4
4
prepared according to literature procedures. [η4-1,5-Cyclooctadiene]-
[triphenylphosphine](N,N-dimethylimidazol-2-ylidene)iridium hexafluoro-
phosphate (1d) was prepared according to the procedure by Buriak et al.23
[η4-1,5-Cyclooctadiene][triphenylphosphine](N,N-bis(2,4,6-trimethylphe-
nyl)imidazol-2-ylidene)iridium hexafluorophosphate (1f) was prepared
according to the procedure of Nilsson et al.27 2-Ethylquinoline and
2-butylquinoline were synthesized using established procedures.39 All other
reagents were commercially available from Sigma-Aldrich, Alfa-Aesar, and
Strem Chemicals and used as received unless otherwise indicated. NMR
[η4-1,5-Cyclooctadiene][triphenylphosphine](N,N-dimethyl-
benzimidazol-2-ylidene)iridium Tetrakis[3,5-bis(trifluoromethyl)-
phenyl]borate (1b). In a flame-dried Schlenk flask was added iodo[η4-
1,5-cyclooctadiene](N,N-dimethylbenzimidazol-2-ylidene)iridium (161.9 mg,
7558
dx.doi.org/10.1021/ja2014983 |J. Am. Chem. Soc. 2011, 133, 7547–7562

Journal of the American Chemical Society
ARTICLE
0.282 mmol), PPh3 (73.8 mg, 0.282 mmol), and anhydrous NaBArF4 (250.0
mg, 0.282 mmol). Dried tetrahydrofuran (5 mL) was added to the flask,
resulting in a red-orange solution. The reaction mixture was stirred for 1 h. The
solvent was removed in vacuo, and dichloromethane was added into the flask.
The reaction mixture was filtered through Celite in vacuo and the solvent was
evaporated to afford an orange solid (430 mg, 97% yield). 1H NMR (500 MHz,
CDCl3): δ 8.02ꢀ7.04 (m, 31H, aromatic), 4.40 (m, 2H, cod CH), 4.03 (m,
2H, cod CH), 3.66 (s, 6H, NꢀCH3), 2.55ꢀ2.01 (m, 8H, cod CH2). 13C{1H}
NMR (126 MHz, CDCl3): δ 186.22 (d, J = 9.6 Hz), 161.59 (q, J = 49.9 Hz),
134.98, 134.70, 133.28 (d, J= 10.9 Hz), 131.41, 129.73, 129.32, 129.22ꢀ128.34
(q, J = 31.3 Hz), 128.77 (d, J = 10.2 Hz), 127.69, 121.19, 117.36, 109.49, 85.87
(d,J= 11.7 Hz), 82.85, 33.89, 31.01, 30.49. 31P{1H} NMR (202 MHz, CDCl3):
δ17.62. Anal. Calcd for C67H49BF24IrN2P: C, 51.19; H, 3.14; N, 1.78. Found: C,
51.09; H, 2.90; N, 1.81.
to the flask. After three freezeꢀpumpꢀthaw cycles, the solution was
cooled to ꢀ78 °C. Hydrogen was introduced, and the reaction mixture
was warmed to room temperature. The reaction mixture was stirred for
22 h, during which time the suspension became a pale yellow solution.
Hydrogen flow was stopped, and the reaction flask was rapidly evacuated
and back-filled with nitrogen gas. The solution was filtered through
Celite in vacuo, and the solvent was evaporated under vacuum. The
material was recrystallized from dichloromethane and diethyl ether to
afford a bright yellow solid (167.7 mg, 37% yield after recrystallization).
1H NMR (500 MHz, CD2Cl2): δ 7.75ꢀ7.61 (m, 12H, PPh3 CH),
7.16ꢀ7.07 (m, 18H, PPh3 CH), 7.00ꢀ6.92 (m, 2H, NHC ꢀCHꢀ),
6.85ꢀ6.78 (m, 2H, NHC ꢀCHꢀ), 3.37 (s, 6H, NꢀCH3), ꢀ11.08 (td,
JPH = 17.6 Hz, JHH = 4.4 Hz, 2H, IrꢀH), ꢀ13.86 (tt, JPH = 19.3 Hz, JHH
= 4.4 Hz, 1H, IrꢀH). 13C{1H} NMR (126 MHz, CD2Cl2): δ 189.45
(m), 139.52 (t, J = 25.6 Hz), 134.87, 134.49 (t, J = 6.5 Hz), 129.14,
127.66 (t, J = 4.8 Hz), 121.54, 109.70, 37.38. 31P{1H} NMR (202 MHz,
CD2Cl2): δ 26.11. IR (CH2Cl2, cmꢀ1): 2001.9 (νIrꢀH), 1760.6 (νIrꢀH).
[η4-1,5-Cyclooctadiene][triphenylphosphine](N,N-di-n-butyl-
benzimidazol-2-ylidene)iridium Hexafluorophosphate (1c). In a
flask was added bromo[η4-1,5-cyclooctadiene](N,N-di-n-butylbenzimidazol-
2-ylidene)iridium (60.0 mg, 0.0952 mmol), PPh3 (25.0 mg, 0.0952 mmol),
and KPF6 (17.5 mg, 0.0952 mmol). Dried tetrahydrofuran was added into the
flask, and the reaction mixture was stirred for 16 h. The solvent was removed
in vacuo, and dichloromethane was added to the flask. The reaction was
filtered through Celite under air, and the solvent was removed. The resulting
orange solid was purified via recrystallization (dichloromethane/diethyl ether)
to afford an orange solid (77 mg, 86% yield). 1H NMR (500 MHz, CD2Cl2)
δ 7.49ꢀ7.38 (m, 3H, PPh3), 7.35ꢀ7.19 (m, 16H, PPh3, NHC ꢀCHꢀ),
4.63ꢀ4.49 (m, 4H, cod CH, NꢀCH2), 3.95 (m, 2H, cod CH), 3.73 (m, 2H,
NꢀCH2ꢀ), 2.43ꢀ2.30 (m, 4H, cod CH2), 2.28ꢀ2.08 (m, 4H, cod CH2),
1.70ꢀ1.58 (m, 4H, ꢀCH2ꢀ), 1.54ꢀ1.43 (m, 4H, ꢀCH2ꢀ), 0.99 (t, J = 7.4
Hz, 6H, ꢀCH3). 13C{1H} NMR (126 MHz, CD2Cl2): δ 185.52 (d, J = 9.4
Hz), 135.51, 134.16 (d, J = 10.7 Hz), 131.99 (d, J = 2.2 Hz), 130.26 (d, J =
51.2 Hz), 129.43 (d, J = 10.1 Hz), 123.97, 111.46, 86.90 (d, J = 11.8 Hz),
82.76, 49.78, 31.74 (d, J = 1.9 Hz), 31.39, 31.06 (d, J = 3.1 Hz), 21.12, 14.01.
31P{1H} NMR (202 MHz, CD2Cl2): δ 17.57, ꢀ144.56 (hept, J = 709.0).
Anal. Calcd for C41H49F6IrN2P2: C, 52.50; H, 5.27; N, 2.99. Found: C, 52.53;
H, 5.22; N, 2.98.
Anal. Calcd for C45H43IrN2P2 H2O: C, 61.14; H, 5.13; N, 3.17. Found:
3
C, 60.89; H, 4.75; N, 3.21.
Bis[triphenylphosphine](N,N-dimethylbenzimidazol-2-ylidene)-
dihydridoiridium Hexafluorophosphate(4). In a flame-dried Schlenk
flask was added [η4-1,5-cyclooctadiene][triphenylphosphine](N,N-di-
methylbenzimidazol-2-ylidene)iridium hexafluorophosphate (1a; 100.0
mg, 0.117 mmol) and PPh3 (30.5 mg, 0.117 mmol). The flask was
evacuated and back-filled with nitrogen three times. Dried degassed
tetrahydrofuran (3 mL) was added to the flask. The solution was cooled
to 0 °C and stirred for 5 min. Hydrogen gas was actively bubbled through
the solution. The reaction mixture was stirred for 2 h, during which time
a white solid precipitated from the solution. Hydrogen flow was stopped,
and the reaction mixture was rapidly evacuated and back-filled with
nitrogen gas. The solution was decanted, and the remaining solid was
rinsed with THF under a nitrogen atmosphere at ꢀ78 °C. The washed
solid was subjected to vacuum, at which point the solid changed color
from white to orange. The crude material was recrystallized from
dichloromethane and diethyl ether at 0 °C to afford the orange product
(50.2 mg, 42% yield after recrystallization). Crystals suitable for X-ray
diffraction studies were grown from a solution of dichloromethane
layered with diethyl ether at 0 °C under a nitrogen atmosphere. 1H NMR
(500 MHz, DCE-d4): δ 7.49ꢀ7.20 (m, 32H, PPh3 CH, NHCꢀCHꢀ),
7.06ꢀ7.00 (m, 2H, NHC ꢀCHꢀ), 3.02 (s, 6H, NꢀCH3), ꢀ20.67 (s,
2H, IrꢀH). 13C{1H} NMR (126 MHz, DCE-d4): δ 134.87, 134.11,
133.54, 132.96 (t, J = 26.6 Hz), 131.84, 131.15, 129.17, 129.03, 123.40,
109.99, 34.38. Attempts to locate the C2 NHC resonance by modifying
the spectrometer decoupling strength and frequency were unsuccessful.
31P{1H} NMR (202 MHz, DCE-d4): δ 27.38, ꢀ144.55 (hept, J = 710.3
Hz). IR (thin film, cmꢀ1): 1981.5 cmꢀ1 (νIrꢀH). Anal. Calcd for
C45H42F6IrN2P3: C, 53.52; H, 4.19; N, 2.77. Found: C, 53.29; H,
3.98; N, 2.86.
[η4-1,5-Cyclooctadiene][triphenylphosphine](N,N-diisopro-
pylimidazol-2-ylidene)iridium Hexafluorophosphate (1e). In a
flask was added chloro[η4-1,5-cyclooctadiene](N,N-diisopropylimidazol-2-
ylidene)iridium43 (43.1 mg, 0.0848 mmol), PPh3 (22.2 mg, 0.0848 mmol),
and KPF6 (15.6 mg, 0.0848 mmol). Dried tetrahydrofuran was added into
the flask, and the reaction mixture was stirred for 16 h. The solvent was
removed in vacuo, and dichloromethane was added into the flask. The
reaction mixture was filtered through Celite under air, and the solvent was
removed. The resulting orange solid was purified via recrystallization
(dichloromethane/diethyl ether) to afford an orange solid (68 mg, 93%
yield). 1H NMR (500 MHz, CD2Cl2): δ 7.52ꢀ7.46 (m, 3H, PPh3), 7.41
(td, J=7.6, 2.2Hz, 6H, PPh3), 7.21ꢀ7.13 (m, 6H, PPh3), 7.11(s, 2H, NHC
CH2), 4.99 (hept, J = 6.7 Hz, 2H, ꢀCHR2), 4.56ꢀ4.36 (m, 2H, cod CH),
3.77ꢀ3.55 (m, 2H, cod CH), 2.42ꢀ2.23 (m, 4H, cod CH2), 2.22ꢀ2.06
(m, 2H, cod CH2), 1.99 (m, 2H, cod CH2), 1.47 (d, J = 6.7 Hz, 6H,
ꢀCH3), 0.60 (d, J = 6.7 Hz, 3H, ꢀCH3). 13C{1H} NMR (126 MHz,
CD2Cl2): δ 170.81 (d, J = 10.0), 134.37 (d, J = 10.9), 131.88, 130.87 (d, J =
50.2), 129.58 (d, J = 10.0), 119.65, 85.99 (d, J = 12.0), 80.25, 31.60 (d, J =
1.8), 30.91 (d, J= 3.0), 25.06, 21.77. 31P{1H} NMR (202 MHz, CD2Cl2):δ
18.22, ꢀ144.52 (hept, J = 710.4 Hz). Anal. Calcd for C35H43F6IrN2P2: C,
48.89; H, 5.04; N, 3.26. Found: C, 49.02; H, 4.77; N, 3.16.
Spectroscopic Observation of 3a. Complex 4 (5.0 mg, 0.0050
mmol) was placed in a J. Young NMR tube in a glovebox. The tube was
subjected to vacuum and back-filled with hydrogen gas. The orange solid
immediately became white. Degassed DCM-d2 (0.7 mL) was added to
the NMR tube under an inert atmosphere, and the sample was
immediately used in spectroscopic studies. 1H NMR (500 MHz,
CD2Cl2): δ 7.57ꢀ7.28 (m, 18H), 7.28ꢀ7.21 (m, 14H), 7.13ꢀ6.93
(m, 2H), 3.11 (s, 6H), ꢀ7.77 (s, 4H). 13C{1H} NMR (126 MHz,
CD2Cl2): δ 177.35 (m), 134.36, 133.86, 133.58 (t, J = 6.3 Hz), 133.40,
131.56, 129.11 (t, J = 5.3 Hz), 124.15, 111.06, 36.98. 31P{1H} NMR
(202 MHz, CD2Cl2): δ 10.33, ꢀ144.55 (hept, J = 710.2 Hz).
mer,trans-Bis[triphenylphosphine](N,N-dimethylbenzimidazol-
2-ylidene)trihydridoiridium (2). In a flame-dried Schlenk flask was
added
[η4-1,5-cyclooctadiene][triphenylphosphine](N,N-dimethyl-
benzimidazol-2-ylidene)iridium hexafluorophosphate (1a; 442.8 mg,
0.5185 mmol) and PPh3 (135.9 mg, 0.5185 mg).The flask was evacuated
and back-filled with nitrogen three times. Freshly distilled and degassed
1,8-diazabicyclo[5.4.0]undec-7-ene (DBU; 77.3 mL, 0.5185 mmol) was
added via syringe to the solid. Dried degassed toluene (3 mL) was added
Spectroscopic Observation of 3b. Complex 2 (5.0 mg, 0.0058
mmol) was placed in a J. Young NMR tube in a glovebox, followed by
degassed DCM-d2 (0.7 mL). [H(Et2O)2][BArF ] (5.8 mg, 0.0057
4
mmol) was then added as a solid to the tube. The sample was
immediately used in inversion recovery experiments. The NMR spectra
7559
dx.doi.org/10.1021/ja2014983 |J. Am. Chem. Soc. 2011, 133, 7547–7562

Journal of the American Chemical Society
ARTICLE
of the cation are identical with those reported for 3a. The BArF4 form of
salt 4 was generated by gently applying a vacuum to this sample; its
identity was confirmed by 1H and 31P NMR.
[H(Et2O)2][BArF ] (9.5 mg, 0. 0094 mmol) were placed in a flame-dried
4
Schlenk tube under a nitrogen atmosphere. A degassed solution of 2-methyl-
quinoline (1.5 µL, 0.0092 mmol) in toluene (0.75 mL) was added to the tube
via syringe. The proton NMR spectrum of the tube indicated that no
2-methylquinoline had been formed. The tube was then placed under vacuum
and rapidly back-filled with hydrogen gas three times. The reaction mixture was
allowed to stand for 16 h at room temperature. Formation of the 2-methylte-
trahydroquinoline product was confirmed by 1H NMR.
Optimized Procedure for Catalytic Quinoline Hydrogenation.
(a) 1 atm Pressure. In a flame-dried Schlenk flask under air was added a
stir bar, catalyst 1a (4.4 mg, 0.0052 mmol), PPh3 (1.36 mg, 0.0052
mmol), and substrate (0.52 mmol). Dried degassed toluene (2 mL) was
added. The flask was rapidly evacuated by vacuum and back-filled with
hydrogen gas three times, and the flask was kept under ambient
hydrogen pressure during the reaction. The reaction mixture was then
heated to 35 °C and stirred for 18 h, after which time the solution was
passed through basic alumina to afford the crude product. Conversion
was assessed by 1H NMR integration of product and any starting
material remaining and normalizing to 100%. The tetrahydroquinoline
products can be isolated by silica gel chromatography (hexanes/ethyl
acetate). The spectra of the known compounds isolated using this
method (2-methyl-1,2,3,4-tetrahydroquinoline,12 2-ethyl-1,2,3,4-tetra-
hydroquinoline,12 2-butyl-1,2,3,4-tetrahydroquinoline,12 2-phenyl-1,2,3,4-
tetrahydroquinoline,12 2,6-dimethyltetrahydroquinoline,12 1,2,3,4-tetrahy-
drobenzo[h]quinoline,4,44 and N-benzylaniline45) matched literature values.
(b) 5 atm Pressure. In a glass reactor sleeve under air was added a stir
bar, catalyst 1a (4.4 mg, 0.0052 mmol), PPh3 (1.36 mg, 0.0052 mmol),
and substrate (0.52 mmol). Dried degassed toluene (2 mL) was added.
The sleeve was inserted into the reactor vessel, and the reactor was
assembled under air. The reactor was purged by pressurizing the
apparatus to 5 atm and venting three times. The reactor was pressurized
to 5 atm and heated to 35 °C and the mixture stirred for 18 h, after which
the reactor was vented. The reaction solution was passed through basic
Stoichiometric Experiment with Complex 2, [H(Et2O)2][BArF ],
4
and Acridine. Complex 2(8.7 mg, 0.0010 mmol) and [H(Et2O)2][BArF ]
4
(10.1 mg, 0.0010 mmol) were placed in a flame-dried J. Young tube under a
nitrogen atmosphere. DCM-d2 (0.75 mL) was placed in the tube, and acridine
(2.0 mg, 0.011 mmol) was added as a solid. The 1H NMR spectrum of the
sample indicated that 9,10-dihydroacridine had been formed.
Decomposition of Complex 4 in Dichloromethane. A
solution of 4 in dichloromethane was stored under nitrogen for 1 week
at room temperature. The solvent was reduced under vacuum, and the
residual yellow solid was recrystallized (dichloromethane/diethyl ether)
to afford a mixture of 4 and X-ray-quality crystals of chlorobis-
[triphenylphosphine](N,N-dimethylbenzimidazol-2-ylide-
ne)hydridoiridium hexafluorophosphate. Spectra of the decomposition
product: 1H NMR (500 MHz, CD2Cl2) δ 7.51ꢀ7.14 (m, 33H), 6.84 (d,
J = 8.1 Hz, 1H), 3.46 (s, 3H), 2.07 (s, 3H), ꢀ40.29 (t, J = 13.3 Hz, 1H);
31P{1H} NMR (202 MHz, CD2Cl2) δ 23.94 (s), ꢀ144.55 (sept).
Inversion Recovery (T1) Experiments. Samples were prepared
in J. Young NMR tubes under an inert atmosphere. All spectra were
collected using a Bruker 500 MHz NMR spectrometer equipped for
variable-temperature studies. Relative integrations were collected, and
individual T1 values determined, for each observed hydride signal. For
each sample, an inversion recovery pulse sequence was performed at
various evolution periods, t, between an initial 180° pulse and a second
90° pulse. This experiment was performed at several different tempera-
tures in order to estimate T1(min). Integration of hydride signals was
performed after a Fourier transform, phase correction, and baseline
correction of the unprocessed NMR spectra. T1 values were obtained
from this data by plotting relative integration versus evolution period
and fitting the data according to
1
alumina to afford the crude product. Conversion was assessed by H
NMR integration of product and any starting material remaining and
normalizing to 100%. The tetrahydroquinoline products can be isolated
by silica gel chromatography (hexanes/ethyl acetate). The spectra of the
known compounds isolated using this method (8-chloro-2-methyl-
1,2,3,4-tetrahydroquinoline,9 6-bromo-1,2,3,4-tetrahydroquinoline,46
and 1,2,3,4-tetrahydroquinoline47) matched literature values.
Catalytic Experiment with Complex 2 without [H(Et2O)2]-
[BArF ]. Complex 2 (3.0 mg, 0.0035 mmol, 0.01 equiv) was placed into
4
a flame-dried Schlenk flask under a nitrogen atmosphere. A degassed
solution of 2-methylquinoline (47 µL, 0.35 mmol) in toluene (3 mL)
was added to the flask via syringe. The flask was placed under vacuum
and rapidly back-filled with hydrogen gas three times. The reaction
IðtÞ ¼ Ið0Þ þ Aeꢀt=T
1
where I(t) is integration at evolution period t, T1 is the spinꢀlattice
relaxation time, and A is a constant.
1
mixture was stirred for 16 h at 35 °C. The H NMR spectrum of an
Computational Details. Two different models of the Ir complex
were used in this study: one using the full system with the DFT/MM
methodology ONIOM, and the other replacing the real carbene C6H4-
(NMe)2C and PPh3 by C2H2(NMe)2C and PH3, respectively. The first
level was used for the structural determination of 2 and the cations of 3a
and 4 and the second level for the mechanistic study of the hydrogena-
tion of 2-methylquinoline. In the ONIOM calculations the Ph groups
of the phosphines were included in the MM part. All optimizations
were performed with the Gaussian09 program48 with the functional
B3PW9149 for the DFT level and the force field UFF for the molecular
mechanics level. For the DFT partition, the basis set used was the ECP-
adapted SDDALL50 with a set of polarization functions for Ir51 and P52
and the all-electron 6-31G(d,p)53 for N, C, and H. All structures were
fully optimized without any geometry or symmetry constraint.
In the mechanistic study using the model system, at full DFT level,
each stationary point was classified as a minimum or transition state by
analytical calculation of the frequencies. The connection between
reactant and product through a given transition state was checked by
optimization of slightly altered geometries of the transition state along
the two directions of the transition state vector associated with the
imaginary frequency. Entropy effects calculated in the gas phase at 298 K
and 1 atm from harmonic approximation of frequencies were included in
order to compare the inner- and outer-sphere mechanisms, since the
aliquot of the reaction mixture confirmed that no conversion to product
occurred.
Catalytic Experiment with Complex 2 and [H(Et2O)2]-
[BArF ]. Complex 2 (3.7 mg, 0.0043 mmol, 0.01 equiv) and
4
[H(Et2O)2][BArF ] (4.0 mg, 0.0043 mmol, 0.01 equiv) were placed
4
in a flame-dried Schlenk flask under a nitrogen atmosphere. A degassed
solution of 2-methylquinoline (39 µL, 0.43 mmol) in toluene (3 mL)
was added to the flask via syringe. The flask was placed under vacuum
and rapidly back-filled with hydrogen gas three times. The reaction
mixture was stirred for 16 h at 35 °C. Formation of the 2-methylte-
1
trahydroquinoline product was confirmed by H NMR of a reaction
aliquot.
Catalytic Experiment with Complex 4. Complex 4 (3.5 mg,
0.0035 mmol, 0.01 equiv) was added to a flame-dried Schlenk flask
under a nitrogen atmosphere. A degassed solution of 2-methylquinoline
(47 µL, 0.347 mmol) in toluene (2 mL) was added to the flask via
syringe. The flask was placed under vacuum and rapidly back-filled with
hydrogen gas three times. The reaction mixture was stirred for 16 h at
35 °C. Formation of the 2-methyltetrahydroquinoline product was
confirmed by 1H NMR of a reaction aliquot.
Stoichiometric Experiment with Complex 2, [H(Et2O)2][BArF ],
4
and 2-Methylquinoline. Complex 2 (8.0 mg, 0.0092 mmol) and
7560
dx.doi.org/10.1021/ja2014983 |J. Am. Chem. Soc. 2011, 133, 7547–7562

Journal of the American Chemical Society
ARTICLE
former is unimolecular, whereas the latter is bimolecular. In addition, the
effect of toluene solvent (ε = 2.379) was included by using the
continuum SMD model54 with single-point 6-311þG** calculations
and the methodology proposed by Maseras et al.55 for the Gibbs energy
in solution.
In the calculations considering the full system, which have been used
to evaluate the relative stability of the different isomers of the cation of
3a, the entropy was not considered, since all isomers have the same
molecularity. In this case, the energies given in the text are potential
energies in solution obtained by single-point 6-311þG** calculations
from the ONIOM geometries using the M06 functional.56 This meth-
odology has been used to consider the dispersion of the aromatic rings
present in the carbene and the phosphine.
Gonzꢀalez, B. Catal. Lett. 2006, 106, 101–105. (i) Lu, S. M.; Han,
X. W.; Zhou, Y. G. J. Organomet. Chem. 2007, 692, 3065–3069.
(4) Fish, R. H.; Thormodsen, A. D.; Cremer, G. A. J. Am. Chem. Soc.
1982, 104, 5234–5237.
(5) Sanchez-Delgado, R. A.; Rondon, D.; Andriollo, A.; Herrera, V.;
Martin, G.; Chaudret, B. Organometallics 1993, 12, 4291–4296.
(6) Yamaguchi, R.; Ikeda, C.; Takahashi, Y.; Fujita, K.-I. J. Am. Chem.
Soc. 2009, 131, 8410–8412.
(7) (a) Eggenstein, M.; Thomas, A.; Theuerkauf, J.; Franciꢁo, G.;
Leitner, W. Adv. Synth. Catal. 2009, 351, 725–731. (b) Li, Z. W.; Wang,
T. L.; He, Y. M.; Wang, Z. J.; Fan, Q. H.; Pan, J.; Xu, L. J. Org. Lett. 2008,
10, 5265–5268.
(8) Zhou, H.; Li, Z.; Wang, Z.; Wang, T.; Xu, L.; He, Y.; Fan,
Q.-H.; Pan, J.; Gu, L.; Chan, A. S. C. Angew. Chem., Int. Ed. 2008,
47, 8464–8467.
(9) Gou, F. R.; Wei, L.; Zhang, Z.; Liang, Y. M. Adv. Synth. Catal.
2010, 352, 2441–2444.
’ ASSOCIATED CONTENT
S
(10) (a) Lu, S. M.; Bolm, C. Adv. Synth. Catal. 2008, 350
1101–1107. (b) Zhou, Y.-G. Acc. Chem. Res. 2007, 40, 1357–1366.
(c) Yamagata, T.; Tadaoka, H.; Nagata, M.; Hirao, T.; Kataoka, Y.;
Ratovelomanana-Vidal, V.; Genet, J.-P.; Mashima, K. Organometallics
2006, 25, 2505–2513. (d) Tadaoka, H.; Cartigny, D.; Nagano, T.;
Gosavi, T.; Ayad, T.; Genet, J.-P.; Ohshima, T.; Ratovelomanana-Vidal,
V.; Mashima, K. Chem. Eur. J. 2009, 15, 9990–9994. (e) Reetz, M. T.; Li,
X. Chem. Commun. 2006, 2159–2160.
(11) (a) Mrsic, N.; Lefort, L.; Boogers, J. A. F.; Minnaard, A. J.;
Feringa, B. L.; de Vries, J. G. Adv. Synth. Catal. 2008, 350, 1081–1089.
(b) Deport, C.; Buchotte, M.; Abecassis, K.; Tadaoka, H.; Ayad, T.;
Ohshima, T.; Genet, J.-P.; Mashima, K.; Ratovelomanana-Vidal, V.
Synlett 2007, 2743–2747. (c) Nꢀu~nez-Rico, J. L.; Fernꢀandez-Pꢀerez,
H.; Benet-Buchholz, J.; Vidal-Ferran, A. Organometallics 2010, 29,
6627–6631.
Supporting Information. Text, tables, figures, and CIF
b
files giving details of the X-ray analysis of 1a, 4, iodo[η4-1,5-
cyclooctadiene](N,N-dimethylbenzimidazol-2-ylidene)iridium,
and chlorobis[triphenylphosphine](N,N-dimethylbenzimidazol-
2-ylidene)hydridoiridium hexafluorophosphate, complete inver-
sion recovery data, the full list of authors for ref 48, and Cartesian
coordinates of all optimized structures with the corresponding
energies. This material is available free of charge via the Internet
at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
robert.crabtree@yale.edu; odile.eisenstein@univ-montp2.fr
(12) Wang, W.-B.; Lu, S.-M.; Yang, P.-Y.; Han, X.-W.; Zhou, Y.-G.
J. Am. Chem. Soc. 2003, 125, 10536–10537.
(13) Wang, D.-W.; Wang, X.-B.; Wang, D.-S.; Lu, S.-M.; Zhou, Y.-G.;
Li, Y.-X. J. Org. Chem. 2009, 74, 2780–2787.
Present Addresses
§Institute of Chemical Research of Catalonia (ICIQ), av. Països
Catalans, 16, 43007 Tarragona, Catalonia, Spain.
(14) Qiu, L.; Kwong, F. Y.; Wu, J.; Lam, W. H.; Chan, S.; Yu, W. Y.;
Li, Y. M.; Guo, R.; Zhou, Z.; Chan, A. S. C. J. Am. Chem. Soc. 2006,
128, 5955–5965.
(15) Bianchini, C.; Meli, A.; Vizza, F. Eur. J. Inorg. Chem. 2001,
2001, 43–68.
’ ACKNOWLEDGMENT
G.E.D., N.D.S., and R.H.C. acknowledge funding from the
Division of Chemical Sciences, Geosciences, and Biosciences,
Office of Basic Energy Sciences of the U.S. Department of
Energy, through Grant DE-FG02-84ER13297. A.N. thanks
the Spanish MICINN for an MEC postdoctoral fellowship.
O.E. thanks the CNRS and the Ministꢁere de l0Enseignement
Supꢀerieur et de la Recherche for funding.
(16) (a) Balcells, D.; Nova, A.; Clot, E.; Gnanamgari, D.; Crabtree,
R. H.; Eisenstein, O. Organometallics 2008, 27, 2529–2535. (b) Noyori,
R.; Hashiguchi, S. Acc. Chem. Res. 1997, 30, 97–102. Comas-Vives, A.;
Ujaque, G.; Lledꢀos, A. Organometallics 2008, 27, 4854–4863. (c) Comas-
Vives, A.; Ujaque, G.; Lledꢀos, A. J. Mol. Struct. (THEOCHEM) 2009,
903, 123–132. (d) Chen, Y.; Tang, Y.; Lei, M. Dalton Trans.
2009, 2359–2364. (e) Wu, X.; Liu, J.; Di Tommaso, D.; Iggo, J. A.;
Catlow, C. R. A.; Bacsa, J.; Xiao, J. Chem. Eur. J. 2008, 14, 7699–7715. (f)
Leyssens, T.; Peeters, D.; Harvey, J. N. Organometallics 2008,
27, 1514–1523. (g) Yamakawa, M.; Ito, H.; Noyori, R. J. Am. Chem.
Soc. 2000, 122, 1466–1478. (h) Samec, J. S. M.; B€ackvall, J.-E.;
Andersson, P. G.; Brandt, P. Chem. Soc. Rev. 2006, 35, 237–248. (i)
Clapham, S. E.; Hadzovic, A.; Morris, R. H. Coord. Chem. Rev. 2004,
248, 2201–2237. (k) Casey, C. P.; Johnson, J. B.; Singer, S. W.; Cui, Q.
J. Am. Chem. Soc. 2005, 127, 3100–3109. (l) Privalov, T.; Samec, J. S. M.;
B€ackvall, J.-E. Organometallics 2007, 26, 2840–2848. (m) Nova, A.;
Balcells, D.; Schley, N. D.; Dobereiner, G. E.; Crabtree, R. H.; Eisenstein,
O. Organometallics 2010, 29, 6548–6558.
(17) Rosales, M.; Vallejo, R.; Soto, J.; Bastidas, L.; Molina, K.;
Baricelli, P. Catal. Lett. 2010, 134, 56–62.
(18) Bianchini, C.; Barbaro, P.; Macchi, M.; Meli, A.; Vizza, F. Helv.
Chim. Acta 2001, 84, 2895–2923.
(19) Rosales, M.; Boves, M.; Soscꢀun, H.; Ruette, F. J. Mol. Struct.
(THEOCHEM) 1998, 433, 319–328.
(20) Crabtree, R. H. The Organometallic Chemistry of the Transition
Metals, 5th ed.; Wiley: New York, 2009.
’ REFERENCES
(1) Katritzky, A. R.; Rachwal, S.; Rachwal, B. Tetrahedron 1996,
52, 15031–15070.
(2) (a) Moores, A.; Poyatos, M.; Luo, Y.; Crabtree, R. H. New J.
Chem. 2006, 30, 1675–1678. (b) Clot, E.; Eisenstein, O.; Crabtree, R. H.
Chem. Commun. 2007, 2231–2233. (c) Crabtree, R. H. Energy Environ.
Sci. 2008, 1, 134–138. (d) Dean, D.; Davis, B.; Jessop, P. G. New J. Chem.
2011, 35, 417–422.
(3) (a) Jardine, I.; McQuillin, F. J. J. Chem. Soc. D 1970, 626a–626a.
(b) Laine, R. M. New J. Chem. 1987, 11, 543–547. (c) Sꢀanchez-Delgado,
R. A.; Gonzꢀalez, E. Polyhedron 1989, 8, 1431–1436. (d) Baralt, E.; Smith,
S. J.; Hurwitz, J.; Horvath, I. T.; Fish, R. H. J. Am. Chem. Soc. 1992,
114, 5187–5196. (e) Rosales, M.; Alvarado, Y.; Boves, M.; Rubio, R.;
Soscꢀun, H.; Sꢀanchez-Delgado, R. A. Transition Met. Chem. (London)
1995, 20, 246–251. (f) Glorius, F.; Spielkamp, N.; Holle, S.; Goddard,
R.; Lehmann, C. W. Angew. Chem., Int. Ed. 2004, 43, 2850–2852. (g)
Legault, C. Y.; Charette, A. B. J. Am. Chem. Soc. 2005, 127, 8966–8967.
ꢀ
(h) Rosales, M.; Vallejo, R.; Soto, J.; Chacꢀon, G.; Gonzꢀalez, A.;
7561
dx.doi.org/10.1021/ja2014983 |J. Am. Chem. Soc. 2011, 133, 7547–7562

Journal of the American Chemical Society
ARTICLE
(21) While this paper was in press, an (NHC)Ru-catalyzed hydro-
genation of quinoxaline carbocyclic rings was reported. Urban, S.;
Ortega, N.; Glorius, F. Angew. Chem. Int. Ed. 2011, 50, 3803ꢀ3806.
(22) (a) Jeletic, M. S.; Jan, M. T.; Ghiviriga, I.; Abboud, K. A.; Veige,
A. S. Dalton Trans. 2009, 2764–2776. (b) Vazquez-Serrano, L. D.;
Owens, B. T.; Buriak, J. M. Inorg. Chim. Acta 2006, 359, 2786–2797.
(c) Page, M. J.; Wagler, J.; Messerle, B. A. Dalton Trans. 2009
7029–7038.
(49) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5662. (b)
Perdew, J. P.; Wang, Y. Phys. Rev. B 1992, 45, 13244–13249.
(50) (a) Andrae, D.; H€aussermann, U.; Dolg, M.; Stoll, H.; Preuss,
H. Theor. Chim. Acta 1990, 77, 123–141. (b) Bergner, A.; Dolg, M.;
K€uchle, W.; Stoll, H.; Preuss, H. Mol. Phys. 1993, 80, 1431–1441.
(51) Ehlers, A. W.; B€ohme, M.; Dapprich, S.; Gobbi, A.; H€ollwarth,
A.; Jonas, V.; K€ohler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G.
Chem. Phys. Lett. 1993, 208, 111–114.
(23) Vazquez-Serrano, L. D.; Owens, B. T.; Buriak, J. M. Chem.
Commun. 2002, 2518–2519.
(24) Crabtree, R. H. Acc. Chem. Res. 1979, 12, 331–337.
(25) K€ocher, C.; Herrmann, W. A. J. Organomet. Chem. 1997,
532, 261–265.
(52) H€ollwarth, A.; B€ohme, H.; Dapprich, S.; Ehlers, A. W.; Gobbi,
A.; Jonas, V.; K€ohler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G.
Chem. Phys. Lett. 1993, 203, 237–240.
(53) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973,
28, 213–222.
(26) Lin, I. J. B.; Vasam, C. S. Coord. Chem. Rev. 2007, 251, 642–670.
(27) Brown, J. A.; Irvine, S.; Kennedy, A. R.; Kerr, W. J.; Andersson,
S.; Nilsson, G. N. Chem. Commun. 2008, 1115–1117.
(28) Lightfoot, A.; Schnider, P.; Pfaltz, A. Angew. Chem., Int. Ed.
1998, 37, 2897–2899. Roseblade, S. J.; Pfaltz, A. C. R. Chim. 2007,
10, 178–187.
(54) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B
2009, 113, 6378–6396.
(55) (a) Balcells, D.; Ujaque, G.; Fernandez, I.; Khiar, N.; Maseras,
F. J. Org. Chem. 2006, 71, 6388–6396. (b) Braga, A. A. C.; Ujaque, G.;
Maseras, F. Organometallics 2006, 25, 3647–3658.
(56) Zhou, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215–241.
(29) Tanaka, R.; Yamashita, M.; Nozaki, K. J. Am. Chem. Soc. 2009,
131, 14168–14169.
(30) (a) Clarke, Z. E.; Maragh, P. T.; Dasgupta, T. P.; Gusev, D. G.;
Lough, A. J.; Abdur-Rashid, K. Organometallics 2006, 25, 4113–4117.
(b) Choualeb, A.; Lough, A. J.; Gusev, D. G. Organometallics 2007,
26, 5224–5229.
(31) Brookhart, M.; Grant, B.; Volpe, A. F. Organometallics 1992,
11, 3920–3922.
(32) (a) Maseras, F.; Lledꢀos, A.; Clot, E.; Eisenstein, O. Chem. Rev.
2000, 100, 601–636. (b) Hebden, T. J.; Goldberg, K. I.; Heinekey, D. M.;
Zhang, X.; Emge, T. J.; Goldman, A. S.; Krogh-Jespersen, K. Inorg. Chem.
2010, 49, 1733–1742.
(33) Crabtree, R. H. Acc. Chem. Res. 1990, 23, 95–101.
(34) Yao, W.; Faller, J. W.; Crabtree, R. H. Inorg. Chim. Acta 1997,
259, 71–76.
(35) (a) Gr€undemann, S.; Kovacevic, A.; Albrecht, M.; Faller, J. W.;
Crabtree, R. H. J. Am. Chem. Soc. 2002, 124, 10473–10481. (b)
Gr€undemann, S.; Kovacevic, A.; Albrecht, M.; Faller, J. W.; Crabtree,
R. H. Chem. Commun. 2001, 2274–2275. (c) Chianese, A. R.; Kovacevic,
A.; Zeglis, B. M.; Faller, J. W.; Crabtree, R. H. Organometallics 2004,
23, 2461–2468.
(36) Peris, E.; Lee, J. C.; Rambo, J. R.; Eisenstein, O.; Crabtree, R. H.
J. Am. Chem. Soc. 1995, 117, 3485–3491.
(37) Herde, J. L.; Lambert, J. C.; Senoff, C. V. Inorg. Synth. 1974,
15, 18–19.
(38) Yakelis, N. A.; Bergman, R. G. Organometallics 2005,
24, 3579–3581.
(39) Lachance, N.; Roy, P.; Leblanc, Y. U.S. Patent 20050277644,
2005.
(40) Zinner, S. C.; Rentzsch, C. F.; Herdtweck, E.; Herrmann, W. A.;
Kuhn, F. E. Dalton Trans. 2009, 7055–7062.
(41) Amyes, T. L.; Diver, S. T.; Richard, J. P.; Rivas, F. M.; Toth, K.
J. Am. Chem. Soc. 2004, 126, 4366–4374.
(42) Huang, W.; Guo, J.; Xiao, Y.; Zhu, M.; Zou, G.; Tang, J.
Tetrahedron 2005, 61, 9783–9790.
(43) Frey, G. D.; Rentzsch, C. F.; von Preysing, D.; Scherg, T.;
M€uhlhofer, M.; Herdtweck, E.; Herrmann, W. A. J. Organomet. Chem.
2006, 691, 5725–5738.
(44) Scheurer, H.; Zsindely, J.; Schmid, H. Helv. Chim. Acta 1973,
56, 478–489.
(45) Hollmann, D.; B€ahn, S.; Tillack, A.; Beller, M. Angew. Chem.,
Int. Ed. 2007, 46, 8291–8294.
(46) Sahin, A.; Cakmak, O.; Demirtas, I.; Okten, S.; Tutar, A.
Tetrahedron 2008, 64, 10068–10074.
(47) Ortiz-Marciales, M.; Rivera, L. D.; De Jesꢀus, M.; Espinosa, S.;
Benjamin, J. A.; Casanova, O. E.; Figueroa, I. G.; Rodríguez, S.; Correa,
W. J. Org. Chem. 2005, 70, 10132–10134.
(48) Frisch, M. J. et al. GAUSSIAN 09 Rev. A.02; Gaussian, Inc.,
Wallingford, CT, 2009.
7562
dx.doi.org/10.1021/ja2014983 |J. Am. Chem. Soc. 2011, 133, 7547–7562