Mechanistic studies on the reversible hydrogenation of carbon dioxide catalyzed by an Ir-PNP complex

Article abstract of DOI:10.1021/om2010172

The PNP-ligated iridium(III) trihydride complex 1 exhibited the highest catalytic activity for hydrogenation of carbon dioxide in aqueous KOH. The catalytic hydrogenation can be tuned to be a reversible process with the same catalyst at the expense of the activity, when triethanolamine was used as a base. Theoretical studies on the hydrogenation of carbon dioxide using DFT calculations suggested two competing reaction pathways: either the deprotonative dearomatization step or the hydrogenolysis step as the rate-determining step. The results nicely explain our experimental observations that the catalytic cycle is dependent on both the strength of the base and hydrogen pressure.

Full text of DOI:10.1021/om2010172

Article
pubs.acs.org/Organometallics
Mechanistic Studies on the Reversible Hydrogenation of Carbon
Dioxide Catalyzed by an Ir-PNP Complex
†
†,§
‡
, ‡
,†
Ryo Tanaka, Makoto Yamashita, Lung Wa Chung, Keiji Morokuma,* and Kyoko Nozaki*
†Department of Chemistry and Biotechnology, Graduate School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku,
1
13-8656, Japan
‡
Fukui Institute for Fundamental Chemistry, Kyoto University, Takano-Nishishiraki-cho, 34-4, Sakyo-ku, Kyoto 606-8103, Japan
*
S Supporting Information
ABSTRACT: The PNP-ligated iridium(III) trihydride complex 1 exhibited the
highest catalytic activity for hydrogenation of carbon dioxide in aqueous KOH. The
catalytic hydrogenation can be tuned to be a reversible process with the same
catalyst at the expense of the activity, when triethanolamine was used as a base.
Theoretical studies on the hydrogenation of carbon dioxide using DFT calculations
suggested two competing reaction pathways: either the deprotonative dearomatiza-
tion step or the hydrogenolysis step as the rate-determining step. The results nicely
explain our experimental observations that the catalytic cycle is dependent on both
the strength of the base and hydrogen pressure.
INTRODUCTION
Formic acid is widely used as a preservative, as an insecticide,
for tanning leathers, etc. The major synthetic route for formic
acid in industry is carbonylation of methanol catalyzed by a
basic catalyst such as sodium methoxide and subsequent
hydrolysis of the obtained methyl formate. Formic acid is also
obtained as a byproduct in the production of acetic acid via
oxidation of butane or naphtha. Although direct carbonylation
the reaction in water solution to be an exergonic process (eqs 1
and 2). Nevertheless, the advantage of the free energy gain for
■
1
the formation of formic acid from aqueous hydrogen and
8
carbon dioxide is only 4 kJ/mol, suggesting that the
equilibrium toward formic acid is only slightly favorable. In
2
fact, catalytic hydrogenation of carbon dioxide under neutral or
21,22
acidic conditions exhibits rather low activities.
Thus, basic
3
conditions are often employed for the hydrogenation, taking
advantage of the stability of a formate salt being higher than
of water is not a practical process, due to the lower
thermodynamic stability of formic acid compared to water
6−8
that of a free acid (eq 3).
4
and CO, an aqueous solution of sodium hydroxide is
effectively carbonylated to give sodium formate, which is
3,5
(1)
(2)
acidified to produce formic acid. A potentially alternative and
attractive process for the formation of formic acid is
hydrogenation of carbon dioxide, the most important green-
6
−10
house gas.
In the past few years, the hydrogenation of carbon dioxide
was spotlighted as a possible method for hydrogen storage, due
1
1−15
to the global energy crisis.
Several transition-metal (e.g.,
1
6−28
22,29,30
24,31−33
34
35
36,37
Ru,
Ir,
Rh,
Mo, Ni, Pd, and Fe
)
catalysts were used for the hydrogenation of carbon dioxide.
When we initiated our project in 2008, the two highest catalytic
(
3)
Dehydrogenation of formic acid or its salt, which is the
reverse reaction of the hydrogenation of CO , is also an
important process for hydrogen storage to release a hydrogen
plati-
complexes were
reported to catalyze the dehydrogenation of formic acid.
Ruthenium catalysts were the most intensely studied catalysts
and led to the highest activity (TOF = 18 000 h ) using a 5:2
mixture of formic acid and triethylamine, as reported by
activities for hydrogenation of CO were documented: Noyori
2
and Jessop used ruthenium phosphine complexes to catalyze
reaction of supercritical carbon dioxide with tertiary amines
2
3
8−47
48−50
48,51−53
−1
19
source. Ruthenium,
num,
rhodium,
iridium,
with a turnover frequency (TOF) of 95 000 h , and Himeda
applied cationic iridium complexes for the first example of the
hydrogenation of carbon dioxide in basic aqueous phase under
5
4,55
56
57,58
molybdenum, and iron
29
ambient pressure with a turnover number (TON) of 222 000.
−1
Formation of formic acid from carbon dioxide and hydrogen in
the gas phase is highly endoergic because of a large entropy loss
6
due to the decreasing number of molecules. Greater
stabilization to the polar product, formic acid, than to the
two gaseous and nonpolar substrates by hydration promotes
Received: October 31, 2011
Published: November 30, 2011
©
2011 American Chemical Society
6742
dx.doi.org/10.1021/om2010172 | Organometallics 2011, 30, 6742−6750

Organometallics
Article
4
6
Morris. For practical hydrogen storage, the reaction in both
directions needs to take place with reasonably high TOFs.
Successful reversible hydrogenation/dehydrogenation of formic
acid or its salt catalyzed by the same ruthenium or iridium
the catalytic cycle by using DFT calculations. Experimental and
theoretical studies both suggested that the two reaction
58
59,60
pathways previously proposed by us and others
can
compete with each other under the reaction conditions.
18,20,24,31,32,47
catalyst was reported (Figure 1).
RESULTS AND DISCUSSION
■1
. Hydrogenation of Carbon Dioxide Catalyzed by the
Ir(III) Complexes 1 and 4. As previously reported, hydro-
genation of carbon dioxide using the Ir(III) complex 1 afforded
potassium formate in aqueous KOH solution with a TOF of
50 000 h− (Table 1, entry 1) and a TON of 3 500 000 (entry
1
1
3
60
), both of the values being the highest ever reported. Here
we further investigated the reaction with varied reaction
conditions. The reaction could be scaled up to a 10 times
larger scale (entry 2). A significant pressure effect was also
observed as, under ambient pressure of dihydrogen and carbon
dioxide, the catalytic activity was decreased to a TON of 43 at
100 °C in 40 h (entry 4). Moreover, the reaction did not
proceed at room temperature, even under pressurized
conditions (entry 5). A higher concentration of KOH (5.0
M) leading to higher pH conditions resulted in lower activity
Figure 1. Recent examples for dehydrogenation of formic acid and its
salts.
In 2009, we communicated that the iridium trihydride
59
complex 1 containing a PNP-pincer ligand catalyzed the
hydrogenation of carbon dioxide. The highest catalytic activity
ever reported was accomplished when aqueous KOH was used
6
0
63
as a base. We also proposed a possible mechanism of
hydrogenation of carbon dioxide catalyzed by 1 by observing
the intermediates 2 and 3 and several elementary reactions in
the proposed catalytic cycle. For the catalytic cycle of this
(entries 1 and 6). In the presence of tertiary amines such as
triethylamine in an ethanol solution, ammonium formate was
obtained, although the catalytic activity was lower than that in
aqueous KOH (entry 7). The presence of a very polar solvent is
essential for the reaction, as the reaction did not proceed in
THF (entry 8). In comparison, the reaction with triethanol-
amine in water showed higher catalytic activity than with
triethylamine in EtOH (entries 9 and 7). Importantly, when
D O was used as a solvent instead of H O, the TOF was
61
62
reaction, Ahlquist and Yang suggested another reaction
pathway of hydrogen activation using theoretical calculations
(vide infra).
Herein we report our detailed studies on this highly active
catalytic hydrogenation reaction. The reversibility of the
reaction was realized by optimizing the reaction conditions.
In order to shed light on the reaction mechanism, the effect of
hydrogen pressure, base, and isotope were further studied. We
also estimated the energy profile of the reaction pathways and
elucidated that the existence of water molecules is essential for
2
2
lowered (entries 10 and 1), indicating that proton transfer is
involved in the key reaction of the catalytic cycle. It is notable
that treatment of 1 in KOH/D O at room temperature for 18.5
2
i
h led to H−D exchange at the py-CH -P Pr position (Scheme
1), suggesting that most of the active species in D O partially
2
2
2
Table 1. Hydrogenation of Carbon Dioxide Catalyzed by Ir(III)-Pincer Complexes 1 and 4
temp (°C) total pressure (MPa) time (h) yield (%) TON (103) TOF (10 h−1)
a
3
entry cat. (amt (μmol))
solvent
base
b
c
,
d
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1 (0.010)
1 (0.1)
H O
KOH
KOH
KOH
KOH
KOH
KOH
200
8.0
6.6
8.0
2
2
60
15
70
1
300
76
150
38
2
e
e
H O
2
200
b
c
,
f
1 (0.0010)
1 (1.0)
H O
2
120
100
25
48
40
48
2
3500
73
H O
2
0.13
0.043
0.001
c
1 (1.0)
H O
2
5.0
8.0
8.0
8.0
8.0
8.0
5.6
8.0
5.6
3.2
0
0
0
g
c
,
,
,
,
,
,
,
,
,
d
1 (0.010)
1 (10)
H O
2
200
200
200
200
200
200
200
200
200
12
15
0
58
29
c
c
c
c
c
c
c
c
d
d
d
d
i
EtOH
THF
Et N
3
2
0.075
0.38
1 (0.010)
1 (0.010)
1 (0.010)
1 (0.010)
4 (0.010)
4 (0.010)
4 (0.010)
N(CH CH OH)
2
0
29
0
14
53
32
2
2
3
3
H O
2
N(CH CH OH)
2
6
2
2
h
0
1
2
3
4
D O
2
KOH
KOH
KOH
KOH
KOH
2
21
110
63
H O
2
2
13
20
13
10
b
d
i
H O
2
13
13
13
100
63
7.7
H O
2
4.8
3.7
j
H O
2
48
a
b
c
Total pressure of inside of the reactor when the system reached the reaction temperature. Data taken from ref 58. A 50 mL autoclave was used,
d
e
and the whole apparatus was heated by an aluminum block. The initial total pressure was 5.0 MPa at 24 °C. A 300 mL autoclave was used, and
only the bottom part was heated by an oil bath. The pressure at 24 °C was 5.0 MPa. The initial total pressure was 6.0 MPa at 24 °C. A 25 mmol
f
g
h
i
portion of KOH was used. Total yield of HCOOK and DCOOK. The pressures of H and of CO were 1.0 and 2.5 MPa, respectively, at 24 °C.
2
2
j
The pressures of H and of CO were 1.0 and 1.0 MPa, respectively, at 24 °C.
2
2
6
743
dx.doi.org/10.1021/om2010172 | Organometallics 2011, 30, 6742−6750

Organometallics
Article
Scheme 1. H−D Exchange Reaction of Complex 1 and D O
Scheme 2. Summary of Our Previous Mechanistic
2
Investigations for Hydrogenation of CO using Iridium
2
Complex 1
incorporates deuterium in the methylene groups. Lowering the
H pressure decelerated the reaction even on maintaining the
2
CO pressure (entries 1/11 and 12/13). Lowering the CO
2
2
pressure also decelerated the reaction under the same H2
pressure (entries 13 and 14). Therefore, it was revealed that
the reaction rate is dependent on the H and CO pressures,
2
2
the strength of the base, and the use of very polar solvents.
. Dehydrogenation of Formic Acid and Its Salt
2
Catalyzed by the Ir(III) Complex 1. Dehydrogenation of
formic acid or its salt using the Ir(III) complex 1 as a catalyst
was also investigated. The aqueous solution of formic acid was
found to decompose and form dihydrogen, which was collected
by water replacement and analyzed by GC. The catalytic
activity was calculated by assuming that all the collected gas was
dihydrogen, and thus the TON of the reaction was estimated to
i
deprotonation from Ar−CH -P Pr and subsequent dearoma-
2
2
6
5−70
tization of the pyridine ring.
(3) Exposure of amidoiridium
dihydride complex 3 to 1 atm of hydrogen resulted in the
regeneration of 1. Therefore, a catalytic cycle which consists of
three elementary reactions, insertion of CO2 into 1,
deprotonative dearomatization with dissociation of formate
ligand in 2, and hydrogenation of 3 to regenerate the trihydride
complex 1, was proposed as shown in Scheme 3. Very
recently, the catalytic cycle of the reaction was studied
theoretically by Ahlquist and Yang, suggesting another
reaction pathway of hydrogen activation (vide infra).
64
be 890 (Table 2, entry 1). However, dehydrogenation did not
proceed when sodium formate was used (entry 2). The use of
triethylammonium formate greatly improved the catalytic
activity, and its initial TOF reached 120 000 h−1 (entry 3),
60
which is 6 times higher than the previously reported value by
46
the best Ru catalyst. The starting material was completely
consumed after 4 h, and the TOF for this condition was
61
62
calculated to be at least 1200 h− (entry 4). The TOF is about
1
38−58
Herein we report our DFT calculations for possible
intermediates, including 1−3 and transition states. The
the same order as those with previously reported catalysts.
When triethanolamine was used, the TON reached 4900 at 60
C (entry 5). It should be noted that, with triethanolamine as
the base, the hydrogenation of carbon dioxide also proceeded
71−73
74−76
B3LYP
method and Lanl2dz
on iridium and 6-31+
°
+
G(d,p) on the other atoms as basis sets were employed for the
calculations. The polarizable continuum model (PCM)
was used to estimate bulk solvent effects using water as the
solvent. The energies in the gas phase were also calculated to
estimate the solvent effects, and this will be discussed in section
3.5.
77−80
−1
smoothly (TOF = 14 000 h ; Table 1, entry 9), showing that
the iridium catalyst 1 can be used for reversible hydrogen
storage.
3
. Theoretical Studies on the Hydrogenation of
Carbon Dioxide using 1. Scheme 2 summarizes the
mechanistic studies of the hydrogenation with the Ir complex
It appears that a reliable method for evaluation of free energy
in solution has not been fully established. The gas-phase free
energy plus solvation free energy is often used. However, some
gas-phase entropy, in particular the translational entropy,
should in part be “quenched” in solution. Several scaling
60
1
in our previous communication. The key findings are briefly
discussed as follows. (1) Exposure of 1 to 1 atm of CO led to a
2
1
:1 mixture of 1 and the dihydridoformatoiridium complex 2,
which were unable to be isolated due to the equilibrium
between 1 and 2. (2) The chloroiridium dihydride complex 4,
81−89
methods have been proposed.
In the present paper, we
an analogue of 2, was treated with CsOH·H O to give the
show the calculated energy and entropy values in Table S1 and
2
amidoiridium dihydride complex 3, which was formed by
evaluated several “scaled” as well as unscaled free energies in
Table 2. Catalytic Dehydrogenation of Formic Acid or Its Salt using Ir Complex 1
TON (103)
TOF (10 h−1)
3
entry
solvent
base
amt of 1 (μmol)
temp (°C)
time (h)
yield (%)
1
2
H O
none
1.0
1.0
1.0
1.0
0.10
60
60
80
80
60
43
48
0.016
4
18
0
0.89
0
0.020
0
2
H O
NaOH
2
a
t
3
BuOH
Et N
3
40
100
10
2.0
5.0
4.9
120
1.2
a
tBuOH
4
a
Et N
3
5
H O
2
N(CH CH OH)
3
5
1.0
2
2
a
A 5 mmol portion of HCOOH·(base) was used.
6
744
dx.doi.org/10.1021/om2010172 | Organometallics 2011, 30, 6742−6750

Organometallics
Article
Scheme 3. Our Proposed Mechanism for Hydrogenation of
carbon dioxide is attacked by the weaker hydride ligand cis to
the nitrogen ligand, is more stable than 1 by 8.4 kcal/mol. The
energy barrier for insertion of carbon dioxide is calculated to be
58
CO using Iridium Complex 1
2
4
.2 kcal/mol. After a rearrangement via TS5/2, the proposed
intermediate 2 is formed and calculated to be more stable than
by 9.9 kcal/mol. In contrast, the formate intermediate 2 is
1
calculated to be higher in energy than 1 by 1.1 kcal/mol in the
gas phase.
3
.2. Deprotonation of Dihydridoformatoiridium Complex
and the Following Elimination of Formate Ligand. The
energetic profile for the reaction of dihydridoformatoiridium
complex 2 and a base is discussed next (Scheme 5).
Experimentally, the dihydridochloroiridium complex 4, a
model compound of 2, was treated with an excess amount of
cesium hydroxide to give the dihydridoamidoiridium complex 3
(
Scheme 2). The following mechanism is suggested for the
reaction of model compound 4: deprotonation on the α-carbon
of the pyridine ring, dearomatization of the pyridine ring to
form an amidoiridium intermediate, and finally elimination of
the chloride ligand. However, the directly deprotonated anionic
complex 6 is calculated to be less stable than 2 by 17.6 kcal/
mol. On the other hand, dissociation of the formate ligand of 2
to give the cationic dihydride complex 7 is a slightly endoergic
process by 1.1 kcal/mol (shown in a red solid line in Scheme
solution in Table S2 (see the Supporting Information). In the
following discussion, we simply adopt the values in which the
gas-phase vibrational and rotational entropies are retained, but
the gas-phase translational entropy is assumed to be totally
quenched (set to zero). Because of these uncertainties, we
avoid quantitative comparison and limit ourselves to qualitative
comparison of reaction free energy profiles within such
uncertainties. The energies for all stationary points were
calculated relative to the iridium trihydride complex 1 and
isolated substrates. In Schemes 4−7, the values in water and in
the gas phase are drawn as solid lines and dotted lines,
respectively. Our theoretical results will be compared with the
experimental data in section 3.6.
5). We were unable to locate transition states connecting 2 and
7
, presumably because of a low barrier. It is likely that there is
an equilibrium between the intermediate 2 and hydroxide
complex 8, which is about 3.0 kcal/mol lower in energy than 2.
The activation energy for proton transfer to give the water-
coordinated dihydridoamido complex 9 from 8 is calculated to
be 14.4 kcal/mol via TS8/9. Dissociation of water from 9 to 3
was slightly endoergic by 0.4 kcal/mol. Thus, this reaction is
more likely to take place via dissociation of the formate ligand
followed by deprotonation: 2 → 7 → 8 → 9 → 3.
3
.3. Addition of Dihydrogen to Dihydridoamidoiridium
3.1. Insertion of Carbon Dioxide into the Trihydridoiri-
Complex. Experimentally, addition of dihydrogen to dihydri-
doamidoiridium complex 3 gave 1 at ambient pressure (Scheme
2). The relative energy of the dihydrogen-coordinated complex
10 is calculated to be −10.6 kcal/mol and is less stable than 8
by 2.3 kcal/mol (Scheme 6). The heterolytic cleavage reaction
of the coordinated dihydrogen of 10, leading to 1, is exoergic by
dium Complex. First, the theoretical investigation of the
reaction of trihydridoiridium complex 1 and CO is discussed
2
(Scheme 4). In our experiment, complex 1 is in equilibrium
with dihydridoformato complex 2 in THF. For this
interconversion, we find that the intermediate 5, in which
a
Scheme 4. Free Energy Change for Insertion of Carbon Dioxide into the Trihydridoiridium-PNP Complex 1
a
The vibrational and rotational entropy terms only are included in solution.
6
745
dx.doi.org/10.1021/om2010172 | Organometallics 2011, 30, 6742−6750

Organometallics
Article
a
Scheme 5. Free Energy Change for Deprotonative Dearomatization of Dihydridoformatoiridium Complex 2
a
The vibrational and rotational entropy terms only are included in solution. The more favorable path is shown in red.
1
7.1 kcal/mol. The activation barrier via TS10/1 is 20.3 kcal/
cationic dihydride complex 7, coordination of the dihydrogen
molecule and hydroxide ion to 7 gives 12, and then heterolytic
mol relative to 8 (shown by black solid lines).
Milstein studied the addition of dihydrogen to a PNP-ligated
cleavage of the dihydrogen molecule assisted by hydrogen bond
90
62
phenylhydridoiridium complex by theoretical methods and
found that the activation energy is 18.2 kcal/mol, which is close
to our current result. In their study, addition of two water
molecules accelerates the cleavage of dihydrogen. We also
examined the addition of two explicit water molecules to 10
and to 1 in order to estimate the acceleration effect of water.
results in the formation of 1·H O. Yang supported the
2
mechanism by calculations using an exact catalyst model
without consideration of the coordination of water molecules.
The reaction from 2 to 1 via 7 and 12 has to overcome the
highest-energy transition state TS12/1 with a barrier of 14.4
kcal/mol relative to 8 (Scheme 7). On the basis of these results,
the two routes 2 → 7 (→8) → 12 → 1 and 2 → 7 → 8 → 9 →
We found the cleavage transition state TS10/11·2H O, which
2
is 5.8 kcal/mol above 10·2H O and is featured mainly as a
3 → 10·2H O → 11 → 1 seem to compete, since the activation
2
2
protonation transition state, as the distance between oxygen
atom of the water molecule and hydrogen atom of the
dihydrogen molecule is still long (red solid lines). TS10/
barrier difference is very small.
3.5. Theoretical Investigation in Gas Phase To Reveal the
Solvent Effect. The relative energies for all stationary points in
the gas phase were also compared to estimate the solvent effect
(Schemes 4−7). The activation energy of insertion of carbon
dioxide to trihydride complex 1 via TS5/2 was 23.1 kcal/mol
higher than that in water. The energy difference between
trihydride complex 1 and dihydrideformate complex 2 was 1.1
kcal/mol, which is less stable than that in water.
1
1·2H O did not directly give the intermediate 1·2H O, but
2
2
intermediate 11 was located in which the pyridylmethyl
position is protonated with the long distance between the
oxygen atom of water and the hydrogen atom of the
dihydrogen molecule (2.092 Å). Attempts to locate a transition
state connecting intermediate 11 and 1·2H O proved to be
2
unsuccessful, probably due to the very flat potential energy
surface and the very early nature of this transition state. It
The deprotonation of 2 by hydroxide ion leading to anion
intermediate 6 is highly exergonic by −45.4 kcal/mol relative to
1, much lower in energy than in water, which originates from
the high instability of the hydroxide ion in the gas phase. The
succeeding elimination of the formate ligand gave the
dihydridoamido complex 3 with an endoergicity of 16.7 kcal/
mol. It is notable that another intermediate 7, which was
generated by dissociation of formate from 2, was considerably
destabilized to 90.8 kcal/mol with reference to 1 in the gas
phase. Thus, the conversion from 2 to 3 or 8 in less polar
solvents seems more likely to proceed via deprotonation to give
should be noted that TS10/11·2H O and 11 are very similar in
2
energy (ΔΔG = 0.7 kcal/mol). Therefore, TS10/11·2H O
2
might be considered as a quasi-concerted, but highly
asynchronous, transition state. The transition state containing
one water molecule, TS10/1·H O, is 8.2 kcal/mol higher in
2
energy than TS10/11·2H O but lower in energy than TS10/1
2
by 5.1 kcal/mol (blue solid lines). The free energy change ΔG
from 10 to 10·H O is −0.7 kcal/mol and ΔG from TS10/1 to
2
TS10/1·H O is −5.1 kcal/mol. In sharp contrast, the additional
2
91
water molecule, participating in the reaction, decreased the
activation barrier for the dihydrogen activation. The overall
intermediate 6 followed by dissociation of formate.
The transition state for the heterolytic cleavage of
dihydrogen on dihydridoamido dihydrogen complex TS10/1
was 36.8 kcal/mol higher in energy than intermediate 6. The
barrier via TS10/11·2H O relative to 8 is 7.0 kcal/mol, lower
2
than that via TS8/9 by 7.4 kcal/mol.
3
.4. Dissociation of Formate from Dihydrideformate
Complex. Using a simplified model with 2,6-
H PCH ) C H N, Ahlquist theoretically investigated an
transition state containing one water molecule, TS10/1·H O, is
2
almost the same in energy as TS10/1. Therefore, the reaction is
unlikely to take place in a nonpolar or gas-phase environment.
3.6. Comparison between the Theoretical and Exper-
imental Results. Now we compare our theoretical results with
(
2
2
2
5
3
alternative route from 2 to 1 via dissociation of the formate
ligand; namely, the formate ligand on 2 first dissociates to give
8
6
746
dx.doi.org/10.1021/om2010172 | Organometallics 2011, 30, 6742−6750

Organometallics
Article
Scheme 6. Free Energy Change for Addition of Dihydrogen to Dihydridoamidoiridium Complex 3 and Effect of Addition of
a
Two Water Molecules
a
The PNP ligand was simplified in the presence of two water molecules for clarity. The vibrational and rotational entropy terms only are included in
solution.
our experimental data. For the initial CO insertion into Ir
trihydride 1, the intermediate 2 was calculated to be 9.9 kcal/
mol more stable than 1 in the water phase. In contrast, formate
2
Scheme 7. Free Energy Change for Dissociation of Formate
from 2 and the Following Heterolytic Cleavage of
Dihydrogen
a
2
was calculated to be higher in energy than 1 by 1.1 kcal/mol
in the gas phase. Experimentally, we observed a 1:1 equilibrium
mixture of 1 and 2 in THF, but it was not in the real reaction
media KOH/H O/THF.
2
For the whole catalytic cycle, we propose that the reaction
can take place via both/either 2 → 7 → 8 → 9 → 3 →
1
0·2H O → 11 → 1 and/or 2 → 7 → 12 → 1 (Scheme 8).
2
The rate-determining step could be the deprotonative
dearomatization step 8 → 9 via TS8/9 for the former reaction
pathway and the hydrogenolysis step 12 → 1 via TS12/1 for
the latter pathway. The relative energies of the two rate-
determining transition states are nearly identical in solution in
the present scaling scheme. The step 8 → 9 should be slowed
down in D O (entry 10, Table 1), since the methylene protons
2
in 8 are easily replaced by deuterium (Scheme 1). In addition,
the equilibrium among 7, 8, 9, and 3 should be strongly affected
by the concentration of base (entries 1 and 6−9). However,
this step should be less affected by the hydrogen pressure. On
a
The vibrational and rotational entropy terms only are included in
solution.
6
747
dx.doi.org/10.1021/om2010172 | Organometallics 2011, 30, 6742−6750

Organometallics
Article
Scheme 8. Proposed Mechanism for the Hydrogenation of
EXPERIMENTAL SECTION
■
CO by the Ir(III) Complex
2
General Procedures. All manipulations involving air- and
moisture-sensitive compounds were carried out by using standard
Schlenk techniques or a glovebox under argon that was purified by
passing through a hot column packed with BASF catalyst R3-11. All
the organic solvents used for reactions were distilled under argon after
drying over an appropriate drying agent or passed through solvent
purification columns. Water was degassed by two freeze−thaw−pump
cycles before use. Most of the reagents were used without further
purification unless otherwise specified. NMR spectra were recorded on
5
00 MHz spectrometers. GC analysis was performed by using N as
2
carrier gas and TCD as a detector.
Standard Procedure for Hydrogenation of Carbon Dioxide.
Iridium trihydride complex 1 (10.7 mg, 20 μmol) was dissolved in
THF (4.00 mL) and diluted to 100 μmol/L. In a 50 mL stainless
autoclave the mixture of catalyst (100 μL, 0.010 μmol), base (5.0
mmol), and solvents was charged and pressurized by a 1:1 mixture of
CO and H . The reaction mixture was heated by an aluminum block
2
2
to 200 °C and stirred for 2 h. Sodium 3-(trimethylsilyl)-1-
propanesulfonate (11.3 mg, 51.8 μmol) was added to the reaction
mixture as an internal standard, and an aliquot of the mixture was
1
dissolved in D O or CDCl for an estimation of the yield by using H
2
3
NMR measurement. DMSO-d was used as an internal standard to
6
2
estimate the yield of DCOOK by using H NMR measurements. The
same reaction was repeated at least twice to confirm the
reproducibility.
Procedure for Hydrogenation of Carbon Dioxide on a Large
Scale. In a 300 mL stainless autoclave the mixture of catalyst (1.0 mL,
0
.10 μmol) and 1 M aqueous solution of KOH (50 mL, 50 mmol)
were charged and pressurized by a 1:1 mixture of CO and H . The
2
2
reaction mixture was heated by an oil bath to 200 °C and stirred for 2
h. The entire reaction vessel cannot be heated, and the temperature at
the top of the autoclave is lower than that estimated. Sodium 3-
(
trimethylsilyl)-1-propanesulfonate (11.5 mg, 52.6 μmol) was added
to the reaction mixture as an internal standard, and an aliquot of the
mixture was dissolved in D O for an estimation of the yield by using
the other hand, the hydrogenolysis step 12 → 1 should be
accelerated by the higher hydrogen pressure, showing good
agreement with our experimental observations (entries 1, 4, and
2
1
H NMR measurements. A 7.64 mmol amount (15%) of HCOOK was
obtained.
1
0), although the kinetic isotope effect we observed in D O is
2
Observation of H−D Exchange of 1. In an NMR sample tube, 1
not reasonably explained by this mechanism. According to the
observations mentioned above, it may be suggested that these
two competing pathways operate concurrently.
(10 mg) in C D (0.5 mL) was introduced and 1 M KOH solution in
6 6
D O (0.1 mL) was added. The solution was stored at room
temperature for 18.5 h. The integral ratio of the py-CH -PiPr
2
2
2
The overall catalytic cycle producing the free formate anion
resulted in a stabilization of 27.7 kcal/mol. If the free formate
anion undergoes the reversed reaction, it is required to
surmount TS 1/12 or TS 9/8, that is 29.2 kcal/mol from 1.
position decreased from 4 to 0.92.
Standard Procedure for Dehydrogenation of Ammonium
Formate. In a 20 mL Schlenk tube was charged a THF solution of 1
(10 mmol/L, 100 μL, 1.0 μmol) and warmed to 60 °C. The reaction
was started by adding a solution of triethylammonium formate (736
t
mg, 5.00 mmol) in BuOH (4.0 mL). The gas that evolved during
CONCLUSION
reaction was collected by water replacement. The catalytic activity was
calculated from the volume of the collected gas, supposing that all the
gas consists of hydrogen only. The content of the collected gas was
confirmed to be hydrogen by GC analysis.
■
The highly active catalytic hydrogenation of carbon dioxide to
form formic acid salts and dehydrogenation of formic acid salt
have been established. Both reactions were catalyzed by the
PNP-ligated iridium trihydride complex 1, and under certain
conditions, the activities were higher than for any other catalyst
previously reported. A theoretical investigation of the hydro-
genation of carbon dioxide in the presence of aqueous base
revealed two competitive pathways. The suggested rate-
determining step for each pathway was the deprotonative
dearomatization step via TS8/9 and the hydrogenolysis step via
TS12/1. The results nicely explained the effect of hydrogen
pressure, base, and solvent. Reversible hydrogen fixation onto
carbon dioxide is an attractive, safe, and inexpensive method for
hydrogen storage. The current highly active catalyst could find
practical application for this purpose.
Theoretical Calculations. Geometry optimization and vibrational
frequency calculation of all reactants, intermediates, transition states,
9
2
and products were performed by using the Gaussian 03 program and
71−73
74−76
the B3LYP
method with combined basis sets (Lanl2DZ
basis sets for iridium atom and 6-31G++(d,p) basis sets for the other
93,94
atoms). IRC
calculations were also performed. The solvent effect
was estimated by single-point energy calculations of all intermediates
77−80
and transition states using PCM
with water as the solvent, in
which the UAHF radius was used to build the cavity of the solute and
the nonelectrostatic terms were also included. The HF/6-31+G(d)
method was tested and gave a solvation energy similar to that for the
key step in the B3LYP method. Since the entropy in the solution is
well-known to be overestimated by the gas-phase calculations, the
vibrational and rotational entropy terms only are included in the free
energy in solution.
6
748
dx.doi.org/10.1021/om2010172 | Organometallics 2011, 30, 6742−6750

Organometallics
Article
(
23) Urakawa, A.; Jutz, F.; Laurenczy, G.; Baiker, A. Chem. Eur. J.
ASSOCIATED CONTENT
■
2
(
4
007, 13, 3886.
24) Himeda, Y.; Miyazawa, S.; Hirose, T. ChemSusChem 2011, 4,
87.
*
S
Supporting Information
Text, tables, and figures giving experimental procedures, the
complete ref 88, and calculated structures of all intermediates
and transition states. This material is available free of charge via
the Internet at http://pubs.acs.org.
(25) Schaub, T.; Paciello, R. A. Angew. Chem., Int. Ed. 2011, 50, 7278.
(26) Federsel, C.; Jackstell, R.; Boddien, A.; Laurenczy, G.; Beller, M.
ChemSusChem 2010, 3, 1048.
(
(
27) Sanz, S.; Azua, A.; Peris, E. Dalton Trans. 2010, 39, 6339.
28) Hao, C.; Wang, S.; Li, M.; Kang, L.; Ma, X. Catal. Today 2011,
AUTHOR INFORMATION
■
1
60, 184.
Corresponding Author
(29) Himeda, Y.; Onozawa-Komatsuzaki, N.; Sugihara, H.; Kasuga,
*
E-mail: nozaki@chembio.t.u-tokyo.ac.jp (K.N.).
K. Organometallics 2007, 26, 702.
30) Schmeier, T. J.; Dobereiner, G. E.; Crabtree, R. H.; Hazari, N. J.
Am. Chem. Soc. 2011, 133, 9274.
(
Present Address
§
Department of Applied Chemistry, Faculty of Science and
(31) Tsai, J. C.; Nicholas, K. M. J. Am. Chem. Soc. 1992, 114, 5117.
Engineering, Chuo University, 1-13-27 Kasuga, Bunkyo-ku,
(32) Leitner, W.; Dinjus, E.; Gaßner, F. J. Organomet. Chem. 1994,
112-8551 Tokyo, Japan.
4
75, 257.
33) Angermund, K.; Baumann, W.; Dinjus, E.; Fornika, R.; Go
H.; Kessler, M.; Kruger, C.; Leitner, W.; Lutz, F. Chem. Eur. J. 1997, 3,
55.
34) Tai, C. C.; Chang, T.; Roller, B.; Jessop, P. G. Inorg. Chem.
003, 42, 7340.
35) Sakamoto, M.; Shimizu, I.; Yamamoto, A. Organometallics 1994,
3, 407.
36) Federsel, C.; Boddien, A.; Jackstell, R.; Jennerjahn, R.; Dyson, P.
(
̈
rls,
ACKNOWLEDGMENTS
̈
■
7
(
2
(
1
(
This work was supported in part (at the University of Tokyo)
by a Funding Program for World-Leading Innovative R&D on
Science and Technology from the Japan Society for the
Promotion of Science (JSPS) and Global COE Program
“Chemistry Innovation through Cooperation of Science and
Engineering”, MEXT, Japan, as well as (at Kyoto University) by
the Japan Science and Technology Agency with a Core
Research for Evolutional Science and Technology grant in the
Area of High Performance Computing for Multiscale and
Multiphysics.
J.; Scopelliti, R.; Laurenczy, G.; Beller, M. Angew. Chem., Int. Ed. 2010,
49, 9777.
(37) Langer, R.; Diskin-Posner, Y.; Leitus, G.; Shimon, L. J. W.; Ben-
David, Y.; Milstein, D. Angew. Chem., Int. Ed. 2011, 50, 9948.
(
38) Laine, R. M.; Rinker, R. G.; Ford, P. C. J. Am. Chem. Soc. 1977,
9, 252.
39) Gao, Y.; Kuncheria, J.; J. Puddephatt, R.; Yap, G. P. A. Chem.
Commun. 1998, 2365.
40) Boddien, A.; Loges, B.; Junge, H.; Beller, M. ChemSusChem
008, 1, 751.
41) Fellay, C.; Dyson, P. J.; Laurenczy, G. Angew. Chem., Int. Ed.
008, 47, 3966.
42) Loges, B.; Boddien, A.; Junge, H.; Beller, M. Angew. Chem., Int.
Ed. 2008, 47, 3962.
43) Fellay, C.; Yan, N.; Dyson, P. J.; Laurenczy, G. Chem. Eur. J.
009, 15, 3752.
44) Junge, H.; Boddien, A.; Capitta, F.; Loges, B.; Noyes, J. R.;
Gladiali, S.; Beller, M. Tetrahedron Lett. 2009, 50, 1603.
45) Loges, B.; Boddien, A.; Junge, H.; Noyes, J. R.; Baumann, W.;
Beller, M. Chem. Commun. 2009, 4185.
46) Morris, D. J.; Clarkson, G. J.; Wills, M. Organometallics 2009, 28,
133.
47) Boddien, A.; Gar
D.; Jackstell, R.; Junge, H.; Beller, M. Angew. Chem., Int. Ed. 2011,
411.
48) Forster, D.; Beck, G. R. J. Chem. Soc. D: Chem. Commun. 1971,
072.
49) Strauss, S. H.; Whitmire, K. H.; Shriver, D. F. J. Organomet.
Chem. 1979, 174, C59.
50) Fukuzumi, S.; Kobayashi, T.; Suenobu, T. ChemSusChem 2008,
, 827.
9
(
REFERENCES
■
(
1) Reutemann, W.; Kieczka, H. Formic Acid. In Ullmann’s
Encyclopedia of Industrial Chemistry, Electronic Release, 7th ed.; Wiley-
(
2
(
2
VCH: Weinheim, Germany, 2005.
(
(
(
(
(
(
(
2) Leonard, J. D. Eur. Patent EP5998, 1978.
3) Aguilo, A.; Horlenko, T. Hydrocarb. Process. 1980, 60, 120.
4) Arsem, W. C., U.S. Patent 1 606 394, 1926.
(
5) Lindkvist, E. Patent GB1049013, 1966.
(
2
(
6) Jessop, P. G.; Ikariya, T.; Noyori, R. Chem. Rev. 1995, 95, 259.
7) Leitner, W. Angew. Chem., Int. Ed. 1995, 34, 2207.
8) Jessop, P. G.; Joo, F.; Tai, C.-C. Coord. Chem. Rev. 2004, 248,
2
(
4
(
3
(
3
425.
9) Federsel, C.; Jackstell, R.; Beller, M. Angew. Chem., Int. Ed. 2010,
9, 6254.
10) Wang, W.; Wang, S.; Ma, X.; Gong, J. Chem. Soc. Rev. 2011, 40,
703.
11) Williams, R.; Crandall, R.; Bloom, A. Appl. Phys. Lett. 1978, 33,
81.
(
(
4
(
̈
tner, F.; Federsel, C.; Sponholz, P.; Mellmann,
6
(
1
(
(
12) Joo
́, F. ChemSusChem 2008, 1, 805.
(
13) Enthaler, S.; von Langermann, J.; Schmidt, T. Energ. Environ. Sci.
2
010, 3, 1207.
14) Johnson, T. C.; Morris, D. J.; Wills, M. Chem. Soc. Rev. 2010, 39,
1.
15) Loges, B.; Boddien, A.; Gar
Catal. 2010, 53, 902.
(
8
(
1
(
̈
tner, F.; Junge, H.; Beller, M. Top.
(51) Coffey, R. S. Chem. Commun. 1967, 923.
(
16) Jessop, P. G.; Ikariya, T.; Noyori, R. Nature 1994, 368, 231.
(
(
52) Himeda, Y. Green Chem. 2009, 11, 2018.
53) Fukuzumi, S.; Kobayashi, T.; Suenobu, T. J. Am. Chem. Soc.
(
17) Jessop, P. G.; Hsiao, Y.; Ikariya, T.; Noyori, R. J. Am. Chem. Soc.
1
(
996, 118, 344.
18) Gao, Y.; Kuncheria, J. K.; Jenkins, H. A.; Puddephatt, R. J.; Yap,
G. P. A Dalton Trans. 2000, 3212.
19) Munshi, P.; Main, A. D.; Linehan, J. C.; Tai, C. C.; Jessop, P. G.
J. Am. Chem. Soc. 2002, 124, 7963.
20) Man, M. L.; Zhou, Z.; Ng, S. M.; Lau, C. P. Dalton Trans. 2003,
727.
2
010, 132, 1496.
54) Yoshida, T.; Ueda, Y.; Otsuka, S. J. Am. Chem. Soc. 1978, 100,
3941.
(
(
(55) Paonessa, R. S.; Trogler, W. C. J. Am. Chem. Soc. 1982, 104,
(
3
3529.
(56) Shin, J. H.; Churchill, D. G.; Parkin, G. J. Organomet. Chem.
2002, 642, 9.
(
21) Hayashi, H.; Ogo, S.; Fukuzumi, S. Chem. Commun. 2004, 2714.
(
22) Ogo, S.; Kabe, R.; Hayashi, H.; Harada, R.; Fukuzumi, S. Dalton
̈
(57) Boddien, A.; Loges, B. R.; Gartner, F.; Torborg, C.; Fumino, K.;
Junge, H.; Ludwig, R.; Beller, M. J. Am. Chem. Soc. 2010, 132, 8924.
Trans. 2006, 4657.
6
749
dx.doi.org/10.1021/om2010172 | Organometallics 2011, 30, 6742−6750

Organometallics
Article
(
58) Boddien, A.; Mellmann, D.; Gar
Dyson, P. J.; Laurenczy, G.; Ludwig, R.; Beller, M. Science 2011, 333,
733.
59) For examples of catalytic reactions using iridium-pincer
̈
tner, F.; Jackstell, R.; Junge, H.;
1
(
complexes, see: Choi, J.; MacArthur, A. H. R.; Brookhart, M.;
Goldman, A. S. Chem. Rev. 2011, 111, 1761.
(
60) Tanaka, R.; Yamashita, M.; Nozaki, K. J. Am. Chem. Soc. 2009,
131, 14168.
(
61) Ahlquist, M. S. G. J. Mol. Catal. A 2010, 324, 3.
(
(
(
62) Yang, X. ACS Catal. 2011, 1, 849.
63) Laurenczy, G.; Joo, F.; Nadasdi, L. Inorg. Chem. 2000, 39, 5083.
64) The purity calculated by GC was 35% in entry 1 because argon
in the headspace of the Schlenk tube was contaminated by the
collected gas. The amount of carbon dioxide and any other gases were
estimated to be less than 100 ppm by S/N ratio of GC spectra
detected by GC analysis.
(
65) Ben-Ari, E.; Leitus, G.; Shimon, L. J. W.; Milstein, D. J. Am.
Chem. Soc. 2006, 128, 15390.
66) Zhang, J.; Leitus, G.; Ben-David, Y.; Milstein, D. Angew. Chem.,
Int. Ed. 2006, 45, 1113.
67) Zhang, J.; Gandelman, M.; Shimon, L. J. W.; Milstein, D. Dalton
Trans. 2007, 107.
68) Schaub, T.; Radius, U.; Diskin-Posner, Y.; Leitus, G.; Shimon, L.
J. W.; Milstein, D. Organometallics 2008, 27, 1892.
69) Vuzman, D.; Poverenov, E.; Shimon, L. J. W.; Diskin-Posner, Y.;
Milstein, D. Organometallics 2008, 27, 2627.
70) Gunanathan, C.; Shimon, L. J. W.; Milstein, D. J. Am. Chem. Soc.
009, 131, 3146.
(
(
(
(
(
2
(
71) Becke, A. D. Phys. Rev. A 1988, 38, 3098.
(
(
(
(
(
(
72) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
73) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
74) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270.
75) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.
76) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284.
77) Cances, E.; Mennucci, B.; Tomasi, J. J. Chem. Phys. 1997, 107,
3032.
(
78) Mennucci, B.; Tomasi, J. J. Chem. Phys. 1997, 106, 5151.
(
79) Cossi, M.; Barone, V.; Mennucci, B.; Tomasi, J. Chem. Phys. Lett.
1
(
2
998, 286, 253.
80) Cossi, M.; Scalmani, G.; Rega, N.; Barone, V. J. Chem. Phys.
002, 117, 43.
(
(
(
81) Amzel, L. M. Proteins: Struct., Funct., Bioinf. 1997, 28, 144.
82) Hermans, J.; Wang, L. J. Am. Chem. Soc. 1997, 119, 2707.
83) Mammen, M.; Shakhnovich, E. I.; Deutch, J. M.; Whitesides, G.
M. J. Org. Chem. 1998, 63, 3821.
84) Strajbl, M.; Sham, Y. Y.; Villa,
Chem. B 2000, 104, 4578.
(
̀
J.; Chu, Z. T.; Warshel, A. J. Phys.
(
(
85) Yu, Z.-X.; Houk, K. N. J. Am. Chem. Soc. 2003, 125, 13825.
86) Liang, Y.; Liu, S.; Xia, Y.; Li, Y.; Yu, Z.-X. Chem. Eur. J. 2008, 14,
4
361.
87) Sugiyama, A.; Ohnishi, Y.-y.; Nakaoka, M.; Nakao, Y.; Sato, H.;
Sakaki, S.; Nakao, Y.; Hiyama, T. J. Am. Chem. Soc. 2008, 130, 12975.
88) Huang, F.; Lu, G.; Zhao, L.; Li, H.; Wang, Z.-X. J. Am. Chem.
Soc. 2010, 132, 12388.
89) Ohnishi, Y.-y.; Nakao, Y.; Sato, H.; Sakaki, S. Organometallics
006, 25, 3352.
90) Iron, M. A.; Cohen, E. B.-A. R.; Milstein, D. Dalton Trans. 2009,
433.
91) The energies of 6 and 7 in THF were calculated to be −8.5 and
0.0 kcal/mol with reference to 1, respectively, which also showed the
stability of 6 over that of 7.
92) Frisch, M. J., et al. Gaussian 03, Revision C.02; Gaussian, Inc.,
Wallingford, CT, 2004.
(
(
(
2
(
9
(
1
(
(
(
93) Gonzalez, C.; Schlegel, H. B. J. Phys. Chem. 1990, 94, 5523.
94) Gonzalez, C.; Schlegel, H. B. J. Chem. Phys. 1989, 90, 2154.
6
750
dx.doi.org/10.1021/om2010172 | Organometallics 2011, 30, 6742−6750