Ruthenium-catalyzed hydrogenation of carbon dioxide to formic acid in alcohols

Article abstract of DOI:10.1002/ejic.200300847

Catalytic hydrogenation of CO₂ to formic acid with the (solvento)metal hydride complex, TpRu(PPh₃)(CH₃CN)H [Tp = hydrotris(pyrazolyl)borate], in various alcohols was studied. High-pressure NMR monitoring of the catalytic reaction in non-acidic methanol shows that the observable intermediate is a formate complex resulting from CO₂ insertion into the Ru-H bond and is stabilized by the hydrogen-bonding interaction between the formato ligand and a methanol molecule. However, in the case of the acidic alcohol, CF₃CH₂OH, the observable intermediates are [TpRu(PPh₃)(CH₃CN)₂] ⁺CF₃CH₂O^- and the alkyl carbonate complex, TpRu(PPh₃)(η²O₂COCH ₂CF₃), which are formed by the reaction of CO₂ with the alkoxide species, TpRu(PPh₃)(CH₃CN)(OCH ₂CF₃), generated by a very facile reaction between TpRu(PPh₃)(CH₃CN)H and CF₃CH₂OH. We propose that the productive catalytic cycles of the reactions conducted in a variety of alcohols are similar to the one we formulated for the catalytic hydrogenation of CO₂ in hydrous THF. The formic acid is produced by the transfer of a hydride and a proton from the transient alcohol hydride intermediate, TpRu(PPh₃)(ROH)H, to an approaching CO₂ molecule. The activity of TpRu(PPh₃)(CH₃CN)H is higher in CF₃CH₂OH than in methanol and other non-acidic alcohols and it is probably due to the enhanced electrophilicity of the carbon atom of CO₂, which results from the strong interaction between the proton of the highly acidic alcohol in TpRu(PPh₃)(CF₃CH ₂OH)H and an oxygen atom of CO₂. The electrophilic carbon atom of CO₂ could in turn abstract the hydride from Ru-H in a more facile manner. Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2004.

Full text of DOI:10.1002/ejic.200300847

SHORT COMMUNICATION
Ruthenium-Catalyzed Hydrogenation of Carbon Dioxide to Formic Acid
in Alcohols
Siu Man Ng,[a] Chuanqi Yin,[a] Chi Hung Yeung,[a] Tak Chung Chan,[a] and Chak Po Lau*[a]
Keywords: Ruthenium / Hydrogenation / Alcohols / Carbon dioxide / Formic acid
Catalytic hydrogenation of CO2 to formic acid with the (solv-
ento)metal hydride complex, TpRu(PPh3)(CH3CN)H [Tp =
hydrotris(pyrazolyl)borate], in various alcohols was studied.
High-pressure NMR monitoring of the catalytic reaction in
non-acidic methanol shows that the observable intermediate
is a formate complex resulting from CO2 insertion into the
Ru−H bond and is stabilized by the hydrogen-bonding inter-
variety of alcohols are similar to the one we formulated for
the catalytic hydrogenation of CO2 in hydrous THF. The
formic acid is produced by the transfer of a hydride and a
proton from the transient alcohol hydride intermediate,
TpRu(PPh3)(ROH)H, to an approaching CO2 molecule. The
activity of TpRu(PPh3)(CH3CN)H is higher in CF3CH2OH
than in methanol and other non-acidic alcohols and it is prob-
action between the formato ligand and a methanol molecule. ably due to the enhanced electrophilicity of the carbon atom
However, in the case of the acidic alcohol, CF3CH2OH, the
observable intermediates are [TpRu(PPh3)(CH3CN)2]+-
CF3CH2O− and the alkyl carbonate complex, TpRu(PPh3)(η2-
O2COCH2CF3), which are formed by the reaction of CO2
with the alkoxide species, TpRu(PPh3)(CH3CN)(OCH2CF3),
of CO2, which results from the strong interaction between
the proton of the highly acidic alcohol in TpRu(PPh3)(CF3-
CH2OH)H and an oxygen atom of CO2. The electrophilic car-
bon atom of CO2 could in turn abstract the hydride from
Ru−H in a more facile manner.
generated
by
a
very
facile
reaction
between
TpRu(PPh3)(CH3CN)H and CF3CH2OH. We propose that the
productive catalytic cycles of the reactions conducted in a
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2004)
Introduction
ent hydroxo species, which then associates an H2 molecule
to regenerate TpRu(PPh3)(H2O)H by σ-metathesis between
the hydroxo and the η2-H2 ligands.[5]
Transition-metal-catalyzed CO2 hydrogenation to formic
acid remains an area of interest, as it provides a promising
approach for utilizing carbon dioxide as a raw material for
large-scale chemical synthesis.[1] Although it has been
known for some time that addition of water to active Pd,[2]
Rh,[3] and Ru[4] systems is effective in enhancing the cata-
lytic hydrogenation of CO2 to formic acid, no compelling
experimental and theoretical data had been provided to ex-
plain the positive ‘‘water effect’’ until the publication of our
recent work on the promoting effect of water in the catalytic
hydrogenation CO2 with the ruthenium complex
TpRu(PPh3)(CH3CN)H (1) [Tp ϭ hydrotris(pyrazolyl)bor-
ate]. With the support of high-pressure NMR spectroscopy
and theoretical calculations, we proposed a reaction mecha-
nism to account for the ‘‘water effect’’. The key intermedi-
ate in the catalytic cycle is an (aquo)ruthenium hydride
species, TpRu(PPh3)(H2O)H, which is generated through li-
gand displacement by H2O. It transfers a hydride and a
proton of the coordinated H2O to the approaching CO2
molecule to give formic acid and is converted into a transi-
Continuing our investigation on the catalytic hydrogen-
ation of CO2 to formic acid with 1, we studied the 1-cata-
lyzed reactions in alcohols. The accelerating effects of cer-
tain alcohols on the hydrogenation of carbon dioxide, cata-
lyzed by (trimethylphosphane)ruthenium complexes, have
recently been reported.[6]
Results and Discussion
Hydrogenation of Carbon Dioxide to Formic Acid with 1
We have previously learned that 1 catalyzes the hydrogen-
ation of carbon dioxide to yield formic acid with a low turn-
over number in anhydrous THF (see Entry 1, Table 1), but
the catalytic activity of 1 was greatly enhanced when 5%
(by volume) of water was added to the system (Entry 2,
Table 1).[5] Perusal of Table 1 reveals significant increase in
the catalytic activity of 1 in alcohols too (relative to that of
1 in anhydrous THF), although the enhancement effects are
not as large as in THF/H2O, except for the acidic alcohol,
CF3CH2OH. In the catalytic reactions using alcohols as sol-
vents, the yields of the ester side products are very low. Also
noteworthy is the trend of decreasing turnover numbers of
[a]
Department of Applied Biology and Chemical Technology, The
Hong Kong Polytechnic University
Hung Hom, Kowloon, Hong Kong, China
Fax: (internat.) ϩ 852-2364-9932
E-mail: bccplau@inet.polyu.edu.hk
1788
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/ejic.200300847
Eur. J. Inorg. Chem. 2004, 1788Ϫ1793

Ruthenium-Catalyzed Hydrogenation of Carbon Dioxide
SHORT COMMUNICATION
TpRu(PPh3)(CH3CN)(η1-OCHO)·HOCH3. The formato li-
gand in 2 is hydrogen-bonded to a methanol molecule. Un-
fortunately, we have not been able to observe any down-
field-shifted OH proton signals due to the hydrogen-bonded
CH3OH, probably because of its undergoing rapid ex-
change with the molecules in the bulk, or because the signal
is masked by the large number of peaks due to the Tp and
phosphane ligands in the region.
Table 1. Hydrogenation of CO2 to formic acid with
TpRu(PPh3)(CH3CN)H (1) in alcohols
Entry[a]
Solvent
Turnover no.
Turnover no.
of ester[b]
of formic acid[b]
1
2
3
4
5
6
7
8
9
THF
30
720
250
388
230
127
213
141
72
Ϫ
Ϫ
THF/H2O[c]
THF/H2O[c,d]
CH3OH
75
11
4
12
Ϫ
15
Ϫ
C2H5OH
C2H5OH[d]
n-C3H7OH
i-C3H7OH
t-C4H9OH
CF3CH2OH
10
1815
[a]
Typical reaction conditions: catalyst 0.013 mmol; NEt3 2 mL;
solvents 20 mL; total pressure 50 atm (CO2/H2 ϭ 1:1); temperature
[b]
100 °C; reaction time 16 h. Turnover no: mol of product/mol of
[c]
[d]
complex. THF/H2O (19 mL/1 mL). Reaction time 4 h.
formic acid with increasing bulkiness of the alcohols (En-
tries 4Ϫ9, Table 1).
Attempts to synthesize 2 on a preparative scale by treat-
ing 1 with pressurized CO2 were frustrated by its decompo-
sition to the hydride species and some intractable materials
during workup, in which the CO2 pressure has to be re-
moved. Similar problems were encountered in our previous
efforts to isolate the formate complexes, TpRu(PPh3)-
In Situ High-Pressure NMR Studies
To gain more insight into the mechanisms of the catalytic
CO2-hydrogenation reactions by 1 in alcohols, we carried
out a series of in situ high-pressure NMR studies.
(CH3CN)(η1-OCHO)
and
TpRu(PPh3)(CH3CN)(η1-
OCHO)·H2O.[5] Treatment of the chloro complex
TpRu(PPh3)(CH3CN)Cl with sodium formate also resulted
in the formation of the hydride complex 1 as the major
product. It seems that all these (Tp)ruthenium formate
species are only stable under CO2 pressure.
(a) Reaction of 1 with CO2 in Methanol/THF
A CH3OH/[D8]THF (1:1 by volume) solution of 1 in a
5-mm Wilmad pressure-valved NMR tube was pressurized
with 6 atm of 13CO2 at room temperature. The 31P{1H}
NMR spectrum of the resulting solution recorded after
10 min shows the disappearance of 1 and the presence of a
new complex 2 which gives a singlet at δ ϭ 52.9 ppm. The
(b) Reaction of 1 with CO2 and H2 in Methanol/THF
A CH3OH/[D8]THF (1:1 by volume) solution of 1 con-
1H{13C} NMR spectrum of the same solution clearly shows taining 20 equiv. of Et3N was pressurized with 6 atm of
the set of nine peaks pertaining to the Tp ligand. A singlet 13CO2 and H2 in a 5-mm Wilmad pressure-valved NMR
at δ ϭ 2.30 ppm, integrated for 3 H, is assignable to the tube; the 31P{1H} and 1H NMR spectra of the solution
coordinated CH3CN, a singlet (1 H) at δ ϭ 7.63 ppm can taken within 10 min at room temperature shows the disap-
be attributed to the hydrogen atom of a formate ligand. pearance of 1 and the formation of 2. A minute amount of
1
This signal is split into a doublet in the 13C-coupled H formic acid, indicated by a very small doublet at δ ϭ
1
NMR spectrum by the formato carbon atom, half of it is 8.60 ppm (JC,H ϭ 189 Hz) in the H NMR spectrum, was
1
unfortunately masked by the peaks of the Tp and PPh3 li- formed. After heating the tube at 100 °C for 3 h, H NMR
gands. The 13C{1H} NMR spectrum of the solution exhib- spectroscopy showed a very significant increase of the for-
its a singlet at δ ϭ 174.7 ppm, pertaining to the formato mic acid signal. In the downfield region, in addition to the
carbon atom, which is split into a doublet in the 1H-coupled set of peaks pertaining to the Tp ligand of 2, some minute
13C NMR spectrum with an H-C coupling constant of peaks were also visible. The 31P{1H} spectrum, however,
1
194.9 Hz. If the 31P, H, and 13C NMR spectroscopic data shows only one singlet due to 2. The CO2-hydrogenation
1
of 2 are taken together, it is obvious that the ligand set of reaction was basically complete within 3 h because the H
the complex consists of a Tp, a PPh3, a formate, and an NMR spectrum taken 3 h later did not show any visible
acetonitrile ligand. With reference to our recently reported increase in the formic acid signal. It is, however, worth not-
formate complex, TpRu(PPh3)(CH3CN)(η1-OCHO), gener- ing that, after heating the solution at 100 °C for 2 d, the
ated by CO2 insertion into the RuϪH bond of 1 in anhy- 1H NMR spectrum showed that the amount of formic acid
drous THF, and TpRu(PPh3)(CH3CN)(η1-OCHO)·H2O, decreased, while the downfield region of the spectrum be-
which is formed by the reaction of 1 with CO2 in hydrous came relatively complex, indicating the presence of a few
THF and stabilized by hydrogen-bonding interactions be- sets of Tp signals. A doublet of doublets (JH,P ϭ 26.4,
tween the formato ligand and a water molecule,[5] we pro- JC,H ϭ 9.3 Hz) is observed in the upfield region at δ ϭ
pose
that
the
molecular
formula
of
2
is Ϫ11.81 ppm, which can be assigned to the carbonyl hydride
Eur. J. Inorg. Chem. 2004, 1788Ϫ1793
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1789

S. M. Ng, C. Yin, C. H. Yeung, T. C. Chan, C. P. Lau
SHORT COMMUNICATION
complex TpRu(PPh3)(CO)H (3), which we have previously bonded intermediate Tp(PPh3)(CH3CN) RuϪH···HϪ
reported.[7] The 31P{1H} spectrum exhibits a major singlet OCH2CF3 and the η2-dihydrogen complex (see Scheme 1).
at δ ϭ 52.9 ppm, which is due to 2. A singlet at δ ϭ The relevancy of dihydrogen bonding to the protonation
66.1 ppm is assignable to 3, while the singlet at δ ϭ reaction is suggested by independent NMR studies involv-
30.9 ppm is identified as the signal of triphenylphosphane ing reactions of 1 with limited amounts of CF3CH2OH.
oxide, which was probably formed by CO2 oxidation of Upon addition of 1.5 and 7.5 equiv. of CF3CH2OH to
PPh3. A number of small, unidentified signals can be ob- [D8]toluene solutions of 1, the 1H NMR spectra of the solu-
served in the region between the signals of 2 and 3. After tions show that the hydride signal of 1 becomes broadened
heating at 100 °C for 3 d, a further decrease in the amount and shifted upfield by 0.16 and 0.63 ppm, respectively. Vari-
of formic acid was evidenced by 1H NMR spectroscopy and able-temperature T1 measurements for the hydrido ligand
a set of Tp signals of TpRu(PPh3)(CO)H (3) was observed of 1 yielded a T1(min) of 252 ms (at 400 MHz and 235 K)
in the downfield region. The hydride signal of 3 also shows and 65 ms (at 400 MHz and 233 K) for the 1/1.5 equiv.
enhanced intensity. The 31P{1H} NMR spectrum was also CF3CH2OH and the 1/7.5 equiv. CF3CH2OH solutions,
simplified, showing a major peak at δ ϭ 66.1 ppm, which respectively. The two T1(min) values are much lower than
is due to 3. The phosphane oxide peak at δ ϭ 30.9 ppm that of the hydrido ligand of 1 in [D8]toluene in the absence
and an unidentified peak at δ ϭ 63.3 ppm are the only two of CF3CH2OH (540 ms at 400 MHz and 257 K). These
minor signals in the spectrum. In our previous work, high- T1(min) measurements thus suggest the presence of
pressure NMR spectroscopy monitoring of the catalytic RuϪH···HϪOCH2CF3 dihydrogen-bonding interaction.
CO2 hydrogenation with 1 in hydrous THF revealed that Reactions of acidic alcohols with transition-metal hydride
the carbonyl hydride complex 3 was the major ruthenium- complexes to generate MϪH···HϪOR dihydrogen-bonded
containing species after prolonged heating at 100 °C. More- species are well documented.[11] For example, a dynamic
over, in the course of the reaction, a very small amount equilibrium between the dihydrogen-bonded ruthenium hy-
1
of TpRu(PPh3)(H2)H (4)[7] was also detected by H NMR dride
species
and
the
η2-dihydrogen
complex
spectroscopy. Independent NMR experiments showed that (dppm)2HRuϪH···HϪOR [(dppm)2HRu(H2)]ϩORϪ is es-
4 reacted slowly with formic acid in [D8]THF to yield tablished in the presence of phenol. The more acidic al-
TpRu(PPh3)(CO)H. We believe that in the course of the 1- cohol, hexafluoro-2-propanol, shifts the equilibrium to the
catalyzed CO2 hydrogenation in CH3OH/[D8]THF, the car- side of the η2-dihydrogen complex and eventually yields the
bonyl hydride complex 3 was formed by the reaction of alkoxide complex (dppm)2Ru[OCH(CF3)2]2.[11c]
TpRu(PPh3)(H2)H (4) with formic acid and the dihydrogen
hydride species was present in such low concentration that
it eluded detection by NMR spectroscopy. Examples rel-
evant to the formation of 3 can be found in the
literature.[8Ϫ10] An authentic sample of 3 was found to be
inactive in catalyzing CO2 hydrogenation in methanol.
Reaction of CF3CH2OH with 1
In view of the superiority of CF3CH2OH in promoting
the CO2-hydrogenation reaction, studies to gain insight into
1
its role in the catalytic reaction were carried out. The H
NMR spectrum taken a few minutes after the dissolution
of 1 in CF3CH2OH/[D8]THF (1:1 by volume) in a 5-mm
Wilmad pressure-valved NMR tube showed that the alkox-
ide complex TpRu(PPh3)(CH3CN)(OCH2CF3) (5) was
formed and 1 was no longer detectable. The spectrum shows
a new set of nine peaks pertaining to the Tp ligand. A sing-
let at δ ϭ 2.27 ppm integrated for 3 H is assignable to the
coordinated acetonitrile. Two multiplets, each integrated for
1 H, are observed at δ ϭ 3.04 and 3.45 ppm; these peaks
can be attributed to the methylene hydrogen atoms of the
coordinated ϪOCH2CF3 group. A singlet at δ ϭ 4.65 ppm
signals the presence of free hydrogen. The trifluoroalkoxo
ligand was verified by 19F NMR spectroscopy, which shows
the signal of ϪOCH2CF3 as a triplet at δ ϭ Ϫ75.05 ppm
(JH,F ϭ 10 Hz). The 31P{1H} NMR spectrum shows a sing-
let at δ ϭ 52.8 ppm.
Scheme 1
Attempts to synthesize 5 on a preparative scale, unfortu-
nately, were not successful, due to its decomposition into
some intractable materials during the workup processes. It
seems that 5 is only stable in CF3CH2OH solutions.
Reaction of TpRu(PPh3)(CH3CN)(OCH2CF3) (5) with
CO2
Et3N (20 equiv.) was added to a CF3CH2OH/[D8]THF
solution of 5; 31P{1H} NMR spectroscopy indicated that 5
was unaffected by the base. The solution was then subjected
to 8 atm of CO2 at room temperature; NMR spectroscopy
showed that 5 was converted into a 1:1 mixture of
Formation of 5 probably resulted from protonation of [TpRu(PPh3)(CH3CN)2]ϩ (6) and another complex 7 after
the hydrido ligand of 1 by CF3CH2OH via a dihydrogen- 2 h. Complex 6, which we have previously reported,[12] can
1790
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 1788Ϫ1793

Ruthenium-Catalyzed Hydrogenation of Carbon Dioxide
1
SHORT COMMUNICATION
be easily identified by H and 31P{1H} NMR spectroscopy. Mechanism of CO2 Hydrogenation in Non-Acidic Alcohols
1
The H NMR spectrum of the mixture clearly shows two
Our high-pressure NMR studies show that reactivity of
sets of signals (six peaks each) in the downfield region, one
1 in the CO2-hydrogenation reaction in methanol is very
similar to that in hydrous THF and it is therefore reason-
able to propose that the mechanism of 1-catalyzed CO2-
hydrogenation reactions in non-acidic alcohols (Entries
of which pertaining to the Tp ligand of complex 6, while
the other set belongs to the Tp ligand of 7. A six-peak pat-
tern for the Tp ligand implies that two of the three legs
trans to the Tp ligand in an octahedral configuration are
4Ϫ9, Table 1) is analogous to that of the catalysis in hy-
equivalent. The assumption that the two legs in 7 are due
Ϫ
drous THF. Therefore, in analogy to the active aquo hy-
dride species TpRu(PPh3)(H2O)H in hydrous THF, the key
intermediate in the catalytic cycle A (see Scheme 2) is the
transient alcohol hydride species 8, which is mainly formed
by displacement of the acetonitrile ligand of 1 by ROH.
The interaction of the OH proton of the alcohol in 8 with
one of the oxygen atoms of the approaching CO2 increases
the electrophilicity of the carbon atom, which is then able
to abstract the hydride from RuϪH. The turnover numbers
of the catalytic reactions decrease with increasing bulkiness
of the alcohols (Entries 3Ϫ7, Table 1) and this observation
supports the fact that the availability of 8 might have a large
effect on the overall efficiency of the catalytic system. It is
more difficult for a hindered alcohol to displace the aceto-
nitrile ligand of 1. In view of the detection of a small
amount of TpRu(PPh3)(H2)H (4) by NMR spectroscopy in
our previous study of 1-catalyzed CO2 hydrogenation in hy-
drous THF, we propose that 4 is present in a minute
amount in the alcohols, although it has not been success-
fully detected by NMR spectroscopy. Reaction of
to the bidentate alkyl carbonato ligand CF3CH2OCO2 is
supported by the 13C NMR study of an identical reaction
using 13CO2 in place of the unlabeled carbon dioxide. The
13C NMR spectrum of the resulting solution shows a
characteristic carbonate singlet resonance at
δ
ϭ
172.3 ppm. Since only one carbonate signal was detected in
the 13C NMR spectrum, the counterion of 6 is not a free
alkyl carbonato anion but it is the alkoxide anion
CF3CH2OϪ, which is confirmed by 19F NMR spectroscopy
showing a triplet signal at δ ϭ Ϫ76.02 (JH,F ϭ 10 Hz). The
methylene hydrogen signals of the CF3CH2OCO2 ligand are
unfortunately masked by a huge peak in the 1H NMR spec-
trum due to the free alcohol. The 31P{1H} spectrum of the
solution shows two singlets: one at δ ϭ 50.8 ppm pertaining
to 6 and another one at δ ϭ 63.9 ppm, ascribable to the
phosphane ligand of 7. When all the NMR spectroscopic
data are taken together, 7 can be identified as the alkyl car-
bonato complex TpRu(PPh3)(η2-O2COCH2CF3). Appar-
ently,
7
was formed by CO2 insertion into the
RuϪOCH2CF3 bond of 5 [see Equation (1)]. Insertion of
CO2 into metalϪoxygen bonds of metal alkoxides and
phenoxides to form metal carbonate species is well docu-
mented.[13]
TpRu(PPh3)(H2)H
with
formic
acid
to
yield
TpRu(PPh3)(CO)H was verified by an independent experi-
ment.
Reaction of 1 with CO2/H2 in CF3CH2OH/[D8]THF
A CF3CH2OH/[D8]THF (1:1 by volume) solution of 1
containing 20 equiv. of Et3N was pressurized with 6 atm of
CO2 and H2 at room temperature in a 5-mm Wilmad press-
ure-valved NMR tube. The 31P{1H} NMR spectrum taken
1 h later at room temperature indicates the formation of a
1
1:1 mixture of 6 and 7. H NMR spectroscopy shows, in
addition to 6 and 7, a small amount of formic acid, evi-
denced by a small singlet at δ ϭ 8.42 ppm. A few hours
1
later, H NMR measurements showed that the amount of
formic acid increased, and it also suggested the presence of
a minute amount of TpRu(PPh3)(CO)H (3), as evidenced
by a very small doublet at δ ϭ Ϫ11.83 ppm (JH,P
ϭ
26.8 Hz). After heating the NMR tube at 100 °C for 1 d, it
1
was found by H and 31P{1H} studies that 6 and 7 disap-
peared and 3 was the only ruthenium complex present in
the solution.
Scheme 2
Eur. J. Inorg. Chem. 2004, 1788Ϫ1793
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1791

S. M. Ng, C. Yin, C. H. Yeung, T. C. Chan, C. P. Lau
SHORT COMMUNICATION
Mechanism of CO2 Hydrogenation in Acidic Alcohol
CF3CH2OH
carbon atom of the molecule, making the hydride abstrac-
tion from RuϪH facile. Again, the final ruthenium complex
TpRu(PPh3)(CO)H results from the reaction of formic acid
with TpRu(PPh3)(H2)H (4) after prolonged heating. The ac-
tion of H2 on the alkyl carbonato intermediate 7 is believed
to be one of the routes to generate 4, although the sequence
of the reaction is unknown.
Unlike the CO2-hydrogenation reaction in methanol, no
metalloformate intermediate is observable in the course of
the catalysis in CF3CH2OH because complex 1 reacts rap-
idly with the acidic alcohol to form the metal alkoxide 5,
which then reacts with CO2 to give the observable inter-
mediate species [TpRu(PPh3)(CH3CN)2]ϩ (6) and
TpRu(PPh3)(η2-O2COCH2CF3) (7). Although the observ-
able intermediates in the catalyses in methanol and
CF3CH2OH are different, we propose that the sequence in
cycle B (see Scheme 3), which is basically identical to the
Conclusion
Continuing our study on the promoting effect of water in
catalytic cycle A of Scheme 2, is the productive route for the 1-catalyzed hydrogenation of CO2 to formic acid, we
the catalytic process in trifluoroethanol. Under H2 pressure, have studied the effect of alcohols on the catalysis. Our
a minute quantity of the active alcohol hydride species study suggests that, although the NMR-detectable inter-
TpRu(PPh3)(CF3CH2OH)H (9) might be generated mediates in the catalysis in methanol (and probably in other
through the alkoxo dihydrogen intermediate TpRu(P- non-acidic alcohols) and in acidic CF3CH2OH are differ-
Ph3)(OCH2CF3)(H2). The much higher catalytic activity of ent, the catalytic cycles of CO2 hydrogenation in both cases
1 in CF3CH2OH than in the non-acidic alcohols is probably are probably identical. The key species of the catalytic pro-
related to the much stronger interaction of the proton of cess, TpRu(PPh3)(ROH)H, is the alcohol analogue of the
the coordinated CF3CH2OH in 9 with an oxygen atom of aquo hydride complex that we have proposed for our pre-
CO2, which enhances significantly the positive charge at the viously studied hydrous THF system. The exceptionally
Scheme 3
1792
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 1788Ϫ1793

Ruthenium-Catalyzed Hydrogenation of Carbon Dioxide
SHORT COMMUNICATION
high promoting effect of CF3CH2OH among the alcohols
studied is due to the stronger interaction of the highly
acidic hydrogen of the trifluoroethanol in TpRu(PPh3)(CF-
3CH2OH)H with an oxygen atom of the approaching CO2,
greatly enhancing the electrophilicity of the carbon atom of
the molecule, and enabling its abstraction of the hydride
from RuϪH.
[1] [1a]
A. Behr, Carbon Dioxide Activation by Metal Complexes,
[1b]
VCH, Weinheim, Germany, 1988.
Enzymatic and Model
Carboxylation and Resuction Reactions for Carbon Dioxide Uti-
lization (Eds.: M. Aresta, J. V. Schloss), NATO ASI Series C,
No. 314, Kluwer Academic, Dordrecht, The Netherlands,
1990. [1c]P. G. Jessop, T. Ikariya, R. Noyori, Chem. Rev. 1995,
95, 259Ϫ272. [1d] W. Leitner, Angew. Chem. Int. Ed. Engl. 1995,
[1e]
34, 2207Ϫ2221.
247Ϫ247.
Y. Inoue, H. Izumida, Y. Sasaki, H. Hashimoto, Chem. Lett.
1976, 863Ϫ864.
J.-C. Tsai, K. M. Nicholas, J. Am. Chem. Soc. 1992, 114,
5117Ϫ5124.
P. G. Jessop, Y. Hsiao; T. Ikariya, R. Noyori, J. Am. Chem.
Soc. 1996, 118, 344Ϫ355.
W. Leitner, Coord. Chem. Rev. 1996, 153,
[2]
[3]
[4]
[5]
Experimental Section
General Remarks: Ruthenium trichloride, RuCl3·3H2O, pyrazole,
and sodium borohydride were obtained from Aldrich. Tri-
phenylphosphane was purchased from Merck and was recrys-
C. Yin, Z. Xu, S.-Y. Yang, S. M. Ng; K. Y. Wong, Z. Lin, C.
P. Lau, Organometallics 2001, 20, 1216Ϫ1222.
tallized
from
ethanol
before
use.
The
complex
[6] [6a]
P. Munshi, A. D. Main, J. C. Linehan, C.-C. Tai, P. G.
TpRu(PPh3)(CH3CN)H was synthesized according to published
procedures.[13] Solvents were distilled under dry nitrogen with ap-
propriate drying agents.[14] High-purity hydrogen gas and carbon
dioxide were supplied by Hong Kong Oxygen. 1H NMR spectra
were obtained with a Bruker DPX 400 spectrometer. Chemical
shifts are reported relative to residual protons of the deuterated
solvents. 31P, 19F, and 13C NMR spectra were recorded with a
Bruker DPX 400 spectrometer at 161.70, 376.31 and 100.60 MHz,
respectively. 31P NMR chemical shifts were externally referenced to
85 % H3PO4 in D2O. 19F NMR chemical shifts were externally
referenced to CFCl3 in CDCl3. 13C NMR chemical shifts were in-
ternally referenced to the residual peak of the deuterated solvents.
High-pressure NMR studies were carried out in commercial 5-mm
Wilmad pressure-valved NMR tubes.
[6b]
Jessop, J. Am. Chem. Soc. 2002, 124, 7963Ϫ7971.
C. A.
Thomas, R. J. Bonilla, Y. Huang, P. G. Jessop, Can. J. Chem.
2001, 79, 719Ϫ724.
Y.-Z. Chen, W. C. Chan, C. P. Lau, H. S. Chu, H. L. Lee, G.
Jia, Organometallics 1997, 16, 1241Ϫ1246.
[7]
[8]
M. A. McLoughlin, N. L. Keder, W. T. Harrison, R. J. Flesher,
H. A. Mayer, W. C. Kaska, Inorg. Chem. 1999, 38, 3223Ϫ3227.
C. S. Yi, N. Liu, Organometallics 1995, 14, 2616Ϫ2617.
H. Koinuma, Y. Yoshida, H. Hirai, Chem. Lett. 1975,
1223Ϫ1226.
[9]
[10]
[11] [11a]
E. S. Shubina, N. V. Belkova, A. A. Krylov, E. V. Vo-
rontsov, L. M. Epstein, D. G. Gusev, M. Niedermann, H.
[11b]
Berke, J. Am. Chem. Soc. 1996, 118, 1105Ϫ1112.
Belkova, E. S. Shubina, A. V. Ionidis, L. M. Epstein, H. Ja-
N. V.
cobsen, A. Messmer, H. Berke, Inorg. Chem. 1997, 36,
[11c]
´
J. A. Ayllon, C. Gervaux, S. Sabo-Etienne, B.
1522Ϫ1525.
Safety Warning: It is advisable not to add more than 12 atm of
gas(es) to the 5-mm Wilmad pressure-valved NMR tubes at room
temperature.
Chaudret, Organometallics 1997, 16, 2000Ϫ2002. [11d] Y. Guari,
´
J. A. Allyon, S. Sabo-Etienne, B. Chaudret, Inorg. Chem. 1998,
37, 640Ϫ644.
[11e]
S. Gründemann, S. Ulrich, H.-H. Limbach,
N. S. Golubev, G. S. Denisov, L. M. Epstein, S. Sato-Etienne,
B. Chaudret, Inorg. Chem. 1999, 38, 2550Ϫ2551. [11f] E. I. Gut-
sul, N. V. Belkova, M. S. Sverdlov, L. M. Epstgein, E. S. Shub-
ina, V. I. Bakhmutov, T. N. Gribanova, R. M. Minyaev, C.
Bianchini, M. Peruzzini, F. Zanobini, Chem. Eur. J. 2003, 9,
2219Ϫ2228.
Catalytic Hydrogenation of CO2 with TpRu(PPh3)(CH3CN)H (1):
The reactions were performed in a 100-mL stainless steel autoclave.
In a typical run, about 0.013 mmol of 1 in 20 mL of alcohol and
2 mL of NEt3 was added to the autoclave. After being flushed with
H2 three times, the system was heated whilst stirring at 100 °C
under 50 atm of CO2/H2 (1:1). At the end of the required length
of time, the autoclave was cooled rapidly and vented carefully. The
formic acid in the resulting solution was analyzed by 1H NMR
spectroscopy. The relative intensities of the formato hydrogen sig-
nal of the product and the formato hydrogen signal of the internal
standard (DMF) were measured, and with the help of a calibration
curve, the turnover number of the catalysis was derived.
[12]
W. C. Chan, C. P. Lau, Y.-Z. Chen, Y.-Q. Fang, S. M. Ng, G.
Jia, Organometallics 1997, 16, 34Ϫ44.
[13] [13a]
H. S. Kim, J. J. Kim, B. G. Lee, O. S. Jung, H. G. Jang, S.
[13b]
O. Kang, Angew. Chem. Int. Ed. 2000, 39, 4096Ϫ4098.
T.
Tsuda, S. I. Sanada, K. Ueda, T. Saegusa, Inorg. Chem. 1976,
[13c]
15, 2329Ϫ2332.
D. J. Darensbourg, M. W. Holtcamp, G.
E. Struck, M. S. Zimmer, S. A. Niezgoda, P. Rainey, J. B. Rob-
ertson, J. D. Draper, J. H. Reibenspies, J. Am. Chem. Soc. 1999,
[13d]
121, 107Ϫ116.
D. J. Darensbourg, K. M. Sanchez, J. H.
Reibenspies, A. L. Rheingold, J. Am. Chem. Soc. 1989, 111,
[13e]
7094Ϫ7103.
D. J. Darensbourg, K. M. Sanchez, A. L.
Rheingold, J. Am. Chem. Soc. 1987, 109, 290Ϫ292.
D. R. Perrin, W. L. Armanego, Purification of Laboratory
Chemicals, 2nd ed., Pergmon, Oxford, U.K., 1980.
Received November 18, 2003
Acknowledgments
We thank the Hong Kong Research Grant Council (Project No.
PolyU5188/00P) for financial support.
[14]
Eur. J. Inorg. Chem. 2004, 1788Ϫ1793
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1793