Direct detection of key reaction intermediates in photochemical CO 2 reduction sensitized by a rhenium bipyridine complex

Article abstract of DOI:10.1021/ja500403e

Photochemical CO₂ reduction sensitized by rhenium-bipyridyl complexes has been studied through multiple approaches during the past several decades. However, a key reaction intermediate, the CO₂-coordinated Re-bipyridyl complex, which should govern the activity of CO₂ reduction in the photocatalytic cycle, has never been detected in a direct way. In this study on photoreduction of CO₂ catalyzed by the 4,4′-dimethyl-2,2′-bipyridine (dmbpy) complex, [Re(dmbpy)(CO) ₃Cl] (1), we successfully detect the solvent-coordinated Re complex [Re(dmbpy)(CO)₃DMF] (2) as the light-absorbing species to drive photoreduction of CO₂. The key intermediate, the CO ₂-coordinated Re-bipyridyl complex, [Re(dmbpy)(CO) ₃(COOH)], is also successfully detected for the first time by means of cold-spray ionization spectrometry (CSI-MS). Mass spectra for a reaction mixture with isotopically labeled ¹³CO₂ provide clear evidence for the incorporation of CO₂ into the Re-bipyridyl complex. It is revealed that the starting chloride complex 1 was rapidly transformed into the DMF-coordinated Re complex 2 through the initial cycle of photoreduction of CO₂. The observed induction period in the time profile of the CSI-MS signals can well explain the subsequent formation of the CO₂- coordinated intermediate from the solvent-coordinated Re-bipyridyl complex. An FTIR study of the reaction mixture in dimethyl sulfoxide clearly shows the appearance of a signal at 1682 cm^-1, which shifts to 1647 cm ^-1 for the ¹³CO₂-labeled counterpart; this is assigned as the CO₂-coordinated intermediate, Re^II-COOH. Thus, a detailed understanding has now been obtained for the mechanism of the archetypical photochemical CO₂ reduction sensitized by a Re-bipyridyl complex.

Full text of DOI:10.1021/ja500403e

Article
pubs.acs.org/JACS
Direct Detection of Key Reaction Intermediates in Photochemical
CO2 Reduction Sensitized by a Rhenium Bipyridine Complex
Youki Kou,† Yu Nabetani,†,‡ Dai Masui,†,§ Tetsuya Shimada,†,‡ Shinsuke Takagi,†,‡ Hiroshi Tachibana,†,‡
and Haruo Inoue*,†,‡
†Department of Applied Chemistry, Graduate School of Urban Environmental Sciences, Tokyo Metropolitan University, 1-1
Minami-osawa, Hachioji, Tokyo 192-0397, Japan
‡Center for Artificial Photosynthesis, Tokyo Metropolitan University, 1-1 Minami-osawa, Hachioji, Tokyo 192-0397, Japan
S
* Supporting Information
ABSTRACT: Photochemical CO2 reduction sensitized by
rhenium−bipyridyl complexes has been studied through
multiple approaches during the past several decades. However,
a key reaction intermediate, the CO2-coordinated Re−
bipyridyl complex, which should govern the activity of CO2
reduction in the photocatalytic cycle, has never been detected
in a direct way. In this study on photoreduction of CO2
catalyzed by the 4,4′-dimethyl-2,2′-bipyridine (dmbpy)
complex, [Re(dmbpy)(CO)3Cl] (1), we successfully detect
the solvent-coordinated Re complex [Re(dmbpy)(CO)3DMF]
(2) as the light-absorbing species to drive photoreduction of
CO2. The key intermediate, the CO2-coordinated Re−
bipyridyl complex, [Re(dmbpy)(CO)3(COOH)], is also
successfully detected for the first time by means of cold-spray ionization spectrometry (CSI-MS). Mass spectra for a reaction
mixture with isotopically labeled 13CO2 provide clear evidence for the incorporation of CO2 into the Re−bipyridyl complex. It is
revealed that the starting chloride complex 1 was rapidly transformed into the DMF-coordinated Re complex 2 through the
initial cycle of photoreduction of CO2. The observed induction period in the time profile of the CSI-MS signals can well explain
the subsequent formation of the CO2-coordinated intermediate from the solvent-coordinated Re−bipyridyl complex. An FTIR
study of the reaction mixture in dimethyl sulfoxide clearly shows the appearance of a signal at 1682 cm−1, which shifts to 1647
cm−1 for the 13CO2-labeled counterpart; this is assigned as the CO2-coordinated intermediate, ReII−COOH. Thus, a detailed
understanding has now been obtained for the mechanism of the archetypical photochemical CO2 reduction sensitized by a Re−
bipyridyl complex.
had led to intense studies on multiple aspects of the
photocatalysis, e.g., synthesis, electrochemical properties,
photochemical reactivity, product analysis, reaction dynamics,
and so on, during the past several decades.9−11 For example,
Hori et al. reported that the reactivity of photochemical CO2
reduction could be dramatically improved with the fac-
[Re(bpy)(CO)3(P(OEt)3)]+ complex, replacing the chloride
ligand with a triethylphosphite (quantum yield ΦCO = 38%).10
In recent years, the efficiency of CO2 reduction by Re−
bipyridyl complexes was amazingly improved by Ishitani’s
group under UV light irradiation,11 even though the electron
donors are still limited to sacrificial reagents such as amines. At
the present stage, how the photoreduction could be induced by
visible light and how the photoreduction could be coupled with
the oxidation of the water molecule should be central subjects
to be resolved from the viewpoint of artificial photosynthesis.
As such, the Ishitani, Inoue, and Perutz groups each reported
INTRODUCTION
■
Carbon dioxide (CO2), well known as a greenhouse gas, which
causes global warming, is also an abundant carbon resource for
fuels and organic materials.1 The development of an effective
method for the fixation of atmospheric CO2 is a global issue to
be achieved for mankind from the viewpoint of realizing a
sustainable society. Artificial photosynthesis, in which CO2 is
reduced with water as an electron donor, is very much
anticipated to be one of the most crucial key technologies to
sustain our society.2
Generally, it is very difficult to directly induce the one-
electron reduction of CO2 to its radical anion, CO2•−, because
the reduction potential is highly unfavorable (−2.14 V vs
SCE).1 Among various studies, including the photoelectro-
chemical reduction of CO2 by p-type semiconductors such as p-
GaP, p-CdTe, p-GaAs, p-Si, p-SiC, p-InP, etc.,3 electrochemical
reduction,4,5 and photochemical reduction,6 a seminal, pioneer-
ing report by Lehn et al. on the rhenium−bipyridyl complex
[Re(bpy)(CO)3X],7,8 which can efficiently induce CO2
reduction to carbon monoxide (CO) upon UV-light irradiation,
Received: January 14, 2014
Published: April 1, 2014
© 2014 American Chemical Society
6021
dx.doi.org/10.1021/ja500403e | J. Am. Chem. Soc. 2014, 136, 6021−6030

Journal of the American Chemical Society
Article
visible-light-induced CO2 reduction, but still with sacrificial
electron donors.12−19 Recently, we reported a visible-light-
induced photoelectrochemical CO2 reduction sensitized by a
rhenium bipyridine/zinc porphyrin dyad on NiO as a non-
sacrificial electron donor.9,17 Even though photoreduction of
CO2 by the dyad provided a novel reaction mechanism through
the upper excited S2 level, it is still necessary to improve the
efficiency of CO2 reduction from a practical viewpoint.
Photochemical CO2 reduction catalyzed by metal complexes
on p-type semiconductors has been reported.20−23 The
formation of formic acid from CO2 and water has also been
reported.23 This system, however, was obliged to use UV light
irradiation of a TiO2 semiconductor to oxidize water.23 The
state of the art in photochemical reduction of carbon dioxide
clearly requires more fundamental studies to get through
various bottlenecks in order to achieve more realistic prospects
for artificial photosynthesis.1,24−28
Scheme 1. Proposed Photochemical CO2 Reduction Cycle
Sensitized by a Rhenium−Bipyridyl Complex
The Re complex has served as one of the benchmark
complexes for photoreduction of CO2, but the reaction
mechanisms are not yet fully understood, despite extensive
studies. Deeper insights into the reaction mechanism, and into
the nature of the key factors controlling the photochemical
reduction of CO2, would be most desirable.
The key intermediate of the Re−bipyridyl complex, after
reaction with CO2 in the reaction pathway, is successfully
observed by positive-ion mode CSI-MS for the first time. An
experiment involving the photochemical reaction with isotopi-
cally labeled 13CO2 also provides concrete evidence for CO2
coordination to the Re−bipyridyl complex in the reaction
pathway. Furthermore, the time-dependent profiles of the
signals are well correlated with the time profile of CO
formation in the photochemical CO2 reduction. An FTIR study
on steady-state light irradiation clearly indicates that the CO2-
coordinated species is ReII−COOH. Thus, the mechanism of
photochemical CO2 reduction sensitized by the Re−bipyridyl
complex is explained in detail.
The mechanism of the photochemical CO2 reduction
sensitized by Re−bipyridyl complexes has been studied so far
by UV−visible absorption spectroscopy,11,29−31 electrospray
ionization mass spectrometry (ESI-MS),32,33 liquid chromatog-
raphy (LC) coupled with ESI-MS (LC-ESI-MS),34 and IR
spectroscopy.35,36 In particular, the Fujita group reported a
transient FTIR study of the Re−bipyridyl complex.35 They
proposed that a CO2-mediated dimer of Re−bipyridyl
complexes could be a key intermediate for CO2 reduction in
acetonitrile (CH3CN) or tetrahydrofuran (THF), although its
reactivity was not mentioned in detail.27,35 Regrettably, CH3CN
and THF are both rather exceptional solvents, in which
photoreduction of CO2 has only a poor reactivity, compared
with a frequently used solvent such as dimethylformamide
(DMF). The formation of a formate Re complex as a minor
side reaction product was studied by NMR8 in the photo-
chemical reduction and by IR spectroscopy in the electro-
chemical reduction.37 A one-electron-reduced species of the Re
complex was also studied by electron paramagnetic resonance
(EPR) spectroscopy.38,39 On the basis of the previous studies, a
generally proposed photochemical CO2 reduction cycle of
[Re(bpy)(CO)3Cl] is summarized in Scheme 1. In a solution of
[Re(bpy)(CO)3Cl] with triethylamine as a sacrificial electron
donor, a one-electron-reduced species is formed through the
triplet metal-to-ligand charge-transfer (3MLCT) state upon
photoexcitation.29−31 After elimination of the Cl− ion, CO2
may be reduced to CO by the activated Re−bipyridyl complex.
The details of the reaction mechanism have not yet been
clarified, since the direct detection of the intermediate, which
could rationalize the observed photoreduction of CO2, has not
yet been achieved.
EXPERIMENTAL SECTION
■
Materials. [Re(dmbpy)(CO)3Cl] (1) was synthesized according to
the literature.42 N,N′-Dimethylformamide (DMF, Kanto Chemical
Co., Inc.) was used as a solvent after dehydration with 4 Å molecular
sieves (Kanto Chemical Co., Inc.) and distillation under reduced
pressure. Triethylamine (TEA, Kanto Chemical Co., Inc.) was used as
a sacrificial electron donor in the photochemical reaction after
dehydration with potassium hydroxide (Kanto Chemical Co., Inc.) and
distillation. Highly pure CO2 (>99.995%, Taiyo Nippon Sanso Co.,
Inc.) was used, and isotopically labeled 13CO2 (chemical purity
specification >99.9%) was purchased from Cambridge Isotope
Laboratories Inc. and used as received.
Photocatalytic Reaction. 1 (0.94 mg, 0.4 mM) was dissolved in
4.8 mL of DMF/TEA solvent (v/v = 5/1). CO2 bubbling was carried
out for 30 min before photoirradiation. The samples in an L-type
quartz cell (optical length, 4 cm, Eikosha) were irradiated with
monochromatic light (365 nm) through an interference filter (KL-
365) and an IR cut-off filter (KG1(EDMOND)) from a 500-W xenon
arc lamp (USHIO 500-DKO). The amount of reduced product, CO,
was measured by gas chromatography with a thermal conductivity
detector (GC-TCD, Shimadzu GC-2014, equipped with a 5 Å
molecular sieves column).
CSI-MS Experiment. CSI-MS was performed on a JMS-T100CS
instrument (JEOL) for analyzing the reaction intermediates. Typical
measurement conditions were as follows: needle voltage, 0 kV; orifice
1 voltage, −90 V; orifice 2 voltage, −10 V; sample flow rate, 0.1 mL/
min; spray temperature, −10 °C; ion source temperature, −5 °C. Each
sample, upon photoirradiation (0, 30, 180, 270, and 1000 min), was
independently prepared and directly injected into the ion chamber of
the CSI mass spectrometer with a syringe.
Against this background, herein we report the direct
detection of the key reaction intermediate in the photochemical
CO2 reduction with the 4,4′-dimethyl-2,2′-bipyridyl (dmbpy)
complex, [Re(dmbpy)(CO)3Cl] (1), by a cold spray ionization
mass spectrometry (CSI-MS) method and FTIR studies. CSI-
MS enables an extremely soft ionization40,41 and is suitable for
detecting rather weakly bound association complexes in
solution, which can easily decompose in the ionization chamber
with ordinary ESI-MS under more severe ionization conditions.
FTIR Experiment. The FTIR spectrum of the reaction mixture in
dimethyl sulfoxide (DMSO) under steady-state light irradiation was
measured by IR monitoring in rapid scan mode on a Shimadzu
IRTracer-100.
6022
dx.doi.org/10.1021/ja500403e | J. Am. Chem. Soc. 2014, 136, 6021−6030

Journal of the American Chemical Society
Article
RESULTS AND DISCUSSION
Time Profile of CO Evolution upon UV Light
Irradiation. Figure 1 shows a time profile of CO evolution
■
Figure 2. Positive-ion mode CSI-MS spectra of the reaction mixture
upon (i) 0, (ii) 30, (iii) 180, (iv) 270, and (v) 1000 min of
▲
photoirradiation. The mass peak at m/z = 528 ( ) is assigned as the
solvent-coordinated species, [Re(dmbpy)(CO)3DMF] (2); m/z = 572
■
( ) as the CO2 adduct, [Re(dmbpy)(CO)3(CO2) + DMF]; and m/z
●
= 592 ( ) as [Re(dmbpy)(CO)3Cl + TEA + H].
Figure 1. CO evolution catalyzed by [Re(dmbpy)(CO)3Cl] (1) in
DMF solvent with TEA (v/v = 5/1) as a sacrificial electron donor
upon UV light irradiation (λ = 365 nm, 2 mW).
dissociation of the Cl ligand after photoexcitation (Figure 2 (i),
(ii)). The solvent-coordinated species 2 (m/z = 528) exhibited
an interesting rise-and-decay profile during the photoirradiation
(Figure 2 (iii)−(v); see also Figure 7b). The peak intensity
gradually rose to a maximum around 270 min of photo-
irradiation. After 1000 min of photoirradiation, the intensity of
the intermediate decreased gradually.
catalyzed by [Re(dmbpy)(CO)3Cl] (1) in DMF including TEA
as a sacrificial electron donor upon UV light irradiation (λ =
365 nm). The CO evolution increased moderately in the initial
period, suggestive of an induction period within the initial 30
min. In the case of photoinduced CO2 reduction catalyzed by
[Re(dmbpy)(CO)3(NCS)], a similar induction period was also
reported to be caused by the dissociation of the NCS ligand in
the early stage of the reaction.43 After 30 min, the CO evolution
was strongly accelerated, with a ΦCO value of 10.5%, and then it
decreased gradually after 500 min, probably owing to
transformation or decomposition of the Re complex under
the prolonged photoirradiation.
CSI-MS Measurement of the Reaction Mixture during
Photoirradiation. To detect reaction intermediates, the
reaction mixtures at appropriate times (0, 30, 180, 270, and
1000 min) covering the induction period, the most reactive
stage, and the saturated stage were analyzed by positive-ion
mode CSI-MS. The spectra obtained by CSI-MS measurement
are shown in Figure 2. In the reaction mixture before the
photoirradiation (Figure 2 (i), 0 min), the starting Re complex
1 exhibited mass peaks at m/z = 592, assigned as [Re(dmbpy)-
(CO)3Cl + TEA + H+], and at m/z = 608, assigned as
[Re(dmbpy)(CO)3Cl + CO2 + DMF + H+]. The mass patterns
were well simulated by characteristic inclusion of one chlorine
atom in each case (Figure S1). In contrast, ordinary positive-ion
mode ESI-MS did not exhibit a peak from the starting Re
complex in solution, even before the photoirradiation (Figure
S2). Only a mass peak at m/z = 528, assigned as
[Re(dmbpy)(CO)3 + DMF], was detected by ESI-MS. The
chloride ion is clearly removed during the ionization process,
and the soft ionization of the CSI-MS enables detection of the
Re complex as it is, without removal of the chloride ion. The
peaks of the starting complex 1 decreased dramatically within
30 min of the photoirradiation, replaced by appearance of the
peak assigned to the solvent-coordinated intermediate, [Re-
(dmbpy)(CO)3DMF] (2, m/z = 528), which is formed by
Among the signals observed in the positive-ion mode CSI-
MS, the peaks at m/z = 403, 415, 427, 442, 455, 476, 500, 515,
528, 556, 570−574, and 616 might be derived from the Re−
bipyridyl complexes including DMF or CO2 as ligands, because
the peak intensities also exhibited a rise-and-decay profile, being
highest at the most reactive period. However, the peaks could
not be characterized in sufficient detail, except those at m/z =
570−574, which will be analyzed in detail later. The other mass
peaks with much smaller intensities, observed at m/z = 556,
584, and 604, showed a different time profile, with a monotonic
increase throughout the photoirradiation. The corresponding
species were clearly accumulating in the reaction mixture.
Those peaks also were not able to be characterized well but
might be ascribed to other Re complexes transformed from the
starting complex 1 or decomposed as side products during the
photoreaction. In summary, the observed mass peaks are
classified into three groups: group I, which exhibited a
monotonic decrease of their intensities (m/z = 592 and 608,
assigned as the starting Re complex 1); group II, which
exhibited a rise-and-decay profile (m/z = 403, 415, 427, 442,
455, 476, 500, 515, 528 (assigned as solvent-coordinated
species), 570−574 (assigned as CO2-cooordinated species, as
analyzed below), and 616); and group III, which exhibited a
monotonic increase but with much smaller intensities (m/z =
556, 584, and 604) (Table S1).
Assignment of the Peaks at m/z = 570−574. Very
interestingly, among the signals in group II, we have
successfully observed the CO2 adducts of the Re−bipyridyl
complex as the key intermediates in the reaction mixture
around m/z = 570−574 in the positive-ion mode CSI-MS. The
observed peaks were carefully assigned as those of [Re-
6023
dx.doi.org/10.1021/ja500403e | J. Am. Chem. Soc. 2014, 136, 6021−6030

Journal of the American Chemical Society
Article
(dmbpy)(CO)3CO2 + DMF] (3, m/z = 570, 572) and
[Re(dmbpy)(CO)3CO2H + DMF] (3′, m/z = 571, 573)
(Figure S3). The isotope distribution caused by 185Re (37.4%)
and 187Re (62.6%) afforded two main peaks for a compound
containing one Re atom, with subpeaks due to natural-
abundance 13C in the ligands within the complex. The patterns
around m/z = 570−574 showed four main peaks with subpeaks,
indicating two species containing one Re atom in each being
superimposed in this region, as shown in Figures 2 and 3
ReI−carboxylic acid complex 5 according to the literature.47
When 5 was employed as the photocatalyst, the ΦCO value was
again observed to be very low, around 0.6%. This indicates that
5 is also not the active species that drives the photocatalytic
reaction, though 5 was reported to form CO and bicarbonate
Re complex in the dark.48 The CSI-MS measurement of 5,
however, afforded a spectral pattern at m/z = 571, 573 fairly
similar to that shown in Figure S4, while smaller peaks were
observed at m/z = 570, 572 that are in different peak patterns
from those of the reaction mixture. This indicates that the ReI−
carboxylic acid complex 5 may be involved in the reaction
mixture and partly contributes to the spectra at m/z = 570−
574, but another species, which shows stronger peaks at m/z =
570, 572, is definitely formed during photoreduction of CO2
and is actually detected in the reaction mixture and assigned as
the CO2 adduct of the Re complex. It should be noted here that
the details of the soft ionization in the chamber are not yet
completely understood; in many cases, a molecule is observed
as an ion, as it is, as an ion plus solvent, as a protonated form,
and as a protonated form plus solvent, irrespective of the total
charges on the original molecules in the positive-ion mode, as
reported above in the case of the starting Re complex
([Re(dmbpy)(CO)3Cl + TEA + H+], m/z = 592).
The time course of the mass patterns assigned to the CO2-
coordinated intermediate of Re−bipyridyl complex (m/z =
570−574) upon photoirradiation is shown in Figure 3. After 30
min of photoirradiation (ii), the mass peaks at m/z = 570−574
appeared, where no peak was detected before the photo-
irradiation (i). The peak intensities at m/z = 570−574 also
exhibited a rise-and-decay profile, without changing the peak
patterns (Figure 3). These results strongly suggest that the
signals at m/z = 570−574 are derived from the same species. As
will be described later, the FTIR study clearly indicated the
ReII−COOH species, which was well correlated with the CSI-
MS experiments. The species with peaks at m/z = 570−574 is
thus assigned as [ReII(dmbpy)(CO)3COOH + DMF] (m/z =
571, 573) and its deprotonated form, [ReII(dmbpy)-
(CO)3(CO2) + DMF] (m/z = 570, 572).
Figure 3. Comparison of the mass patterns assigned to a CO2-
coordinated intermediate at each photoirradiation time: (i) 0, (ii) 30,
(iii) 180, (iv) 270, and (v) 1000 min.
without chlorine. It has been reported that the formate complex
with oxygen atom coordination to Re, [Re(dmbpy)-
(CO)3(OCHO)] (4), was formed as a minor reaction
product.44−46 The formate complex 4 should have exactly the
same mass number of m/z = 571, 573; thus, the observed peaks
at m/z = 571, 573 may be ascribed to 4. To check this
possibility, we have synthesized the formate complex 4
according to the literature37 and measured its CSI-MS signals
(Figure S4). In fact, the mass pattern of 4 showed the peaks at
m/z = 571, 573 with weak intensities, along with much more
intense peaks, by a factor of ca. 7, at m/z = 600, 602. The
intensity pattern, however, is very much different from that
observed for the reaction mixture, where the peaks at m/z =
571, 573 had, in contrast, stronger intensity, by a factor of ca. 2,
than those at m/z = 600, 602. Judging from the peak intensity
at m/z = 600, 602 in the reaction mixture, the contribution of 4,
thus, should be smaller than 8% (1/2 × 1/7 = 1/14) compared
to the peaks at m/z = 571, 573, even when 4 is involved in the
reaction mixture. Moreover, the formate complex 4 had no
peak at m/z = 570, whereas the reaction mixture definitely had
peaks at m/z = 570, 572. These results strongly suggest that the
peaks around m/z = 570−574 are derived mostly from a
species different from 4. Actually, when 4 was used as the
starting photocatalyst for UV light irradiation, the quantum
yield of CO evolution was very modest, only ∼2.1%, indicating
that 4 should not be the active catalyst in the current system.
Another probable candidate for the peaks at m/z = 571, 573
would be the Re−carboxylic acid complex, [Re(dmbpy)-
(CO)3(COOH)] (5). Here, we have also synthesized the
13C Isotope Effect on the CSI Mass Spectra. To obtain a
deeper insight into the spectral analysis of the CSI-MS
experiment, a photochemical reaction of CO2 reduction
sensitized by [Re(dmbpy)(CO)3Cl] (1) was carried out with
isotope-labeled CO2 (13CO2) under the same reaction
conditions as above. The mass peaks of the reaction mixture
with 13CO2 were also detected in positive-ion mode CSI-MS.
As a typical example, the mass spectra of the reaction mixture
upon 180 min of photoirradiation in both 12CO2 and 13CO2 are
compared in Figure 4. Very interestingly, among the group II
Figure 4. Positive-ion mode CSI-MS spectra of reaction mixtures
under 12CO2 and 13CO2 atmosphere upon 180 min of photo-
irradiation, respectively.
6024
dx.doi.org/10.1021/ja500403e | J. Am. Chem. Soc. 2014, 136, 6021−6030

Journal of the American Chemical Society
Article
mass peaks in the case of the 12CO2 experiment, as classified
above, which exhibited a rise-and-decay profile (m/z = 403,
415, 427, 443, 454, 476, 500, 528, 556, 570−574, 604, 616),
only three peaks, m/z = 528 (solvent-coordinated species 2),
570−574 (CO2 adduct of Re complex, 3/3′), and 616 (not well
characterized), showed distinct shifts of their mass numbers,
strongly suggesting that these are derived from the actual
intermediates in the reaction pathway of CO2 reduction
photocatalyzed by the Re complex. Figure 5 shows how the
explained as further 13CO2 incorporation as the sixth ligand to
form the composition [Re(dmbpy)(13CO)313COOH + DMF].
Both peak shifts (from m/z = 572 to 576, and from m/z = 528
to 531) in the CSI-MS experiment proceeded gradually with
photoirradiation, as seen in Figure 5, indicating that the peak
shifts originate from a mixture of singly 13C-incorporated
species up to the fully incorporated species. Their compositions
among the non-13C-incorporated species and each 13C species
(one-13C species, two-13C species, three-13C species for the
DMF-coordinated Re complex, and one-13C, two-13C,
three-13C, four-13C species for the CO2 adduct of the Re
complex) were carefully estimated by subtracting the
contribution of each 13C-incorporated species stepwise from
the original spectra.
The detailed procedure for the analysis is described in the
Supporting Information (Figure S5). The relative ratios among
the 13C species at each photoirradiation time are summarized in
Figure 6. The incorporation of 13C into the CO ligand indicates
Figure 5. Mass pattern variation for (a) the DMF-coordinated and (b)
13CO2-coordinated intermediates in the photochemical reaction cycle.
The upper spectra are for 12CO2, and the lower three are for 13CO2.
The arrows indicate the mass peak shifts upon photoirradiation.
mass peaks around m/z = 528 (solvent-coordinated species 2)
and 570−574 (CO2 adduct of Re complex 3/3′) change in the
case of 13CO2 when compared with those for 12CO2.
As clearly seen in Figure 5, both peaks shifted to larger mass
number, and the degree of the shifts became more pronounced
during the photoirradiation. A careful inspection of Figure 5a
indicates that, of the two main peaks (m/z = 526, 528), the
intensity of the one at m/z = 526 gradually decreased and
disappeared after 1000 min of photoirradiation, when the other
strongest peak, m/z = 528, reached m/z = 531. This indicates
that the solvent-coordinated Re complex clearly had an uptake
of 13C during the photoreaction, and the degree of 13C uptake
increased upon prolonged photoirradiation. The three mass
unit shift (+3) of the main peak (from m/z = 528 to 531)
indicates that three 13C atoms were incorporated in the DMF-
coordinated Re complex. This result coincides well with the
previous NMR study, which showed that all three of the CO
ligands of the Re complex were replaced with 13CO during the
photoreaction under 13CO2 conditions.8 The other peaks
around m/z = 570−574 exhibited a more interesting feature:
the most intense peak at m/z = 572 shifted to 576 after 1000
min photoirradiation (Figure 5b). The four mass unit shift (+4)
Figure 6. 13C uptake into the key intermediates, (a) the DMF-
coordinated Re complex 2 and (b) the CO2 adduct of the Re complex
3, upon photoirradiation. The turnover numbers are 1 (t = 30 min),
2.5 (t = 180 min), and 7 (t = 1000 min).
that the reduced product 13CO generated from the CO2 adduct
of the Re complex replaces one of the three other CO ligands
within the complex through axial−equatorial isomerization, or
the 13CO liberated into solution simply attacks the complex
again to replace one of the three CO ligands. Since the
solubility of CO in DMF is nearly negligible, this extremely low
CO concentration may exclude the latter possibility.
An illustration of the ligand replacement with 13CO2 in the
CO2 reduction cycle is shown in Scheme 2. In the CO2
reduction cycle, a tetracarbonyl complex, 6, would be
transiently formed, although it was not detected by the CSI-
MS experiment. One of the four CO ligands present just after
the reaction of CO2 reduction may be immediately eliminated
from the Re complex, followed by axial−equatorial isomer-
ization. The 13C uptake distributions among the two
intermediates, the DMF-coordinated species [Re(dmbpy)-
(CO)3DMF] (2) and the CO2 adduct [Re(dmbpy)-
(CO)3(COO)] (3), provide very interesting mechanistic
clearly indicates an additional 13C uptake, in addition to the 13
C
uptake into the three CO ligands within the Re complex. This
should be the most convincing evidence for the assignment of
the peaks at m/z = 570−574 to the CO2 adduct with the Re
complex, since the uptake of four 13C atoms can only be
6025
dx.doi.org/10.1021/ja500403e | J. Am. Chem. Soc. 2014, 136, 6021−6030

Journal of the American Chemical Society
Article
Scheme 2. 13C Uptake through Stepwise Replacement of CO
Ligands of Re−Bipyridyl Complexes and Addition of CO2 in
the Photochemical Reduction Cycle
Time Course of the Peak Intensity of Each Inter-
mediate. As summarized above, the CSI-MS measurement of
the reaction mixture afforded three kinds of peak signals. Group
I exhibited a monotonic decay, and group II exhibited a rise and
decay, while group III exhibited a monotonic increase with
minor intensity. Since group III clearly is associated with side
reactions such as decomposition of the Re complex on
prolonged light irradiation, attention was focused on groups I
and II, especially on the specific signal peaks, m/z = 592
(starting Re complex, [Re(dmbpy)(CO)3Cl], 1) in group I,
and m/z = 528 (solvent-coordinated Re complex, [Re(dmbpy)-
(CO)3DMF], 2) and m/z = 570, 572 (CO2 adduct of the Re
complex [Re(dmbpy)(CO)3(CO2) + DMF, 3) in group II. The
time course of the mass peaks m/z = 592, 528, and 572, which
can be assigned as the key intermediates, is shown in Figure 7,
information. As shown in Figure 6, the relative distributions of
13C uptakes of n = 1, 2, 3 in 2 are very similar to those of n = 2,
3, 4 for 3 at each time of photoirradiation, strongly suggesting
that CO2 simply adds to the DMF-coordinated complex 2 to
form a species with an additional 13C, the CO2 adduct 3. The
DMF-coordinated Re complex 2 is thus supposed to be the
actual precursor of the CO2 adduct 3 in the reaction pathway,
as discussed in the reaction mechanism shown below in Scheme
3. After CO2 incorporation within the Re complex, the CO2
adduct may have two options when it accepts the second
electron to form CO: (1) replacing one of the tricarbonyls
(CO)3 within the Re complex to induce CO liberation from the
tricarbonyl (CO)3 or (2) escaping from the Re complex into
solution, without the replacement as the evolving CO. In either
case, CO is liberated into the gas phase over the reaction
mixture, since the solubility of CO in DMF is nearly negligible.
The former case induces 13C uptake into the Re complex when
13CO2 is used, while in the latter case the Re complex has no
13C uptake, simply evolving CO. Among the data of Figure 6,
those at 180 min photoirradiation, when the starting Re
complex 1 has been nearly transformed into the DMF-
coordinated species 2 and the CO2 adduct 3/3′, as can be
observed in Figure 7, and the highest CO evolution was
observed, were chosen for further analysis. The turnover
number (TON) of CO formation means how many cycles of
reaction each Re complex can drive. Thus, the relative ratio
between the replacement of the tricarbonyls within the complex
and the simple escape from the complex can be estimated from
the TON and the non-13C uptake distribution (n = 0) in Figure
6, as indicated in eq 1 and Scheme 2,, where α denotes the ratio
Figure 7. Time variations of the mass peaks, which can be assigned to
the key species in the positive-ion mode CSI-MS and the resulting CO
evolution: (a) m/z = 592 ([Re(dmbpy)(CO)3Cl + TEA + H+]) as 1,
(b) m/z = 528 ([Re(dmbpy)(CO3)DMF]) as 2, (c) m/z = 572
([Re(dmbpy)(CO3)(CO2) + DMF]) as 3, and (d) the amount of CO
evolution.
αTON = ratio of non‐13C uptake species(n = 0)
(1)
of the escaping versus the replacement of CO with tricarbonyls
(CO)3 in each CO formation. From the data, TON = 2.5, the
relative percentage of the non-13C uptake species for the DMF-
coordinated complex 2 (0.20), the ratio α is calculated to be
0.53. A similar calculation for the one-13C uptake species (n =
1) for the CO2 adduct (0.16) yields α = 0.48. These results
indicate that nearly half of the CO formed within the Re
complex escapes from the complex, while the other half remains
within the complex to replace the tricarbonyls (CO)3. This is
the first example to reveal the microscopic coordination
chemistry of the Re complex, which provides an interesting
deeper insight to understand the behavior.
along with the CO evolution in the photoreaction. In Figure 7a,
the intensity of the peak m/z = 592, assigned to the starting
photocatalyst, 1, decreased drastically within 180 min of
photoirradiation, while the CO evolution reaction of CO2
reduction maintained the reactivity (Figure 7d). This means
that 1 is not the real active photocatalyst but forms an active
catalyst in the reaction pathway. The DMF-coordinated
intermediate 2 (m/z = 528) had appeared just after the
photoirradiation in response to the decay of the starting Re
complex, 1 (m/z = 592), as shown in Figure 7b. This
corresponds to the dissociation of the Cl− ligand from the Re−
6026
dx.doi.org/10.1021/ja500403e | J. Am. Chem. Soc. 2014, 136, 6021−6030

Journal of the American Chemical Society
Article
bipyridyl complex by photoexcitation.35,49 As regards the rapid
decrease of 1 and the efficient production of 2, Kutal et al. have
reported that the strong reductant formed by deprotonation of
the generated radical cation of triethanolamine ([TEOA]•+) by
an electron transfer from TEOA to the excited Re complex
might again reduce the original Re complex to induce an
efficient removal of the halide ion followed by formation of the
solvent-coordinated Re complex.29 Under the current con-
ditions of the photoreaction, it is similarly considered that the
[TEA]•+ formed by an electron transfer from TEA to the
excited Re complex would also afford a strong reductant by
deprotonation to further reduce the Re complex 1 in the dark.
Interestingly, the time course of the peak intensity for the
CO2 adduct of the Re complex 3 (m/z = 572) shown in Figure
7c was different from that for the solvent-coordinated Re
complex 2 in Figure 7b. The intensity for the CO2 adduct of the
Re complex 3 (m/z = 572, Figure 7c) increased in a delayed
timing as compared with the solvent-coordinated species 2 (m/
z = 528, Figure 7b) in the initial photoirradiation period.
Careful inspection of the ratio of the peak intensity between the
two signals at m/z = 528 (Figure 7b) and 572 (Figure 7c) at
each time of observation clearly indicates that they exhibit
different time profiles; the relative ratio of the intensity of (b) at
m/z = 528 against that of (c) at m/z = 572 decreased
systematically with the time of observation as 19 (t = 30 min),
8.6 (t = 180 min), 6.3 (t = 270 min), down to 3.9 (t = 1000
min). This indicates again that the species at m/z = 572 is not
the CO2-solvated one [Re(dmbpy)(CO3)DMF + CO2], but
the CO2-coordinated one [Re(dmbpy)(CO3)(CO2) + DMF].
If it were the CO2-solvated species, it should exhibit exactly the
same time profile as with [Re(dmbpy)(CO3)DMF] (m/z =
528). The delayed timing may be reasonable, because the CO2
adduct of the Re complex 3 would be formed through the
reaction between the 2 and CO2 in the reaction system.
More interestingly, the rise-and-decay profiles of the two
intermediates, the solvent-coordinated species 2 and the CO2
adduct of the Re complex 3, correspond well to the CO
evolution, as shown in Figure 7d; the time regions for the initial
slow evolution, moderate stage, the most reactive stage, and
decelerated stage correspond well to the time profile of each
intermediate. This strongly suggests that the two species, the
solvent-coordinated Re complex 2 and the CO2 adduct 3/3′,
are the respective intermediates in the reaction pathway of
photoreduction of CO2 catalyzed by the Re complex. Since the
current experiments were carried out under steady-state light
irradiation conditions, the time profiles of each component and
CO evolution only express qualitative features of the reaction.
Even in this case, Figure 7 clearly indicates that the
photoreaction should proceed through the following consec-
utive processes as 1 → 2 → [3/3′] → CO evolution.
which photoreduction of CO2 has only poor reactivity,
compared to a frequently used solvent such as DMF. Thus,
we searched thoroughly for appropriate solvents that would
allow both efficient CO2 reduction and IR measurements, and
we found that DMSO is a good candidate. The CO2
photoreduction catalyzed by 1 proceeds with reactivity similar
to that in DMF (Figure S6), and IR measurements are possible
in the region down to 1500 cm−1 by choosing a proper optical
path length within 0.2 mm. We have succeeded in obtaining an
FTIR spectrum by means of a specially designed coaxial
cylindrical IR cell with an Si window. During photoreduction of
CO2 under steady-state light irradiation, the FTIR signal was
monitored by a rapid scan FTIR spectrometer under
atmospheres of nitrogen, 12CO2, and 13CO2 (Figure 8). Very
Figure 8. FTIR spectra of the reaction system of CO2 reduction
catalyzed by [Re(dmbpy)(CO)3Cl] (1, 2 mM) in DMSO/TEA
solvent (v/v = 15/1) under steady-state light irradiation: (a) under
nitrogen atmosphere without CO2, (b) under 12CO2 atmosphere, and
(c) under 13CO2 atmosphere.
interestingly, the following distinct characteristic spectra were
observed. Under nitrogen atmosphere in DMSO/TEA (15:1),
we observed a clear depletion of the three carbonyl (CO triple
bond) absorptions at 2017, 1910, and 1888 cm−1 of the starting
Re complex 1 with appearance of the one-electron-reduced
form of the Re complex around 2000 and 1870 cm−1 (Figure
8a). In the 1600−1700 cm−1 region, however, almost no signal
was observed under nitrogen atmosphere (Figure 8a). Under
12CO2 atmosphere, on the other hand, a clear signal appeared at
1682 cm−1 along with those of the one-electron-reduced form
of the Re complex (Figure 8b). This band clearly shifted to a
smaller wavenumber (1647 cm−1) under 13CO2 (Figure 8c),
and the peaks of the one-electron-reduced form also exhibited
substantial shifts and became broader. These results clearly
indicate that the 1682 cm−1 band (Figure 8b) is due to the
introduction of CO2 and should be ascribed to the key
intermediate in the reaction pathway of the photoreduction
catalyzed by the Re complex. The spectral broadening of the
CO region under 13CO2 is well correlated with the 13C uptake
into CO ligands during the photoreaction (Figure 6). The
wavenumber 1682 cm−1 corresponds exactly to the region for
the CO vibration. Comparisons with the synthesized formate
complex, ReI−OCHO (4, 2016, 1908, 1886, 1625 cm−1), the
carboxylic acid, ReI−COOH (5, 2003, 1890, 1605 cm−1) in
DMSO/TEA (15:1), and DFT calculations have led to the
FTIR Study of the Reaction Mixture during Photo-
irradiation. The CSI-MS results described above clearly
indicate the detection of the key intermediate, the CO2 adduct
with the Re complex. The mass spectra, however, afford only
information on the mass number, which is determined by the
atomic composition. We were thus prompted to study further
the FTIR spectra of the reaction mixture during the
photoirradiation. The vibrational spectra should afford definite
evidence of the nature of the bond. In the studies previously
reported on IR measurements, the solvents employed were
mostly CH3CN or THF, due to their wide optical window for
IR measurements. As described in the Introduction, however,
CH3CN and THF are both rather exceptional solvents, for
6027
dx.doi.org/10.1021/ja500403e | J. Am. Chem. Soc. 2014, 136, 6021−6030

Journal of the American Chemical Society
Article
conclusion that the transient species is definitely different from
4/5 and is assigned as ReII−COOH, which could be formed
through the electrophilic attack of CO2 against the one-
electron-reduced form of the Re complex. This process should
be a two-electron oxidative addition for the Re atom; that is,
the Re0 species (one-electron-reduced form of the ReI
complex) would be converted to a ReII species upon addition
of CO2. The second electron transfer, from either the one-
electron-reduced form of the Re complex accumulated within
the system or a carbon radical possibly formed through the
deprotonation of [TEA]•+, would evolve CO and regenerate
the solvent-coordinated ReI species 2. It should be noted here
that CSI-MS experiment in DMF/TEA (v/v = 5/1) exhibited
both the deprotonated form of the CO2 adduct, [ReII(dmbpy)-
(CO)3(CO2)] (3, m/z = 570, 572), and the protonated form,
[ReII(dmbpy)(CO)3COOH] (3′, m/z = 571, 573), while only
the protonated form, 3′ (1682 cm−1), was detected by FTIR
measurement in DMSO/TEA (15/1). The difference might be
caused by the characteristic soft ionization in the CSI-MS
chamber, the different concentration of the base TEA in DMF
and DMSO, or the nature of the solvents themselves.
(4) The CO2 adduct of the Re complex, 3′, accepts another
electron, probably from the one-electron-reduced form
of the starting Re complex 1′, the one-electron-reduced
form of the solvent-coordinated Re complex 2′, or a
carbon radical possibly formed through the deprotona-
tion of [TEA]•+ to evolve CO, along with the
regeneration of the solvent-coordinated Re complex 2.
(5) The solvent-coordinated Re complex 2 absorbs UV light
to form the one-electron-reduced solvent-coordinated Re
complex, 2′, driving the continuous reaction cycle.
The DMF-coordinated Re complex 2, thus, would be the
major species for the light absorption to drive the reaction after
the induction period of 30 min.
One of the most curious subjects to be clarified in the CO2
photoreduction is the nature of the electron donor for the
second electron, which reduces the CO2 adduct of the Re
complex ([ReII(dmbpy)(CO)3COOH] (3′). In the CSI-MS
and FTIR studies, there was no indication of a Re complex
dimer formation, which was postulated by Fujita et al.27,35
Recently, the Kubiak group reported formation of ReI−COOH,
the synthesis and chemical behavior of which had been studied
by Gibson,47 in an electrochemical simultaneous two-electron
reduction process of the Re−bipyridyl complex.50 This is also
negative evidence for the dimer mechanism. In the photo-
chemical system, the total two-electron reduction of CO2 to
form CO is proceeding in a stepwise manner, as described
above. The second electron, thus, should come from either a
carbon radical possibly formed through the deprotonation of
[TEA]•+ or a one-electron-reduced Re complex species such as
2′. The highly efficient CO2 reduction reported by the Ishitani
group may be due to the relatively long lifetime of the one-
electron species for [Re(dmbpy)(CO)3{P(OEt)3}].11 Even-
tually, the photocatalytic cycle will be maintained until the Re−
bipyridyl complex is decomposed through side reactions during
prolonged photoirradiation. The FTIR study has afforded clear,
conclusive information about the assignment of the CO2 adduct
of the Re complex, 3/3′, observed in the CSI-MS. The
observed species at m/z = 571, 573 is not a simple CO2-
solvated one such as [Re(dmbpy)(CO)3DMF + CO2 + H+] or
[Re(dmbpy)(CO)3 + DMF + CO2 + H+] but the chemical-
bonded CO2 adduct, [Re(dmbpy)(CO)3COOH + DMF].
Reaction Mechanism of the Photoreduction Cata-
lyzed by the Re Complex. These observations are well
correlated with the previous speculation on the reaction
mechanism of photoreduction of CO2 catalyzed by the Re
complex shown in Scheme 1, leading to a deeper insight. The
direct detection of the two key intermediates, 2 and 3/3′, in
this study definitely enables us to depict the reaction
mechanism in a more detailed manner in Scheme 3.
Scheme 3. Photochemical CO2 Reduction Cycle Sensitized
by [Re(dmbpy)(CO)3Cl] (1)
CONCLUSIONS
■
By means of sophisticated CSI-MS measurements of the
reaction mixture, we have successfully detected the DMF-
coordinated Re complex [Re(dmbpy)(CO)3DMF] (2) as the
light-absorbing species to drive photoreduction of CO2. The
key intermediate, the CO2-coordinated Re−bipyridyl complex,
[Re(dmbpy)(CO)3COOH] (3), was also successfully detected
for the first time by CSI-MS. The mass spectra for the reaction
mixture with the isotopically labeled 13CO2 provided clear
evidence for the incorporation of CO2 into the Re−bipyridyl
complex. It was revealed that the starting chloride complex,
[Re(dmbpy)(CO)3Cl] (1), was rapidly transformed into the
DMF-coordinated Re complex through the initial cycle of the
photoreduction of CO2. The observed induction period in the
time profile of the CSI-MS signals can well explain the
subsequent formation of the CO2-coordinated intermediate 3
from the DMF-coordinated Re−bipyridyl complex 2. FTIR
spectra in DMSO clearly indicated that the CO2-coordinated
intermediate should be the ReII−COOH complex. Thus, a
detailed understanding of the mechanism of photochemical
The primary reaction cycles under steady-state light
irradiation proceed as follows:
(1) The starting Re complex 1 undergoes one-electron
reduction by a reductive quenching of the excited Re
complex 1 in its 3MLCT state by TEA to form the radical
anion of 1 (1′ in Scheme 3).
(2) Chloride ion is subsequently eliminated, along with
replacement with the solvent DMF to form the one-
electron-reduced form of the solvent-coordinated Re
complex, 2′.
(3) The one-electron-reduced form of the solvent-coordi-
nated Re complex, 2′, undergoes an electrophilic attack
by CO2 along with protonation to form the CO2 adduct
of the Re complex, 3′.
6028
dx.doi.org/10.1021/ja500403e | J. Am. Chem. Soc. 2014, 136, 6021−6030

Journal of the American Chemical Society
Article
ChemSusChem 2011, 4, 191. (d) Barton, E. E.; Rampulla, D. M.;
Bocarsly, A. B. J. Am. Chem. Soc. 2008, 130, 6342.
(5) Keith, J. A.; Carter, E. A. J. Am. Chem. Soc. 2012, 134, 7580.
(6) Iizuka, K.; Wato, T.; Miseki, Y.; Saito, K.; Kudo, A. J. Am. Chem.
Soc. 2011, 133, 20863.
(7) Hawecker, J.; Lehn, J.-M.; Ziessel, R. J. Chem. Soc., Chem.
Commun. 1983, 0, 536.
(8) Hawecker, J.; Lehn, J.-M.; Ziessel, R. Helv. Chim. Acta 1986, 69,
1990.
(9) Kou, Y.; Nakatani, S.; Sunagawa, G.; Tachikawa, Y.; Masui, D.;
Shimada, T.; Takagi, S.; Tryk, D. A.; Nabetani, Y.; Tachibana, H.;
Inoue, H. J. Catal. 2014, 310, 57.
(10) Hori, H.; Johnson, F. P. A.; Koike, K.; Ishitani, O.; Ibusuki, T. J.
Photochem. Photobiol. A: Chemistry 1996, 96, 171.
(11) Takeda, H.; Koike, K.; Inoue, H.; Ishitani, O. J. Am. Chem. Soc.
2008, 130, 2023.
CO2 reduction sensitized by the Re−bipyridyl complex as a
benchmark reaction has now been obtained.
ASSOCIATED CONTENT
■
S
* Supporting Information
Mass pattern for [Re(dmbpy)(CO)3Cl + TEA + H] in the
positive-ion mode CSI-MS; mass spectrum for 1 in the ordinary
ESI-MS; mass pattern analysis for 3 (m/z = 570, 572) and 3′
(m/z = 571, 573); CSI-MS spectra for alternatively synthesized
4 and 5; analysis of 13C uptake into 2; CO evolution catalyzed
by 1 in DMSO/TEA solvent (v/v = 15/1); FTIR spectra of 4
and 5; and characteristic peaks observed in the positive-ion
mode CSI-MS spectra of the reaction mixture. This material is
available free of charge via the Internet at http://pubs.acs.org.
(12) Gholamkhass, B.; Mametsuka, H.; Koike, K.; Tanabe, T.; Furue,
M.; Ishitani, O. Inorg. Chem. 2005, 44, 2326.
AUTHOR INFORMATION
■
(13) Sato, S.; Koike, K.; Inoue, H.; Ishitani, O. Photochem. Photobiol.
Sci. 2007, 6, 454.
(14) Bian, Z.-Y.; Sumi, K.; Furue, M.; Sato, S.; Koike, K.; Ishitani, O.
Inorg. Chem. 2008, 47, 10801.
(15) Gabrielsson, A.; Lindsay Smith, J. R.; Perutz, R. N. Dalton Trans.
2008, 4259.
Corresponding Author
inoue-haruo@tmu.ac.jp
Present Address
§School of Medicine, Tokyo Medical University, 6-1-1
Shinjuku, Shinjuku-ku, Tokyo 160-8402, Japan
(16) Koike, K.; Naito, S.; Sato, S.; Tamaki, Y.; Ishitani, O. J.
Photochem. Photobiol. A: Chemistry 2009, 207, 109.
(17) Kiyosawa, K.; Shiraishi, N.; Shimada, T.; Masui, D.; Tachibana,
H.; Takagi, S.; Ishitani, O.; Tryk, D. A.; Inoue, H. J. Phys. Chem. C
2009, 113, 11667.
(18) Schneider, J.; Vuong, K. Q.; Calladine, J. A.; Sun, X.-Z.;
Whitwood, A. C.; George, M. W.; Perutz, R. N. Inorg. Chem. 2011, 50,
11877.
(19) Sato, S.; Morikawa, T.; Kajino, T.; Ishitani, O. Angew. Chem.
2013, 125, 1022.
(20) (a) Sato, S.; Morikawa, T.; Saeki, S.; Kajino, T.; Motohiro, T.
Angew. Chem., Int. Ed. 2010, 49, 5101. (b) Suzuki, T. M.; Tanaka, H.;
Morikawa, T.; Iwaki, M.; Sato, S.; Saeki, S.; Inoue, M.; Kajino, T.;
Motohiro, T. Chem. Commun. 2011, 47, 8673.
(21) Arai, T.; Sato, S.; Uemura, K.; Morikawa, T.; Kajino, T.;
Motohiro, T. Chem. Commun. 2010, 46, 6944.
(22) Arai, T.; Tajima, S.; Sato, S.; Uemura, K.; Morikawa, T.; Kajino,
T. Chem. Commnun. 2011, 47, 12664.
(23) Sato, S.; Arai, T.; Morikawa, T.; Uemura, K.; Suzuki, T. M.;
Tanaka, H.; Kajino, T. J. Am. Chem. Soc. 2011, 133, 15240.
(24) Takeda, H.; Ishitani, O. Coord. Chem. Rev. 2010, 254, 346.
(25) Doherty, M. D.; Grills, D. C.; Muckerman, J. T.; Polyansky, D.
E.; Fujita, E. Coord. Chem. Rev. 2010, 254, 2472.
(26) Windle, C. D.; Perutz, R. N. Coord. Chem. Rev. 2012, 256, 2562.
(27) Agarwal, J.; Fujita, E.; Schaefer, H. F.; Muckerman, J. T. J. Am.
Chem. Soc. 2012, 134, 5180.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was partially supported by a Grant-in-Aid for
Scientific Research on Innovative Areas from the Ministry of
Education, Culture, Sports, Science and Technology of Japan.
The authors thank Dr. Tsuyoshi Tsuchibuchi at Shimadzu
Corp. for kind help with the FTIR measurements.
REFERENCES
■
(1) Morris, A. J.; Meyer, G. J.; Fujita, E. Acc. Chem. Res. 2009, 42,
1983.
(2) Inoue, H.; Shimada, T.; Kou, Y.; Nabetani, Y.; Masui, D.; Takagi,
S.; Tachibana, H. ChemSusChem 2011, 4, 173.
(3) (a) Halmann, M. Nature 1978, 275, 115. (b) Ito, K.; Ikeda, S.;
Yoshida, M.; Ohta, S.; Iida, T. Bull. Chem. Soc. Jpn. 1984, 57, 583.
(c) Taniguchi, I.; Aurian-Blajeni, B.; Bockris, J. O’. M. Electrochim. Acta
1984, 29, 923. (d) Taniguchi, I.; Aurian-Blajeni, B.; Bockris, J. O’. M. J.
Electroanal. Chem. 1984, 161, 385. (e) Yoneyama, H.; Sugimura, K.;
Kuwabata, S. J. Electroanal. Chem. 1988, 249, 143. (f) Bockris, J. O’.
M.; Wass, J. C. J. Electrochem. Soc. 1989, 136, 2521. (g) Ueda, J.;
Nakabayashi, S.; Ushizaki, J.; Uosaki, K. Chem. Lett. 1993, 1747.
(h) Flaisher, H.; Tenne, R.; Halmann, M. J. Electroanal. Chem. 1996,
402, 97. (i) Bradley, M. G.; Tysak, T.; Graves, D. J.; Viachopoulos, N.
A. J. Chem. Soc., Chem. Commun. 1983, 349. (j) Cabrera, C. R.; Abruna,
H. D. J. Electroanal. Chem. 1986, 209, 101. (k) Liu, J. F.; ChunYu, B. Z.
J. Electroanal. Chem. 1992, 324, 191. (l) Hinogami, R.; Nakamura, Y.;
Yae, S.; Nakato, Y. J. Phys. Chem. 1998, 102, 974. (m) Aurian-Blajeni,
B.; Halmann, M.; Manassen, J. Sol. Energy Mater. 1983, 8, 425.
(n) Sears, W. M.; Morrison, S. R. J. Phys. Chem. 1985, 89, 3295.
(o) Parkinson, B. A.; Weaver, P. F. Nature 1984, 309, 148.
(p) Canfield, D.; Frese, K. W., Jr. J. Electrochem. Soc. 1983, 130,
1772. (q) Noda, H.; Ikeda, S.; Saito, Y.; Nakamura, T.; Maeda, M.; Ito,
K. Denki Kagaku 1989, 57, 1117. (r) Hashimoto, K.; Fujishima, A. In
Carbon Dioxide Chemistry: Environmental Issues; Pradier, J. P., Pradier,
C. M., Eds.; The Royal Institute of Technology: Stockholm, 1994; pp
388. (s) Hirota, K.; Tryk, D. A.; Yamamoto, T.; Hashimoto, K.;
Okawa, M.; Fujishima, A. J. Phys. Chem. B 1998, 102, 9834.
(28) Schneider, J.; Jia, H.; Muckerman, J. T.; Fujita, E. Chem. Soc. Rev.
2012, 41, 2036.
(29) Kutal, C.; Weber, M. A.; Ferraudi, G.; Geiger, D. Organometallics
1985, 4, 2161.
(30) Kutal, C.; Corbin, A. J.; Ferraudi, G. Organometallics 1987, 6,
553.
(31) Kalyanasundaram, K. J. Chem. Soc., Faraday Trans. 2 1986, 82,
2401.
(32) Hori, H.; Koike, K.; Suzuki, Y.; Ishizuka, M.; Tanaka, J.;
Takeuchi, K.; Sasaki, Y. J. Mol. Catal. A: Chemical 2002, 179, 1.
(33) Hori, H.; Ishihara, J.; Koike, K.; Takeuchi, K.; Ibusuki, T.;
Ishitani, O. Chem. Lett. 1997, 26, 273.
(34) Hori, H.; Ishihara, J.; Koike, K.; Takeuchi, K.; Ibusuki, T.;
Ishitani, O. J. Photochem. Photobiol. A: Chemistry 1999, 120, 119.
(35) Hayashi, Y.; Kita, S.; Brunschwig, B. S.; Fujita, E. J. Am. Chem.
Soc. 2003, 125, 11976.
(4) (a) Seshadri, G.; Lin, C.; Bocarsly, A. B. J. Electroanal. Chem.
1994, 372, 145. (b) Barton, C. E.; Lakkaraju, P. S.; Rampulla, D. M.;
Morris, A. J.; Abelev, E.; Bocarsly, A. B. J. Am. Chem. Soc. 2010, 132,
11539. (c) Morris, A. J.; McGibbon, R. T.; Bocarsly, A. B.
(36) Shinozaki, K.; Hayashi, Y.; Brunschwig, B.; Fujita, E. Res. Chem.
Intermed. 2007, 33, 27.
6029
dx.doi.org/10.1021/ja500403e | J. Am. Chem. Soc. 2014, 136, 6021−6030

Journal of the American Chemical Society
Article
(37) Johnson, F. P. A.; George, M. W.; Hartl, F.; Turner, J. J.
Organometallics 1996, 15, 3374.
(38) Scheiring, T.; Klein, A.; Kaim, W. J. Chem. Soc., Perkin Trans. 2
1997, 2569.
(39) Klein, A.; Vogler, C.; Kaim, W. Organometallics 1996, 15, 236.
(40) Kunimura, M.; Sakamoto, S.; Yamaguchi, K. Org. Lett. 2002, 4,
347.
(41) Yamaguchi, K. J. Mass Spectrom. 2003, 38, 473.
(42) Worl, L. A.; Duesing, R.; Chen, P.; Ciana, L. D.; Meyer, T. J. J.
Chem. Soc., Dalton Trans. 1991, 11, 849.
(43) Bruckmeier, C.; Lehenmeier, M. W.; Reithmeier, R.; Rieger, B.;
Herranz, J.; Kavakli, C. Dalton Trans. 2012, 41, 5026.
(44) Sullivan, B. P.; Meyer, T. J. J. Chem. Soc., Chem. Commnun. 1984,
1244.
(45) Gibson, D. H.; He, H. Chem. Commnun. 2001, 2082.
(46) Hori, H.; Takano, Y.; Koike, K.; Sasaki, Y. Inorg. Chem. Commun.
2003, 6, 300.
(47) (a) Gibson, D. H.; Yin, X. J. Am. Chem. Soc. 1998, 120, 11200.
(b) Gibson, D. H.; Yin, X.; He, H.; Mashuta, M. S. Organometallics
2003, 22, 337.
(48) Hori, H.; Johnson, F. P. A.; Koike, K.; Takeuchi, K.; Ibusuki, T.;
Ishitani, O. J. Chem. Soc., Dalton Trans. 1997, 1019.
(49) Agarwal, J.; Sanders, B. C.; Fujita, E.; Schaefer, H. F., III;
Harrop, T. C.; Muckerman, J. T. Chem. Commun. 2012, 48, 6797.
(50) Sampaon, M. D.; Froelich, J. D.; Smieja, J. M.; Benson, E. E.;
Sharp, I. D.; Kubiak, C. P. Energy Environ. Sci. 2013, 6, 3748.
6030
dx.doi.org/10.1021/ja500403e | J. Am. Chem. Soc. 2014, 136, 6021−6030