Development of an efficient photocatalytic system for CO2 reduction using rhenium(I) complexes based on mechanistic studies

Article abstract of DOI:10.1021/ja077752e

The reaction mechanism of photocatalytic CO₂ reduction using rhenium(I) complexes has been investigated by means of a detailed comparison of the photocatalyses of three rhenium(I) complexes, fac-[Re(bpy)(CO)₃L] (L = SCNT (1-NCS), Cl^- (1-CI), and CN^- (1-CN)). The corresponding one-electron-reduced species (OER) of the complexes play two important roles in the reaction: (a) capturing CO₂ after loss of the monodentate ligand (L) and (b) donation of the second electron to CO₂ by another OER without loss of L. In the case of 1-NCS, the corresponding OER has both of the capabilities in the photocatalytic reaction, resulting in more efficient CO formation (with a quantum yield of 0.30) than that of 1-CI (quantum yield of 0.16), for which OER species have too short a lifetime to accumulate during the photocatalytic reaction. On the other hand, 1-CN showed no photocatalytic ability, because the corresponding OER species does not dissociate the CN^- ligand. Based on this mechanistic information, the most efficient photocatalytic system was successfully developed using a mixed system with fac-[Re(bpy)(CO)₃(CH₃CN)]⁺ and fac-[Re-{4,4′-(MeO)₂bpy}(CO)₃{P(OEt) ₃}]⁺, for which the optimized quantum yield for CO formation was 0.59.

Full text of DOI:10.1021/ja077752e

Published on Web 01/19/2008
Development of an Efficient Photocatalytic System for CO
Reduction Using Rhenium(I) Complexes Based on
Mechanistic Studies
2
Hiroyuki Takeda, Kazuhide Koike, Haruo Inoue, and Osamu Ishitani*,†,|
†
‡
§,|
Department of Chemistry, Graduate School of Science and Engineering, Tokyo Institute of
Technology, 2-12-1-E1-9 O-okayama, Meguro-ku, Tokyo, 152-8551, Japan, National Institute of
AdVanced Industrial Science and Technology, 16-1 Onogawa, Tsukuba, Ibaraki, 305-8569,
Japan, Department of Applied Chemistry, Graduate Course of Urban EnVironmental Sciences,
Tokyo Metropolitan UniVersity, 1-1 Minamiosawa, Hachioji, Tokyo, 192-0364, Japan, and
SORST, Japan Science and Technology Agency (JST), Japan
Received October 9, 2007; E-mail: ishitani@chem.titech.ac.jp
Abstract: The reaction mechanism of photocatalytic CO
2
reduction using rhenium(I) complexes has been
investigated by means of a detailed comparison of the photocatalyses of three rhenium(I) complexes, fac-
-
-
-
[
Re(bpy)(CO)
species (OER) of the complexes play two important roles in the reaction: (a) capturing CO
the monodentate ligand (L) and (b) donation of the second electron to CO by another OER without loss
3
L] (L ) SCN (1-NCS), Cl (1-Cl), and CN (1-CN)). The corresponding one-electron-reduced
2
after loss of
2
of L. In the case of 1-NCS, the corresponding OER has both of the capabilities in the photocatalytic reaction,
resulting in more efficient CO formation (with a quantum yield of 0.30) than that of 1-Cl (quantum yield of
0.16), for which OER species have too short a lifetime to accumulate during the photocatalytic reaction.
On the other hand, 1-CN showed no photocatalytic ability, because the corresponding OER species does
-
not dissociate the CN ligand. Based on this mechanistic information, the most efficient photocatalytic
+
system was successfully developed using a mixed system with fac-[Re(bpy)(CO)
3
(CH
3
CN)] and fac-[Re-
+
{
2
4,4′-(MeO) bpy}(CO)
3
{P(OEt)
3
}] , for which the optimized quantum yield for CO formation was 0.59.
Introduction
a photosensitizer, can reduce CO with relatively high quantum
2
yield and high selectivity of products. In particular, the rhenium-
Development of visible light driven photocatalysts for CO2
reduction might provide a solution for both the shortage of fossil
fuels and the global warming problem. It may give a new carbon
source in place of fossil resources.
I
n+
(I) bipyridine (bpy) complexes fac-[Re (bpy)(CO)3(L)] (L )
Cl (n ) 0); PR3 (n ) 1)) have unique and high yield
photocatalyses, so that they can work not only as a catalyst but
also as a photosensitizer, and CO is the principal product of
Although semiconductors such as TiO2, ZnS, and CdS have
been reported to act as photocatalysts for CO2 reduction, their
7,11-14
+
CO2 reduction.
fac-[Re(bpy)(CO)3{P(OEt)3}] had previ-
1
-5
ously been the most efficient photocatalyst for CO2 reduction
in a homogeneous system that selectively produces CO with a
quantum yields and selectivities of products are low.
Pho-
tocatalytic systems, including transition metal complexes, such
1
2
6
quantum yield of 0.38. Many researchers have looked at the
as ruthenium(II) polypyridine carbonyl complex, cobalt(II)
I
n+
7
8-10
mechanism of CO2 reduction using fac-[Re (bpy)(CO)3(L)]
as a photocatalyst.
trisbipyridine, and cobalt(III) macrocycles
as catalyst with
1
3,15-21
The triplet metal-to-ligand charge
†
Tokyo Institute of Technology.
National Institute of Advanced Industrial Science and Technology.
Tokyo Metropolitan University.
SORST.
‡
(
10) Matsuoka, S.; Yamamoto, K.; Ogata, T.; Kusaba, M.; Nakashima, N.; Fujita,
§
E.; Yanagida, S. J. Am. Chem. Soc. 1993, 115, 601-609.
|
(11) Hawecker, J.; Lehn, J.-M.; Ziessel, R. HelV. Chim. Acta 1986, 69, 1990-
(
1) Inoue, T.; Fujishima, A.; Konishi, S.; Honda, K. Nature 1979, 277, 637-
2012.
6
38.
(12) Hori, H.; Johnson, F. P. A.; Koike, K.; Ishitani, O.; Ibusuki, T. J. Photochem.
Photobiol. A: Chem. 1996, 96, 171-174.
(
2) Thampi, K. R.; Kiwi, J.; Gr a¨ tzel, M. Nature 1987, 327, 506-508.
3) Ishitani, O.; Inoue, C.; Suzuki, Y.; Ibusuki, T. J. Photochem. Photobiol.
A: Chem. 1993, 72, 269-271.
(
(13) Koike, K.; Hori, H.; Ishizuka, M.; Westwell, J. R.; Takeuchi, K.; Ibusuki,
T.; Enjouji, K.; Konno, H.; Sakamoto, K.; Ishitani, O. Organometallics
1997, 16, 5724-5729.
(
(
(
(
(
(
4) Fujiwara, H.; Hosokawa, H.; Murakoshi, K.; Wada, Y.; Yanagida, S.;
Okada, T.; Kobayashi, H. J. Phys. Chem. B 1997, 101, 8270-8278.
5) Ikeue, K.; Yamashita, H.; Anpo, M.; Takewaki, T. J. Phys. Chem. B 2001,
05, 8350-8355.
6) Ishida, H.; Terada, T.; Tanaka, K.; Tanaka, T. Organometallics 1990, 6,
81-186.
7) Hawecker, J.; Lehn, J.-M.; Ziessel, R. J. Chem. Soc., Chem. Commun. 1983,
(14) Kurz, P.; Probst, B.; Spingler, B.; Alberto, R. Eur. J. Inorg. Chem. 2006,
2966-2974.
1
(15) Kutal, C.; Weber, M. A.; Ferraudi, G.; Geiger, D. Organometallics 1985,
4, 2161-2166.
1
(16) Kutal, C.; Corbin, A. J.; Ferraudi, G. Organometallics 1987, 6, 553-557.
(17) Kalyanasundaram, K. J. Chem. Soc., Faraday Trans. 1986, 2, 2401-2415.
(18) Gibson, D. H.; Yin, X.; He, H.; Mashuta, M. S. Organometallics 2003,
22, 337-346.
5
36-538.
8) Grodkowski, J.; Dhanasekaran, T.; Neta, P.; Hambright, P.; Brunschwig,
B. S.; Shinozaki, K.; Fujita, E. J. Phys. Chem. A 2000, 104, 11332-11339.
9) Ogata, T.; Yanagida, S.; Brunschwig, B. S.; Fujita, E. J. Am. Chem. Soc.
(19) Gibson, D. H.; Yin, X. Chem. Commun. 1999, 1411-1412.
(20) Hayashi, Y.; Kita, S.; Brunschwig, B. S.; Fujita, E. J. Am. Chem. Soc.
2003, 125, 11976-11987.
1
995, 117, 6708-6716.
10.1021/ja077752e CCC: $40.75 © 2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 2023-2031
9
2023

A R T I C L E S
Takeda et al.
sodium ion as an ionization reagent,22 were measured with an Applied
Biosystems Mariner System 5231 mass spectrometer.
Table 1. Photophysical and Electrochemical Properties of the
Rhenium(I) Complexes, and Reductive Quenching Rates of Their
Emission by TEOA
Materials. Acetonitrile was distilled three times over P
2 5
O and then
emissiona
quenching by TEOAa
distilled over CaH just before use. Dimethylformamide (DMF) was
2
b
τd
ns
f
g
λ
nm
max
k
q
k
q
τ
E
1
/
2
dried over molecular sieves of size 4 Å and distilled at reduced pressure.
Triethanolamine (TEOA) was distilled at reduced pressure. Tetraethy-
c
M- s-
1
1
M-
1
complex
Φ
r
V
8
lammonium tetrafluoroborate was prepared according to standard
1
1
1
-NCS
-CN
-Cl
635
611
637
0.003
0.013
0.003
30
87
25
3.7 × 10
2.5 × 10
8.0 × 10
11
22
2.1
-1.61
-1.67
-1.67
8
23,24
methods
and dried in vacuo at 100 °C for one night before use.
e
7 e
Tetra-n-butylammonium thiocyanate and tetraethylammonium chloride
were dried in vacuo at 100 °C for one night before use. The rhenium-
a
Measured in DMF at room temperature under an Ar atmosphere.
(
I) complexes, fac-[Re(bpy)(CO)
(CO) Cl] (1(Me)-Cl), and fac-[Re(bpy)(CO)
MeCN] ), were prepared according to standard methods.
fac-[Re(bpy)(CO) (NCS)] (1-NCS). An ethanol/water (1:1 v/v)
3
Cl] (1-Cl), fac-[Re(4,4′-Me
2
bpy)-
b
e
c
d
Emission maxima. Emission quantum yields. Excited-state lifetimes.
f
3
3
(CH CN)](PF ) ([1-
3
6
Value measured in MeCN solution (ref 17). Quenching rate constants of
the emission by TEOA. Redox potentials vs Ag/AgNO3 (0.01 M) for (1-
g
+
25,26
-
•
x/1-x ).
3
mixed solution (300 mL) containing 1-Cl (500 mg) and NaSCN (8.8
g) was refluxed under an Ar atmosphere for 12 h. The complex 1-NCS
was extracted from the reaction solution with CH Cl three times. The
3
transfer ( MLCT) excited state of the rhenium complexes was
quenched by tertiary amine (such as triethanolamine, TEOA)
as the first step of the photocatalytic reaction, generating the
one-electron reduced (OER) species of the rhenium complexes,
2
2
resulting organic layer was washed with water and evaporated to give
1-NCS as yellow solids. Purification of 1-NCS was achieved by
recrystallization with an acetone-water mix twice and then with
15-17
in laser flush photolysis studies.
However, the following
acetone-Et
acetone-d ): δ/ppm, 7.87 (2H, ddd, J ) 1.4 Hz, 5.6 Hz, 7.8 Hz, bpy
O. The typical yield was 85%. 1H NMR (270 MHz,
2
processes are not yet understood in detail: how the OER species
react with CO2, what is the second electron source for the two-
electron reduction of CO2 to CO, and how the photocatalyst is
recovered after CO is produced.
There are some clues. In the photocatalytic reaction with fac-
Re(bpy)(CO)3Cl in the presence of Br , fac-Re(bpy)(CO)3Br
was formed. On the other hand, under a CO2 atmosphere,
the complex converted to fac-Re(bpy)( CO)3Cl during the
photocatalytic reaction.11 The OER species fac-[Re(bpy)(CO)3-
6
5
5′
4
4′
H , H ), 8.42 (2H, ddd, J ) 1.6 Hz, 7.8 Hz, 8.1 Hz, bpy H , H ), 8.76
3
3′
(
2H, ddd, J ) 1.1 Hz, 1.1 Hz, 8.1 Hz, bpy H , H ), 9.15 (2H, ddd, J
6
6′
-1
)
0.8 Hz, 1.6 Hz, 5.6 Hz, bpy H , H ). IR (MeCN): ν(CO)/cm
,
-1
+
2
027, 1919(br); ν(CN)/cm , 2098. ESI-MS: m/z, 508 [M + Na] .
-
-1
-1
2 2
UV-vis (CH Cl ): λmax/nm (ꢀ / M cm ), 297 (19700), 396 (2900).
7
13
14 8 3 3
Anal. Calcd (%) for C H O N SRe: C, 34.71; H, 1.66; N, 8.67; S,
6.62. Found: C, 34.51; H, 1.61; N, 8.48; S, 6.73.
13
-
There are two linkage isomers with a SCN ligand, i.e., thiocyanato
(
PR3)], which are produced by the photoinduced electron-
complex M-SCN and isothiocyanato complex M-NCS. Recently,
1-NCS was synthesized by Vl cˇ ek et al.27 and was found to be the
isothiocyanato isomer by X-ray single crystal analysis. The spectral
and analytical data of 1-NCS prepared in the present study were
consistent with those reported by Vl cˇ ek et al.27
transfer reaction with triethanolamine, reacted with CO2 in the
-
4
-2
-1 -1
dark, with rate constants 3.5 × 10 -1.9 × 10
M
s , and
disproportionation of the OER species is not likely to be a major
process for CO2 reduction.13 Some “CO2 adducts” with reduced
fac-[Re(bpy)(CO)
tion of the method of that reported by Leasure et al., as follows. An
ethanol/water (1:1 v/v, 50 mL) mixture containing [1-MeCN] (PF
100 mg) and KCN (1.06 g) was refluxed under an Ar atmosphere for
h in dim light. The complex was extracted with CH Cl three times,
3
(CN)] (1-CN). We prepared 1-CN by modifica-
rhenium complexes, such as a CO2-bridged rhenium dimer and
metalocarboxylates, have been proposed as key intermediates
for the photocatalytic reduction;1
have been directly observed under photocatalytic reaction
conditions.
We report below a detailed mechanistic study of photocata-
lytic CO2 reduction using three rhenium(I) complexes with
anionic ligand fac-[Re(bpy)(CO)3(L)] (L ) NCS , Cl , and
CN ) which have similar photophysical properties (shown in
Table 1). This has clarified which ligand is eliminated from
the OER species before reaction with CO2, the identification
of the second-electron donor, and how the starting complexes
are reproduced in the photocatalytic cycle.
A most efficient homogeneous photocatalytic system has been
developed by the molecular design of a new photocatalytic
system based on the information about the reaction mechanism
found in this study.
2
8
+
-
6
)
8-21
none of these complexes
(
6
2
2
and the organic layer was washed with water three times. The solvent
was evaporated, to give yellow solids. The complex 1-CN was isolated
by column chromatography on aluminum oxide with CH Cl /CH CN
2
2
3
-
-
(3:1 v/v) eluent. Two bands appeared in the column; the first band
contained 1-CN. Further purification took place by recrystallization
-
with an acetone-water mix twice and then with acetone-Et
2
O. The
): δ/ppm, 7.80 (2H,
ddd, J ) 1.0 Hz, 5.6 Hz, 7.6 Hz, bpy H , H ), 8.35 (2H, ddd, J ) 1.4
1
yield was 64%. H NMR (200 MHz, acetone-d
6
5
5′
4
4′
Hz, 7.6 Hz, 8.5 Hz, bpy H , H ), 8.73 (2H, ddd, J ) 1.0 Hz, 1.0 Hz,
3
3′
8
.5 Hz, bpy H , H ), 9.13 (2H, ddd, J ) 1.0 Hz, 1.2 Hz, 5.6 Hz, bpy
6
6′
-1
-1
H , H ). IR (MeCN): ν(CO)/cm , 2023, 1925, 1915; ν(CN)/cm
124. ESI-MS: m/z, 476 [M + Na] . UV-vis (CH
,
+
2
M
2
Cl
2
): λmax/nm (ꢀ/
-1
-1
cm ), 252 (16300), 287 (17400), 315 (7800), 374 (4000). Anal.
Calcd (%) for C H O N Re: C, 37.17; H, 1.78; N, 9.29. Found: C,
14
8
3
3
3
7.37; H, 1.85; N, 9.15.
Experimental Section
General Procedures. IR spectra were recorded on a JASCO FT/
(22) Hori, H.; Ishihara, J.; Koike, K.; Takeuchi, K.; Ibusuki, T.; Ishitani, O.
-
1
Chem. Lett. 1997, 273-274.
IR-610 spectrometer at 2-cm resolution. UV-vis absorption spectra
were recorded on a JASCO V-565 or Photal MCPD-2000 photodiode-
array spectrometer. The emission spectra were recorded with a JASCO
(
23) Wheeler, Charles M. J.; Sandstedt, R. A. J. Am. Chem. Soc. 1955, 77, 2025-
2026.
(24) Moe, N. S. Acta Chem. Scand. 1965, 19, 1023-1024.
(
(
25) Caspar, J. V.; Meyer, T. J. J. Phys. Chem. 1983, 87, 952-957.
26) Worl, L. A.; Duesing, R.; Chen, P.; Ciana, L. D.; Meyer, T. J. J. Chem.
Soc., Dalton Trans. 1991, 849-858.
1
FP6600 fluorescence spectrometer. H NMR spectra were recorded on
a Bruker AC200, JEOL EX270, or AL300 NMR spectrometer.
Electrospray ionization mass spectra of the neutral complexes, using
(27) Rodr ´ı guez, A. M. B.; Gabrielsson, A.; Motevalli, M.; Matousek, P.; Towrie,
M.; Sebera, J.; Z a´ li sˇ , S.; Anntonin. Vl cˇ ek, J. J. Phys. Chem. A 2005, 109,
5016-5025.
(
21) Shinozaki, K.; Hayashi, Y.; Brunschwig, B. S.; Fujita, E. Res. Chem.
Intermed. 2007, 33, 27-36.
(28) Leasure, R. M.; Sacksteder, L.; Nesselrodt, D.; Reitz, G. A.; Demas, J. N.;
DeGraff, B. A. Inorg. Chem. 1991, 30, 3722-3728.
2024 J. AM. CHEM. SOC.
9
VOL. 130, NO. 6, 2008

Photocatalytic CO
2
Reduction Using Re(I) Complexes
A R T I C L E S
fac-[Re{4,4′-(MeO)
50 mL) containing Re(CO)
bipyridine (4,4′-(MeO) bpy) (0.19 g) was refluxed under an Ar
2 3
bpy}(CO) Cl] (1(MeO)-Cl). A toluene solution
(
5
Cl (0.30 g) and 4,4′-dimethoxy-2,2′-
2
atmosphere for 3 h in dim light. After the solution cooled to room
temperature, the resulting yellow precipitates were filtered and washed
with toluene, hot hexane, and then n-pentane. The yield was 0.40 g
1
(
7
93%). H NMR (300 MHz, acetone-d
6
3
): δ/ppm, 4.13 (6H, s, CH O),
5
5′
.29 (2H, dd, J ) 2.7 Hz, 6.6 Hz, bpy H , H ), 8.17 (2H, d, J ) 2.7
Hz, bpy H , H ), 8.83 (2H, d, J ) 6.6 Hz, bpy H , H ). IR (MeCN):
ν(CO)/cm , 2020, 1913, 1893. Anal. Calcd (%) for C15
3
3′
6
6′
-
1
H
12
O
5
N
2
ClRe:
C, 34.52; H, 2.32; N, 5.37. Found: C, 34.53; H, 2.58; N, 5.62.
+
-
([1(MeO)-P]+-
fac-[Re{4,4′-(MeO)
2
bpy}(CO)
3
{P(OEt)
3
}] (PF
6
)
-
(
PF
6
) ). A tetrahydrofuran solution (50 mL) containing Re{4,4′-
+
-
(
MeO) bpy}(CO) Cl (0.30 g) and Ag CF SO (0.15 g) was refluxed
2 3 3 3
Figure 1. Irradiation-time dependence of CO formation using 1-NCS (2),
under an Ar atmosphere for 1.5 h in dim light. The precipitated silver
chloride was filtered off, and triethyl phosphite (1 mL) was added to
the filtrate. The solution was refluxed under an Ar atmosphere for 8 h
in dim light. The solvent was evaporated, and the resulting yellow solids
were washed several times with n-pentane. The salts were dissolved in
1
-Cl (b), and 1-CN (9): irradiation with 365-nm light to a 4-mL DMF-
TEOA (5:1 v/v) solution containing the complex (0.5 mM) under a CO2
atmosphere, with light intensity 8.5 × 10 einstein s
-
9
-1
.
)
10 nm) purchased from Asahi Spectra Co. and a CuSO solution
4
methanol and a saturated methanolic solution of NH
4
PF
salt of [1(MeO)-P] . Purification was achieved by
recrystallization with CH Cl /Et O several times, with drying under
vacuo. The yield was 0.46 g (70%). H NMR (300 MHz, acetone-d
δ/ppm, 1.09 (9H, t, J ) 7.1 Hz, POCH CH ), 3.94 (6H, quint, J ) 7.1
Hz, POCH CH ), 4.16 (6H, s, CH O), 7.42 (2H, dd, J ) 2.8 Hz, 6.5
6
was added,
-
1
-
+
(250 g l , 5-cm pass length) filter. Neutral density (ND) glass filters
were used in order to reduce the light intensity. The temperature of
the reaction solution was maintained at 25 ( 0.2 °C using a TAITEC
LabBath LB-21JR cooling thermo pump during the irradiation. The
to give the PF
6
2
2
2
1
6
):
2
3
incident light intensity was determined using a K
3 2 4 3
[Fe(C O )
] acti-
, were
2
5
3
5′
3
3
3′
nometer.30 The gaseous reaction products, i.e., CO and H
Hz, bpy H , H ), 8.34 (2H, d, J ) 2,8 Hz, bpy H , H ), 8.93 (2H, d,
J ) 6.5 Hz, bpy H , H ). IR (MeCN): ν(CO)/cm , 2044, 1956, 1923.
2
6
6′
-1
analyzed using GC-TCD (Shimadzu GC-9A) with an active carbon
column. Quantum yields for CO formation were calculated as the
number of CO molecules formed, divided by the number of absorbed
photons. The complexes were analyzed using an HPLC system with a
Shimadzu LC-10AD pump, a Nomura ODS-HG-5 column (25 cm), a
Shimadzu SPD-10AV detector (wavelength 320 nm), and a Rheodyne
-
+
-1
ESI-MS: m/z, 653 [M- PF
6
] . UV-vis (CH
cm ), 251 (33900), 294sh (10300), 301 (9260), 336 (5090). Anal. Calcd
%) for C21 Re: C, 31.62; H, 3.41; N, 3.51. Found: C,
1.71; H, 3.51; N, 3.56.
Electrochemistry. Cyclic voltammograms were measured in an
acetonitrile solution containing complex (0.5 mM) and Et NBF (0.1
2 2
Cl ): λmax/nm (ꢀ/M -
-1
(
3
27 8 2 2 6
H O N P F
4
4
7
125 injector. The mobile phase was a mixture (3:2 v/v) of MeOH
and a KH PO -NaOH buffer solution (0.05 M, pH 5.9). The anions
M) as the supporting electrolyte under an Ar atmosphere. We used a
BAS CHI620 electrochemical analyzer with a glassy-carbon working
electrode (diameter 3 mm), a Ag/AgNO
2
4
in the solution were analyzed using an Otsuka Electronics CAPI-3300I
Photal capillary electrophoresis system with a buffer solution (pH
3
(0.01 M) reference electrode,
-
1
and a Pt counter electrode at a scan rate of 100 mV s
.
8
.5) consisting of ammonium molybdate, diethylenetriamine, and
trishydroxyaminomethane as the electrolyte.
13C NMR Study of Photocatalytic CO
2
9
Flow electrolysis was used to measure IR and UV-vis absorption
spectra of the OER species of the complexes. The working electrode
was contained in a porous glass tube (2 mm i.d., 3 mm o.d., 40 mm
length) and consisted of many strands of 0.1 mm diameter Pt wire.
2
6
Reduction. A DMF-d -
TEOA (1:5 v/v) solution (0.5 mL) containing 1-NCS (30 mM) was
degassed by the freeze-pump-thaw method and transferred into an
NMR tube (total volume ∼4 mL; i.d. ) 5 mm). After 0.72 atm of
The counter electrode was a Pt wire coiled around the porous glass
-
tube. The I
2
/I
3
reference electrode, which consists of a Pt wire and an
and 0.1 M Et NI, was equipped
near the counter electrode. An acetonitrile solution containing a rhenium
complex (0.5 mM), and Et NBF (0.1 M) as electrolyte, was purged
13
13
acetonitrile solution containing 0.1 M I
2
4
CO gas ( C content 99.8 atom%) was introduced through a vacuum
line, the NMR tube was shaken for several minutes and the tube was
sealed with a flame torch. The solution was irradiated with the merry-
go-round apparatus described above. The 13C NMR spectra of the
solution after 24-h irradiation were measured on a JEOL Lambda 500
4
4
with Ar for 20 min and was then passed into the flow-through cell at
a rate of 0.5 mL/min using a JASCO PU-980 pump. Immediately after
electrolysis using a BAS CHI 620 electrochemical analyzer, the solution
was transferred to a CaF
cell (path length 1.5 mm) for spectral measurements.
Photocatalytic Reactions. A 4-mL DMF- TEOA (5:1 v/v) solution
containing a rhenium complex in a pyrex test tube (i.d. ) 8 mm; 11.4
mL volute) or a quartz cubic cell (1-cm pass length; 11.0 mL volume)
system (125.65 MHz), using DMF-d
of the peaks obtained by the NOE complete H-decoupling method
6
as an internal standard. Integration
1
2
cell (path length ca. 1.5 mm) and/or a quartz
(
EXMOD: nne) was used to determine the relative concentrations of
the 13CO ligands. The carbons in the bpy ligand were used as an internal
standard.
2
was bubbled through with CO for 20 min. The pyrex test tubes were
equipped with a merry-go-round apparatus with a 500-W high-pressure
Results and Discussion
mercury lamp and a solution cutoff filter (<330 nm; 1-cm path length);
Photocatalyses. Figure 1 shows the amounts of CO formed
in the photocatalytic CO2 reduction using 1-NCS, 1-Cl, and
this is a 300 mL aqueous solution containing NiSO
4
2
‚6H O (75 g) and
CoSO ‚7H O (72 g).
4
2
1
-CN. It is clear that the photocatalytic activities of these three
To determine the quantum yield of a photochemical reaction, or to
record spectral changes of the reaction solution, the sample solution
complexes are very different. In the case of 1-NCS, 60 µmol
of CO were produced after 25 h of irradiation, whereas 1-Cl
produced only 30 µmol of CO under the same conditions.
(in a quartz cubic cell) was irradiated using an Ushio Optical Moduley
high-pressure Hg lamp BA-H500 with a 365-nm band-pass filter (fwhm
(
29) Ishitani, O.; George, M. W.; Ibusuki, T.; Johnson, F. P. A.; Koike, K.;
Nozaki, K.; Pac, C.; Turner, J. J.; Westwell, J. R. Inorg. Chem. 1994, 33,
(30) Calvert, J. G.; Pitts, J. N. Photochemistry; Wiley and Sons: NewYork,
4
712-4717.
1966.
J. AM. CHEM. SOC.
9
VOL. 130, NO. 6, 2008 2025

A R T I C L E S
Takeda et al.
Figure 2. At left, UV-vis spectra changes of DMF-TEOA (5:1 v/v) solutions of (a) 1-NCS, (b) 1-CN, and (c) 1-Cl (0.5 mM) during irradiation with
65-nm monochromic light under a CO2 atmosphere (before the irradiation; ∼1 s; 5 s; 10 s; 15 s). At right, UV-vis spectral changes of CH3CN solutions
containing (d) 1-NCS and (e) 1-CN (0.5 mM) with 0.1 M Et4NBF4 during flow electrolysis under an Ar atmosphere. The following potentials vs Ag/AgNO3
0.01 M) were employed: (d) before the electrolysis; -1.35; -1.55; -1.65; -1.70; -1.75; -1.80 V and (e) before the electrolysis; -1.25; -1.45; -1.50;
1.65; -1.70; -1.75; -1.80; -1.85; -1.95; -2.00 V. The arrows indicate increasing absorbance.
3
(
-
However, 1-CN cannot act as a photocatalyst.31 The turnover
numbers for CO formation (TNCO) were 30 and 15 using 1-NCS
and 1-Cl, respectively. The quantum yields of CO formation
reaction. Similar results were obtained using 1-CN (Figure 2b
and 2e) instead of 1-NCS.
3
As shown in Table 1, the MLCT excited states of both
(ΦCO) were 0.30 (1-NCS) and 0.16 (1-Cl), where the concentra-
1-NCS and 1-CN are efficiently quenched by TEOA. Reductive
tion of the complex was 2.5 mM and the light intensity was
quenching of the excited state should therefore give the OER
species of these complexes. Although similar reductive quench-
ing has been reported in the case of 1-Cl, a far smaller amount
of the OER species accumulated in the reaction solution even
in the first stage of the photocatalytic reaction, as shown in
Figure 2c, because of the much lower stability of the OER
species of 1-Cl than of the other complexes.
-
9
-1
7
.5 × 10 einstein s ; the incident light was almost completely
absorbed by the complex. Only small amounts of H2 (less than
0
.1 µmol) were formed in all cases.
Photochemical Production and Reactivities of One-Electron-
Reduced Species. Figure 2a shows the UV-vis absorption
spectral change of the reaction solution containing 1-NCS during
the photocatalytic reaction. A sharp absorption band near 500
nm, and a broad band in the near IR region, appeared just after
irradiation commenced. These new bands were characteristic
The reactivities of the OER species of both 1-NCS and 1-CN
with CO2 were investigated as follows. The OER species were
-7
-1
produced by photoirradiation (1.0 × 10 einstein s ) of a DMF
solution containing the complex in the presence of TEOA (1.26
M) for 20-90 s under CO2, CO2-Ar (1 : 1), or Ar. Immediately
after stopping irradiation, the decay profiles of the OER species
were followed in the dark by measuring the absorbance at 506
nm, a wavelength which can be absorbed by the OER species
but not by the starting complexes. Figure 3 shows the results
for 1-NCS. The decay rate of the OER species of 1-NCS
increased in the presence of higher concentrations of CO2: we
found that τ1/2 ) 98 s under Ar, 49 s under CO2-Ar (1:1)
mixture, and 39 s under CO2. Decay of the OER species of
1-Cl was also accelerated in the presence of CO2: τ1/2 ) 39 s
-
•
for the anion radical for the bpy ligand (bpy ), and a very
similar absorption spectrum was observed by the quantitative
one-electron reduction of 1-NCS using a flow-electrolysis
technique, as shown in Figure 2d. Consequently, the one-electron
-•
-
reduced (OER) species of 1-NCS, fac-[Re(bpy )(CO)3(NCS)] ,
should be produced in the first stage of the photocatalytic
(
31) Although photocatalysis of 1-CN which was produced in the photoreaction
3 2
of fac-[Re(bpy)(CO) atmo-
(4-cyanopyridine)]+ with TEOA under a CO
sphere has been reported, it is clear that very pure 1-CN has no
photocatalytic activity. A small amount of other complex(es) should work
2
as a catalyst for CO reduction: Hori, H.; Ishihara, J.; Koike, K.; Takeuchi,
K.; Ibusuki, T.; Ishitani, O. J. Photochem. Photobiol. A: Chem. 1999, 120.
19-124.
1
2026 J. AM. CHEM. SOC.
9
VOL. 130, NO. 6, 2008

Photocatalytic CO
2
Reduction Using Re(I) Complexes
A R T I C L E S
1
-NCS and 1-CN the corresponding OER species accumulated
immediately after a short period of irradiation, as described
above. The behavior of the OER species varied. Further
-
•
irradiation caused a rapid decrease of [1-NCS] to one-sixth
of the maximum amount after 3 min of irradiation, but a
-
•
photostationary state of [1-NCS] was reached after 30 min
-
•
of irradiation. After that, a very slow decrease of [1-NCS]
-
•
occurred, but [1-NCS] was still observed until CO formation
ceased (Figure 4a). The rapid decrease of [1-NCS] in the first
-•
stage of the photoreaction should correspond to elimination of
-
-
the SCN ligand, because addition of excess SCN caused a
-
•
dramatic increase of [1-NCS] in the photostationary state
Figure 4d). No comparable rapid decrease of [1-CN]-• was
observed in the first stage of the photoreaction, but [1-CN]
(
-
•
decreased gradually during 5 h of irradiation (Figure 4c). The
Figure 3. Decay curves of the absorbance at 506 nm of DMF-TEOA
-
•
accumulation and decrease of [1-Cl] were similar to those of
(
5:1 v/v) solutions with 1-NCS (0.5 mM) under Ar (O), CO2-Ar (1:1)
-
•
(
]), and CO2 atmosphere (b) after irradiation with 365-nm monochromic
[1-NCS] , but the amount that accumulated in the photosta-
tionary state was much less (Figure 4b).
-
8
-1
light having intensity of 1.0 × 10 einstein s
.
There have been reports of the dissociation of a ligand from
OERs of rhenium diimine complexes in electrospectroscopic
3
2
studies. Stufkens et al. shows that the elimination capabilities
of ligands from the OER species is in the order halide . PPh3,
n-PrCN > P(OMe)3 that corresponds to the order of weakness
3
3
of π acceptabilities of the ligands. Kaim et al. looked at the
-
•
electron paramagnetic resonance spectroscopy of [1-Cl] and
-
-•
suggested that the dissociation of Cl from [1-Cl] is due to
the electron density population on the Re-Cl dσ* orbital,
through hyperconjugation from bpy π*. Bignozzi et al.34
reported the CV behavior of 1-CN and 1-Cl at low temperature
and concluded that, at low temperature, the doubly reduced
-
species of 1-Cl eliminates Cl , whereas that of 1-CN does not
-
eliminate CN . They pointed out that the difference in the
reactivities is attributable to the different energies of the Re-L
-
-
-
dσ* orbitals, i.e., dσ*(Cl ) < dσ*(CN ), because the CN
ligand induces stronger ligand field splitting. The stability of
[
-
•
-•
-•
1-NCS] is between that of [1-Cl] and [1-CN] , probably
because of its moderate ligand field and π acceptability.
Change of the Rhenium Complexes during Photocatalytic
Reactions. The rhenium complexes in the solutions before and
after the photocatalytic reaction were analyzed using high-
performance liquid chromatography (Figure S1 in the Supporting
Information). In all cases the peaks of the original complexes
decreased, and new peaks appeared. For both 1-NCS and 1-Cl,
Figure 4. Yields of the OER species of the complex (b) and CO (O) during
photocatalytic reactions using (a) 1-NCS, (b) 1-Cl, and (c) 1-CN (2.5 mM)
with 365-nm monochromic light (intensity 7.5 × 10 einstein s ) under
a CO2 atmosphere. (d) Comparison of yields of OER of 1-NCS in the
absence (O) and presence (b) of n-Bu4NSCN (50 mM), where the reaction
conditions were as described above.
-
9
-1
under Ar and < 1 s under CO2. In contrast, the decay of the
OER species of 1-CN was not affected by CO2.
These results clearly show that the photochemical formation
rates of all three OER complexes are similar (eq 1), but their
the new peaks were observed at retention times of 6.7, 7.1, and
+
8
(
.6 min, which were attributed to [Re(bpy)(CO)3(DMF)] , [Re-
+
bpy)(CO)3(TEOA)] , and [Re(bpy)(CO)3(O2CH)], respec-
35
-
-
tively. In the same solutions, free SCN and Cl were detected
by capillary electrophoresis (CE). Figure 5a shows the time
profiles of CO formation, the decrease of 1-NCS, and the
-
•
-•
decay profiles are quite different, i.e., [1-NCS] and [1-CN]
are relatively stable in the dark and accumulate in solution, but
-
•
-•
-•
[
1-Cl] is not stable; also, [1-NCS] and [1-Cl] can react
-
formation of free SCN during the photocatalytic reaction. A
-
•
with CO2, but [1-CN] does not.
rapid decrease of 1-NCS was observed within 1 h of irradiation,
-
followed by a much slower decrease. The amount of free SCN
hν
fac-[Re(bpy)(CO) L] in DMF/TEOA (5:1 v/v)8
was almost the same as that of the decrease of the starting
complex. Similar phenomena were observed in the case of 1-Cl,
but the decrease of 1-Cl was more rapid, and the amount of
3
-•
-
fac-[Re(bpy )(CO) L] (1)
3
Figure 4 shows the time profiles of the amount of the OER
species, during the photocatalytic reactions. Note the experi-
mental differences between Figures 3 and 4: the former was
obtained in the dark immediately after brief irradiation, whereas
for the latter the solution was continuously irradiated. For both
(32) Stor, G. J.; Hartl, F.; van Outersterp, J. W. M.; Stufkens, D. J. Organo-
metallics 1995, 14, 1115-1131.
(
33) Klein, A.; Vogler, C.; Kaim, W. Organometallics 1996, 15, 236-244.
(34) Paolucci, F.; Marcaccio, M.; Paradisi, C.; Roffia, S.; Bignozzi, C. A.;
Amatre, C. J. Phys. Chem. B 1998, 102, 4759-4769.
(
35) Hori, H.; Johnson, F. P. A.; Koike, K.; Takeuchi, K.; Ibusuki, T.; Ishitani,
O. J. Chem. Soc., Dalton Trans. 1997, 1019-1023.
J. AM. CHEM. SOC.
9
VOL. 130, NO. 6, 2008 2027

A R T I C L E S
Takeda et al.
Figure 5. Irradiation time dependence of CO formation (O), and amounts
of the starting complex (2) and the dissociated anion (4) in the reaction
solutions containing (a) 1-NCS, (b) 1-Cl, and (c) 1-CN. A 4-mL DMF-
TEOA (5:1 v/v) solution was used containing 2.5 mM of the complex using
Figure 6. (a) Irradiation-time dependence of CO formation (O), and
amounts of 1-NCS (2) and 1-Cl (4) in the photocatalytic reaction solution
-
9
-1
3
65-nm monochromic light at a high intensity of 7.25 × 10 einstein s
-
under a CO2 atmosphere.
in the presence of Cl ; a 4-mL DMF-TEOA (5:1 v/v) solution containing
1-NCS (0.5 mM) and Et4NCl (10 mM) was irradiated under a CO2
residual 1-Cl after 1 h of irradiation was less than that with
-NCS (Figure 5b). On the other hand, no such rapid decrease
atmosphere. (b) Ratios of concentrations of 1-Cl and 1-NCS in the
photocatalytic reaction solution in the presence of n-Bu4NSCN, 0.5 mM
(2) and 1.0 mM (9), and in its absence (b). The reaction conditions were
the same as above.
1
was observed in the case of 1-CN, which decreased gradually
during 5 h of irradiation, as shown in Figure 5c. Although new
peaks were observed in the chromatogram of the solution after
2
. The decrease of [1-NCS]-• under CO2 in the dark became
1
h of irradiation, the solvent complexes and the formate
-
slower in the presence of excess SCN : τ1/2 ) 39 s in the
absence of SCN , but τ1/2 ) 64 and 83 s in the presence of 50
and 100 mM SCN . On the other hand, no slow decrease of
complex were not detected in the case of 1-CN (Figure S1c in
-
-
the Supporting Information). Free CN was not detected at all
-
by CE analysis. These results clearly indicate that elimination
of the CN- ligand does not occur from [1-CN]-• under the
-•
+
-
[1-NCS] was observed in the presence of 100 mM Et4N BF4 .
. Although 1-CN is efficiently reduced under the photo-
3
photocatalytic reaction conditions. The yellow color of the
solution, attributed to 1-CN, was bleached after 5 h of
irradiation, but three ν(CO) stretching bands were still observed
in the FT-IR spectrum. Structural changes of the bpy ligand,
such as reduction and elimination, may occur during irradiation.
The observed behaviors of the complexes and of the corre-
-
•
catalytic reaction conditions to give [1-CN] , elimination of
-
-•
CN from [1-CN] does not occur, and 1-CN cannot function
as a photocatalyst for CO2 reduction.
-
Although the elimination of Cl is a key step of electrocata-
3
6,37
lytic reduction of CO2 by fac-Re(bpy)(CO)3Cl,
the present
-
-
sponding OER species clearly show that the SCN and Cl
observations constitute the first evidence that reaction of 1 with
ligands are eliminated from [1-NCS]-• and [1-Cl] , respec-
-•
CO2 causes CO formation in the photocatalytic reaction.
-•
tively, in the first stage of the photocatalytic reactions, while
Fujitaetal.reportedthereactivitywithCO2 of[Re(bpy )(CO)3],
-
-•
20,21
elimination of CN from [1-CN] does not occur; there should
made by photoirradiation to [Re(bpy)(CO)3]2.
Their results
-
•
-•
be other relatively slow decomposition processes of [1-CN]
eq 2). The equilibrium, shown in eq 3, should be achieved by
show that [Re(bpy )(CO)3] rapidly accommodates a solvent
molecule(S)suchasTHFandCH3CNtoproduce[Re(bpy )(CO)3S],
in which the unpaired electron is located mainly on the bpy
-•
(
further irradiation in the cases of 1-NCS and 1-Cl, because of
the appearance of the photostationary states of 1-NCS, 1-Cl
-
•
ligand. The reaction rate of [Re(bpy )(CO)3S] with CO2 was
-1
(Figure 5), and the corresponding OER species (Figure 4).
relatively slow (kobs ≈ 0.003 s under 0.8 atm of CO2) in THF.
-• -•
It follows that the OER species [1-NCS] and [1-Cl] ,
3
which were formed via reductive quenching of the MLCT
excited state by TEOA, dissociates into X (SCN and Cl )
andtheOERspeciesofthesolventocomplexes[Re(bpy )(CO)3S]
-
-
-
-•
(S ) DMF, TEOA), and the OER species of the solvento
complexes should react with CO2.
The following evidence strongly supports the hypothesis that
It is again noteworthy that, although the solvento complexes
and a small amount of the formate complex [Re(bpy)(CO)3-
-
•
-
1
reacts with CO2 to give CO but the [1-X] (X ) SCN ,
-
Cl ) does not:
-
(36) Sullivan, B. P.; Bolinger, C. M.; Conrad, D.; Vining, W. J.; Meyer, T. J.
J. Chem. Soc., Chem. Commun. 1985, 1414-1416.
1. The presence of an excess amount of SCN in the reaction
solution did not increase the CO formation, but a dramatic
(37) Johnson, F. P. A.; George, M. W.; Hartl, F.; Turner, J. J. Organometallics
1996, 15, 3374-3387.
-•
increase of [1-NCS] was observed at the photostationary state
Figure 4d).
(
38) New peaks at 198.6, 194.8, 198.3, 194.2, and 191.2 ppm were observed
(
but have not yet been identified.
2028 J. AM. CHEM. SOC.
9
VOL. 130, NO. 6, 2008

Photocatalytic CO
2
Reduction Using Re(I) Complexes
A R T I C L E S
Figure 7. 13C NMR spectra of a DMF-d7-TEOA (5:1 v/v) solution
containing 1-NCS (30 mM) (a) after and (b) before 24 h of irradiation under
13
38
a
CO2 (0.72 atm) atmosphere. Integration curves are shown as red lines.
Figure 8. (a) Photocatalytic CO formation (b) by a mixed system of 1-
The spectra were recorded under the following conditions: EXMODE
nne: acquisition time 2.4 s; pulse delay 5 s; 10 000 times integration. The
solution was irradiated using a merry-go-round apparatus with a high-
pressure mercury lamp and the solution filter described in the Experimental
Section.
(Me)-Cl (0.11 mM) and 1-CN (0.11 mM) and (O) by only 1(Me)-Cl (0.11
mM) in the presence of Et4NCl (11 mM). (b) Changes in absorbance of
reaction solutions at 513 nm during irradiation; 4-mL DMF-TEOA (5:1
v/v) solutions containing (O) 1(Me)-Cl (0.11 mM) and Et4NCl (11 mM),
(b) 1(Me)-Cl (0.11 mM), 1-CN (0.11 mM), and Et4NCl (11 mM), and (4)
1
-CN (0.11 mM) and Et4NCl (11 mM) were irradiated under CO2 by 365-
(
OCHO)] were generated during the photocatalytic reaction
-9
-1
nm monochromic light of intensity 7.25 × 10 einstein s
.
using 1-NCS and 1-Cl, these complexes have only low
photocatalytic ability: ΦCO ) 0.04 for the solvento complexes,
and ΦCO ) 0.05 for the formate complex.35 This is because the
excited-state lifetimes of these species are too short to be
photochemically reduced by TEOA. Consequently, a recovering
process of the starting complex, 1-NCS or 1-Cl, is necessary
for efficient photocatalytic CO2 reduction (eq 4). Formation of
Scheme 1
-
much less than that of SCN . It is noteworthy that 1-Cl did not
-
+
form by dark reaction of Cl with fac-[Re(bpy)(CO)3(DMF)]
+
and fac-[Re(bpy)(CO)3(TEOA)] in a DMF-TEOA mixed
solution. There are two possible formation processes of 1-Cl
during the photocatalytic reaction involving 1-NCS in the
presence of Cl : (1) coordination of Cl to 1 produced by
elimination of SCN from [1-NCS] ; (2) coordination of Cl
immediately after elimination of CO from the rhenium center.
-
-
the solvento complexes should be a competitive process versus
-
-
-
-•
-
the recoordination of SCN and Cl to the rhenium complex.
Accordingly, both dissociation and recoordination abilities of
the anionic ligands strongly affect the photocatalysis.
-
•
Since [1-Cl] is too unstable to accumulate during the
photocatalytic reaction, process (1) is probably minor.
-
Figure 6a shows the effects of addition of Cl (10 mM) on
the photocatalytic reaction by 1-NCS (0.5 mM). In the first stage
of the photocatalytic reaction, 1-Cl formed but 1-NCS decreased
Figure 7 shows 13C NMR spectra of a DMF-d7-TEOA
solution containing 1-NCS before and after irradiation for 24
h. We observe two important facts: (a) CO2 is the main source
of CO, and (2) the three CO ligands are quantitatively substituted
(
see Figure 6a). After 5 h of irradiation, a photostationary state
between 1-NCS and 1-Cl was reached, in the ratio 1:3 as shown
in Figure 6b. Although further irradiation caused a gradual
decrease of the complexes, CO was formed continuously in
proportion to the total amount of 1-NCS and 1-Cl (Figure 6a).
1
3
13
with CO formed from CO2 during irradiation. GC analysis
of the gas phase in the NMR tube after irradiation shows that
TNCO was 2.5, while 25% of the 1-NCS disappeared. Lehn and
co-workers have reported similar phenomena with 1-Cl pho-
-
The addition of small amounts of SCN (0.5 and 1.0 mM)
1
1
caused a dramatic decrease of the population of 1-Cl in the
photostationary state (Figure 6b). The ratio [1-Cl]/[1-NCS] in
the photostationary state was proportional to the ratio of the
tocatalyst. The rhenium dimer with CO2 as a bridge ligand
[Re(bpy)(CO)3-C(O)O-Re(bpy)(CO)3] and the carboxylate
complex [Re(bpy)(CO)3(COOH)] have been proposed as a
precursor for CO formation. However, the experimental results
-
-
anions [Cl ]/[SCN ] (Figure S2 in the Supporting Information),
-
13
and the gradient was 0.16. Addition of excess Cl did not affect
the concentration of [1-NCS] in the photostationary state
using CO2 indicate that simple CO elimination from the bridge
-
•
CO2 or carboxylate ligand are not the main processes in the
photocatalytic reaction. These ligands should be converted to a
CO ligand before CO formation.
(
Figure S3 in the Supporting Information). These results indicate
-
that the ability of Cl to coordinate to the rhenium center is
J. AM. CHEM. SOC.
9
VOL. 130, NO. 6, 2008 2029

A R T I C L E S
Takeda et al.
Scheme 2. Photocatalytic Reaction Mechanism by 1-NCS
As described above, 1-NCS has much greater photocatalytic
ability than 1-Cl. However, elimination of the Cl anion from
reaction using 1-NCS, it is reasonable that [1-NCS]-• also acts
as an electron donor to the reaction intermediate.
-
-
•
-
[
[
1-Cl] is much more efficient than that of SCN from
We conclude that the mechanism of photocatalytic CO2
reduction with 1-NCS is as shown in Scheme 2. The lowest
-•
1-NCS] , both of which give 1, the precursor of CO2 reduction.
In addition, [1-NCS] accumulated during the photocatalytic
-
•
3
excited state MLCT of 1-NCS is reductively quenched by
-•
-•
reaction, whereas only a small amount of [1-Cl] accumulated
even in the initial stage of the photoreaction. These results
suggest that the OER of 1-NCS is not only the precursor of 1
but plays a further role in the photocatalytic reduction of CO2.
We infer that its role is as an electron donor to the CO2 adduct
which should be produced by the reaction of 1 with CO2,
because CO is the two-electron reduced product of CO2 but 1
has only a single extra electron. The following experiments
found that the OER species can function as an electron donor
for CO2 reduction.
TEOA, giving the OER species [1-NCS] . Elimination of
-
SCN from the OER species is a key step in the photocatalytic
reaction. The resulting “17-electron species”, for which a
-•
20,21
proposed structure is [Re(bpy )(CO)3] (1),
reacts with CO2
to give the CO2 adduct(s). A further key role of the OER species
of 1-NCS is expected to be electron donation to the CO2 adduct,
+
giving rise to CO, [Re(bpy)(CO)3] , and 1-NCS. The eliminated
-
anion SCN efficiently coordinates to the species [Re(bpy)-
+
(CO)3] , to recover 1-NCS.
-
The action of SCN in this process is much greater than that
A DMF-TEOA (5:1 v/v) solution containing an equimolar
amount (0.11 mM) of both fac-[Re(4,4′-Me2bpy)(CO)3Cl] (1-
of Cl- (Figure 6b). The steady-state concentration of 1-NCS
therefore remained higher than that of 1-Cl during irradiation
(Figure 5a and 5b). This is one reason why 1-NCS is a better
photocatalyst for CO2 reduction than 1-Cl. In the case of 1-CN,
the nonelimination of CN from the OER species [1-CN]
(Figure 5c) is responsible for its absence of photocatalytic
activity.
(Me)-Cl) and 1-CN was irradiated under a CO2 atmosphere.
The photocatalytic activity of the mixed system was 1.2 times
-
-•
higher than that of the system containing only 1(Me)-Cl (0.22
-•
mM) (Figure 8a). On the other hand, accumulation of [1-CN]
in the mixed system during irradiation was almost half that using
only 1-CN (Figure 8b). As described above, 1-CN cannot act
as a photocatalyst for CO2 reduction by itself. The OER species
of 1-CN is unable to reduce 1(Me)-Cl because it has a lower
redox potential (Table 1) than 1(Me)-Cl (E1/2 ) -1.77 V vs
Ag/AgNO3). We therefore expect that the mixed system should
have lower photocatalysis than 1-Cl, because of an inner-filter
effect by 1-CN. This contradiction is resolved by observing that
Development of More Efficient Photocatalytic Systems.
New architecture for constructing more efficient photocatalytic
CO2 reduction using rhenium complexes can be based on the
present mechanistic investigation in this study. The photocata-
lytic system should have the following properties:
(1) Efficient formation of the OER species by quenching of
3
the MLCT excited state by the reducing reagent.
-
•
-•
[
1-CN] produced during irradiation acts as an electron donor
(2) Effective production of [Re(LL )(CO)3] by dissociation
to the CO2 adduct made from 1(Me) and CO2 (Scheme 1).
Figure 8b shows the accumulation of the OER species during
irradiation in three photocatalytic systems: (1) a mixed system
of 1(Me)-Cl (0.11 mM) and 1-CN (0.11 mM), (2) 1-CN (0.11
mM), and (3) 1(Me)-Cl (0.11 mM). Comparison of these
systems shows that, in the mixed system, the accumulated
amount of the OER species was reduced compared with the
system containing only 1-CN, while 1-CN itself was stabilized
during the photocatalytic reaction.
of the ligand from the OER species.
(3) Efficient reduction of the CO2 adduct(s) by a further OER
species.
(4) High-yield recovery of the starting complex by recoor-
dination of a ligand after CO formation.
Based on these criteria, we sought to construct two-component
photocatalytic systems, as follows. As a catalyst for CO2
+
+
reduction, fac-[Re(bpy)(CO)3(MeCN)] ([1-MeCN] ) was used,
because it reacts with the solvent molecule to give the solvento
+
In view of the reducing power (Table 1) and accumulation
complexes, fac-[Re(bpy)(CO)3(S)] (S ) DMF and TEOA), in
-
•
of [1-NCS] in the reaction solution during the photocatalytic
which Re-solvent bonds should be weak, within several minutes
2030 J. AM. CHEM. SOC.
9
VOL. 130, NO. 6, 2008

Photocatalytic CO
2
Reduction Using Re(I) Complexes
A R T I C L E S
+
Figure 9. Dependence of ΦCO (b) on the concentration of [1-MeCN] . A
+
+
DMF-TEOA (5:1 v/v) solution containing [1(MeO)-P] and [1-MeCN] ,
of which the total concentration was adjusted to 2.5 mM, was irradiated
under a CO2 atmosphere using 365-nm monochromic light of intensity 1.11
-
9
-1
+
×
10 einstein s . The proportion of absorbance by [1(MeO)-P] at 365
nm is also shown (4).
of dissolving in a DMF-TEOA solution. The excited states have
too short lifetimes to react efficiently with free TEOA, and their
photocatalytic activity was low (ΦCO ) 0.04) in the absence of
a redox photosensitizer.35
We chose fac-[Re{4,4′-(MeO)2bpy}(CO)3{P(OEt)3}]+
+
(
[1(MeO)-P] ) as a photosensitizer, because the generation
quantum yield of the corresponding OER species, [1(MeO)-
P], is very high (ΦOER ) 1.6)39 and the reduction power of
[
1(MeO)-P] (E1/2 ) -1.67 V vs Ag/AgNO3) is high enough to
Figure 10. At left are the UV-vis spectral changes of solutions containing
+
+
+
(
(
[
a) [1(MeO)-P] (2.5 mM), (b) [1(MeO)-P] (2.5 mM) and [1-MeCN]
reduce the solvento complexes (E1/2 ) -1.57 V for the DMF
complex and -1.64 V for the TEOA complex). Although only
[
+
+
0.01 mM), (c) [1(MeO)-P] (2.4 mM) and [1-MeCN] (0.1 mM), and (d)
1(MeO)-P] (2.2 mM) and [1-MeCN]+ (0.3 mM). To the right is the
+
+
1(MeO)-P] also displayed relatively high photocatalytic
dependence of the absorbance at 500 nm on the irradiation time. Details of
the photochemical reaction are shown in Figure 9.
+
activity (ΦCO ) 0.33), the presence of [1-MeCN] increased
the quantum yield for CO formation. Figure 9 shows the
photocatalytic CO2 reduction was obtained by using a mixture
of [1(MeO)-P] and [1-MeCN] in the ratio 25:1, i.e., the
+
dependence of ΦCO on the ratio of [1-MeCN] to [1(MeO)-
+
+
+
P] when the total concentration of the rhenium complexes is
lowest overall inner-filter effects due to both [1(MeO)-P] and
2
1
.5 mM. The maximum value of ΦCO was 0.59, at a ratio of
:25.
+
[
1-MeCN] .
Figure 10 shows the accumulation of the OER species during
Conclusion
the photocatalytic reaction using complexes with different ratios.
The reaction mechanism of efficient photocatalytic CO2
reduction using fac-[Re(bpy)(CO)3(NCS)] (1-NCS) has been
determined as follows. The OER species of 1-NCS, which is
made by photochemical electron transfer from TEOA, plays two
+
Addition of [1-MeCN] led to lower accumulation of [1(MeO)-
+
P]; at greater concentrations of [1-MeCN] than 0.1 mM (Figure
1
0c and d), only very low accumulation of [1(MeO)-P] was
observed.
These results clearly show that the rate-determining process
-
important roles: elimination of the SCN ligand gives 1, which
reacts with CO2 to give the CO2 adduct(s), and the OER species
also works as an electron donor to the CO2 adduct(s), leading
in the reduction of CO2 by the mixed system is electron transfer
+
from [1(MeO)-P] to [1-MeCN] when the concentration of
-
+
to CO formation. Recombination of SCN to 1 , which should
form by CO formation, causes recovery of 1-NCS.
+
[
1-MeCN] is less than 0.1 mM under the photocatalytic
-9
reaction conditions (the light intensity was 1.11 × 10 einstein
Based on this mechanistic investigation, the most efficient
photocatalytic system for CO2 reduction in the reported
homogeneous photocatalysts has been developed, i.e., the 25:1
-1
s ). Accumulation of the OER species lowers the photocatalytic
activity of the system because of an inner-filter effect by the
OER species, which has strong absorption at the irradiation
wavelength, and decomposition of the accumulated [1(MeO)-
P] due to photoexcitation. On the other hand, too high a
+
+
mixture of [1(MeO)-P] as photosensitizer and [1-MeCN] as
catalyst, for which the quantum yield of CO formation is 0.59
-
9
-1
at an irradiated light intensity of 1.11 × 10 einstein s .
+
concentration of [1-MeCN] also lowered photocatalysis (Figure
), because of another inner-filter effect by [1-MeCN] ; direct
excitation of [1-MeCN] gives much lower photocatalysis as
+
9
Supporting Information Available: HPLC chromatograms
of the reaction solutions after the photocatalytic reaction, ratios
between 1-Cl and 1-NCS at photostationary states, and time
+
described above. It is therefore reasonable that the most efficient
-
•
profiles of [1-NCS] during the photocatalytic reaction in the
presence of excess amount of Et4NCl. This material is available
free of charge via the Internet at http://pubs.acs.org.
1
(
39) The extinction coefficient of [1(MeO)-P] at 506 nm (4900 M- cm-1) was
determined by the flow-electrolysis technique. TEOA can act as a two-
electron donor. After photochemical oxidation of TEOA, deprotonation from
its R-carbon gives an R-amino radical which has enough reducing power
to reduce the rhenium(I) complexes. This is a reason why ΦOER exceeds 1.
JA077752E
J. AM. CHEM. SOC.
9
VOL. 130, NO. 6, 2008 2031