A ruthenium-based biomimetic hydrogen cluster for efficient photocatalytic hydrogen generation from formic acid

Article abstract of DOI:10.1002/chem.201406381

A ruthenium-based biomimetic hydrogen cluster, [Ru₂(CO)₆(μ-SCH₂CH₂CH₂S)] (1), has been synthesized and, in the presence of the P ligand tri(o-tolyl)phosphine, demonstrated efficient photocatalytic hydrogen generation from formic acid decomposition. Turnover frequencies (TOFs) of 5500 h^-1 and turnover numbers (TONs) over 24 700 were obtained with less than 50 ppm of the catalyst, thus representing the highest TOFs for ruthenium complexes as well as the best efficiency for photocatalytic hydrogen production from formic acid. Moreover, 1 showed high stability with no significant degradation of the photocatalyst observed after prolonged photoirradiation at 90 C. Chemical copycat: A diruthenium complex with efficient photocatalytic hydrogen evolution from formic acid decomposition was developed by mimicking an [FeFe]-based hydrogen cluster (see figure). The catalytic activities obtained with less than 50 ppm of the catalyst amount represented the highest turnover frequency for ruthenium complexes as well as the best efficiency for photocatalytic hydrogen production from formic acid.

Full text of DOI:10.1002/chem.201406381

DOI: 10.1002/chem.201406381
Full Paper
&
Hydrogen Generation
A Ruthenium-Based Biomimetic Hydrogen Cluster for Efficient
Photocatalytic Hydrogen Generation from Formic Acid
Chin-Hao Chang, Mei-Hua Chen, Wan-Shan Du, Jacek Gliniak, Jia-Hoa Lin, Hsin-Hua Wu,
Hsin-Fang Chan, Jen-Shiang K. Yu, and Tung-Kung Wu*[a]
Abstract: A ruthenium-based biomimetic hydrogen cluster,
[Ru2(CO)6(m-SCH2CH2CH2S)] (1), has been synthesized and, in
the presence of the P ligand tri(o-tolyl)phosphine, demon-
strated efficient photocatalytic hydrogen generation from
formic acid decomposition. Turnover frequencies (TOFs) of
5500 hꢀ1 and turnover numbers (TONs) over 24700 were ob-
tained with less than 50 ppm of the catalyst, thus represent-
ing the highest TOFs for ruthenium complexes as well as the
best efficiency for photocatalytic hydrogen production from
formic acid. Moreover, 1 showed high stability with no sig-
nificant degradation of the photocatalyst observed after pro-
longed photoirradiation at 908C.
Introduction
has prompted the synthesis of model complexes that mimic
the hydrogen cluster in structure, function, and reactivity.[4] In
the past several years, significant effort has been put forth to
design and synthesize [FeFe]-based biomimetic hydrogen clus-
ters and to study the mechanism of proton reduction.[5] In
comparison to the highly efficient hydrogen cluster present in
nature, however, none or a small amount of H2 was produced
upon photoirradiation, and ultimately they complete their pho-
tochemical H2 production in organic solutions or a mixture of
organic solvents and water. More strikingly, these synthetic hy-
drogen-cluster mimics were not stable and would decompose
generally within one hour of irradiation. Thus the creation of
an artificial hydrogen-cluster system that can produce H2 with
high catalytic activity and stability in an aqueous or organic so-
lution remains a significant scientific challenge and an impor-
tant objective that is pursued to this day.
The development of clean and sustainable energy, such as hy-
drogen production driven by water splitting, has been a world-
wide goal and challenge for scientists for more than four de-
cades.[1] To date, more than 95% of generated hydrogen is
from steam reformation and coal gasification, which originates
from limited fossil resources such as coal and oil. Nevertheless,
safe and reversible hydrogen generation and storage is a signif-
icant obstacle for the utilization of hydrogen as transportation
fuel. Alternatively, formic acid, one of the main products
formed in biomass processes such as pyrolysis, fermentation,
and supercritical reactions, is a liquid at ambient conditions.
The ease of storing, transporting, and handling makes formic
acid a highly competitive and low-cost source for hydrogen
generation. Although greenhouse gas carbon dioxide is pro-
duced during hydrogen production from formic acid, recycling
is possible through natural photosynthesis or industrial tech-
nologies such as carbon capture.[2]
Various approaches, including homogeneous and heteroge-
neous systems to generate hydrogen from formic acid with
heat treatment, have been reported in recent years. For exam-
ple, Puddephatt et al. reported a binuclear ruthenium–phos-
phine complex capable of selective H2 production from formic
acid with a turnover frequency (TOF) of approximately 500 hꢀ1
at ambient temperature.[6] Fukuzumi et al. reported a heteronu-
clear iridium–ruthenium complex capable of producing hydro-
Notably, the natural iron-based hydrogenase (H2ase) family
of proteins produces hydrogen or utilizes it as an energy
source under ambient conditions. The dinuclear iron–iron cata-
lytic site, referred to as the hydrogen cluster, is responsible for
reversible reduction of protons to hydrogen, with a remarkable
efficiency of 6000–9000 molecules of hydrogen per site under
optimal conditions.[3] In spite of the high efficiency of hydro-
gen generation, the low availability and stability of the iron–
iron hydrogen cluster during light-driven hydrogen production
gen with a TOF of approximately 426 hꢀ1 [7]
Similarly, Boddien
.
et al. showed that hydrogen is generated from formic acid and
a ruthenium–phosphine catalyst system under mild conditions
and visible light.[8] Recently, Himeda investigated an iridium–
azole-containing complex for H2 generation from formic acid
in water, thereby obtaining a TOF of 34000 hꢀ1 at high temper-
ature.[9] In contrast to ruthenium–polypyridine complexes, pho-
tocatalytic hydrogen-generation catalysts based on a dirutheni-
um-based biomimetic hydrogen cluster have rarely been stud-
ied.[10]
[a] Dr. C.-H. Chang,+ M.-H. Chen,+ W.-S. Du,+ J. Gliniak, J.-H. Lin, H.-H. Wu,
H.-F. Chan, Prof. J.-S. K. Yu, Prof. T.-K. Wu
Department of Biological Science and Technology
National Chiao Tung University, 75, Po-Ai Street, Hsin-Chu (Taiwan)
E-mail: tkwmll@mail.nctu.edu.tw
[+] These authors contributed equally to this work.
The higher hydrogen-generation efficiency of ruthenium-
based complexes relative to that of [FeFe]-based hydrogen
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201406381.
Chem. Eur. J. 2015, 21, 1 – 7
1
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
clusters and the increasing demand for efficient hydrogen-gen-
eration techniques prompted our investigation of the photoca-
talytic decomposition of formic acid by an [RuRu]-based bio-
mimetic hydrogen cluster. Herein, two [RuRu]-based biomimet-
ic hydrogen clusters, [Ru2(CO)6(m-SCH2CH2CH2S)] (1) and [{(m-
H)Ru3(CO)10}2(m-SCH2CH2CH2S)] (2), were synthesized and char-
acterized by spectroscopic techniques and X-ray diffraction.[11]
Furthermore, the optical characteristics and electrochemical
properties of these ruthenium-based hydrogen clusters were
analyzed by UV/Vis spectroscopy and cyclic voltammetry. The
photocatalytic hydrogen-generation activity of 1 was evaluated
in the organic phase with various phosphine ligands
(Scheme 1).
Scheme 1. Hydrogen-generation reaction catalyzed by the ruthenium-based
biomimetic hydrogen cluster and P ligands.
Figure 1. Solid-state structure of: A) [Ru2(CO)6(m-SCH2CH2CH2S)] (1), and
B) [{(m-H)Ru3(CO)10}2(m-SCH2CH2CH2S)] (2) shown by a thermal ellipsoid plot.
Hydrogen atoms are omitted for clarity.
Results and Discussion
Reflux treatment of [Ru3(CO)12] with 1,3-propanedithiol in tolu-
ene gave two [RuRu]-based biomimetic hydrogen clusters:
a major product [Ru2(CO)6(m-SCH2CH2CH2S)] (1) and a minor
The UV/Vis absorption spectra of solutions of [Ru3(CO)12], 1,
product [{(m-H)Ru3(CO)10}2(m-SCH2CH2CH2S)] (2). The two com- and 2 were analyzed in ethyl acetate (EA) at room tempera-
pounds were characterized by mass spectrometry (MS), infra- ture. As shown in Figure 2, [Ru3(CO)12] (black line), 1 (dashed
red (IR) spectroscopy, 1H and 13C NMR spectroscopy, and line), and 2 (dotted line) exhibited characteristic absorption
single-crystal X-ray crystallography (Figures S1–S3 and Ta- bands at 254/393, 275/347, and 253/282/352/421 nm, respec-
bles S1–S4 in the Supporting Information).[11,12] Compounds tively. The electronic absorption bands of [Ru3(CO)12], 1, and 2
1 and 2 formed light yellow and yellow-orange crystals, re- observed in the UV region could be designated as p!p* tran-
spectively, from hexane and dichloromethane at ꢀ208C. The sitions and attributed to a metal-to-ligand charge-transfer
crystallographic data and select bond lengths and angles for (MLCT) transition.[14] Similarly, the broad and low-energy bands
1 and 2 are summarized in Tables S1 to S3 of the Supporting at 393, 347, and 352/421 nm observed for [Ru3(CO)12], 1, and 2,
Information. The molecular structures of 1 and 2 are depicted respectively, could be assigned to s!s* transitions in metal–
in Figure 1. Compound 1 exhibited a butterfly-like geometry of metal d orbitals.[14b,c] The electrochemical properties of 1 and 2
the Ru2S2 moiety, and each Ru atom was coordinated with were analyzed by cyclic voltammetry (CV) and differential
three linear carbonyl groups and two bridging sulfur ligands to pulse voltammetry (DPV; Figure S4 in the Supporting Informa-
form a “sawhorse” arrangement. The RuꢀRu distance of tion). The redox potentials were measured with an Ag/AgNO3
1
(2.6592(6) ꢁ) is longer than the FeꢀFe distance of electrode. Compound 1 displayed two reversible oxidation po-
[Fe2(CO)6(m-SCH2CH2CH2S)] (2.5103(11) ꢁ), but shorter than the tentials and one reversible reduction potential at +0.772,
RuꢀRu distance of [Ru3(CO)12] (2.854 ꢁ).[13] The structure of 2 +1.676, and ꢀ1.856 V, attributed to RuIRuI to RuIRuII, RuIRuII to
shows that two [Ru3(CO)12] groups were connected by a pro- RuIIRuII, and RuIRuI to RuIRu0, respectively. Because the standard
panedithiolate ligand with the loss of two carbonyl groups. potential of Ag/AgNO3 relative to the normal hydrogen elec-
The average RuꢀRu distances of 2 (2.8315 ꢁ) is slightly shorter trode (NHE) is +0.56 V, the oxidation and reduction potentials
than that of [Ru3(CO)12] (2.854 ꢁ) but is longer than that of were recalculated to be +1.282, +2.236, and ꢀ1.296 V, respec-
1 (2.6592(6) ꢁ). This result indicates that the RuꢀRu distance in tively. Compound 2 displayed two reversible oxidation poten-
1 is affected by the nature of the terminal ligand. The bridging tials and one reversible reduction potential at +0.716, +1.312,
RuꢀS distance of 1, which ranges from 2.3569(13) to and ꢀ1.55 V, respectively. The oxidation and reduction poten-
2.3650(14) ꢁ, is also longer than the bridging FeꢀS distance of tials relative to NHE were recalculated to be +1.276, +1.872,
[Fe2(CO)6(m-SCH2CH2CH2S)] (2.2491(10) to 2.2542(10) ꢁ). The and ꢀ0.992 V, respectively. The oxidation and reduction poten-
bridging RuꢀS distance of 2 (range from 2.384(4) to 2.398(4) ꢁ) tials could be attributed to the loss of two electrons and re-
is also longer than that of 1 (2.3569(13) to 2.3650(14) ꢁ).
duction in the triruthenium catalytic center, respectively.
&
&
Chem. Eur. J. 2015, 21, 1 – 7
www.chemeurj.org
2
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Figure 2. UV/Vis absorption spectrum of [Ru3(CO)12] (black line), 1 (dashed
line), and 2 (dotted line).
We started our photocatalytic hydrogen-generation studies
with 1 in combination with different phosphorous-containing
ligands. Compound 2 was omitted from photocatalytic hydro-
gen-generation experiments owing to insufficient amounts of
compound. The preliminary reactions were performed with
catalyst (20 mmol) in dimethylformamide (DMF; 1 mL) and
formic acid/triethylamine (FA/TEA, 5:2; 4 mL) under a 500 W Xe
lamp and kept at 408C for six hours. Subsequently, a number
of P ligands with different electronic or steric properties, in-
cluding triphenylphosphine (PPh3, 3), tri(o-tolyl)phosphine (P(o-
tol)3, 4), tri(m-tolyl)phosphine (P(m-tol)3, 5), tri(p-tolyl)phos-
phine P(p-tol)3, (6), tris(2-methoxyphenyl)phosphine (P(o-
Figure 3. Photocatalytic hydrogen-generation efficiency of 1 as catalyst in
the absence and presence of various P ligands including PPh3 (3), P(o-tol)3
(4), P(m-tol)3 (5), P(p-tol)3 (6), P(o-C6H4OCH3)3 (7), P(p-C6H4OCH3)3 (8), P(p-
C6H4F)3 (9), and P(p-C6H4CF3)3 (10).
ligands tested, the electron-withdrawing P ligand such as 10
showed a low activity comparable to that of reactions per-
formed in the absence of P ligand. The addition of a P ligand
such as 4 provided the highest activity, with the efficiency of P
ligands decreasing in the order of 4>7>8ꢁ6>5>3>9>10.
We next evaluated the influence of light, catalyst-to-sub-
strate ratio, and temperature on H2-generation efficiency
(Table 1). When compound 1 was heated at 40 or 608C in the
presence of P ligand 4 without photoirradiation for six hours,
a slight amount of hydrogen was generated and detected at
a turnover number (TON) of 47.4 and 61.7 and a TOF of 8.0
and 10.3 hꢀ1, respectively. Interestingly, when compound 1 was
photoirradiated without temperature control in the presence
of P ligand 4, a higher TON of 2149.4 and TOF of 358.2 hꢀ1 was
detected, thus indicating that light has a strong enhancing
effect on the H2-generation efficiency.
C6H4OCH3)3,
7),
tris(4-methoxyphenyl)phosphine
(P(p-
C6H4OCH3)3, 8), tris(4-fluorophenyl)phosphine (P(p-C6H4F)3, 9),
and tris(4-trifluoromethylphenyl)phosphine (P(p-C6H4CF3)3, 10),
were investigated for their ability to increase hydrogen-genera-
tion efficiency. Figure 3 shows H2 production at 408C over time
using 1 as the catalyst with the addition of different P ligands.
As expected, only gas mixtures that contained H2/CO2 and not
CO were detected in cases of significant gas production. Con-
trol experiments in the absence of catalysts showed no detect-
able production of H2, thus verifying that the catalyst is essen-
tial for the photocatalytic generation of H2. Reaction with 1 in
the absence of the P ligand gave very slight activity. The addi-
tion of electron-donating P ligands provided higher activity
than addition of electron-withdrawing P ligands. Among the P
To determine the optimal catalyst-to-substrate ratio, the cat-
alyst was photoirradiated under a 500 W Xe lamp and kept at
608C for six hours at concentrations of 40, 20, 10, 5, or 2 mmol
in DMF (1 mL) and FA/TEA (5:2; 4 mL). Interestingly, the system
with 20 mmol of catalyst generated a higher total amount of
Table 1. Influence of light, catalyst-to-substrate ratio, and temperature of the ruthenium complex 1 and P ligand 4 on hydrogen-generation efficiency.
Entry
Amount
[mmol]
T
[8C]
TON 1 h
(TOF [hꢀ1])
TON 2 h
(TOF [hꢀ1])
TON 3 h
(TOF [hꢀ1])
TON 4 h
(TOF [hꢀ1])
TON 5 h
(TOF [hꢀ1])
TON 6 h
(TOF [hꢀ1])
Conversion
[%]
1[a]
2[a]
3[a]
4[a]
5[a]
6[a]
7[a]
8[a]
40
20
10
5
2
2
60
60
60
60
60
40
80
90
42 (42)
150 (75)
584 (292)
280 (93)
957 (319)
449 (112)
552 (110)
625 (104)
50 (6 h)
70 (6 h)
66 (6 h)
56 (6 h)
28 (6 h)
10 (6 h)
90 (6 h)
98 (5 h)
227 (227)
184 (184)
522 (522)
308 (308)
242 (242)
3655 (3655)
4705 (4705)
1189 (297)
2431 (608)
4074 (1019)
4654 (1164)
1879 (470)
16117 (4029)
21004 (5251)
1469 (294)
2810 (562)
5105 (1021)
6142 (1228)
2128 (426)
19328 (3866)
24700 (4940)
1748 (291)
3276 (546)
5626 (938)
7492 (1249)
2563 (427)
22428 (3738)
1035 (518)
1848 (924)
2075 (1038)
845 (422)
8212 (4106)
10672 (5336)
1894 (631)
2910 (970)
3248 (1083)
1330 (443)
13194 (4398)
16595 (5532)
2
2
[a] For entries 1–8, the reaction mixtures were photoirradiated in 4 mL of 5HCO2H·2NEt3 and 1 mL of DMF by using a 500 W xenon lamp.
Chem. Eur. J. 2015, 21, 1 – 7
www.chemeurj.org
3
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
hydrogen than that with 40 mmol of catalyst. In addition, the
TOF activity is reversely proportional to the concentration of
catalyst (Table 1), whereby a higher TOF is observed in
2 mmolꢂ5 mmol>10 mmol>20 mmol>40 mmol. In the subse-
quent evaluation of the temperature effect, 2 mmol of com-
pound 1 (50 ppm) and P ligand 4 in DMF (1 mL) and 4 mL of
FA/TEA (5:2) was photoirradiated under a 500 W Xe lamp for
five or six hours at 40, 60, 80, and 908C. The hydrogen-genera-
tion efficiency was observed to be proportional to the temper-
ature increase with the highest efficiency observed at 908C. As
shown in Table 1, the highest TON and TOF of 24700 and
5500 hꢀ1, fourfold relative to that measured at 608C, was ob-
served for the system photoirradiated at 908C for 5 h (Table 1,
entries 5 and 8). In lieu with this finding, the hydrogen yield
calculated for formic acid dehydrogenation by 1 was close to
98%. These results indicate that H2-generation efficiency is de-
pendent on both light and temperature. Furthermore, approxi-
mately 0.4 L of H2 per hour was generated from the system,
thus suggesting the feasibility of the direct application of the
system for H2 generation. This is the highest catalyst productiv-
ity known to date for any homogeneous hydrogen generation
from formic acid with an [RuRu]-based biomimetic hydrogen
cluster.
eration increase was observed during the irradiation period
but not between two irradiation periods. Consistent with this
observation is that when different wavelengths ((280 nm
cutoff, 280–397 nm (T=346 nm), 301–384 nm (T=352 nm),
355–469 nm (T=397 nm)) were tested for hydrogen-genera-
tion efficiency, the percentage of formic acid consumed in six
hours was 100, 92, 60, 9% for the 280, 280–397, 301–384, and
355–469 nm wavelengths of light cutoffs, respectively. Further-
more, by calculating the total number of moles of photon ab-
sorbed at 355 nm and the net number of moles of hydrogen
generated by photoirradiation, the measured quantum yield is
about 0.331. These results indicate the photoinduced and pho-
toassisted effects on hydrogen generation.
To investigate the stability of the catalyst, the system was
photoirradiated and kept at 908C for 24 h (Figure 4). During
the reaction, 0.8 mL of FA/TEA (5:2) per hour was added to the
reaction vessel to maintain the substrate concentration. Hydro-
gen-generation efficiency remained linear for approximately
ten hours before a deceleration occurred. A total turnover
number of 44102 in 24 h and an averaged turnover frequency
of 1838 hꢀ1 was achieved. To the best of our knowledge, this
represents the highest stability of the catalyst and the TON
ever reported for photocatalytic formic acid decomposition.
The photoinduced and photoassisted effects were investi-
gated by performing a series of experiments in which the tem-
perature was kept at 608C and the light was turned on and off
at a constant time interval. As shown in Figure 5, only a slight
increase in hydrogen production was observed during the first
hour in the absence of light. Then, an apparent hydrogen-gen-
Figure 5. The photoinduced and photoassisted effect of hydrogen genera-
tion was investigated with 1 (2 mmol)+4 (2 mmol) in DMF (1 mL)+FA/TEA
(4 mL, 5:2). The solution was heated at 608C for 1 h before the light was
switched on and off every 1 h. Gas (500 mL) was harvested for every 15 min
by a gas-tight syringe and analyzed by means of GC-TCD.
To identify active catalyst species, we performed time-de-
pendent 31P NMR spectroscopic analyses of 1+4+FA/TEA+hn
(1 h; Figure 6). The 31P NMR spectroscopic signal observed at
d=ꢀ30.1 ppm corresponds to the P ligand 4 before photoirra-
diation. In the absence of FA/TEA and before photoirradiation,
one small singlet at d=39.8 ppm, which corresponds to the P
ligand coordinated to the metal center to form a complex, was
detected. After photoirradiation for 30 min, two singlets at d=
37.2 and ꢀ21.1 ppm were observed in the 31P NMR spectrum,
with concordant diminishing of the singlet at d=39.8 ppm
and decrease in the free P ligand 4 at d=ꢀ30.1 ppm. The sin-
glet at d=37.2 ppm corresponds to protonated tri(o-tolyl)-
phosphine oxide, which was also observed after photoirradia-
tion of 4+FA/TEA (data not shown). After one hour of photoir-
radiation, the signal at d=ꢀ21.1 ppm increased and a new sin-
glet at d=ꢀ26.1 ppm appeared. The intensity of the signal at
d=37.2 ppm remained unchanged, whereas the signals at d=
ꢀ21.1 and ꢀ26.1 ppm increased with prolonged photoirradia-
tion, thus indicating an increase in P ligands coordinated to
the metal center and subsequently treated with FA. Consistent
with this observation is the acceleration of hydrogen genera-
tion after one hour of photoirradiation. These results also sug-
Figure 4. Stability of ruthenium complex 1 and P ligand 4 for continuous hy-
drogen generation.
&
&
Chem. Eur. J. 2015, 21, 1 – 7
www.chemeurj.org
4
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
tivated to generate a new species (II). The carbonyl group was
then replaced with a formate (III) or a hydride (IV) to result in
the shift of the singlet at d=39.8 to ꢀ21.1 ppm and the newly
generated singlet signal at d=ꢀ26.1 ppm. Protonation by FA
resulted in generation of H2, the release of CO2, and the regen-
eration of the active species III and IV, which was supported
by the detection of the 31P NMR spectroscopic signals from
prolonged photoirradiation reactions.
Conclusion
Figure 6. 31P NMR spectra of the time courses of the 1+4+FA/TEA+hn
(0 min, 30 min, 1 h) catalyst solution.
In conclusion, we have utilized a novel ruthenium-based bio-
mimetic hydrogen cluster for efficient photocatalytic H2 gener-
ation from formic acid decomposition. The addition of elec-
tron-donating P ligands provides higher activity than electron-
withdrawing P ligands, from which the tri(o-tolyl)phosphine ex-
hibits the highest activity. Remarkably, the highest TON and
TOF of 24700 and 5500 hꢀ1, respectively, and up to 0.4 L of H2
per hour could be generated from a 5 mL reaction solution
that contained 50 ppm of catalyst-to-substrate ratio. In addi-
tion, the catalyst is highly stable and shows constant catalytic
activity under prolonged photoirradiation at 908C; it is puta-
tively feasible that the system could be directly applied for hy-
drogen generation. Accompanying light on and off experi-
ments and 31P NMR spectroscopic studies suggest that the H2
generation reaction is photoinduced and photoassisted, and in
the presence of FA/TEA the binding of the P ligand to 1 is pre-
ferred over its own photo-oxidation. Finally, the overall activity
of the [RuRu]-based biomimetic hydrogen cluster is more than
20-fold higher than that of [FeFe]-based hydrogen clusters.[15]
In further studies we will focus on improving the activity of co-
valent incorporation of P ligands to the metal center.
gest that in the presence of FA/TEA the binding of P ligands to
the ruthenium complex is preferred over the photo-oxidation.
In parallel, after photoirradiation for six hours and comple-
tion of the reaction, the reaction solution was subjected to
column chromatography for product characterization. The iso-
lated product was characterized by both 31P NMR spectroscopy
and X-ray crystallography. The only isolated product is the
[Ru2(CO)5(m-SCH2CH2CH2S)(P(o-tol)3)] complex, which has been
characterized by X-ray crystallography previously in our labora-
tory (data not shown). In addition, when the isolated product
was subjected to photocatalytic hydrogen generation, both
TON and TOF similar to those of 1+4 were obtained (Figure S5
in the Supporting Information). These results showed that no
decomposition product could be detected after six hours of
photoirradiation, which indicates that all the formic acid was
completely decomposed to H2 and CO2.
On the basis of the results described above and the catalytic
cycle described for the [FeFe]-
based hydrogen cluster, a pro-
posed mechanism for the photo-
generation of H2 by an [RuRu]-
based biomimetic hydrogen
cluster and P ligands can be put
forth, as in Scheme 2.[15] In the
absence of FA/TEA and before
photoirradiation, the addition of
the P ligand to the [RuRu]-based
hydrogen cluster generated
a small amount of ruthenium/P-
ligand complex (I). Upon photo-
irradiation and in the presence
of FA/TEA, a minor amount of
free P ligand was initially oxi-
dized to form inactive tri(o-tolyl)-
phosphine oxide and the subse-
quent
protonated
product,
which showed a singlet at d=
37.2 ppm. Simultaneously, the
carbonyl group of the rutheni-
um/P-ligand complex (I) was ac- Scheme 2. Proposed hydrogen-generation reaction catalyzed by 1 and P ligand 4.
Chem. Eur. J. 2015, 21, 1 – 7 www.chemeurj.org ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5
&
&
These are not the final page numbers! ÞÞ

Full Paper
009-017-MY3. We are grateful to Prof. Chien-Chung Cheng and
Prof. Hsiu-Fu Hsu for helpful discussion.
Experimental Section
Synthesis and spectroscopic characterization of [Ru2(CO)6(m-
SCH2CH2CH2S)] and [{(m-H)Ru3(CO)10}2(m-SCH2CH2CH2S)]
Keywords: biomimetic synthesis · hydrogen · P ligands ·
photocatalysis · ruthenium
A suspension of [Ru3(CO)12] (0.32 mmol, 204 mg) in toluene (80 mL)
was treated with 1,3-propanedithiol (2 equiv) under argon. The re-
action mixture was stirred at 808C until its color changed from
yellow-orange to deep orange. The solvent and an excess amount
of 1,3-propanedithiol were removed under vacuum, and the resi-
due was purified through a silica gel column eluting with 100%
hexane and hexane/CH2Cl2 (98:2 v/v) to develop two bands. The
two compounds were characterized by spectroscopic data and X-
ray diffraction and compared to previously published literature re-
ports.[10–12]
[1] A. Fujishima, K. Honda, Nature 1972, 238, 37–38.
[2] a) I. Ntaikou, H. N. Gavala, M. Kornaros, G. Lyberatos, Int. J. Hydrogen
Energy 2008, 33, 1153; b) R. S. Haszeldine, Science 2009, 325, 1647–
1652.
[3] a) B. H. Huynh, M. H. Czechowski, H. J. Krꢃger, D. V. DerVartanian, H. D. J.
Peck, J. LeGall, Proc. Natl. Acad. Sci. USA 1984, 81, 3728–3732; b) M. W.
Adams, Biochim. Biophys. Acta Bioenerg. 1990, 1020, 115–145; c) M. W.
Adams, E. I. Stiefel, Curr. Opin. Chem. Biol. 2000, 4, 214–220.
[4] a) J. W. Peters, W. N. Lanzilotta, B. J. Lemon, L. C. Seefeldt, Science 1998,
282, 1853–1858; b) Y. Nicolet, C. Piras, P. Legrand, C. E. Hatchikian, J. C.
Fontecilla-Camps, Structure 1999, 7, 13–23.
[5] a) Y. Na, J. Pan, M. Wang, L. Sun, Inorg. Chem. 2007, 46, 3813–3815;
b) Y. Na, M. Wang, J. Pan, P. Zhang, B. Akermark, L. Sun, Inorg. Chem.
2008, 47, 2805–2810.
[6] a) Y. Gao, J. K. Kuncheria, R. J. Puddeph, G. P. A. Yap, Chem. Commun.
1998, 2365–2366; b) Y. Gao, J. K. Kuncheria, H. A. Jenkins, R. J. Pudde-
phatt, G. P. A. Yap, J. Chem. Soc. Dalton Trans. 2000, 3212–3217.
[7] S. Fukuzumi, T. Kobayashi, T. Suenobu, J. Am. Chem. Soc. 2010, 132,
1496–1497.
Spectroscopic and crystal structure determination of
[Ru2(CO)6(m-SCH2CH2CH2S)] and [{(m-H)Ru3(CO)10}2(m-
SCH2CH2CH2S)]
[Ru2(CO)6(m-SCH2CH2CH2S)] was crystallized as yellow crystals from
hexane and CH2Cl2 at ꢀ208C, and [{(m-H)Ru3(CO)10}2(m-
SCH2CH2CH2S)] was crystallized as yellow-orange crystals also with
hexane and CH2Cl2 at ꢀ208C. Each crystal was formed onto the
end of a thin glass tube and was measured by using single-crystal
X-ray determination. The crystal structures were analyzed and con-
firmed by the instrumentation at the National Taiwan Normal Uni-
versity. Crystallographic data were recorded at 200 K using
a Bruker Nonius Kappa CCD X-ray diffraction.
[8] B. Loges, A. Boddien, H. Junge, J. R. Noyes, W. Baumann, M. Beller,
Chem. Commun. 2009, 4185–4187.
[9] a) Y. Himeda, Green Chem. 2009, 11, 2018–2022; b) Y. Manaka, W. H.
Wang, Y. Suna, H. kambayashi, J. T. Muckerman, E. Fujita, Y. Himeda,
Catal. Sci. Technol. 2014, 4, 34–37.
Photocatalytic hydrogen generation
[10] T. E. Bitterwolf, J. A. Cabeza, J. Organomet. Chem. 2004, 689, 2947–
2951.
Formic acid was used as a proton source. A quartz Schlenk tube
that contained the photocatalyst/ruthenium-based hydrogen-clus-
ter/[Ru3(CO)12] (20 mmol), phosphine ligands (20 mmol) in DMF
(1 mL), and proton source (4 mL) aliquot of 5HCO2H·2NEt3 (FA/
TEA) was degassed under pure Ar atmosphere. The reaction solu-
tion was stirred and irradiated at 608C using a Xe lamp (500 W)
without a filter. For each hour, 500 mL of the gas in the headspace
of the sample tube was harvested by means of a gas-tight syringe
and analyzed using a Bruker 450 GC gas chromatograph instru-
ment equipped with a Carboxen 1010 PLOT-fused silica capillary
column (Supelco, 30 mꢂ0.53 mm) in 408C oven temperature,
a thermal conductivity detector (TCD), and with Ar as carrying gas
to measure the volume of H2. In addition, the amount of hydrogen
was determined and calculated by an external standard method.
[11] a) K. M. Hanif, S. E. Kabir, M. A. Mottalib, M. B. Hursthouse, K. M. A. Malik,
E. Rosenberg, Polyhedron 2000, 19, 1073–1080; b) S. E. Kabir, K. M. A.
malik, E. Rosenberg, T. A. Siddiquee, Inorg. Chem. Commun. 2000, 3,
140–142; c) A. K. Justice, R. C. Linck, T. B. Rauchfuss, S. R. Wilson, J. Am.
Chem. Soc. 2004, 126, 13214–13215; d) A. K. Justice, R. C. Linck, T. B.
Rauchfuss, Inorg. Chem. 2006, 45, 2406–2412.
[12] a) F. A. Cotton, C. S. Kraihanzel, J. Am. Chem. Soc. 1962, 84, 4432–4438;
b) R. Hoffmann, M. M. L. Chen, D. L. Thorn, Inorg. Chem. 1977, 16, 503–
511.
[13] M. R. Churchill, F. J. Hollander, J. P. Hutchinson, Inorg. chem. 1977, 16,
2655–2659.
[14] a) R. A. Levenson, H. B. Gray, J. Am. Chem. Soc. 1975, 97, 6042–6047;
b) D. R. Tyler, R. A. Levenson, H. B. Gray, J. Am. Chem. Soc. 1978, 100,
7888–7893; c) J. G. Bentsen, M. S. Wrighton, J. Am. Chem. Soc. 1987,
109, 4530–4544.
[15] A. Boddien, B. Loges, F. Gꢄrtner, C. Torborg, K. Fumino, H. Junge, R.
Ludwig, M. Beller, J. Am. Chem. Soc. 2010, 132, 8924–8934.
Acknowledgements
Received: December 8, 2014
Revised: February 1, 2015
Published online on && &&, 0000
We thank the National Science Council of the Republic of
China for financial support under contract no. NSC-102-2113M-
&
&
Chem. Eur. J. 2015, 21, 1 – 7
www.chemeurj.org
6
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
FULL PAPER
Chemical copycat: A diruthenium com-
plex with efficient photocatalytic hydro-
gen evolution from formic acid decom-
position was developed by mimicking
an [FeFe]-based hydrogen cluster (see
figure). The catalytic activities obtained
with less than 50 ppm of the catalyst
amount represented the highest turn-
over frequency for ruthenium com-
plexes as well as the best efficiency for
photocatalytic hydrogen production
from formic acid.
&
Hydrogen Generation
C.-H. Chang, M.-H. Chen, W.-S. Du,
J. Gliniak, J.-H. Lin, H.-H. Wu, H.-F. Chan,
J.-S. K. Yu, T.-K. Wu*
&& – &&
A Ruthenium-Based Biomimetic
Hydrogen Cluster for Efficient
Photocatalytic Hydrogen Generation
from Formic Acid
Chem. Eur. J. 2015, 21, 1 – 7
www.chemeurj.org
7
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ