Visible-light-driven hydrogen evolution from water using a noble-metal-free polyoxometalate catalyst

Article abstract of DOI:10.1016/j.jcat.2013.06.028

In an effort to address the need to develop hydrolytically more stable, molecular water reduction catalysts (WRCs) amenable to in-depth investigation, we report here one prototype: a tetra-manganese-containing V-centered polyoxotungstate, Na₁₀[Mn₄(H₂O) ₂(VW₉O₃₄)₂] (1). The electronic structure of 1 was elucidated using the DFT approach. Complex 1 is readily prepared by a one-pot procedure in aqueous solution and catalyzes the reduction of water using visible light irradiation (λ = 455 nm) with [Ru(bpy) ₃]²⁺ and triethanolamine (TEOA) as a photosensitizer and sacrificial electron donor, respectively. Upon irradiation, the excited state [Ru(bpy)₃]^2+* is oxidatively quenched by 1, as confirmed by steady-state and time-resolved fluorescence decay studies, to form [Ru(bpy)₃]³⁺ and a reduced form of the catalyst. The [Ru(bpy)₃]²⁺ is rapidly regenerated by reaction with TEOA. The reduced form of the catalyst, 1, reacts with water to generate hydrogen. Isotope labeling experiments demonstrate that the hydrogen comes from water. The stability of the catalyst was assessed using different spectroscopic methods. A mechanism based on experimental results is proposed.

Full text of DOI:10.1016/j.jcat.2013.06.028

Journal of Catalysis 307 (2013) 48–54
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Visible-light-driven hydrogen evolution from water using a
noble-metal-free polyoxometalate catalyst
Hongjin Lv a, Jie Song a, Haiming Zhu a, Yurii V. Geletii a, John Bacsa b, Chongchao Zhao a, Tianquan Lian a,
Djamaladdin G. Musaev c, Craig L. Hill a,
⇑
a Department of Chemistry, Emory University, 1515 Dickey Dr., Atlanta, GA 30322, USA
b X-ray Crystallography Center, Emory University, 1515 Dickey Dr., Atlanta, GA 30322, USA
c Cherry L. Emerson Center for Scientific Computation, Emory University, 1515 Dickey Dr., Atlanta, GA 30322, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
In an effort to address the need to develop hydrolytically more stable, molecular water reduction cata-
lysts (WRCs) amenable to in-depth investigation, we report here one prototype: a tetra-manganese-con-
taining V-centered polyoxotungstate, Na10[Mn4(H2O)2(VW9O34)2] (1). The electronic structure of 1 was
elucidated using the DFT approach. Complex 1 is readily prepared by a one-pot procedure in aqueous
solution and catalyzes the reduction of water using visible light irradiation (k = 455 nm) with
[Ru(bpy)3]2+ and triethanolamine (TEOA) as a photosensitizer and sacrificial electron donor, respectively.
Upon irradiation, the excited state [Ru(bpy)3]2+ꢀ is oxidatively quenched by 1, as confirmed by steady-
state and time-resolved fluorescence decay studies, to form [Ru(bpy)3]3+ and a reduced form of the cat-
alyst. The [Ru(bpy)3]2+ is rapidly regenerated by reaction with TEOA. The reduced form of the catalyst, 1,
reacts with water to generate hydrogen. Isotope labeling experiments demonstrate that the hydrogen
comes from water. The stability of the catalyst was assessed using different spectroscopic methods. A
mechanism based on experimental results is proposed.
Received 6 May 2013
Revised 26 June 2013
Accepted 26 June 2013
Keywords:
Visible-light-driven hydrogen evolution
Polyoxometalate
Noble-metal-free photocatalyst
Steady-state and time-resolved fluorescence
Isotope labeling
Ó 2013 Elsevier Inc. All rights reserved.
1. Introduction
the catalyst [11,12,18–20], with a separate chromophore [21–
25]; systems based exclusively on earth-abundant elements are
Currently, the conversion of sunlight energy into chemical en-
ergy in the form of convenient, sustainable, and non-fossil-based
fuels such as dihydrogen is the focus of numerous studies [1,2].
The general principle of water splitting into H2 and O2 is adapted
from Nature, where reductive equivalents are generated in Photo-
systems I and II producing O2 as a by-product. The promise of a
low-cost, sustainable, environmentally benign fuel using sunlight
as the energy source and water as the reducing agent is motivating
a large international community to develop novel and effective
heterogeneous [3–8] or homogeneous [5,9–13] structures or de-
vices to facilitate water splitting.
Photocatalysis based on dispersion of heterogeneous semicon-
ductor particles is one promising approach for solar water splitting
because such particles are robust under some conditions and read-
ily prepared at low cost [3,5,14,15]. To fully understand the struc-
tures and dynamics of solar-driven water-splitting architectures as
well as optimizing their performance, it is valuable to have molec-
ular systems that facilitate detailed experimental and computa-
tional studies [16,17]. Most of systems employ noble metals as
very rare [26–28]. In this context, the development of catalysts that
exhibit the advantages of both heterogeneous semiconductor cata-
lysts and homogeneous molecular catalysts are of clear and grow-
ing interest. Polyoxometalates (POMs), showing similar
compositions to semiconductor metal oxides, are a large family
of transition metal oxygen anion clusters with d0 electronic config-
urations. POMs combine the stability advantages of heterogeneous
catalysts with the tractability and other advantages of homoge-
neous catalysts. While there are many very recent reports of tran-
sition-metal-substituted POMs catalyzing the efficient oxidation of
water under thermal and photochemical conditions [29–36], there
are very few reports of such complexes catalyzing the reduction of
water [37–41]. We and others have studied hydrogen evolution
from reduced POMs, usually in context with UV-light-driven pho-
tooxidation of organic substrates (alcohols, alkanes, and others),
but all these studies involve conventional (non-d-metal-contain-
ing) POMs [42–46]. Many of these H2 evolution studies required
the presence of Pt(0) as a co-catalyst to achieve rates amenable
to detailed studies. Recently, Artero and coworkers [47] developed
a covalent-linked Ir(III)-photosensitized polyoxometalate complex,
which shows very efficient photoreduction of polyoxometalate in
the presence of sacrificial electron donor upon visible light
⇑
Corresponding author.
E-mail address: chill@emory.edu (C.L. Hill).
0021-9517/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2013.06.028

H. Lv et al. / Journal of Catalysis 307 (2013) 48–54
49
irradiation. This system performs photocatalytic H2 production
without obvious loss of activity over more than 1 week; however,
a TON of only 41 is obtained after 7 days of irradiation [47]. Clearly,
the development of an inexpensive and efficient WRC that is also
water compatible would be a breakthrough in artificial photosyn-
thesis. In this context, we report a Mn-substituted polyoxometa-
late, [Mn4(H2O)2(VW9O34)2]10ꢁ (1), that is isostructural to the
efficient water oxidation catalyst, [Co4(H2O)2(PW9O34)2]10ꢁ. While
1 is inactive as a water oxidation catalyst, it is catalytically active
for water reduction. We describe the reduction of water catalyzed
by 1 under visible light irradiation using [Ru(bpy)3]2+ as photosen-
sitizer with TEOA as a sacrificial electron donor.
dark block crystals of Na1 (yield 2.45 g; ca. 70% based on tung-
state). One single crystal was submitted to structural analysis by
X-ray crystallography and the bulk sample was analyzed by ele-
mental analysis. Elemental Analysis, Calcd. (Found%) for Na101:
calc for Mn, 4.03; V, 1.87; W, 60.68; found for Mn, 4.01; V, 1.81;
W, 59.89. FT-IR (2% KBr pellet, 1000–400 cmꢁ1): 951(m), 874(s),
825(s), 752(m), 698(s), 490(m). 51V NMR: ꢁ505.2 ppm,
Dm1/2 =
73.7 Hz. Thermogravimetric analysis (TGA): weight loss 8.5%,
corresponding to 26 water molecules. Molecular weight:
5453.04 g molꢁ1
.
The tetrabutylammonium (TBA) salt was prepared according to
general literature procedure of Katsoulis and Pope with some mod-
ifications [50]. Typically, an aqueous solution of Na1 (0.1 mmol,
0.545 g) in 10 mL H2O was added to a solution of tetrabutylammo-
nium bromide (15 mmol, 0.483 g) in CH2Cl2. The mixture was then
shaken heavily to transfer the polyanions to the organic layer. The
dark brown organic layer was separated and washed with deion-
ized water (15 mL ꢄ 6 times) to remove excess TBA bromide. Crys-
talline material was obtained by dissolving the TBA salt in 10 mL
acetonitrile and allowing diethyl ether vapor to diffuse into the
solution. The product was then dried in vacuo. The purity was con-
firmed by FT-IR, 51V NMR, and EA.
2. Experimental
2.1. Materials and Instrumentation
All reagents and solvents were purchased from commercial
sources unless otherwise noted. [Ru(bpy)3]Cl2ꢂ6H2O was purified
before use by recrystallizing from 5 mL of warm water; this
[Ru(bpy)3](ClO4)3 was prepared by literature procedure with minor
modifications [48]. All other chemicals and salts were used as re-
ceived without further purification unless otherwise noted. Water
for the preparation of solutions was obtained from a Barnstead
NanopureÒ water-purification system. D2O and CDCl3 for isotope
labeling experiment or NMR studies were obtained from Cam-
bridge Isotope Inc.
2.3. X-ray crystallography
The title compound Na101 was isolated as orange–brown crys-
tals. A complete set of diffraction intensities was collected at the
X-ray Crystallography facility at Emory University. A single crystal
0.23 ꢄ 0.14 ꢄ 0.1 mm3 with high-order faces was used. Single crys-
tal X-ray data were collected at 173(2) K on a Bruker APEX2 dif-
The FT-IR spectra were measured on a Thermo Nicolet 6700
spectrometer. UV–Vis spectra were acquired using Agilent 8453
spectrophotometer equipped with a diode-array detector and an
Agilent 89090A cell temperature controller unit. 51V NMR
(151.6 MHz) spectra were obtained at 298 K in 5 mm O.D. NMR
tubes on a Unity Plus 600 spectrometer equipped with a Varian
600 SW/PF6 probe head. All the chemical shifts were referenced
to neat VOCl3 (reference as 0 ppm at 25 °C). Elemental analyses
were performed by Galbraith Lab Inc., Knoxville, TN, 37921. Ther-
mogravimetric data were collected on Instrument Specialists
Incorporated TGA 1000 instruments. Analysis of hydrogen in the
reaction headspace was performed using a HP7890A model gas
chromatograph equipped with thermal conductivity detector
(TCD) and a GC column packed with 5 Å molecular sieves. The stea-
dy-state luminescence quenching spectra were recorded on a Flu-
oroMax 3 spectrofluorimeter. For time-resolved fluorescence decay
measurements, femtosecond laser pulses (ꢃ100 fs, 80 MHz repeti-
tion rate) were generated with a mode-locked Ti:sapphire laser
(Tsunami oscillator pumped by 10 W Millennia Pro, Spectra-Phys-
ics). Excitation pulses at 400 nm were generated by second har-
monic generation of the 800 nm pulses in a BBO crystal. The
repetition rate of output pulse centered at 800 nm was reduced
to 1.6 MHz using a pulse picker (Conoptics, USA). The 580–
fractometer with graphite-monochromated Mo K
a (0.71073 Å)
radiation. The data were collected using scans with different
x
u
values and optimal frame exposure times and widths yielding
43,235 reflections in the h range 1.37–29.15°, of which 12,472 were
unique. The strategy for the data collection, cell parameters, and
symmetry were evaluated using APEXII software [51]. The frames
were integrated with the SAINT v7.68a [52]. The distances of the
faces from the center of the crystal were measured for a numerical
absorption correction. A combination of a numerical and a multi-
scan absorption correction was carried out using the program SAD-
ABS V2008-1 [53]. The structure was solved with SHELXS and re-
fined with Olex2 [54], a graphical interface to SHELXL [55]. The
results are summarized in Table S1.
2.4. General procedure for light-driven catalytic experiments
The light-driven water reduction experiment was performed in
a cylindrical cuvette (NSG, 32UV10) with a total volume of
ꢃ2.5 mL. In a typical experiment, the cell was filled with 2.0 mL
DMF/H2O (1.86/1 v/v) solution containing 0.67 mM Ru(bpy)3Cl2,
0.25 M TEOA, and 22.9 lM catalyst. The pH of the reaction solution
2þ
620 nm emissions of RuðbpyÞ3 were detected by a microchannel
was adjusted using 6 M HCl. The reaction cell was sealed with a
rubber septum, carefully degassed, and filled with Ar. All proce-
dures were performed with a minimum exposure to ambient light.
The reaction samples were irradiated by a LED-light source
(k = 455 nm; light intensity 30 mW, beam diameter ꢃ0.4 cm) at
room temperature with constant stirring (4 ꢄ 103 RPM) using a
magnetically coupled stirring system (SYS 114, SPECTROCELL).
The post-reaction solution was collected and left for a few days
in order for crystals to grow. Needle-like crystals were isolated
by centrifugation and used in the FT-IR stability evaluations
(2.0 wt% samples in KBr) for comparison with complex 1 before
catalysis.
plate photomultiplier tube (Hamamatsu R3809U-51), whose out-
put was amplified and analyzed by a TCSPC board (Becker & Hickel
SPC 600).
2.2. Preparation of polyanion [Mn4(H2O)2(VW9O34)2]10ꢁ (1)
Na10[Mn4(H2O)2(VW9O34)2]ꢂ26H2O (Na1) was synthesized
according to modified literature method [49] as follows:
Mn(NO3)2ꢂ4H2O (1.0 g) and Na2WO4ꢂ2H2O (5.8 g) were dissolved
in 0.5 M sodium acetate buffer (120 mL, pH 4.8) and vigorously
stirred for several minutes before NaVO3 (0.27 g) was added in
small portions. The resulting turbid mixture was then heated to
80 °C for 1.5 h. The hot brown mixture was filtered to remove
any precipitate and left to crystallize for around 1 week to give
Control experiments were carried out under the same condi-
tions in the absence of each component (e.g., TEOA, Ru(bpy)3Cl2)
of the hydrogen generating samples as described above. More

50
H. Lv et al. / Journal of Catalysis 307 (2013) 48–54
control experiments were performed by replacing complex 1 with
TBA3[VW12O40] [56], freshly prepared MnO2, VO2, and WO3 nano-
powders under otherwise the identical conditions.
was characterized by X-ray crystallography (Fig. 1), FT-IR
(Fig. S1), TGA (Fig. S2), UV–Vis (Fig. S3), 51V NMR (Fig. S4) and ele-
mental analysis (Section 2.1). The sodium salt of 1 crystallizes in
the triclinic space group Pꢁ1 (Table S1). The sandwich type cluster,
9ꢁ
[Mn4(H2O)2(VW9O34)2]10ꢁ, consists of two [VW9O34
]
trivacant
2.5. Isotope labeling experiment
3ꢁ
subunits which derive from the Keggin structure ([VW12O40
]
)
by removal of three adjacent edge-sharing octahedral WO6 units.
Isotope-labeling experiments were conducted under the same
conditions as described in Section 2.4 using D2O instead of H2O
to verify the source of H2 production. The gas produced in the
headspace was analyzed by GC-TCD with helium as the carrier gas.
A tetra-manganese core is stabilized in the central belt position
9ꢁ
by two heptadentate [VW9O34
]
units (Fig. 1). Two of the Mn(II)
centers reside in distorted octahedra with one solvent accessible
site and thus one terminal labile water ligand each. The vanadium
atoms occupying the center of each POM ligand reside in an
approximately tetrahedral O4 coordination environment.
2.6. Steady-state and time-resolved fluorescence decay measurement
of [Ru(bpy)3]2+in the presence of 1 or TEOA
Computational studies of the electronic structure of 1 provide
additional support for the nature of Mn-centers. It was shown that
1 has several lower-lying electronic states among which the high-
spin ferromagnetically coupled 21A and the antiferromagnetically
coupled 11A and 1A states are the lowest in energy and degenerate:
the 11A and 1A states are calculated to be only 0.10 and 0.21 kcal/
mol higher in energy than the 21A state. In the other words, the
antiferromagnetic interaction between the Mn-centers is insignifi-
cant. Spin density analyses indicate that in the 21A electronic state
of 1, its Mn1, Mn2, Mn3 and Mn4 centers have 4.82, 4.82, 4.78 and
The mixed solvent of DMF/H2O (1.86/1) was used for fluores-
cence decay measurements. Before each experiment, the
samples with different concentrations of either 1 or TEOA were
purged with argon for 10 min to remove soluble oxygen. The
strong emission band of [Ru(bpy)3]2+ꢀ at 500–800 nm with
kmax = 612 nm (kexcitation = 455 nm) was measured using the
instruments described in Section 2.1.
2.7. Computational Procedure
4.78 |e| unpaired
a-spins, respectively (see Fig. S10 for notation of
atoms in 1). In its 11A state, the Mn1, Mn3 and Mn4 centers are
Calculations were performed using the Gaussian 09 program
[57]. The geometries of the [Mn4(H2O)2(VW9O34)2]10ꢁ (1) polyoxo-
anion were optimized without any symmetry constraint at the
density functional (M06L) level of theory [58]. In these calcula-
tions, we used Hay–Wadt effective core potentials (ECPs) with
the associated Lanl2dz basis set for transition metals [59–61] and
the standard 6–31Gꢀ split-valence-polarization basis set for all
other atoms. The solvent effects were estimated by using the
self-consistent reaction field IEF-PCM method [62] with water as
high-spin Mn(II) centers with 4.82, 4.79 and 4.79 |e|
a-spins,
respectively, while Mn2 has 4.83 |e| b-spins. In the fully antiferro-
magnetically coupled 1A state of 1, the Mn1 and Mn3 centers are
high-spin Mn(II) centers with 4.82 and 4.79 |e|
a-spins, respec-
tively, while the Mn2 and Mn4 centers have 4.83 and 4.79 |e| b-
spins. We should mention that almost 0.80 |e| spins are located
on the O-centers coordinated with Mn-centers. The frontier orbital
analysis (HOMOs and LUMOs) is consistent with the above elec-
tronic-state analyses. The several top singly occupied MOs of the
high-spin ground state of 1 are mostly Mn core orbitals, and there
is small involvement of tungstate orbitals.
a solvent (dielectric constant
e = 78.39).
3. Results and discussion
3.1. Preparation and characterization of complex 1
3.2. Photocatalytic H2 formation
Complex 1 was synthesized using a one-pot method in acidic
buffer solution starting from salts of earth-abundant elements
(Co, W, and V) by a modification of the literature method [49]. It
The catalytic activity was initially studied by the commonly
used system comprising [Ru(bpy)3]2+ (photosensitizer), methyl
viologen (MV2+
; electron relay), and a H2-evolution catalyst
Fig. 1. X-ray crystal structure of 1. Left: a combined ball-and-stick (central unit) and polyhedral (POM ligands) representation; Right: ball-and-stick representation. Red:
oxygen; Pink: hydrogen; Magenta: manganese; Yellow: VO4; Gray: WO6. (For interpretation of the references to color in this figure legend, the reader is referred to the web
version of this article.)

H. Lv et al. / Journal of Catalysis 307 (2013) 48–54
51
[5,20,25,63]. However, only small amount of H2 was produced with
1 as a catalyst under such experimental conditions. Interestingly,
in the absence of MV2+, the yield of H2 increases. The low activity
of 1 with MV2+ as a relay is related to the low reduction potential
of the MV2+/MV+ꢅ couple, ꢁ450 mV vs. NHE relative to the reduc-
tion potential of the primary reducing agent, [Ru(bpy)3]2+ꢀ
:
[Ru(bpy)3]3+/2+ꢀ = ꢁ840 mV vs. NHE. As a consequence, the reduced
methyl viologen, unlike the photoexcited [Ru(bpy)3]2+ꢀ, is not suf-
ficiently reducing to reduce complex 1 in the present system.
Therefore, all the studies reported here involve a methyl violo-
gen-free, three-component system: [Ru(bpy)3]2+ as photosensi-
tizer, TEOA as sacrificial electron donor and complex
catalyst. In this system, the catalyst is directly and readily reduced
by [Ru(bpy)3]2+ꢀ
1 as
.
As shown in Fig. 2, no H2 is formed in dark. After exposure to
light for 0.5 h, a peak at 0.98 min was observed, corresponding to
the formation of H2. The contamination of the reaction vessel
with air is seen as an increase in the N2 peak at 1.2 min. Residual
amounts of dioxygen give a peak at 1.08 min; however, O2 is hard
to detect because it is consumed under the reducing experimen-
tal conditions. Under minimally optimized conditions, 1.92
lmol
H2 gas per 0.046 mol 1, corresponding to a turnover number
l
(TON) of 42, was obtained after 5.5 h of irradiation. Control
experiments revealed that the presence of 1, TEOA, and
[Ru(bpy)3]2+ is essential for H2 evolution: the absence of any
one of these species results in only a trace of H2 formation
(Fig. 3A). Additional control experiments using TBA3[VW12O40],
freshly prepared MnO2, VO2, and WO3 nanopowders (as possible
products of catalyst decomposition) gave only negligible amounts
of H2 (Fig. 3A).
Fig. 3. (A) Photocatalytic H2 evolution with (a) no catalyst; (b) no [Ru(bpy)3]2+; (c)
no TEOA; (d) 0.5 mg freshly prepared MnO2; (e) 3 mg WO3; (f) 0.8 mg VO2; (g)
55 lM TBA3[VW12O40]; (h) 22.9 lM 1 and (B) photocatalytic H2 evolution with
The H2 yield also depends on pH and the concentrations of
[Ru(bpy)3]2+, TEOA and 1. As shown in Fig. 3B, increasing the con-
centration of 1 from 4.6 to 23
0.67 mM [Ru(bpy)3]2+) results in an increase in the H2 yield from
0.18 to 1.51 mol after 5 h of illumination. The concentration of
lM at pH 9.5 (0.25 M TEOA and
different concentrations of 1. Conditions: LED light (30 mW, 455 nm, beam
diameter ꢃ0.4 cm), [Ru(bpy)3]2+ (0.67 mM), TEOA (0.25 M), catalyst
1 (0–
l
22.9 lM), 2 mL DMF/H2O (1.86/1).
the photosensitizer, [Ru(bpy)3]2+, strongly affects the rate and final
yield of H2 (Fig. S5). The amount of H2 formed increases with an in-
crease in [Ru(bpy)3]2+ concentration from 0.13 mM to 0.67 mM. In
contrast, the concentration of TEOA has a minor effect on final H2
yield (Fig. S6). The rate of photocatalytic H2 production is pH
dependent. The maximum rate of H2 evolution and the final yield
of H2 are observed at higher pH (Fig. S7), despite that H+ reduction
becomes less thermodynamically favorable at higher pH. H2 pro-
duction rates and yields maximize when TEOA is fully deproto-
nated (the pKa of TEOAH+ is 7.74) [10,25,64].
3.3. Determination of the origin of H2
To determine the source of H2, isotope labeling experiments
with either H2O or D2O as solvent were conducted. The formation
of H2 and D2 can be distinguished by GC-TCD if helium is used as a
carrier gas due to their thermal conductivity difference [65–67].
Fig. 4 shows the normalized GC signal of H2 and D2 generated in
the full photocatalytic system. Specifically, H2, indicated by a posi-
tive peak at ꢃ0.84 min, forms when H2O is the proton source, and
D2, indicated by a negative peak at ꢃ0.84 min, forms when D2O is
the proton source. These results clearly demonstrate that water is
the source of hydrogen during the photocatalytic reaction.
3.4. Catalyst stability tests
A central point of this study is to address the possibility of for-
mulating a homogeneous WRC or family of WRCs that are thermo-
dynamically stable to hydrolytic decomposition under turnover
conditions, a challenge not met by nearly all known homogeneous
WRCs. In this context, we have examined the stability of 1 using
multiple spectroscopic methods under aqueous catalytic turnover
conditions. First, the electronic absorption spectrum of 1 under
these conditions shows less than 1% change after 3 days (Fig. S3).
Fig. 5 gives the time-dependent data of the UV–Vis absorbance in
first 21 h. Second, the FT-IR spectra (Fig. 6) confirm the integrity
of 1 before and after catalysis. As shown in Fig. 6, the characteristic
IR bands of 1 are basically retained after 24 h of irradiation. Control
experiment (Fig. 6c) shows that the other IR bands result from
TEOA and bpy oxidation products.
Fig. 2. GC/TCD signal of H2 production over time. Conditions: LED light (30 mW,
455 nm, beam diameter ꢃ0.4 cm), TEOA (0.25 M),
1 (22.9 l
M), [Ru(bpy)3]2+
(0.67 mM), 2 mL DMF/H2O (1.86/1). The peak at 1.2 min is N2.

52
H. Lv et al. / Journal of Catalysis 307 (2013) 48–54
Fig. 4. Normalized GC signal of H2 and D2 evolved as monitored by GC-TCD using
helium as carrier gas. Conditions: LED light (30 mW, 455 nm, beam diameter
ꢃ0.4 cm), [Ru(bpy)3]Cl2 (0.67 mM), TEOA (0.25 M), 1 (22.9
lM), 2 mL DMF/H2O (or
DMF/D2O) (1.86/1).
Fig. 7. (A) Emission spectra of [Ru(bpy)3]2+ as a function of added 1 and (B) Stern–
Volmer plot for emission quenching of [Ru(bpy)3]2+ by 1. The concentration of
[Ru(bpy)3]2+ was 2 ꢄ 10ꢁ5 M.
at kmax = 455 nm results in a luminescence at kmax = 612 nm with
emission quantum yield of 0.062 [68]. The excited state can be
either reductively or oxidatively quenched by an electron donor
or acceptor [69–72], resulting in a decrease in the fluorescence
intensity [69–72]. Fig. 7A shows the luminescence spectra of
deaerated solutions of [Ru(bpy)3]2+ (DMF/H2O = 1.86/1, v/v). The
excited state [Ru(bpy)3]2+ꢀ is dramatically quenched by 1. The
Stern–Volmer plot deviates from linearity (Fig. 7B), indicating the
presence of static as well as dynamic quenching. Approximate fit-
ting of the Stern–Volmer plot by linear function gives an apparent
quenching rate constant of 2.8 ꢄ 1011 Mꢁ1 sꢁ1. This is faster than
diffusion control (ꢃ1010 Mꢁ1 sꢁ1) consistent with a contribution
of static quenching via formation of a {1. . .[Ru(bpy)3]2+ꢀ} [17] com-
plex [73–75]. Although the efficient quenching of excited states
may also happen by energy transfer [72,76,77], the small spectral
overlap between the absorption of the catalyst, 1, (Fig. S3) and
the emission of [Ru(bpy)3]2+ (Fig. 7A) makes this pathway unim-
portant [78]. Importantly, at [1] < [Ru(bpy)3]2+ the drop of emis-
sion intensity, (1 ꢁ I/Io) is about twice higher than (1 ꢁ [1]/
[Ru(bpy)3]2+). This is consistent with reduction of 1 by more than
one electron. The reductive quenching of [Ru(bpy)3]2+ꢀ by TEOA
is ruled out because the addition of TEOA in concentrations up to
0.25 M does not affect the luminescence intensity (Fig. S8) [69].
The reduction of [Ru(bpy)3]3+ by TEOA proceeds very rapidly, con-
firming the ability of TEOA to serve as an efficient sacrificial elec-
tron donor (Fig. S9).
Fig. 5. Time-dependent data of UV–Vis absorbance of 0.38 mM TBA1 in the mixed
solvent of DMF/H2O (3/1) containing 0.25 M TEOA.
Fig. 6. FT-IR spectra of (a) TBA1 before light-driven reaction; (b) the same sample
after reaction (re-isolated from post-catalytic reaction solution, illumination time
24 h, TON ꢃ 87) and (c) isolated sample before light-driven reaction from the
solution containing [Ru(bpy)3]3+ and TEOA. Control experiments show that the
additional peaks in the ‘‘after catalysis’’ reaction derive from bpy and TEOA
oxidation products. All FT-IR spectra were 2 wt% samples in KBr.
To further investigate the reaction mechanism, we used time-
resolved fluorescence spectroscopy to follow the kinetics of fluo-
rescence decay of [Ru(bpy)3]2+ (kexcitation = 400 nm). Fig. 8 clearly
3.5. Mechanistic studies
shows that TEOA does not affect the fluorescence decay (
are 0.60 and 0.57 s in the absence and presence of TEOA,
respectively) in agreement with the steady-state fluorescence
s1/2 values
The direct reduction of 1 by [Ru(bpy)3]2+ꢀ was first confirmed by
l
steady-state fluorescence spectroscopy. Excitation of [Ru(bpy)3]2+

H. Lv et al. / Journal of Catalysis 307 (2013) 48–54
53
(CHE-0958205) and the use of the resources of the Cherry Emerson
Center for Scientific Computation.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2013.06.028.
References
[1] N.S. Lewis, D.G. Nocera, Proc. Natl. Acad. Sci. USA 103 (2006) 15729.
[2] E. Fujita, J.T. Muckerman, K. Domen, ChemSusChem 4 (2011) 155–157.
[3] A. Kudo, Y. Miseki, Chem. Soc. Rev. 38 (2009) 253.
[4] A.J. Bard, M.A. Fox, Acc. Chem. Res. 28 (1995) 141.
[5] M. Kirch, J.-M. Lehn, J.-P. Sauvage, Helv. Chim. Acta 62 (1979) 1345.
[6] H. Zhu, N. Song, H. Lv, C.L. Hill, T. Lian, J. Am. Chem. Soc. 134 (2012) 11701.
[7] H.J. Lv, L. Ma, P. Zeng, D.N. Ke, T.Y. Peng, J. Mater. Chem. 20 (2010) 3665.
[8] Z. Han, F. Qiu, R. Eisenberg, P.L. Holland, T.D. Krauss, Science 338 (2012) 1321.
[9] P. Keller, A. Moradpour, J. Am. Chem. Soc. 102 (1980) 7193.
[10] K. Kalyanasundaram, J. Kiwi, M. Grätzel, Helv. Chim. Acta 61 (1978) 2720.
[11] E.D. Cline, S.E. Adamson, S. Bernhard, Inorg. Chem. 47 (2008) 10378.
[12] P. Du, K. Knowles, R. Eisenberg, J. Am. Chem. Soc. 130 (2008) 12576.
[13] H. Kunkely, A. Vogler, Angew. Chem. Int. Ed. 48 (2009) 1685.
[14] Z. Zou, J. Ye, K. Sayama, H. Arakawa, Nature 414 (2001) 625.
[15] T. Peng, X. Zhang, H. Lv, L. Zan, Catal. Commun. 28 (2012) 116.
[16] A.J. Esswein, D.G. Nocera, Chem. Rev. (Washington, DC, U.S.) 107 (2007) 4022.
[17] J.T. Muckerman, E. Fujita, Chem. Commun. (Cambridge, U.K.) 47 (2011).
[18] G. Singh Bindra, M. Schulz, A. Paul, S. Soman, R. Groarke, J. Inglis, M.T. Pryce,
W.R. Browne, S. Rau, B.J. Maclean, J.G. Vos, Dalton Trans. 40 (2011) 10812.
[19] A.F. Heyduk, D.G. Nocera, Science 293 (2001) 1639.
Fig. 8. Normalized fluorescence decay of [Ru(bpy)3]2+ (dark gray rectangles) and
[Ru(bpy)3]2+ with TEOA (gray triangles) and [Ru(bpy)3]2+ with catalyst 1 (light gray
circles). Conditions: 20 l lM catalyst 1. The
M [Ru(bpy)3]2+, 20 mM TEOA, and 27
black solid lines are best fits according to single exponential decay (for the former
two) and bi-exponential decay (for the latter). The y-axis is in log scale.
[20] K. Yamauchi, S. Masaoka, K. Sakai, J. Am. Chem. Soc. 131 (2009) 8404.
[21] H. Ozawa, M.-a. Haga, K. Sakai, J. Am. Chem. Soc. 128 (2006) 4926.
[22] M. Nielsen, A. Kammer, D. Cozzula, H. Junge, S. Gladiali, M. Beller, Angew.
Chem. Int. Ed. 50 (2011) 9593.
[23] W.M. Singh, D. Pegram, H. Duan, D. Kalita, P. Simone, G.L. Emmert, X. Zhao,
Angew. Chem. Int. Ed. 51 (2012) 1653.
[24] J. Zhang, P. Du, J. Schneider, P. Jarosz, R. Eisenberg, J. Am. Chem. Soc. 129 (2007)
7726.
Scheme 1. Proposed mechanism for light-driven hydrogen evolution catalyzed by 1.
[25] P. Du, J. Schneider, P. Jarosz, R. Eisenberg, J. Am. Chem. Soc. 128 (2006) 7726.
[26] F. Wang, W.-G. Wang, H.-Y. Wang, G. Si, C.-H. Tung, L.-Z. Wu, ACS Catal. 2
(2012) 407.
[27] D.G. Nocera, Acc. Chem. Res. 45 (2012) 767.
[28] V. Artero, M. Chavarot-Kerlidou, M. Fontecave, Angew. Chem. Int. Ed. 50 (2011)
7238.
[29] Y.V. Geletii, B. Botar, P. Kögerler, D.A. Hillesheim, D.G. Musaev, C.L. Hill, Angew.
Chem. Int. Ed. 47 (2008) 3896.
[30] A. Sartorel, M. Carraro, G. Scorrano, R.D. Zorzi, S. Geremia, N.D. McDaniel, S.
Bernhard, M. Bonchio, J. Am. Chem. Soc. 130 (2008) 5006.
[31] Q. Yin, J.M. Tan, C. Besson, Y.V. Geletii, D.G. Musaev, A.E. Kuznetsov, Z. Luo, K.I.
Hardcastle, C.L. Hill, Science 328 (2010) 342.
[32] P.-E. Car, M. Guttentag, K.K. Baldridge, R. Alberto, G.R. Patzke, Green Chem.
(2012).
[33] S. Tanaka, M. Annaka, K. Sakai, Chem. Commun. (Cambridge, U.K.) 48 (2012)
1653.
[34] G. Zhu, Y.V. Geletii, P. Kogerler, H. Schilder, J. Song, S. Lense, C. Zhao, K.I.
Hardcastle, D.G. Musaev, C.L. Hill, Dalton Trans. 41 (2012) 2084.
[35] G. Zhu, E.N. Glass, C. Zhao, H. Lv, J.W. Vickers, Y.V. Geletii, D.G. Musaev, J. Song,
C.L. Hill, Dalton Trans. 41 (2012) 13043.
[36] H. Lv, Y.V. Geletii, C. Zhao, J.W. Vickers, G. Zhu, Z. Luo, J. Song, T. Lian, D.G.
Musaev, C.L. Hill, Chem. Soc. Rev. 41 (2012) 7572.
[37] E. Papaconstantinou, J. Chem. Soc. Chem. Commun. (1982).
[38] E. Papaconstantinou, M.T. Pope, Inorg. Chem. 9 (1970) 667.
[39] E. Papaconstantinou, M.T. Pope, Inorg. Chem. 6 (1967) 1152.
[40] E. Papaconstantinou, D. Dimoticali, A. Ioannidis, P. Argitis, J. Photochem. 17
(1981) 171.
quenching results. The kinetics of fluorescence decay in the pres-
ence of 1 is bi-exponential with lifetimes of 21 and 540 ns. The
slower component is assigned to the light emission by
[Ru(bpy)3]2+ꢀ, while the faster component (21 ns) is attributed to
the efficient electron transfer process from excited [Ru(bpy)3]2+
to the proximal catalyst 1. This electron transfer rate
(4.7 ꢄ 107 sꢁ1) is much faster than the diffusion-limited bimolecu-
lar quenching rate (estimated to be 2.7 ꢄ 105 sꢁ1 based on approx-
imated rate constant of ꢃ1010 Mꢁ1 sꢁ1 with the catalyst, 1, at a
concentration of 27 lM). These data, together with the bi-expo-
nential decay behavior, suggest that electron transfer occurs with-
in a {1. . .[Ru(bpy)3]2+ꢀ} complex present in DMF/H2O solutions.
A proposed qualitative mechanism for this visible-light-in-
duced, aqueous-medium catalytic H2 evolution is shown in
Scheme 1. Upon visible light irradiation, the excited state
[Ru(bpy)3]2+ꢀ is oxidatively quenched by catalyst 1 through oxida-
tive electron transfer, producing [Ru(bpy)3]3+ which is rapidly re-
duced by the sacrificial electron donor, TEOA, to regenerate the
photosensitizer [Ru(bpy)3]2+. In the second cycle, the catalyst gains
one more electron and then reduces water to produce H2. Optimi-
zation of the catalytic system and further investigation of the reac-
tion mechanism are in progress.
[41] E.N. Savinov, S.S. Saidkhanov, V.N. Parmon, K.I. Zamaraev, React. Kinet. Catal.
Lett. 17 (1981) 407.
[42] C.L. Hill, D.A. Bouchard, J. Am. Chem. Soc. 107 (1985) 5148.
[43] R.F. Renneke, C.L. Hill, J. Am. Chem. Soc. 108 (1986) 3528.
[44] C.L. Hill, R.F. Renneke, L. Combs, Tetrahedron 44 (1988) 7499.
[45] R.F. Renneke, C.L. Hill, Angew. Chem. Int. Ed. 27 (1988) 1526.
[46] R.F. Renneke, C.L. Hill, J. Am. Chem. Soc. 110 (1988) 5461.
[47] B. Matt, J. Fize, J. Moussa, H. Amouri, A. Pereira, V. Artero, G. Izzet, A. Proust,
Energy Environ. Sci. 6 (2013) 1504.
4. Conclusions
We have developed a new noble-metal-free polyoxometalate-
based and hydrolytically stable (i.e., water compatible) catalyst
for visible-light-induced reduction of water to H2.
[48] P.K. Ghosh, B.S. Brunschwig, M. Chou, C. Creutz, N. Sutin, J. Am. Chem. Soc. 106
(1984) 4772.
[49] B. Li, Y. Yan, F. Li, L. Xu, L. Bi, L. Wu, Inorg. Chim. Acta 362 (2009) 2796.
[50] D.E. Katsoulis, M.T. Pope, J. Am. Chem. Soc. 106 (1984) 2737.
[51] Bruker, APEX2, BRUKER AXS Inc., Madison, WI, USA, 2012.
[52] Bruker, SAINT V7.68a, BRUKER AXS Inc., Madison, WI, USA, 2009.
[53] Bruker, SADABS V2008-1, BRUKER AXS Inc., Madison, WI, USA, 2008.
[54] O.V. Dolomanov, L.J. Bourhis, R.J. Gildea, J.A.K. Howard, H. Puschmann, J. Appl.
Crystallogr. 42 (2009) 339.
Acknowledgments
We thank the NSF (Grant Number CHE-0911610) for support.
The authors gratefully acknowledge NSF MRI-R2 Grant

54
H. Lv et al. / Journal of Catalysis 307 (2013) 48–54
[55] G. Sheldrick, Acta Crystallogr. A 64 (2008) 112.
[56] S. Himeno, M. Takamoto, A. Higuchi, M. Maekawa, Inorg. Chim. Acta 348
(2003) 57.
[57] M.J. Frisch, G.W. Trucks, et al., Gaussian 09, Revision A.1, Gaussian, Inc.,
Wallingford CT, 2009.
[68] J.M. Calvert, J.V. Caspar, R.A. Binstead, T.D. Westmoreland, T.J. Meyer, J. Am.
Chem. Soc. 104 (1982) 6620.
[69] H. Sun, M.Z. Hoffman, J. Phys. Chem. 98 (1994) 11719.
[70] Y.V. Geletii, Z. Huang, Y. Hou, D.G. Musaev, T. Lian, C.L. Hill, J. Am. Chem. Soc.
131 (2009) 7522.
[58] Y. Zhao, D.G. Truhlar, J. Chem. Phys. 125 (2006) 194101.
[59] P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1985) 270.
[60] W.R. Wadt, P.J. Hay, J. Chem. Phys. 82 (1985) 284.
[61] P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1985) 299.
[62] E. Cances, B. Mennucci, J. Tomasi, J. Chem. Phys. 107 (1997) 3032.
[63] E. Borgarello, J. Kiwi, E. Pelizzetti, M. Visca, M. Graetzel, J. Am. Chem. Soc. 103
(1981) 6324.
[64] D.S. Miller, G. McLendon, J. Am. Chem. Soc. 103 (1981) 6791.
[65] K. Snavely, B. Subramaniam, J. Chromatogr. Sci. 36 (1998) 191.
[66] E.L. Garwin, A. Roder, J. Chromatogr. Sci. 14 (1976) 541.
[67] O.F. Degtyareva, L.T. Bondareva, J. Anal. Chem. 59 (2004) 442.
[71] A. Juris, V. Balzani, F. Barigelletti, S. Campagna, P. Belser, A. von Zelewsky,
Coord. Chem. Rev. 84 (1988) 85.
[72] N. Sutin, J. Photochem. 10 (1979) 19.
[73] J.J. Walsh, D.-L. Long, L. Cronin, A.M. Bond, R.J. Forster, T.E. Keyes, Dalton Trans.
40 (2011) 2038.
[74] D. Dung, J. Ramsden, M. Graetzel, J. Am. Chem. Soc. 104 (1982) 2977.
[75] R. Ballardini, M.T. Gandolfi, V. Balzani, Inorg. Chem. 26 (1987) 862.
[76] N. Sutin, C. Creutz, Pure Appl. Chem. 52 (1980) 2717.
[77] D.G. Whitten, Acc. Chem. Res. 13 (1980) 83.
[78] M.D. Symes, D.A. Lutterman, T.S. Teets, B.L. Anderson, J.J. Breen, D.G. Nocera,
ChemSusChem 6 (2013) 65.