Hydrophilicity control of visible-light hydrogen evolution and dynamics of the charge-separated state in dye/TiO2/Pt hybrid systems

Article abstract of DOI:10.1002/chem.201201500

Visible-light-driven H₂ evolution based on Dye/TiO ₂/Pt hybrid photocatalysts was investigated for a series of (E)-3-(5'-{4-[bis(4-R¹-phenyl)amino]phenyl}-4,4'-(R²) ₂-2,2'-bithiophen-5-yl)-2-cyanoacrylic acid dyes. Efficiencies of hydrogen evolution from aqueous suspensions in the presence of ethylenediaminetetraacetic acid as electron donor under illumination at λ>420 nm were found to considerably depend on the hydrophilic character of R¹, varying in the order MOD (R¹=CH ₃OCH₂, R²=H)≈MO4D (R¹=R ²=CH₃OCH₂)>HD (R¹=R ²=H)>PD (R¹=C₃H₇, R ²=H). In the case of MOD/TiO₂/Pt, the apparent quantum yield for photocatalyzed H₂ generation at 436 nm was 0.27A±0.03. Transient absorption measurements for MOD- or PD-grafted transparent films of TiO₂ nanoparticles dipped into water at pH 3 commonly revealed ultrafast formation (<100 fs) of the dye radical cation (Dye^.+) followed by multicomponent decays, which involve minor fast decays (<5 ps) almost independent of R¹ and major slower decays with significant differences between the two samples: 1) the early decay of the major components for MOD is about 2.5 times slower than that for PD and 2) a redshift of the spectrum occurred for MOD with a time constant of 17 ps, but not for PD. The substituent effects on H₂ generation as well as on transient behavior have been discussed in terms substituent-dependent charge recombination (CR) of Dye^.+ with electrons in bulk, inner-trap, and/or interstitial-trap states, arising from different solvent reorganization. Photosensitized H₂ generation by a series of Dye/TiO₂/Pt hybrids revealed clear dependence on the hydrophilic/hydrophobic character of the substituent R of Dye (see figure; I(H₂)=apparent quantum yield for H₂ generation.), which was shown by femtosecond transient spectroscopy to arise from the hydrophilic control of charge recombination (CR) between Dye^.+ and the injected electron TiO₂(e ^-). Copyright

Full text of DOI:10.1002/chem.201201500

DOI: 10.1002/chem.201201500
Hydrophilicity Control of Visible-Light Hydrogen Evolution and Dynamics
of the Charge-Separated State in Dye/TiO2/Pt Hybrid Systems
Won-Sik Han,[a] Kyung-Ryang Wee,[a] Hyun-Young Kim,[a] Chyongjin Pac,*[b]
Yu Nabetani,[c] Daisuke Yamamoto,[c] Tetsuya Shimada,[c] Haruo Inoue,*[c]
Heesung Choi,[d] Kyeongjae Cho,[d] and Sang Ook Kang*[a]
In memory of Kenichi Honda
Abstract: Visible-light-driven H2 evolu-
tion based on Dye/TiO2/Pt hybrid pho-
tocatalysts was investigated for a series
of (E)-3-(5’-{4-[bis(4-R1-phenyl)amino]-
phenyl}-4,4’-(R2)2-2,2’-bithiophen-5-yl)-
2-cyanoacrylic acid dyes. Efficiencies of
hydrogen evolution from aqueous sus-
pensions in the presence of ethylene-
diaminetetraacetic acid as electron
donor under illumination at l>420 nm
were found to considerably depend on
the hydrophilic character of R1, varying
in the order MOD (R1 =CH3OCH2,
MOD/TiO2/Pt, the apparent quantum
yield for photocatalyzed H2 generation
at 436 nm was 0.27ꢂ0.03. Transient ab-
sorption measurements for MOD- or
PD-grafted transparent films of TiO2
nanoparticles dipped into water at
pH 3 commonly revealed ultrafast for-
mation (<100 fs) of the dye radical
of R1 and major slower decays with sig-
nificant differences between the two
samples: 1) the early decay of the
major components for MOD is about
2.5 times slower than that for PD and
2) a redshift of the spectrum occurred
for MOD with a time constant of 17 ps,
but not for PD. The substituent effects
on H2 generation as well as on transi-
ent behavior have been discussed in
cation (Dye +) followed by multicom-
C
ponent decays, which involve minor
fast decays (<5 ps) almost independent
terms substituent-dependent charge re-
+
C
combination (CR) of Dye with elec-
trons in bulk, inner-trap, and/or inter-
stitial-trap states, arising from different
solvent reorganization.
Keywords: charge transfer · dyes/
pigments · hydrogen · photosynthe-
sis · water chemistry
R2 =H)ꢁMO4D
(R1 =R2 =
CH3OCH2)>HD (R1 =R2 =H)>PD
(R1 =C3H7, R2 =H). In the case of
Introduction
[a] W.-S. Han, K.-R. Wee, H.-Y. Kim, Prof. S. O. Kang
Department of Advanced Materials Chemistry
Sejong, Korea University
For light-to-chemical energy conversion, the photochemical
splitting of water into hydrogen and oxygen is a key issue in
artificial photosynthesis.[1–6] In 1972, Fujishima and Honda
reported the first example of photosplitting of water based
on a photoelectrochemical cell using a rutile TiO2 electrode
and a counterelectrode of Pt black [Eq. (1)], the so-called
Honda–Fujishima effect.[1] On illumination of the TiO2 elec-
trode with UV light, charge separation occurs in TiO2 to
give electrons in the conduction band and leave holes in the
valence band. The photogenerated electrons move to the Pt
electrode by application of an external bias to reduce water
to H2, while holes in the valence band oxidize water to form
O2 on the TiO2 electrode. After this pioneering work, many
researchers have been intensively studying photochemical
water splitting using a variety of semiconductor electrodes,
particulate semiconductors, or molecular catalysts.[5b,7–11]
208 Seochang, Chochiwon, Chung-nam 339-700 (Korea)
Fax : (+82)44-867-5396
E-mail: sangok@korea.ac.kr
[b] Prof. C. Pac
Yulchon Research Center, Sejong, Korea University
208 Seochang, Chochiwon, Chung-nam 339-700 (Korea)
Fax : (+82)44-867-5396
E-mail: jjpac@korea.ac.kr
[c] Y. Nabetani, D. Yamamoto, T. Shimada, Prof. H. Inoue
Department of Applied Chemistry
Graduate Course of Urban Environmental Sciences
Tokyo Metropolitan University
Minami-ohsawa 1-1, Hachiohji, Tokyo 192-0397 (Japan)
Fax : (+81)-426-77-2525
E-mail: inoue-haruo@tmu.ac.jp
[d] H. Choi, Prof. K. Cho
Department of Materials Science and Engineering
University of Texas at Dallas
Richardson, Texas 75080 (USA)
and
H2O þ hn ðTiO2 ꢀ PtÞ ! H2 þ 0:5 O2
ð1Þ
WCU Multiscale Mechanical Design Division
School of Mechanical and Aerospace Engineering
Seoul National University, Seoul 151-742 (Korea)
However, TiO2 has a crucial drawback for use in artificial
photosynthesis due to its wide bandgap, which limits light
absorption at l<420 nm and is unfavorable for utilization
of solar light. For this reason, investigations have been per-
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201201500.
15368
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 15368 – 15381

FULL PAPER
formed to develop visible-light-absorbing semiconductors in
place of TiO2.[8,9,11] Nonetheless, TiO2 is still an attractive
material of wide applicability,[12–15] because of its high photo-
catalytic activity, excellent photostability, low toxicity, and
rich abundance. An effective way to overcome the drawback
of TiO2 is to utilize photosensitization by a molecular dye
which acts as an effective visible light harvesting antenna
and as an efficient excited-state electron-transfer compo-
nent. In such Dye/TiO2 systems, TiO2 may function as a
chemically stable charge transporter and as a potential
charge reservoir to achieve one-electron-to-multi-electron
conversion chemistry. From such a viewpoint, numerous
studies have been carried out for visible-light-driven H2 gen-
eration by water reduction [Eq. (2)], the half-reaction of
water splitting, with aqueous dispersions of Pt-loaded TiO2
particles (TiO2/Pt) or related solid catalysts in the presence
of a dye photosensitizer, typically an RuII complex, and a
sacrificial electron donor (SD).[16–18]
has been demonstrated by multiple-component decays of
transients formed by electron injection from a photoexcited
[26]
dye into TiO2
as well as by direct photoexcitation of
TiO2.[24b,27,28] Therefore, a crucial question emerges about
what CR process(es) would be essential in determining the
net efficiencies of H2 generation.
Another important factor taken into consideration for
Dye/TiO2 hybrids is the water miscibility of the dye mole-
cule grafted on the TiO2 surface. In general, the neutral
chromophoric cores of conventional organic dyes are intrins-
ically hydrophobic. Consequently, the poor water miscibility
inevitably gives rise to complex features at the Dye/TiO2 in-
terface in contact with the water phase as well as in the ori-
entation/aggregation features of the adsorbed dyes on the
TiO2 surface, and thus exerts critical effects on the electron-
transfer and chemical processes involved in net H2 genera-
tion. Recently, we demonstrated that H2 generation based
on Dye/TiO2/Pt using (diphenylaminophenyl)dithiophenea-
crylic acid as the dye and ethylenediaminetetraacetic acid
(EDTA) as SD is controlled by the hydrophilic CH3O-
H2O þ SD þ hn=ðDye=TiO2=PtÞ ! H2 þ SDox
ð2Þ
AHCUTNERTG(GNNUN CH2CH2O)nCH2 substituent introduced into the amino-
Shimidzu et al. reported that H2 generation can be sensi-
tized by xanthene dyes under visible-light irradiation by
using Pt-loaded metal oxide particles in the presence of trie-
thanolamine as SD.[19] In recent years, considerable improve-
ments in H2-generation efficiency have been achieved with
similar reaction systems in which the surfaces of TiO2/Pt
particles or related solid catalysts are modified with various
dyes.[20,21] In these systems, the visible-light-absorbing dye
can function as an electron donor in the excited state(s) for
phenyl end group.[29] It has been suggested that the hydro-
philic character of the substituent coupled with its steric ef-
fects should play an essential role in maximizing the effi-
ciency of photocatalytic H2 generation; greater hydrophilic
character of the substituent with lower steric bulkiness
would allow easier access of SD to the radical-cation center
localized on the end group of Dye +, and thus enhance the
C
+
C
reduction of Dye in competition with charge recombina-
tion to leave greater amounts of long-lived electrons in the
TiO2 particle.
electron injection into the conduction band of TiO2 to form
+
C
charge-separated state(s), dye radical cation (Dye
This interpretation seems reasonable but rather simple,
because the CR behavior in TiO2 catalysts reveals complex
features. To obtain closer insights into essential pathways
controlling the H2-generation efficiency in our Dye/TiO2/Pt
hybrid system, we have performed further investigations on
photocatalytic H2 generation using a series of dyes with a
hydrophilic or hydrophobic substituent in the dye core,
namely, hydrophobic parent dye HD with no substituent,
PD with hydrophobic propyl substituent at the aminophenyl
end, MOD with hydrophilic methoxymethyl substituent at
the aminophenyl end, and MO4D with hydrophilic methox-
ymethyl substituents at both the outer diphenylamino and
inner dithiophene units. We confirmed that the photocatalyt-
ic H2-generation efficiencies show clear dependence on the
hydrophilic/hydrophobic character of the substituent attach-
ed to the diphenylamino end group. Transient absorption ex-
periments in different time domains spanning from subpico-
seconds to microseconds on hydrophilic MOD and hydro-
phobic PD as representative dyes demonstrated that the
transient decays involve substituent-independent events oc-
curring within 5 ps followed by substituent-dependent be-
havior in a longer time domain extending to microseconds.
Herein we report the findings and discuss implications of
the photodynamic results in relation to essential factors that
control the efficiency of H2 generation photosensitized by
the Dye/TiO2/Pt hybrid.
)/TiO2(eꢀ). However, there has been little exploration of de-
tails of essential factors controlling net efficiencies of H2
generation in such hybrid systems. Actually, dye-sensitized
H2 generation should involve a variety of competitive path-
ways[22] including, at least, electron injection in competition
with the decay of excited-state dye, charge recombination
+
(CR) between Dye and TiO2(eꢀ) as a crucial energy-wast-
C
+
C
ing process, reduction of Dye by SD to leave long-lived
electrons, accumulation of persistent electrons on the Pt
site, and reduction of protons (or water) to H2. According to
this simple scheme, the H2-generation efficiency should be
primarily determined by kinetics of the competitive process-
es, particularly CR process(es) competing with electron
transfer from SD to Dye +, though little direct evidence
C
proving the relationship of net H2-generation efficiencies
with the kinetics has appeared. In reality, TiO2 nanoparticles
have complex chemical and morphological features on their
surfaces[14c] and a variety of energetically distributed trap
sites, so that electrons injected from an excited-state dye
may reveal complex kinetic behavior.[23] It was reported that
very fast trapping of electrons occurs with subpicosecond
time constants on bandgap excitation of TiO2 particles,[24]
whereas other investigations indicated the existence of long-
lived electrons.[25] This means that the electrons injected into
a TiO2 particle should show complex kinetic behavior, as
Chem. Eur. J. 2012, 18, 15368 – 15381
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
15369

C. Pac, H. Inoue, S. O. Kang et al.
Results and Discussion
Synthesis and electronic properties of the dyes: The dyes
were prepared according to Scheme 1 as previously report-
Figure 1. UV/Vis absorption (empty symbols) and fluorescence (full sym-
bols, normalized) spectra for THF solutions of PD (black), HD (red),
MOD (blue), and MO4D (pink).
Table 1. Spectroscopic properties of the dyes.
Dye
Absorption
Fluorescence
DRS[e]
[nm]
[d]
lmax [nm][a] e [mꢀ1 cmꢀ1
]
lF [nm][b] tF [ns][c] FF
Scheme 1. Synthetic route for the dyes. 1) [PdCl
308C; 2) nBuLi, DMF, THF, ꢀ788C; 3) p-toluenesulfonic acid, neopentyl
glycol; 4) nBuLi, iPrOBO2C6H12, 5) [Pd(PPh3)4], K2CO3, THF/H2O,
reflux; 6) [Pd(dba)3] (dba=trans,trans-dibenzylideneacetone), Xanthphos,
NaOtBu; 7) Trifluoroacetic acid, THF, RT; 8) NCCH2COOH, piperidine,
MeCN, reflux.
2ACHTUNGTREN(NGNU PhCN)2], AgF, DMSO,
HD
PD
MOD
MO4D 456
433
442
432
27360
27378
26378
29846
552
568
561
645
1.18
1.22
1.25
1.04
0.23 430
0.21 450
0.24 448
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
0.15
–
[a] Absorption maxima in THF. [b] Fluorescence maxima in THF. [c] Flu-
orescence lifetimes. [d] Fluorescence quantum yields. [e] Absorption
maxima in diffuse reflectance spectra of Dye/TiO2 powders.
ed.[29] A key compound in the synthesis is 5,5’-dibromo-4,4’-
(R2)-2,2’-bithiophene (R2 =H or CH3OCH2) which was ob-
tained by the PdCl
(PhCN)2/AgF-catalyzed coupling reac-
emission maxima of MO4D are redshifted by 23 and 93 nm,
respectively, compared to those of parent dye HD. The four
electron-donating methoxymethyl groups substituted at the
whole donor part should enhance the charge-transfer char-
acter. On the other hand, the spectroscopic properties of the
other three dyes are very similar to each other, which sug-
gests that substitution with propyl or methoxymethyl groups
at the diphenylamino donor end causes little perturbation of
the excited singlet state.
tion of 2-bromo-3-R2-thiophene. 2-Bromo-3-methoxyme-
thylthiophene was obtained by bromination of the methyl
group of 2-bromo-3-methylthiophene with N-bromosuccini-
mide followed by heating a methanol solution in the pres-
ence of K2CO3. Formylation at the 2-position of the dibro-
mobithiophene followed by acetal protection gave 1, from
which dithiopheneboronic acid pinacol esters 2 were pre-
pared by replacing the 5-bromo substituent with the dioxa-
boranyl group. The Suzuki cross-coupling reaction of 4-bro-
moaniline with 2 afforded 3 in 72–85% yield, which was
then subjected to palladium-catalyzed Buchwald–Hartwig
reaction with 4-R1-bromobenzene to give 4. The target dyes
were thus obtained by selective deprotection of the acetal
group of 4 with trifluoroacetic acid followed by Knoevena-
gel condensation of the corresponding aldehydes with cya-
noacetic acid.
Cyclic voltammetry of the dyes revealed common features
with quasi-reversible oxidation and reduction waves (Fig-
ure S1, Supporting Information), from which the half-wave
ox
1=2
red
1=2
oxidation and reduction potentials (E and E vs. Fc+/Fc)
were estimated (Table 2). The excited-singlet state oxidation
ox
potentials (Eꢃox) of the dyes calculated from the E values
1=2
and the 0–0 transition energies (E0–0) are ꢀ(1.96ꢂ0.06) V
versus Fc+/Fc, which are substantially more negative than
the flat-band potential of TiO2 at pH 3 of ꢀ0.58 V[30a] or
Figure 1 shows UV/Vis absorption and emission spectra
of the dyes in THF, and the corresponding data are summar-
ized in Table 1. The absorption spectra show strong maxima
at 432–456 nm (emax =2.7–3ꢁ104) with an end tail extending
to 530–560 nm, and relatively intense fluorescence occurs at
552–645 nm with significantly short lifetimes (tF =1.04–
1.26 ns). It was reported on the basis of TD-DFT calcula-
tions that the electronic transitions can be attributed to in-
tramolecular charge transfer between the cyanoacrylic acid
acceptor and the (diphenylaminophenyl)bithiophene
donor.[29] In line with this assignment, the absorption and
Table 2. Cyclic voltammetric data [V] of the dyes in THF.[a]
Dye
Oxidation
Reduction
ox
1=2
red
1=2
[b]
Epa
Epc
E
ꢀEpc
ꢀEpa
ꢀE
ꢀEꢃox
PD
HD
MOD
MO4D
0.39
0.52
0.56
0.54
0.34
0.46
0.44
0.47
0.37
0.49
0.50
0.51
2.21
2.10
2.23
2.17
2.12
2.00
2.06
2.09
2.17
2.05
2.14
2.13
2.04
1.99
1.95
1.76
[a] Epa =anodic peak potential, Epc =cathodic peak potential. [b] Oxida-
tion potential in the excited singlet state; Eꢃox =EoxꢀE0–0
.
15370
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 15368 – 15381

Hydrogen Evolution in Dye/TiO2/Pt Hybrid Systems
FULL PAPER
ꢀ0.47 V[30b] versus SCE (ꢀ0.20 V or ꢀ0.09 V vs. Fc+/Fc);
the E0–0 values were estimated from the intersectional wave-
length between the absorption and fluorescence spectra
ox
1=2
red
(Figure 1). Either the E or E1/2 values of the dyes are
similar with differences less than or equal to about 50 mV
ox
1=2
red
1=2
for E and less than or equal to about 20 mV for E , that
is, the substituents of the dyes appear to have only minor ef-
fects on the redox processes. As previously discussed on the
basis of TD-DFT calculations,[29] one-electron reduction
should occur on the cyanoacrylic acid acceptor where the
LUMO is largely localized, whereas oxidation should pro-
ceed mainly on the diphenylaminophenyl donor end where
the HOMO is mainly populated.
Figure 2. Diffuse-reflectance spectra (DRS, normalized) of 1 mmol dye on
10 mg of s-TiO2 particles for PD (black), HD (red), and MOD (blue).
Dye/TiO2/Pt hybrid: According to the providerꢂs technical
data, the as-received TiO2 particles have anatase modifica-
tion, primary crystal size of less than 10 nm, and BET sur-
face area of greater than or equal to 250 m2 gꢀ1. The as-re-
ceived TiO2 sample was washed with distilled water before
use. The washed TiO2 particles (abbreviated as w-TiO2) are
micrometer-sized agglomerates with a size distribution from
500 nm to ꢁ10 mm and a peak size of about 3 mm, as con-
firmed by dynamic light scattering analysis of a suspension
in water. The observed BET surface area of w-TiO2 was
271.7 m2 gꢀ1, which is close to the providerꢂs data. On soni-
cating w-TiO2 powder in water for 4 h at room temperature,
the agglomerates were partially smashed to give smaller par-
ticles with a peak size of about 1 mm and a size distribution
from 200 nm to 2 mm. The observed BET surface area of
this sonicated sample (abbreviated as s-TiO2) was
304.1 m2 gꢀ1, which is about 12% greater than that of w-
TiO2. In the present investigation, we used s-TiO2 and, to a
limited extent, w-TiO2 for comparison.
Platinum loading of TiO2 particles was performed by a
photodeposition method.[31] Adsorption of the dyes on TiO2
was carried out by stirring the Pt-loaded TiO2 particles
(TiO2/Pt) in a solution of dye in acetonitrile/tert-butyl alco-
hol (1:1) overnight. After the adsorption process, the dye-
loaded TiO2/Pt particles (Dye/TiO2/Pt) were collected from
the slurry by centrifugation, thoroughly washed with the sol-
vent, dried, and stored under N2 in the dark. The superna-
tants from the centrifugation step were transparent. Further-
more, no indication of liberation of the dyes from Dye/TiO2/
Pt was obtained after the collected Dye/TiO2/Pt particles
were stirred in the organic solvent or in water overnight.
The stable adsorption behavior of the dyes indicates their
strong anchoring on the TiO2 surface, due to chemical bond-
ing of the carboxylic acid group with the TiO2 surface.
shift in the case of parent HD (Table 1). The redshifts cou-
pled with spectral broadening could arise, at least in part,
from electronic interactions of the adsorbed dyes with
TiO2.[32] In this respect, it is noteworthy that the Dye/TiO2
particles exhibit virtually no fluorescence. Laser-pulse irradi-
ation of the Dye/s-TiO2 powders gave no extra emission-
decay component at all other than the laser-pulse profile
(ꢁ30 ps). The very fast fluorescence quenching that occurs
within 30 ps in this system indicates efficient electron injec-
tion from the excited singlet state of the dyes into the con-
duction band of TiO2 (see below), probably due to possible
electronic couplings between the dyes and TiO2.
Hydrogen generation: Hydrogen-generation experiments
were carried out for 3 mL suspensions of 10 mg Dye/TiO2/Pt
catalysts in water at pH 3 in the presence of approximately
5 mm ethylenediaminetetraacetic acid (EDTA, saturated
concentration in water at pH 3). After bubbling with an N2
stream for 30 min, the suspensions were irradiated with visi-
ble light at l>420 nm under stirring. The amounts of H2
generated were determined by gas chromatography (ꢄ5%
error). To confirm optimum loadings of Pt and dye, we per-
formed preliminary experiments with varying amounts of Pt
(0.1–1.0 wt%) and MOD (0.1–10 mmol/10 mg s-TiO2). With
a fixed amount of MOD (1.0 mmol/10 mg s-TiO2), the H2-
generation efficiency was found to be maximal at 0.3 wt%
Pt loading. Presumably, excess Pt loading would lead to
more extensive quenching of excited-state dye molecules
and/or to formation of larger Pt particles of lower catalytic
activity.[33] For 0.3 wt% Pt-loaded s-TiO2, the optimum ad-
sorbed amount of MOD was found to be 1.0 mmol/10 mg s-
TiO2, an amount assumed to be sufficient for efficient har-
vesting of the incident light, but low enough to avoid signifi-
cant molecular aggregation associated with fast decay of the
excited state. Comparable experiments with s-TiO2 and w-
TiO2 loaded with 0.3 wt% Pt and 1.0 mmol dye/10 mg TiO2
showed that H2 generation proceeds 3.4 times more effi-
ciently with the s-TiO2-based catalyst than with the w-TiO2-
based one. Accordingly, we used s-TiO2 particles loaded
with 0.3 wt% Pt and 1.0 mmol dye/10 mg TiO2 for the pres-
ent H2-generation experiments unless otherwise stated. In
the case of 1.0 mmol MOD/10 mgs-TiO2/0.3 wt% Pt, per-
Figure 2 shows the absorption spectra of Dye/s-TiO2 parti-
cles taken by diffuse-reflectance spectroscopy (DRS), which
reveal similar features for HD, PD, and MOD with distinct
maxima but with broad end tails reaching to about 700 or
about 750 nm, substantially longer than those in solution
(ꢁ530 nm). Although the broad spectra may be mainly due
to multiple reflections in the powder samples, the spectra
for PD or MOD with electron-donating propyl or methoxy-
methyl groups show significant redshifts as opposed to little
Chem. Eur. J. 2012, 18, 15368 – 15381
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
15371

C. Pac, H. Inoue, S. O. Kang et al.
forming four repeated runs for H2 generation under identi-
cal conditions gave sufficiently reproducible results (Sup-
porting Information Figure S2).
Figure 3 shows typical plots of H2 generation versus irra-
diation time for the Dye/s-TiO2/Pt photocatalysts. It was
confirmed that the aqueous phase after 240 min irradiation
dipped in water adjusted to pH 3, were irradiated with laser
pulses. For these experiments, hydrophobic PD and hydro-
philic MOD were used as representative dyes, because they
have similar molecular shape and size but showed a consid-
erable difference in H2 generation efficiency.
As a reference experiment, transient spectra for homoge-
neous ethanol solutions of the dyes were taken in a time
range of 0–250 ps after laser-pulse irradiation at 400 nm
(Supporting Information Figure S4). Both PD and MOD
showed evolution of very broad spectra with a maximum at
about 730 nm within about 100 fs followed by decays with
time constants beyond 250 ps. The transient spectra should
be attributable to S1!Sn (nꢅ2) transition(s). The evolution
of the transient absorption seems to involve an early part
within a couple of picoseconds (indicated by 0 ps in Sup-
porting Information Figure S4) and a major component of
10 ps. This behavior can be reasonably explained in terms of
the Franck–Condon state population in Sn followed by vi-
brational relaxation coupled with solvent reorganization.
As shown in Figure 4, on the other hand, laser-pulse irra-
diation of Dye/np-TiO2 films dipped in water at pH 3
showed simultaneous growth of broad absorption in the
wavelength region of about 580–800 nm accompanied by
negative absorption due to bleaching of the ground-state ab-
sorption below about 570 nm. The transient spectra immedi-
ately after laser irradiation (at 0 ps in Figure 4) are almost
Figure 3. Temporal plots for photocatalyzed H2 evolution from water sus-
pensions of Dye/s-TiO2/Pt (1 mmol dye/10 mg TiO2) in the presence of
ꢁ5 mm EDTA at pH 3 under visible-light irradiation (l> 420 nm) at
293 K.
turns less acidic (pH 4–4.5). It is noteworthy that H2-genera-
tion efficiency significantly depends on the dye substituent,
and the turnover number (TON=micromoles of H2 generat-
ed per micromole of dye) after 4 h of irradiation varied
from 17 for MOD, to 15 for MO4D, to 8.0 for HD, and to
5.4 for PD. Hydrophilic methoxymethyl substitution at the
4,4’-positions of the diphenylamino end group enhances the
photosensitization activity compared with the parent HD
catalyst, in contrast to the significant diminishing effect of
the hydrophobic propyl substituent. On the other hand, ad-
ditional substitution with methoxymethyl groups in the
inner dithiophene unit had no particular effect, but led to a
slight decrease in H2 generation. The apparent quantum
yield (AQY) at 436 nm for MOD/s-TiO2/Pt was determined
in
a
linear time–conversion region to be 0.27ꢂ0.03
[Eq. (3)].[34]
amount of H2 generated per unit time
number of incident photons per unit time
ð3Þ
AQY ¼
Transient absorption spectra: We attempted to measure
transient spectra for Dye/s-TiO2 particles dispersed in water
at pH 3. However, no reliable results were obtained. There-
fore, transient-absorption experiments were performed for
transparent films of dye-loaded anatase-TiO2 nanoparticles
(Dye/np-TiO2) coated on a quartz plate. The visible absorp-
tion spectra of the Dye/np-TiO2 films (Supporting Informa-
tion Figure S3) are almost identical with the DRS of Dye/s-
TiO2 particles shown in Figure 2, that is the isoelectronic
nature of the two different Dye/TiO2 systems allows use of
the films as an effective substitute of the dispersion system
for spectroscopic measurements. The films, which were
Figure 4. Transient absorption spectra in the time region of <20 ps for
films of MOD/np-TiO2 (top) and PD/np-TiO2 (bottom) dipped in water
at pH 3.
15372
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 15368 – 15381

Hydrogen Evolution in Dye/TiO2/Pt Hybrid Systems
FULL PAPER
identical for the two samples. The spectral feature is consid-
erably different from those of the S1!Sn (nꢅ2) transition of
the dyes in Figure S3 (Supporting Information) and also
from those reported for electrons in the conduction band/
trap site(s) of TiO2.[35] The transient absorptions at 600–
800 nm shown in Figure 4 should be mainly attributable to
The transient absorption spectra taken at a 20 ns delay
after irradiation with nanosecond laser pulses at 416 nm
have two maxima at about 650 nm and 730 nm with a spec-
tral shape similar to those observed in the picosecond
region, and reveal a difference in absorbance of the surviv-
ing transients between the two samples (Supporting Infor-
mation Figure S6). Figure 6 shows decays of the long-lived
the dye radical cation (Dye +), because a similar spectrum
C
was reported for the radical cation of a dye having the same
donor and acceptor ends connected through a thiophenyl-
ethylene unit.[36]
Figure 5 (top) shows decay profiles of the transients moni-
tored at 730 nm within 3.5 ps, which are very similar for the
two samples, though minor differences can be seen. The
Figure 6. Decay profiles in the <17 ms time region monitored at 730 nm
for films of MOD/np-TiO2 (left) and PD/np-TiO2 (right) dipped in water
at pH 3.
transients monitored at 730 nm. Although the kinetic behav-
ior is complex, it can be clearly seen that the transient of
MOD/np-TiO2 survives for a longer time than that of PD/
np-TiO2 in the time region of less than 4 ms.
Theoretical calculations of the orientation of dye molecules
on the TiO2 surface: Dyes with a stretched rectangular
shape of the core may take on various molecular orienta-
tions on anchoring to the TiO2 surface, which should be re-
lated to different distances between the dye donor part and
the TiO2 surface as well as to different electronic couplings
in the charge-separated state(s).
Therefore, molecular orientations of the dyes would affect
CR processes and play a possible controlling role in H2 gen-
eration efficiency. Thus, we performed preliminary theoreti-
cal calculations for possible orientations of MOD and PD
on the (101) surface of TiO2 in vacuo. The carboxyl group of
the dyes can bind with the TiO2 surface through a monoden-
tate binding mode between one oxygen atom of the COO
group and one Ti atom, a bidentate chelating mode between
the two oxygen atoms of the COO group and one Ti atom,
and a bidentate bridging mode between the two oxygen
atoms and two adjacent Ti atoms,[37] among which the last-
named was reported to be relatively favorable.[38] In the
present study, therefore, we assumed the bidentate bridging
mode for anchoring of the dyes on the TiO2 surface. The
structural optimizations were initially performed for binding
of the cyanoacrylate root part (H2C=C(CN)COO) to the
two converged positions of TiO2 shown by the blue and
green boxes in Figure 7a.
Figure 5. Decay profiles of transients monitored at 730 nm in a <3.5 ps
time domain (top) and for a longer time region extending to 1.5 ns
(bottom) for films of MOD/np-TiO2 (blue) and PD/np-TiO2 (red) dipped
in water at pH 3.
decays apparently consist of a minor component (<0.5 ps)
and a major one (>0.5 ps). When the decays were moni-
tored in a much longer time region extending to 1.5 ns, on
the other hand, we observed significantly different behavior
in the decay kinetics, such that the MOD/np-TiO2 film
shows a slower decay than the PD/np-TiO2 film, as shown in
Figure 5 (bottom). The decays monitored in the time region
of 100–1500 ps gave time constants of 4.6ꢁ108 sꢀ1 (or 2.2 ns)
for MOD/np-TiO2 and 1.1ꢁ109 sꢀ1 (or 0.91 ns) for PD/TiO2.
Starting from the initial result, we proceeded with further
calculations to optimize the configurations for the dithio-
phene-bonded cyanoacrylate unit and then for the whole
molecules. It was confirmed that the optimized configura-
tion of the cyanoacrylate part is kept at each calculation
Chem. Eur. J. 2012, 18, 15368 – 15381
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
15373

C. Pac, H. Inoue, S. O. Kang et al.
for both MOD and PD reveal similar dependences of the
energy change on the angle of inclination. In the case where
the two oxygen atoms are fixed in the optimized configura-
tion, the energy change slightly decreases with a variation of
ꢄ0.05 eV on inclining to the left-hand side. On inclining to
the right-hand side, on the other hand, the energy change
continuously increases but within 0.2 eV. In contrast, when
the whole carboxylate part is fixed, the energy increases to
much greater extent on inclining towards either direction
from the orientation at 08. The latter behavior can be rea-
sonably understood, because the carboxylate carbon atom
should be involved in conjugation with the upper p-electron
system. These findings suggest that the molecular plane of
the p-electron system including the carboxylate carbon
atom may be capable of inclining towards either of the di-
rections with no or small energy changes.
Discussion: The spectroscopic and electrochemical behavior
of the dyes in organic solvents indicates that the electronic
properties in the ground or excited singlet state are not sig-
nificantly affected by the R1 substituents at the aminophenyl
end, whereas the redshifted fluorescence of MO4D indicates
considerable stabilization of the excited singlet state with
charge-transfer character due to the methoxymethyl substi-
tution of the whole donor part. On excitation of the dyes,
charge transfer from the donor part to the cyanoacrylic acid
acceptor site should occur to result in localization of the ex-
cited electron on the acceptor part, a situation favorable for
efficient injection of the excited electron into the conduction
band of TiO2. This was strongly indicated by complete
quenching of dye fluorescence on adsorption on the TiO2
particle surface as well as by the ultrafast appearance of
transients on laser-pulse irradiation of the Dye/np-TiO2
films. Alternatively, fluorescence quenching in the Dye/s-
TiO2 particles could be due to aggregation of dye molecules.
As shown in Figure 2, however, the DRS shows distinct ab-
sorption maxima which are much different from a feature-
less spectrum characteristic of molecular aggregation report-
ed for an indoline dye adsorbed on TiO2,[38b,39] and suggest
that little dye aggregation occurs in the present Dye/s-TiO2
system. In the case of HD, for which intermolecular p–p
Figure 7. a) Top view of the TiO2 (101) surface. The blue and green boxes
show the geometric positions for mode 1 and mode 2 bonding of the dyes
on the TiO2 surface, respectively. b) Side view of MOD on the TiO2 sur-
face with energy-optimized configuration based on mode 1. The same
molecular configuration was obtained for PD. Energy changes were cal-
culated by inclining the molecular plane towards the left- and right-hand
sides. c) Front view of the molecular configuration on the TiO2 surface.
The colored balls represent oxygen (red), titanium (pink), carbon (cyan),
sulfur (yellow), nitrogen (blue), and hydrogen (white). d) Enlarged front
view of the dye-anchoring part on the TiO2 surface. The red and blue
lines indicate the two positions fixed while inclining the upper part of the
molecular plane.
step and that dye adsorption on the blue-box position
(mode 1) is more stable by 0.20 eV for MOD and by
0.15 eV for PD than that on the green-box position
(mode 2). Accordingly, we describe here only the calculation
results for the bidentate-bridging adsorption of the dyes on
the mode 1 position, which give the energy-optimized con-
figuration of the dyes shown in Figure 7b and c. The side
view indicates an approximately perpendicular orientation
of the molecular plane with respect to the TiO2 surface,
whereas the molecule-core axis
is inclined along the direction
of the Ti-O-Ti bond. Energy
changes caused by inclining the
molecular plane from the con-
figuration in Figure 7b towards
the left- (negative angles) and
right-hand (positive angles)
sides were calculated while
fixing the two oxygen atoms of
the carboxylate group (shown
by the red line in Figure 7d) or
the whole carboxylate part
(blue line in Figure 7d) over
the TiO2 surface. As shown in
Figure 8. Energy changes versus angles of inclination of the molecular plane calculated while fixing the two
oxygen atoms of the carboxylate group (left) and the whole carboxylate part (right) over the TiO2 surface for
Figure 8, the calculated results MOD (blue line) and PD (red line).
15374
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 15368 – 15381

Hydrogen Evolution in Dye/TiO2/Pt Hybrid Systems
FULL PAPER
stacking may occur with least steric repulsion among the
dyes, calculations using the estimated molecular area
(ca. 1.5 nm2) and the observed BET surface area of s-TiO2
gave approximately 15% coverage for 1 mmol of HD on
10 mg of s-TiO2 particles, which is perhaps low enough to
avoid molecular aggregation. In addition, the two methoxy-
methyl or propyl groups of MOD or PD attached to the dye
core might exert more or less steric interference on intermo-
lecular p–p stacking of the dyes.
While the steady-state spectroscopic behavior of the dyes
in organic solvents and on the TiO2 surface is almost inde-
pendent of the substituent R1, catalytic activities of Dye/s-
TiO2/Pt in photosensitized H2 generation are considerably
controlled by the substituent attached to the donor end, as
demonstrated by threefold more efficient H2 generation in
the case of the hydrophilic substituent (MOD) compared to
the hydrophobic substituent (PD). A major difference be-
tween the two substituents is clearly attributable to the hy-
drophilic and hydrophobic properties. Solubility of the free
dyes and their methyl esters in water can be used as a
simple measure for estimating their hydrophilicity. Saturated
concentrations of the dyes in water at ambient temperature
were determined from the absorption spectra to be about
15 mm for MOD and about 8 mm for its methyl ester, while
HD and PD as well as their methyl esters are virtually in-
soluble in water (!1 mm). This indicates that the dye core is
highly hydrophobic and that the two hydrophilic methoxy-
methyl groups attached to the aminophenyl ends can make
the MOD molecule hydrophilic, albeit to a slight extent,
and probably the limited part of the diphenylamino end
group should be more hydrophilic.
The photodynamic behavior of MOD and PD adsorbed
on the films of anatase TiO2 nanoparticles again showed
similar dependences on the dye substituent, which suggests
that CR processes associated with the photocatalytic H2 gen-
eration efficiency are controlled by the hydrophilic/hydro-
phobic character of the substituent. Although the nanocrys-
talline film system is morphologically different from the dis-
persion used for H2 generation, the observations made with
the film system should provide essential information on CR
processes occurring between the oxidized dye and injected
electrons in conduction-band/trap states of TiO2 particles.
The multicomponent decay behavior of transients observed
in the present investigation parallels that reported for a vari-
ety of TiO2 systems involving dye-loaded nanocrystalline
anatase films covered with organic solvents,[23,36,40] colloidal
dispersions of dye-loaded amorphous TiO2 particles in
water,[26,32a] and anatase powders.[24b,27]
tation of TiO2 particles are long-lived[25] and that most of
photogenerated electrons remain in the conduction band.[25b]
According to these reported observations, a basic mecha-
nism working in the present system can be represented by
Equations (3)–(11). Following ultrafast (<100 fs) electron
injection from the excited dye into the conduction band of
TiO2 [Eq. (3)], very fast trapping of the injected electrons
occurs, at least in part, to give trapped electrons (TiO2(eꢀ)ꢂ
and TiO2(eꢀ)’’) [Eq. (4)]; TiO2(eꢀ) in Equation (3) repre-
sents the electrons immediately after injection from the pho-
toexcited dye. The fast substituent-independent decays of
transients occurring within a few picoseconds should be due
to CR [Eq. (5)] of the dye radical cation (Dye +) with elec-
C
trons (TiO2(eꢀ)’) in surface and/or shallow trap sites, where-
as the substituent-dependent decays observed in longer time
domains would involve participation of electrons
(TiO2(eꢀ)’’) in inner-trap, interstitial-trap, and/or conduc-
tion-band states [Eq. (6)].[14c,23,24b,25,36] The former CR proc-
ess is too fast to allow significant participation of other com-
peting pathways such as electron collection at Pt sites
+
C
[Eq. (7)] and electron transfer from SD (EDTA) to Dye
,
both of which are essential for photocatalytic H2 generation.
However, this very fast decay is a minor component. The
major slower-decay components are certainly due to longer-
lived electrons (TiO2(eꢀ)’’), which should play crucial roles
+
C
in net H2 generation [Eq. (11) following reduction of Dye
by SD [Eqs. (9) and (10); SDox stands for the oxidized prod-
ucts of SD (EDTA)].
ꢀ
Cþ
Dye=TiO2=Pt þ hn ! Dye =TiO2ðe Þ=Pt t < 100 fs
ð3Þ
ꢀ
ꢀ
0
ꢀ 00
Cþ
Cþ
Dye =TiO2ðe Þ=Pt ! Dye =TiO2ðe Þ and TiO2ðe Þ =Pt
ð4Þ
ð5Þ
ð6Þ
ð7Þ
ꢀ
0
Cþ
Dye =TiO2ðe Þ =Pt ! Dye=TiO2=Pt t < 5 ps
ꢀ
00
Cþ
Dye =TiO2ðe Þ =Pt ! Dye=TiO2=Pt t > 5 ps
ꢀ
00
ꢀ
Cþ
Cþ
Dye =TiO2ðe Þ =Pt ! Dye =TiO2=Ptðe Þ
ꢀ
00
ꢀ 00
Cþ
Dye =TiO2ðe Þ =Pt þ SD ! Dye=TiO2ðe Þ =Pt þ SDox
ð8Þ
ð9Þ
ꢀ
ꢀ
Cþ
Dye =TiO2=Ptðe Þ þ SD ! Dye=TiO2=Ptðe Þ þ SDox
Dye=TiO2ðeꢀÞ00=Pt ! Dye=TiO2=PtðeꢀÞ
ð10Þ
ð11Þ
Dye=TiO2=PtðeꢀÞ þ Hþ ! Dye=TiO2=Pt þ 0:5 H2
Extensive investigations of fast events occurring after
photosensitized electron injection from a dye into TiO2 or
on bandgap excitation of TiO2 exemplified the dynamic be-
havior of injected or excited electrons that involves very fast
trapping of the electrons followed by multimode transient
decays due to various CR processes by participation of “var-
ious” electrons in energetically distributed traps and in the
conduction band. On the other hand, Hoffman et al. report-
ed that electrons in the bulk phase formed by bandgap exci-
The substituent-dependent decay behavior of the MOD
and PD systems indicates that the CR process of Equa-
tion (6) is controlled by the hydrophilic/hydrophobic differ-
ence of the substituent, because the other conditions are es-
sentially identical. This means that solvent reorganization
should be an essential factor controlling the CR process. In
the case of MOD, water molecules can form hydrogen
bonds with the substituent to specifically solvate the dye
+
C
donor part where the positive charge of Dye is dominantly
Chem. Eur. J. 2012, 18, 15368 – 15381
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
15375

C. Pac, H. Inoue, S. O. Kang et al.
populated, while such solvation cannot be expected for fully
hydrophobic PD. In relation to solvent reorganization, the
spectral shape of the MOD sample in Figure 4 is slightly,
but appreciably, changed within a few tens of picoseconds,
unlike the PD sample, which shows little change in spectral
shape. This change can be explicitly seen in Figure 9 (top), a
and AOT ionic reverse micelles or lamellae (18 ps).[41] Al-
though much shorter time constants were reported for iner-
tial (librational) motions of water molecules accompanied
by the electronic transitions of coumarin dyes,[42a,b] it is inter-
esting that water reorganization associated with the excita-
tion of eosin Y adsorbed on the surface of ZrO2 particles
was reported to occur with two time components (ꢁ300 fs
and 10 ps), the latter of which is twice as large as that for an
aqueous solution (ꢁ5 ps).[42c] Presumably, the water mole-
cules that have already formed solvation shells around the
two hydrophilic groups of a neutral MOD molecule by hy-
drogen bonding may be able to further solvate the positively
+
C
charged donor part of MOD and thus perturb the elec-
tronic property (e.g., enhanced localization of the positive
+
C
charge on the diphenylamino donor end). By contrast, PD
would still be highly hydrophobic, and not susceptible to ap-
preciable solvation-induced stabilization. Such different sol-
+
C
vation features in the relaxed state of Dye might give rise
+
C
to different electronic couplings between Dye
and
TiO2(eꢀ)’’ for the CR process of Equation (6).
Another important factor for the different transient-decay
behavior of the two samples is the orientation of the dye
molecule on the TiO2 surface. As shown in Figure 8, the dye
molecular plane is not orientationally fixed but more or less
+
C
mobile, so that the positively charged donor part of Dye
may be close to or distant from the TiO2 surface depending
on whether the orientation is inclined or upright. Since elec-
trons trapped in surface or inner traps are not fixed in a spe-
cific site of TiO2 but can rapidly move with a low barrier
(1.4 or 3.3 meV each),[43] CR can be expected to occur any-
where around the electron-injected site on the TiO2 surface
depending on orientation mobility and distributions. In par-
ticular, the substituent-dependent “slow” decay [Eq. (6)]
should be more relevant to such effects than the substituent-
independent “fast” decay [Eq. (5)]. In this respect, an inter-
esting observation was reported that a porphyrin molecule
bearing two cationic anchoring groups at adjacent meso po-
sitions and two hydrophobic phenyl groups at the opposite
two meso positions is adsorbed on the anionic sites of syn-
thetic-clay surface with a “flat orientation” under water, in
contrast to a “standing-up orientation” under dioxane/wa-
ter.[44a,b] For the system of fully hydrophobic PD to be stabi-
lized, the dye molecule would tend to minimize interactions
with the bulk water phase but interact with the TiO2 surface
Figure 9. Time-dependent changes of transient spectra redrawn by nor-
malizing the spectra of Figure 4 at 740 nm for the MOD/np-TiO2 sample
(top) and for the PD/np-TiO2 sample (bottom).
modified version of Figure 4 that has been redrawn after
normalizing the spectra at 740 nm. While the normalized
spectra of the PD sample at different delay times are essen-
tially superimposed (Figure 9, bottom), the MOD sample re-
veals a decrease of relative absorbance at 600–700 nm
within 50 ps. This indicates that a redshift of the transient
spectra occurs with an elapsed time of less than50 ps.
by adopting a lying-down orientation. If this were the case,
+
C
the positive hole of PD should be located close to the
Analysis of the time-dependent change of relative absorb-
ance at 640 nm in the normalized spectra of the MOD
sample gave a time constant of about 17 ps (Supporting In-
formation Figure S7). This dynamic spectral change should
TiO2 surface, a situation certainly favorable for CR with
ꢀ
TiO2(e )’’. In the case of MOD +, on the other hand, the hy-
C
drophilic methoxymethyl-substituted donor end is sufficient-
ly solvated with water molecules to be exposed to the bulk
water phase. As a consequence, the dye molecule may adopt
a standing configuration in which the donor end should be
distant from the TiO2 surface. Since molecular orientations
on solid surfaces are of general interest and related to pho-
tophysical behavior at molecule/solid interfaces,[44c,d] we are
now investigating possible orientations of MOD, PD, and re-
lated dye molecules on the TiO2 surface under water.
be related to solvent reorganization accompanied by the
+
C
change from the Franck–Condon state of MOD to the re-
laxed state. The value of about 17 ps observed in the present
study is very similar to the time constants reported for ori-
entational relaxation of water molecules at the interfaces
with
a tetraethylene glycol dimethyl ether molecule
(ꢁ19 ps), Igepal co-520 neutral reverse micelles (ꢁ13 ps),
15376
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 15368 – 15381

Hydrogen Evolution in Dye/TiO2/Pt Hybrid Systems
FULL PAPER
Another essential process associated with H2 generation is
nophenyl donor end; the efficiency is highest with hydro-
philic MOD but about three times less efficient with hydro-
phobic PD. The apparent quantum yield of photocatalyzed
H2 generation at 436 nm for the MOD/TiO2/Pt hybrid was
determined to be 0.27ꢂ0.03. Transient spectroscopic investi-
gations on MOD- and PD-grafted transparent films of TiO2
+
C
reduction of Dye by EDTA [Eqs. (9) and (10)]. Therefore,
we attempted to observe possible effects of EDTA on the
transient behavior of the Dye/np-TiO2 films. However, the
decay behavior of transients was not affected by EDTA,
+
C
that is, reduction of Dye by about 5 mm EDTA in the film
system should proceed in a microsecond time domain. This
nanoparticles dipped into water at pH 3 showed ultrafast (<
+
+
C
C
is not unexpected. If the reduction of Dye by EDTA in
100 fs) formation of Dye followed by multimode decays
homogeneous aqueous solution can occur at the diffusion-
controlled limit (ꢁ5ꢁ109 mꢀ1 sꢀ1), the maximum reaction
rate at about 5 mm of EDTA should be around 2.5ꢁ107 sꢀ1.
This is, however, clearly not the case for the nanoparticle
film. Under the present conditions, where the absorbance of
including minor substituent-independent components of
<5 ps and major substituent-dependent components with
longer time constants extending to microseconds. The major
substituent-dependent decay components in longer time do-
+
C
mains were attributed to slower CR of Dye with electrons
the films is relatively low (ꢁ0.13 or 0.18), the excitation of
in inner-trap, interstitial-trap, and/or bulk states, which
occurs with approximately threefold difference between the
two Dye/np-TiO2 systems. The observed differences in H2
generation and in transient decay behavior between the
MOD and PD systems were discussed as the consequences
of different solvent reorganization around the hydrophilic or
hydrophobic substituent, as indicated by a dynamic redshift
of the transient spectra for the MOD/np-TiO2 film in con-
trast to little shift in the case of PD/np-TiO2; the redshift,
which occurs within 17 ps, is attributable to solvation-in-
+
C
Dye should equally occur through the film, giving Dye
+
C
mostly inside the film. For reduction of Dye to occur,
EDTA molecules must penetrate into mesopores inside the
+
C
film and hence must diffuse there to find Dye fixed to the
heterogeneous TiO2 surface. Consequently, the real event
occurring in the film system would proceed with a time dis-
tribution much longer than the time expected from the
above rate constant in homogeneous solution. In the case of
the dispersion system, in which all dye molecules are ex-
posed to bulk water, on the other hand, it is not unreasona-
duced relaxation of MOD +. On the basis of the present
C
+
C
ble to expect that reduction of Dye by EDTA would occur
findings, we are aiming to achieve further improvements in
photocatalysis efficiency of Dye/TiO2/Pt hybrid systems by
molecular design of the dyes and engineering of TiO2 parti-
cles.
faster than that in the film system. Furthermore, it was re-
ported for TiO2 dispersions in water that, on Pt loading of
TiO2 particles, trapped electrons can rapidly move to the Pt
site [Eq. (8)] in a picosecond time region before and/or in
competition with CR.[45] This is in line with the low activa-
tion barriers for diffusion of trapped electrons.[43] According
Experimental Section
to these reported results, Equation (10) might be an impor-
+
C
tant process for the reduction of Dye by EDTA that
General procedures: All manipulations were performed under a dry ni-
trogen or argon atmosphere by using standard Schlenk techniques. THF
was dried over sodium/benzophenone. The 1H and 13C NMR spectra
were recorded on a Varian Mercury 300 spectrometer operating at 300.1
and 75.4 MHz, respectively, for CDCl3 or [D6]DMSO solutions. Proton
chemical shifts were referenced relative to the corresponding solvent sig-
nals (dH =7.26 for CDCl3 and dH =2.49 for [D6]DMSO). Column chro-
matography was performed on Merck 70D silica gel (230 mesh), and
TLC on silica-gel plates (Merck, Kieselgel 60 F254, 0.20 mm). Elemental
analyses and high-resolution tandem mass spectrometry were performed
respectively on a Carlo Erba Instruments CHNS-O EA 1108 analyzer
and on a Jeol LTD JMS-HX 110/110A at the Korean Basic Science Insti-
tute (Ochang). The absorption and photoluminescence spectra were re-
corded on a Shimadzu UV-3101PC UV/Vis/NIR scanning spectropho-
tometer and a VARIAN Cary Eclipse fluorescence spectrophotometer,
respectively. The diffuse-reflectance UV/V absorption spectra of powder
samples were recorded on a Scinco spectrophotometer S-3100. Particle
sizes of the TiO2 samples were measured on a Nicomp 380 dynamic light
scattering analyzer. Cyclic voltammetry (CV) measurements were carried
out for THF solutions of 1 mm organic dye and 0.1m tetrabutylammoni-
um hexafluorophosphate at room temperature by using a BAS 100B elec-
trochemical analyzer equipped with a Pt working electrode, a platinum
wire counterelectrode, and an AgjAgNO3 (0.1m) reference. All poten-
tials were calibrated against a ferrocenium/ferrocene (Fc+/Fc) redox
couple (ꢀ0.63 V).
occurs after fast electron collection at the Pt site. Further-
more, it can be expected that hydrophilic EDTA molecules
+
C
can approach the hydrophilic donor end of MOD more fa-
vorably than the hydrophobic PD +, as previously suggest-
C
ed.[21]
Finally, we note that the observed AQY value is not the
real quantum yield with scientific definition, but a relative
measure for the estimation of H2 generation efficiency, be-
cause the number of photons absorbed by Dye cannot be
properly estimated. Moreover, the so-called current-dou-
bling effect of EDTA[49] should be taken into consideration
for the relatively high AQY. Anyway, more than 54% of the
total electrons injected into TiO2 from excited Dye as well
as from oxidized EDTA radical[49a] can be utilized for net H2
generation by the MOD/TiO2/Pt hybrid.
Conclusion
We have investigated the visible-light-driven H2 evolution
by Dye/TiO2/Pt hybrid photocatalysts in the presence of
about 5 mm EDTA. Photocatalyzed H2 evolution from water
at pH 3 revealed clear dependence on the hydrophilic/hy-
drophobic character of the substituent attached to the ami-
Fluorescence lifetimes: Lifetime were measured by
a single-photon
counting method by using a streakscope (Hamamatsu Photonics, C4334–
01), a polychromator (Acton Research, SpectraPro150), and a Ti:sap-
phire laser (Spectra-Physics, Tsunami 3941M1BB, FWHM 100 fs)
pumped with a diode solid-state laser (Spectra-Physics, Millennia VIIIs).
Chem. Eur. J. 2012, 18, 15368 – 15381
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
15377

C. Pac, H. Inoue, S. O. Kang et al.
For excitation of the samples, the output of the Ti:sapphire laser was con-
verted to the second harmonic (400 nm) with a harmonic generator
(Spectra-Physics, GWU-23FL). The instrument response function was de-
termined by analyzing temporal profiles of the scattered laser light to
give a time resolution of approximately 20 ps after deconvolution. The
observed temporal emission profiles were well fitted to single- or double-
exponential functions. The residuals were less than 1.1% for each
system.
trifluoroacetic acid (30 mL) were added to acetal 4 (R1 =CH3CH2CH2,
R2 =H; 0.7 g, 1.15 mmol). The resulting mixture was stirred for 2 h at
room temperature, quenched with saturated aqueous sodium bicarbonate,
and extracted with dichloromethane. The combined dichloromethane
phase was washed with aqueous sodium bicarbonate and dried over
MgSO4. After removal of the solvent under reduced pressure, silica-gel
column chromatography of the residue with ethyl acetate/hexane (1:3) as
eluent gave the product as an orange oil. Yield: 0.46 g (77%). 1H NMR
(CDCl3): d=9.85 (s, 1H, COH), 7.67 (d, 1H, ArH), 7.43 (d, 2H, ArH),
7.30 (d, 1H, ArH), 7.23 (d, 1H, ArH), 7.15 (d, 1H, ArH), 7.00–7.11 (m,
10H, ArH), 2.56 (t, 4H, CH2), 1.66 (t, 4H, CH2), 0.97 (t, 6H, CH3);
HRMS (FAB) calcd for C33H31NOS2: 521.1847; found: 521.1851 [M]+; el-
emental analysis (%) calcd for C33H31NOS2: C 75.97, H 5.99, N 2.68;
found: C 75.93, H 5.97, N 2.66.
Transient absorption spectroscopy: The experiments were performed for
transparent thin films of dye-loaded TiO2 nanoparticles on quartz plates
dipped into water at pH 3 and for homogeneous dye solutions in a rotary
cell with an optical path length of 1 mm. For the observation of short-
lived transient absorption (1 ps to 1.5 ns), a double-beam subpicosecond
pump–probe technique was adopted. The pump and probe pulses were
generated from the fundamental of an amplified Ti-sapphire laser (CDP,
Tissa50+MPA50; 800 nm, 1 kHz). The second harmonic of the funda-
mental (400 nm, FWHM 100 fs, 1 kHz) was used for excitation of the
samples. The continuum probe pulse was generated by focusing the fun-
damental onto a sapphire plate. After laser-pulse irradiation of the
sample, the probe light was recorded with a dual photodiode array
(CDP) combined with a polychromator (SOLAR TII, MS 3504i). Transi-
ents in a longer time region (10 ns to 15 ms) were observed by irradiation
of the samples at 416 nm with the third harmonic of a Q-switched
Nd:YAG laser (Continuum, Surelite II, pulse width of 4.5 ns) combined
with a home-made Raman shifter and a xenon lamp (ILC Technology, PS
300–1) as probe light. Temporal profiles were measured by a photomulti-
plier (Zolix Instruments Co., CR 131) and a digital oscilloscope (Tektro-
nix, TDS-784D) equipped with a monochromator (DongWoo Optron,
Monora 500i). Reported signals were obtained as averages of 500 events.
(E)-3-(5’-{4-[Bis(4-propylphenyl)amino]phenyl}-2,2’-bithiophen-5-yl)-2-
cyanoacrylic acid (PD): An acetonitrile solution (10 mL) containing 5
(R1 =CH3CH2CH2, R2 =H; 0.45 g, 0.86 mmol), cyanoacetic acid (73 mg,
0.86 mmol), and piperidine (0.1 mL, 1.05 mmol) was heated to reflux for
6 h, cooled to room temperature, and then quenched with 1n HCl. After
removal of acetonitrile under reduced pressure, the crude product was
extracted with dichloromethane and dried over anhydrous MgSO4. After
removal of the solvent under reduced pressure, silica-gel column chroma-
tography of the residue with ethyl acetate/methanol (10:1) as eluent gave
the product PD as
a
red powder. Yield: 0.35 g (82%). 1H NMR
([D6]DMSO): d=8.45 (s, 1H, =CH), 7.95 (d, 1H, ArH), 7.59 (t, 4H,
ArH), 7.44 (d, 1H, ArH), 7.15 (d, 4H, ArH), 6.98 (d, 4H, ArH), 6.87 (d,
2H, ArH), 2.51 (t, 4H, CH2), 1.58 (t, 4H, CH2), 0.90 (t, 6H, CH3);
HRMS (FAB) calcd for C36H32N2O2S2: 588.1905; found: 588.1902 [M]+;
elemental analysis (%) calcd for C33H31NOS2: C 73.44, H 5.48, N 4.76;
found: C 73.41, H 5.47, N 4.75.
The films of dye-loaded TiO2 nanoparticles (Dye/np-TiO2) used for tran-
sient spectroscopy were prepared as previously reported.[46] Quartz plates
(8ꢁ15 mm, 1.0 mm thick) for nanosecond measurements and round
quartz disks (30 mm diameter, 1.0 mm thick) for femtosecond measure-
ments were ultrasonically treated in a dilute solution of Merck Extran
MA02 for 30 min, thoroughly washed with distilled water, and then dried
under vacuum. After spin-coating a titanium(IV) bis(ethylacetoacetato)
diisopropoxide solution (2 wt% in 1-butanol) onto the cleaned quartz
plate, the coated plate was heated stepwise to 4508C and maintained at
this temperature for 20 min. A TiO2 nanoparticle paste (15–20 nm, Ti-
Nanoxide T, Solaronix) was cast onto the treated quartz plates by a
doctor-blade technique and then sintered at 4508C for 30 min to give
nanocrystalline anatase TiO2 films with a thickness of 13–18 mm. The
TiO2 films on quartz were dipped into a THF solution of PD or MOD
(0.3 mm) overnight, washed three times with THF, and then dried in an
oven in the dark. Adsorption of the dyes on the TiO2 films was con-
firmed by UV/Vis absorption spectroscopy.
5,5’-Dibromo-4,4’-bis(methoxymethyl)-2,2’-bithiophene: 2-Bromo-3-(me-
thoxymethyl)thiophene (2.0 g, 9.66 mmol), [PdCl2ACHTNUTRGNE(UNG PhCN)2] (0.11 g,
0.29 mmol), potassium fluoride (1.12 g, 19.3 mmol), DMSO (60 mL), and
then AgNO3 (3.28 g, 19.3 mmol) in one portion were added to a screw-
capped glass tube equipped with a magnetic stirring bar. The mixture
was heated with stirring at 608C for 3 h. Additional AgNO3 (3.28 g,
19.3 mmol) and potassium fluoride (1.12 g, 19.3 mmol) were added, and
then the reaction mixture was stirred for 6 h. The reaction mixture was
passed through a Celite pad to remove a solid residue and washed re-
peatedly with diethyl ether. The filtrate was washed with water and the
combined organic layer was dried over anhydrous MgSO4. Evaporation
under reduced pressure left a crude solid, which was purified by chroma-
tography on silica gel to give the product. Yield: 1.75 g (88%). 1H NMR
(CDCl3): d=6.97 (s, 2H, ArH), 4.35 (s, 4H, CH2), 3.39 (s, 6H, CH3);
HRMS (FAB) calcd for C12H12Br2O2S2: 409.8645; found: 409.8642 [M]+;
elemental analysis (%) calcd for C12H12Br2O2S2: C 34.97, H 2.93; found:
C 34.93, H 2.91.
Synthesis: Benzyl bromide, 4-bromoaniline, and 1-bromo-4-propylben-
zene were purchased from Aldrich. The preparation of HD, MOD, and
bithiophene 3 (R2 =H in Scheme 1) was performed according to the
method previously reported.[29] Synthetic details for PD and MO4D are
described below.
5’-Bromo-4,4’-bis(methoxymethyl)-2,2’-bithiophene-5-carbaldehyde:
Phosphorus oxychloride (2.21 g, 14.4 mmol) and DMF (75 mL) were
added to a round-bottom flask kept at 08C, and the mixture was stirred
for 30 min. To the mixture was added 5,5’-dibromo-4,4’-bis(methoxymeth-
yl)-2,2’-bithiophene (1.5 g, 3.6 mmol), and the solution was stirred at
room temperature for 30 min and then at 508C until the reaction was
complete (checked by TLC). Dilute hydrochloric acid was added to the
reaction mixture at 08C, and the organic layer was extracted with ethyl
acetate. After usual workup, column chromatography of the mixture on
silica gel with ethyl acetate/hexane (1:3) as eluent gave the product.
Yield: 0.95 g (73%). 1H NMR (CDCl3): d=10.01 (s, 1H, COH), 7.19 (s,
1H, ArH), 7.16 (s, 1H, ArH), 4.72 (s, 2H, CH2), 4.36 (s, 2H, CH2), 3.45
(s, 3H, CH3), 3.40 (s, 3H, CH3); HRMS (FAB) calcd for C13H13BrO3S2:
359.9489; found: 359.9486 [M]+; elemental analysis (%) calcd for
C13H13BrO3S2: C 43.22, H 3.63; found: C 43.19, H 3.62.
4-[5’-(5,5-Dimethyl-1,3-dioxan-2-yl)-2,2’-bithiophen-5-yl]-N,N-bis(4-pro-
pylphenyl)benzenamine (4, R1 =CH3CH2CH2, R2 =H): A toluene solu-
tion (20 mL) containing a mixture of 3 (R2 =H)[29] (0.50 g, 1.35 mmol), 1-
bromo-4-propylbenzene (0.62 g, 3.1 mmol), [Pd2ACHTNUTRGNE(UNG dba)3] (0.085 g,
3 mol%), Xantphos (0.054 g, 3 mol%), and NaOtBu (1.04 g, 10.8 mmol)
was stirred under N2 at 1108C for 16 h, hydrolyzed with water, extracted
with dichloromethane, and dried over MgSO4. After removal of the sol-
vent under reduced pressure, silica-gel column chromatography of the
residue with ethyl acetate/hexane (1:5) as eluent gave 4 (R2 =H) as a
1
yellow oil. Yield: 0.71 g (87%). H NMR (CDCl3): d=7.41 (d, 2H, ArH),
7.00–7.09 (m, 14H, ArH), 5.62 (s, 1H, CH), 3.77 (d, 2H, CH2), 3.65 (d,
2H, CH2), 2.55 (t, 4H, CH2), 1.64 (q, 4H, CH2), 1.29 (s, 3H, CH3), 0.96
(t, 6H, CH3), 0.81 (s, 3H, CH3); HRMS (FAB) calcd for C38H41NO2S2:
607.2579; found: 607.2584 [M]+; elemental analysis (%) calcd for
C38H41NO2S2: C 75.08, H 6.80, N 2.30; found: C 75.04, H 6.79, N 2.29.
5-Bromo-5’-(5,5-dimethyl-1,3-dioxan-2-yl)-4,4’-bis(methoxymethyl)-2,2’-
bithiophene (1, R2 =CH3OCH2): 5’-Bromo-4,4’-bis(methoxymethyl)-2,2’-
bithiophene-5-carbaldehyde (0.9 g, 2.5 mmol), neopentyl glycol (0.31 g,
3.0 mmol), p-toluenesulfonic acid (0.05 g), and benzene (10 mL) were
added to a round-bottom flask, and the mixture was heated to reflux for
5 h. After cooling to room temperature, the solution was washed with
5’-{4-[Bis(4-propylphenyl)amino]phenyl}-2,2’-bithiophene-5-carbaldehyde
(5, R1 =CH3CH2CH2, R2 =H): THF (100 mL), water (25 mL), and then
15378
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 15368 – 15381

Hydrogen Evolution in Dye/TiO2/Pt Hybrid Systems
FULL PAPER
aqueous sodium bicarbonate and then dried over Na2SO4. After removal
of the solvent under reduced pressure, silica-gel column chromatography
of the residue gave
(s, 4H, CH2), 4.39 (s, 2H, CH2), 3.44 (s, 3H, CH3), 3.41 (s, 9H, CH3);
HRMS (FAB) calcd for C35H35NO5S2: 613.1957; found: 613.1950 [M]+;
elemental analysis (%) calcd for C35H35NO5S2: C 68.49, H 5.75, N 2.28;
found: C 68.41, H 5.73, N 2.27.
1
(R1 =R2 =CH2OCH3). Yield: 0.91 g (81%).
1H NMR (CDCl3): d=7.00 (s, 2H, ArH), 5.67 (s, 1H, CH), 4.42 (s, 2H,
CH2), 4.33 (s, 2H, CH2), 3.72 (d, 2H, CH2), 3.60 (d, 2H, CH2), 3.35 (s,
6H, CH3), 1.26 (s, 3H, CH3), 0.76 (s, 3H, CH3); HRMS (FAB) calcd for
C18H23BrO4S2: 446.0221; found: 446.0220 [M]+; elemental analysis (%)
calcd for C18H23BrO4S2: C 48.32, H 5.18; found: C 48.29, H 5.16.
(E)-3-(5’-{4-[N,N-Bis(4-methoxymethylphenyl)amino]phenyl}-4,4’-bis(me-
thoxymethyl)-2,2’-bithiophen-5-yl)-2-cyanoacrylic acid (MO4D): This
compound was prepared by a procedure similar to that used for PD. The
product MO4D was obtained as a red powder. Yield: 0.21 g (75%).
1H NMR ([D6]DMSO): d=8.08 (s, 1H, =CH), 7.42 (s, 1H, ArH), 7.37 (d,
2H, ArH), 7.34 (s, 1H, ArH), 7.24 (d, 4H, ArH), 7.02 (d, 4H, ArH), 6.94
(d, 2H, ArH), 4.47 (s, 2H, CH2), 4.32 (s, 6H, CH2), 3.27 (s, 6H, CH3),
3.25 (s, 6H, CH3); HRMS (FAB) calcd for C38H36N2O6S2: 680.2015;
found: 680.2009 [M]+; elemental analysis (%) calcd for C38H36N2O6S2: C
67.04, H 5.33, N 4.11; found: C 67.00, H 5.31, N 4.10.
2-[5’-(5,5-Dimethyl-1,3-dioxan-2-yl)-4,4’-bis(methoxymethyl)-2,2’-bithio-
phen-5-yl]-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (2, R2 =CH3OCH2): A
2.5m solution of n-BuLi in hexane (0.8 mL, 1.98 mmol) was added to a
solution of 1 (R2 =CH3OCH2; (0.9 g, 1.8 mmol) in THF (10 mL) at
ꢀ788C, and the mixture was stirred for 1 h. After adding 2-isopropoxy-
4,4,5,5-tetramethyl-1,3,2-dioxaborolane (0.41 mL, 1.98 mmol) to the mix-
ture, the solution was warmed to room temperature and stirred for 12 h.
After quenching with H2O followed by extraction with dichloromethane,
the combined dichloromethane layer was washed with aqueous sodium
bicarbonate and dried over MgSO4. After removal of the solvent under
reduced pressure, silica-gel column chromatography of the residue with
ethyl acetate/hexane (1:3) as eluent gave 2 (R2 =CH3OCH2). Yield:
0.67 g (76%). 1H NMR (CDCl3): d=7.29 (s, 1H, ArH), 7.12 (s, 1H,
ArH), 5.70 (s, 1H, CH), 4.64 (s, 2H, CH2), 4.46 (s, 2H, CH2), 3.75 (d,
2H, CH2), 3.63 (d, 2H, CH2), 3.38 (s, 3H, CH3), 3.36 (s, 3H, CH3), 1.56
(s, 3H, CH3), 1.33 (s, 12H, CH3), 0.79 (s, 3H, CH3); HRMS (FAB) calcd
for C24H35BO6S2: 494.1968; found: 494.1965 [M]+; elemental analysis (%)
calcd for C24H35BO6S2: C 58.30, H 7.13; found: C 58.27, H 7.11.
Preparation of Dye/TiO2/Pt: Surface-platinized TiO2 particles (TiO2/Pt,
0.3–1 wt% Pt loading) were prepared by a photodeposition method as
described elsewhere.[31] Commercially available anatase TiO2 particles
(Hombikat UV-100, primary particle size <10 nm, BET surface area
>250 m2 gꢀ1) were used after washing with distilled water or after ultra-
sonic treatment of the washed TiO2 in water. The particles were collected
by filtration or by centrifugation and then dried in an oven under N2. A
water suspension of the TiO2 particles (0.5 gLꢀ1) containing ꢁ1m metha-
nol and 0.03–0.1 mm chloroplatinic acid was irradiated with a 200 W mer-
cury lamp for 30 min. After irradiation, the Pt-loaded TiO2 powders were
separated by centrifugation, washed with distilled water, dried, and
stored under N2 in the dark. The Pt-loaded TiO2 powders (0.1 g) were
dispersed in an acetonitrile/tert-butanol solution of the organic dyes (1.0–
100 mmol) and stirred overnight. The dye-loaded photocatalysts (Dye/
TiO2/Pt) were separated by centrifugation, washed with acetonitrile/tert-
butanol, dried in an oven (858C), and stored under N2 in the dark.
4-[5’-(5,5-Dimethyl-1,3-dioxan-2-yl)-4,4’-bis(methoxymethyl)-2,2’-bithio-
phen-5-yl]benzenamine (3, R2 =CH3OCH2): Compound
2
(R2 =
CH3OCH2; 0.65 g, 1.31 mmol), potassium carbonate (0.72 g, 5.24 mmol),
and water (3 mL) were added to a stirred solution of 4-bromoaniline
(0.22 g, 1.31 mmol) and [PdACHTNUTRGNEUNG(PPh3)4] (45 mg, 0.04 mmol) in THF (10 mL).
The mixture was refluxed for 12 h, cooled to room temperature, extract-
ed with dichloromethane, and dried over anhydrous MgSO4. After re-
moval of the solvent under reduced pressure, flash chromatography of
the residue over silica gel with ethyl acetate/hexane (1:2) as eluent gave
3 (R2 =CH3OCH2) as a yellow powder. Yield: 0.45 g (75%). 1H NMR
(CDCl3): d=7.29 (d, 2H, ArH), 7.17 (s, 1H, ArH), 7.05 (s, 1H, ArH),
6.72 (s, 2H, ArH), 5.71 (s, 1H, CH), 4.46 (s, 2H, CH2), 4.35 (s, 2H, CH2),
3.78 (s, 2H, NH2), 3.76 (d, 2H, CH2), 3.64 (d, 2H, CH2), 3.38 (s, 3H,
CH3), 3.37 (s, 3H, CH3), 1.56 (s, 3H, CH3), 0.80 (s, 3H, CH3); HRMS
(FAB) calcd for C24H29NO4S2: 459.1538; found: 459.1533 [M]+; elemental
analysis (%) calcd for C24H29NO4S2: C 62.72, H 6.36, N 3.05; found: C
62.69, H 6.34, N 3.04.
Photocatalyzed hydrogen evolution: Aqueous suspensions (3 mL) of
Dye/TiO2/Pt particles (10 mg) containing EDTA (ꢁ5 mm) were placed in
a quartz cell and adjusted to pH 3 with 1n HClO4, bubbled with an N2
stream for 30 min, sealed with a septum, and then irradiated with stirring
by a xenon lamp (450 W, model 66924, Newport corporation); the inci-
dent light (l>420 nm) was obtained by passing the light from the xenon
lamp through a 10 cm water layer and a cutoff glass filter. The amounts
of H2 evolved were determined by gas chromatography (HP6890A GC
equipped with a TCD detector) on a 5 ꢃ molecular sieve column. Appa-
rent quantum yield F(H2) for H2 generation was determined for the
MOD/TiO2/Pt suspensions, a band-pass filter (415–450 nm) was used to
isolate the 436 nm light from the emission light of a high-pressure mercu-
ry lamp (1000 W, model 6171, Newport Corporation), and the incident
light flux was determined by using a 0.2m ferrioxalate actinometer solu-
tion.[47]
4-[5’-(5,5-Dimethyl-1,3-dioxan-2-yl)-4,4’-bis(methoxymethyl)-2,2’-bithio-
phen-5-yl]-N,N-bis[4-(methoxymethyl)phenyl]benzenamine (4, R1 =R2 =
CH3OCH2): A toluene solution (10 mL) containing a mixture of 1-
bromo-4-(methoxymethyl)benzene (0.50 g, 2.5 mmol), 3 (R2 =CH3OCH2)
(0.44 g, 0.95 mmol), [Pd2ACHTNUGTRNEUNG(dba)3] (52 mg, 3 mol%), Xantphos (33 mg,
Theoretical methods: DFT calculations for energetically stabilized molec-
ular orientations of MOD and PD adsorbed on the (101) facet of anatase
TiO2 were performed by changing the angle of the molecular plane with
respect to the TiO2 surface by using the VASP code with the generalized
gradient approximation (GGA), periodic boundary condition, plane
wave basis sets, and projector augmented wave (PAW) ultrasoft pseudo-
potentials.[48] The electronic wave functions of the model structures were
represented by plane wave basis with a cutoff energy of 400 eV. The size
of the unit cell for an isolated dye was 15.83ꢁ4.28ꢁ21.08 ꢃ. The TiO2
surface structure was modeled for three layers with 144 atoms (48 Ti
atoms and 96 O atoms), in which one TiO2 layer consists of one titanium
layer and two oxygen layers. The size of unit cell was 10.18ꢁ15.08ꢁ
9.32 ꢃ. DFT calculations of the dyes adsorbed on TiO2 surface were per-
formed with a much larger super cell of size 10.18ꢁ15.08ꢁ42.00 ꢃ and
with two different initial geometric positions of the dyes. For the energet-
ically favorable structures obtained by structure optimization calculation,
the dye molecular plane was subsequently inclined on the fixed TiO2 sur-
face in 108 steps to 908 to both the right- and left-hand sides. All total en-
ergies at each angle were measured and the most stable case was found
when the dyes have a specific angle on the TiO2 surface.
3 mol%), and NaOtBu (0.64 g, 6.65 mmol) was stirred under N2 at 1108C
for 12 h, hydrolyzed with water, extracted with dichloromethane, and
dried over MgSO4. After removal of the solvent under reduced pressure,
silica-gel column chromatography of the residue with ethyl acetate/
hexane (1:4) as eluent gave 4 (R1 =R2 =CH3OCH2) as a yellow oil.
Yield: 0.45 g (68%). 1H NMR (CDCl3): d=7.33 (d, 2H, ArH), 7.25 (d,
2H, ArH), 7.19 (s, 2H, ArH), 7.05–7.13 (m, 8H, ArH), 5.71 (s, 1H, CH),
4.47 (s, 2H, CH2), 4.41 (s, 4H, CH2), 4.38 (s, 2H, CH2), 3.76 (d, 2H,
CH2), 3.64 (d, 2H, CH2), 3.42 (s, 6H, CH3), 3.39 (s, 3H, CH3), 3.37 (s,
3H, CH3), 1.56 (s, 3H, CH3), 0.79 (s, 3H, CH3); HRMS (FAB) calcd for
C40H45NO6S2: 699.2688; found: 699.2681 [M]+; elemental analysis (%)
calcd for C40H45NO6S2: C 68.64, H 6.48, N 2.00; found: C 68.60, H 6.46,
N 2.01.
5’-{4-[N,N-Bis(4-methoxymethylphenyl)amino]phenyl}-4,4’-bis(methoxy-
methyl)-2,2’-bithiophene-5-carbaldehyde (5, R1 =R2 =CH3OCH2): This
compound was prepared from 4 (R1 =R2 =CH3OCH2) by a procedure
similar to that used for 5 (R1 =CH3CH2CH2, R2 =H). The product 5
(R1 =R2 =CH3OCH2) was obtained as a yellow oil. Yield: 0.28 g (71%).
1H NMR (CDCl3): d=10.01 (s, 1H, COH), 7.31–7.37 (m, 4H, ArH),
7.22–7.27 (m, 4H, ArH), 7.05–7.12 (m, 6H, ArH), 4.73 (s, 2H, CH2), 4.40
Chem. Eur. J. 2012, 18, 15368 – 15381
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
15379

C. Pac, H. Inoue, S. O. Kang et al.
Brennaman, P. G. Hoertz, A. O. T. Patrocinio, N. Y. Murakami Iha,
J. L. Templeton, T. J. Meyer, Acc. Chem. Res. 2009, 42, 1954; g) W. J.
Youngblood, S.-H. A. Lee, K. Maeda, T. J. Meyer, Acc. Chem. Res.
2009, 42, 1966.
Acknowledgements
This article is dedicated with deepest gratitude and respect to the
memory of the late Professor Dr. Kenichi Honda, one of the earliest pio-
neers of artificial photosynthesis, as a result of the seminal finding of
water splitting into hydrogen and oxygen by means of light irradiation of
semiconducting titanium oxide in 1972 with Professor Akira Fujishima.
This seminal finding has thenceforth become known as the “Honda-Fu-
jishima Effect.” This research was supported by Basic Science Research
Program through the National Research Foundation of Korea (NRF)
funded by the Ministry of Education, Science and Technology (No. 2012-
0006248) and in part by Korea Research Institute of Science and Tech-
nology (No. 2E22142-11-206). H.C and K.C are supported by the NRF of
Korea through WCU program (Grant No. R-31-2009-000-10083-0).
[17] K. Kalyanasundaram, J. Kiwi, M. Grꢄtzel, Helv. Chim. Acta 1978,
61, 2720.
[18] a) J. Zhang, P. Du, J. Schneider, P. Jarosz, R. Eisenberg, J. Am.
Chem. Soc. 2007, 129, 7726; b) E. Reisner, D. J. Powell, C. Cabazza,
J. C. Fontecilla-Campus, F. A. Armstrong, J. Am. Chem. Soc. 2009,
131, 18457.
[19] T. Shimidzu, T. Iyoda, Y. Koide, J. Am. Chem. Soc. 1985, 107, 35.
[20] a) R. Abe, K. Hara, K. Sayama, K. Domen, H. Arakawa, J. Photo-
chem. Photobiol. A 2000, 137, 63; b) R. Abe, K. Sayama, H. Araka-
wa, J. Photochem. Photobiol. A 2004, 166, 115; c) Y. Astuti, E. Palo-
meres, S. A. Haque, J. R. Durrant, J. Am. Chem. Soc. 2005, 127,
15120; d) Q. Li, Z. Jin, Z. Peng, Y. Li, S. Li, G. Lu, J. Phys. Chem. C
2007, 111, 8237; e) Y. Li, C. Xie, S. Peng, G. Lu, S. Li, J. Mol. Catal.
A 2008, 282, 117; f) P. Jarosz, P. Du, S.-H. Lee, D. McCamant, R. Ei-
senberg, Inorg. Chem. 2009, 48, 9653.
[1] A. Fujishima, K. Honda, Nature 1972, 238, 37.
[2] a) M. Grꢄtzel, Acc. Chem. Res. 1981, 14, 376; b) A. J. Bard, M. A.
Fox, Acc. Chem. Res. 1995, 28, 141.
[21] R. Abe, K. Shinmei, K. Hara, B. Ohtani, Chem. Commun. 2009,
3577.
[22] G. J. Meyer, Inorg. Chem. 2005, 44, 6852.
[23] a) J. R. Durrant, S. A. Haque, E. Palomaras, Coord. Chem. Rev.
2004, 248, 1247; b) B. OꢂRegan, J. R. Durrant, P. Sommeling, J. N.
Bakker, J. Phys. Chem. C 2007, 111, 14001.
[24] a) D. E. Skinner, D. P. Colombo, Jr., J. J. Cavaleri, R. Bowman, J.
Phys. Chem. B 1995, 99, 7853; b) D. P. Colombo Jr., R. M. Bowman,
J. Phys. Chem. B 1995, 99, 11752.
[3] a) C. A. Grims, O. K. Vorghese, S. Ranjan, Light, Water, Hydrogen-
The Solar Generation of Hydrogen by Water Photoelectrolysis,
Springer, New York, 2008; b) S. Licht, G. Holdes, in Series on Pho-
toconversion of Solar Energy, Vol. 3 (Eds.: M. D. Archer, A. J.
Nozik, Imperial College Press, Singapore, 2008, pp. 591–632; c) N. S.
Lewis, A. G. Nocera, Proc. Natl. Acad. Sci. USA 2006, 103, 15729.
[4] a) F. M. MacDonnell, in Solar Hydrogen Generation-Toward a Re-
newable Energy Future (Eds.: K. Rajeshwar, R. McConnel, S.
Licht), Springer, New York, 2008, pp. 123–166; b) A. J. Esswein,
D. G. Nocera, Chem. Rev. 2007, 107, 4022; c) T. Shimada, A. Kuma-
gai, S. Funyu, S. Takagi, D. Masui, Y. Nabetani, H. Tachibana, D. A.
Tryk, H. Inoue, Faraday Discuss. 2012, 155, 145.
[25] a) S. H. Szczepankiewicz, A. J. Colussi, M. R. Hoffmann, J. Phys.
Chem. B 2000, 104, 9842; b) S. H. Szczepankiewicz, J. A. Moss,
M. R. Hoffmann, J. Phys. Chem. B 2002, 106, 2922.
[5] a) J. H. Alstrum-Acevedo, M. K. Brennaman, T. J. Meyer, Inorg.
Chem. 2005, 44, 6802; b) P. G. Hoertz, T. E. Mallouk, Inorg. Chem.
2005, 44, 6828; c) S. Chakraborty, T. J. Wadas, H. Hester, R.
Schmehl, R. Eisenberg, Inorg. Chem. 2005, 44, 6865.
[6] a) H. G. Kim, P. H. Borse, J. S. Jang, C. W. Ahn, E. D. Jeong, J. S.
Lee, Adv. Mater. 2011, 23, 2088; b) J. S. Hong, S. Lee, J. S. Jang, J. S.
Lee, Energy Environ. Sci. 2011, 4, 1781.
[7] K. Rajeshwar, in Solar Hydrogen Generation-Toward a Renewable
Energy Future (Eds.: K. Rajeshwar, R. McConnel, S. Licht), Spring-
er, New York, 2008, pp. 167–228.
[8] a) A. Kudo, Y. Miseki, Chem. Soc. Rev. 2009, 38, 253; b) M. Kitano,
M. Hara, J. Mater. Chem. 2010, 20, 627.
[9] a) K. Domen, J. N. Kondo, M. Hara, T. Takata, Bull. Chem. Soc.
Jpn. 2000, 73, 1307; b) K. Maeda, M. Higashi, D. Lu, R. Abe, K.
Domen, J. Am. Chem. Soc. 2010, 132, 5858; c) K. Maeda, R. Abe,
K. Domen, J. Phys. Chem. C 2011, 115, 3057; d) R. Abe, Bull.
Chem. Soc. Jpn. 2011, 84, 1000.
[10] H.-I. Kim, J. Kim, W. Kim, W. Choi, J. Phys. Chem. C 2011, 115,
9797.
[11] a) Y. H. Iwase, Y. Ng, A. Ishiguro, A. Kudo, R. Amal, J. Am. Chem.
Soc. 2011, 133, 11054; b) Y. Miseki, A. Kudo, ChemSusChem 2011,
4, 245.
[26] a) H. N. Ghosh, J. B. Asbury, T. Lian, J. Phys. Chem. B 1998, 102,
6482; b) G. Ramakrishna, A. Das, H. N. Ghosh, Langmuir 2004, 20,
1430.
[27] A. Furube, T. Asahi, H. Masuhara, H. Yamashita, M. Anpo, J. Phys.
Chem. B 1999, 103, 3120.
[28] N. Serpone, D. Lawless, R. Khairutdinorv, E. Pellizzetti, J. Phys.
Chem. 1995, 99, 16655.
[29] S. H. Lee, Y. Park, K. R. Wee, H. J. Son, D. W. Cho, C. Pac, W.
Choi, S. O. Kang, Org. Lett. 2010, 12, 460–463.
[30] a) G. Redmond, D. Fitzmaurice, J. Phys. Chem. 1993, 97, 1426;
b) M. D. Ward, J. M. White, A. J. Bard, J. Am. Chem. Soc. 1983, 105,
27.
[31] J. M. Herrmann, J. Disdier, P. Pichat, J. Phys. Chem. 1986, 90, 6028.
[32] a) G. Ramakrishna, H. N. Ghosh, J. Phys. Chem. B 2001, 105, 7000;
b) T. Tachikawa, S. Tojo, M. Fujitsuka, T. Majima, J. Phys. Chem. B
2004, 108, 5859.
[33] G. R. Bamwenda, S. Tsubota, T. Nakamura, M. Haruta, J. Photo-
chem. Photobiol. A 1995, 89, 177.
[34] We assume that one photon is consumed to produce one hydrogen
molecule for calculations of apparent quantum yield.
[35] T. Yoshihara, R. Katoh, A. Furube, Y. Tamaki, M. Murai, K. Hara,
S. Murata, H. Arakawa, M. Tachiya, J. Phys. Chem. B 2004, 108,
3817.
[36] J. Wiberg, T. Marinado, D. P. Hagberg, L. Sun, A. Hagfeldt, B. Al-
binsson, J. Phys. Chem. C 2009, 113, 3881.
[37] A. Vittadini, A. Selloni, F. P. Rotzinger, M. Grꢄtzel, J. Phys. Chem.
B 2000, 104, 1300.
[38] a) H. Tian, X. Yang, R. Chen, R. Zhang, A. Hagfeldt, L. Sun, J.
Phys. Chem. C 2008, 112, 11023; b) M. Pastore, F. De Angelis, ACS
nano 2010, 4, 556.
[39] T. Horiuchi, H. Miura, S. Uchida, Chem. Commun. 2003, 3036.
[40] M. Ziꢅłek, I. Tacchini, M. T. Martꢆnes, X. Yang, L. Sun, A. Douhal,
Phys. Chem. Chem. Phys. 2011, 13, 4032.
[12] K. Hashimoto, I. Irie, A. Fujishima, Jpn. J. Appl. Phys. 2005, 44,
8269.
[13] a) A. Mills, S. Le Hunte, J. Photochem. Photobiol. A 1997, 108, 1;
b) M. Grꢄtzel, Chem. Lett. 2005, 34, 8.
[14] a) M. R. Hoffmann, S. T. Martin, W. Choi, D. W. Bahnemann,
Chem. Rev. 1995, 95, 69; b) A. L. Linsebigler, G. Lu, J. T. Yates,
Chem. Rev. 1995, 95, 735; c) T. L. Thompson, J. T. Yates, Chem. Rev.
2006, 106, 4428; d) X. Chen, S. S. Mao, Chem. Rev. 2007, 107, 2891.
[15] D. C. Schmelling, K. A. Gray, P. V. Kamat, Environ. Sci. Technol.
1996, 30, 2547.
[16] a) J.-M. Lehn, J.-P. Sauvage, Nouv. J. Chim. 1977, 1, 449; b) A. Mor-
adpour, E. Amouyal, P. Keller, H. Kagan, Nouv. J. Chim. 1978, 2,
547; c) M. Kirch, J.-M. Lehn, J.-P. Sauvage, Helv. Chim. Acta 1979,
62, 1345; d) P. Keller, A. Moradpour, E. Amouyal, H. B. Kagan,
Nouv. J. Chim. 1980, 4, 377; e) D. Gust, T. A. Moore, A. L. Moore,
Acc. Chem. Res. 2009, 42, 1890; f) J. J. Concepcion, J. W. Jurss, M. K.
[41] M. D. Fayer, Acc. Chem. Res. 2012, 45, 3.
[42] a) R. Jimenez, G. R. Fleming, P. V. Kumar, M. Maroncelli, Nature
1994, 369, 471; b) Y. Rao, N. J. Turro, K. B. Eisenthal, J. Phys. Chem.
C 2010, 114, 17703; c) E. Portuondo-Campa, A. Tortschanoff, F. van
15380
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 15368 – 15381

Hydrogen Evolution in Dye/TiO2/Pt Hybrid Systems
FULL PAPER
Mourik, J.-E. Moser, A. Kornherr, M. Chergui, J. Phys. Chem. B
2006, 110, 7835.
[43] S. C. Ke, T. C. Wang, M. S. Wong, N. O. Gopal, J. Phys. Chem. B
2006, 110, 11628.
[44] a) M. Eguchi, S. Takagi, H. Inoue, Chem. Lett. 2006, 35, 14; b) M.
Eguchi, H. Tachibana, S. Takagi, T. A. Tryk, H. Inoue, Bull. Chem.
Soc. Jpn. 2007, 80, 1350; c) T. Egawa, H. Watanabe, T. Fujimura, Y.
Ishida, M. Yamato, O. Masui, T. Shimada, H. Tachibana, H. Yoshi-
da, H. Inoue, S. Takagi, Langmuir 2011, 27, 10722; d) Y. Ishida, T.
Shimada, D. Masui, H. Tachibana, H. Inoue, S. Takagi, J. Am.
Chem. Soc. 2011, 133, 14280.
[47] A. M. Braun, M.-T. Maurette, E. Oliveros, Photochemical Technolo-
gy, Wiley & Sons, New York, 1991; pp. 76–80 and references there-
in.
[48] a) R. Dronskowski, P. E. Blçchl, J. Phys. Chem. 1993, 97, 8617; b) G.
Kresse, J. Hafner, J. Phys. Condens. Matter 1994, 6, 8245; c) G.
Kresse, J. Furthmꢀller, Phys. Rev. B 1996, 54, 11169.
[49] a) K. Maeda, M. Eguchi, S-H. A. Lee, W. J. Youngblood, H. Hata,
T. E. Mallouk, J. Phys. Chem. C 2009, 113, 7962; b) S. R. Morrison,
T. Freund, J. Chem. Phys. 1967, 47, 1543; c) K. Micka, H. Gerischer,
J. Electroanal. Chem. 1972, 38, 397; d) S. N. Frank, A. J. Bard, J. Am.
Chem. Soc. 1977, 99, 4667; e) M. Matsumura, H. Ohnishi, K. Hana-
fusa, H. Tsubomura, Bull. Chem. Soc. Jpn. 1987, 60, 2001.
[45] B. Ohtani, R. M. Bowman, D. P. Colombo Jr., H. Kominami, H. No-
guchi, K. Uosaki, Chem. Lett. 1998, 579.
[46] W. S. Han, J. K. Han, H. Y. Kim, M. J. Choi, Y. S. Kang, C. Pac, S. O.
Kang, Inorg. Chem. 2011, 50, 3271.
Received: April 30, 2012
Revised: August 13, 2012
Published online: October 5, 2012
Chem. Eur. J. 2012, 18, 15368 – 15381
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
15381