Significance of hydrophilic characters of organic dyes in visible-light hydrogen generation based on TiO2

Article abstract of DOI:10.1021/ol9026182

[Chemical equaction presented] A series of dyes were synthesized to examine the roles of the hydrophilic characteristics of R in sensitized hydrogen generation by dyegrafted Pt/TiO₂ under visible light irradiation. The hydrogen-generation efficiencies and optimum amounts of the dyes grafted to Pt/TiO₂ were affected substantially by the hydrophilic and steric effects of R; moderately hydrophilic DE01 and DE02 showed higher sensitization activity at a lower loading than hydrophobic D-H.

Full text of DOI:10.1021/ol9026182

ORGANIC
LETTERS
2010
Vol. 12, No. 3
460-463
Significance of Hydrophilic Characters
of Organic Dyes in Visible-Light
Hydrogen Generation Based on TiO2
Su-Hyun Lee,† Yiseul Park,‡ Kyung-Ryang Wee,† Ho-Jin Son,† Dae Won Cho,†
Chyongjin Pac,†,* Wonyong Choi,‡ and Sang Ook Kang*,†
Department of Materials Chemistry, Sejong Campus, Korea UniVersity, Chung-Nam
339-700, South Korea, and School of EnVironmental Science and Engineering, Pohang
UniVersity of Science and Technology, Pohang, 790-784, South Korea
jjpac@korea.ac.kr; sangok@korea.ac.kr
Received November 12, 2009
ABSTRACT
A series of dyes were synthesized to examine the roles of the hydrophilic characteristics of R in sensitized hydrogen generation by dye-
grafted Pt/TiO2 under visible light irradiation. The hydrogen-generation efficiencies and optimum amounts of the dyes grafted to Pt/TiO2 were
affected substantially by the hydrophilic and steric effects of R; moderately hydrophilic DEO1 and DEO2 showed higher sensitization activity
at a lower loading than hydrophobic D-H.
7
Considerable attention has been paid to semiconductor TiO2
in connection to the creation of clean energy,1 typically
hydrogen evolution from water2 and electricity generation
by solar cells under solar-light illumination.3 Due to little
visible-light absorption by TiO2, dye modification has been
used to make TiO2 active to visible-light excitation, typically
in dye-sensitized solar cells (DSSCs)4 and in dye-sensitized
H2 generation using platinized TiO2 particles (Pt/TiO2).5 Such
modifications are usually performed by chemical bonding
of a functional group (e.g., CO2H6 or PO3H2 ) of dyes to
the TiO2 surface. While ruthenium(II) polypyridine com-
plexes have been used for such modifications, organic dyes
are also considered to be potentially attractive due to the
availability of versatile functional molecules associated with
tuning of the electronic and chemical properties. While
DSSCs using organic dyes have been developed,8 little
(4) (a) O’Regan, B.; Gra¨tzel, M. Nature 1991, 353, 737. (b) Huang,
S. Y.; Schlichtho¨rl, G.; Nozik, A. J.; Gra¨tzel, M.; Frank, A. J. J. Phys.
Chem. B 1997, 101, 2576. (c) Gra¨tzel, M. Nature 2001, 414, 338. (d) Gra¨tzel,
M. J. Photochem. Photobiol. A 2004, 164, 3. (e) Gra¨tzel, M. Chem. Lett.
2005, 34, 8.
† Korea University.
‡ Pohang University of Science and Technology.
(1) Fujishima, A.; Honda, K. Nature 1972, 37, 238.
(2) (a) Sato, S.; White, J. M. Chem. Phys. Lett. 1980, 72, 83. (b)
Duonghong, D.; Borgarello, E.; Gra¨tzel, M. J. Am. Chem. Soc. 1981, 103,
4685. (c) Kajiwara, T.; Hasimoto, K.; Kawai, T.; Sakata, T. J. Phys. Chem.
1982, 86, 4516. (d) Tabata, S.; Nishida, H.; Masaki, Y.; Tabata, K. Catal.
Lett. 1995, 34, 245. (e) Sayama, K.; Arakawa, H. J. Chem. Soc., Faraday
Trans. 1997, 93, 1647.
(5) (a) Kiwi, J.; Gra¨tzel, M. J. Am. Chem. Soc. 1979, 101, 7214. (b)
Malinka, E. A.; Kamalov, G. L.; Vodzinskii, S. V.; Melnik, V. I.; Zhilina,
Z. I. J. Photochem. Photobiol. A 1995, 90, 153. (c) Bae, E.; Choi, W.;
Park, J.; Shin, H. S.; Kim, S. B.; Lee, J. S. J. Phys. Chem. B 2004, 108,
14093. (d) Kudo, A.; Miseki, Y. Chem. Soc. ReV. 2009, 38, 253.
(6) Dunghong, D.; Serpone, N.; Gra¨tzel, M. HelV. Chim. Acta 1984,
67, 1012.
(3) (a) Vlachopoulos, N.; Liska, P.; Augustynski, J.; Gra¨tzel, M. J. Am.
Chem. Soc. 1988, 110, 1216. (b) Ardo, S.; Meyer, G. J. Chem. Soc. ReV.
2009, 38, 115.
(7) Pe´chy, P.; Rotzinger, F. P.; Nazeeruddin, M. K.; Kohle, O.;
Zakeeruddin, S. M.; Humphry-Baker, R.; Gra¨tzel, M. J. Chem. Soc., Chem.
Commun. 1995, 65.
10.1021/ol9026182  2010 American Chemical Society
Published on Web 12/30/2009

systematic investigation has been carried out on the applica-
tion of organic dyes to sensitized H2 generation from water.
In 1985, Shimidzu et al. reported that H2 generation can be
sensitized by xanthene dyes under visible-light irradiation
using Pt-loaded metal-oxide particles and triethanolamine as
a sacrificial electron donor.9 In recent years, considerable
improvements in H2-generation efficiency have been achieved
with similar reaction systems, in which Pt/TiO2 or related
solid catalysts are modified with organic dyes, particularly
Eosin Y.10 In general, neutral chromophoric cores of
conventional organic dyes are intrinsically hydrophobic.
Consequently, the poor water miscibility should give rise to
complex features in the dye/TiO2 interface with the water
phase as well as in the conformation/aggregation behavior
of the dyes adsorbed on the hydrophilic TiO2 surface, which
have critical effects on the electron-transfer and chemical
processes involved in H2 generation.
with the efficiencies depending on the substituent R. Mod-
erately hydrophilic DEO1 and DEO2 showed substantially
higher sensitization activities at a lower loading than
hydrophobic D-H, whereas intermediate behavior was ob-
served in the cases of slightly hydrophilic DMOM and most
hydrophilic DEO3.
Table 1. Photophysical Properties of the Dyes in THF
a
c
d
e
dye
λabs (nm) εb (M-1 cm-1
)
λF (nm) τF (ns) ΦF
D-H
433
432
433
433
432
27,360
26,730
31,580
27,550
35,460
552
561
559
557
557
1.18
1.25
1.26
1.31
1.19
0.23
0.24
0.22
0.24
0.21
DMOM
DEO1
DEO2
DEO3
a Absorption maxima. b Molar absorption coefficients. c Fluorescence
maxima. d Fluorescence lifetimes. e Fluorescence quantum yield was
measured using rhodamine B as a reference at rt.
Scheme 1. Synthetic Routes to the Dyes 6R
Figure 1. CV trace in THF using [n-Bu4N]PF6 (0.1 M) as a
supporting electrolyte and Fc/Fc+ reference and DFT-generated
population densities of the HOMO and LUMO for DEO2.
Scheme 1 summarizes the synthetic routes for the dyes;
the experimental details for the synthetic procedures and
spectroscopic data for identification are given in the Sup-
porting Information (SI). The dyes in tetrahydrofuran (THF)
commonly reveal the visible absorption maxima at 432-433
nm and relatively strong fluorescence at 552-561 nm with
lifetimes of 1.18-1.31 ns (Table 1). Cyclic voltammetry
(CV) measurements for THF solutions of the dyes show clear
one-electron oxidation/reduction peaks, as shown in Figure
1 for a typical CV trace of DEO2. The apparent half-wave
oxidation and reduction potentials of the dyes are commonly
+(0.46-0.49) V and -(2.06-2.08) V, respectively, vs the
Fc/Fc+ reference (see Figure S4 in SI). From the CV data
coupled with the longest absorption edge, the HOMO and
LUMO levels of the dyes were estimated to be (5.26-5.29)
eV and (2.78-2.82) eV below the vacuum level, respectively.
DFT MO calculations of the dyes indicated that the HOMO
and LUMO largely populate on the diphenylaminophenyl-
2,2′-bithiophene donor and cyanoacrylic acid acceptor,
respectively, as shown in Figure 1 for DEO2 as a typical
example. Therefore, the electronic properties of the dyes are
essentially identical independent of the substituent R.
This study aims to explore the significance of hydrophilic
and steric characteristics of organic dyes in sensitized H2
generation based on Pt/TiO2 catalysts using a series of (E)-
3-[5-(4-(p,p′-bis(R-phenyl)amino)phenyl)-2,2′-bithiophen-2′-
yl]-2-cyanoacrylic acid dyes, which are the hydrophobic
parent dye (R ) H, D-H), slightly hydrophilic DMOM (R
)
CH2OCH3,), and hydrophilic DEO1-3 (R
)
CH2-(OCH2CH2)n-OCH3, n ) 1-3). It was found that Pt/
TiO2 particles modified with the dyes work as effective
catalysts for visible-light-induced H2 evolution from water
(8) (a) Kim, S.; Lee, J. K.; Kang, S. O.; Ko, J.; Yum, J.-H.; Fantacci,
S.; De Angelis, F.; Di Censo, D.; Nazeeruddin, M. K.; Gra¨tzel, M. J. Am.
Chem. Soc. 2006, 128, 16701. (b) Mishra, A.; Fischer, M. K. R.; Ba¨uerle,
P. Angew. Chem., Int. Ed. 2009, 48, 2474. (c) Moon, S.-J.; Yum, J.-H.;
Humphry-Baker, R.; Karlsson, K. M.; Hagberg, D. P.; Marinado, T.;
Hagfeldt, A.; Sun, L.; Gra¨tzel, M.; Nazeeruddin, M. K. J. Phys. Chem. C
2009, 113, 16816.
(9) Shimidzu, T.; Iyoda, T.; Koide, Y. J. Am. Chem. Soc. 1985, 107,
35.
(10) (a) Abe, R.; Hara, K.; Sayama, K.; Domen, K.; Arakawa, H. J.
Photochem. Photobiol. A 2000, 137, 63. (b) Abe, R.; Sayama, K.; Arakawa,
H. J. Photochem. Photobiol. A 2004, 166, 115. (c) Li, Q.; Jin, Z.; Peng, Z.;
Li, Y.; Li, S.; Lu, G. J. Phys. Chem. C 2007, 111, 8237. (d) Li, Y.; Xie,
C.; Peng, S.; Lu, G.; Li, S. J. Mol. Catal. A 2008, 282, 117. (e) Reisner,
E.; Fontecilla-Camps, J. C.; Armstrong, F. A. Chem. Commun. 2009, 550.
Org. Lett., Vol. 12, No. 3, 2010
461

After Pt had been loaded on TiO2 particles (Hombikat UV-
100, primary particle size of 16 nm, BET surface area of
348 m2/g) according to a published method,11 the Pt/TiO2/
Dye catalysts were obtained by adsorption of the dyes
(0.5-10 µmol) on Pt/TiO2 (0.1 g) in 1:1 MeCN:t-BuOH (10
mL) at rt. The dye adsorption on the Pt/TiO2 particles was
complete because the filtrates obtained after the adsorption
treatment were transparent. Figure 2a shows diffuse reflec-
tance spectra (DRS) of the Pt/TiO2/DEO2 catalyst, which
are significantly broader than the solution spectrum but
showed little dependence of the spectral shape on the amount
of dye-adsorption. This spectral broadening might be at-
tributable to charge transfer between the adsorbed dyes and
TiO2,12 and/or to multiple reflections. In the case of Pt/TiO2/
D-H, however, a further red shift in the DRS was observed
(Figure 2b), suggesting that the hydrophobic dye molecules
should more or less undergo molecular stacking on the
hydrophilic TiO2 surface, whereas the ethylene-oxide chains
of DEO2 might sterically hinder the dye-core stacking.
Figure 3. (a) Temporal plots for H2 generation by irradiation of
water suspensions of the catalysts (0.3 µmol dye/30 mg Pt/TiO2)
at >420 nm and (b) the amounts of H2 generated after 5 h irradiation;
pHi ) 3 and [EDTA]i ) 10 mM.
the conduction band of TiO2 (process 1) in competition with
the excited-state decay to the ground state (process 2),
recombination between the injected electron and the dye
radical cation (process 3), and the reduction of the dye radical
cation by EDTA (process 4) that facilitates the collection of
electrons injected at Pt sites to result in net H2 generation
(process 5).
Pt/TiO2/Dye + hν f Pt/TiO2/1Dye*
Pt/TiO2/1Dye* f Pt/TiO2(e-)/Dye·+
Pt/TiO2/1Dye* f Pt/TiO2/Dye + hνF and heat (2)
Pt/TiO2(e-)/Dye·+ f Pt/TiO2/Dye
(3)
Pt/TiO2(e-)/Dye·+(+EDTA) f Pt/TiO2(e-)/Dye (4)
Pt/TiO2(e-)/Dye + H+ f Pt/TiO2/Dye + 1/2H2
(5)
(1)
Figure 2. (a) DRS of Pt/TiO2/DEO2 and absorption spectrum of a
THF solution. (b) DRS of Pt/TiO2/D-H compared to that of Pt/
TiO2/DEO2 (1.5 µmol dye/30 mg Pt/TiO2).
Water suspensions containing the catalysts (30 mg/30 mL)
and EDTA (10 mM) were adjusted at pH ) 3, bubbled with
an N2 stream, and then irradiated at λ > 420 nm. The
generated amounts of H2 were determined by gas chroma-
tography with e3 µmol error. Figure 3 shows temporal plots
for the H2 generation using the catalysts of 0.3 µmol dye/30
mg Pt/TiO2. This clearly shows that the DEO1 and DEO2
catalysts bearing the hydrophilic ethylene-oxide substituents
are ∼3 times more active than the hydrophobic D-H.
Moreover, even the slightly hydrophilic methoxymethyl
groups in DMOM are significantly effective in enhancing
the sensitization activity. Interestingly, the catalyst activity
uniquely depends on the hydrophilic chain length such that
DEO1 and DEO2 showed significantly higher catalyst
activities than DEO3 with the longest hydrophilic chain.
These observations clearly suggest that the hydrophilic
chains of DEO1 and DEO2 are particularly effective in
enhancing H2 generation. In general, the H2-generation
efficiencies should be controlled by the following major
processes; electron injection from the excited-state dyes into
The dyes adsorbed on Pt/TiO2 exhibited virtually no
fluorescence. The fluorescence decay occurred in a time
domain of the excitation laser pulse (∼30 ps). The fluores-
cence quenching can be mainly attributed to the dominant
occurrence of process 1 over process 2. Process 5 depends
primarily on the electron mobility in TiO2, electron-collection
efficiency at the Pt sites, and H2 formation rate on the Pt
surface. These factors should be not affected significantly
by the dye substituent unless the Pt sites are covered with
dye molecules. Consequently, the observed effects of sub-
stituent R on the H2-generation effciency should mainly arise
from the competition between processes 3 and 4. However,
process 3 should be essentially independent of the substituent
R because the driving force for electron transfer between
the dye-core radical cation and injected electron in TiO2 are
certainly identical for all the dyes.
Process 4 requires the diffusional approach of hydrophilic
EDTA molecules to the dye radical cation at an effective
distance. In this regard, it should be noted that the positive
charge of the dye radical cation largely develops on the
(11) (a) Herrmann, J. M.; Disdier, J.; Pichat, P. J. Phys. Chem. 1986,
90, 6028. (b) Lee, J.; Choi, W. EnViron. Sci. Technol. 2004, 38, 4026.
(12) Tachikawa, T.; Tojo, S.; Fujitsuka, M.; Majima, T. J. Phys. Chem.
B 2004, 108, 5859.
462
Org. Lett., Vol. 12, No. 3, 2010

diphenylaminophenyl moiety to which the substituents (R)
attach. In the cases of DEO1 and DEO2, the hydrophilic
ethylene-oxide chains should make the dye surface miscible
with the aqueous phase so that process 4 might be facilitated
by approaching of EDTA molecules close to the radical-
cation center. Such enhancement effects should be weaker,
but not negligible, with the slightly hydrophilic methoxym-
ethyl groups of DMOM. Another benefit of the hydrophilic
chains might be due to possible hydrogen bonding of EDTA
molecules with the ethylene-oxide chains. This is certainly
favorable for process 4 to proceed. In contrast, the interface
between hydrophobic D-H and the water phase should be
substantially different, thus forming an unfavorable environ-
ment for electron transfer from hydrophilic EDTA to the
hydrophobic dye radical cation. Moreover, the molecular
stacking of D-H might open a rapid decay channel from the
excited state, but this is less important for the hydrophilic
dyes. However, in the case of DEO3, irrespective of the
hydrophilic character, the long triglyme chains might steri-
cally inhibit the approach of EDTA to the radical-cation core
and/or cover the Pt sites.
and limit the effective utilization of incident light by the
whole reaction system. Moreover, the dye coverage over the
Pt/TiO2 surface should become high enough to have unfavor-
able effects on process 5 proceeding over the Pt site. At high
levels of dye adsorption, stabilization of the dye radical cation
can occur by charge delocalization over the closely packed
dye molecules to retard process 4. Such counterbalancing
effects would control the optimum dye loading for the
maximum H2-generation efficiency, for example, 1.5 µmol/
30 mg Pt/TiO2 for D-H and 0.9 µmol/30 mg Pt/TiO2 for
DMOM. In the cases of DEO1 and DEO2, the ethylene-
oxide chains are intrinsically capable of enhancing H2
generation due to their hydrophilic nature, but can form a
dense layer at high dye loadings to impart steric effects on
processes 4 and 5. The hydrophilic effect should prevail at
lower dye loadings to allow rapid increases in H2 generation
with increasing dye loading from 0.15 µmol to 0.3 µmol,
whereas the steric effects might be increasingly dominated
at higher dye loadings. On the other hand, the longer triglyme
chains of DEO3 form a dense layer, even at lower concen-
trations, giving greater steric effects that might cancel the
hydrophilic effect at g0.15 µmol.
In summary, a series of dyes (hydrophobic D-H, slightly
hydrophilic DMOM, and hydrophilic DEO1-DEO3) were
synthesized and applied to sensitized hydrogen generation
based on Pt/TiO2/Dye catalysts. The moderately hydrophilic
DEO1 and DEO2 catalysts showed substantially higher H2-
generation efficiencies at a low dye adsorption than the
hydrophobic D-H catalyst, whereas either slightly hydrophilic
DMOM or DEO3 with the longest ethylene-oxide chains
showed intermediate sensitization ability. The optimum
grafted amounts of the dyes (µmol/30 mg Pt/TiO2) for
maximizing H2 generation were considerably different: e0.15
for DEO3, ∼0.3 for DEO1 and DEO2, 0.15-0.9 for
DMOM, and ∼1.5 for D-H. The dependence of the catalyst
activity on the substituent was interpreted in terms of the
intrinsic enhancement effects of the hydrophilic character
canceled by the steric and/or coverage effects. These results
strongly suggest that the optimization of hydrophilic character
coupled with the minimization of steric effects is a key issue
associated with the general guideline for designing organic
dyes with high sensitization capability in H2 generation based
on TiO2 and related semiconductors.
Figure 4. Dependences of the sensitized H2 generation on the dye
loading (µmol on 30 mg Pt/TiO2). The experimental conditions are
the same as those in Figure 3.
Another interesting observation is the complex dependence
of the catalyst activity on the amounts of dye fixed to Pt/
TiO2 particles, as shown in Figure 4. The catalyst activity
for D-H increases with increasing amount of dye to show a
maximum at a dye amount of 1.5 µmol/30 mg Pt/TiO2 and
a slight decrease at 3 µmol/30 mg Pt/TiO2. On the other hand,
the DMOM catalyst exhibits a weak dependence of the
activity on the dye loading at e1.5 µmol/30 mg Pt/TiO2 but
a relatively large drop at 3 µmol/30 mg Pt/TiO2. In the cases
of DEO1 and DEO2, the H2-generation efficiency reaches
a maximum at a significantly lower dye loading (0.3 µmol/
30 mg Pt/TiO2) and decreased rapidly at higher amounts. In
contrast, DEO3 showed a continuous decrease in catalyst
activity from the minimum dye amount (0.15 µmol/30 mg
Pt/TiO2).
Acknowledgment. This work was supported by the
National Research Foundation of Korea (NRF) grant funded
by the Korea government (MEST) (No. 2009-0083181). We
also acknowledge supports from Korea Basic Science
Institute on HR-ESI-MS, LTQ-FT, and 300 NMR.
Supporting Information Available: Experimental details
and characterization data (1H, 13C NMR), CVs, and DFT-
calculation data. This material is available free of charge
via the Internet at http://pubs.acs.org.
In general, the fraction of incident light absorbed by the
dyes increases with increasing dye loading to saturate at a
certain level. However, as the dye loading is increased
further, the penetration depth of incident light should decrease
OL9026182
Org. Lett., Vol. 12, No. 3, 2010
463