Photophysical and (photo)electrochemical properties of a coumarin dye

Article abstract of DOI:10.1021/jp044851v

A new coumarin dye, cyano-{5,5-dimethyl-3-[2-(1,1,6,6-tetramethyl-10-oxo-2, 3,5,6-tetrahydro-1H,4H,10H-11-oxa-3a-aza-benzo[de]anthracen-9-yl)vinyl] cyclohex-2-enylidene}-acetic acid (NKX-2753), was prepared and characterized with respect to photophysical and electrochemical properties. It was employed as a dye sensitizer in dye-sensitized solar cells and showed efficient photon-to-electron conversion properties. The photocurrent action spectrum exhibited a broad feature with a maximum incident photon-to-electron conversion efficiency (IPCE) of 84% at 540 nm, which is comparable to that for the famous red dye RuL₂(NCS)₂ (known as N3), where L stands for 2,2′-bipyridyl-4,4′-dicarboxylic acid. The sandwich-type solar cell with NKX-2753, under illumination of full sun (AM1.5, 100 mW cm^-2), produced 16.1 mA cm^-2 of short-circuit photocurrent, 0.60 V of open-circuit photovoltage, and 0.69 of fill factor, corresponding to 6.7% of overall energy conversion efficiency using 0.1 M Lil, 0.05 M I₂, 0.1 M guanidinium thiocyanate, and 0.6 M 1,2-dimethyl-3-n-propyl-imidazolium iodide in dry acetonitrile as redox electrolyte. In comparison with its analogue NKX-2586 (Langmuir 2004, 20, 4205), NKX-2753 with an extra side ring on the alkene chain produced much higher IPCE values at the same conditions. The side ring acted as a spacer to efficiently prevent dye aggregation when adsorbed on the TiO₂ surface, resulting in significant improvements of short-circuit photocurrent, open-circuit photovoltage, and fill factor compared with NKX-2586 that aggregated on the TiO₂ surface.

Full text of DOI:10.1021/jp044851v

J. Phys. Chem. B 2005, 109, 3907-3914
3907
Photophysical and (Photo)electrochemical Properties of a Coumarin Dye
Zhong-Sheng Wang,*,† Kohjiro Hara,§ Yasufumi Dan-oh,| Chiaki Kasada,| Akira Shinpo,|
Sadaharu Suga,| Hironori Arakawa,† and Hideki Sugihara*,†
Solar Light Energy ConVersion Team, Energy Technology Research Institute, National Institute of AdVanced
Industrial Science and Technology (AIST), 1-1-1, Higashi, Tsukuba, Ibaraki 305-8565, Japan, Photonics
Research Institute, National Institute of AdVanced Industrial Science and Technology (AIST), 1-1-1, Higashi,
Tsukuba, Ibaraki 305-8565, Japan, and Hayashibara Biochemical Laboratories, Incorporated, 564-176 Fujita,
Okayama 701-0221, Japan
ReceiVed: NoVember 10, 2004; In Final Form: January 3, 2005
A new coumarin dye, cyano-{5,5-dimethyl-3-[2-(1,1,6,6-tetramethyl-10-oxo-2,3,5,6-tetrahydro-1H,4H,10H-
11-oxa-3a-aza-benzo[de]anthracen-9-yl)vinyl]cyclohex-2-enylidene}-acetic acid (NKX-2753), was prepared
and characterized with respect to photophysical and electrochemical properties. It was employed as a dye
sensitizer in dye-sensitized solar cells and showed efficient photon-to-electron conversion properties. The
photocurrent action spectrum exhibited a broad feature with a maximum incident photon-to-electron conversion
efficiency (IPCE) of 84% at 540 nm, which is comparable to that for the famous red dye RuL2(NCS)2 (known
as N3), where L stands for 2,2′-bipyridyl-4,4′-dicarboxylic acid. The sandwich-type solar cell with NKX-
2753, under illumination of full sun (AM1.5, 100 mW cm-2), produced 16.1 mA cm-2 of short-circuit
photocurrent, 0.60 V of open-circuit photovoltage, and 0.69 of fill factor, corresponding to 6.7% of overall
energy conversion efficiency using 0.1 M LiI, 0.05 M I2, 0.1 M guanidinium thiocyanate, and 0.6 M 1,2-
dimethyl-3-n-propyl-imidazolium iodide in dry acetonitrile as redox electrolyte. In comparison with its analogue
NKX-2586 (Langmuir 2004, 20, 4205), NKX-2753 with an extra side ring on the alkene chain produced
much higher IPCE values at the same conditions. The side ring acted as a spacer to efficiently prevent dye
aggregation when adsorbed on the TiO2 surface, resulting in significant improvements of short-circuit
photocurrent, open-circuit photovoltage, and fill factor compared with NKX-2586 that aggregated on the
TiO2 surface.
Introduction
Among the family of dye sensitizers pure organic dyes are
also promising due to their rich photophysical properties.
Dye-sensitized solar cells (DSSCs) have attracted considerable
attention since the landmark report by Gra¨tzel’s group.1 In
DSSCs, the sensitizer, chemically adsorbed on the TiO2 surface,
absorbs visible light and after excitation injects an electron to
the conduction band of TiO2, generating electric current. The
photovoltage generated by the cell is theoretically the difference
between the Fermi level in TiO2 under illumination and the
Nernst potential of the redox couple in the electrolyte.2 Among
the dyestuff tested so far, ruthenium polypyridyl complexes have
shown outstanding charge-transfer properties. Overall energy
conversion efficiency greater than 10% was achieved with two
excellent dyes cis-di(thiocyanato)bis(2,2′-bipyridyl-4,4′-dicar-
boxylate)ruthenium(II) (known as N3)3 and tri(thiocyanato)-
2,2′,2′′-terpyridyl-4,4′,4′′-tricarboxylate)ruthenium(II)4 (known
as the black dye). In particular, N3 has been used as the
paradigm of a heterogeneous charge-transfer sensitizer in
DSSCs. In addition to ruthenium complexes,3-10 other metal
complexes comprising Os11 and Fe12 also showed promising
results.
Organic dyes have two major advantages in DSSCs. One is the
high molar extinction coefficient for the organic dye due to its
much higher oscillator strengths than those of metal complexes.
The use of organic dyes would allow a thinner TiO2 film to
harvest the incident light efficiently. The other advantage is that
no noble metal like Ru is concerned in organic dyes. This
reduces the overall cost of the cell production. For these reasons
organic dyes may find applications in DSSCs. Recently, several
groups reported solar cell performance based on organic dyes,
such as cyanines,13 merocyanines,14 hemicyanines,15 perylenes,16
xanthenes,17 porphyrins,18 phthalocyanines,19 polyenes,20 poly-
thiophenes,21 coumarins,22 etc. Among the organic dyestuff
tested so far, coumarin dye may be the most promising
sensitizer. Coumarin 343, a laser dye, displayed very high IPCEs
in the spectral sensitization of TiO2 film but produced low
photocurrents due to the lack of red and near-IR light absorp-
tion.23 Our group has systematically studied the coumarin dyes
and successfully shifted the maximum absorption to the red
region by way of expanding π-conjugation through molecular
design.22 As a result of improved photoresponse in the red
region, short-circuit photocurrent and overall energy conversion
efficiency have been improved significantly. After we obtained
an efficiency of 7.7% with a coumarin derivative NKX-2677,24
Horiuchi et al. achieved an efficiency of 8.00% using an indoline
* Corresponding authors. E-mail: zs.wang@aist.go.jp, sugihara-hideki@
aist.go.jp, Tel: +81-29-861-4643. Fax: +81-29-861-6771.
† Energy Technology Research Institute, National Institute of Advanced
Industrial Science and Technology (AIST).
§ Photonics Research Institute, National Institute of Advanced Industrial
Science and Technology (AIST).
| Hayashibara Biochemical Laboratories, Inc.
10.1021/jp044851v CCC: $30.25 © 2005 American Chemical Society
Published on Web 02/12/2005

3908 J. Phys. Chem. B, Vol. 109, No. 9, 2005
Wang et al.
SCHEME 1: Schematic Structure of NKX-2586 and
NKX-2753
washed with basic solution, ethanol, and acetone successively
under supersonication for 10 min each. NKX-2753 was provided
by Hayashibara Biochemical Laboratories, and the purity was
confirmed by NMR measurement (see Supporting Information).
The synthesis of NKX-2586 was reported previously.22
Fabrication of the Dye-Sensitized TiO2 Films. A screen-
printing method was used to fabricate TiO2 films on TCO glass.
TiO2 nanoparticles (23 nm) and large particles (100 nm) were
prepared by the method reported previously.27 TiO2 nanopar-
ticles (23 nm) and a mixture of nanoparticles (23 nm) and
scattering particles (100 nm) at a ratio of 6:4 were dispersed in
ethanol. The TiO2 pastes were prepared by mixing TiO2
suspension, ethyl cellulose, and R-terpinol in ethanol followed
by removal of the solvent ethanol by rotary evaporator at 40
°C under a vacuum of 20 hPa. Pastes N and M consisted of
nanoparticles and mixed particles (23 and 100 nm particles at
a ratio of 6:4, respectively). Transparent TiO2 films, fabricated
on TCO glass using paste N, were used for characterizations
such as spectroscopic and electrochemical measurements.
Double-layer TiO2 films (12 µm), employed for photovoltaic
measurements, were prepared by printing the transparent layer
(6 µm) first with paste N followed by further coating with paste
M. The TiO2 films were fired at 525 °C for 2 h with a rising
rate of 10 °C/min. All the films were immersed in 0.05 M TiCl4
solution for 30 min at 70 °C followed by calcination at 450 °C
for 30 min.4 The film thickness was measured with a Tencor
Alpha-Step 500 Surface Profiler. The film size is apparently
0.5 × 0.5 cm2, whose precise area was determined by a digital
CCD camera controlled by a computer. The films were
immersed in a dye solution for at least 6 h when they were at
∼100 °C cooled from the heating. The dye solution (0.3 mM)
was prepared in a mixed solvent of AN and 4-tert-butyl alcohol
with a volume ratio of 1/1.
dye.25 These findings indicate that organic dyes are hopefully
becoming alternative sensitizers to ruthenium complexes in
DSSCs.
One strategy to extend the photoresponse of coumarin to the
red region is to introduce methine (-CHdCH-) units to the
coumarin framework with a carboxylic acid (-COOH) group
as the anchoring group to TiO2 and a cyano (-CN) group as
the electron-withdrawing group.22 Because of the expanded
π-conjugation, the lowest energy absorption was shifted to the
longer wavelength and the dyes absorb widely in the visible
region.22 Using this strategy we synthesized a series of coumarin
derivatives with various numbers of methine units and com-
paratively studied their photovoltaic properties.22 It has been
found that the maximum absorption was gradually red shifted
by increasing the number of methine units from 1 to 3. Although
NKX-2586 (Scheme 1), with three methine units, showed the
broadest feature of photocurrent action spectrum among the
series of dyes tested, IPCE values around the maximum
absorption are much lower than those of the other analogues
mainly due to dye aggregation.22 It is well known that dye
aggregation is disadvantageous to photocurrent generation
because π-stacked aggregation on TiO2 may decrease electron
injection due to intermolecular energy transfer.22,26 The strong
interaction between NXK-2586 dye molecules on the TiO2
surface resulted in aggregation and diminished both photocurrent
and photovoltage.22
Although coadsorption of NKX-2586 with deoxycholic acid
was able to prevent aggregation and hence improve the
photovoltaic performance by means of improving both photo-
current and photovoltage,22 the significantly decreased dye
adsorption unavoidably limits the photovoltaic performance. To
reduce the intermolecular interaction or dye aggregation and
meanwhile keep the adsorbed dye amount, a new dye, NKX-
2753 (Scheme 1), was designed and synthesized. When a side
ring was introduced to the alkene chain, dye aggregation was
prevented efficiently according to the UV-vis absorption and
the photocurrent action spectra. In this study we report the
photophysical and electrochemical properties of NKX-2753 and
the photovoltaic performance of dye-sensitized solar cells. It
was found that NKX-2753 is superior to NKX-2586 in terms
of cell efficiency because of the prevented dye aggregation.
Photovoltaic Measurement. Open (unsealed) and sealed cells
were employed for photovoltaic measurements and stability
tests, respectively. An open cell was assembled by clipping dye-
sensitized TiO2 film (working electrode) and Pt-coated TCO
glass (counter electrode) together; the two electrodes were
separated by a polyethylene spacer. The electrolytes were
introduced into the interelectrodes by capillary force. For the
sealed cells the two electrodes were separated by a Surlyn spacer
(30 µm thick) and sealed up by heating the polymer frame. The
redox electrolyte was introduced into the space of interelectrodes
through the two holes predrilled on the back of the counter
electrode. The two holes were sealed up using a Surlyn film,
on which a glass slide was pressed under heat.27 The redox
electrolyte used was 0.1 M LiI, 0.05 M I2, 0.6 M DMPImI, and
TBP or GT with various concentrations in dry AN. The current-
voltage characteristics of the cells were measured on a PC-
controlled source meter (Advantest, R6243) under illumination
of simulated AM1.5 solar light from an AM1.5 solar simulator
(Wacom Co., Japan, WXS-80C-3 with a 300-W Xe lamp and
an AM1.5 filter). The incident light intensity was calibrated with
a standard Si solar cell produced by Japan Quality Assurance
Organization (JQA). Photocurrent action spectra were recorded
on a CEP-99W system (Bunkoh-Keiki Co., Ltd.).
Experimental Section
Materials and Reagents. 4-tert-Butylpyridine (TBP) and
tetrabutylammonium perchlorate (TBAP) were bought from
Aldrich and Fluka, respectively. The other reagents used in this
study, such as guanidinium thiocynate (GT), deoxycholic acid
(DCA), acetonitrile (AN), etc., were purchased from Wako Pure
Chemical Industries Ltd. 1,2-Dimethyl-3-n-propylimidazolium
iodide (DMPImI) was purchased from Tomiyama Pure Chemical
Industries Ltd. Transparent conducting oxide (TCO, F-doped
SnO2, 10 Ω/0) was provided by Nippon Sheet Glass Co. and
Characterization. The UV-vis absorption spectra of the dye-
loaded transparent film and the dye solutions were recorded on
a Shimadzu UV-3101PC spectrophotometer. The attenuated total
reflection Fourier transform infrared (ATR-FTIR) spectra were
measured with a Perkin-Elmer Spectrum One spectrometer.
Cyclic voltammograms were measured with CH 610B Electro-
chemical Analyzer. Dye-loaded film, platinum, and Ag/Ag+
(0.01 M AgNO3 + 0.1 M TBAP in AN) were employed as

Photoelectrochemical Properties of a Coumarin Dye
J. Phys. Chem. B, Vol. 109, No. 9, 2005 3909
Figure 1. (a) UV-vis absorption spectra for 2.6 × 10-5 M NKX-2753 in ethanol (dotted line) and on 3.5 µm TiO2 transparent film (solid line).
Ethanol and TiO2-coated conducting glass were used as references, respectively. (b) Normalized spectra.
working, counter, and reference electrodes, respectively. The
supporting electrolyte was 0.1 M TBAP in AN, which was
degassed with N2 for 20 min prior to scan. The potential of the
reference electrode is 0.49 V vs NHE and calibrated with
ferrocene.
Results
UV-vis Absorption Spectra. Figure 1a depicts the absorp-
tion spectra for NKX-2753 both in ethanol solution and on the
TiO2 surface. The normalized spectra are shown in Figure 1b.
The strong peak at 492 nm is attributed to the π-π* electron
transition. The molar extinction coefficient in ethanol was
determined to be 5.03 × 104 M-1 cm-1 at 492 nm; such a high
value is beneficial to light harvesting. After the dye was
adsorbed from solution to the surface of TiO2, the absorption
spectrum did not show a big difference and the maximum
absorption was slightly blue shifted from 492 to 487 nm, as
seen in Figure 1b. The 5 nm blue shift was ascribed to the
interaction between the dye molecules and TiO2. The absorption
threshold at 650 nm for the dye in ethanol is red shifted to 700
nm upon adsorption on the TiO2 surface. The interaction
between the carboxylate group and the surface Ti4+ ions may
lead to increased delocalization of the π* orbital of the
conjugated framework. The energy of the π* level is decreased
by this delocalization, which explains the red shift for the
Figure 2. ATR-FTIR spectra for NKX-2753 powder (top) and NKX-
2753 adsorbed on a TiO2 film (bottom).
absorption threshold.3
The absorption for NKX-2753 (maximum absorption at 492
SCHEME 2: Possible Binding Modes for Carboxylate
Group on TiO2
nm) is similar to that for NKX-2586 (maximum absorption at
513 nm) in ethanol22c due to their similar structure (Scheme 1).
However, NKX-2586 exhibited a much bigger blue shift than
NKX-2753 after adsorption on TiO2, which will be detailed in
the Discussion section.
FTIR Spectra. The binding modes of the dye attached to
the TiO2 surface, which are related to the interfacial electron
injection process, were characterized by ATR FTIR.28 Figure 2
shows FTIR spectra for the free NKX-2753 powder and the
dye-loaded TiO2 film. For the free dye characteristic bands for
cyano (CtN), carbonyl (CdO), and carboxylic acid (COOH)
were clearly observed at 2204, 1712, and 1666 cm-1, respec-
tively.29 After the dye was adsorbed on the TiO2 surface, while
the IR peaks for cyano and carbonyl groups were also detected
at the same frequencies, the 1666 cm-1 peak for carboxylic acid
ion in several modes (Scheme 2) such as unidentate (i.e., ester),
bidentate chelating, bidentate bridging, etc.30 The frequency
separation, ∆ν ) νas(COO-) - νs(COO-), between asymmetric
and symmetric stretching modes of the carboxylate group can
be used to determine the binding mode of the dye on the TiO2
surface according to an empirical rule proposed by Deacon and
Philips.31 The splitting values are in the order of unidentate >
ionic form = bidentate bridging > bidendate chelating.31 The
frequency separations were 242 and 169 cm-1 for the ionic form
of the free dye and the dye-loaded film, respectively, suggesting
disappeared and instead two new peaks at 1540 and 1371 cm-1
,
assignable to the asymmetric and symmetric mode of carboxy-
late,29 appeared in the spectrum. This observation indicates
deprotonation of the COOH group taking place on the TiO2
surface. The formed carboxylate ion may coordinate to a metal

3910 J. Phys. Chem. B, Vol. 109, No. 9, 2005
Wang et al.
Figure 4. IPCE action spectrum for NKX-2753 dye-sensitized solar
cell. The electrolyte used was 0.1 M LiI, 0.05 M I2, 0.1 M GT, and 0.6
M DMPImI in dry AN.
Figure 3. Cyclic voltammograms of NKX-2753 adsorbed on 6 µm
TiO2 nanocrystalline film. Scan rates (from inter to outer) are 50, 100,
200, 300, 400, and 500 mV s-1, respectively. The inset shows the
relationship between the anodic peak currents and scan rates.
a bidentate chelating mode for the anchoring of the dye on the
TiO2 surface. A similar coordination mode for NKX-258622c
and a merocyanine dye14b was also observed. However, a
bidentate bridging mode was suggested for ruthenium complexes
adsorbed on TiO2.32
Cyclic Voltammetry (CV). Figure 3 shows typical cyclic
voltammograms of NKX-2753 dye-sensitized 6 µm thick TiO2
nanocrystalline film on TCO glass in an AN solution containing
0.1 M TBAP as supporting electrolyte. When the potential was
scanned between -0.2 to 0.8 V vs Ag+/Ag, quasi-reversible
waves were observed. The formal redox potential, (Eox + Ered)/
2, was calculated to be 0.41 V, corresponding to 0.9 V vs NHE.
The anodic and cathodic peak potentials shifted symmetrically,
resulting in a constant formal potential at various scan rates.
The linear relationship between the anodic peak current and
the scan rate, as shown in the inset of Figure 3, indicates a
surface-confined electrochemical process.
Figure 5. Photocurrent-photovoltage characteristics of a NKX-2753
dye-sensitized solar cell under illumination of simulated solar light (AM
1.5, 100 mW cm-2). Electrolyte used was the same as that shown in
Figure 4.
The formal redox potential is regarded as the highest occupied
molecular orbital (HOMO) energy level, and the absorption
threshold (700 nm, 1.77 eV) for the dye-loaded TiO2 film is
roughly taken as the 0-0 transition energy (E0-0).22 Therefore,
the lowest unoccupied molecular orbital (LUMO) energy level
is deduced to be -0.87 V vs NHE by
above 600 nm toward the long-wavelength region stems from
the decrease in the extinction coefficient of this dye. On the
other hand, the drop of IPCE blow 450 nm is mainly ascribed
to the filtering of light by the triiodide anions. The peak located
at ∼350 nm in the action spectrum is due to the band gap
excitation of TiO2 under UV light. Taking the light losses due
to light absorption and reflection by the conducting glass and
the filtering of light by the electrolyte, IPCEs at wavelengths
below 600 nm can be considered to be unity. NKX-2753 showed
a broad feature in the action spectrum with IPCE values
exceeding 80% below 600 nm. From this point of view, NKX-
2753 rivals N3.3
Photovoltaic Performance. A sandwich-type DSSC com-
posed of NKX-2753 dye-sensitized TiO2 film was illuminated
under 100 mW cm-2 of simulated AM 1.5 solar light, and the
typical current-voltage curve is shown in Figure 5. The
electrolyte used is 0.1 M LiI, 0.05 M I2, 0.6 M DMPImI, and
0.1 M GT in dry AN. Like TBP, GT can also improve Voc as
Gra¨tzel reported.33 The cell generated 16.1 mA cm-2 of Jsc,
0.60 V of open-circuit photovoltage (Voc), and 0.69 of fill factor
(FF), corresponding to an energy conversion efficiency (η) of
6.7%. The obtained Jsc (16.1 mA cm-2) almost equals to the
integrated photocurrent (16.06 mA cm-2) from the action
spectrum and the standard global AM1.5 solar emission
spectrum. This means the spectral mismatch is very small.
Figure 6 shows the dependence of Jsc and Voc on light
intensity for the NKX-2753-based cell. Jsc increases linearly
with light intensity, while Voc increases linearly with the
LUMO ) HOMO - E0-0
(1)
The CV for NKX-2586-sensitized TiO2 was also measured
at the same conditions. The HOMO and LUMO for NKX-2586
were determined to be 0.91 and -0.79 V vs NHE, respectively.
It is apparent that the LUMO levels for both dyes are more
negative than the conduction band edge of TiO2 (-0.5 V vs
NHE at pH 7),2 rendering electron injection thermodynamically
possible.
Photocurrent Action Spectrum. Incident monochromatic
photon-to-electron conversion efficiency (IPCE) is calculated
by
1240 × Jph[µA/cm2]
IPCE(λ) )
(2)
λ(nm) × P[W/m2]
where Jph is the photocurrent density generated by monochro-
matic light with wavelength λ and intensity of P. IPCE as a
function of excitation wavelength is plotted in Figure 4. IPCEs
exceed 80% in the range of 450 and 600 nm, attaining a
maximum value of 84% at 540 nm. The decrease of the IPCE

Photoelectrochemical Properties of a Coumarin Dye
J. Phys. Chem. B, Vol. 109, No. 9, 2005 3911
Figure 6. (Left) Dependence of short-circuit photocurrent (Jsc) on the
light intensity (P). (Right) Dependence of open-circuit photovoltage
(Voc) on the logarithm of light intensity for NKX-2753-based solar cell.
Figure 7. UV-vis absorption spectra for NKX-2586 in ethanol (‚ ‚ ‚)
and on a 3.5 µm TiO2 nanocrystalline film (- - -), and IPCE action
spectra for dye-sensitized TiO2 film (double layer, 12 µm) without
(- ‚ -) and with DCA (- ‚ ‚ -). IPCE action spectrum (s) for NKX-
2753-loaded TiO2 film (double layer, 12 µm) is also shown in this
figure for convenience of comparison.
logarithm of the light intensity. With increasing light intensity,
the amount of active dye molecules becomes larger and the
transport kinetics of the redox couples plays an important role
in the regeneration of dye and hence the photocurrent generation.
The linear increase of Jsc with light intensity indicates that the
absorption band is observed for an H-aggregate.35 We previously
reported the photovoltaic performance of NKX-2586 and found
that dye aggregation limited the cell efficiency. Figure 7 shows
the UV-vis spectra for NKX-2586 both in solution and on TiO2
film and IPCE action spectra for NKX-2586 and DCA/NKX-
2586. The action spectrum of NKX-2753 is also included in
Figure 7 for comparison.
NKX-2586 showed a broad feature in the action spectrum
with IPCE values less than 60% as seen in Figure 7. IPCE can
be expressed by the product of light harvesting efficiency (LHE),
the quantum yield of electron injection (φinj), and the efficiency
of collecting the injected electron (φcoll) at the conducting glass
substrate.3
-
transport of I-/I3 ions from and to the counter electrode is
fast enough to regenerate the dye. It is therefore concluded that
photocurrent generation in this system is not limited by the
diffusion of the redox species in the electrolyte up to one sun.3
In liquid-junction DSSCs Voc is determined by33b,34
Voc ) (nRT/F)ln[Jsc/J0 - 1]
(3)
where n is the ideality factor, Jsc and J0 are the short-circuit
photocurrent and the reverse saturation current, respectively,
and R and F are the ideal gas and Faraday constants, respec-
tively. If Jsc is much larger than J0, Voc should be proportional
to the logarithm of Jsc, and thus the light intensity because of
the linear relationship between the latter two factors. This is
the case of the system reflected from the linear dependence of
Voc on the logarithm of light intensity. A similar dependence
of Jsc and Voc on light intensity was also observed for NKX-
2586 (not shown).
Stability. The poor stability of the organic dye limits its
practical applications. Without electrolyte, most of the organic
dyes are easy to decompose at the surface of TiO2 but are stable
in the presence of redox species because of the quick regenera-
tion of the dye by the iodide ions. In this study sealed cells
with electrolyte of 0.1 M LiI, 0.05 M I2, 0.1 M GT, and 0.6 M
DMPImI in dry AN were used to conduct stability tests at room
temperature. The cells were covered with an L42 cutoff filter
(λ > 420 nm) and illuminated under simulated AM1.5 solar
light (100 mW cm-2) in the open-circuit mode for 1-2 h each
day. After 2 months of operation (about 100 h illumination)
the efficiencies were decreased by 5-10%, indicating good
stability. The cell stability depends on the nature of the dye,
the solvent in the electrolyte, temperature, and the sealing
condition. To obtain a stable cell, detailed study on the long-
term stability is highly needed.
IPCE(λ) ) LHE(λ) × φinj(λ) × φcoll(λ)
(4)
If all three factors in eq 4 are unity, the maximum IPCE should
fall into the range of 80-85% taking the absorption and
reflection by TCO glass into account. Obviously, the maximum
IPCE value obtained for NKX-2586 is much lower than the
maximum. Since the LHE of NKX-2586-adsorbed film at the
maximum absorption, where the absorbance is higher than 5,
is unity, either φinj or φcoll or both should be less than unity.
H-Aggregation usually quenches fluorescence efficiently and
would consequently be a loss mechanism competing with
electron injection, resulting in lower IPCEs in the visible
region.26 We attribute the lower IPCEs for NKX-2586 to the
formation of H-aggregates on the TiO2 surface, as discussed
below.
As we can see from the absorption spectra in Figure 7, the
maximum absorption of NKX-2586 was blue shifted by 50 nm
from 513 nm in ethanol solution to 463 nm on TiO2 surface. In
addition, the absorption peak became much broader upon
adsorption on TiO2 film from a solution. In agreement with the
absorption peak for the dye-loaded film, a peak at 445 nm was
observed in the action spectrum. The blue shift by 50 nm of
the maximum absorption band, broadening of the absorption
spectrum, and low IPCEs generated suggest the formation of
H-aggregates when NKX-2586 was adsorbed on the TiO2
surface. For the dyes tending to aggregate, DCA or the like is
usually used as a coadsorbate to prevent aggregation for the
purpose of improving cell performance.4,22,24,36 When 20 mM
DCA was included in 0.3 mM NKX-2586, the surface concen-
tration of the dye on TiO2 was decreased by 70%, but
Discussion
Dye Aggregation. Dyes are known to form aggregation,
which is due to the strong coupling between the molecules that
causes either a blue shift or a red shift in the absorption band
of the aggregate. For a J-type aggregate the absorption band is
red shifted relative to the monomer, while a blue shift in the

3912 J. Phys. Chem. B, Vol. 109, No. 9, 2005
Wang et al.
Figure 8. Current-potential curves obtained in the dark for NKX-
2586 and NKX-2753 dye-sensitized solar cells. The electrolyte used
was 0.1 M LiI, 0.05 M I2, and 0.6 M DMPImI in AN.
Figure 9. Photocurrent-photovoltage characteristics for NKX-2753
dye-sensitized solar cells at various concentrations of DCA in the dye
solutions. Electrolyte was 0.1 M LiI, 0.05 M I2, 0.2 M TBP, and 0.6
M DMPImI in dry AN.
surprisingly the IPCEs between 430 and 680 nm were improved
significantly (Figure 7). It is reasonable to attribute this
remarkable improvement in IPCE to the prevention of dye
aggregation due to DCA.36 Narrowing of the action spectrum
with DCA also supports the formation of aggregation in NKX-
2586 because H-aggregation usually broadens the absorption
spectrum as compared to the monomer. Although coadsorption
of DCA increased IPCE to some extent, the remarkable loss of
dye adsorption unavoidably limits the photocurrent generation.36
To solve this problem NKX-2753 with a side ring, linked to
the alkene chain, was designed and prepared (Scheme 1). In
contrast to NKX-2586, the UV-vis absorption spectrum of
NKX-2753 was not changed much upon adsorption on TiO2
film from solution (Figure 1). It is concluded from the spectral
shifts that NKX-2586 aggregated on the TiO2 surface whereas
NKX-2753 did not. The side ring, acting as a spacer, reduces
the coupling between the molecules and therefore prevented
aggregation. From this point of view the side ring has a similar
function as DCA. However, coadsorption of DCA decreased
dye adsorption significantly, while the side ring in the dye did
not influence surface concentration of the dye so much. By
desorbing the dye into a basic solution we estimated the dye
amount by measuring the absorption spectrum of the resultant
solution. The surface concentrations were determined to be (7.6
( 0.2) × 10-8 and (9.3 ( 0.2) × 10-8 M cm-2 for NKX-2753
and NKX-2586, respectively. NKX-2753 is bigger than NKX-
2586 in molecular size, and therefore, it is reasonable to observe
such a difference in the adsorbed amount of dye. Although
NKX-2753 has a lower surface concentration on TiO2 film
compared to that of NKX-2586, it generated much higher IPCEs
than NKX-2586 in the visible region (Figure 7), which can be
explained by the prevention of dye aggregation.
90 mV increase in Voc. The dark current represents the reaction
between injected electrons and the triiodides. Compared with
NKX-2586, the reduced dark current and negative shift of the
onset potential for dark current indicate that NKX-2753 favors
electron injection and impairs the escape of photoinjected
electrons,37 accounting for the improvements of Jsc and Voc. This
significant improvement of cell performance from NKX-2586
to NKX-2753 supports the prevention of dye aggregation in
the latter dye due to the side ring.
Another piece of evidence for the prevention of the aggrega-
tion in NKX-2753 comes from the effect of DCA on the
photovoltaic performance. As we reported previously, co-
grafting of DCA with NKX-2586 improves IPCE and hence
Jsc significantly despite the remarkable loss of dye adsorption.36
The mixed DCA/NKX-2586 seems better than the dye alone in
terms of cell efficiency, which was explained by the observation
of reduced dark current and negative shift of dark current onset
due to the coadsorbate.36 Contrasting to NKX-2586, DCA gave
a different effect on the performance of NKX-2753. Figure 9
shows the effect of DCA on I-V curves of NKX-2753. With
increasing the concentration of DCA, Jsc decreased dramatically
while Voc remained almost same. This behavior is quite different
from that of NKX-2586, where suppression of charge recom-
bination due to DCA overrides the significant loss of dye
adsorption, resulting in increased Jsc and Voc.36 Compared with
NKX-2586, DCA had a larger influence on adsorption of NKX-
2753 on the TiO2 surface. At 20 mM DCA, the surface
concentration of NKX-2586 was decreased by about 70%
whereas NKX-2753 was not adsorbed on the TiO2 surface. This
suggests that NKX-2753 has a weaker interaction than NKX-
2586 between molecules, consistent with the spectral shifts
(Figures 1 and 7). The presence of DCA with NKX-2753 did
not influence the dark current noticeably, resulting in a similar
Voc. The decreased Jsc was ascribed to the decreased dye
adsorption. Unlike NKX-2586, DCA is not needed for NKX-
2753 to improve cell performance. This means that charge
recombination in NKX-2753 is much less significant than that
in NKX-2586. Interestingly, the increase in Voc of 90 mV
induced by coadsorption of DCA36 is equal to that by the side
ring. This suggests that the side ring has a similar effect on Voc
as DCA.
NKX-2753 is superior to NKX-2586 in terms of photovoltaic
performance. While NKX-2586 produced an overall efficiency
of 3.1% (Jsc ) 13.7 mA cm-2, Voc ) 0.45 V, FF ) 0.50), NKX-
2753 generated an overall efficiency of 5.7% (Jsc ) 16.9 mA
cm-2, Voc ) 0.54 V, FF ) 0.63) using 0.1 M LiI, 0.05 M I2,
and 0.6 M DMPImI in AN as electrolyte. All three parameters
were improved significantly upon introduction of a side ring to
the alkene chain, resulting in an increase in efficiency by 84%.
Dark current measurement can give a plausible explanation for
these observations. The dark current as a function of applied
potential is plotted in Figure 8. It is evident that introduction
of the side ring suppressed the dark current significantly and
shifted the onset potential of dark current from -0.27 to -0.37
V. The 100 mV negative shift is in good agreement with the
Of course, there are many other factors, such as HOMO and
LUMO energy levels, influencing photocurrent generation.38
Assuming that 0.2 eV of the energy difference between the
LUMO of the dye and the conduction band edge of TiO2 is

Photoelectrochemical Properties of a Coumarin Dye
J. Phys. Chem. B, Vol. 109, No. 9, 2005 3913
TABLE 1: Effect of TBP on the Cell Performance for
NKX-2753a
of the GT effect on NKX-2586 was observed. Like NKX-2753,
the best efficiency for NKX-2586 (5.4%) was obtained at 0.1
M GT too.
c (TBP)/M
Jsc/mA cm-2
Voc/V
FF
η (%)
The GT effect is similar to the TBP effect. However, the
increase in Voc by GT (Table 2) is less remarkable than that by
TBP (Table 1), while the decrease in Jsc by TBP is larger than
that by GT. As a consequence, GT treatment gave a little higher
efficiency than TBP treatment.
0
16.9
14.6
14.1
12.5
11.2
10.3
0.54
0.61
0.62
0.64
0.66
0.68
0.63
0.70
0.74
0.73
0.74
0.75
5.7
6.2
6.5
5.9
5.5
5.3
0.1
0.2
0.3
0.4
0.5
a The cell was illuminated under 100 mW cm-2 of simulated AM1.5
solar light. The electrolyte used was 0.1 M LiI, 0.05 M I2, 0.6 M
DMPImI, and TBP in dry AN.
Conclusions
To prevent dye aggregation of NKX-2586 and hence improve
the cell efficiency, a new coumarin dye with a side ring, NKX-
2753, was designed and synthesized. The side ring linked to
the alkene chain was proved to be effective to prevent dye
aggregation as DCA. For this reason, DCA is not needed to
coadsorb with NKX-2753 on the TiO2 surface, ensuring a high
surface concentration of dye and efficient photon-to-electron
conversion. From NKX-2586 to NKX-2753 the extra side ring
resulted in a remarkable improvement in Jsc, Voc, and FF. This
finding may be useful to design and synthesize new efficient
dye sensitizers in DSSCs. In particular, the cell based on NKX-
2753 generated an efficiency of 5.7% in the absence of additive
such as TBP and GT; such a high value is very important for
long-term stability.
TABLE 2: Effect of GT on the Cell Performance for
NKX-2753a
c(GT)/M
Jsc/mA cm-2
Voc/V
FF
η (%)
0
16.9
16.0
15.7
15.5
13.6
13.4
0.54
0.57
0.58
0.62
0.63
0.65
0.63
0.66
0.70
0.70
0.70
0.70
5.7
6.0
6.4
6.7
6.0
6.1
0.01
0.05
0.1
0.3
0.5
a The cell was illuminated under 100 mW cm-2 of simulated AM1.5
solar light. The electrolyte used was 0.1 M LiI, 0.05 M I2, 0.6 M
DMPImI, and GT in dry AN.
necessary for efficient electron injection,38 the driving force for
electron injection of NKX-2753 and NKX-2586 should be
similar. On the basis of the spectral shifts, IPCE values, and
energy levels we analyzed the side ring in NKX-2753 acted as
a spacer to prevent dye aggregation, leading to remarkable
improvement in device performance compared with NKX-2586.
Acknowledgment. This work was supported by the New
Energy and Industrial Technology Development Organization
(NEDO) under the Ministry of Economy of Japan.
Supporting Information Available: Synthesis procedure
and NMR data for NKX-2753. This material is available free
of charge via the Internet at http://pubs.acs.org.
Effect of TBP and GT. Treating the dye-sensitized TiO2
electrode with TBP significantly improves both Voc and the
energy conversion efficiency for N3.3 Frank and co-workers
studied the effect of TBP on charge recombination in DSSCs
based on N3. TBP was found to lower the rate constant for
triiodide reduction by 1-2 orders of magnitude.34 The sup-
pressed charge recombination between injected electrons and
triiodide ions and the negative movement of conduction band
edge due to TBP are responsible for the improved Voc.34 TBP
is also effective in improving the Voc of some organic dyes.22,24,36
For example, 1 M TBP increases Voc from 0.47 to 0.73 V and
the overall efficiency from 3.6% to 7.7% for NKX-2677.35
Contrasting to NKX-2677, NKX-2753 is less sensitive to TBP.
The effect of TBP content in the electrolyte on cell performance
is summarized in Table 1. With increasing TBP, Jsc decreased,
References and Notes
(1) O’Regan, B.; Gra¨tzel, M. Nature 1991, 353, 737.
(2) Hagfeldt, A.; Gra¨tzel, M. Chem. ReV. 1995, 95, 49.
(3) Nazeerudin, M. K.; Kay, A.; Rodicio, I.; Humphry-Baker, R.;
Mu¨ller, E.; Liska, P.; Vlachopoulos, N.; Gra¨tzel, M. J. Am. Chem. Soc.
1993, 115, 6382.
(4) Nazeeruddin, M. K.; Pe´chy, P.; Renouard, T.; Zakeeruddin, S. M.;
Humphry-Baker, R.; Comte, P.; Liska, P.; Cevey, L.; Costa, E.; Shklover,
V.; Spiccia, L.; Deacon, G. B.; Bignozzi, C. A.; Gra¨tzel, M. J. Am. Chem.
Soc. 2001, 123, 1613.
(5) (a) Yanagida, M.; Islam, A.; Tachibana, Y.; Fujihashi, G.; Katoh,
R.; Sujihara, H.; Arakawa, H. New J. Chem. 2002, 26, 963. (b) Islam, A.;
Sujihara, H.; Yanagida, M.; Hara, K.; Fujihashi, G.; Tachibana, Y.; Katoh,
R.; Murata, S.; Arakawa, H. New J. Chem. 2002, 26, 966. (c) Hara, K.;
Sujihara, H.; Tachibana, Y.; Islam, A.; Yanagida, M.; Sayama, K.; Arakawa,
H. Langmuir 2001, 17, 5992.
(6) (a) Wang, Z.-S.; Huang, C.-H.; Zhang, B.-W.; Hou, Y.-J.; Xie, P.-
H.; Qian, H.-J.; Ibrahim, K. New J. Chem. 2000, 24, 567. (b) Wang, Z.-S.;
Huang, C.-H.; Huang, Y.-Y.; Zhang, B.-W.; Xie, P.-H.; Hou, Y.-J.; Ibrahim,
K.; Qian, H.-J.; Liu, F.-Q. Sol. Energy Mater. Sol. Cells 2002, 71, 261.
(7) (a) Amadelli, R.; Argazzi, R.; Bignozzi, C. A.; Scandola, F. J. Am.
Chem. Soc. 1990, 112, 7099. (b) Argazzi, R.; Bignozzi, C. A.; Yang, M.;
Hasselmann, G. M.; Meyer, G. J. Nano Lett. 2002, 2, 625.
(8) (a) Wang, P.; Zakeeruddin, S. M.; Moser, J. E.; Nazeeruddin, M.
K.; Sekiguchi, T.; Gra¨tzel, M. Nat. Mater. 2003, 2, 402. (b) Renouard, T.;
Fallahpour, R.-A.; Nazeeruddin, M. K.; Humphry-Baker, R.; Gorelsky, S.
I.; Lever, A. B. P.; Gra¨tzel, M. Inorg. Chem. 2002, 41, 367.
(9) Hou, Y.-J.; Xie, P.-H.; Zhang, B.-W.; Cao, Y.; Xiao, X.-R.; Wang,
W.-B. Inorg. Chem. 1999, 38, 6320.
(10) Kilsa, K.; Mayo, E. I.; Kuciauskas, D.; Villahermosa, R.; Lewis,
N. S.; Winkler, J. R.; Gray, H. B. J. Phys. Chem. A 2003, 107, 3397.
(11) Sauve, G.; Cass, M. E.; Doig, S. J.; Lauermann, I.; Pomykal, K.;
Lewis, N. S. J. Phys. Chem. B 2000, 104, 3488.
(12) Ferrere, S.; Gregg, B. A. J. Am. Chem. Soc. 1998, 120, 843.
(13) (a) Sayama, K.; Tsukugoshi, S.; Mori, T.; Hara, K.; Ohga, Y.;
Shinpou, A.; Abe, Y.; Suga, S.; Arakawa, H. Sol. Energy Mater. Sol. Cells
2003, 80, 47. (b) Ehret, A.; Stuhl, L.; Spitler, M. T. J. Phys. Chem. B 2001,
105, 9960.
V
oc increased, and FF increased until 0.2 M and then remained
almost unchanged. The best efficiency was obtained at 0.2 M
TBP for NKX-2753. Without TBP, NKX-2677, one of the best
organic dye sensitizers, only generated an efficiency of 3.6%
but NKX-2753 produced an efficiency of 5.7%, which is
comparable to N3 with an efficiency of 5.8% at similar
conditions.34 As TBP is able to desorb the dye from the TiO2
surface, it is very important to get high efficiency at low TBP
content or without TBP for the benefit of long-term stability.
GT as additive was first reported by Gra¨tzel.33 Coadsorption
of GT with N719 results in a remarkable improvement in Voc
due to a reduction in dark current.33 A new record efficiency
of 11% was achieved recently with this approach.33b We also
checked the effect of GT on the cell performance for NKX-
2753 and list the results in Table 2. With increasing GT
concentration, Jsc decreased, Voc increased, and FF increased
until 0.05 M GT. Consequently, the maximum energy conver-
sion efficiency (6.7%) was reached at 0.1 M GT. A similar trend
(14) (a) Sayama, K.; Hara, K.; Tsukugoshi, S.; Abe, Y.; Mori, T.;
Satsuski, M.; Suga, S.; Sugihara, H.; Arakawa, H. Chem. Commun. 2000,

3914 J. Phys. Chem. B, Vol. 109, No. 9, 2005
Wang et al.
1173. (b) Sayama, K.; Shingo, T.; Hara, K.; Ohga, Y.; Shinpo, A.; Abe,
Y.; Suga, S.; Arakawa, H. J. Phys. Chem. B 2002, 106, 1363.
(15) (a) Wang, Z.-S.; Li, F.-Y.; Huang, C.-H.; Wang, L.; Wei, M.; Jin,
L.-P.; Li, N.-Q. J. Phys. Chem. B 2000, 104, 9676. (b) Wang, Z.-S.; Li,
F.-Y.; Huang, C.-H. Chem. Commun. 2000, 2063. (c) Wang, Z.-S.; Li, F.-
Y.; Huang, C.-H. J. Phys. Chem. B 2001, 105, 9210. (d) Wang, Z.-S.; Huang,
Y.-Y.; Huang, C.-H.; Zheng, J.; Cheng, H.-M.; Tian, S.-J. Synth. Met. 2000,
114, 201. (e) Yao, Q.-H.; Meng, F.-S.; Li, F.-Y.; Tian, H.; Huang, C.-H. J.
Mater. Chem. 2003, 13, 1048.
A.; Ohga, Y.; Shinpo, A.; Suga, S.; Sayama, K.; Sugihara, H.; Arakawa,
H. J. Phys. Chem. B 2003, 107, 597.
(23) Enea, O.; Moser, J.; Gra¨tzel, M. J. Electroanal. Chem. 1989, 259,
59.
(24) Hara, K.; Kurashige, M.; Dan-oh, Y.; Kasada, C.; Shinpo, A.; Suga,
S.; Sayama, K.; Arakawa, H. New J. Chem. 2003, 27, 783.
(25) Horiuchi, T.; Miura, H.; Sumioka, K.; Uchida, S. J. Am. Chem.
Soc. 2004, 126, 12218.
(26) Khazraji, A. C.; Hotchandani, S.; Das, S.; Kamat, P. V. J. Phys.
(16) (a) Ferrere, S.; Zaban, A.; Gregg, B. A. J. Phys. Chem. B 1997,
101, 4490. (b) Ferrere, S.; Gregg, B. A. New J. Chem. 2002, 26, 1155. (c)
Wang, Z.-S.; Huang, C.-H.; Li, F.-Y.; Weng, S.-F.; Ibrahim, K.; Liu, F.-Q.
J. Phys. Chem. B 2001, 105, 4230. (d) Wang, Z.-S.; Huang, C.-H.; Li, F.-
Y.; Weng, S.-F.; Yang, S.-M. J. Photochem Photobiol. A 2001, 140, 255.
(e) Wang, Z.-S.; Huang, C.-H.; Li, F.-Y.; Yang, S.-M.; Weng, S.-F.; Fu,
X.-Y.; Wu, N.-Z.; Ibrahim, K.; Liu, F.-Q.; Qian, H.-J. Sci. China, Ser. B
2001, 44, 268.
(17) (a) Gerischer, H.; Tributsch, H. Ber. Bunsen-Ges. Phys. Chem. 1968,
72, 437. (b) Tsubomura, H.; Matsumura, M.; Nomura, Y.; Amamiya, T.;
Nature 1976, 261, 402. (c) Sayama, K.; Sugino, M.; Sugihara, H.; Abe,
Y.; Arakawa, H. Chem. Lett. 1998, 753.
(18) (a) Jasieniak, J.; Johnston, M.; Waclawik, E. R. J. Phys. Chem. B
2004, 108, 12962. (b) Watson, D. F.; Marton, A.; Stux, A. M.; Meyer, G.
J. J Phys. Chem. B 2004, 108, 11680. (c) Kroeze, J. E.; Savenije, T. J.;
Warman, J. M. J. Am. Chem. Soc. 2004, 126, 7608.
(19) (a) He, J.; Benko, G.; Korodi, F.; Polivka, T.; Lomoth, R.;
Akermark, B.; Sun, L.; Hagfeldt, A. J. Am. Chem. Soc. 2002, 124, 4922.
(b) Amao, Y.; Komori, T. Langmuir 2003, 19, 8872. (c) Huisman, C. L.;
Goossens, A.; Schoonman, J. J. Phys. Chem. B 2002, 106, 10578.
(20) (a) Hara, K.; Kurashige, M.; Ito, S.; Shinpo, A.; Suga, S.; Sayama,
K.; Arakawa, H. Chem. Commun. 2003, 252. (b) Kitamura, T.; Ikeda, M.;
Shigaki, K.; Inoue, T.; Anderson, N. A.; Ai, X.; Lian, T.; Yangagida, S.
Chem. Mater. 2004, 16, 1806.
Chem. B 1999, 103, 4693.
(27) Wang, Z. S.; Kawauchi, H.; Kashima, T.; Arakawa, H. Coord.
Chem. ReV. 2004, 248, 1381.
(28) Argazzi, R.; Binozzi, C. A.; Heimer, T. A.; Castellano, F. N.; Meyer,
G. J. Inorg. Chem. 1994, 33, 5741.
(29) Socrates, G. Infrared Characteristic Group Frequencies, 2nd ed.;
Wiley & Sons Ltd.: Baffins Lane, Chichester, England, 1994.
(30) Shklover, V.; Ovchinnikov, Y. E.; Braginsky, L. S.; Zakeeruddin,
S. M.; Gra¨tzel, M. Chem. Mater. 1998, 10, 2533.
(31) Deacon, G. B.; Phillips, R. J. Coord. Chem. ReV. 1980, 33, 227.
(32) (a) Finnie, K. S.; Bartlett, J. R.; Woolfrey, J. L. Langmuir 1998,
14, 2744. (b) Duffy, N. W.; Dobson, K. D.; Gordon, K. C.; Robinson, B.
H.; Quillan, A. J. M. Chem. Phys. Lett. 1997, 266, 451. (c) Klein, C.;
Nazeeruddin, M. K.; Censo, D. D.; Liska, P.; Gra¨tzel, M. Inorg. Chem.
2004, 43, 4216. (d) Bauer, C.; Boschloo, G.; Mukhtar, E.; Hagfeldt, A. J.
Phys. Chem. B 2002, 106, 12693.
(33) (a) Gra¨tzel, M. J. Photochem. Photobiol. C 2003, 4, 145. (b) Gra¨tzel,
M. J. Photochem. Photobiol. A 2004, 164, 3.
(34) Huang, S. Y.; Schlichtho¨rl, G.; Nozik, A. J.; Gra¨tzel, M.; Frank,
A. J. J. Phys. Chem. B 1997, 101, 2576.
(35) (a) McRae, E. G.; Kasha, M. J. Chem. Phys. 1958, 28, 721. (b)
Kasha, M.; Rawls, H. R.; El-Bayoumi, M. A. Pure Appl. Chem. 1965, 11,
371.
(36) Hara, K.; Dan-oh, Y.; Kasada, C.; Ohga, Y.; Shinpo, A.; Suga, S.;
Sayama, K.; Arakawa, H. Langmuir 2004, 20, 4205.
(37) Wang, P.; Zakeeruddin, S. M.; Comte, P.; Charvet, R.; Humphry-
Baker, R.; Gra¨tzel, M. J. Phys. Chem. B 2003, 107, 14336.
(38) Katoh, R.; Furube, A.; Yoshihara, T.; Hara, K.; Fujihashi, G.;
Takano, S.; Murata, S.; Arakawa, H.; Tachiya, M. J. Phys. Chem. B 2004,
108, 4818.
(21) (a) Senadeera, G. K. R.; Nakamura, K.; Kitamura, T.; Wada, Y.;
Yanagida, S. Appl. Phys. Lett. 2003, 83, 5470. (b) Kim, Y.-G.; Walker, J.;
Samuelson, L. A.; Kumar, J. Nano Lett. 2003, 3, 523.
(22) (a) Hara, K.; Sayama, K.; Ohga, Y.; Shinpo, A.; Suga, S.; Arakawa,
H. Chem. Commun. 2001, 569. (b) Hara, K.; Tachibana, Y.; Ohga, Y.;
Shinpo, A.; Suga, S.; Sayama, K.; Sugihara, H.; Arakawa, H. Sol. Energy
Mater. Sol. Cells 2003, 77, 89. (c) Hara, K.; Sato, T.; Katoh, R.; Furube,