Dye-sensitized solar cells based on donor-acceptor π-conjugated fluorescent dyes with a pyridine ring as an electron-withdrawing anchoring group

Article abstract of DOI:10.1002/anie.201102552

Firmly attached: The short-circuit photocurrent density and conversion yields of solar cells based on pyridine-containing donor-acceptor π-conjugated (D-π-A) dyes (see scheme) are greater than those for conventional D-π-A dye sensitizers that bear a carbo

Full text of DOI:10.1002/anie.201102552

DOI: 10.1002/anie.201102552
Solar Cells
Dye-Sensitized Solar Cells Based On Donor–Acceptor p-Conjugated
Fluorescent Dyes with a Pyridine Ring as an Electron-Withdrawing
Anchoring Group**
Yousuke Ooyama, Shogo Inoue, Tomoya Nagano, Kohei Kushimoto, Joji Ohshita, Ichiro Imae,
Kenji Komaguchi, and Yutaka Harima*
Since the first report of dye-sensitized solar cells (DSSCs) by
Grꢀtzel and co-workers in 1991,[1] DSSCs have received
considerable attention because of their high incident solar
light to electricity conversion efficiency and low cost of
production. Recently, in order to increase power conversion
efficiency, much research has focused on the development of
new metal-free organic dye sensitizers.[2–9] In particular,
donor–acceptor p-conjugated (D-p-A) dyes with both the
electron-donating (D) and electron-accepting (A) groups
linked by a p-conjugated bridge that exhibits broad and
intense absorption, are proposed as one of the most promising
organic dye sensitizers. The spectral features of the d-p-A
dyes are associated with the intramolecular charge transfer
(ICT) excitation from the donor to the acceptor moiety of the
dye, thus leading to efficient electron transfer from the
excited dye through the acceptor moiety into the conduction
band (CB) of TiO2. Most of the D-p-A dyes, such as
coumarins, polyenes, thiophenes, and indolines, have dialkyl-
amine or diphenylamine moieties as electron donor, and
carboxylic acid, cyanoacrylic acid, or rhodanine-3-acetic acid
moieties that act as electron acceptors as well as anchoring
groups for attachment to a TiO2 surface.[2–9] The carboxyl
group enables good electronic communication between the
dye and TiO2 by forming a strong ester linkage with the TiO2
surface. However, development of new d-p-A dyes for
DSSCs is limited as long as the carboxyl group is employed
as an anchor and electron acceptor. To create new and
efficient d-p-A dyes for DSSCs, an new molecular design such
as formation of a strong interaction between the electron-
accepting moiety of sensitizers and TiO2 surface is required.
Herein, we propose the use of a pyridine ring as an
electron-withdrawing and anchoring group in place of a
conventional carboxy group. We have demonstrated that the
newly developed d-p-A fluorescent dyes NI3–6 with a
pyridine ring can adsorb onto a TiO2 surface by strong
coordinate bonding between the lone pair of electrons on the
nitrogen atom of the pyridine, and the Lewis acid sites of TiO2
to give electron injection efficiencies comparable to or higher
than those for NI1 and NI2 where the pyridine ring is replaced
with carboxyphenyl group (Scheme 1; see the Supporting
Information for the detailed synthetic procedures).
The absorption and fluorescence spectra of NI1–6 in 1,4-
dioxane are shown in Figure 1 and their spectral data are
Scheme 1. Chemical structures of d-p-A fluorescent dyes NI1–6.
[*] Dr. Y. Ooyama, S. Inoue, T. Nagano, K. Kushimoto,
Prof. Dr. J. Ohshita, Dr. I. Imae, Dr. K. Komaguchi, Prof. Dr. Y. Harima
Department of Applied Chemistry
Graduate School of Engineering, Hiroshima University
Higashi-Hiroshima 739-8527 (Japan)
E-mail: harima@mls.ias.hiroshima-u.ac.jp
[**] This work was supported by Grants-in-Aid for Young Scientists (B)
(22750179) from the Japan Society for the Promotion of Science
(JSPS).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201102552.
Figure 1. Absorption (a) and fluorescence (c) spectra of NI1–6
in 1,4-dioxane.
Angew. Chem. Int. Ed. 2011, 50, 7429 –7433
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7429

Communications
Table 1: Optical and electrochemical data, HOMO and LUMO energy levels, and DSSC performance parameters of NI1–6.
Dye labs[nm][a]
(e [mꢀ1 cmꢀ1]) (FF)
lem [nm][a] E1/2 [V][b] HOMO [V][c] LUMO [V][c] Molecules/cm2 [d] Jsc [mAcmꢀ2 [e]
]
Voc [mV][e] Fill
factor[e]
h [%][e]
ox
NI1 374 (34900)
NI2 376 (34300)
438 (0.89) 0.30
442 (0.87) 0.37
0.93
1.00
ꢀ2.12
ꢀ2.03
5.3ꢀ1016
10.4ꢀ1016
4.8ꢀ1016
10.8ꢀ1016
4.9ꢀ1016
4.7ꢀ1016
7.9ꢀ1016
8.0ꢀ1016
1.99
2.96
1.80
3.07
3.16
3.35
5.80
5.63
516
503
517
520
524
522
540
548
0.59
0.61
0.60
0.61
0.63
0.62
0.60
0.60
0.60
0.91
0.56
0.97
1.04
1.15
1.89
1.84
NI3 372 (30200)
NI4 375 (33000)
NI5 394 (48100)
NI6 396 (49600)
423 (0.83) 0.34
423 (0.84) 0.39
465 (0.60) 0.30
464 (0.58) 0.34
0.97
1.02
0.93
0.97
ꢀ2.15
ꢀ2.07
ꢀ1.93
ꢀ1.87
[a] In 1,4-dioxane. [b] Half-wave potentials for oxidation (E1/2ox) vs. Fc/Fc+ were recorded in CH2Cl2/Bu4NClO4 (0.1m) solution. [c] Potentials recorded
vs. the NHE. [d] Adsorption amount per unit area of TiO2 film was controlled by the immersion time of TiO2 electrode in the dye solution. [e] The
photocurrent–voltage characteristics were measured under a simulated solar light conditions (AM 1.5, 100 mWcmꢀ2).
summarized in Table 1. All the dyes show two absorption
maxima: the band at around 300–315 nm is ascribed to a p!
p* transition, and the band at around 370–400 nm is assigned
to the ICT excitation from the donor (diphenylamino group)
to the acceptor (carboxyphenyl group for NI and NI2 and
pyridine ring for NI3–6). The ICT bands of NI5 and NI6 occur
at a wavelength that is approximately 20 nm longer than those
of NI1–4. Furthermore, the molar extinction coefficients (e)
for the ICT bands of NI5 and NI6 are approximately
50000mꢀ1 cmꢀ1 higher than the values of 30000–
35000mꢀ1 cmꢀ1 for NI1–4. These results show that the
introduction of thiophene unit onto the carbazole skeleton
expands the p conjugation in the dye and thus results in the
red-shift of absorption maximum and enhancement of the
extinction coefficient. The corresponding fluorescence max-
imum (lem) occurs at around 420–465 nm. The fluorescent
dyes NI1–4 (FF ꢁ 0.85) exhibit a higher fluorescence quantum
yield FF than those of NI5 and NI6 (FF ꢁ 0.6).
lengths are blue-shifted by approximately 10 nm for NI3 and
NI4 and 20 nm for NI5 and NI6, although the peak wave-
lengths are still red-shifted compared with those in 1,4-
dioxane. In contrast, the absorption peak wavelengths of NI1
and NI2 adsorbed on TiO2 with coadsorption of CDCA are
similar to those in 1,4-dioxane. These results show that the
red-shifts of NI3–6 adsorbed on TiO2 are due to the strong
interaction between the dyes and TiO2 surface.
The electrochemical properties of all the dyes were
determined by cyclic voltammetry (CV; see Figure S1 and
Table S1 in the Supporting Information for the electrochem-
ical properties). The oxidation peaks of NI1–6 were observed
at 0.34–0.42 V versus ferrocene/ferrocenium (Fc/Fc+) and the
corresponding reduction peaks appeared at 0.26–0.35 V, thus
showing that the oxidized states of all the dyes are stable. The
highest occupied molecular orbital (HOMO) and lowest
unoccupied molecular orbital (LUMO) energy levels of all
the dyes were evaluated from the spectral analyses and the
half-wave potentials for oxidation of NI1–6 (E1/2ox = 0.30–
0.39 V). The HOMO energy levels for NI1–6 were 0.93–
1.02 V versus the normal hydrogen electrode (NHE), thus
indicating that all the dyes have similar HOMO energy levels
that are more positive than the I3ꢀ/Iꢀ redox potential (0.4 V).
This property assures an efficient regeneration of the oxidized
Absorption spectra of the dyes adsorbed on TiO2 nano-
particles are shown in Figure 2 (see the Supporting Informa-
tion for details of the measurements). The absorption peak
wavelengths (labs) are red-shifted by approximately 10 nm for
NI1 and NI2, 25 nm for NI3 and NI4, and 30 nm for NI5 and
NI6 compared with those in 1,4-dioxane. Chenodeoxycholic
acid (CDCA) has been employed as a coadsorbent to prevent
dye aggregation on the TiO2 surface. When CDCA is
coadsorbed with NI3–6 on TiO2, the absorption peak wave-
dyes by electron transfer from Iꢀ/I3 redox couple in the
ꢀ
electrolyte. The LUMO energy levels of the dyes were
ox
estimated from E1/2 and an intersection of absorption and
fluorescence spectra (407 and 409 nm (3.05 and 3.03 eV) for
NI1 and NI2, 398 and 401 nm (3.12 and 3.09 eV) for NI3 and
NI4, and 434 and 436 nm (2.86 and 2.84 eV) for NI5 and NI6),
which correspond to the energy gap between the HOMO and
the LUMO. The LUMO energy levels of NI1–6 were ꢀ2.12,
ꢀ2.03, ꢀ2.15, ꢀ2.07, ꢀ1.93, and ꢀ1.87 V, respectively. Evi-
dently, these levels are higher than the energy level of the CB
of TiO2 (ꢀ0.5 V), so that these dyes can efficiently inject
electrons to the TiO2 electrode.
The MO calculations (AM1 and INDO/S)[10,11] showed
that the absorption bands of these dyes were mainly assigned
to the HOMO–LUMO transition, where HOMOs were
mostly localized on the diphenylamino–carbazole moiety for
NI1–4 and the diphenylamino–thiophenyl–carbazole moiety
for NI5 and NI6, and LUMOs were mostly localized on the
carboxyphenyl–carbazole moiety for NI1 and NI2, the
pyridinyl–carbazole moiety for NI3 and NI4, and the pyr-
Figure 2. Absorption spectra of a) NI1, NI3, and NI5 and b) NI2, NI4,
and NI6 adsorbed on TiO2 nanoparticles with (c) and without
(a) CDCA as coadsorbent. The y axis is expressed in terms of the
Kubelka–Munk equation K/S=(1-R)2/2R, where K is the absorption
coefficient, S is the scattering coefficient, and R is the fractional
reflectance.
7430 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 7429 –7433

idinyl–thiophenyl–carbazole moiety for NI5 and NI6 (Fig-
ure 3a), thus suggesting a strong ICT nature upon irradiation
(Figure 3b).
Figure 4 shows the FTIR spectra of the dye powders and
the dyes adsorbed on TiO2. For the powders of NI1 and NI2,
=
the C O stretching band of carboxy group was observed at
1680 cmꢀ1. When the dyes were adsorbed on TiO2 surface, the
ꢀ1
=
C O stretching band at 1680 cm disappeared. The obser-
vations indicate that the carboxy groups of the dyes form an
ester linkage with TiO2 surface.[12] On the other hand, the
=
=
characteristic stretching bands for C N or C C were clearly
observed at around 1590, 1490, and 1460 cmꢀ1 for all the dye
powders. In the FTIR spectra of NI3–6 adsorbed on TiO2, a
new band appeared at around 1615 cmꢀ1, which is assigned to
the pyridine ring coordinated to the Lewis acid sites of the
TiO2 surface.[13,14] This band indicates that the dyes NI1 and
NI2 are adsorbed on the TiO2 surface by the ester linkage
alone at Brønsted acid sites (surface-bound hydroxy groups),
whereas the dyes NI3–6 are predominantly adsorbed on the
TiO2 surface by coordinate bonding at the Lewis acid sites
(exposed Tin+ cations). The strong coordinate bonding is
responsible for the large red-shift of the absorption peak for
NI3–6 adsorbed on TiO2 (Figure 2).
Figure 3. a) HOMO and LUMO of NI1, NI3, and NI5. The red and
blue lobes denote the positive and negative phases of the coefficients
of the molecular orbitals. The size of each lobe is proportional to the
MO coefficient. b) Calculated electron density changes accompanying
the first electronic excitation of NI1, NI3, and NI5. The black and
white lobes signify decrease and increase in electron density accom-
panying the electronic transition, respectively. The areas indicate the
magnitude of the electron density change. H light blue, C green,
N blue, O red, and S gold.
The DSSCs were prepared by using the dye-adsorbed
TiO2 electrode, platinum-coated glass as a counter electrode,
and an solution of 0.05m iodine, 0.1m lithium iodide, and 0.6m
1,2-dimethyl-3-propylimidazolium iodide in acetonitrile as
electrolyte. The photocurrent–voltage (I–V) characteristics
were measured under simulated solar light conditions
(AM 1.5, 100 mWcmꢀ2). The I–V curves and the incident
photon-to-current conversion efficiency (IPCE) spectra are
shown in Figure 5. The photovoltaic performance parameters
of DSSCs based on the dyes NI1–6 are given in Table 1. The
short-circuit photocurrent density (Jsc) and solar energy-to-
electricity conversion yield (h, %) for NI3 (3.16 mAcmꢀ2,
1.04%) and NI4 (3.35 mAcmꢀ2, 1.15%) are similar to those
for NI1 (2.96 mAcmꢀ2, 0.91%) and NI2 (3.07 mAcmꢀ2,
0.97%), when comparisons are made at maximum adsorption
amounts of dyes adsorbed on TiO2 (10.4 ꢁ 1016, 10.8 ꢁ 1016,
4.9 ꢁ 1016, and 4.7 ꢁ 1016 molecules per cm2 for NI1, NI2, NI3,
and NI4, respectively) when using 1 ꢁ 10ꢀ4 m dye solutions in
THF. The open-circuit photovoltages (Voc) for NI1–4 are
503 mV, 520 mV, 524 mV, and 552 mV, respectively, which
were slightly different among the four dyes. The maximum
IPCE values of NI1–4 (47–55%) are in good agreement.
Interestingly, when comparisons are made for similar
amounts of dyes adsorbed on TiO2 (5.3 ꢁ 1016, 4.8 ꢁ 1016,
4.9 ꢁ 1016, and 4.7 ꢁ 1016 moleculescmꢀ2 for NI1, NI2, NI3, and
NI4, respectively), the Jsc and h values increase in the order of
NI2 (1.80 mAcmꢀ2, 0.56%) ꢁ NI1 (1.99 mAcmꢀ2, 0.60%) !
NI3 (3.16 mAcmꢀ2, 1.04%) ꢁ NI4 (3.35 mAcmꢀ2, 1.15%). As
shown in Figure 6, this result indicates that the formation of
strong coordinate bonding between the pyridine ring of dyes
NI3 and NI4 and the Lewis acid sites of the TiO2 surface leads
to an efficient electron injection owing to good electronic
communication, rather than the formation of an ester linkage
between the dyes NI1 and IN2 and the Brønsted acid sites of
TiO2 surface. However, the maximum amounts of dyes
adsorbed on TiO2 for NI3 and NI4 are about half the
Figure 4. FTIR spectra of dye powders (a) and dyes adsorbed on
TiO2 nanoparticles (c) for a) NI1, b) NI2, c) NI3, d) NI4, e) NI5,
and f) NI6.
Angew. Chem. Int. Ed. 2011, 50, 7429 –7433
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7431

Communications
amounts of those for NI1 and NI2 because there are fewer
Lewis acid sites on the TiO2 surface than Brønsted acid
sites. On the other hand, the Jsc and h values for NI5
(5.80 mAcmꢀ2, 1.89%) and NI6 (5.63 mAcmꢀ2, 1.84%)
are larger than those for NI3 and NI4. The maximum
IPCE values of 65–70% for both NI5 and NI6 were
observed in the range from 410 to 470 nm. The relatively
high photovoltaic performances of NI5 and NI6 are
attributed to both the red-shift of the absorption band
and the good balance between the LUMO level of the dye
and the energy level of the CB of TiO2 by the introduction
of a thiophene unit to the p-conjugation system of the
dye. Thus, our results demonstrate that the dyes NI3–5
can inject electrons efficiently from the pyridine ring as
electron-withdrawing anchoring group to the CB of the
TiO2 electrode through the strong coordinate bonding
with the Lewis acid site of the TiO2 surface.
In conclusion, we have designed and synthesized
fluorescent dyes NI3–NI6 with a pyridine ring as electron-
withdrawing anchoring group as new D-p-A dye sensi-
tizers for DSSCs. We demonstrate that the formation of
coordinate bonding between the pyridine ring of NI3–NI6
and the Lewis acid sites of TiO2 surface leads to an
efficient electron injection arising from good electronic
communication, rather than the formation of an ester
linkage between NI1 and NI2 and the Brønsted acid sites
of TiO2 surface. Thus, we propose that the formation of a
coordinate bond between the pyridine ring and the Lewis
acid site of TiO2 surface can be a promising candidate for
electron-withdrawing anchoring group in new D-p-A dye
sensitizers for DSSCs.
Received: April 13, 2011
Revised: May 30, 2011
Published online: June 29, 2011
Keywords: donor–acceptor systems · dyes/pigments ·
.
fluorescence · titanium dioxide · solar cells
Figure 5. a) IPCE spectra and b) I–V curves of DSSCs based on NI1–6. The
amounts of adsorbed dyes on TiO2 film are 10.4ꢀ1016 (NI1), 10.8ꢀ1016
(NI2), 4.9ꢀ1016 (NI3), 4.7ꢀ1016 (NI4), 7.9ꢀ1016 (NI5), and
8.0ꢀ1016 molecules per cm2 (NI6). 9 mm thick TiO2 electrodes were used.
CDCA was not employed. The same color code is used in (a) and (b).
[1] a) B. OꢂRegan, M. Grꢀtzel, Nature 1991, 353, 737; b) A.
Hagfeldt, M. Grꢀtzel, Chem. Rev. 1995, 95, 49; c) U. Bach,
D. Lupo, P. Comte, J. E. Moser, F. Weissꢃrtel, J. Salbeck, H.
Spreitzer, M. Grꢀtzel, Nature 1998, 395, 583; d) M. Grꢀtzel,
Nature 2001, 414, 338.
[2] a) N. Robertson, Angew. Chem. 2006, 118, 2398; Angew. Chem.
Int. Ed. 2006, 45, 2338; b) N. Robertson, Angew. Chem. 2008,
120, 1028; Angew. Chem. Int. Ed. 2008, 47, 1012.
[3] Z. Chen, F. Li, C. Huang, Curr. Org. Chem. 2007, 11, 1241.
[4] Z. Ning, H. Tian, Chem. Commun. 2009, 5483.
[5] Y. Ooyama, Y. Harima, Eur. J. Org. Chem. 2009, 2903.
[6] A. Mishra, M. K. R. Fischer, P. Bꢀuerle, Angew. Chem. 2009,
121, 2510; Angew. Chem. Int. Ed. 2009, 48, 2474.
[7] A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo, H. Pettersson, Chem.
Rev. 2010, 110, 6595.
[8] T. Bessho, S. M. Zakeeruddin, C.-Y. Yeh, E. W.-G. Diau, M.
Grꢀtzel, Angew. Chem. 2010, 122, 6796; Angew. Chem. Int. Ed.
2010, 49, 6646.
[9] H. Choi, I. Raabe, D. Kim, F. Teocoli, C. Kim, K. Song, J.-H.
Yum, J. Ko, M. K. Nazeeruddin, M. Grꢀtzel, Chem. Eur. J. 2010,
16, 1193.
Figure 6. Configurations of a) NI1 and b) NI3 on the TiO2 surface.
7432
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 7429 –7433

[10] M. J. S. Dewar, E. G. Zoebisch, E. F. Healy, J. J. Stewart, J. Am.
Chem. Soc. 1985, 107, 3902.
b) A. A. Davydov in Infrared Spectroscopy of Adsorbed Species
on the Surface of Transition Metal Oxide, (Ed.: C. H. Rochester),
Wiley, Chichester, 1984; c) M. W. Urban, Vibrational Spectros-
copy of Molecules and Macromolecules on Surfaces, Wiley,
Chichester, 1993, pp. 171 – 185.
[11] a) J. E. Ridley, M. C. Zerner, Theor. Chim. Acta 1973, 32, 111;
b) J. E. Ridley, M. C. Zerner, Theor. Chim. Acta 1976, 42, 223;
c) A. D. Bacon, M. C. Zerner, Theor. Chim. Acta 1979, 53, 21;
d) H. A. Kurtz, J. J. P. Stewart, D. M. Dieter, J. Comput. Chem.
1990, 11, 82.
[12] Z.-S. Wang, K. Hara, Y. Don-oh, C. Kasada, A. Shinpo, S. Suga,
H. Arakawa, H. Sugihara, J. Phys. Chem. B 2005, 109, 3907.
[13] a) J. B. Peri, Catalysis: Science and Technology, Vol. 5 (Eds.: J. R.
Anderson, M. Boudart), Springer, Berlin, 1984, pp. 171 – 220;
[14] a) M. I. Zaki, M. A. Hasan, F. A. Al-Sagheer, L. Pasupulety,
Colloids Surf. A 2001, 190, 261; b) V. Vishwanathan, H.-S. Roh,
J.-W. Kim, K.-W. Jun, Catal. Lett. 2004, 96, 23; c) M. M.
Mohamed, W. A. Bayoumy, M. Khairy, M. A. Mousa, Micro-
porous Mesoporous Mater. 2007, 103, 174.
Angew. Chem. Int. Ed. 2011, 50, 7429 –7433
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7433