the triphenylamine and cyanoacrylic acid fragments has been already
demonstrated as a good strategy to produce stable and efficient
organic sensitizers.10 DEK1 and DEK2 contain the same donor and
acceptor groups, while the length and structure of the thiophenic
bridges differ. The DEK1 spacer is a tetrathiophene, while DEK2
contains a dithieno[3,2-b:20,30-d]thiophene condensed ring with a high
intermolecular p stacking and aggregation tendency, inserted
between thiophene rings.11 The conjugated bridge of both dyes
contains 3,4-dibutyl-thiophene ring with high sterical hindrance.
In the literature, D–p–A dyes containing 3-alkyl-thiophene moie-
ties12 and a disubstituted ring9 in the spacer giving efficient solar cells
are reported, but these sensitizers have different donor and acceptor
groups with respect to DEK1 and DEK2. The introduction of
a dialkyl substituted moiety enhances the dye solubility and hinders
both intermolecular interaction and aggregation allowing the prep-
aration of a DSC without a coadsorbant agent.
edge of titania, assuring the electron injection from the excited dyes to
the metal oxide semiconductor. In addition, the ground-state oxida-
tion potentials of these sensitizers are higher than the redox potential
(0.35 V vs. NHE) of the electrolyte iodine/triiodide couple, which
could lead to a fast dye-regeneration.
Dye loading on TiO2 surface was quantitatively evaluated by
UV-Vis spectrophotometry (see ESI† Fig. S3).
The adsorbed dyes were completely removed by means of a trie-
thylamine/chloroform (9/1, v/v) solvent mixture. DEK1 was found
to be loaded onto the titania surface with average values of
about 4.4 ꢂ 10ꢁ7 mol mmꢁ3, a higher amount than the loading of
3.5 ꢂ 10ꢁ8 mol mmꢁ3 found for DEK2.
Interaction of DEK1 and DEK2 with the titania surface was
investigated by diffuse reflectance infrared spectroscopy (DRIFT)
(see ESI† Fig. S4). Spectral features of dye-loaded photoanodes
suggest a unidentate coordination of carboxylic groups towards the
titania surface. The cyanide group is not involved in the dye
anchoring on the titania surface, its stretching vibration is completely
preserved in shape and position in both unloaded and loaded dye
spectra.
Experimental
Both dyes have been prepared using classical reactions: the synthetic
procedures and the characterization are described in the ESI†
(Fig. S1).
J–V characteristic curves of solar cells under simulated sunlight
irradiation are reported in Fig. 2 for DEK1 and DEK2 dyes with the
addition of DCA at various concentrations. As a comparison, the
functional properties of a cell fabricated using the same photoanode
sensitized by the commercial N719 dye is reported. The short-circuit
current density (JSC) linearly increases with the light power density,
while a logarithmic behaviour has been found for the open-circuit
voltage (VOC), as expected for a well operating DSC (see ESI† Fig. S5
and S6). Photoconversion efficiency (PCE) values up to 7.17% and
6.27% have been obtained for the DEK1 and DEK2 molecules
respectively, without DCA. Addition of DCA has no beneficial effect
The standard mesoporous double layer titania film was employed
for solar cell fabrication.13 A 100 nm TiO2 compact layer was
preliminarily spray pyrolized on the fluorine-doped tin dioxide (FTO)
conducting glass electrode, to inhibit charge recombination between
the FTO and the electrolyte. A 6 mm thick film of TiO2 nanoparticles
(Solaronix HT, 20 nm in size), was first tape cast on the FTO con-
ꢀ
ducting glass. After 15 min drying at 150 C, a 4 mm thick second
layer of light scattering particles, with bimodal size distribution (20
nm and 400 nm in diameter, Solaronix D paste) was added. Firing in
ambient atmosphere at 450 ꢀC for 30 min followed. The TiO2 elec-
trode was sensitized by a 20 h immersion into a dye solution (10ꢁ4
M
in toluene for both the dyes). A platinized FTO conducting glass was
applied as the counter electrode. The cell was assembled by separating
the electrodes by a 25 mm thick plastic spacer and injecting the elec-
trolyte by means of capillary force.
The cell performances were measured by using a ABET 2000 solar
simulator at AM 1.5G (100 mW cmꢁ2). The impedance spectroscopy
was done using a SOLARTRON 1260A Impedance/Gain-Phase
Analyzer, with an AC signal 20 mV in amplitude in the frequency
range between 10 mHz and 300 kHz. The applied bias was between
0 V and 100 mV above the open circuit voltage of the solar cell.
Results and discussion
DEK1 and DEK2 show absorption maxima in the visible region (477
nm and 465 nm respectively) and have high molar extinction coeffi-
cients (3 ¼ 53.5 ꢂ 103 Mꢁ1 cmꢁ1 and 3 ¼ 57.6 ꢂ 103 Mꢁ1 cmꢁ1
respectively) which are extremely important for DSC applications
since they will allow the use of thinner TiO2 films contributing to
a better charge separation and lower charge recombination. More-
over, the comparison of UV spectra in the solid state and in solution
evidences similar maxima for both DEK1 and DEK2, suggesting
a poor aggregation in film. The cyclic voltammogram of both DEK1
and DEK2 shows a reversible oxidation wave at 1.09 V and 1.03 V
respectively versus NHE due to the oxidation of the triphenylamine
moiety. The LUMO energy levels of DEK1 and DEK2 are ꢁ3.55 eV
and ꢁ3.50 eV, respectively and are higher than the conduction band
Fig. 2 (a) J–V characteristic curve of DEK1 and DEK2 solar cells under
simulated sunlight irradiation (AM 1.5G, 100 mW cmꢁ2) at different
DCA concentrations. Solid line: no DCA; dashed line: 50 mM; dash/
dotted line: 100 mM.
13786 | J. Mater. Chem., 2011, 21, 13785–13788
This journal is ª The Royal Society of Chemistry 2011