Synthesis of a novel dinuclear ruthenium polypyridine dye for dye-sensitized solar cells application

Article abstract of DOI:10.1016/j.poly.2013.09.023

A new dinuclear ruthenium(II) polypyridine complex has been successfully synthesized. The new compound has been characterized by spectroscopic and electrochemical methods. Its potential application as a sensitizing dye in dye-sensitized solar cells has been checked under AM 1.5 G irradiation conditions (100 mW cm^-2) and its performance was compared to that of a commercially available mononuclear analogous dye. The overall light-to-electricity conversion efficiency of the photovoltaic device sensitized by the new dinuclear dye has been found to be over 2.5 times lower than that sensitized by the commercial analogue, despite a much higher extinction coefficient of the former dye. The probable reasons for the lower photovoltaic activity are discussed.

Full text of DOI:10.1016/j.poly.2013.09.023

Polyhedron 67 (2014) 381–387
Contents lists available at ScienceDirect
Polyhedron
journal homepage: www.elsevier.com/locate/poly
Synthesis of a novel dinuclear ruthenium polypyridine dye
for dye-sensitized solar cells application
a
b
c
a
a
M. Zalas a, , B. Gierczyk , M. Klein , K. Siuzdak , T. Pe˛dzinski , T. Łuczak
⇑
´
a Adam Mickiewicz University in Poznan´, Faculty of Chemistry, Umultowska 89 b, 61-614 Poznan´, Poland
b Faculty of Applied Physics and Mathematics, Gdansk University of Technology, G. Narutowicza 11/12, 80-233 Gdan´sk, Poland
c
´
Polish Academy of Sciences, The Szewalski Institute of Fluid-Flow Machinery, Fiszera 14, 80-231 Gdansk, Poland
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 8 June 2013
Accepted 3 September 2013
Available online 2 October 2013
A new dinuclear ruthenium(II) polypyridine complex has been successfully synthesized. The new com-
pound has been characterized by spectroscopic and electrochemical methods. Its potential application
as a sensitizing dye in dye-sensitized solar cells has been checked under AM 1.5 G irradiation conditions
(100 mW cmꢀ2) and its performance was compared to that of a commercially available mononuclear
analogous dye. The overall light-to-electricity conversion efficiency of the photovoltaic device sensitized
by the new dinuclear dye has been found to be over 2.5 times lower than that sensitized by the commer-
cial analogue, despite a much higher extinction coefficient of the former dye. The probable reasons for the
lower photovoltaic activity are discussed.
Keywords:
Ruthenium dye
Polypyridine complexes
Dye-sensitized solar cells
TiO2 sensitization
Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction
works in this area were published by scientific groups all over the
world. Nowadays, everyday a few new research articles on the sub-
Growing energy demand along with fossil fuel depletion and
the need for global environment protection has made the develop-
ment of non-polluting renewable energy technologies of a highest
priority in science and technology [1]. Solar technologies, espe-
cially photovoltaics, are the most promising ones. The last 35 years
have brought an explosion of popularity of silicone solar cells. The
demand for solar energy has grown rapidly, with growth rates be-
tween 20% and 25% per every year [2]. Today the standard effi-
ciency of a commercial silicone solar cell is between 15% and
20%. However, silicone solar cells still account for less than 0.1%
of the total world energy production. The relatively high produc-
tion costs of silicone photovoltaics and the fact that their produc-
tion process needs ultraclean materials and toxic chemicals are
the main reasons preventing their widespread use [2–4]. A low
cost alternative for silicon solar cells appeared when O’Regan and
Grätzel developed a highly efficient solar cell based on dye-sensi-
tized titania [5]. The photoconversion efficiency of this device
was 7.1–7.9%. The use of commercially available materials and
chemicals without any specific treatment and purification, to-
gether with the relatively high efficiency (at that time) made
dye-sensitized solar cells (DSSCs) an interesting alternative to the
conventional silicon devices. Enormous interest in DSSCs may be
illustrated by the fact that over the 20 years since the first
announcement by O’Regan and Grätzel, more than 10,000 original
ject are published [6].
In short, the DSSC device is made of two conductive glass elec-
trodes [1,3,4,6–10]. The first electrode is a transparent photoanode,
which consists of a tin conducting oxide glass (TCO) substrate
coated with a thin mesoporous layer of a nanocrystalline semicon-
ducting oxide with a wide band gap (mostly TiO2). The surface of
the semiconductor is modified with a monolayer of chemisorbed
dye molecules. The second electrode, the counter electrode, is
made of TCO coated with an ultrathin layer of platiꢀnum. The elec-
trolyte containing redox medium (usually the Iꢀ/I3 couple) is in-
jected between these two electrodes to transfer charges between
them. There are different opinions about which part of the DSSC
device, the semiconductor [7] or the dye [1,11], plays the main role
in the photoelectric process taking place in the cell upon its illumi-
nation. There is no doubt, however, that the whole operating pro-
cess starts when the dye molecule absorbs a photon of visible light.
After this process the electron is excited from the highest occupied
molecular orbital (HOMO) to the lowest unoccupied molecular
orbital (LUMO) of the dye molecule, which leads to generation of
the excited state of the dye. The excited electron is consequently
injected from the LUMO orbitals of the dye to the conduction band
of the semiconductor and in consequence the dye undergoes oxi-
dation. The oxidized dye is regenerated through the electron trans-
ferred from the reduced form of redox mediator present in the
electrolyte and simultaneously the oxidized form of the redox
mediator is formed. At the same time, the photoinjected electron
is transported through the nanoparticle network of the semicon-
⇑
Corresponding author. Tel.: +48 618296754; fax: +48 618291505.
E-mail address: maciej.zalas@amu.edu.pl (M. Zalas).
0277-5387/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.poly.2013.09.023

382
M. Zalas et al. / Polyhedron 67 (2014) 381–387
ductor to reach the TCO substrate. The electrons collected on TCO
are transported through the external circuit, give electric power,
and finally get into the counter electrode, where they regenerate
the reduced form of the redox mediator via reducing the oxidized
form and close the circuit. In an ideal system the whole operation
process takes place without consumption nor permanent transfor-
mation of any chemical species and theoretically it can occur until
illumination is present.
The most important issue of DSSC research is to design effective
sensitizing dyes. The main conditions an effective sensitizer should
meet are as follows [12]:
ꢁ wide range of visible light adsorption spectra;
ꢁ strong anchoring on the semiconductor surface;
ꢁ efficient injection of the photoexcited electrons into the semi-
conductor conduction band;
ꢁ the energy of oxidation potential of excited state being higher
than that of the conduction band edge of the semiconductor;
ꢁ high redox potential of the ground state;
ꢁ long term stability in the solar cell working conditions.
The ruthenium dyes cis-bis(isothiocyanato)bis(2,20-bipyridyl-
4,40-dicarboxylato)-ruthenium(II), its bis(tetrabutylammonium)
salt and triisothiocyanato-(2,20:6,600-terpyridyl-4,40,400-tricarboxy-
lato) ruthenium(II) tris(tetrabutylammonium), widely known as
N3, N719 and ‘‘black dye’’, respectively, meet all the above condi-
tions and have become a benchmark for other sensitizers. The pho-
ton to current conversion efficiencies reached by the devices
sensitized with these three dyes are 10.0%, 11.2% and 10.4%
[13,14], respectively, though these are still not high enough as ex-
pected for commercial application in DSSCs. The mainstream re-
search in ruthenium sensitizers is focused on structure
modification of the ligands to improve light harvesting and electron
injection efficiency and, in consequence, the efficiency of the DSSCs.
The vast majority of work has concerned the modifications of ancil-
lary ligands, which allowed a record efficiency of 12.3% for Z991
dye, which is an N3 analogue with an ancillary ligand substituted
with bithiophene containing moieties [1,15]. Anchoring ligand
modifications are rarely reported in the literature and most of the
works have concerned the substitution of the carboxyl or phosphate
groups with other ones capable of bonding the dye to the semicon-
ductor surface [16]. A different approach to the anchoring ligand
modifications has been presented by Funaki et al. [17], who have
Scheme 1. The structures of the B1 (a) and 455PF6 (b) dyes.
dark. Then the solvent was evaporated in vacuo. The obtained com-
plex was dissolved in dry methanol (10 mL) and a concentrated
solution of ammonium hexafluorophosphate (2 g in 4 mL of water)
was added. The obtained dark red precipitate was filtered off,
washed with water, methanol and diethyl ether and dried in vac-
uum to obtain 0.73 g of dye B1 as a red amorphous powder
(Scheme 2). Yield 81%. Anal. Calc. for C71H50N12O2F24P4Ru2: C,
45.23; H, 2.67; N 8.92. Found: C, 45.41; H, 2.76; N, 8.78%. ESI-MS
m/z: 326 [Mꢀ4PF6]4+
.
1H NMR ([2H]3-acetonitrile) d: 7.41 (m, 8H;
H-19); 7.43 (ddd, 2H; 7.4, 5.6, 1.3 Hz; H-16); 7.47 (dd, 2H; 5.9,
1.8 Hz; H-9); 7.72 (m; 6H; H-18); 7.75 (ddd, 2H; 5.6, 1.5, 0.8 Hz;
H-17); 7.76 (dd, 2H; 5.9, 0.8 Hz; H-10); 7.80 (ddd, 2H; 5.6, 1.5,
0.8 Hz; H-18); 8.03 (t, 1H; 1.6 Hz; H-5); 8.07 (m, 8H; H-20); 8.09
(ddd, 2H; 8.3, 7.4, 1.5 Hz; H-15); 8.31 (d, 2H; 1.6 Hz; H-3); 8.51
(m, 8H; H-21); 8.54 (ddd, 2H; 8.3, 1.3, 0.8 Hz; H-14); 8.66 (dd,
2H; 1.8, 0.7 Hz; H-11); 11.0 (br, 1H; H-1). 13C NMR ([2H]3-acetoni-
trile) d: 87.91 (C-6); 96.12 (C-7); 123.61 (C-4); 125.39 (C-21);
125.55 (C-14); 127.13 (C-11); 128.68, 128.73 (C-19); 129.01 (C-
9); 129.75 (C-16); 132.60 (C-8); 133.19 (C-2); 135.17 (C-3);
138.99 (C-15); 139.03 (C-20); 139.14 (C-5); 152.67 (C-17);
152.78 (C-10); 152.84, 152.86, 152.90 (C-18); 157.57 (C-13);
157.90, 157.99, 158.00 (C-22); 158.52 (C-12); 166.52 (C-1).
synthesized a ‘‘black dye’’ analogue with a
p expanded ligand hav-
ing a phenylene-ethynylene moiety. The modification led to an in-
crease in the molar extinction coefficient, but unfortunately did
not translate into better conversion efficiencies in the DSSCs.
In this work we present a new dinuclear dendritic-like ruthe-
nium dye which consist of two trisbipyridyl ruthenium(II) deriva-
tives clipped with an ethyl 3,5-diethynylbenzoate group
(Scheme 1). Full spectroscopic and electrochemical characteriza-
tion of the new compound is presented and the effects of DSSC sen-
sitization in comparison with that achieved with a commercially
available mononuclear analogue (Ruthenizer 455PF6, Solaronix;
Scheme 1) are discussed.
2.2. Electrochemical and photochemical studies
Cyclic voltammograms were collected using an Autolab electro-
chemical analyzer (Eco Chemie, B. V., Utrecht, The Netherlands).
The working electrode was a Pt wire, the counter electrode was a
Pt plate and the reference electrode was a saturated calomel elec-
trode (SCE). 10ꢀ3 M solutions of the dyes in acetonitrile (Aldrich)
with 0.1 M LiClO4 (Fluka) as a supporting electrolyte were used.
The experiments were carried out in an electrochemical cell under
a pure Ar gas flow. UV–Vis absorbance spectra were measured on a
Lambda 35 UV–Vis spectrometer (Perkin Elmer, Waltham MA,
USA) and the amounts of adsorbed dye were determined on Cary
50 Probe UV–Vis spectrometer (Varian Inc., Palo Alto CA, USA).
Photoluminescence spectra were recorded on a LS 55 fluorescence
spectrometer (Perkin Elmer, Waltham MA, USA). In both cases, the
2. Experimental
2.1. Synthesis
0.500 g of Ru(bpy)2Cl2 (0.96 mmol, as a dihydrate) and 0.230 g
(0.48 mmol) of dinucleating ligand 1 (the synthetic procedure for
ligand 1 has been published elsewhere [18] and for convenience
also attached to this paper as Supplementary data) were dissolved
in ethanol (200 mL) and refluxed under nitrogen over 10 h in the

M. Zalas et al. / Polyhedron 67 (2014) 381–387
383
concentration of the dyes was 1.5 ꢂ 10ꢀ5 M in ethanol. The dyes
nol to a volume of 10 mL) for 10 min. Concentration of the des-
orbed dye in the solution left was determined by a UV–Vis
technique.
were excited at the wavelengths of the maximum absorption. Fluo-
rescence lifetime measurements were carried out with 3 ꢂ 10ꢀ6
M
ethanol dyes solutions on a time-resolved photon counting spec-
trofluorometer (FluoTime 300, PicoQuant, Berlin, Germany) using
a 440 nm pulsed diode laser (PicoQuant, LDH440) with a repetition
rate set automatically by the instrument to allow a 5-times longer
time window compared to the measured lifetime. The detected
light was spectrally selected using a monochromator and detected
using a cooled photomultiplier tube. The instrument response
function (IRF) had a full-width at half-maximum of 200 ps. The
fluorescence decay curves were analyzed using the FluoFit soft-
ware (PicoQuant, version 4.5) based on either a mono- or double-
exponential model with IRF or simply by a mono-exponential tail
fitting.
The photovoltaic characteristics of the cells were measured
using a Sun 2000 class A Solar Simulator (Abet Technologies, Mil-
ford CT, USA) equipped with an AM 1.5 G filter, with the light
intensity adjusted to 100 mW cmꢀ2 using a silicon reference cell
(ReRa Systems, Wijchen, The Netherlands). J–V curves were re-
corded on a Keithley 2400 SourceMeter (Keithley, Cleveland OH,
USA). The incident photon to current conversion efficiency (IPCE)
was measured using a set consisting of a light source – 450 W xe-
non lamp (Osram, Munich, Germany), a prismatic monochromator
SPM-2 (Zeiss, Jena, Germany), silicon reference photodiode S1336-
18BQ (Hamamatsu Photonics, Hamamatsu, Japan) and a picoam-
meter 485 (Keithley, Cleveland OH, USA). The short circuit current
was measured using a picoammeter 485 (Keithley, Cleveland OH,
USA). During the measurements the cell was illuminated by a stea-
dy stream of photons of 1014 cmꢀ2 sꢀ1 intensity. The measure-
ments were performed for the wavelength range 400–800 nm.
IPCE was calculated according to the following equation:
2.3. Photovoltaic performance
All the chemicals used were of analytical grade and were used
as received without any additional purification. The procedure
used for preparation of the titania electrodes was similar to those
described elsewhere [19] and was as follows: 3 mL titanium tetra-
isopropoxide (Aldrich) were added to 13.5 mL of ethylene glycol
(Chempur) magnetically stirred at 333 K. The mixture, after addi-
tion of 12.6 g of citric acid monohydrate (POCh), was heated under
stirring at 363 K until clear. The transparent sol obtained was
mixed with 5.6 g P25 TiO2 (Degussa) by grinding in an agate mor-
tar for 1 h. The viscous titania paste obtained was spread on a fluo-
rine doped tin oxide (FTO) conductive glass substrate (Solaronix)
using the ‘‘doctor blading’’ technique and sintered in air at 723 K
for 1 h. To prepare the working electrodes for DSSCs, titania elec-
trodes were immersed in a 1 ꢂ 10ꢀ4 M solution of B1 dye or Ruth-
enizer 455PF6 (Solaronix) in absolute ethanol at 5 °C overnight.
After dye adsorption, the electrodes were washed with absolute
ethanol and dried in a hot air stream. A platinum film coated
FTO was used as a counter electrode. A typical cell was assembled
1240ðeV ꢃ nmÞ ꢃ JscmðmA=cm2Þ
IPCE ¼
;
kðnmÞ ꢃ
U
ðmW=cm2Þ
where Jscm is the short circuit photocurrent density for the mono-
chromatic irradiation, k is the wavelength and is the monochro-
U
matic light intensity. Electrochemical impedance spectroscopy
measurements were carried out using a 150 W Optel 57–003 light
source (Optel, Opole, Poland), potentiostat–galvanostat Solartron
1287 (Solartron, Torrance CA, USA) equipped with an impedance
analyzer Solartron 1260 (Solartron, Torrance CA, USA). The imped-
ance spectra were taken at frequencies ranging from f = 1 MHz to
0.1 Hz with a Vac = 5 mV amplitude of the AC signal at open circuit
voltage. During spectra collection, the cells were under continuous
light illumination (AM 1.5, 100 mW cmꢀ2). The impedance data
were analyzed on the basis of an electric equivalent circuit (EE
QC) using an EIS Spectrum Analyser [20]. The modified Powell algo-
rithm [21] was used with amplitude weighting:
using a 25 lm thick, hot-melted, ionomeric foil (Solaronix) as a
sealant and a spacer between the electrodes and an electrolyte (a
mixture of 0.6 M 1-propyl-3-methyl-imidazolium iodide (Aldrich),
0.03 M iodine (POCh), 0.1 M guanidine thiocyanate (Fluka) and
0.5 M 4-tert-butylpiridine (Aldrich) in acetonitrile (Merck)) were
injected within two holes predrilled in the counter electrode. The
final sealing was realized with the use of hot melted sealant and
a microscope cover slide. The typical active area of the obtained
DSSC was approximately 0.125 cm2. Each type of DSSC prepared
was reproduced 5 times and photoelectric measurements were re-
peated 5 times for every cell prepared, so that each final result of
photovoltaic performance presented is an average of 25 photoelec-
tric experiments. To determine the amount of dye absorbed, the
sensitized electrodes were immersed in 5 mL of NH4OH solution
(obtained by diluting 1 mL of 25% NH4OH in water with 95% etha-
rað
x
; P1 . . . PMÞ ¼ r2c =ðN ꢀ MÞ
where N is the number of points, M is the number of parameters,
x
is the angular frequency and P1. . .PM are parameters. Parameter rc is
defined as
2
2
0
0
ðZi ꢀ Zi Þ þ ðZ0i0 ꢀ Zi00
Þ
N
X
rc2
¼
calc
calc
Z0i2 þ Z02
i¼1
icalc
where i corresponds to the measured values of impedance and icalc
is attributed to the calculated values; N is the number of points.
Scheme 2. Synthesis route for the B1 dye.

384
M. Zalas et al. / Polyhedron 67 (2014) 381–387
2.4. Other analytical methods
signals probably come from moieties C and F, whose electronic
environments are very similar.
The NMR spectra were recorded at 298 K on an Agilent 800 MHz
spectrometer (Agilent Technologies, Santa Clara CA, USA),
equipped with an 1H{13C/15N} triple resonance probe. The reso-
nance frequencies were 799.926 and 201.162 MHz for 1H and
13C, respectively. Signal assignments were made on the basis of
homonuclear COSY and NOESY experiments as well as heteronu-
clear 1H–13C HSQC and HMBC spectra. Standard acquisition param-
eters were used for recording the 1D and 2D spectra. The mixing
time for the NOESY spectra was set to 0.5 s. Elementary analyses
were obtained on a Vario EL III (Elementar Analysensysteme
GmbH, Hanau, Germany) analyser. ESI Mass spectra were recorded
on a Waters/Micromass Q-tof Premier (ESI-MS) mass spectrometer
(Waters Ltd., Eltree, UK).
3.2. Electrochemical, absorption and emission properties
The electrochemical characteristics of B1 and 455PF6 were
investigated by cyclic voltammetry and the stable cyclic voltam-
mograms (the j–E curves) recorded after the 3rd potential scans
are presented in Fig. 2. Three quasi-reversible redox peaks are ob-
served at half wave potentials E1/2 = 0.76, 0.53 and 0.28 V versus
SCE on the j–E curve for B1. Moreover a couple of reversible redox
wide waves with peaks at E1/2 = 0.72 V and one peak at 0.40 V ver-
sus SCE in the reduction scan are observed in the cyclic voltammo-
gram recorded for 455PF6. The two redox peaks observed for B1 at
E1/2 = 0.76 and 0.53 V correspond to the two step redox process:
B1+4MB1+5MB1+6, connected with the oxidation and reduction of
both ruthenium atoms. There is no doubt that the third peak at
E1/2 = 0.28 V corresponds to the redox processes involving the mul-
tiple bonds of ligand 1 [12,16]. The reversible peak observed for
455PF6 corresponds to the one step redox process of the RuII/RuIII
couple and the reduction peak is probably due to the irreversible
reduction of the 455PF6 ligands [22]. The redox potentials of both
Ru atoms of B1 are more positive than that of the iodine electron
donor, providing sufficient impellent for efficient dye regeneration
3. Results and discussion
3.1. NMR characterization
2+
In the Ru(bpy)3 ion the bipyridine molecules are equivalent.
As a consequence, all six pyridine moieties are also equivalent in
this cation. Monosubstitution of Ru(bpy)32+, e.g. at C-4 of one pyr-
idine group, results in dissymetrisation of this ion. Such a unit con-
tains six inequivalent pyridyl groups (A–F; See Fig. 1), therefore up
to 30 signals in the 13C NMR spectrum and 23 signals in the 1H
ꢀ
by the Iꢀ/I3 electrolyte operating in the DSSC [17]. The electro-
chemical measurements of 455PF6 were performed in various sol-
vents (such as DMF or acetonitrile) and supporting electrolytes
(TBAPF6 or LiClO4), but no better intensity of the peaks could be ob-
tained. One can underline that the CV curves presented in this pa-
per are fully repeatable.
2+
spectrum are expected. Moreover, the monosubstituted Ru(bpy)3
unit is chiral. As the B1 molecule contains two Ru-containing moi-
eties, the dye may exist as a mixture of two diastereoisomers,
which may results in increasing the number of signals in the
NMR spectra. However, as only one set of phenyl ring and ethyne
carbon signals has been observed in the 13C NMR spectrum of B1,
one could deduce that only one of the expected diastereoisomers
forms during the synthesis. The signals of the hydrogen and carbon
atoms of the substituted bipyridine ligand (A, B) are well separated
from the other ones. Unfortunately, the assignment of the two lat-
ter bpy units is not possible due to overlapping of the signals. The
H-19, H-20 and H-21 protons give three signals of 8H intensity. For
H-18, two signals are observed, one has the intensity 2H and the
other 6H. The NOE measurements indicate that the separate H-
18 signal (7.80 ppm) corresponds to the pyridine moiety D in the
trans position to the 4-substituted one (A), but the assignment is
uncertain. In the 13C NMR spectrum all C-21 carbon atoms give
one slightly asymmetric but unresolved peak. The same situation
is observed for the C-20 carbon atom. The signals of the other car-
bon atoms are more sensitive to changes of the electron density
and are better resolved, but the chemical shift differences are small
(less than 0.1 ppm). For C-19 three signals are observed, one of
them (128.73 ppm) has a double intensity and contains two over-
lapping resonance lines. Similarly, the C-18 and C-22 carbons give
sets of two separated and two unresolved signals. The overlapping
The absorption spectra of 1.5 ꢂ 10ꢀ5 ethanolic solutions of B1
and 455PF6 are shown in Fig. 3. The absorbance of B1 is signifi-
cantly higher than 455PF6 in the whole spectrum, which indicates
that B1 should show more efficient light harvesting ability. The
absorption bands between 200 and 350 nm, corresponding to the
ligands
p–
p⁄ electron transfer, are observed for both dyes, but in
the spectra of B1 an additional strong band is present at 325 nm,
assigned to the more extended delocalized bond structure of the
B1 anchoring ligand, playing the role of an antenna. The broad
bands with maxima at 453 and 460 nm for 455PF6 and B1, respec-
tively, correspond to the metal-to-ligand charge transfer (MLCT)
transition, characteristic of ruthenium polypyridine complexes
[11,16]. The molar extinction coefficients of the MLTC bands are
94174 and 37380 Mꢀ1 cmꢀ1 for B1 and 455PF6, respectively, so
Fig. 2. Cyclic voltammograms of the B1 and 455PF6 dyes compared with that of the
Fig. 1. Labelling of pyridine units in substituted Ru(bpy)3.
pure 0.1 M supporting electrolyte (LiClO4) solution (SES). dE/dt = 0.1 Vsꢀ1
.

M. Zalas et al. / Polyhedron 67 (2014) 381–387
385
of the excited electron over the bulky anchoring ligand with a
p-
expanded bond structure.
3.3. Photovoltaic performance
The incident photon to current conversion efficiencies (IPCE)
spectra, plotted as a function of excitation wavelength, for the pho-
tovoltaic devices sensitized with B1 and 455PF6 dyes are shown in
Fig. 5. The sensitization range corresponds to the visible light
absorption spectra of both investigated dyes. The maximum IPCE
values were obtained at 400 nm (3.6%) and 450 nm (23.7%) for
B1 and 455PF6, respectively. The IPCE value of the B1 sensitized
device at 460 nm, which is a MLTC band of B1, was equal to 2.8%.
The photovoltage–photocurrent characteristic curves of B1 and
455PF6 sensitized cells are presented in Fig. 6 and their typical
parameters are collected in Table 1. A comparison of B1 and
455PF6 cell parameters shows that the open circuit photovoltage
(Voc) is slightly lower for the B1 cell, the short circuit current (Jsc)
is much lower for the B1 cell and the fill factor (FF) is higher for
the B1 cell. The overall photon-to-current conversion efficiencies
are 0.32% and 0.72% for B1 and 455PF6, respectively. The general
lesser photovoltaic activity of the B1 sensitized cell is in good
agreement with the IPCE measurements described above. The
dye loading parameter (Ndye), which has a significant influence
on the cell performance [6,25], determined for both investigated
cells (see Table 1) shows that the adsorption of B1 on the TiO2 elec-
trode is less efficient than that of 455PF6. This phenomena may be
related to the bulky molecule of B1 and/or the presence of only one
anchoring –COOH group in the B1 molecule, which may decrease
the binding abilities of this dye. The lower Ndye value is definitely
not the only reason for the poorer performance of the B1 cell, espe-
cially in view of the light absorption abilities of these dyes, which
are much better for B1. An additional effect leading to the low
activity of the B1 dye sensitized cell may be easier aggregation of
these dye molecules on the electrode surface [17]. Moreover, the
meta position of the alkyne groups of the B1 molecule, relative
to the anchoring carboxylic moiety, may lead to poor electronic
communication with the TiO2 orbitals. Sensitizers based on such
a substitution system have also been found to be less efficient at
converting visible photons into electrons than para substituted
ones (like 455PF6) because of the non-radiative decay of excited
states [26–28]. To get a deeper insight into the reasons for the
poorer performance of the cell sensitized with B1, both types of
cells were subjected to electrochemical impedance spectroscopy
studies, as this technique is regarded as a diagnostic tool to
Fig. 3. Absorption and emission (insert) spectra of the B1 and 455PF6 dyes.
the former for the dinuclear complex is 2.5 times higher than that
for the mononuclear one. The MLCT band of B1 is slightly red
shifted in relation to that in the spectrum of 455PF6. The insert
of Fig. 3 shows the emission spectra recorded at excitation wave-
lengths corresponding to the MLCT maxima of the dyes investi-
gated. The emission maxima were obtained at 629.5 and
609.5 nm and the calculated Stokes shifts were 169.5 and
156.5 nm for B1 and 455PF6, respectively. The red-shifted emis-
sion of B1 may be due to stronger p-acceptor properties of the li-
gand 1 and, in consequence, a lower energy of its p⁄ orbital [12].
The fluorescence decay traces for B1 and 455PF6 are shown in
Fig. 4. The fitting was performed using a simple one-exponential
tail fitting method as the most appropriate one to measure the
long-lived component. However, the generally more accurate
deconvolution method with IRF (instrument response function)
was also used, and the short component with the lifetime was
measured in addition to the long-lived component of 360 and
289 ns for B1 and 455PF6, respectively [23]. These fast dynamics
may be interpreted as an intersystem crossing (ISC) from the sin-
glet state to the lower energy triplet state of the excited dye mol-
ecule, while the longer lifetime is most likely due to the triplet
lifetime decay of the complex [24]. The longer fluorescence lifetime
observed for B1 is probably caused by more effective delocalization
Fig. 4. Time-resolved emission decays for the B1 (a), and 455PF6 (b) dyes measured
in 10ꢀ6 M ethanolic solutions. Excitation wavelength = 440 nm, emission monitored
at 650 nm.
Fig. 5. Photocurrent action spectra of photovoltaic devices sensitized with the B1
and 455PF6 dyes.

386
M. Zalas et al. / Polyhedron 67 (2014) 381–387
Two resistive elements, R2 and R3, are in parallel with a modi-
fied capacitive component – a constant phase element (CPE)
describing the capacitance of each interface [34]. This element re-
places a typical capacitor because the electrodes used do not have
an ideal geometry and the electrode/electrolyte interface cannot be
modelled as an ideal capacitor. The impedance of the CPE is char-
acterised by the impedance Z = Qꢀ1(i ꢀn, where
x) x is the angular
frequency, Q is the CPE parameter and n is the exponential. The n
values for CPE elements vary from 0.87 to 0.98.
The ohmic serial resistance element R1 in the high frequency re-
gime represents the conductive FTO electrode resistance and takes
similar values irrespective of the dye used. The resistance R2 ob-
served in the high-frequency region is associated with the redox
charge-transfer process at the Pt electrode of the solar cell [35].
The R2 value calculated for the DSSC with the B1 dye is in the range
4–12
X, which is regarded as the optimum range for DSSC work
[30]. The R2 value obtained for the DSSC with the 455PF6 dye is
a little higher. It could be caused by a thinner Pt film [36] or discon-
tinuities of the platinum layer [33]. The resistance R3 describes the
TiO2, dye and electrolyte interface and the Warburg element W
corresponds to Nernstian diffusion inside the electrolyte at low-
frequencies [37,38]. As shown in Table 2, a significant difference
between the dye sensitized solar cells was found in the R3 resis-
tance element. The radius of the semicircle (R3) corresponding to
the photoelectrode is much greater for the DSSC with the B1 dye
than that with the 455PF6 dye. This indicates an increase in the
electron recombination resistance when the B1 dye covers porous
titania [39]. This phenomenon could be responsible for the worse
performance, i.e. the photocurrent-generation of FTO/TiO2/B1 is
less efficient, so the short circuit current is lower and thus lower
photoconversion efficiency of the B1 sensitized solar cell is ob-
served. We speculate that the longer distance from the metallic
centers of the dye molecule, at which the electron excitation by
an incident photon take place, to the semiconducting electrode
surface can be a factor contributing to a more effective recombina-
tion process within the dye molecule.
Fig. 6. Photocurrent density-photovoltage characteristic curves of the solar cells
sensitized with the B1 and 455PF6 dyes.
Table 1
Photovoltaic performance of the B1 and 455PF6 dye-sensitized TiO2 cells.
Dye
Jsc (mA cmꢀ2
)
Voc (mV) FF (%)
g )
(%) Ndye (10ꢀ7 mol cmꢀ2
B1
1.01
520
556
60.5
51.4
0.32
0.72
3.31
5.69
455PF6 2.53
The electron lifetime
xmax from the equation
lue for the DSSC with the B1 dye was 6 ms, whereas
s
s
e was calculated using the peak frequency
e = 1/xmax [30]. The electron lifetime va-
e for the cell
s
with the 455PF6 dye was equal to 11 ms. This means that the
application of the B1 dye remarkably decreases the electron life-
time in titania before recombination.
4. Conclusions
The dye under study was synthesized with high purity and yield
using a simple procedure. The high electrochemical stability and
the redox potential values of the processes occurring in the B1
molecule, as well as a very high extinction coefficient, indicate that
the new complex should be a good candidate as a sensitizing dye in
DSSC. Unfortunately the IPCE measurements show a low photon-
to-current conversion ability of B1 and, as a consequence, the de-
vices sensitized with B1 show poor photovoltaic parameters when
compared to those of the commercially available mononuclear
analogue. The impedance spectroscopy measurements show that
a more effective electron recombination process takes place on
the photoanode sensitized with B1, moreover the electron lifetime
in the porous electrode decreases when compared with that for the
Fig. 7. Impedance spectra for DSSC with the B1 and 455PF6 dyes.
estimate electron recombination resistance and dye regeneration
efficiency [29]. Fig. 7 presents the impedance spectra of the cells
sensitized with the B1 and 455PF6 dyes. For analysis, an electrical
equivalent circuit R1(R2CPE1)(R3CPE2)W was used, considering each
interface as a parallel RCPE circuit [30]. Similar EEQC has been pro-
posed by Yoon et al. [31,32] and Jiang et al. [33]. The fitting proce-
dure gives normalized fitting errors in the optimal solution r2c on
the level of 10ꢀ5. The values of each parameter are listed in the
Table 2.
Table 2
The calculated values of electric equivalent circuit elements for DSSCs with the B1 and 455PF6 dyes.
Dye
R1
(X
)
R2
(X)
CPE1 (10ꢀ5
X
ꢀ1 cmꢀ2 sn)
R3
(X)
CPE2 (10ꢀ5
X
ꢀ1 cmꢀ2 sn)
W (X )
sꢀ1
B1
455PF6
12.09
9.74
8.09
13.50
1.28
2.28
1120.00
862.32
3.76
6.85
2.23
2.11

M. Zalas et al. / Polyhedron 67 (2014) 381–387
387
[12] M.K. Nazeeruddin, S.M. Zakeeruddin, J.J. Lagref, P. Liska, P. Comte, C. Barolo, G.
Viscardi, K. Schenk, M. Graetzel, Coord. Chem. Rev. 248 (2004) 1317.
[13] M.K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker, E. Mueller, P. Liska, N.
Vlachopoulos, M. Graetzel, J. Am. Chem. Soc. 115 (1993) 6382.
[14] M.K. Nazeeruddin, P. Pechy, M. Gratzel, Chem. Commun. (1997) 1705.
[15] M. Grätzel, Acc. Chem. Res. 42 (2009) 1788.
[16] U. Veikko, X. Zhang, T. Peng, P. Cai, G. Cheng, Spectrochim. Acta, Part A 105
(2013) 539.
[17] T. Funaki, M. Yanagida, N. Onozawa-Komatsuzaki, Y. Kawanishi, K. Kasuga, H.
Sugihara, Sol. Energy Mater. Sol. Cells 93 (2009) 729.
photoanode sensitized with the 455PF6 dye. The significantly low-
er adsorption of B1 on the photoanode surface and the potential
formation of dye molecule aggregates have been found as addi-
tional factors for the insufficient photovoltaic efficiency of the
DSSCs investigated. However, these preliminary results indicate
that dinuclear ruthenium dyes may potentially find application
in DSSCs after some cell optimization for better dye loading on
the photoelectrode and structure modifications of the dye mole-
cule to improve electron injection and prevent the recombination
process.
[18] M. Zalas, B. Gierczyk, M. Cegłowski, G. Schroeder, Chem. Pap. 66 (2012) 733.
[19] U. Opara Krasovec, M. Berginc, M. Hocevar, M. Topic, Sol. Energy Mater. Sol.
Cells 93 (2009) 379.
[20] A.S.R. Bondarenko, G.A. Ragoisha, in: A.L. Pomerantsev (Ed.), Progress in
Chemometrics Research, Nova Science Publishers, New York, 2005, p. 89.
[21] L. Lu, B.H. Brown, D.C. Barber, A.D. Leathard, Physiol. Meas. 16 (1995) A39.
[22] G. Wolfbauer, A.M. Bond, G.B. Deacon, D.R. MacFarlane, L. Spiccia, J. Am. Chem.
Soc. 122 (1999) 130.
Acknowledgment
This work was supported by the Polish National Science Centre,
Grant No. N N204 023538.
[23] D. Phillips, R.C. Drake, D.V. O’Connor, R.L. Christensen, Anal. Instrum. 14 (1985)
267.
[24] S.E. Koops, P.R.F. Barnes, B.C. O’Regan, J.R. Durrant, J. Phys. Chem. C 114 (2010)
8054.
[25] M. Zalas, M. Klein, Int. J. Photoen., (2012) Art. ID 927407.
[26] Thomas J. Meyer, Gerald J. Meyer, Brian W. Pfennig, Jon R. Schoonover, Cliff J.
Timpson, Jennifer F. Wall, Claus Kobusch, Xiaohong Chen, Brian M. Peek, Craig
G. Wall, Ou Wei, Bruce W. Erickson, Carlo A. Bignozzit, Inorg. Chem. 33 (1994)
3952.
[27] Y. Hou, P. Xie, B. Zhang, Y. Cao, X. Xiao, W. Wang, Inorg. Chem. 38 (1999) 6320.
[28] R. Argazzi, N.Y. Murakami Iha, H. Zabri, F. Odobel, C.A. Bignozzi, Coord. Chem.
Rev. 248 (2004) 1299.
[29] K. Kalyanasundaram, Dye-sensitized Solar Cells, EFPL Press, Lausanne,
Switzerland, 2010.
[30] K.B. Fadadu, S.S. Soni, Electrochim. Acta 88 (2013) 270.
[31] J. Yoon, M. Jin, M. Lee, Adv. Mater. 23 (2011) 3974.
[32] S.J. Yoon, M. Lee, E.-K. Kim, W. Lee, Y.-C. Nah, N.K. Shrestha, S.-H. Lee, S.-H. Han,
New J. Chem. 35 (2011) 2521.
[33] W. Jiang, L. Yin, H. Liu, Y. Ding, J. Power Sources 218 (2012) 405.
[34] Q. Wang, J.-E. Moser, M. Grätzel, J. Phys. Chem. B 109 (2005) 14945.
[35] S. Cai, X. Hu, J. Han, Z. Zhang, X. Li, C. Wang, J. Su, Tetrahedron 69 (2013) 1970.
[36] M. Adachi, M. Sakamoto, J. Jiu, Y. Ogata, S. Isoda, J. Phys. Chem. B 110 (2006)
13872.
[37] K. Kalyanasundaram, M. Grätzel, Coord. Chem. Rev. 177 (1998) 347.
[38] J. van de Lagemaat, N.G. Park, A.J. Frank, J. Phys. Chem. B 104 (2000) 2044.
[39] A.S. Hart, K.C. Chandra Bikram, Navaneetha K. Subbaiyan, P.A. Karr, F. D’Souza,
ACS Appl. Mater. Interfaces 4 (2012) 5813.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.poly.2013.09.023.
References
[1] G.C. Vougioukalakis, A.I. Philippopoulos, T. Stergiopoulos, P. Falaras, Coord.
Chem. Rev. 255 (2011) 2602.
[2] H. Spanggaard, F.C. Krebs, Sol. Energy Mater. Sol. Cells 83 (2004) 125.
[3] M. Gratzel, Nature 414 (2001) 338.
[4] B. Li, L. Wang, B. Kang, P. Wang, Y. Qiu, Sol. Energy Mater. Sol. Cells 90 (2006)
549.
[5] B. O’Regan, M. Gratzel, Nature 353 (1991) 737.
[6] A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo, H. Pettersson, Chem. Rev. 110 (2010)
6595.
[7] A. Hagfeldt, M. Grätzel, Acc. Chem. Res. 33 (2000) 269.
[8] S. Günes, N.S. Sariciftci, Inorg. Chim. Acta 361 (2008) 581.
[9] M. Grätzel, J. Photochem. Photobiol., C 4 (2003) 145.
[10] M. Grätzel, J. Sol–Gel Sci. Technol. 22 (2001) 7.
[11] A.S. Polo, M.K. Itokazu, N.Y. Murakami Iha, Coord. Chem. Rev. 248 (2004) 1343.