Synthesis and characterization of cross-linkable ruthenium dye with ion coordinating property for dye-sensitized solar cells

Article abstract of DOI:10.1016/j.polymer.2011.05.035

Crosslinkable ruthenium complex dye, Ru(2,2′-bipyridine-4,4′- bicarboxylic acid)(4,4′-bis((4-vinyl benzyloxy)methyl)-2,2′- bipyridine)(NCS)₂ (denoted as Ru-S dye), was synthesized and characterized using ¹H-NMR, ¹³C-NMR, Fourier transform infrared (FTIR) and UV/vis spectroscopies.The power conversion efficiency of dye-sentitized solar cell (DSSC) using Ru-S and liquid electrolyte containing lithium iodide (LiI) reached 7.53% under standard global AM 1.5 full sunlight, which is partly attributed to Li⁺ being coordinated by Ru-S as verified by ATR-FTIR spectroscopy. As Ru-S was further crosslinked with glycerol propoxylate triacrylate (GPTA), not only 89% of dye retained on TiO₂ mesoporous surface after rinsed by 0.1 N NaOH aqueous solution, the power efficiency was also increased to 7.88%. As poly(methyl acrylate) was used to gel the electrolyte system, the power efficiency of DSSC with Ru-S dye was 6.96% but increased to 7.57% after crosslinking with GPTA. Notably, both DSSCs showed a good long-term stability after one month storage.

Full text of DOI:10.1016/j.polymer.2011.05.035

Polymer 52 (2011) 3318e3324
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Synthesis and characterization of cross-linkable ruthenium dye with ion
coordinating property for dye-sensitized solar cells
,b,
Ken-Yen Liu a, Chih-Yang Ko b, Kuo-Chuan Ho b, King-Fu Lin a
*
a Department of Materials Science and Engineering, National Taiwan University, Taipei, Taiwan
b Institute of Polymer Science and Engineering, National Taiwan University, Taipei, Taiwan
a r t i c l e i n f o
a b s t r a c t
Crosslinkable ruthenium complex dye, Ru(2,20-bipyridine-4,40-bicarboxylic acid)(4,40-bis((4-vinyl ben-
zyloxy)methyl)-2,20-bipyridine)(NCS)2 (denoted as Ru-S dye), was synthesized and characterized using
1H-NMR, 13C-NMR, Fourier transform infrared (FTIR) and UV/vis spectroscopies.The power conversion
efficiency of dye-sentitized solar cell (DSSC) using Ru-S and liquid electrolyte containing lithium iodide
(LiI) reached 7.53% under standard global AM 1.5 full sunlight, which is partly attributed to Liþ being
coordinated by Ru-S as verified by ATR-FTIR spectroscopy. As Ru-S was further crosslinked with glycerol
propoxylate triacrylate (GPTA), not only 89% of dye retained on TiO2 mesoporous surface after rinsed by
0.1 N NaOH aqueous solution, the power efficiency was also increased to 7.88%. As poly(methyl acrylate)
was used to gel the electrolyte system, the power efficiency of DSSC with Ru-S dye was 6.96% but
increased to 7.57% after crosslinking with GPTA. Notably, both DSSCs showed a good long-term stability
after one month storage.
Article history:
Received 3 April 2011
Received in revised form
14 May 2011
Accepted 21 May 2011
Available online 6 June 2011
Keywords:
Ruthenium dye
Dye-sensitized solar cells
Crosslink
Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction
measurement of dye content on the TiO2 mesoporous layer.Besides,
because the aliphatic side chains are too long, they tend to retard
Mesoscopic dye sensitized solar cells (DSSC) reported in 1991 by
O’Regan and Grätzel [1] have attracted intensive interest due to
their flexibility of device design, relatively high conversion effi-
ciency and inexpensive fabrication processes compared to the
amorphous silicon solar cells. To improve the photovoltaic perfor-
mance of the DSSC, great effort has been paid to the development of
sensitizers, electrolytes, and photoanode materials [2e7]. Various
structural modifications to the ruthenium dye have been attemp-
ted, such as tethering additional conjugate moieties on bipyridyl
ligand for increasing the light absorption and tethering hydro-
phobic aliphatic long chains on bipyridyl ligand for prolonging the
service life [8e14]. Recently, to overcome the desorption problem
of ruthenium dyes from TiO2 mesoporous surface, a crosslinkable
ruthenium dye with aliphatic side chains carrying allyl end groups
on one of the bipyridyl ligands has been synthesized [15]. However,
owing to the low reactivity of allyl groups, as it was crosslinked
with methyl acrylic acid (MAA), the dye only retained 77% on the
TiO2 mesoporous surface after immersion in 0.1 N NaOH solution,
which is typically employed to remove the ruthenium dye for the
the dye adsorption on TiO2 and hinder the dye regeneration process
upon excitation by sunlight. The power conversion efficiency of
DSSC using Ru-C dye is only 5.18% [15].
It has been frequently reported that the presence of lithium
cations (Liþ) in the liquid electrolyte is capable of increasing the
photocurrent of DSSCs [16e20]. However, because the surface of
TiO2 mesoporous film is negatively charged during operation of
DSSC, Liþ may intercalate or adsorb onto the surface of TiO2 layer,
resulting in a positive shift of the TiO2 conduction band and
consequently causing a dramatic drop of open-circuit voltage for
the cell [21]. To refraining Liþ from the TiO2 surface adsorption, the
ruthenium sensitizers were often designed with the ethylene oxide
groups tethering on one of their bipyridyl ligands to coordinate Liþ
[22e24]. For example, the ruthenium dye with a bipyridine ligand
tethering two triethyleneoxide methyl ether groups, denoted as
K51 [21]. However, because the triethyleneoxide methyl ether
groups are hydrophilic, K51 dye is highly soluble in polar solvent.
In this contribution, we designed a novel ion-coordinating
crosslinkable ruthenium dye system, which contains an ion-
coordinating crosslinkable Ru(2,20-bipyridine- 4,40-bicarboxylic
acid)(4,40-bis((4-vinyl benzyloxy)methyl)-2,20- bipyridine)(NCS)2,
denoted as Ru-S and glycerol propoxylate triacrylate (GPTA), an
ion-coordinating crosslinker with the chemical structure shown
below
* Corresponding author. Department of Materials Science and Engineering,
National Taiwan University, Taipei, Taiwan. Tel.: þ886 2 3366 1315; fax: þ886 2
2363 4562.
E-mail address: kflin@ccms.ntu.edu.tw (K.-F. Lin).
0032-3861/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymer.2011.05.035

K.-Y. Liu et al. / Polymer 52 (2011) 3318e3324
3319
2.2. Fabrication of DSSCs
O
H2C
H2C
O(C3H6O)
O(C3H6O)
O
DSSCs were fabricated using fluorine-doped tin oxide (FTO)
glass (20e25 /sq) as a substrate for photoelectrode. A mesoporous
O
CH2
U
(OC3H6)O
anatase TiO2 layer coated on the FTO glass was prepared by sol-gel
process as suggested in the literature [27]. In general, 72 mL tita-
nium tetraisopropoxide (TTIP) was added to 430 mL of 0.1 M HNO3.
After vigorously stirring for 30 min, the solution was heated to
85e90 ꢀC and kept stirring for 8 h for peptization. After filtering to
remove the large particles, the solution was placed in an autoclave
for reaction at 240 ꢀC for 12 h. After cooling, it was concentrated to
13 wt% of TiO2. Poly(ethylene glycol) was then added to the solution
in 30 wt% of TiO2 weight. After further stirring for 5 h to obtain TiO2
paste, it was spread on the FTO glass subtract using a glass rod and
an adhesive tape was used as a spacer to control the thickness to
Ru-S carries two styryl crosslinkable groups which are an elec-
tron donor, whereas GPTA carries three acrylic groups, an electron
acceptor. The crosslinking reaction rate will be increased if they
form a chargeetransfer complex [25]. The power conversion effi-
ciency of dye-sentitized solar cell (DSSC) using Ru-S and liquid
electrolyte containing lithium iodide (LiI) reached 7.53% under
standard global AM 1.5 full sunlight. As Ru-S was crosslinked with
GPTA (the product is designated as Ru-S-cross-GPTA), not only 89%
of dye retained on TiO2 surface as compared to 28% of uncrosslinked
Ru-S after immersion in 0.1 N NaOH aqueous solution, but the
power efficiency was also increased to 7.88%.
w120 mm. The photoelectrode was obtained by heating the dried
TiO2 paste on the FTO glass in an oven at 450 ꢀC for 30 min and
repeating the coating and heating process two more times to
complete the sintering process.
An active area of 0.25 cm2 was selected from the TiO2 sintered
electrode, immersed in a solution of acetonitrile and tertiary
butanol (volume ratio of 1:1) containing 3 ꢁ 10ꢂ4 M of Ru-S for
24 h, rinsed with acetone, and dried. To crosslink Ru-S with GPTA in
optimal practice condition, two drops of 1 ꢁ 10ꢂ3 M GPTA mixing
with 1 ꢁ 10ꢂ4 M azobisisobutyronitrile (AIBN) in acetonitrile were
applied on top of the Ru-S layer. The crosslinking reaction was
immediately performed under UV exposure at 300 W mꢂ2 for 1 min
followed by heating at 80 ꢀC for 30 min. The counter electrode was
fabricated by sputtering a thin Pt layer with a thickness of w100 nm
2. Experimental section
2.1. Materials
All solvents and chemicals were obtained from Acros and Aldrich.
Ru-S was synthesized on the basis of typical one-pot synthetic method
developed for heteroleptic polypyridyl ruthenium complexes [26]. The
synthesis scheme is illustrated in Fig. 1 with the detail synthesis
procedure shown in the supporting information.
Fig. 1. Synthesis scheme of Ru-S.

3320
K.-Y. Liu et al. / Polymer 52 (2011) 3318e3324
The gelled electrolyte system after heating to fluid was injected
from a small filling hole which was pre-drilled on the counter
electrode. At last, the filling hole was sealed with hot melt foil.
2.3. Analytical methods
Nuclear magnetic resonance spectra were recorded on a Bruker
AVANCE 500 spectrometer at 298 K by utilizing deuterated chlo-
roform and dimethyl sulphoxide (d6-DMSO) as solvents. Elemental
analysis was conducted by HERAEUS VarioEL EA, NCH elemental
analysis instrument. UV-vis absorption spectra were recorded on
a Jasco model 555 spectrometer. Fourier-transform infrared (FT-IR)
spectra were recorded on a Jasco model 480 Fourier-transform
infrared spectrometer and ATR-FTIR spectra were recorded on
a ThermoNicolet NEXUS470 Fourier-transform infrared spectrom-
eter. Photovoltaic characterization of all the DSSCs was carried out
by illuminating the cell at full sunlight (100 mW cmꢂ2) with
a 1000 W ozone-free Xenon lamp equipped with a water-based IR
filter and AM 1.5 filter (Newport). The open-circuit voltage (Voc) and
photocurrent-voltage characterization plots were recorded with
a potentiostat/galvanostat (PGSTAT 302N, Autolab, Eco-Chemie, the
Netherlands).
Fig. 2. 1H-NMR spectrum of Ru-S.
on indium tin oxide (ITO, with a sheet resistivity of 8e10 U/sq) glass
plate.
The liquid electrolyte systems containing 0w0.2 M LiI and
0.5w0.7 M 1-methyl-3-propyl-imidazolium iodide (PMII), 0.05 M
I2, 0.5 M 4-tert-butylpyridine (TBP) and 0.1 M guanidinium thio-
cyanate (GuNCS) in 3-methoxypropylnitrile (MPN) were applied to
the dye-sensitized electrode. To fix the total concentration of
Iꢂ ions in MPN at 0.7 M, the concentration of PMII was varied
responding to the alteration of LiI concentration. The thickness of
3. Result and discussion
3.1. Characterization and crosslinking property of Ru-S
Fig. 2 shows the 1H NMR spectrum of Ru-S with the enlarged
aromatic region and asigned proton peaks. Due to different elec-
tronical environment, the symmetrical pyridyl ring protons in
a bipyridine ligand have two different reasonance peaks, indicating
that the ligands are coordinated to ruthenium ion in cis-hetero-
leptic complex form [31]. The vinyl protons of styryl groups can be
liquid electrolytes was controlled at 60
mm using a spacer. The
assembly of DSSCs was done by pressing the counter electrode
against the electrolyte-filled dye-sensitized electrode.
On the other hand, PMA latex powder prepared by soap-free
emulsion polymerization [28e30] was used to gel the liquid elec-
trolyte system at an amount of w25 vol% of MPN. The final
concentrations of PMA-gelled electrolyte system were 0.5 M LiI,
0.05 M I2, and 0.5 M TBP. To seal the DSSCs, a hot melt sealing foil
SX1170-60 (SOLARONIX) was center-cut and placed on top of the
counter electrode to form a frame for enclosing the coated dye, and
then the photoelectrode was placed on top of the frame. Both
electrodes were assembled by melting the sealing foil with heat.
clearly observed at
d
¼ 5.24, 5.82, and 6.72 ppm. Fig. 3 shows the
13C NMR spectrum of Ru-S with the asigned carbons for the
aromatic, carboxlic acid, eNCS and styryl groups. Twenty resonance
peaks in the aromatic region from two bipyridine ligands, twelve
peaks from two stryrl benzen rings, two peaks from two carboxylic
acids, and two peaks from two eNCS groups also supported that the
ligands are coordinated to ruthenium ion in cis form. The resonance
Fig. 3. 13C-NMR spectrum of Ru-S.

K.-Y. Liu et al. / Polymer 52 (2011) 3318e3324
3321
Fig. 6. UV-vis absorption spectra of Ru-S and Ru-S-cross-GPTA before and after
immersion in 0.1 N NaOH aqueous solution.
Fig. 4. FTIR spectra of Ru-S before and after crosslinked with AIBN initiator.
the TiO2 mesoporous layer. The possessed additional ester and ether
peaks at 1732 cmꢂ1 and 1234 cmꢂ1 contributed by GPTA for Ru-S-
cross-GPTA compared to Ru-S were observed. The extent of cross-
linking density for Ru-S and Ru-S-cross-GPTA on the TiO2 meso-
porous layer respectivly were investigated by immersing the
crosslinked samples in the 0.1 N NaOH aqueous solution for 1 min.
After fully rinsed with water and acetonitrile, they were subjected to
measure the UV-vis absorption spectroscopy. Fig. 6 shows the results
in comparion with their UV-vis absorption spectra before immer-
sion. 89% of Ru-S-cross-GPTA dye still remained on the TiO2 meso-
porous layer as estimated by the absorption peak at 500 nm, whereas
the pristine Ru-S had only 28% remained and the crosslinked Ru-S by
itself with AIBN initiator has 62% remained [32]. Because the
ruthenium dyes anchored to the TiO2 mesoporous surface using two
carboxylic groups through the coordination modes such as uniden-
tate mode, chelating mode, and bridging bidentate mode [33e35],
the styryl group away from the TiO2 surface would be easy to
crosslink with GPTA.
peaks at
of styryl groups.
d
¼ 114 and 136 ppm were contribued by the vinyl carbons
FTIR spectrum of Ru-S in KBr pallet is shown in Fig. 4. The
absorption band observed at 2100 cmꢂ1 was contributed by the
stretching of eNSC groups and that at 1730 cmꢂ1 was contributed
by carboxylic acid groups. The absorption bands contributed by the
vinyl groups of Ru-S dye can be clearly seen at 910 and 3070 cmꢂ1
.
To investigate the reactivity of vinyl groups, a drop of the Ru-S
solution with AIBN initiator (1 mol% of Ru-S) in acetonitrile was
placed on the glass slide, which was immediately polymerized
under UV exposure at 300 W mꢂ2 for 1 min and followed by
heating at 80 ꢀC for 30 min to remove the remaining solvent. The
crosslinked Ru-S which could not be dissolved in any solvent was
ground with KBr for FTIR measurement. Compared to the untreated
Ru-S by normalizing the spectrum to the major absorption band at
2100 cmꢂ1 (see Fig. 4), the vinyl peaks at 910 and 3070 cmꢂ1 were
significantly reduced, and the peaks at 2852 cmꢂ1 and 2930 cmꢂ1
contributed by eCH- and eCH2- appeared, indicating that the vinyl
groups have been open for crosslinking reactions.
3.2. Liþ coordinating property of Ru-S
Moreover, Ru-S is more facile to crosslink with GPTA than by itself
owing to the possible formation of “charge-transfer complexes” [25].
Fig. 5 shows the ATR-FTIR spectra of Ru-S-cross-GPTA and Ru-S on
Ru-S and Ru-S-cross-GPTA are capable of coordinating Liþ ions
because both contain ether linkages. Fig. 7 shows the ATR-FTIR
Fig. 7. ATR-FTIR spectra of (a) Ru-S dye adsorbed on TiO2 mesoporous film, (b) the
Ru-S dye in (a) after adding a few drops of 0.25 M LiI in acetonitrile on top and dried,
and (c) the Ru-S dye in (b) after washed with pure acetonitrile.
Fig. 5. ATR-FTIR spectra of Ru-S and Ru-S-cross-GPTA, respectively adsorbed on TiO2
mesoscopic film.

3322
K.-Y. Liu et al. / Polymer 52 (2011) 3318e3324
Table 2
Photovoltaic parameters of DSSCs with Ru-S-cross-GPTA dye containing various
concentrations of LiI in the liquid electrolyte.
LiI (M)
h
(%)
Voc (V)
Jsc (mA cmꢂ2
)
FF
0
5.80 ꢃ 0.11
7.24 ꢃ 0.03
7.88 ꢃ 0.02
7.69 ꢃ 0.06
7.75 ꢃ 0.08
0.742 ꢃ 0.002
0.733 ꢃ 0.005
0.720 ꢃ 0.001
0.713 ꢃ 0.003
0.704 ꢃ 0.002
10.97 ꢃ 0.15
14.25 ꢃ 0.22
18.73 ꢃ 0.06
19.49 ꢃ 0.08
18.69 ꢃ 0.06
0.71 ꢃ 0.01
0.69 ꢃ 0.01
0.58 ꢃ 0.01
0.55 ꢃ 0.01
0.59 ꢃ 0.02
0.05
0.1
0.15
0.2
photocurrent-voltage characteristics of DSSC with Ru-S and Ru-S-
cross-GPTA dyes, respectively containing various concentrations of
LiI in the liquid electrolytes. Their related photovoltaic parameters
such as Voc, short circuit current density (Jsc), fill factor (ff), and
power conversion efficiency (h) are listed in Tables 1 and 2. For the
concentration of LiI lower than 100 mM, Voc of the DSSC with Ru-S
dye was almost unchanged. Until the concentration of LiI was
increased to 150 mM, Voc began to drop from 743 to 722 mV. On the
other hand, Jsc was steadily increased from 11.99 to 19.04 mA cmꢂ2
as the concentration of LiI was increased from 0 to 150 mM. The
highest
h
of DSSC with Ru-S dye was 7.53 ꢃ 0.04%. The increase of Jsc
with the presence of Liþ in the liquid electroltyte has been thor-
oughly studied by the transient absorbance decay kinetics of the K-
51 dye and attributed to the higher dye regeneration rate compared
to the absent of Liþ [21]. It has been suggested that because the
surface of TiO2 mesoporous film is negatively charged, Liþ would be
adsorbed, which increased the local concentration of the iodide and
hence increased the dye regeneration rate. However, if LiI is over
supplied, it would increase the charge recombination rate and
deteriorate the photovoltaic performance of DSSC. It explains why
all the Voc, Jsc and
h decreased as the concentration of LiI increased
to 200 mM for the DSSC with Ru-S dye.
For the DSSC with Ru-S-cross-GPTA and without containing LiI
in the liquid electrolyte, the power conversion efficiency was only
5.80 ꢃ 0.11%, which is lower than that with Ru-S dye and without
containing LiI. The crosslinked GPTA may retard the regeneration of
Ru-S dye so that the power conversion efficiency was decreased.
However, it steadily increased to 7.8 ꢃ 0.02% as the concentration of
LiI was increased to 100 mM. Jsc was also increased from
10.97 ꢃ 0.15 to 19.49 ꢃ 0.08 mA cmꢂ2, whereas Voc was slightly
decreased from 0.742 ꢃ 0.002 to 0.720 ꢃ 0.001 V. Further increase
of LiI concentration decreased the photovoltaic performance of
DSSC. Basically, the DSSC with Ru-S-cross-GPTA dye showed the
similar trend of photovoltaic performance as that with Ru-S dye.
The CeOeC linkages of Ru-S-cross-GPTA dye are likely to
Fig. 8. Photocurrent-voltage plots of the DSSCs with (a) Ru-S and (b) Ru-S-cross-GPTA,
respectively with various concentrations of LiI in the liquid electrolyte.
spectrum of Ru-S dye adsorbing on the surface of the TiO2 mesoporous
film, which has been subtracted a spectrum of “blank” TiO2 meso-
porous reference film for the background correction. The broad peak at
1050e1200 cmꢂ1 is derived from the
n(CeO) stretch of the methylene
oxide group. A few drops of 0.25 M lithium iodide in acetonitrile
solution were then added onto the Ru-S dye and fully dried. The
change of IR spectrum was monitored, which was also shown in Fig. 7.
The peak at the
owing to the appearance of the additional peak at 1070 cmꢂ1
According to the literature [21], the
(CeO) stretch at 1106 cmꢂ1
shifted to lower energy (1078 cmꢂ1) after coordination with lithium
ions
n
(CeO) stretch (1050e1200 cmꢂ1) became broader
.
n
n
(CeO.Liþ). The similar trend was also observed in our case. As
the sample was washed with acetonitrile to remove LiI and dried, the
original peak is restored (see Fig. 7).
In order to investigate the effect of Liþ coordination with the
dyes on the performance of DSSC, the liquid electrolyte systems
containing various concentrations of LiI were used. Fig. 8 shows the
Table 1
Photovoltaic parameters of DSSCs with Ru-S dye containing various concentrations
of LiI in the liquid electrolyte.
LiI (M)
h
(%)
Voc (V)
Jsc (mA cmꢂ2
)
FF
0
6.25 ꢃ 0.055
7.39 ꢃ 0.056
7.53 ꢃ 0.040
7.48 ꢃ 0.069
7.29 ꢃ 0.038
0.743 ꢃ 0.005
0.739 ꢃ 0.002
0.740 ꢃ 0.001
0.722 ꢃ 0.004
0.715 ꢃ 0.002
11.99 ꢃ 0.129
14.91 ꢃ 0.083
18.24 ꢃ 0.040
19.04 ꢃ 0.183
16.71 ꢃ 0.162
0.702 ꢃ 0.008
0.671 ꢃ 0.001
0.558 ꢃ 0.003
0.544 ꢃ 0.008
0.610 ꢃ 0.001
0.05
0.1
0.15
0.2
Fig. 9. Photocurrent-voltage plots of Ru-S and Ru-S-cross-GPTA, respectively with the
PMA-gelled electrolyte.

K.-Y. Liu et al. / Polymer 52 (2011) 3318e3324
3323
Fig. 10. Time-course changes of photovoltaic performance for PMA-gelled DSSCs with Ru-S and Ru-S-cross-GPTA dyes.
coordinate Liþ preventing it from intercalation onto the surface of
TiO2 mesoporous layer.
After 30 days of storage, the power efficiency of DSSC with Ru-S
retained 99%, whereas that with Ru-S-cross-GPTA increased 1%. The
Voc for both of the devices with Ru-S and Ru-S-cross-GPTA dyes
showed a little increase in the initial stage and stayed unchanged
up to the end of storage time. The Jsc of the device with Ru-S dye
showed an initial increase at first 10 days, and then slightly
decreased. The increase of photocurrent in the initial storage of the
DSSCs might be due to the excess penetration of gelled electrolytes
into the pores of TiO2 mesoporous layer [37]. After 30 days, Jsc
decreased to 93% relative to the original value. The Jsc of the device
with Ru-S-cross-GPTA dye also showed a similar trend. The fill
factors of both devices were quite steady in the initial stage and
then slightly increased when Jsc decreased. The increase of power
efficiencey for the DSSCs with Ru-S-cross-GPTA dye was due to the
increase of fill factor. Voc barely changed.
3.3. Photovoltaic performance of PMA-gelled DSSCs with Ru-S and
Ru-S-cross-GPTA dyes
Because the liquid electrolyte in DSSCs has problems of leakage
and evaporation of solvent, polymers and their nanocomposites
have been frequently used to gel the electrolyte systems [27,28,36].
It has been reported that by utilizing PMA with Tg of 6 ꢀC to gel the
electrolyte system, the power efficiency of fabricated DSSC is
superior to that gelled by the polymers with higher Tg [28].
Therefore, in this study PMA was chosen to gel the electrolyte
system of DSSCs with Ru-S and Ru-C-cross-GPTA dyes. Their typical
photocurrent density-voltage characteristics were shown in Fig. 9.
The average Jsc and Voc of DSSCs with Ru-S were
15.83 ꢃ 0.02 mA cmꢂ2 and 0.723 ꢃ 0.001 V with a fill factor of
4. Conclusion
0.61 ꢃ 0.01, resulting the average
h
¼ 6.96 ꢃ 0.02%. Whereas, the
DSSC with Ru-S-cross-GPTA has Jsc ¼ 16.83 ꢃ 0.06 mA cmꢂ2
,
A new crosslinkable Ru-S dye was synthesized and well charac-
terized. Its capability of coordinating Liþ was verified by ART-FTIR
spectroscopy. By crosslinking with GPTA to form a Ru-S-cross-
GPTA layer on TiO2 mesoporous film, it is almost insoluble in 0.1 N
NaOH aqueous solution. By increasing the amount of LiI in the liquid
electrolyte, the DSSCs with Ru-S and Ru-S-cross-GPTA both showed
small drop of Voc but significant increase of Jsc. The highest power
conversion efficiency of 7.88% was achieved for the DSSC with Ru-S-
cross-GPTA dye containing 150 mM LiI in the liquid electrolyte. By
using the PMA-gelled electrolyte system, the DSSC with Ru-S-cross-
GPTA attained a power efficiency of 7.57%. For the time-course
change test of power efficiency in storage, the DSSCs with Ru-S
and Ru-S-cross-GPTA respectively showed an initial increase of
power efficiencies and then stayed almost unchanged for a month.
Voc ¼ 0.701 ꢃ 0.001 V, fill factor of 0.64 ꢃ 0.01 and ¼ 7.53 ꢃ 0.03%.
h
The incorporation of GPTA to crosslink with Ru-S has brought
higher Jsc and better fill factor, leading to higher power efficiency.
The PMA-gelled DSSCs with Ru-S and Ru-S-cross-GPTA dyes,
respectively were then subjected to measure the long-term
stability. All the samples were stored at ambient environment. At
chosen time intervals, the samples were taken to measure the
power efficiency at room temperature under AM 1.5 and 100
mW cmꢂ2 illumination. All the photovoltaic paramenters such as
power efficiency, photovoltage, photocurrent, and fill factors of
samples at different storage time shown in Fig. 10 were normalized
to their initial measured values. The results show that the power
efficiences of all cells in the storge period fluctuated with time.

3324
K.-Y. Liu et al. / Polymer 52 (2011) 3318e3324
[15] Liu KY, Hsu CL, Chang SH, Chen JG, Ho KC, Lin KF. J Polym Sci Part A Polym
Acknowledgment
Chem 2010;48:366e72.
[16] Pelet S, Moser JE, Grätzel M. J Phys Chem B 2000;104:1791e5.
[17] Nakade S, Kambe S, Kitamura T, Wada Y, Yanagida S. J Phys Chem B 2001;105:
9150e2.
[18] Nakade S, Kanzaki T, Kubo W, Kitamura T, Wada Y, Yanagida S. J Phys Chem B
2005;109:3480e7.
The authors thank for the financial support of the National
Science Council in Taiwan through Grants NSC95-2221-E-002-206
and NSC 96-2120-M-002-016.
[19] Kambe S, Kitamura T, Wada Y, Yanagida S. J Phys Chem B 2002;106:2967e72.
[20] Frank AJ, Kopidakis N, van de Lagemaat J. Coord Chem Rev 2004;248:
1165e79.
Appendix. Supplementary material
[21] Kuang D, Klein C, Snaith HJ, Moser JE, Humphry-Baker R, Comte P, et al. Nano
Lett 2006;6:769e73.
[22] Snaith HJ, Moule AJ, Klein C, Meerholz K, Friend RH, Grätzel M. Nano Lett
2007;7:3372e6.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.polymer.2011.05.035.
[23] Snaith HJ, Zakeeruddin SM, Schmidt-Mende L, Klein C, Grätzel M. Angew
Chem Int Ed 2005;44:6413e7.
References
[24] Kuang D, Klein C, Snaith HJ, Humphry-Baker R, Zakeeruddin SM, Grätzel M.
Inorg Chim Acta 2008;361:699e706.
[1] O’Regan B, Grätzel M. Nature 1991;353:737e40.
[2] Nazeeruddin MK, Kay A, Rodicio I, Humphrey-Baker R, Müller E, Liska P, et al.
J Am Chem Soc 1993;115:6382e90.
[25] Billmeyer Jr FW. Texbook of polymer science. 3rd ed. New York: Wiey; 1984
[Chapter 5].
[26] Wang P, Zakeeruddin SM, Humphry-Baker R, Moser JE, Grätzel M. Adv Mater
2003;15:2101e4.
[3] Bach U, Lupo D, Comte P, Moser JE, Weissörtel F, Salbeck J, et al. Nature 1998;
395:583e5.
[27] Tu CW, Liu KY, Chien AT, Yen MH, Weng TH, Ho KC, et al. J Polym Sci Part A
Polym Chem 2008;46:47e53.
[4] Grätzel M. Nature 2001;414:338e44.
[5] Grätzel M. Acc Chem Res 2009;42:1788e98.
[28] Tu CW, Liu KY, Chien AT, Lee CH, Ho KC, Lin KF. Eur Polym J 2008;44:608e14.
[29] Lin KJ, Weng TH, Lee CH, Lin KF. J Polym Sci Part A Polym Chem 2009;47:
5891e7.
[6] Grätzel M. J Photochem Photobio B 2003;4:145e53.
[7] Fan SQ, Kim C, Fang B, Liao KX, Yang GJ, Li CJ, et al. J Phys Chem C 2011;115:
7747e54.
[30] Lin KJ, Dai CA, Lin KF. J Polym Sci Part A Polym Chem 2009;47:459e66.
[31] Kong FT, Dai SY, Wang KJ. Chin J Chem 2007;25:168e71.
[32] Ko CY. Applications of crosslinkable ruthenium complex and its crosslinker on
dye-sensitized solar cells. Master Thesis, National Taiwan University, 2010.
[33] Murakoshi K, Kano G, Wada Y, Yanagida S, Miyazaki H, Matsumoto M, et al.
J Electroanal Chem 1995;396:27e34.
[34] Finnie KS, Bartlett JR, Woolfrey JL. Langmuir 1998;14:2744e9.
[35] Fillinger A, Parkinson BA. J Electrochem Soc 1999;146:4559e64.
[36] De Freitas JN, Nogueira AF, De Paoli MA. J Mater Chem 2009;19:5279e94.
[37] Wang P, Klein C, Humphry-Baker R, Zakeeruddin SM, Grätzel M. J Am Chem
Soc 2005;127:808e9.
[8] Jang SR, Lee C, Choi H, Ko JJ, Lee J, Vittal R, et al. Chem Mater 2006;18:
5604e8.
[9] Chen CY, Wu SJ, Wu CG, Chen JG, Ho KC. Angew Chem Int Ed 2006;45:5822e5.
[10] Karthikeyan CS, Wietasch H, Thelakkat M. Adv Mater 2007;19:1091e5.
[11] Chen CY, Wu SJ, Li JY, Wu CG, Chen JG, Ho KC. Adv Mater 2007;19:3888e91.
[12] Kuang D, Klein C, Ito S, Moser JE, Humphry-Baker R, Evans N, et al. Adv Mater
2007;19:1133e7.
[13] Wang P, Zakeeruddin SM, Moser JE, Nazeeruddin MK, Sekiguchi T, Grätzel M.
Nat Mater 2003;2:402e7.
[14] Kuang D, Klein C, Ito S, Moser JE, Humphry-Baker R, Zakeeruddin SM, et al.
Adv Funct Mater 2007;17:154e60.