Photocurrents of solar cells sensitized by aggregate-forming polyenes: Enhancement due to suppression of singlet-triplet annihilation by lowering of dye concentration or light intensity

Article abstract of DOI:10.1016/j.cplett.2006.01.003

Titania-based Gr?tzel-type solar cells were fabricated by the use of polyene dyes with various transition dipole moments. In the dye having the largest transition dipole among the samples, an aggregate was readily formed through dispersive interaction, an

Full text of DOI:10.1016/j.cplett.2006.01.003

Chemical Physics Letters 420 (2006) 309–315
www.elsevier.com/locate/cplett
Photocurrents of solar cells sensitized by aggregate-forming
polyenes: Enhancement due to suppression of singlet–triplet
annihilation by lowering of dye concentration or light intensity
a
a,
b
c
Xiao-Feng Wang , Yasushi Koyama *, Hiroyoshi Nagae , Yumiko Yamano ,
c
d
Masayoshi Ito , Yuji Wada
a
Faculty of Science and Technology, Kwansei Gakuin University, Department of Chemistry, 2-1 Gakuen, Sanda 669-1337, Japan
b
Kobe City University of Foreign Studies, Gakuen Higashim achi, Nishiku, Kobe 651-2187, Japan
c
Kobe Pharmaceutical University, 4-19-1 Motoyama-Kitamachi, Kobe 658-8558, Japan
Department of Material and Life Science, Graduate School of Engineering, Osaka University, Suita 565-0871, Japan
d
Received 3 December 2005; in final form 31 December 2005
Available online 25 January 2006
Abstract
Titania-based Gra¨tzel-type solar cells were fabricated by the use of polyene dyes with various transition dipole moments. In the dye
having the largest transition dipole among the samples, an aggregate was readily formed through dispersive interaction, and the photo-
current was increased when the dye concentration or the light intensity was lowered. This observation was ascribed to the suppression of
the singlet–triplet annihilation reaction. In the dye having the smallest transition dipole, there was no sign of aggregate formation, and
the photocurrent was decreased when the dye concentration or the light intensity was lowered.
Ó 2006 Elsevier B.V. All rights reserved.
1. Introduction
current conversion efficiency (IPCE) as well as the solar
energy-to-electricity conversion efficiency (to be called con-
version efficiency, g) were the highest for the CA having 7
conjugated double bonds (CA7), and the IPCE and the g
values declined toward n = 5 and n = 13 (from 1.5% to
1.0% and 0.3%, respectively) (see Table 1 of Supplementary
Information of Ref. [7]). For CA7, we examined the effect
of dilution with deoxycholic acid as a spacer on the perfor-
mance. At various concentrations, we scaled the above-
mentioned values to those per unit concentration and then
divided by the value at 100% to obtain the enhancement
factors. They monotonically increased with the decreasing
concentration and reached a value as high as ꢀ9–10 times
at the 10% dye concentration (see Fig. 4 of Ref. [7]). The
result indicated that isolated excitation was the best to
obtain the highest performance per unit dye molecule.
In another study [8], we examined the excited-state
dynamics of RA5 and CA6–CA11 bound to TiO2 nanopar-
ticles in suspension by sub-ps and sub-ls time-resolved
absorption spectroscopy. The electron-injection efficiency
The geometrical arrangement and the excited-state
dynamics of the dye molecules on the surface of the TiO2
layer must strongly affect the performance of dye-sensitized
solar cells (DSSCs): It is well documented that formation
of dye aggregates suppresses the performance [1] and that
prevention of aggregate formation by co-adsorption of
spacers [2,3] or by structural modification of dye sensitizer
[4] can solve this problem. It is also shown that high light-
intensity can sometimes suppress the performance of
DSSCs [5,6]. All those observations have never been
explained systematically, as far as we know.
In our previous study [7], we fabricated TiO2-based
DSSCs using a series of retinoic acid and carotenoic acids
(RA and CAs) as dye sensitizers having the number of con-
jugated double bonds, n = 5–13. The incident photon-to-
*
Corresponding author. Fax: +81 79 565 8408.
E-mail address: ykoyama@kwansei.ac.jp (Y. Koyama).
0009-2614/$ - see front matter Ó 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2006.01.003

310
X.-F. Wang et al. / Chemical Physics Letters 420 (2006) 309–315
(U) immediately after excitation to the 1Bþu singlet state was
as high as 98% in CA7, and declined toward n = 11 (29%).
However, the U value declined only slightly toward n = 5
(92%), contrary to the above-mentioned large decline of
g. On the other hand, the efficiency of triplet generation
increased systematically from 75% to 92% on going from
CA7 to RA5.
COOH
COOH
-6-CA
Based on these results, we ascribed the decline of the
performance of DSSCs from CA7 to RA5 to an increasing
singlet–triplet annihilation reaction in the aggregate of the
dye molecules [7].
MeO- -6-C A
MeO
OM
e
The present investigation has been planned to establish
this idea: We have synthesized a series of polyene dyes hav-
ing various transition-dipole moments, and as a result, dif-
ferent potentials of aggregate formation through the
dispersive (van der Waals) interaction. Then, we fabricated
DSSCs by the use of those polyenes as sensitizers at various
concentrations. In order to prove the idea, we have exam-
ined the effect of dilution of each dye sensitizer with deoxy-
cholic acid on the performance of the DSSCs, addressing
two questions: First, how does the lowering of the dye con-
centration (to suppress singlet–triplet annihilation) affect
the photocurrent and the conversion efficiency?
COOH
COOH
OM
e
M
e
O
(MeO)3- -6-CA
M
e2
N
- -6-C A
N
Me2
Scheme 1.
To achieve definitive evidence for the mechanism of sin-
glet–triplet annihilation, we have also examined the effect
of light intensity. We have addressed the second question:
How does the lowering of the light intensity (to suppress
singlet–triplet annihilation) affect the photocurrent in the
DSSCs using those dye sensitizers?
deoxycholic acid (spacer) in THF solution was conserved
after deposition on the TiO2 layer.
3. Results and discussion
3.1. Characterization of dye sensitizers
2. Experimental
The dye sensitizers used in this study (Scheme 1) are
analogues of carotenoic acids, consisting of a conjugated
chain having 6 double bonds, and one of the phenyl, meth-
oxyphenyl, tri-methoxyphenyl and N-dimethylphenyl
groups attached to it. According to their chemical struc-
tures, we call them ‘/-6-CA’, ‘MeO-/-6-CA’, ‘(MeO)3-/-
6-CA’ and ‘Me2N-/-6-CA’, respectively. Table 1 compares
the observed and calculated l values. The lobs values
increase systematically in the order, /-6-CA < MeO-/-6-
CA 6 (MeO)3-/-6-CA < Me2N-/-6-CA. The lcalc values
exhibit much smaller differences.
In the electronic spectra of the dye sensitizers in THF
solution, shown in Fig. 1a, the 1Bþu profile manifests a sys-
tematic red shift, and its intensity decreases in the order, /-
6-CA, MeO-/-6-CA, (MeO)3-/-6-CA and Me2N-/-6-CA;
the absorption maxima are indicated in the figure and listed
in Table 1. As for the dye sensitizers deposited on the TiO2
layer, all the absorption profiles are broadened, as shown
in Fig. 1b. The absorption maxima exhibit a systematic
red shift and decrease in the case of /-6-CA, MeO-/-6-
CA and (MeO)3-/-6-CA; the profile of Me2N-/-6-CA
may be slightly shifted to the blue due to partial aggrega-
tion on the TiO2 surface. (The aggregated states are
expected to give rise to the 1Bþu absorption below
350 nm.) Fig. 1c shows the absorption spectra of the dye-
The set of polyene dyes in Scheme 1 was synthesized
starting from commercially available or synthesized [9]
benzaldehyde derivatives via the Emmons–Horner reaction
with two different kinds of C5-phosphonate [10] and a C10-
phosphonate [11], as shown in Scheme S-1 in supplementary
information, where /-6-CA, MeO-/-6-CA, (MeO)3-/-6-
CA and Me2N-/-6-CA are named 6a, 6b, 6c and 6d,
respectively. The details of syntheses and spectroscopic
characterizations of relevant compounds are also described
in Supplementary Information.
The transition dipole moment, lobs, of each compound
was determined spectroscopically from the molar extinc-
tion coefficient as a function of energy by [12]
Z
ꢀ
eðmÞ
2
jlobsj ¼ 91:86 ꢁ 10ꢂ4 ꢃ n ꢃ
ð1Þ
ꢀ
dm;
ꢀ
m
ꢀ
where n is the refractive index of solvent, m is the energy in
cmꢂ1 and lobs is the transition dipole in debye. The corre-
sponding value, lcalc, was calculated by the INDO/S SCI
method [13] after the optimization of the molecular geom-
etry by the MOPAC PM3 method [14]. Similar values were
also obtained by the INDO/S method.
Each DSSC was fabricated, and its IPCE and I–V curves
were recorded [15]. Concerning the concentration depen-
dence, we assumed that the composition of each dye and

X.-F. Wang et al. / Chemical Physics Letters 420 (2006) 309–315
311
Table 1
the trend is shown that the dyes having larger transition
dipoles have higher potentials of aggregate formation in
the dyes.
1Bþu energies, molar extinction coefficients (emax) both at absorption
maxima, oscillator strengths (f), transition dipole moments determined
spectroscopically (lobs), and those determined theoretically (lcalc) for a set
of dye sensitizers
3.2. Dependence of photocurrent and conversion efficiency on
the dye concentration
Dye
1Bþu
energy/cmꢂ1
e
max/Mꢂ1
fa
lobs/D
lcalc/D
cmꢂ1
/-6-CA
24067
23696
22779
22321
89100
85800
75300
72100
2.3
2.5
2.5
2.6
14.2
15.1
15.2
15.6
15.3
15.4
15.4
15.6
Fig. 2 shows the dependence of the IPCE profiles on
the dye concentration in DSSCs: These profiles, as action
spectra for generation of the photocurrent, indicate that
/-6-CA and MeO-/-6-CA are rather in the monomeric
state, whereas (MeO)3-/-6-CA and Me2N-/-6-CA are
in the aggregated state at 100%. The latter two transform
to the monomeric state at 5%. The effect of dilution can
be summarized as follows: (i) The photocurrent systemat-
ically decreases in /-6-CA, whereas it increases in
Me2N-/-6-CA on going from 100% to 5%; (ii) MeO-/-
6-CA and (MeO)3-/-6-CA exhibit two phases of rise
and decay with decreasing dye concentration, and the
highest IPCE profiles appear at 70% and 10%,
respectively.
MeO-/-6-CA
(MeO)3-/-6-CA
Me2N-/-6-CA
a
2
Calculated by f ¼ 4:71 ꢁ 10ꢂ7 ꢃ ꢀmmax ꢃ jlobsj .
a
100,000
50,000
Fig. 3 shows the dependence of the I–V curves on the
dye concentration in DSSCs: In /-6-CA, the Jsc value
decreases monotonically with the dye concentration low-
ered from 100% to 5%. In MeO-/-6-CA, the Jsc value
jumps up from 100% to 70% and then declines gradually
toward 5%. For these sensitizers, the Voc values are similar
to one another at different concentrations.
0
3
b
For (MeO)3-/-6-CA, the Jsc value jumps up from 100%
to 70% and stays almost unchanged, whereas the Voc value
increases at lower concentrations, which probably reflects
an improvement of the molecular packing supported by
the spacer, because the bulky head group, (MeO)3-/,
may prevent close packing of the dye molecules by them-
selves (see Scheme 1). For Me2N-/-6-CA, however, the
Jsc and Voc values both increase systematically at lower
concentrations (except for an anomaly at 30%, which
may be an artifact). It is highly likely that the Jsc value
increases due to the suppression of singlet–triplet annihila-
tion, while the Voc value increases due to the better molec-
ular packing supported by the spacer.
2
1
0
4
c
2
0
Table 2 lists the values relevant to the performance of
DSSCs extracted from the I–V curves. The highest g values,
2.2%, 2.6%, 1.8% and 1.2%, are obtained for /-6-CA,
MeO-/-6-CA, (MeO)3-/-6-CA and Me2N-/-6-CA at the
dye concentrations of 100%, 70%, 10% and 5%, respec-
tively. The conversion efficiency scaled to a unit concentra-
tion, gu, increases monotonically toward the lowest
concentration, 5%, for all the sensitizers. The enhancement
factor of the conversion efficiency per unit concentration,
ge, with reference to the value at 100%, increases as 12,
17, 31 and 60 times for /-6-CA, MeO-/-6-CA,
(MeO)3-/-6-CA and Me2N-/-6-CA, respectively.
400
500
600
700
Wavelength / nm
Fig. 1. Electronic absorption spectra of dye sensitizers (a) in THF
solution, and those adsorbed on the TiO2 layer at the concentrations of (b)
5% and (c) 100%.
sensitizers deposited on the TiO2 layer without dilution
(100%). The profiles are similar to one another, probably
reflecting the blue shift of absorption due to the formation
of aggregates in (MeO)3-/-6-CA and Me2N-/-6-CA. Thus,
The conversion efficiency, g, as observed, decreases
monotonically with the decreasing dye concentration, as
shown in Fig. 4. The trend for /-6-CA is obviously
ascribed to the decrease in the number of dye molecules.

312
X.-F. Wang et al. / Chemical Physics Letters 420 (2006) 309–315
80
60
40
20
8
50%
70%
5%
70%
6
50%
30%
10%
100%
10%
30%
4
5%
2
0
0
80
8
70%
70%
10%
50%
30%
60
100%
6
4
2
30%
50%
5%
10%
40
20
5%
0
0
8
80
30%
60
40
20
70%
6
4
2
10%
50%
5%
10%
5%
70%
100%
30%
50%
0
80
0
8
60
40
20
0
6
4
2
0
50%
30%
5%
10%
50%
10%
5%
70%
100%
30
%
70%
400
500
Wavelength / nm
600
700
0
0.2
0.4
0.6
Fig. 2. Dependence of the IPCE profiles on the concentration of dye
sensitizers in DSSCs.
Photo-voltage / V
Fig. 3. Dependence of the I–V curves on the dye concentration in DSSCs.
On the other hand, the monotonic increase in g for
Me2N-/-6-CA is most probably due to the suppression
of singlet–triplet annihilation. For MeO-/-6-CA and
(MeO)3-/-6-CA, the effect of lowering concentration con-
sists of two phases: at an early rising phase of g by the sup-
pression of singlet–triplet annihilation, and
a later
declining phase due to the decreasing number of dye mol-
ecules available for excitation. As a result, they reach max-
ima at 70% and 10%, respectively.

X.-F. Wang et al. / Chemical Physics Letters 420 (2006) 309–315
313
Table 2
Concentration dependence of parameters relevant to photo-conversion efficiency, inclusing the values of short-circuit photocurrent (Jsc), open-circuit
photovoltage (Voc), Fill factor (FF) and conversion efficiency (g)
a
b
Concentration%
Jsc/mAcmꢂ2
Voc/V
FF
g%
gu
ge
/-6-CA
100
70
50
30
10
5
100
70
50
30
10
5
100
70
50
30
10
5
100
70
50
30
10
5
7.0
6.2
6.1
5.9
4.7
4.2
5.9
7.4
6.3
6.5
6.0
5.3
3.6
4.9
4.6
4.7
5.3
4.7
1.8
2.7
3.2
3.0
3.4
3.7
0.54
0.57
0.57
0.55
0.55
0.57
0.59
0.59
0.59
0.57
0.59
0.59
0.51
0.52
0.55
0.56
0.57
0.59
0.41
0.45
0.45
0.48
0.49
0.50
0.59
0.60
0.55
0.50
0.59
0.55
0.63
0.61
0.62
0.62
0.61
0.63
0.60
0.59
0.61
0.61
0.60
0.61
0.55
0.59
0.59
0.61
0.60
0.63
2.2
2.1
1.9
1.6
1.5
1.3
2.2
2.6
2.3
2.3
2.2
1.9
1.1
1.5
1.6
1.6
1.8
1.7
0.4
0.7
0.8
0.9
1.0
1.2
2.2
3.0
3.8
5.3
15
26
2.2
3.7
4.6
7.7
22
38
1.1
2.1
3.2
5.3
18
34
0.4
1.0
1.6
3.0
10
24
1
1.4
1.7
2.4
6.8
12
1
1.7
2.1
3.5
10
17
1
1.9
2.9
4.8
16
31
1
2.5
4.0
7.5
25
60
MeO-/-6-CA
(MeO)3-/-6-CA
Me2N-/-6-CA
a
gu = 100g/concn.
ge (x%) = gu(x%)/gu(100%).
b
As for the enhancement factor, ge, the monotonic
increase with the concentration, shown in Table 2, suggests
that the conversion efficiency is strongly suppressed by
aggregate formation, and the resultant singlet–triplet annihi-
lation, in the order: /-6-CA < MeO-/-6-CA < (MeO)3-/-
6- CA < Me2N-/-6-CA.
potential of aggregate formation, and the resultant singlet–
triplet annihilation.
4. Concluding remarks
To the questions addressed in Section 1, the following
answers have been obtained:
3.3. Dependence of the photocurrent on the light intensity
As for the first question: As we expected, the photocur-
rent and the conversion efficiency increases (decreases)
for Me2N-/-6-CA (/-6-CA), which has the largest (the
smallest) transition dipole. For the dyes having medium
transition dipoles, the optimum dye concentration is found
to be determined by two factors: one, the singlet–triplet
annihilation reaction to quench singlet excitation, and the
other, the number of dye molecules to be excited. The for-
mer factor becomes predominant for (MeO)3-/-6-CA,
which has the second largest transition dipole, whereas
the latter factor becomes predominant for MeO-/-6-CA,
which has the second smallest transition dipole (see Fig. 4).
As for the second question: In Me2N-/-6-CA, which has
the largest transition dipole, the lowering of light intensity
increases the photocurrent, whereas in /-6-CA, which has
the smallest transition dipole the lowering light intensity,
decreases the photocurrent. The former observation can
be ascribed to the suppression of the singlet–triplet annihi-
lation reaction in the dye aggregate, whereas the latter
observation can be ascribed to the suppression of the sin-
glet excitation in the dye monomers.
The observed effect at the dye concentrations of 100%
and 5%, shown in Fig. 5, can be summarized as follows:
(i) The Voc value systematically decreases in all the dye sen-
sitizers tested at these concentrations; (ii) Most impor-
tantly, the Jsc value systematically decreases for /-6-CA,
but it increases for Me2N-/-6-CA at both concentrations;
(iii) For both MeO-/-6-CA and (MeO)3-/-6-CA, the Jsc
values increase even at 100%. At 5%, however, the Jsc value
for the former decreases, whereas that for the latter stays
unchanged. The dependence on the light intensity is hardly
affected by the concentration in /-6-CA and Me2N-/-6-
CA, but it is slightly affected in MeO-/-6-CA and
(MeO)3-/-6-CA.
On lowering the light intensity, a decrease in the photo-
current is clearly observed for /-6-CA having a small tran-
sition dipole (naturally expected), whereas an increase in
the photocurrent is seen for Me2N-/-6-CA, which has a
large transition dipole (rather unexpected, see Table 1).
The latter trend can be explained in terms of much higher

314
X.-F. Wang et al. / Chemical Physics Letters 420 (2006) 309–315
8
2.2
2
<100%>
6
4
1.8
1.6
1.4
<5%>
2
0
8
2.6
2.4
2.2
2
<100%>
6
<5%>
4
100% light intensity
50% light intensity x 2
20% light intensity x 5
10% light intensity x 10
2
0
8
1.8
1.6
1.4
1.2
6
4
2
<5%>
<100%>
1
0
8
1.2
1
0.8
0.6
0.4
6
4
2
<5%>
<100%>
0
50
100
0
0.2
0.4
0.6
Dye-concentration / %
Photo-voltage / V
Fig. 4. Dependence of the conversion efficiency (g) on the dye concen-
Fig. 5. Effect of light intensity on the I–V curves of DSSCs, at the dye
concentration of 100% and 5%, after the scaling shown in inset: AM 1.5,
100 mWcmꢂ2, is defined as the light intensity of 100%.
tration in DSSCs.
In summary, the dependence of the conversion efficiency
and the photocurrent on the dye concentration and the
light intensity observed in the present study have estab-
lished the mechanism of singlet–triplet annihilation that
lowers the performance of DSSCs.
characterization of the dye systems, e.g. by precise
measurements of the efficiencies of electron-injection and
triplet-generation, the one-electron oxidation potentials
etc., and exploration of well-designed TiO2 surfaces to
accommodate a large number of the dye and spacer
molecules.
For making the best use of the potential of polyene sen-
sitizers to build a high-performance DSSC, we need further

X.-F. Wang et al. / Chemical Physics Letters 420 (2006) 309–315
315
[4] Z.-S. Wang, K. Hara, Y. Dan-oh, C. Kasada, A. Shinpo, S. Suga, H.
Arakawa, H. Sugihara, J. Phys. Chem. B 109 (2005) 3907.
[5] M.K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker, E.
Mu¨ller, P. Liska, N. Vlachopoulos, M. Gra¨tzel, J. Am. Chem. Soc.
115 (1993) 6382.
Acknowledgment
This work has been supported by a grant from the Min-
istry of Education, Culture, Sports, Science and Technol-
ogy (an Open Research Center Project, ‘The Research
Center of Photo-Energy Conversion’).
[6] T. Trupke, P. Wurfel, I. Uhlendorf, J. Phys. Chem. B 104 (2000)
¨
11484.
[7] X.-F. Wang, R. Fujii, S. Ito, Y. Koyama, Y. Yamano, M. Ito, T.
Kitamura, S. Yanagida, Chem. Phys. Lett. 416 (2005) 1.
[8] J. Xiang, F.S. Rondonuwu, Y. Kakitani, R. Fujii, Y. Watanabe, Y.
Koyama, H. Nagae, Y. Yamano, M. Ito, J. Phys. Chem. B 109 (2005)
17066.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.cplett.2006.
01.003.
˚
[9] N. Kann, T. Rein, B. Akermark, P. Helquist, J. Org. Chem. 55 (1990)
5312.
[10] F.J. Jansen, J. Lugtenburg, in: G. Briton, S. Liaaen-Jensen, H.
Pfander (Eds.), Carotenoids, Synthesis, vol. 2, Birkha¨user Verlag,
Basel, 1996, p. 232.
References
[11] Y.Yamano,M.Yoshizawa,M.Ito,J.Nutr.Sci.Vitaminol.45(1999)49.
[12] W. Liptay, Z. Naturforsch. 21A (1966) 1605.
[13] J. Ridley, M. Zerner, Theort. Chim. Acta. 32 (1973) 111.
[14] J.J.P. Stewart, J. Comput. Chem. 10 (1989) 209.
[15] X.-F. Wang, J. Xiang, P. Wang, Y. Koyama, S. Yanagida, Y. Wada,
K. Hamada, S. Sasaki, H. Tamiaki, Chem. Phys. Lett. 408 (2005) 409.
[1] K. Kalyanasundaram, M. Gra¨tzel, Coord. Chem. Rev. 77 (1998)
347.
[2] A. Kay, M. Gra¨tzel, J. Phys. Chem. 97 (1993) 6272.
[3] K. Hara, H. Sugihara, Y. Tachibana, A. Islam, M. Yanagida, K.
Sayama, H. Arakawa, Langmuir 17 (2001) 5992.