Photochemical &

Photobiological Sciences
View Article Online
PAPER View Journal | View Issue

Effects of bis-carbazole based D–π-A sensitizers on

Cite this: Photochem. Photobiol. Sci., 2013, solar energy capture in DSSCs
12, 421
Bo Hyung Kim and Harold S. Freeman*
Published on 13 November 2012 on http://pubs.rsc.org | doi:10.1039/C2PP25262G

A new series of molecular engineered dyes having a carbazole (CA) based bulky donor was synthesized
Downloaded by University of Massachusetts - Amherst on 24 March 2013

Received 23rd July 2012,
Accepted 24th October 2012
DOI: 10.1039/c2pp25262g
www.rsc.org/pps
and compared to a previously studied set of dyes having a diphenyl amine (DP) donor for their use in
dye sensitized solar cells. In this work, their photophysical, electrochemical, and kinetic properties were
measured and molecular modeling methods were used to predict their electronic and physical properties.
Regarding the latter, TDDFT studies adopting different exchange correlation functionals were performed
to predict solar capture behavior of these sensitizers. The calculated electronic energies showed that the
introduction of the bulky CA donor gave a greater negative shift on EHOMO rather than ELUMO, imparting
the CA series of dyes a red shifted absorption. On the other hand, tuning of the spacer unit by using
different heterocylic rings was sensitive to the shift of ELUMO shift. Better light absorption and driving
force of devices based on the CA series were counteracted with negative effects related to the electron
injection yield and electron recombination lifetime. The origin of the lower efficiency of the CA dye
based devices was investigated in terms of dye load, charge recombination lifetime, in tandem with mod-
eling studies. Results from assessing the effects of different heterocyclic rings, as a π conjugation unit on
photovoltaic performance showed that the thiophene unit gave the highest current. On the other hand,
CA- and DP-based dyes having an N-methyl pyrrole unit afforded the highest Voc. The best DSSC perform-
ance in the CA series was achieved by the thiophene derivative, where η = 3.05% with Jsc = 6.29 mA
cm−2, Voc = 673 mV and FF = 0.72 under 100 mW m−2 irradiation.

Introduction desired spectral properties can conflict with factors affecting

Dye sensitized solar cells (DSSCs) have attracted significant
attention as a new renewable energy source and become one of
the most promising alternatives to the conventional semicon-
ductor photovoltaic devices.1 At the heart of DSSC is wide
band gap metal oxide semiconductor (typically, TiO2)
anchored to a molecular sensitizer. Among the various types of
dye sensitizers for DSSCs, Ru-poly-pyridyl-complex type sensi-
tizers have shown solar-to electrical power conversion efficien-
cies up to 11% under standard solar illumination (AM 1.5G).2
While the performances of DSSCs based on organic dyes have
not exceeded those based on Ru-based dyes, organic dyes have
been adopted as sensitizers for DSSCs due to their practical
advantages including (i) large molar absorption coefficients
resulting from intramolecular π–π* transition, (ii) simple syn-
thesis, color tuning, and structural modification for desired
physical and photochemical properties, (iii) their economy.3
Efforts devoted to developing promising metal-free sensi-
tizers have shown that structural modifications that impart the
Fiber and Polymer Science Program, North Carolina State University, Raleigh, NC
27695, USA. E-mail: bkim6@ncsu.edu, hfreeman@ncsu.edu
the high efficiency of DSSCs. For example, extended conju-
gation gives the red-shifted absorption spectrum needed for
good light harvesting but can also induce aggregation and/or a
positive shift in the LUMO of the dye molecules as well. There-
fore, a systematic study involving the design of DSSC dyes for
enhanced efficiency is still of interest. To date, the structures
of sensitizers in DSSCs have been largely based on push–pull
structures for efficient charge separation.3a,b Regarding such
structures, variations in the donor group have been studied.4
For example, Hagberg et al.4c reported the effects of alkoxy-sub-
stituted donor groups on photovoltaic properties and showed
that this structural feature plays an important role in control-
ling aggregation. Similarly, Ko and co-workers showed that a
twisted non-planar geometry plays a significant role in the
control of aggregation and high rates of charge separation.5
Tuning of a julolidine based donor5 to bulky stable bis-
dimethylfluoreneaniline unit afforded enhanced efficiency.6
As part of a study aimed at investigating molecular struc-
ture-performance relationships in DSSCs, we report the syn-
thesis and photo/electrochemical, kinetics properties and
photovoltaic performance of D–π-A dye sensitizers having car-
bazole (CA) and diphenylamine (DP) donors in combination
with aromatic spacers. The CA dyes (Fig. 1) contained the

This journal is © The Royal Society of Chemistry and Owner Societies 2013 Photochem. Photobiol. Sci., 2013, 12, 421–431 | 421

Paper
Published on 13 November 2012 on http://pubs.rsc.org | doi:10.1039/C2PP25262G
Downloaded by University of Massachusetts - Amherst on 24 March 2013
Fig. 1 Synthetic pathway for CA series.
9-(4-Methoxyphenyl)-N-(9-(4-methoxyphenyl)-9H-carbazol-6-yl)-
9H-carbazol-3-amine (CA donor) and an anchoring unit of cya-
noacrylic acid connected by different spacers: namely CA-T,
CA-P, and CA-B for thiophene, N-methyl pyrrole, and
benzene spacer, respectively. A previously studied DP series
(DP-T, DP-P, and DP-B)7 was employed in the present study to
investigate the effects of donor groups on photovoltaic
performance.
Experimental
All reagents were purchased from Sigma-Aldrich, TCI, or
Fisher Scientific. All chemicals were of reagent-grade quality
and solvents purchased from commercial suppliers were used
without further purification. Moisture-sensitive reactions were
performed under either nitrogen or argon gas.
422 | Photochem. Photobiol. Sci., 2013, 12, 421–431
View Article Online
Photochemical & Photobiological Sciences
Synthesis
Synthetic procedures for DP based dyes and key intermediates
(1a–3a, 2a–2d, and 3a–3c in Fig. 1) can be found elsewhere.7
9-(4-Methoxyphenyl)-9H-carbazole (I).8 Anisole (9.1 g,
38.9 mmol), carbazole (5 g, 29.9 mmol), K2CO3 (33.1 g,
239.2 mmol), copper powder (7.6 g, 119.6 mmol), and
18-crown-6 (0.74 g, 1.8 mmol) were added into flask under Ar gas
followed by adding o-dichlorobenzene (50 ml). The reaction
mixture was stirred under for 24 h and mixture was allowed to
cool to room temperature before filtering copper and inorganic
salts. The crude product was collected by evaporation of
solvent and recrystallized from methanol to yield I as a white
solid. Yield: 84%, 1H NMR (300 MHz, DMSO-d6) δ = 8.23 (d,
2H, J = 7.50 Hz), 7.53–7.50 (m, 2H), 7.44–7.39 (m, 2H),
7.29–7.20 (m, 6H), 3.87 (s, 3H).
3-Bromo-9-(4-methoxyphenyl)-9H-carbazole (II).9 To a stirred
solution of I (0.8 g, 2.93 mmol) in tetrahydrofuran at 0 °C was
This journal is © The Royal Society of Chemistry and Owner Societies 2013

Photochemical & Photobiological Sciences
added N-bromosuccinimide (0.36 g, 2.05 mmol) in small por-
tions. The reaction mixture was then allowed to warm to room
temperature and stir overnight. The solvent was evaporated
and the residue was extracted with ether. Crude II was col-
lected by ether evaporation and recrystallized from hexane to
produce a white solid. Yield: 67%, 1H NMR (300 MHz DMSO-
d6) δ = 8.49 (d, 1H, J = 2.1 Hz), 8.29 (dd, 1H, J = 8.4 Hz, 0.9 Hz),
7.54 (dd, 1H, J = 8.7 Hz, 2.1 Hz), 7.51 (AA′BB′, 2H), 7.44 (td,
1H, J = 7.2 Hz, 1.2 Hz), 7.31–7.20 (m, 5H), 3.87 (s, 3H).
9-(4-Methoxyphenyl)-3-nitro-9H-carbazole (III). To a stirred
solution of I (4.5 g, 16.46 mmol) in acetic acid (70 ml) was
added urea nitrate (2.03 g, 16.4 mmol) and stirring continued
Published on 13 November 2012 on http://pubs.rsc.org | doi:10.1039/C2PP25262G
at 70 °C. The precipitate formed by addition of water was col-
lected by filtration and dried. Pure III was obtained by recrys-
Downloaded by University of Massachusetts - Amherst on 24 March 2013
tallization from acetic acid, resulting in a yellow solid. Yield:
89%, 1H NMR (300 MHz, DMSO-d6) δ = 9.28 (d, 1H, J = 1.8 Hz),
8.51 (d, 1H, J = 7.8 Hz), 8.31 (dd, 1H, J = 9 Hz, 2.1 Hz),
7.59–7.52 (m, 3H), 7.42–7.23 (m, 5H), 3.89 (s, 3H).
9-(4-Methoxyphenyl)-9H-carbazol-3-amine (IV).10 III (0.4 g,
1.26 mmol), Sn(II) chloride dihydrate (2.83 g, 12.5 mmol), and
ethanol (10 ml) were combined and reaction mixture was
stirred under reflux for 48 h under N2 gas flow. The reaction
mixture was allowed to cool to room temperature followed by
evaporation of solvent. NaOH (40%) solution was added to the
reaction mixture while the residue cooled in an ice bath and
stirred vigorously. The resulting solid was collected by fil-
tration and extracted with toluene several times and the com-
bined extracts were dried using anhydrous MgSO4. Pure IV was
obtained by recrystallization from toluene, resulting in an off-
white solid. Yield: 75%, 1H NMR (300 MHz, DMSO-d6) δ = 8.00
(dd, 1H, J = 7.8 Hz, 0.6 Hz), 7.45 (from AA′BB′, 2H), 7.31 (td, J =
7.35 Hz, 0.9 Hz), 7.23–7.12 (m, 4H), 7.04 (d, 1H, J = 8.4 Hz),
6.77 (dd, 1H, J = 8.7 Hz, 2.1 Hz), 4.85 (s, 2H), 3.33 (s, 3H).
9-(4-Methoxyphenyl)-N-(9-(4-methoxyphenyl)-9H-carbazol-
6-yl)-9H-carbazol-3-amine (V). Tris(dibenzylideneacetone)-
dipalladium(0) (0.14 g, 0.15 mmol), 2-dicyclohexylphosphino-
2′,4′,6′-triisopropylbiphenyl (0.28 g, 0.59 mmol), sodium tert-
butoxide (0.99 g, 0.33 mmol), II (2.6 g, 7.38 mmol), IV (2.13 g,
7.38 mmol) were added to a flask followed by addition of
toluene (30 ml) using a syringe. The flask was sealed under a
positive pressure of Ar, using a Teflon screw cap, and heated at
80 °C for 24 h. The solution was cooled to room temperature
and diluted by adding ethyl acetate. The mixture was filtered
through Celite and the solvent was removed by evaporation.
Purification of crude V by column chromatography (DCM–
hexane, 3 : 1, v/v) on silica gel produced a pale greenish yellow
solid. Yield: 45%, 1H NMR (300 MHz, DMSO-d6) δ = 8.11 (d,
2H, J = 7.8 Hz), 7.93 (s, 1H), 7.88 (d, 2H, J = 0.6 Hz), 7.52 (from
AA′BB′, 4H), 7.37 (td, 2H, J = 7.50 Hz, 0.9 Hz), 7.28–7.16 (m,
12H), 3.87 (s, 6H).
9-(4-Methoxyphenyl)-N-(9-(4-methoxyphenyl)-9H-carbazol-6-yl)-
N-(3,3-dimethyl-2-(thiophen-2-yl)-3H-indol-5-yl)-9H-carba-
zol-3-amine (1c). Tris(dibenzylideneacetone) dipalladium(0)
(0.02 g, 0.02 mmol), 2-dicyclohexylphosphino-2′,4′,6′-triisopro-
pylbiph-enyl (0.04 g, 0.09 mmol), sodium tert-butoxide (0.16 g,
1.69 mmol), 1b (0.37 g, 1.20 mmol), V (0.67 g, 1.20 mmol)
This journal is © The Royal Society of Chemistry and Owner Societies 2013
View Article Online
Paper
were stirred under Ar gas and 12 ml of toluene was added
using a syringe. The flask was sealed under a positive pressure
of argon with a Teflon screw cap and stirring was continued at
80 °C. The solution was cooled to room temperature and
diluted by adding ethyl acetate. The diluted mixture was fil-
tered through Celite and was evaporated. Purification by
column chromatography (ethyl acetate–hexane, 1 : 2, v/v) on
silica gel produced a brownish yellow solid. Yield: 66%, 1H
NMR (300 MHz, DMSO-d6) δ = 8.11 (d, 2H, J = 7.5 Hz), 8.05 (s,
2H), 7.72–7.70 (m, 2H), 7.52 (AA′BB′, 4H), 7.39–7.33 (m, 3H),
7.27–7.13 (m, 9H), 7.19 (AA′BB′, 4H), 7.10 (d, 1H, J = 1.8 Hz),
6.78 (dd, 1H, J = 8.1 Hz, 2.1 Hz), 3.84 (s, 6H), 1.40 (s, 6H). 13C
NMR (500 MHz, DMSO-d6) δ = 158.49, 148.48, 147.86, 146.36,
141.15, 140.93, 137.58, 137.50, 129.65, 129.35, 128.61, 128.49,
128.11, 126.27, 125.02, 123.48, 122.23, 120.62, 120.31, 119.62,
117.34, 115.31, 112.99, 110.64, 109.56, 55.46, 52.71, 24.45.
5-(5-(9-(4-Methoxyphenyl)-N-(9-(4-methoxypheny l)-9H-carba-
zol-6-yl)-9H-carbazol-3-amino)-3,3-dimethyl-3H-indol-2-yl)thio-
phene-2-carbaldehyde (1d). A solution of 1c (0.4 g, 0.51 mmol)
in THF (15 ml) was cooled to −78 °C while the system was
purged with N2. n-BuLi (0.79 ml, 1.27 mmol) was added drop-
wise over 10 min and the reaction mixture was stirred at
−78 °C for 1 h. The solution kept at −78 °C for a dropwise
addition of dry DMF (0.20 ml, 2.55 mmol). The reaction
mixture was stirred for additional 1 h at −78 °C before allowing
it warm to 0 °C and the solution was neutralized using 10%
HCl. The crude product obtained from extractions with DCM
was dried over anhydrous MgSO4 and concentrated under
vacuum. Purification was performed by column chromato-
graphy (ethyl acetate–hexane, 1 : 2, v/v) on silica gel to give a
violet-red solid. Yield: 80%, 1H NMR (700 MHz, DMSO-d6) δ =
9.96 (s, 1H), 8.14 (d, 2H, J = 7.7 Hz, 0.7 Hz), 8.12 (dd, 2H, J =
2.1 Hz, 0.7 Hz), 8.03 (d, 1H, J = 4.2 Hz), 7.88 (d, 1H, J = 3.5 Hz),
7.54 (AA′BB′, 4H), 7.45 (d, 1H, J = 8.4 Hz), 7.40 (ddd, 2H, J = 8.8
Hz, 6.8 Hz, 0.7 Hz), 7.33–7.21 (m, 8H), 7.21 (AA′BB′, 4H), 7.11
(d, 1H, J = 2.8 Hz), 6.81 (dd, 1H, J = 8.4 Hz, 2.1 Hz), 3.87 (s,
6H), 1.43 (s, 6H).
(2E)-2-Cyano-3-(5-(5-(9-(4-methoxyphenyl)-N-(9-(4-methoxy-
phenyl)-9H-carbazol-6-yl)-9H-carbazol-3-amino)-3,3-dimethyl-
3H-indol-2-yl)thiophen-2-yl)acrylic acid (CA-T). To a solution
of compound 1d (0.068 g, 0.08 mmol) and 2-cyanoacetic acid
(0.008 g, 0.10 mmol) in acetonitrile (3 ml) was added a few
drops of piperidine under N2 gas flow. The reaction mixture was
then stirred under reflux at 80 °C. The precipitate formed
during reaction was collected by filtration and dried. Pure CA-T
was obtained by column chromatography (DCM–methanol,
11 : 1, v/v) on silica gel, to give a reddish violet solid. Yield: 24%,
1
H NMR (300 MHz, DMSO-d6) δ = 8.15 (d, 2H, J = 7.8 Hz), 8.107
(s, 2H), 8.105 (s, 1H), 7.79 (d, 1H, J = 3.9 Hz), 7.73 (d, 1H, J =
3.9 Hz), 7.55 (AA′BB′, 4H), 7.45 (d, 1H, J = 8.1 Hz), 7.40 (td, 2H,
J = 8.1 Hz, J = 1.2 Hz), 7.31–7.16 (m, 12H), 7.12 (d, 1H, J =
1.8 Hz), 6.81 (dd, 1H, J = 8.4 Hz, 2.1 Hz),) 3.87 (s, 6H), 1.43 (s,
6H). 13C NMR (500 MHz, DMSO-d6) δ = 174.46, 163.09, 160.28,
158.63, 148.84, 148.48, 142.40, 141.75, 141.16, 140.67, 138.87,
137.68, 136.15, 129.31, 129.31, 128.84, 128.18, 126.34, 125.18,
123.51, 122.24, 120.97, 120.71, 119.64, 118.97, 117.67, 115.27,
Photochem. Photobiol. Sci., 2013, 12, 421–431 | 423

Paper
112.44, 110.74, 109.57, 55.42, 52.85, 24.24. HRMS-ESI: calcu-
lated for C56H41N5O4S (M + H)/z 880.2952, Found: 880.2940.
5-(5-(9-(4-Methoxyphenyl)-N-(9-(4-methoxyphenyl)-9H-carbazol-
6-yl)-9H-carbazol-3-amino)-3,3-dimethyl-3H-indol-2-yl)-1-methyl-
1H-pyrrole-2-carbaldehyde (2e). Compound 2e was prepared
using the same procedure as for 1c except that 2d was used in
lieu of 1b and Cs2CO3 was used as a base. The crude product
collected after evaporation of solvent was purified by column
chromatography on silica gel (ethyl acetate–hexane, 1 : 3, v/v)
to yield a yellowish brown solid, Yield: 31%, 1H NMR
(300 MHz, DMSO-d6) δ = 10.42 (s, 1H), 8.12 (d, 2H, J = 7.5 Hz),
8.06 (d, 2H, J = 1.5 Hz), 7.54 (AA′BB′, 4H), 7.41–7.36 (m, 3H),
Published on 13 November 2012 on http://pubs.rsc.org | doi:10.1039/C2PP25262G
7.29–7.12 (m, 9H), 7.21 (AA′BB′, 4H), 7.11 (d, 1H, J = 2.4 Hz),
6.81 (dd, 1H, J = 8.4 Hz, 2.1 Hz), 6.75 (d, 1H, J = 2.7 Hz), 3.87
Downloaded by University of Massachusetts - Amherst on 24 March 2013
(s, 6H), 1.37 (s, 6H).
(2E)-2-Cyano-3-(5-(5-(9-(4-methoxyphenyl)-N-(9-(4-methoxyphenyl)-
9H-carbazol-6-yl)-9H-carbazol-3-amino)-3,3-dimethyl-3H-indol-2-yl)-
1-methyl-1H-pyrrol-2-yl)acrylic acid (CA-P). To a stirred sol-
ution of 2e (0.07 g, 0.09 mmol) and 2-cyanoacetic acid
(0.015 g, 0.18 mmol) in ethanol (12 ml) was added NaOH
(0.017 g, 0.42 mmol) under N2 gas flow. The reaction mixture
was then stirred under reflux at 80 °C. The mixture formed
during reaction was filtered and product was dried. Pure CA-P
was obtained by column chromatography (DCM–methanol,
10 : 1, v/v) on silica gel, to yield a red brown solid. Yield: 53%,
1
H NMR (300 MHz, DMSO-d6) δ = 8.11–8.07 (m, 4H), 7.92 (s,
1H), 7.53 (AA′BB′, 4H), 7.38 (ddd, 2H), 7.35–7.09 (m, 15H), 6.81
(dd, 1H, J = 8.4 Hz, 2.4 Hz), 6.70 (dd, 1H, J = 3.4 Hz, 1.5 Hz),
3.86 (s, 6H), 3.34 (s, 3H), 1.36 (s, 6H). 13C NMR (500 MHz,
DMSO-d6) δ = 177.62, 158.49, 147.77, 147.69, 147.03, 141.24,
141.11, 137.36, 129.32, 128.80, 128.16, 127.69, 126.22, 124.84,
123.41, 122.18, 120.61, 119.93, 119.57, 119.43, 119.27, 117.06,
115.29, 113.33, 110.59, 109.50, 108.96, 55.38, 53.39, 35.97,
24.26. HRMS-ESI: calculated for C57H44N6O4 (M + H)/z
877.3497, Found: 877.3485.
4-(5-(9-(4-Methoxyphenyl)-N-(9-(4-methoxyphenyl)-9H-carbazol-
6-yl)-9H-carbazol-3-amino)-3,3-dimethyl-3H-indol-2-yl)benzaldehyde
(3d). Compound 3d was prepared using the same procedure
as for 2e except that 3c was used in lieu of 2d. The crude
product collected after evaporation of solvent was purified by
column chromatography on silica gel (ethyl acetate–hexane,
1 : 2, v/v) to yield a red solid, Yield: 30%, 1H NMR (300 MHz,
DMSO-d6) δ = 10.07 (s, 1H), 8.32 (d, 2H, J = 8.7 Hz), 8.15–8.12
(m, 4H), 8.00 (d, 2H, J = 8.4 Hz), 7.54 (AA′BB′, 4H), 7.48 (d, 1H,
J = 8.4 Hz), 7.40 (t, 2H, J = 8.1 Hz), 7.31–7.16 (m, 12H), 7.12 (d,
1H, J = 2.1 Hz), 6.83 (dd, 1H, J = 8.4 Hz, 2.4 Hz), 3.87 (s, 6H),
1.46 (s 6H). 13C NMR (500 MHz, DMSO-d6) δ = 192.73, 178.24,
158.65, 149.68, 148.85, 145.82, 141.17, 140.62, 137.86, 137.70,
136.47, 129.69, 129.29, 128.19, 128.15, 126.34, 125.29, 123.51,
122.22, 121.42, 120.70, 119.70, 118.64, 117.74, 115.32, 112.10,
110.75, 109.57, 55.46, 52.94, 24.00.
(2E)-2-Cyano-3-(4-(5-(9-(4-methoxyphenyl)-N-(9-(4-methoxyphenyl)-
9H-carbazol-6-yl)-9H-carbazol-3-amino)-3,3-dimethyl-3H-indol-2-yl)-
phenyl)acrylic acid (CA-B). CA-B was prepared using the same
procedure as for CA-P except that 3d was used as a starting
material instead of 2e. Pure CA-B were obtained by column
424 | Photochem. Photobiol. Sci., 2013, 12, 421–431
View Article Online
Photochemical & Photobiological Sciences
chromatography on silica gel (DCM–methanol, 10 : 1, v/v) to
produce a red solid, Yield: 36%, 1H NMR (300 MHz, DMSO-d6)
δ = 8.21 (d, 2H, J = 9 Hz), 8.15–8.10 (m, 4H), 7.90 (d, 2H, J =
8.7 Hz), 7.96 (s, 1H), 7.54 (AA′BB′, 4H), 7.46 (d, 1H, J = 8.4 Hz),
7.39 (td, 2H, J = 8.1 Hz, 1.2 Hz), 7.31–7.16 (m, 12H), 7.11 (d,
1H, J = 2.1 Hz), 6.82 (dd, 1H, J = 8.7 Hz, 2.4 Hz), 3.87 (s, 6H),
1.46 (s, 6H). 13C NMR (500 MHz, DMSO-d6) δ = 178.56, 158.62,
149.50, 148.50, 145.90, 141.14, 140.76, 137.69, 134.47, 129.59,
128.22, 128.04, 126.30, 125.21, 123.46, 122.20, 121.19, 120.70,
119.66, 118.99, 188.81, 117.67, 115.29, 122.33, 110.71, 109.53,
55.41, 52.76, 24.11. HRMS-ESI: calculated for C58H43N5O4
(M + H)/z 874.3388, Found: 874.3396.
Analytical instrumentation and measurements
1
H NMR and 13C NMR spectra were recorded on a Bruker
Advance 300 or 500 MHz spectrometer using DMSO-d6 or
CDCl3 as a solvent. HRMS (high resolution mass spectra) were
obtained using electrospray ionization (ESI) on an Agilent
Technologies (Santa Clara, California) 6210 LC-TOF mass spec-
trometer with accuracy <5 ppm. UV-Vis spectra in THF and
TiO2 film were recorded on a Varian Cary 300 UV-Vis spectro-
photometer. Photoluminescence of dye solutions was recorded
using a fluorometer (Fluorolog®-3, HORIBA, USA) equipped
with a 450 W xenon lamp (FL-1039/40, HORIBA, USA). Dye
loads on the photoanode were removed by dissolving dyes
adsorbed on the TiO2 film from 0.1 M NaOH in THF–H2O (v/v,
1 : 1). Thickness of dye loaded TiO2 film for all characteriz-
ations was 6 μm.
The oxidation potentials of dye adsorbed on TiO2 films
were measured by cyclovoltammetry using a three electrode
electrochemical cell at a scan rate of 100 mV s−1. TBAPF6
(tetrabutylammonium hexafluorophosphate, 0.1 M) in CH3CN
was used as a supporting electrolyte. The excited oxidation
potential was calculated from Eox − E0–0. Dye coated TiO2 film
was used as a working electrode and Ag/Ag+ electrode and
Pt wire were employed as a reference and counter electrode,
respectively. The potential of the working electrode was
calibrated with Fc/Fc+ couple. Electrochemical impedance
spectroscopy (EIS) was recorded with an impedance analyzer
connected to a potentiostat (reference 600TM, Gamry instru-
ments, USA) in a frequency range of 0.1 Hz–105 Hz under dark
condition. The applied forward bias was −0.65 V and AC
amplitude set to 10 mV.
Theoretical studies
DFT (density functional theory)11 calculations were conducted
with Gaussian 03 software. Geometry was optimized using
B3LYP hybrid functional (Beck’s three-parameter functional
and Lee–Yang–Parr functional) as an exchange–correlation
functional12 and 3-21G(d) was used as the basis set. HOMO
and LUMO orbitals were determined using optimized geome-
tries. Transition energies were calculated using TDDFT (time-
dependent density functional theory).13 B3LYP and BHandH
(designed by Becke with 50% of LSDA exchange)12c were
implemented to study electronic transitions at 3-21G(d) and
6-31G(d). The polarizable continuum model (PCM)14 was
This journal is © The Royal Society of Chemistry and Owner Societies 2013
 View Article Online

Photochemical & Photobiological Sciences Paper

performed to investigate solvent effects on λmax values. The

polarizabilites and dipole moments were calculated by single
point calculation based on the optimized geometries.

DSSCs fabrication15
To prepare photoanode, a thin layer of TiO2 anatase nanoparti-
cles (Ti-Nanoxide D/SP, Solaronix, Switzerland) was first coated
on pretreated FTO glass, with Ti(iv) bis(ethylacetoacetato)-di-
isopropoxide solution (7.5% in BuOH), followed by deposition
of a scattering layer of TiO2 nanoparticles (Ti-Nanoxide R/SP,
Solaronix, Switzerland) by doctor-blade method. The resulting
thin film was composed of a 12 μm thick TiO2 nanocrystalline
Published on 13 November 2012 on http://pubs.rsc.org | doi:10.1039/C2PP25262G

layer and a 6 μm thick scattering layer, which was measured by

a profilometer. The prepared TiO2 electrode was immersed
Downloaded by University of Massachusetts - Amherst on 24 March 2013

into 0.3 mM dye solution in THF and kept at room tempera-

ture for 24 h. Cells were built with and without 1 mM DCA
(deoxycholic acid) as coadsorbate. The photo active area of the
TiO2 electrode was 0.5 × 0.4 cm. A photocathode was prepared
by coating a drop of H2PtCl6 solution (2 mg Pt in 1 ml ethanol)
on punctured FTO glass. The dye loaded TiO2 electrode and
thermally platinized counter electrode were sealed together
with a 60 μm hot-melt gasket (Suryln 1702, Dupont). After
injecting electrolyte into a cell, cell was encapsulated using
sealant (Suryln 1702, Dupont, 60 μm thick) and a cover glass.
Two types of electrolytes were employed, which were composed
of 0.6 M DMPII (1,2-dimethyl-3-n-propylimidazolium iodide),
0.1 M LiI, 0.05 M I2 or 0.6 M DMPII (1,2-dimethyl-3-n-propyli-
midazolium iodide), 0.1 M LiI, 0.05 M I2, 0.5 M TBP (4-tert
butyl pyridine) in anhydrous CH3CN.
Photovoltaic measurements
Photocurrent–voltage characteristics of DSSCs were measured
with a Keithley 2400 source meter under illumination of AM
1.5 G solar light coming from a solar simulator (SOL3A, Oriel)
equipped with a 450 W xenon lamp (91160, Oriel). The inci-
dent light intensity was calibrated using a reference Si solar
cell (Newport Oriel, 91150V) to set 1 sun (1 mW cm−2). The
current–voltage curve of the cell was obtained by applying
external voltage bias and then measuring the generated photo-
current. A black mask (0.6 × 0.5 cm) was covered on a testing
cell to control illumination area. IPCE (incident monochro-
matic photon to current conversion efficiency) was measured
on a system (QEX10, PV Measurements, USA) equipped with a
75 W short arc xenon lamp (UXL-75XE, USHIO, Japan) as a
light source connected to a monochrometer. Monochromatic
quantum efficiency was recorded at short circuit condition
under AC mode with white-light bias. The AC chopping speed
was set to 10 Hz.
Results and discussion
Optical properties
The absorption spectra of CA and DP-based dyes in THF are
shown in Fig. 2 and the characteristic data are listed in
Table 1. All dyes gave more than two absorption bands in the
Fig. 2 Absorption spectra (a) of CA (solid lines) and DP (dashed lines) based
dyes, (upper; in solution, bottom; on TiO2). Optimized geometries for dyes DP-T
(b) and CA-(c).
UV/visible region. The ICT (internal charge transfer) band was
reported to be dependent not only on the acceptor (or donor)
strength of the molecules but also solvents used.16 Localized
π–π* band was affected by conjugation path length. Provided
that the differences in the electron pushing/pulling effects of
each heterocyclic unit in the D–π-A system are considered, the
more sensitive band shift involves the band close to ICT band.
As can be seen in Table 1, λmax values obtained using THF–
H2O (v/v, 1 : 1) as a solvent versus THF alone were almost
unchanged in UV region but differed in the visible region,
indicating that the latter are due to charge transfer (CT). The
observed solvatochromism (here, hypsochromism) probably
stems from deprotonation of the carboxylic group, which
reduces the electron pulling effect of the acceptor group. Com-
pared to DP-based dyes, dyes having the carbazole donor gave
a bathochromic shift by an average of ∼40 nm, leading to a
preferred absorption spectrum for efficient light harvesting of
solar energy. The red-shifts caused by changing the donor
from a diphenylamine to a carbazole type amine probably
arise from more planar geometries and increased polarizabil-
ities. From MO calculations (cf. Fig. 2b–c), the CA-T dyes
showed a dihedral angle of 25° between the CA unit and
indole moiety, and the corresponding angle in DP-T was 34°.
Concerning the relative energies corresponding to the CT
among CA dyes, λmax was either hypso or batho shifted
depending on the electronegativity of heteroatom rather than
its delocalization energy. CA-T showed the most red-shifted
absorption spectrum (λmax = 513 nm). On the other hand, CP-P

This journal is © The Royal Society of Chemistry and Owner Societies 2013 Photochem. Photobiol. Sci., 2013, 12, 421–431 | 425
 View Article Online

Paper Photochemical & Photobiological Sciences

Table 1 Spectroscopic and electrochemical properties obtained from CA-dyes

Dyes λmax a/nm εmax a/M−1 cm−1 λmax_emia/nm λmax b/nm λmax_TiO2 c/nm λmax_TiO2_w/D c/nm Eox d [V] vs. NHE Eox* e [V] vs. NHE

CA-T 367, 513 25 165, 26 013 688 365, 499 519 514 1.079, 1.568 −0.988, −0.499
CA-P 318, 417 56 128, 19 567 606 318, 414 N/A N/A 0.864, N/A −1.656, N/A
CA-B 318, 477 59 237, 26 046 661 316, 465 480 477 1.094, 1.550 −1.108, −0.652
a
Absorption and emission spectra recorded in THF. b Absorption spectra recorded in THF–water (v/v, 1 : 1). c λmax values of dye and dye-loaded
TiO2 film, respectively. d Oxidation potentials. e Excited state oxidation potentials (Eox*) = Eox − E0–0, where E0–0 is intercept of absorption and
emission spectrum.

afforded a 36 nm hypsochromic shift relative to CA-T. This

strong dependence of λmax on the electronic nature of hetero-
Published on 13 November 2012 on http://pubs.rsc.org | doi:10.1039/C2PP25262G

atom is consistent with our previous work involving DP-based

dyes,7 where the use of heterocylic rings for π conjugation
Downloaded by University of Massachusetts - Amherst on 24 March 2013

resulted in increased/decreased energy for CT. In the case of

absorption strength in visible region, CA dyes conferred
increased εmax (19 576–26 046) compared to the corresponding
DP dyes (εmax = 17 062–25 491).
Dye adsorption on the TiO2 surface gave rise to broadened
and shifted spectra due to dye–dye and/or dye–TiO2 inter-
actions (Fig. 2).3f,17 Unfortunately, CA-P (DP-P) absorptions
below 420 nm on TiO2 film could not be resolved due to an
overlap with the absorption band arising from the bare TiO2
layer. However, a red-shifted absorption spectrum would be
expected from the absorption edge which was not related to
the bare TiO2 layer. Except for CA-B (DP-B), DP and CA-based
dyes seem to form low level of J-aggregates on the TiO2
surface, judging from the bathochromic shift. Based on the
typically accepted assumption that the smaller λmax shift is
associated with a lower tendency to from aggregates,2a,18 the
modest λmax shift involving CA-T (cf. DP-T) could be due to the
bulky amine donor, leading to suppression of aggregate for-
mation. However, the results were different for CA-P and CA-B
which afforded spectral changes similar to DP-P and DP-B,
respectively. For example, CA-P and DP-P produced a similar
spectral shift (∼50 nm) on TiO2 surface and in solution. Fur-
thermore, dyes having a benzene spacer showed almost
unchanged spectra on TiO2 surface relative to the solution
state regardless of donor geometry. Therefore, it can be con-
cluded that spectral changes on film state are affected by
spacer nature as well as donor geometry, giving rise to
different dye adsorption behavior on the TiO2 surface.

Modeling studies
Molecular modeling has been employed as a screening pro-
cedure to predict spectroscopic and electronic properties.3d,g,4c,19
Fig. 3 Electronic densities (a) of frontier orbitals of CA dyes. (Values in parenth-
Therefore, TDFT and time-dependent DFT (TDDFT) cal-
esis are the corresponding HOMO and LUMO values for DP dyes). (b) It shows a
culations were performed to predict ground and excited states, representative example of electron densities of DP-T (upper) and electronic
respectively. Isodensity surface plots of frontier orbitals are configurations (bottom) of DP-T and CA-T after TDDFT-PCM/3-21G(d).
shown in Fig. 3. In each case, HOMO was delocalized through
the CA donor and indole moiety while LUMO has a sizable
electron distribution involving the spacer and anchoring unit, exchange correlational functionals, B3LYP12b and BHandH,12c
indicating effective intramolecular charge transfer. As a start- were performed using the optimized geometries at the 3-21G
ing point for predicting vertical electronic transitions, TDDFT (d) or 6-31G(d) level. The following general features were found
was studied for the corresponding DP dyes. Two different as follow: (1) vertical electronic transitions were dependent

426 | Photochem. Photobiol. Sci., 2013, 12, 421–431 This journal is © The Royal Society of Chemistry and Owner Societies 2013

Photochemical & Photobiological Sciences
Table 2 Calculated absorption parameters using different exchange corre-
lation functional and basis set for DP-dyes
λmax a values from TDDFT in vacuum
Functional/basis set Dyes
DP-T DP-P DP-B
B3LYP/3-21G(d) 624(402) 562(387) 598(378)
BHandH/3-21G(d) 446(322) 415(316) 404(296)
BHandH/6-31G(d) 470(333) 428(322) 433(307)
λmax values from TDDFT-PCM/THF
BHandH/3-21G(d) 466(330) 432(325) 423(304)
BHandH/6-31G(d) 494(343) 449(333) 448(315)
Published on 13 November 2012 on http://pubs.rsc.org | doi:10.1039/C2PP25262G
a
Values and those in parenthesis are λmax values at S0 → S1 and S0 →
Downloaded by University of Massachusetts - Amherst on 24 March 2013
S2 transition state, respectively.
upon the hybrid functional employed, (2) there were two main
transitions corresponding to absorption bands in the visible
region and in the UV region around 300 nm. (Table 2). Judging
from the electron configurations (cf. Fig. 3b), S0 → S1 is
assigned to ICT transition while S0 → S2 transition is con-
sidered as a localized π–π* transition. For both series, the low
energy absorption was mainly responsible for the HOMO to
LUMO transition, whereas the high energy absorption showed
somewhat different main configuration contributions from
which DP dyes (CA dyes) showed main contribution of the
HOMO − 1 to LUMO transition (increased ratio of HOMO − 2
to LUMO). A comparison of experimental and calculated λmax
values showed that the B3LYP functional overestimated the
transition energy compared to the BHandH functional, which
was much more evident for the ICT (internal charge transfer)
band. These results are consistent with related research, where
a fraction of HF (Hartree–Fock) incorporated in the exchange
correlation functional was found to have a significant effect on
predicted λmax.20
Based on the results in Table 2, the BHandH method
having a higher HF (50%) was selected for the investigation of
solvent (THF) effects by adopting the PCM model. Two
different basis sets, 3-21G(d) and 6-31G(d), were adopted. For
small and expended basis set, small deviations below 0.2 V
(≤30 nm) from experimental ICT absorption energies were
observed except for DP-P. The expanded basis set 6-31G(d)
gave better results among the dyes studied. Consequently,
BHandH hybrid functional seems to be a good candidate for
predicting ICT bands with small deviations whereas B3LYP
could be used to determine relative trends among different
structures in a D–π-A system (Table 3).
Electrochemical properties
Oxidation potentials (Eox) of dye-loaded TiO2 film were
measured by cyclic voltammetry. All redox values for CA dyes
are summarized in Table 1. Unlike DP dyes showing one elec-
tron transfer process, CA dyes showed two oxidation peaks. A
low first oxidation potential is responsible for the oxidation of
the aryl amine moiety while the second oxidation peak
This journal is © The Royal Society of Chemistry and Owner Societies 2013
View Article Online
Paper
Table 3 Comparison of experimental and calculated λmax for CA dyes
TDDFT/BHandH/3-21G(d)
In vacuum PCM-THF
Exp’t Dyes S0 → S1 S0 → S2 S0 → S1 S0 → S2
513, 367 CA-T 477 344 497 352
417, 318 CA-P 435 330 448 338
477, 318 CA-B 437 317 450 323
observed at higher potential could arise from oxidation of the
electron donor which is incorporated into carbazole group.
The oxidation potentials of CA dyes were 1.07, 0.86, and 1.09 V
versus NHE for CA-T, CA-P, and CA-B, respectively. Notably,
CA-P showed the highest Eox due to the electron rich N-methyl
pyrrole unit. Likewise, the higher lied Eox level for CA dyes
(average 0.29 V more negative shift) compared to those of DP
dyes is responsible for the strong electron donating properties
of the CA donor. All Eox values lie below the redox potential of
I−/I3− (0.4 V versus NHE),2a,21 indicating that electronic match-
ing occurs to accept electrons from electrolyte followed by the
regeneration of oxidized dye. On the other hand, all excited
oxidation potentials were in the range of −0.99 to −1.65 V vs.
NHE for CA dyes (cf. −0.92 V and −1.06 V for DP-T and DP-B,
respectively). Based on the Eox* values obtained above Ecb (con-
duction edge of TiO2; −0.5 V versus NHE),2a photo-generated
electrons can be thermodynamically injected to into the TiO2
semiconductor. Also, it was found that the introduction of a
strong donor had a greater effect on the shift in HOMO, which
plays a role in red shifted spectrum of CA dyes arising from
the decreased energy gap between HOMO and LUMO. On the
other hand, varying the spacer unit had a greater affect on the
LUMO level rather than HOMO. The observed shifts were also
predicted from modeling studies, where CA-T shifted the
HOMO and LUMO negatively by 0.46 eV and 0.16 eV, respect-
ively, relative to DP-T.
Photovoltaic performance
The current–potential characteristic of DSSCs was obtained
under simulated AM 1.5 G irradiation (100 mW cm−2). The J–V
curves of CA series are shown in Fig. 4. Data from DP dye
based DSSCs, which were fabricated under similar conditions,
are taken from our previous research22 for a comparison. A CA
dye sensitized DSSCs showed somewhat lowered efficiencies
(η = 2.05, 1.55, 1.49% for CA-T, CA-P, CA-B, respectively) com-
pared to DSSCs from DP dyes (η = 3.53, 2.86, 3.50% for DP-T,
DP-P, DP-B, respectively). Both current densities and open-
circuit voltage parameters of CA dye based solar cells could not
exceed those of DP dye based devices. Notably, the IPCE action
spectra (Fig. 5) showed lower photocurrent densities for CA
dyes.
It is interesting to note that the maximum quantum
efficiencies from IPCE did not follow the relative hyperchro-
mism (or hypsochromism) induced from different types of
donor or spacer units.
Photochem. Photobiol. Sci., 2013, 12, 421–431 | 427

Paper
Published on 13 November 2012 on http://pubs.rsc.org | doi:10.1039/C2PP25262G
Downloaded by University of Massachusetts - Amherst on 24 March 2013
Fig. 4 IV curves of DSSCs sensitized with CA and DP dyes. (w/D and w/T mean
the devices containing DCA and TBP with electrolyte of 0.6 M DMPII/0.1 M
LiI/0.05M I2, respectively.)
For example, the plateaus of IPCE (reaching around 29%)
for devices based on the CA dyes were lower than those from
DP dye sensitized devices having corresponding values of 58,
55, 64% for DP-S, DP-P, DP-B, respectively (cf. Fig. 5a). Among
CA dyes, DSSCs sensitized with CA-P showed highest plateau
of IPCE but the spectrum was not broad enough to exceed the
Jsc (Jsc = 5.35 mA cm−2) of the CA-T sensitized DSSCs. The
highest open-circuit voltage (Voc) was obtained from CA-P, at
0.582 V. The best performance was achieved by CA-T where
Jsc = 5.35 mA cm−2, Voc = 0.564 V, ff = 0.681 resulting in an
overall conversion efficiency (η) of 2.05%.
Photovoltaic performances were influenced by the additives
used (cf. Fig. 4). By the addition of DCA (1 mM), Jsc increased
33–49% and there was a Voc gain (3.22, 7.81, 8.06% for CA-T,
CA-P, CA-B, respectively) leading to corresponding η with 2.80,
2.62, 2.24%. These current densities changes for CA dye sensi-
tized devices in the presence of DCA were more significant
than those based on the DP dyes which gave 0.2–7% increases.
The significant Jsc increase by adding DCA could arise from
the suppression of dye aggregate formation on the TiO2
surface. However, this would not account for the small shifts
in absorption spectra from CA dyes, which suggests a smaller
tendency of dye aggregate formation.2a,15 The other possible
explanation is greater effect of DCA on suppression of charge
recombination for CA dyes versus DP dyes. This was supported
by electron lifetime measurements. Furthermore, Voc gain
from CA dyes by the addition of DCA could be another reason
for suppression of charge recombination.4d,23 On the other
hand, overall efficiencies of device based on the CA dyes were
increased in the presence of TBP. This is due to the increase in
Voc (average of 17% in Voc) at the expense of slightly decreased
Jsc. In the case of CA-P, Voc increased from 0.582 to 0.691 V
but the Jsc decreased from 3.86 to 3.67 mA cm−2, giving
η of 1.84%. The slight decrease in current densities of
DSSCs based on the CA dyes (versus DP dyes) in the presence
of TBP is attributed to the larger driving force for electron
injection.
428 | Photochem. Photobiol. Sci., 2013, 12, 421–431
View Article Online
Photochemical & Photobiological Sciences
Fig. 5 IPCE curves (left) and dark currents (right) for selective DSSCs con-
structed by CA and DP dyes, (containing no additives (a,b), DCA (c), and TBP (d).
Investigation of photocurrent densities
Replacing the DP donor with the CA donor afforded a red spec-
tral shift and higher driving force for electron injection.
However, these advantages for solar energy capture of CA
series did not lead to an increased Jsc. In DSSCs, the photocur-
rent of the cell is determined by the IPCE, which is defined as
the number of electrons produced by light in the external
circuit divided by the number of photons absorbed as
follow:2b,24
1240Isc ðμA cm2 Þ
IPCEðλÞ ¼ LHEðλÞΦinj Φc ¼ LHEðλÞΦðλÞet ¼
λðnmÞPinput ðW m2 Þ
where Φet is electron transfer yield defined as a product of Φinj
(charge injection efficiency) and Φc (charge collection
efficiency) and LHE is light harvesting efficiency. LHE is
affected by the spectral properties of the dye sensitizer and
expressed by 1–10−f (f = oscillator strength). Assuming that
LHE is treated as unity for CA and DP dyes, Φet remains an
important consideration. In general, the larger the driving
force, the larger the Φinj.25 Here, driving force of electron injec-
tion (ΔGinject) can be defined as Edye* − Ecb,25b where Edye* is
the oxidation potential of excited dyes. The Edye* can be deter-
mined from either relaxed or unrelaxed geometries of excited
states. However, experimental results have been reported
which indicate the latter type of electron transfer is much
faster than vibrational relaxed electron transfer in DSSCs.20 To
determine ΔGinject, the values of −0.5 V vs. NHE and reduction
potential values were used as Ecb and Edye*, respectively. The
resultant ΔGinject values were −0.49, −1.16, and −0.61 V for
CA-T, CA-P, and CA-B, versus values of −0.42 and −0.56 V for
DP-T and DP-B, giving CA dyes higher driving forces. Nonethe-
less, this advantage of CA dyes was not enough to lead to
increased current density. It seems that electron transfer yield
is influenced by charge recombination affecting Φc as well as
increasing Eox*.26
In conjunction with the present dye molecules, two factors
might be related to lower Φc for CA dyes: (1) dye loads and (2)
charge recombination. From the comparison of relative
This journal is © The Royal Society of Chemistry and Owner Societies 2013
 View Article Online

Photochemical & Photobiological Sciences Paper

Table 4 Adsorbed dye amount and charge recombination lifetimes (τ), and calculated physical parameters

CA dyes DP dyes

CA-T CA-P CA-B DP-T DP-P DP-B

Ads.amounta 9.04 7.67 8.27 11.6 8.79 9.73

τb 2.5 20.2 10.1 8 40.0 15.9
τc 5.0 31.8 15.9 10.1 50.4 15.9
τd 20.2 79.8 15.9 25.2 63.4 40.0
Polarizabilities (Å3)e 846.2 800.5 832.6 458.9 443.7 458.8
Dipole moment(μ)e 7.86 4.88 9.09 8.62 7.76 8.84
μnormal e 7.33 4.66 7.67 4.97 2.88 3.42
a
Adsorbed amounts; unit of ×10−8 mol cm−2. b Recombination lifetimes; unit of ms. Devices having no additives. c Recombination lifetimes; unit
of ms. Devices having DCA. d Recombination lifetimes; unit of ms. Devices having TBP. e Obtained from single point calculation based on
Published on 13 November 2012 on http://pubs.rsc.org | doi:10.1039/C2PP25262G

optimized geometries; μnormal was calculated at the axis where carboxyl group of the molecule was aligned out of perpendicular to the TiO2
surface.
Downloaded by University of Massachusetts - Amherst on 24 March 2013

adsorbed amount of dyes on TiO2 film (Table 4), introduction
of the CA donor induced dye loss by an average value of 17%
compared to DP dyes. The more bulky nature of CA dyes (cf.
Fig. 2b–c) seems to influence not only dye loads but also
charge recombination, affecting charge collection efficiency
and giving lower IPCE plateaus for CA dyes. Results of previous
research showed that high dye loads lead to improved surface
blocking, resulting in less electron recombination, and con-
stantly increasing electron transfer yield. Also, a less tightly
packed dye layer could result in a more accessible TiO2
surface, resulting in increased recombination of injected elec-
trons.27 From this combination of key factors (e.g. higher εmax,
Eox*, low level of aggregation28) for IPCE, donor group geome-
try, which may affect efficient surface blocking of dye uptake
on the TiO2 surface, stands out.
Effects of charge recombination, dye load, dipole moment,
and polarizabilities on open-circuit voltage
The open-circuit voltage (Voc) in DSSCs is determined by the
absolute difference between the standard reduction potential
of redox coupling and the fermi level of the TiO2 electrode.21
The equations below shows the dependence of Voc on the elec-
tron recombination,29
 
RT AI kB T
Voc ¼ ln ¼ Ered  E cb  γ
βF n0 kb ½I  þ n 0 k r ½D þ e n
3
where β represents the order for I3− and electrons, I is incident
photon flux, A is electrode area, n0 is concentration of accessi-
ble electronic states in the Ecb, and kb and kr are rate constant
for recombination of electrons with I3− and oxidized dye,
respectively. The right side of the equation shows the depen-
dence of Voc on Ecb and electron density in TiO2. Here, γ is a
characteristic constant of TiO2 tailing states, kB is Boltzmann
constant, T is absolute temperature, e is elementary charge,
and Ne is the effective density states at Ecb edge. Even though
the adsorption behavior of a dye sensitizer on the TiO2 surface
can change the surface charge on TiO2,30 Ecb is assumed to be
−0.5 V versus NHE for similar DSSC fabrication conditions. In
practice, the interfacial charge recombination has been
reported to have a greater effect on Voc than Ecb changes
caused by different types of dyes.4d,31 As an approach to study-
ing kinetics in the present DSSCs, EIS (electrochemical impe-
dance spectroscopy)32 was conducted in the dark using −0.65 V
of forward bias. The electron recombination lifetimes (τ),
extracted from the reciprocal of highest frequency of Nyquist
plot, are listed in Table 4. In conjunction with the equation
shown above, by ignoring n0 kr [D+] term followed by approxi-
mation of 1/kr to electron lifetime (τ),33 Voc is logarithmically
proportional to τ. Specifically, Lifetime (τ) is expressed as a
product of chemical capacitance (C) and recombination resist-
ance (Rrec). Here, the EIS measurements used to obtain Rrec or
τ involve a similar number of electrons. In cases leading to
possible Ecb changes of TiO2 induced by different types of sen-
sitizers, Ecb shifts can be estimated from C as a function of
open circuit voltage.34 The present τ values were not deter-
mined using adjusted Ecb values. Although Voc results include
a contribution from the Ecb position, the Ecb of TiO2 should
not be a determinant on Voc for the present dye sensitizers. If
Ecb shifts were largely affected by the polarity of the present
dyes, the larger dyes would exhibit higher polarity and induce
a more negative Ecb shift than the smaller dye system, resulting
in a Voc increase.27 However, in each case, CA type dyes having
a two-fold higher polarizability gave a lower Voc relative to DP
  types. While the present Voc values do not reflect a precise esti-
Ne
ln mation about the effect of the Ecb change on lifetime among
the investigated dyes, τ values without adjustment of Ecb posi-
tions permit a useful relative comparison of the present dye
sensitizers.35 The obtained τ values for DSSCs based on CA
dyes were shorter than those sensitized by DP dyes. Specifi-
cally, CA-T (DP-T), CA-P (DP-T), and CA-B (DP-B) recorded τ of
2.5 (8), 20 (40), and 10 (16) ms, respectively. Concerning the
charge resistance (Rct), CA dyes showed lower Rct. For example,
CA-T showed Rct of 25 Ω compared to 40 Ω for DP-T (Fig. 6a).
Results indicating smaller Rct, and shorter τ, provide evi-
dence of increased interfacial charge recombination, leading
to lower Voc values for CA-based dyes versus DP-based dyes.
Increased charge recombination for CA dye based devices was
also supported by dark current measurements (Fig. 5b). The
onset voltages at 1 mA cm−2 for DP dyes based devices were
higher than those for devices based on CA dyes. Furthermore,

This journal is © The Royal Society of Chemistry and Owner Societies 2013 Photochem. Photobiol. Sci., 2013, 12, 421–431 | 429

Paper
Fig. 6 Nyquist phase plots of CA and DP dyes based DSSCs (a) and CA-P based
DSSCs depending on additives (b).
Published on 13 November 2012 on http://pubs.rsc.org | doi:10.1039/C2PP25262G
Rct well followed the relative trend for Voc among different
Downloaded by University of Massachusetts - Amherst on 24 March 2013
types of spacers. For example, CA-N having the largest Rct
among CA dyes gave the highest Voc (cf. Fig. 6b)
Recently, it was reported that the origin of different electron
recombination kinetics is related to the molecular size, surface
blocking,31,36 and/or interaction of dyes with oxidizing species
in electrolyte.27 Judging from our study, higher dye load on the
TiO2 electrode and/or polarizabilities can enhance electron
recombination lifetimes. From the calculated dipole moments
(μ) and polarizabilities (Table 4), polarizability seems to corre-
late better with τ values. The CA-P and DP-P, which recorded
highest τ and Voc, showed the lowest polarizability. Also, mod-
eling studies on CA dyes having a bulky donor predicted an
almost 2-fold higher polarizability compared to those of DP-
based dyes, which might have negative effects on τ. However,
another recent study emphasized the effects of a bulky
umbrella-shaped donor on suppressing charge recombina-
tion.4d In that case, the larger protected area arising from the
bulky donor compensated for a lower dye loads, giving
increased electron lifetime, and consequently, producing a
higher Voc. In our case, efficient surface blocking was not
evident from the introduction of the CA donor. It is likely that
molecular conformation plays a significant role in the surface
blocking behavior of dye molecules on TiO2 surface. Further-
more, the number of hetero atoms on the donor group
might increase the local concentration of I3−, resulting in a
lower Voc.
In conclusion, new three sensitizers containing a bis-carba-
zole (CA) donor in a D–π-A system were synthesized and com-
pared with diphenyl amine based analogs (DP series) for their
use in DSSCs. The effects of heterocyclic rings as a π conju-
gation unit in the CA-based dyes were compared to DP-based
dyes having the same heterocyclic rings as π units. From the
modeling studies, TDDFT-PCM calculations based on the
BHandH functional proved reliable for predicting λmax. The
higher excited oxidation potential and better spectral coverage
for solar capture for devices based on CA dyes versus DP-based
dyes were not enough to compensate for lower dye loads and
recombination electron lifetimes, resulting in lower DSSC
efficiency. When employing a bulky donor unit in a D–π-A
system to give a red shift, its conformation should be con-
sidered, to give effective surface blocking leading to high
efficiency DSSCs.
430 | Photochem. Photobiol. Sci., 2013, 12, 421–431
View Article Online
Photochemical & Photobiological Sciences
References
1 B. O’Regan and M. Gratzel, Nature, 1991, 353, 737.
2 (a) A. Hagfeldt and M. Graetzel, Chem. Rev., 1995, 95, 49;
(b) M. K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-
Baker, E. Mueller, P. Liska, N. Vlachopoulos and
M. Graetzel, J. Am. Chem. Soc., 1993, 115, 6382; (c) P. Péchy,
T. Renouard, S. M. Zakeeruddin, R. Humphry-Baker,
P. Comte, P. Liska, L. Cevey, E. Costa, V. Shklover,
L. Spiccia, G. B. Deacon, C. A. Bignozzi and M. Grätzel,
J. Am. Chem. Soc., 2001, 123, 1613; (d) A. Hagfeldt and
M. Grätzel, Acc. Chem. Res., 2000, 33, 269;
(e) M. K. Nazeeruddin, F. De Angelis, S. Fantacci, A. Selloni,
G. Viscardi, P. Liska, S. Ito, B. Takeru and M. Grätzel, J. Am.
Chem. Soc., 2005, 127, 16835; (f ) M. Grätzel, Inorg. Chem.,
2005, 44, 6841; (g) P. Wang, C. Klein, R. Humphry-Baker,
S. M. Zakeeruddin and M. Grätzel, J. Am. Chem. Soc., 2004,
127, 808.
3 (a) X. Ma, J. Hua, W. Wu, Y. Jin, F. Meng, W. Zhan and
H. Tian, Tetrahedron, 2008, 64, 345; (b) M. Velusamy,
K. R. Justin Thomas, J. T. Lin, Y.-C. Hsu and K.-C. Ho, Org.
Lett., 2005, 7, 1899; (c) K. Hara, T. Sato, R. Katoh,
A. Furube, Y. Ohga, A. Shinpo, S. Suga, K. Sayama,
H. Sugihara and H. Arakawa, J. Phys. Chem. B, 2002, 107,
597; (d) D. P. Hagberg, T. Marinado, K. M. Karlsson,
K. Nonomura, P. Qin, G. Boschloo, T. Brinck, A. Hagfeldt
and L. Sun, J. Org. Chem., 2007, 72, 9550; (e) K. Hara,
M. Kurashige, S. Ito, A. Shinpo, S. Suga, K. Sayama and
H. Arakawa, Chem. Commun., 2003, 252; (f ) T. Kitamura,
M. Ikeda, K. Shigaki, T. Inoue, N. A. Anderson, X. Ai,
T. Lian and S. Yanagida, Chem. Mater., 2004, 16, 1806;
(g) C. Kim, H. Choi, S. Kim, C. Baik, K. Song, M.-S. Kang,
S. O. Kang and J. Ko, J. Org. Chem., 2008, 73, 7072;
(h) H. Zollinger, Color Chemistry: Syntheses, Properties,
and Applications of Organic Dyes and Pigments, Verlag
Helvetica Chimica Acta, Wiley-VCH, Zürich Weinheim,
2003.
4 (a) Z. Ning, Q. Zhang, W. Wu, H. Pei, B. Liu and H. Tian,
J. Org. Chem., 2008, 73, 3791; (b) S. W. Park, K.-I. Son,
M. J. Ko, K. Kim and N.-G. Park, Synth. Met., 2009, 159,
2571; (c) D. P. Hagberg, J.-H. Yum, H. Lee, F. De Angelis,
T. Marinado, K. M. Karlsson, R. Humphry-Baker, L. Sun,
A. Hagfeldt, M. Grätzel and M. K. Nazeeruddin, J. Am.
Chem. Soc., 2008, 130, 6259; (d) M. Lu, M. Liang, H.-Y. Han,
Z. Sun and S. Xue, J. Phys. Chem. C, 2011, 115, 274;
(e) T. Horiuchi, H. Miura and S. Uchida, Chem. Commun.,
2003, 3036.
5 H. Choi, J. K. Lee, K. H. Song, K. Song, S. O. Kang and
J. Ko, Tetrahedron, 2007, 63, 1553.
6 S. Kim, J. K. Lee, S. O. Kang, J. Ko, J. H. Yum, S. Fantacci,
F. De Angelis, D. Di Censo, M. K. Nazeeruddin and
M. Grätzel, J. Am. Chem. Soc., 2006, 128, 16701.
7 B. H. Kim, PhD thesis, North Carolina State University,
2012.
8 C.-C. Lee, M.-k. Leung, P.-Y. Lee, T.-L. Chiu, J.-H. Lee,
C. Liu and P.-T. Chou, Macromolecules, 2012, 45, 751.
This journal is © The Royal Society of Chemistry and Owner Societies 2013

Photochemical & Photobiological Sciences
9 C.-L. Ho, W.-Y. Wong, B. Yao, Z. Xie, L. Wang and Z. Lin,
J. Organomet. Chem., 2009, 694, 2735.
10 S. Hayashida, M. Taya and M. Morita, Jap. Pat. 08041327,
Composition for Active Optical Waveguide, Hitachi Chemi-
cal Co Ltd, Japan, 1996, pp. 11.
11 E. Runge and E. K. U. Gross, Phys. Rev. Lett., 1984, 52,
997.
12 (a) C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens.
Matter, 1988, 37, 785; (b) A. D. Becke, J. Chem. Phys.,
1993, 98, 1372; (c) A. D. Becke, J. Chem. Phys., 1993, 98,
5648.
13 (a) N. N. Matsuzawa, A. Ishitani, D. A. Dixon and T. Uda,
Published on 13 November 2012 on http://pubs.rsc.org | doi:10.1039/C2PP25262G
J. Phys. Chem. A, 2001, 105, 4953; (b) R. E. Stratmann,
G. E. Scuseria and M. J. Frisch, J. Chem. Phys., 1998, 109,
Downloaded by University of Massachusetts - Amherst on 24 March 2013
8218.
14 J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005,
105, 2999.
15 S. Ito, T. N. Murakami, P. Comte, P. Liska, C. Grätzel,
M. K. Nazeeruddin and M. Grätzel, Thin Solid Films, 2008,
516, 4613.
16 (a) P. Shen, X. Liu, S. Jiang, Y. Huang, L. Yi, B. Zhao and
S. Tan, Org. Electron., 2011, 12, 1992; (b) Y.-S. Yen,
Y.-C. Hsu, J. T. Lin, C.-W. Chang, C.-P. Hsu and D.-J. Yin,
J. Phys. Chem. C, 2008, 112, 12557.
17 (a) T. Horiuchi, H. Miura, K. Sumioka and S. Uchida, J. Am.
Chem. Soc., 2004, 126, 12218; (b) D. P. Hagberg,
T. Edvinsson, T. Marinado, G. Boschloo, A. Hagfeldt and
L. Sun, Chem. Commun., 2006, 2245; (c) G. Ramakrishna
and H. N. Ghosh, J. Phys. Chem. A, 2002, 106, 2545.
18 (a) C. Teng, X. Yang, C. Yang, H. Tian, S. Li, X. Wang,
A. Hagfeldt and L. Sun, J. Phys. Chem. C, 2010, 114, 11305;
(b) S. Roquet, A. Cravino, P. Leriche, O. Alévêque, P. Frère
and J. Roncali, J. Am. Chem. Soc., 2006, 128, 3459.
19 (a) S. Paek, H. Choi, H. Choi, C.-W. Lee, M.-s. Kang,
K. Song, M. K. Nazeeruddin and J. Ko, J. Phys. Chem. C,
2010, 114, 14646; (b) K. Srinivas, K. Yesudas,
K. Bhanuprakash, V. J. Rao and L. Giribabu, J. Phys. Chem.
C, 2009, 113, 20117.
20 J. Preat, C. Michaux, D. Jacquemin and E. A. Perpète,
J. Phys. Chem. C, 2009, 113, 16821.
21 L. M. Peter, PCCP, 2007, 9, 2630.
22 B. H. Kim and H. S. Freeman, J. Mater. Chem., 2012, 22,
20403.
This journal is © The Royal Society of Chemistry and Owner Societies 2013
View Article Online
Paper
23 K.-M. Lee, S.-J. Wu, C.-Y. Chen, C.-G. Wu, M. Ikegami,
K. Miyoshi, T. Miyasaka and K.-C. Ho, J. Mater. Chem.,
2009, 19, 5009.
24 H. Tian, X. Yang, R. Chen, R. Zhang, A. Hagfeldt and
L. Sun, J. Phys. Chem. C, 2008, 112, 11023.
25 (a) R. Katoh, A. Furube, T. Yoshihara, K. Hara, G. Fujihashi,
S. Takano, S. Murata, H. Arakawa and M. Tachiya, J. Phys.
Chem. B, 2004, 108, 4818; (b) J. Zhang, H.-B. Li, S.-L. Sun,
Y. Geng, Y. Wu and Z.-M. Su, J. Mater. Chem., 2012, 22, 568.
26 (a) N. J. Cherepy, G. P. Smestad, M. Grätzel and J. Z. Zhang,
J. Phys. Chem. B, 1997, 101, 9342; (b) M. K. R. Fischer,
S. Wenger, M. Wang, A. Mishra, S. M. Zakeeruddin,
M. Grätzel and P. Bäuerle, Chem. Mater., 2010, 22, 1836.
27 T. Marinado, K. Nonomura, J. Nissfolk, M. K. Karlsson,
D. P. Hagberg, L. Sun, S. Mori and A. Hagfeldt, Langmuir,
2010, 26, 2592.
28 K. Hara, Z.-S. Wang, T. Sato, A. Furube, R. Katoh,
H. Sugihara, Y. Dan-oh, C. Kasada, A. Shinpo and S. Suga,
J. Phys. Chem. B, 2005, 109, 15476.
29 (a) M. Koelsch, S. Cassaignon, C. Ta Thanh Minh,
J. F. Guillemoles and J. P. Jolivet, Thin Solid Films, 2004,
451–452, 86; (b) A. Usami, S. Seki, Y. Mita, H. Kobayashi,
H. Miyashiro and N. Terada, Sol. Energy Mater. Sol. Cells,
2009, 93, 840.
30 H. Ishii, K. Sugiyama, E. Ito and K. Seki, Adv. Mater., 1999,
11, 605.
31 Y. Liang, B. Peng and J. Chen, J. Phys. Chem. C, 2010, 114,
10992.
32 (a) Q. Wang, J.-E. Moser and M. Grätzel, J. Phys. Chem. B,
2005, 109, 14945; (b) A. J. Bard and L. R. Faulkner, Electro-
chemical Methods: Fundamentals and Applications, Wiley,
New York, 2nd edn, 2001.
33 M. Adachi, M. Sakamoto, J. Jiu, Y. Ogata and S. Isoda,
J. Phys. Chem. B, 2006, 110, 13872.
34 (a) J. Bisquert, F. Fabregat-Santiago, I. n. Mora-Seró,
G. Garcia-Belmonte and S. Giménez, J. Phys. Chem. C, 2009,
113, 17278; (b) E. M. Barea, V. Gónzalez-Pedro, T. Ripollés-
Sanchis, H.-P. Wu, L.-L. Li, C.-Y. Yeh, E. W.-G. Diau and
J. Bisquert, J. Phys. Chem. C, 2011, 115, 10898.
35 T. Ono, T. Yamaguchi and H. Arakawa, Sol. Energy Mater.
Sol. Cells, 2009, 93, 831.
36 Z. Ning, Q. Zhang, H. Pei, J. Luan, C. Lu, Y. Cui and
H. Tian, J. Phys. Chem. C, 2009, 113, 10307.
Photochem. Photobiol. Sci., 2013, 12, 421–431 | 431