Single and double branched organic dyes based on carbazole and red-absorbing cationic indolium for p-type dye-sensitized solar cells: A combined experimental and theoretical investigation

Article abstract of DOI:10.1016/j.dyepig.2017.09.042

A series of novel single and double branched carbazole-based red-absorbing cationic dyes without (CSI and CDI) and with (CST and CDT) a thiophene spacer have been synthesized for p-type dye-sensitized solar cells (p-DSSCs). The introduction of the red-absorbing cationic acceptor/thiophene spacer led to the broadening and bathochromic shift of the absorption maxima from an onset of 600 nm for CSI to 800 nm for CDT as well as improved molar absorptivity. The introduction of the double branching/thiophene spacer lowered the lowest unoccupied molecular orbital (LUMO) levels of CDI, CST, and CDT, making their potentials more positive than that of CSI. Among these, the double branched CDI exhibited the highest conversion efficiency of 0.112%. Furthermore, all of the dyes examined outperformed the standard C343 dye (0.062%), measured under similar fabrication conditions. Despite the decreased photovoltaic performance as a result of the introduction of the thiophene spacer, overall, the double branched dyes exhibited better interfacial charge transfer that led to higher J_SC and V_OC values compared to those of singly branched dyes. Electrochemical impedance spectroscopy analysis showed that double branched dyes have much lower charge transfer resistance and increased hole lifetime than single branched dyes. Density functional theory (DFT) and time-dependent DFT calculations were performed to theoretically characterize the optical and electrochemical properties of the synthesized dyes.

Full text of DOI:10.1016/j.dyepig.2017.09.042

Accepted Manuscript
Single and double branched organic dyes based on carbazole and red-absorbing
cationic indolium for p-type dye-sensitized solar cells: A combined experimental and
theoretical investigation
Le Quoc Bao, Phuong Ho, Ramesh Kumar Chitumalla, Joonkyung Jang, Suresh
Thogiti, Jae Hong Kim
PII:
S0143-7208(17)31558-9
10.1016/j.dyepig.2017.09.042
DYPI 6267
DOI:
Reference:
To appear in: Dyes and Pigments
Received Date: 18 July 2017
Accepted Date: 17 September 2017
Please cite this article as: Bao LQ, Ho P, Chitumalla RK, Jang J, Thogiti S, Kim JH, Single and double
branched organic dyes based on carbazole and red-absorbing cationic indolium for p-type dye-
sensitized solar cells: A combined experimental and theoretical investigation, Dyes and Pigments
(2017), doi: 10.1016/j.dyepig.2017.09.042.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical Abstract
Single and double branched organic dyes based on carbazole and red-absorbing cationic
indolium for p-type dye-sensitized solar cells: a combined experimental and theoretical
investigation
Le Quoc Bao, Phuong Ho, Ramesh Kumar Chitumalla, Joonkyung Jang, Suresh Thogiti* and Jae
Hong Kim*
Doule branched dyes showed better interfacial charge transfer and hole lifetime over single
branched dyes leading to enhanced photovoltaic performance.

ACCEPTED MANUSCRIPT
Single and double branched organic dyes based on carbazole and
red-absorbing cationic indolium for p-type dye-sensitized solar cells:
a combined experimental and theoretical investigation
1
,2
1
3
3
1
Le Quoc Bao, Phuong Ho, Ramesh Kumar Chitumalla, Joonkyung Jang, Suresh Thogiti*
1
and Jae Hong Kim*
1
School of Chemical Engineering, Yeungnam University, 214-1, Dae-dong, Gyeongsan,
Gyeongbuk 712-749, Republic of Korea.
2
Faculty of Applied Sciences, Ton Duc Thang University, 19 Nguyen Huu Tho Str., Tan Phong
Ward, District 7, Ho Chi Minh City 700000, Vietnam
3
Department of Nanoenergy Engineering, Pusan National University, Busan, 609-735, Republic
of Korea
*
Prof. Suresh Thogiti, Department of Chemical Engineering, Yeungnam University, Korea
(ROK). Email: sureshyu@ynu.ac.kr, Tel : (+82)53- 810-3537, Fax : (+82)53-810-4631.
*
Prof. Jae Hong Kim, Department of Chemical Engineering, Yeungnam University, Korea
(
ROK). Email : jaehkim@yumail.ac.kr, Tel : (+82)53- 810-3537, Fax : (+82)53-810-4631.

ACCEPTED MANUSCRIPT
Abstract:
A series of novel single and double branched carbazole-based red-absorbing cationic dyes
without (CSI and CDI) and with (CST and CDT) a thiophene spacer have been synthesized for p-
type dye-sensitized solar cells (p-DSSCs). The introduction of the red-absorbing cationic
acceptor/thiophene spacer led to the broadening and bathochromic shift of the absorption
maxima from an onset of 600 nm for CSI to 800 nm for CDT as well as improved molar
absorptivity. The introduction of the double branching/thiophene spacer lowered the lowest
unoccupied molecular orbital (LUMO) levels of CDI, CST, and CDT, making their potentials
more positive than that of CSI. Among these, the double branched CDI exhibited the highest
conversion efficiency of 0.112%. Furthermore, all of the dyes examined outperformed the
standard C343 dye (0.062%), measured under similar fabrication conditions. Despite the
decreased photovoltaic performance as a result of the introduction of the thiophene spacer,
overall, the double branched dyes exhibited better interfacial charge transfer that led to higher JSC
and VOC values compared to those of singly branched dyes. Electrochemical impedance
spectroscopy analysis showed that double branched dyes have much lower charge transfer
resistance and increased hole lifetime than single branched dyes. Density functional theory
(DFT) and time-dependent DFT calculations were performed to theoretically characterize the
optical and electrochemical properties of the synthesized dyes.
Key words: p-type dye-sensitized solar cells; carbazole; cationic dye; charge transfer;
photocurrent density.

ACCEPTED MANUSCRIPT
Introduction
Dye-sensitized solar cells (DSSCs), developed by Grätzel et al. in the 1990s, are an attractive
source of renewable energy and are typically perceived as a low-cost alternative to conventional
silicon-based solar cells [1–3]. Two types of DSSCs exist, namely, DSSCs that employ n-type
2
semiconductors such as TiO as the photo-anode (n-DSSCs) [4], and their inverse model, p-
DSSCs, which employ p-type semiconductors such as NiO [5]. Various solar cells have been
prepared using n-type sensitizers, attaining enhanced power conversion efficiencies (PCEs) [4,6–
1
2]. The dye sensitizers typically utilized in these n-DSSCs can absorb only in the visible or
near-infrared (NIR) region of the solar spectrum. Currently, it is still challenging to synthesize
dyes that are capable of covering both the visible and NIR regions. On the other hand, p-type
devices have rarely been studied and their PCEs are generally significantly lower than those of n-
type DSSCs [5,13–23]. The investigation of p-type DSSCs is motivated by the desire to combine
them with traditional n-type DSSCs into tandem pn-DSSCs (TDSSCs), which can provide higher
photo-voltage and potentially also a high PCE. Until now, however, the efficiencies of TDSSCs
have been hampered by the poor performance of p-type DSSCs [24–28].
The operational principles of p-type DSSCs are illustrated in Fig. 1. In contrast to the n-type
DSSC, the hole in the p-type DSSC is injected into the semiconductor and an electron is
transferred to the electrolyte through a photo-excited sensitizer [5,29]. Among all of the
components present, the sensitizer plays a key role in capturing photons, generating electron and
hole pairs, and subsequently transferring them. For this reason, the design and development of
new dye materials is in high demand. Owing to the fact that the efficiency of DSSCs depends
significantly on the optimization and compatibility of each of its constituents, the sensitizers are
expected to capture as many photons from the sunlight as possible. With the objective of

ACCEPTED MANUSCRIPT
developing new sensitizers for p-type DSSCs, we synthesized in this study four sensitizers (CSI-
CDT) based on a well-known aromatic heterocycle, namely carbazole, which has been shown to
provide potentially increased efficiencies in n-DSSCs [30–35]. The red-absorbing cationic 1-
octyl-2,3,3-trimethyl-3H-indolium iodide was used as the acceptor in order to facilitate the
fabrication of p-type DSSCs with improved performance, owing to the recent reports that
showed that this compound can enable dyes to absorb long-wavelength visible light [36–39].
Earlier efforts to shift the absorption towards longer wavelengths by increasing the number of
cyano moieties have led to diminished photocurrents in p-DSSCs. This outcome was attributed to
the decrease in the driving force for dye regeneration, which resulted from the decrease in the
level of the lowest unoccupied molecular orbital (LUMO). By introducing an indolium cationic
acceptor, Wood et al. succeeded in maintaining sufficient driving force for dye regeneration
while increasing the absorption in the red region [36].
Organic dyes with donor–acceptor-based double anchoring/acceptor were reported to display
improved efficiencies in n- and p-type DSSCs [37,40–45]. In 2014, Park et al. synthesized a
donor–acceptor carbazole-based MCBZ (herein referred to as CSI) and DCBZ dye sensitizers for
p-DSSCs, and obtained enhanced conversion efficiency of 0.035% for the double branched di-
anchoring dye DCBZ. Taking into consideration the above described points, the single (CSI and
CST) and double branched (CDI and CDT) carbazole-based photosensitizers are applied to p-
type DSSCs and compared with standard C343 dye under similar fabrication conditions. The
experimental results of UV-vis spectroscopy, cyclic voltammetry (CV), and electrochemical
impedance spectroscopy (EIS), together with the incident photo-to-current conversion
efficiencies (IPCEs), and the outcome of theoretical density functional theory (DFT) and time-

ACCEPTED MANUSCRIPT
dependent DFT (TDDFT) calculations were examined to explain their structure–property
relationships.
2
2
. Experimental
.1 Materials
1
All H NMR spectra were recorded on a Varian Mercury NMR 300 MHz spectrometer using
1
CDCl or DMSO-d as the solvent (purchased from Alfa Aesar). The H NMR chemical shifts
3
6
were referenced relative to TMS. The absorption spectra were recorded on an Agilent 8453 UV-
vis spectrophotometer. The active areas of the dye-absorbed NiO films were estimated using a
digital microscope camera equipped with an image-analysis-software (Moticam 1000). The
photovoltaic I-V characteristics of the prepared DSSCs were measured under 1 sunlight intensity
2
(
100 mW/cm , AM 1.5), which was verified with an AIST-calibrated Si-solar cell (PEC-L11,
Peccell Technologies, Inc., City, State, Country). The incident monochromatic photon-to-current
efficiencies (IPCEs) were plotted as a function of the light wavelength using an IPCE
measurement instrument (PEC-S20, Peccell Technologies, Inc.). The redox properties of three
dyes were examined using cyclic voltammetry (Model: CV-BAS-Epsilon). The electrolyte
solution employed in this analysis was 0.10 M tetrabutylammonium hexafluorophosphate
(
TBABF ) in freshly dried acetonitrile. The Ag/AgCl and Pt wire (0.5 mm in diameter)
6
electrodes were used as the reference and counter electrodes, respectively. The scan rate was set
at 30 mV/s. Electrochemical impedance spectroscopy (EIS) was performed using an electronic-
chemical analyzer (Iviumstat Tec.).

ACCEPTED MANUSCRIPT
2
.2. Preparation of sensitizers
Some of the starting compounds, i.e., 1-bromohexane, 1-iodooctane, pyridine, and methanol
were obtained from Sigma-Aldrich, whereas others such as carbazole, N,N-dimethylformamide
(
DMF), 4-hydrazinobenzoic acid, 3-methyl-2-butanone, 2-(tri-n-butylstannyl)thiophene, trans-
dichlorobis(triphenyl-phosphine)palladium(II) (PdCl (PPh ) were purchased from Alfa Aesar.
Phosphorus oxychloride (POCl ) was obtained from DeaJung Co., Ltd., and all other solvents
and components were purchased from Duksan Pure Chemicals Co., Ltd.
-Carboxyl-2,3,3-trimethyl-3H-indoline (1): a mixture of 4-hyrazinobenzoic acid (10 g, 65.72
2
3 2
)
3
5
mmol), 3-methyl-2-butanone (8.49 g, 98.58 mmol) and sodium acetate (10.78 g, 131.449 mmol)
in acetic acid were refluxed under nitrogen at 120 °C for 12 h. The resulting brown suspension
was cooled to room temperature, the solvent was removed under reduced pressure, and a mixture
of MeOH:H O (1:9) was added to the residue. The flask was cooled on ice to induce
2
precipitation. The obtained residue was filtered and washed several times with hexane and
1
diethyl ether to give 1 as a white powder (8.62 g, 64.57%). H NMR (DMSO-d , 300 MHz) δ
6
1
2.79 (s, 1H), 7.99 (s, 1H), 7.93 (d, J = 7.8 Hz, 1H), 7.52 (d, J = 7.8 Hz, 1H), 2.25 (s, 3H), 1.28
s, 6H).
-Carboxyl-2,3,3-trimethyl-1-octyl-3H-indoline, iodide (1:1) (2): 5-carboxyl-2,3,3-trimethyl-3H-
(
5
indoline 1 (5.5 g, 36.14 mmol) was dissolved in acetonitrile (100 mL) and refluxed at 80 °C
under nitrogen atmosphere and 1-iodooctane (21.12 g, 87.95 mmol) was added to this solution.
After 72 h, the flask was cooled to room temperature and the solvent was removed under reduced
pressure. The crude product was recrystallized from a mixture of acetonitrile and diethyl ether.
The precipitate was filtered and washed several times with ether, hexane, and ethyl acetate to

ACCEPTED MANUSCRIPT
1
give 2 as a white solid (4.68 g, 40%). H NMR (DMSO-d
6
, 300 MHz) δ 8.40 (s, 1H), 8.17 (d, J =
7
.8 Hz, 2H), 4.48 (s, 2H), 2.89 (s, 3H), 1.84 (s, 2H), 1.58 (s, 6H), 1.33 (m, 10H), 0.85 (s, 3H).
-Hexylcarbazole (3): Carbazole (15 g, 89.70 mmol) was dissolved in DMF (100 mL) and 1-
9
bromohexane (23.90 g, 144 mmol) was added directly to the solution. The mixture was stirred
under nitrogen at 80 °C for 6 h. After cooling to room temperature, the solvent was removed
under reduced pressure. Next, MeOH was added to the solution and the mixture was kept on ice
1
to yield the product 3 as a white crystalline precipitate (16.24 g, 72.75%). H NMR (Acetone-d ,
6
3
00 MHz) δ 8.17 (d, J = 7.8 Hz, 2H), 7.60 (d, J = 8.1 Hz, 2H), 7.50 (t, J = 8.1 Hz, 2H), 7.24 (t, J
7.8 Hz, 2H) 4.47 (t, J = 7.2 Hz, 2H), 1.94 (p, J = 7.5 Hz, 2H), 1.45 (m, 6H), 0.89 (t, J = 6.9 Hz,
=
3
H).
9
-Hexylcarbazole-3-carboxaldehyde (4): 9-hexylcarbazole 3 (6 g, 23.88 mmol) was dissolved in
1
,2-dichloroethane (50 mL, DCE) in a three-necked round-bottom flask and the mixture was
cooled on an ice bath. A mixture of DMF (2.8 g, 38.3 mmol) and POCl (8 g, 52.17 mmol) in
3
EtOH was added to the solution. The reaction was stirred at 80 °C under nitrogen. After 6 h, the
reaction was cooled to room temperature and an aqueous solution of NaOH was added slowly to
quench the reaction. The organic fraction was extracted with chloroform, dried over anhydrous
2
CaCl , filtered, and the solvent was removed under reduced pressure. The residue was purified
by silica gel column chromatography using hexane:DCM (1:1) as the eluent. The desired product
1
4
8
1
was obtained as a white solid (5.3 g, 80.34%). H NMR (CDCl
3
, 300 MHz) δ 10.09 (s, 1H),
.61 (s, 1H), 8.17 (d, J = 7.8 Hz, 1H), 8.02 (d, J = 8.4 Hz, 1H), 7.56 (m, 3H), 7.35 (t, J = 7.5 Hz,
H), 4.35 (t, J = 7.2 Hz, 2H), 1.91 (t, J = 7.2 Hz, 2H), 1.39 (m, 6H), 0.88 (t, J = 6.9 Hz, 3H).
9
-Hexylcarbazole-3-[2-(5-carboxyl-2,3,3-trimethyl-1-octyl-3H-indoline)ethenyl], iodide (CSI): a
mixture of 9-hexylcarbazole-3-carboxaldehyde 4 (0.5 g, 1.79 mmol) and 5-carboxyl-2,3,3-

ACCEPTED MANUSCRIPT
trimethyl-1-octyl-3H-indoline 1 (1.26 g, 1.6 equiv.) was dissolved in MeOH, and refluxed under
nitrogen for 7 h at 70 °C. After this time, the solution was cooled to room temperature and the
solvent was removed under reduced pressure. The residue was recrystallized from diethyl ether,
and the precipitate was washed several times with ethyl acetate to give CSI as a red solid (0.92 g,
1
7
8
7
2.93%). H NMR (DMSO-d , 300 MHz) δ 13.36 (s, 1H), 9.13 (s, 1H), 8.82 (d, J = 15.6 Hz, 1H),
6
.43 (d, J = 6.6 Hz, 2H), 8.25 (d, J = 7.2 Hz, 1H), 8.19 (d, J = 8.4 Hz, 1H), 7.99 (d, J = 9 Hz, 1H),
.87 (d, J = 8.7 Hz, 1H), 7.77–7.69 (t, 2H), 7.61 (t, J = 7.5 Hz, 1H), 7.41 (t, J = 7.5 Hz, 1H), 4.75
(
t, J = 6.9 Hz, 2H), 4.54 (t, J = 6.9 Hz, 2H), 1.90 (m, 10H), 1.45 (m, 16H), 0.82 (q, 6H).
-Hexylcarbazole-3,6-dicarboxaldehyde (5): 9-hexylcarbazole-3-carboxaldehyde 4 (2 g, 7.96
mmol) was dissolved in DCE (50 mL) and kept at 0 °C while DMF (16 g, 2.18 mmol) and POCl3
20 g, 13.04 mmol) were added slowly. After 2 h of stirring under nitrogen, the solution was
9
(
heated to 90 °C and maintained at this temperature for 7 h. The reaction was cooled to room
temperature and quenched by the slow addition of NaOH (200 mL, 2 M) solution. The organic
2
fractions were extracted with chloroform, dried over anhydrous CaCl , and filtered. Subsequently,
the solvent was removed under reduced pressure and the crude residue was purified by silica gel
column chromatography using a mixture of hexane:DCM (1:1) as the eluent. The final product 5
1
was obtained as a white solid (0.67 g, 27.27%). H NMR (CDCl , 300 MHz) δ 10.14 (s, 2H),
3
8
6
.68 (s, 2H), 8.11 (d, J = 8.1 Hz, 2H), 4.41 (t, J = 7.2 Hz, 2H), 1.94 (m, J = 7.5 Hz, 2H), 1.43 (m,
H), 0.89 (t, J = 7.2 Hz, 3H).
9
H- Carbazole, 9-hexyl-3,6-bis-[2-(5-carboxyl-2,3,3-trimethyl-1-octyl-3H-indoline)ethenyl]-,
iodide (CDI): 9-hexylcarbazole-3,6-dicarboxaldehyde 5 (0.5 g, 1.62 mmol) and 5-carboxyl-
,3,3-trimethyl-1-octyl-3H-indoline, iodide (1:1) 2 (2 g, 4.51 mmol) were dissolved in MeOH.
The subsequent reaction steps were identical to those described for the synthesis of CSI. The
2

ACCEPTED MANUSCRIPT
1
target CDI was obtained as a dark red solid (0.93 g, 49.36%). H NMR (DMSO-d
6
, 300 MHz) δ
1
3.37 (s, 2H), 9.24 (s, 2H), 8.82 (d, J = 15.6 Hz, 2H), 8.54 (d, J = 7.8 Hz, 2H), 8.45 (s, 2H), 8.21
(d, J = 8.4 Hz, 2H), 8.04 (t, J = 8.4 Hz, 4H), 7.85 (d, J = 15.6 Hz, 2H), 4.80 (t, J = 6.9 Hz, 4H),
4
.63 (t, J = 6.9 Hz, 2H), 1.92 (m, 18H), 1.47–1.21 (m, 26H), 0.82 (q, 9H).
3
-Bromo-9-hexylcarbazole (6): NBS (2.76 g, 15.52 mmol) was dissolved in a mixture of DCE
(80 mL) and THF (20 mL) and added drop-wise to the solution 9-hexylcarbazole 3 (3 g, 11.94
mmol) dissolved in DCE (50 mL) in a three-necked round-bottom flask and stirred under
nitrogen from 0 °C to ambient temperature. After 6 h, the solution was extracted with DCM,
dried over anhydrous CaCl , and filtered. The solvent was removed under reduced pressure to
2
yield the crude product, which was purified by silica gel column chromatography with 90%
1
hexane:DCM as the eluent to give 6 as a white solid (3.2 g, 81.41%). H NMR (Acetone-d , 300
6
MHz) δ 8.41–8.19 (q, 2H), 7.62–7.49 (m, 4H), 7.28 (t, J = 7.5 Hz, 1H), 4.48 (t, J = 7.2 Hz, 2H),
1
.91 (m, J = 6.6 Hz, 2H), 1.31 (m, 6H), 0.88 (t, J = 6.6 Hz, 3H).
9
-Hexylcarbazole-3-(2-thienyl) (7): 3-bromo-9-hexylcarbazole 6 (2.6 g, 7.9 mmol), 2-(tri-n-
butylstannyl)thiophene (4.7 g, 12.6 mmol) and PdCl (PPh ) (0.55 g, 0.79 mmol) were dissolved
2
3 2
in pyridine (30 mL) and refluxed at 90 °C under nitrogen. After 18 h, the mixture was cooled to
room temperature, the organic components were extracted with ethyl acetate, dried over
2 4
anhydrous Na SO , and filtered. The residue obtained after the solvent was evaporated under
reduced pressure was purified by silica gel column chromatography with 5–10% ethyl acetate in
1
hexane as the eluent. The final product 7 was acquired as a yellow solid (0.83 g, 31.53%). H
NMR (CDCl , 300 MHz) δ 8.34 (d, J = 9 Hz, 1H), 8.14 (d, J = 7.2 Hz, 1H), 7.74 (d, J = 8.4 Hz,
3
1
1
H), 7.50 (t, J = 7.8 Hz, 1H), 7.41–7.33 (m, 3H), 7.24 (d, J = 7.5 Hz, 2H), 7.12 (t, J = 3.6 Hz,
H), 4.32 (t, J = 7.2 Hz, 2H), 1.92 (m, J = 7.2 Hz, 2H), 1.42 (m, 6H), 0.89 (t, J = 7.2 Hz, 3H).

ACCEPTED MANUSCRIPT
5
-(9-hexyl-9H-carbazole-3-yl)thiophene-2-carbaldehyde (8): 9-hexylcarbazole-3-(2-thienyl) 7
(
0.5 g, 1.5 mmol) was dissolved in DCE (20 mL) and stirred in a three-necked round-bottom
flask from 0 °C to ambient temperature under nitrogen atmosphere. A mixture of DMF (0.15 g,
.1 mmol) and POCl (0.46 g, 3 mmol) in DCE (20 mL) was added slowly to this solution and
2
3
the mixture was maintained for 1 h before being heated to 70 °C. After 2 h, the solution was
cooled to room temperature and an aqueous solution of NaOH (2 M) was added slowly to quench
2
the reaction. The mixture was extracted with DCM, dried over anhydrous CaCl , filtered, and the
solvent removed under reduced pressure. The crude product was purified by silica gel column
chromatography using 10% ethyl acetate in hexane as the eluent to give 8 as a yellow solid (0.36
1
g, 66.41%). H NMR (Acetone-d , 300 MHz) δ 9.93 (s, 1H), 8.63 (s, 1H), 8.29 (d, J = 7.2 Hz,
6
1
1
H), 7.97–7.89 (q, 2H), 7.70–7.61 (m, 3H), 7.54 (t, J = 7.2 Hz, 1H), 4.50 (t, J = 7.2 Hz, 2H),
.93 (m, J = 7.5 Hz, 2H), 1.42 (m, 6H), 0.87 (t, J = 6.9 Hz, 3H).
5
-(9-hexyl-9H-carbazole-3-yl)thiophene-2-[5-carboxyl-2,3,3-trimethyl-1-octyl-3H-indoline],
iodide (CST): 5-(9-hexyl-9H-carbazole-3-yl)thiophene-2-carbaldehyde 8 (0.3 g, 0.83 mmol) was
dissolved in DCE (30 mL) and mixed with 5-carboxyl-2,3,3-trimethyl-1-octyl-3H-indoline,
iodide (1:1) 2 (0.55 g, 1.24 mmol) dissolved in MeOH (70 mL) in a three-necked round-bottom
flask. The mixture was refluxed at 70 °C for 12 h under nitrogen atmosphere. After this time, the
mixture was cooled to room temperature. The subsequent work-up and purification steps were
identical to those described for the preparation of CSI. The final product CST was obtained as a
1
dark blue solid (0.3 g, 45.93%). H NMR (DMSO-d
6
, 300 MHz) δ 8.85–8.73 (t, 2H), 8.40 (s, 1H),
8
7
.32 (d, J = 3.9 Hz, 1H), 8.27 (d, J = 7.5 Hz, 1H), 8.17 (d, J = 8.4 Hz, 1H), 7.98–7.93 (m, 3H),
.78 (d, J = 8.7 Hz, 1H), 7.69 (d, J = 8.4 Hz, 1H), 7.56 (t, J = 7.2 Hz, 1H), 7.32–7.20 (m, 2H),

ACCEPTED MANUSCRIPT
4
6
.64 (t, J = 7.2 Hz, 2H), 4.48 (t, J = 7.2 Hz, 2H), 1.85 (m, 10H), 1.44–1.21 (m, 17H), 0.83 (t, J =
.9 Hz, 6H).
3
,6-Dibromo-9-hexylcarbazole (9): NBS (8.5 g, 47.76 mmol) was dissolved in a mixture of DCE
(80 mL) and THF (20 mL) and added to a solution 9-hexylcarbazole 3 (3 g, 11.94 mmol) in DCE
(50 mL). The mixture was stirred under nitrogen, while the temperature was allowed to increase
from 0 °C to ambient temperature over 6 h. After this time, the mixture was extracted with DCM,
and the organic layer was dried over CaCl and filtered. The residue obtained after the solvent
2
was removed under reduced pressure was purified using silica gel column chromatography with
1
5
0% DCM:hexane mixture as the eluent to give 9 as a white solid (4.2 g, 86.4%). H NMR
(
CDCl , 300 MHz) δ 8.14 (s, 2H), 7.56 (d, J = 7.2 Hz, 2H), 7.28 (d, J = 7.2 Hz, 2H), 4.26 (t, J =
3
7
.2 Hz, 2H), 1.87 (m, J = 7.5 Hz, 2H), 1.52 (m, 6H), 1.26 (t, J = 7.5 Hz, 3H).
9
-Hexylcarbazole-3,6-di-(2-thienyl) (10): 3,6-dibromo-9-hexylcarbazole 9 (2.5 g, 6.14 mmol), 2-
(
tri-n-butylstannyl)thiophene (7.35 g, 19.66 mmol), and PdCl (PPh ) (0.43 g, 0.61 mmol) were
2 3 2
mixed in pyridine and refluxed at 90 °C under nitrogen. After 30 h, the mixture was cooled to
room temperature and extracted with DCM. The organic layer was dried over CaCl and the
2
solvent was removed under reduced pressure. The crude residue was purified by silica gel
column chromatography with 80% hexane:DCM as the eluent to give 10 as a yellow solid (1.07
1
g, 41.95%). H NMR (CDCl
3
, 300 MHz) δ 8.34 (s, 2H), 7.27 (d, 2H), 7.75 (d, J = 7.2 Hz, 2H),
7
2
.40 (m, 4H), 7.13 (t, J = 4.8 Hz, 2H), 4.32 (t, J = 7.2 Hz, 2H), 1.93 (m, J = 7.2 Hz, 2H), 1.39 (m,
H), 0.89 (t, J = 6.9 Hz, 3H).
5
,5’-(9-Hexyl-9H-carbazole-3,6-diyl)thiophene-2-carbaldehyde (11): 9-hexylcarbazole-3,6-di-
(2-thienyl) 10 (0.5 g, 1.2 mmol) was dissolved in DCE and stirred under nitrogen while the
temperature was allowed to increase from 0 °C to ambient temperature. DMF (0.95 g, 13 mmol)

ACCEPTED MANUSCRIPT
and POCl
3
(3.2 9 g, 21.45 mmol) were dissolved in DCE and added slowly to the solution. After
1
h of stirring, the mixture was heated to 70 °C and refluxed for 8 h before being quenched by
cooling to room temperature and slow addition of an aqueous solution of NaOH (2 M). The
mixture was extracted with DCM, the organic layer was dried over anhydrous CaCl , and the
2
solution filtered. The residue obtained after the solvent was removed under reduced pressure was
purified using silica gel column chromatography with 70% hexane:DCM as the eluent. The
1
target product 11 was obtained as a yellow solid (0.46 g, 81%). H NMR (CDCl
3
, 300 MHz) δ
9
2
.90 (s, 2H), 8.44 (s, 2H), 7.84 (m, 4H), 7.49 (t, 4H), 4.36 (t, J = 7.2 Hz, 2H), 1.92 (p, J = 6.9 Hz,
H), 1.40 (m, 6H), 0.89 (t, J = 6.9 Hz, 3H).
5
,5’-(9-Hexyl-9H-carbazole-3,6-diyl)thiophene-2-[5-carboxyl-2,3,3-trimethyl-1-octyl-3H-
indoline], iodide (CDT): 5,5’-(9-hexyl-9H-carbazole-3,6-diyl)thiophene-2-carbaldehyde 11 (0.3
g, 0.64 mmol) was dissolved in DCE (30 mL) in a three-necked round-bottom flask. To this
solution, 5-carboxyl-2,3,3-trimethyl-1-octyl-3H-indoline, iodide (1:1) 2 (0.85 g, 1.9 mmol)
dissolved in MeOH (50 mL) was added and the mixture was refluxed under nitrogen for 12 h at
7
0 °C. After this time, the mixture was cooled to room temperature. The subsequent purification
steps were identical to those described for the synthesis of CDI. The final product CDT was
1
obtained as a dark blue solid (0.53 g, 62.99%). H NMR (DMSO-d , 300 MHz) δ 8.88 (t, 4H),
6
8
2
.41 (s, 2H), 8.35 (d, J = 3.9 Hz, 2H), 8.18 (d, J = 8.1 Hz, 2H), 8.02 (t, 6H), 7.81 (d, J = 8.7 Hz,
H), 7.28 (d, J = 15.3 Hz, 2H), 4.63 (b, 4H), 4,52 (b, 4H), 1.86 (b, 18H), 1.40 (m, 26H), 0.84 (t,
J = 6.3 Hz, 9H).

ACCEPTED MANUSCRIPT
2
.3 Details of computational analysis
All calculations reported in this work were carried out using the Gaussian 09 suite of programs
46]. Ground state geometries of the four dyes were fully optimized without imposing any
[
symmetry constraints by employing the B3LYP functional [47–49], which represents a
combination of Becke’s three-parameter hybrid treatment with the Lee–Yang–Parr correlation.
To model the dyes, we replaced all of the long hexyl and octyl chains with simple methyl groups
for computational simplicity and speed. As the alkyl chains have hardly any influence on the
electrochemical and spectral properties of the dyes, we opted for small methyl groups. The effect
of the solvent (acetonitrile) was taken into account using the integral equation formalism for the
polarizable continuum model (PCM) [50–51] within the self-consistent reaction field theory. For
all atoms, we used the standard 6-31G(d) polarized double-ξ quality basis set. Based on the
optimized ground state geometries, we performed TDDFT simulations to obtain the excitation
energies of the dyes in the UV-vis absorption region. The optical properties of the dyes were also
calculated using a long-range corrected functional such as CAM-B3LYP.

ACCEPTED MANUSCRIPT
2
.4 Assembly and characterization of DSSCs
The NiO paste was prepared by mixing a slurry of NiO nanopowder (6 g) in ethanol with
ethanolic ethyl cellulose solution (20 mL, 10 wt%), and terpinol (40 mL), and subsequently
removing ethanol under reduced pressure. The conducting glass substrates (composed of
fluorine-doped tin oxide, FTO) were cleaned successively with soapy water, ethanol, and acetone
using an ultrasonic bath for 20 minutes. The mesoporous NiO layer was deposited using the
doctor-blade technique on top of the bare FTO and compact NiO-blocking layer coated on top of
the FTO. The films were dried at 70 °C for 45 minutes and sintered at 450 °C for 30 minutes.
2
Subsequently, the resultant films with an active cell area of 0.24 cm were immersed in 0.3 mM
solution of sensitizers in dry acetonitrile for 48 hours at 25 °C for dye absorption.
2 6
The counter electrodes were produced using a single drop of H PtCl in isopropanol solution.
This prepared mixer was applied to a clean FTO glass, which was subsequently thermally
decomposed by annealing at 450 °C for 30 minutes. The complete device was assembled by
placing a platinum-coated FTO substrate and a dyed NiO working electrode face-to-face using a
6
0 µm thick Surlyn (Dupont 1702). A solution of 1.4 M LiI and 0.1 M iodine in acetonitrile was
used as the electrolyte, which was injected through a pre-drilled hole on the counter electrode.
3
3
. Results and discussion
.1 Synthesis
Generally, the multi-branching concept in organic dye-sensitizers is expected to facilitate
increased semiconductor surface, thus leading to a better photovoltaic performance. This effect
has been observed for many organic dyes. In the majority of p-type dyes, the extension of π-

ACCEPTED MANUSCRIPT
conjugation can result in improved photovoltaic performance. In this work, the red-absorbing
cationic indolium was selected as an acceptor owing to the recent reports that showed that the
use of this compound can enable dyes to absorb long wavelength visible light, leading to
improved performance of p-type DSSCs [36–39]. With these points in mind, we have designed
and synthesized in the present work four novel carbazole-based single and double branched dye
molecules with and without thiophene spacers. The synthetic routes to these four dyes are shown
in Scheme 1, and are based on the procedure published by Park et al.[37] The 5-carboxy-2,3,3-
trimethyl-1-octyl-3H-indolium iodide acceptor (2) was synthesized following the procedure
published by Park et al.[37] Intermediate 3 was prepared from a simple carbazole via alkylation
reaction with 1-bromohexane. In order to prepare the thiophene-conjugated dyes, intermediate 3
was subjected to mono and di-bromination, followed by Stille coupling to obtain intermediates 7
and 10, respectively. In the next step, intermediates 3, 7, or 10 were subjected to single or double
Vilsmeier-Haack formylation to obtain intermediate 4 and 5 or 8 and 11, respectively. In the
final step, the target dye molecules, CSI, CDI, CST, and CDT, were obtained in reasonably good
yields by Knoevenagel condensation reactions of aldehydes 4, 5, 8, and 11 with 5-carboxy-2,3,3-
trimethyl-1-octyl-3H-indolium iodide 2, respectively.
3
.2 Optical properties
UV-vis absorption spectra of the synthesized dye-sensitizers in acetonitrile are shown in Fig. 2a,
and the corresponding data are listed in Table 1. Two major absorption bands were observed,
namely a relatively weak band in the near-UV region (340–365 nm) corresponding to the π–π*
electron transition, and a strong absorption in the visible region (440–600 nm) that could be
assigned to an intramolecular charge transfer (ICT) between the donor unit and the acceptor end

ACCEPTED MANUSCRIPT
−
1
−1
group. For the above four dyes, the double anchoring dyes CDI (541 nm and 63,101 M cm )
−
1
−1
and CDT (601 nm and 77,577 M cm ) exhibited considerably more red-shifted and improved
molar extinction coefficients (ε) than the congener of mono anchoring dyes CSI (520 nm and
−
1
−1
−1
−1
5
0,171 M cm ) and CST (578 nm and 54,811 M cm ), respectively. In addition, irrespective
of the anchoring number, the introduction of a thiophene spacer in CST and CDT resulted in a
higher and broader absorption pattern. Overall, the double anchoring dyes CDI and CDT dyes
displayed higher values of ε, while the mono anchoring dyes CSI and CST dyes showed lower
values. The absorption peak wavelengths and the onset wavelengths could be ordered from long
to short as CDT > CST > CDI > CSI. Fig. 2b shows the absorption spectra of the four dyes on
NiO films, where the NiO films were soaked in the dye solutions for 24 h. The relevant values
are summarized in Table 1. The absorption maxima peaks for CSI, CDI, CST, and CDT on NiO
films were at 545, 560, 580, and 632 nm, respectively. The absorption spectra of these dyes were
broadened and red-shifted by 25, 19, 2, and 31 nm, respectively, when compared to their spectra
in solution (Fig. 2a). These changes are indicative of strong interactions formed between the dyes
and the semiconductor. Especially, CDT provided a panchromatic absorption wavelength.
3
.3 Electrochemical properties
The ground-state oxidation and reduction potentials of the four prepared dyes were measured by
CV in dry acetonitrile containing 0.1 M TBA(BF ) at a 30 mV/s scan rate (Fig. 3). In each case,
4
the highest occupied molecular orbital (HOMO) and the LUMO of the synthesized dye-
sensitizers should provide adequately large driving force when compared to the NiO valence
band and redox couple for successful charge transfer to occur. Table 1 shows the oxidation
potentials (estimated HOMO levels) of the four sensitizers, which were found to be similar to

ACCEPTED MANUSCRIPT
each other (0.86–0.89 V) vs. NHE. These values are higher than the valence band potential of
NiO (0.54 V) [38].
These higher values of HOMO levels could easily lead to the injection of electrons from the
valence band of the NiO semiconductor into the sensitizers, or in other words, the hole could be
distributed from the sensitizer to the NiO valence band (Fig. 4). The increase in conjugation as a
result of the insertion of a hetero-aromatic thiophene unit or changing from single branched to
double branched resulted in red-shifted absorption pattern and subsequently, a decrease in the
energy bandgap E0-0 [13,38]. The values of LUMO levels, which were extracted from the
difference between E and E , were found to be in the range from −0.91 to −1.37 V vs. NHE.
ox
0-0
−
·
−
These values are more negative than the redox potential of E (I / I ) (−0.19 V vs. NHE). It is
3
2
well known that the number of anchoring/electron accepting group has an influence on the
LUMO level of dyes, and strong electron-accepting ability generally corresponds to lower lying
LUMO. This will result in a lower thermodynamic driving force for dye regeneration. However,
a strong electron-accepting group should increase the push-pull nature of the dye and improve
the efficiency of charge separation within the molecule. The driving forces for hole injection (Eq.
1
) and dye regeneration (Eq. 2) can be calculated according to the following equations [52]:
−
ΔG = e[E (NiO) − (E (S*) + E (S/S ))]
(1)
(2)
inj
VB
0−0
red
−
−
ΔG = e[E(M/M ) − E (S/S )]
reg
red
These higher values of HOMO levels could easily lead to the injection of electrons from the
valence band of the NiO semiconductor into the sensitizers, or in other words, the hole could be

ACCEPTED MANUSCRIPT
distributed from the sensitizer to the NiO valence band (Fig. 4). The increase in conjugation as a
result of the insertion of a hetero-aromatic thiophene unit or changing from single branched to
double branched resulted in red-shifted absorption pattern and subsequently, a decrease in the
energy bandgap E0-0 [13,38]. The values of LUMO levels, which were extracted from the
difference between E and E , were found to be in the range from −0.91 to −1.37 V vs. NHE.
ox
0-0
−
·
−
3 2
These values are more negative than the redox potential of E (I / I
) (−0.19 V vs. NHE). It is
well known that the number of anchoring/electron accepting group has an influence on the
LUMO level of dyes, and strong electron-accepting ability generally corresponds to lower lying
LUMO. This will result in a lower thermodynamic driving force for dye regeneration. However,
a strong electron-accepting group should increase the push-pull nature of the dye and improve
the efficiency of charge separation within the molecule. The driving forces for hole injection (Eq.
1
) and dye regeneration (Eq. 2) can be calculated according to the following equations [52]:
3
.4 Theoretical calculations
DFT and TDDFT studies were performed on the four dyes to obtain a deeper understanding of
their structural, electronic, electrochemical, and optical properties. The experimental optical and
electrochemical properties of the dyes were corroborated by these theoretical results. While dyes
CSI and CDI are planar in nature, the introduction of the thiophene ring in CST and CDT dyes
disturbs this planarity. Fig. 5 shows the optimized ground state geometries of the four sensitizers,
and Fig. 6 displays the electron density distributions in the frontier molecular orbitals of the
modeled sensitizers. In the HOMO of all sensitizers, the electron density was predominantly
localized over the carbazole moiety and partly delocalized over the rest of the molecule, with the
exception of the anchoring group. By contrast, the electron densities of the LUMOs were shifted

ACCEPTED MANUSCRIPT
away from the carbazole moiety and spread onto the anchoring groups instead. Similar electron
density distribution patterns have been observed previously for reported p-type organic
sensitizers [53].
For effective hole injection in a p-type DSSC, the HOMO of the dye should be located well
below the valence band edge of the semiconductor (−4.98 eV for NiO). The calculated HOMOs
of the four dyes were below the valence band edge of NiO, which should ensure efficient hole
injection in each case. Another requirement for p-type DSSC is that the LUMO of the dye should
−
−
be well above the redox potential of the redox couple (−4.80 eV for I /I ) for efficient dye
3
regeneration process. It can be seen from Table 2 that the calculated LUMO levels of all
−
−
sensitizers were above the redox potential of I /I , which facilitates effective dye regeneration.
3
All simulated HOMO, LUMO eigenvalues, and HOMO–LUMO energy gaps are provided in
Table 2, and are found to be in good agreement with those obtained experimentally using cyclic
voltammetry. The HOMO–LUMO energy gaps decreased on going from CSI, to CDI, CST, and
CDT, producing a bathochromic shift in the UV-vis absorption spectra of these dyes.
Simulated UV-vis absorption spectra. The UV-vis absorption spectra of the four sensitizers
were simulated at the TD-B3LYP/6-31G(d) level of theory in acetonitrile. In these TDDFT
simulations, 25 of the lowest singlet–singlet transitions were evaluated to obtain the optical
properties of the four dyes. The simulated absorption spectra of the dyes are depicted in Fig. 7,
and the calculated excitation wavelengths, oscillator strengths, and coefficients of configuration
interaction are summarized in Table 3. The TDDFT simulations reproduced the main bands that
were observed in the experimental UV-vis spectra of the four dyes. The spectrum of each dye
exhibited two main bands, one below 500 nm and the second above 500 nm, with a tail
extending to 800 nm. The calculated results revealed that the most intense transitions in the low-

ACCEPTED MANUSCRIPT
energy region were found at 499, 541, 615, and 660 nm for CSI, CDI, CST, and CDT,
respectively. The theoretical results obtained for dyes CSI and CDI were in excellent agreement
with the experimental absorption data. For dyes CST and CDT, however, the simulated
absorption maxima varied somewhat from the experimental data, a fact that may be attributed to
the inherent limitations of the TDDFT methodology to reproduce extended π systems [54].
The calculated absorption maxima of the dyes followed the same trend as observed
experimentally. A considerable bathochromic shift in the absorption spectra was observed on
going from CSI to CDT, which arises as a result of decreased HOMO–LUMO energy gaps as the
extent of π-conjugation increased. We also calculated the absorption spectra of the dyes using the
range-separated functional CAM-B3LYP, but the results obtained using this method were not in
a good agreement with the experimental data. The optical properties of the dyes obtained using
the CAM-B3LYP method are nonetheless provided in Table S1. The introduction of the
thiophene ring produced a bathochromic shift on going from CSI to CST and from CDI to CDT.
Because of the improvements in optical properties observed as a result of the introduction of the
thiophene ring, the overall performances of DSSCs prepared using CST and CDT dyes also
improved. The most intense low-energy excitations of all the dyes examined in this work
occurred from H ‰ L transition.
3
.5 Photovoltaic properties
−
−
The four sensitizers, CSI, CDI, CST and CDT, were incorporated into DSSCs with an I /I
3
redox couple, and Fig. 8 shows the incident photon-to-current conversion efficiencies (IPCEs) of
the corresponding DSSCs. The differences in photovoltaic performances of these DSSCs, as well
as their power conversion efficiencies (PCEs), short-circuit photocurrent densities (JSC), open-
circuit photo-voltages (VOC), and fill factors (FFs) were investigated under full sunlight and the

ACCEPTED MANUSCRIPT
results summarized in Table 4, and the J–V curves displayed in Fig. 9. Usually, an increase in the
π-conjugation of the dyes broadens and shifts the spectral response towards the red region, which
is crucial for enhanced light harvesting. As shown in Fig. 8, the devices lacking the thiophene
spacer, CSI and CDI, showed higher values of IPCE of 10% and 14%, respectively, than CST
and CDT. In the case of CST and CDT, the introduction of the thiophene spacer improved the
width of their corresponding IPCE up to 800 nm. However, this introduction did not lead to an
improvement in the magnitude of the IPCE values, producing lower photo-current responses
relative to those obtained with devices lacking the thiophene spacers (CSI and CDI).
As shown in Fig. 9, switching from single to double branched dyes produced an increase in both
JSC and VOC values of the corresponding devices. In contrast, introducing the thiophene spacer
between the donor and acceptor groups led to a prominent decrease in J . These outcomes are in
SC
good agreement with the calculated dye regeneration values discussed above. It is well known
that the decrease in the LUMO energy of structurally similar dyes could decrease the Greg
values to restore the ground state of the dye molecules. This could, in turn, result in severe
reduction in IPCE values. In our results, the structurally similar single branched dyes CSI and
CST exhibited a shifted bathochromically max. In particular, the IPCE dropped from 10% for the
single branched dye absorbing at a max. of 487 nm to 8% for the single branched dye absorbing
up to 770 nm (max. at 532 nm). A similar trend was observed for the double branched dyes.
However, despite the decrease in the Greg values of the double branched dyes, overall, the
current density and photo-voltage values were higher compared to those of single branched dyes
owing to the fact that the double branched di-anchoring dyes could better insulate and attain
higher adsorption on the NiO surface.

ACCEPTED MANUSCRIPT
The power conversion efficiencies from highest to lowest were CDI > CDT > CSI > CST. It is
clear that the DSSC devices with double branched di-anchoring dyes showed both higher JSC and
VOC values, and in turn also PCEs compared to the singly branched dyes. Usually, the current-
density values are positively correlated with the light-harvesting abilities of dye sensitizers.
Nevertheless, contrary to previous reports, although the CDT exhibited the highest and broadest
absorption spectrum, the value of J of the p-DSSC based on CDT was lower than that achieved
SC
with CDI, and comparable with that of CSI. This result can be attributed to the lower dye
regeneration driving force of CDT. Additionally, this result can be also explained by the fact that
the CDT-based DSSC exhibited lower IPCE than those prepared with CSI and CDI, despite the
broader nature of its IPCE. However, the presence of diethyl and dioctyl chains is highlighted in
terms of the changes in VOC values. This may be caused by the double branching, which could
act as a better insulator. Subsequently, the increased number of alkyl chains as a result of the
double branching could facilitate better use of inhibiting molecular aggregation on the
photocathode film, which decreases the possibility of hole recombination processes at the
photoelectrode/dye/electrolyte interface, and enhances photo-voltage.
3
.6 Electrochemical impedance spectroscopy
We further employed electrochemical impedance analysis (EIS) to investigate the working
principles of the prepared p-DSSCs. This analysis is a helpful tool for the interpretation of charge
transport processes in electrochemical devices such as n-DSSCs and p-DSSCs [37,45]. Fig. 10
shows the results obtained for DSSCs comprised of FTO/NiO/dye electrodes. The major
semicircles (RCT) observed in the Nyquist plots (Fig. 10a) of all sensitizers were related to the
S
resistance at the NiO/Dye/electrolyte interface. In our case, R and RCE represent the FTO

ACCEPTED MANUSCRIPT
interface resistance and the resistance at the interface between the counter electrode/electrolyte,
respectively. Owing to the fact that we employed the same Pt-coated counter electrode and the
same electrolyte, the RCE values of all devices did not show any significant variation. Similarly,
the values of R of all sensitizers, which correspond to the intercept of the first semicircle in each
S
case, showed similar results. In the present work, Warburg diffusion was not apparent and thus
taken to be insignificant, most likely as a result of the fact that we employed a thin spacer (60
µm) and a low-viscosity solvent (acetonitrile). The results of fitting are listed in Table 5.
Overall, the analysis showed that the RCT values changed in accordance with the applied
sensitizers, and the RCT values for devices employing double branched dyes (44.70 Ω for CDI
and 48.90 Ω for CDT) were indeed smaller than those of single branched dyes (66.75 Ω for CSI
and 63.67 Ω for CST) owing to their increased charge transfer ability. Consequently, such
decrease in the values of RCT could result in improved photovoltaic performances. Fig. 10b
shows the Bode phase plots of the corresponding devices. The frequency of the characteristic
peak in the Bode phase plot increased in the order of CDT < CDI ≤ CSI < CST. A lower
characteristic frequency in the Bode phase plot signifies a slower rate of charge recombination
and higher V . The reciprocal of the characteristic frequency is related to the hole lifetime and
OC
therefore, the electron lifetime increased in the order of CST < CSI ≤ CDI < CDT. This
sequence of hole lifetimes obtained from the Bode phase plots showed increasing resistance to
recombination. The lower RCT values and longer electron lifetimes observed for double branched
dyes (CDT/CDI) compared to the single branched dyes (CST/CSI) indicates more effective
suppression of charge recombination. In turn, this suppression is reflected in the improved JSC
and VOC values, thus yielding a substantially enhanced power conversion efficiency.

ACCEPTED MANUSCRIPT
By changing from single branched to double branched dyes, we noticed an enhancement in the
values of VOC for double branched dyes (CDI and CDT) relative to those of the single branched
dyes (CSI and CST), which might arise as a result of their better surface coverage of NiO films.
The obtained VOC and electron lifetime values suggest that double branching with increased
number of alkyl chains in CDI and CDT hampers the contact between the electrolyte and the
charges generated in the devices. Therefore, the charge recombination rate on the NiO film
decreases considerably, which in turn increases the electron lifetime on the NiO film, thus
corroborating the higher VOC values for CDI and CDT compared to CSI and CST.
4
. Conclusions
In summary, we have synthesized four novel carbazole-based single and double branched
cationic p-type dyes (CSI, CDI, CST, and CDT) and studied how branching and the presence of
a thiophene spacer affect the photophysical and electrochemical properties in DSSC devices
employing these dyes. These red-absorbing cationic p-type dyes compare favorably with the
state-of-the-art dyes in terms of performance, which indicates that they could be utilized to
enhance the photovoltaic performance of tandem devices in the future. We observed that both the
introduction of a thiophene spacer, as well as switching from single to double branched dyes,
caused the absorption maximum to shift to the red region of the visible spectrum. Consequently,
2
2
the current density values increased from 2.15/2.09 mA/cm (CSI/CST) to 2.92/2.15 mA/cm
CDI/CDT). The EIS analysis showed that this difference stems to an extent from the differences
in the interfacial charge transfer resistance values of these dyes, which were determined as
6.75/63.67 Ω and 44.70/48.90 Ω for single (CSI/CST) and double (CDI/CDT) branched dyes,
respectively. We also found that branching or introduction of the thiophene spacer (in
(
6

ACCEPTED MANUSCRIPT
structurally similar dyes) lowered the LUMO levels and diminished the driving forces for dye
regeneration. In addition, the photo-voltage values increased from 0.117/0.109 V (CSI/CST) to
0
.119/0.135 V (CDI/CDT). Overall, increased branching led to increased JSC and VOC values,
and ultimately produced increased power conversion efficiencies. This outcome can be most
likely attributed to the effectiveness of interfacial charge transfer processes and the increased
electron lifetime values of double anchored dyes. The electronic structures and both
electrochemical and optical properties of the synthesized dyes were systematically studied
through DFT/TDDFT simulations and the obtained results corroborated the experimental data.
This study showed that the LUMO energy level of the dye sensitizer plays a key role in
determining the photovoltaic performance of a p-DSSC..
Acknowledgements
This research was supported by Basic Science Research Program through the National Research
Foundation of Korea (NRF) funded by the Ministry of Education (2016R1D1A3B01011514). J.J.
wishes to thank the Korea Institute of Science and Technology Information for the use of the
PLSI supercomputing resources.
References:
[
[
[
1] O’Regan B, Grätzel M. A low-cost, high-efficiency solar cell based on dye-sensitized
colloidal TiO films. Nature, 1991; 353:737-40.
2
2] Hagfeldt A, Boschloo G, Sun L, Kloo L, Pettersson H. Dye-Sensitized Solar Cells.
Chem. Rev., 2010; 110:6595-663.
3] Clifford J.N. Martinez-Ferrero E, Viterisi A, Palomares E. Sensitizer molecular
structure-device efficiency relationship in dye sensitized solar cells. Chem. Soc. Rev.,
2011; 40:1635-46.

ACCEPTED MANUSCRIPT
[
4] O’Regan BC, Durrant JR. Kinetic and energetic paradigms for dye-sensitized solar cells:
moving from the ideal to the real. Acc. Chem. Res., 2009; 42:1799-08.
5] Odobel F, Le Pleux L, Pellegrin Y, Blart E. New Photovoltaic Devices Based on the
Sensitization of p-type Semiconductors: Challenges and Opportunities. Acc. Chem.
Res., 2010; 43:1063-71.
[
[
[
[
[
[
6] Kakiage K, Aoyama Y, Yano T, Oya K, Fujisawa J, Hanaya M. Highly-efficient dye-
sensitized solar cells with collaborative sensitization by silyl-anchor and carboxy-anchor
dyes. Chem. Commun., 2015; 51:15894-7.
7] Mathew S, Yella A, Gao P, Humphrey-Baker R, Curchod BFE, Ashari-Astani N. et al.
Dye-sensitized solar cells with 13% efficiency achieved through the molecular
engineering of porphyrin sensitizers. Nat. Chem., 2014; 6:242-7.
8] Zhang M, Wang Y, Xu M, Ma W, Li R, Wang P. Design of high-efficiency organic dyes
for titania solar cells based on the chromophoric core of cyclopentadithiophene-
benzothiadiazole. Energ. Environ. Sci., 2013; 6:2944-9.
9] Chandrasekharam M, Suresh T, Singh SP, Priyanka B, Bhanuprakash K, Islam A. et al.
Functionalized styryl bipyridine as a superior chelate for a ruthenium sensitizer in dye
sensitized solar cells. Dalton Trans., 2012; 41:8770-2.
10] Cuesta V, Vartanian M, Cruz de la P, Singhal R, Sharma GD, Langa, F.
Comparative study on the photovoltaic characteristics of A–D–A and D–A–D molecules
based on Zn-porphyrin; a D–A–D molecule with over 8.0% efficiency. J. Mater. Chem.
A, 2017; 5:1057-65.

ACCEPTED MANUSCRIPT
[
[
[
[
11] Odobel F, Pellegrin Y, Gibson EA, Hagfeldt A, Smeigh AL, Hammarström L.
Recent advances and future directions to optimize the performances of p-type dye-
sensitized solar cells. Coord. Chem. Rev., 2012; 256:2414-23.
12] Ren Y, Li Y, Chen S, Liu J, Zhanga J, Wang P. Improving the performance of dye-
sensitized solar cells with electron-donor and electron-acceptor characteristic of planar
electronic skeletons. Energy Environ. Sci., 2016; 9:1390-9.
13] Nattestad A, Mozer AJ, Fischer MKR, Cheng YB, Mishra A, Bäuerle P. et al.
Highly efficient photocathodes for dye-sensitized tandem solar cells. Nat. Mater., 2010;
9:31-5.
14] Le Pleux L, Smeigh AL, Gibson E, Pellegrin Y, Blart E, Boschloo G. et al.
Synthesis, photophysical and photovoltaic investigations of acceptor-functionalized
perylene monoimide dyes for nickel oxide p-type dye-sensitized solar cells. Energ.
Environ. Sci., 2011; 4:2075-84.
[
[
[
15] Bonomo M. Sabuzi F. Carlo AD, Conte V, Dini D, Galloni P. KuQuinones as
sensitizers for NiO based p-type dye-sensitized solar cells. New J. Chem., 2017;
41:2769-79.
16] Liu Z, Xiong D, Xu X, Arooj Q, Wang H, Yin L. et al. Modulated Charge Injection
in p-Type Dye-Sensitized Solar Cells Using Fluorene-Based Light Absorbers. ACS
Appl. Mater. Interfaces, 2014; 6:3448-54.
17] Gibson EA, Smeigh AL, Le Pleux L, Fortage J, Boschloo G, Blart E. et al. A p-Type
NiO-Based Dye-Sensitized Solar Cell with an Open-Circuit Voltage of 0.35 V. Angew.
Chem. Int. Ed., 2009; 48:4402-5.

ACCEPTED MANUSCRIPT
[
[
[
[
[
[
[
18] Zhu H, Wu Y, Liu J, Zhang W, Wu W, Zhu WH. D–A–π–A featured sensitizers
containing an auxiliary acceptor of benzoxadiazole: molecular engineering and co-
sensitization. J. Mater. Chem. A, 2015; 3:10603-9.
19] Maufroy A, Favereau L, Anne FB, Pellegrin Y, Blart E, Hissler M. et al. Synthesis
and properties of push-pull porphyrins as sensitizers for NiO based dye-sensitized solar
cells. J. Mater. Chem. A, 2015; 3:3908-17.
20] Bonomo M, Barbero N, Matteocci F, Carlo AD, Barolo C, Dini D. Beneficial Effect
of Electron-Withdrawing Groups on the Sensitizing Action of Squaraines for p-Type
Dye-Sensitized Solar Cells. J. Phys. Chem. C, 2016; 120:16340-53.
21] Lefebvre JF, Sun XZ, Calladine JA, George MW, Gibson A. Promoting charge-
separation in p-type dye-sensitized solar cells using bodipy. Chem. Commun., 2014;
50:5258-60.
22] Kong W, Li S, Chen Z, Wei C, Li W, Li T. et al. p-Type Dye-Sensitized Solar Cells
with a CdSeS Quantum-Dot-Sensitized NiO Photocathode for Outstanding Short-Circuit
Current. Part. Part. Syst. Charact., 2015; 32:1078-82.
23] Park JY, Suresh T, Yun HJ, Lee CH, Ahn KS, Kim JH. Synthesis and Investigation
of the Donor Effect in Novel Hemicyanine Organic Dyes for p-Type Dye-Sensitized
Solar Cells. Sci. Adv. Mater., 2016; 8:589-95.
24] He J, Lindström H, Hagfeldt A, Lindquist SE. Dye-sensitized nanostructured tandem
cell-first demonstrated cell with a dye-sensitized photocathode. Sol. Energy Mater. Sol.
Cells, 2000; 62:265-73.