Anchoring stability and photovoltaic properties of new D(-π-A) 2 dyes for dye-sensitized solar cell applications

Article abstract of DOI:10.1016/j.dyepig.2013.02.012

This study deals with the synthesis and characterization of two new di-anchoring dyes for applications in dye-sensitized solar cells. The materials were designed with a branched D(-π-A)₂ structure containing (i) a rigid alkyl-functionalized carbazole core as the donor part, (ii) one (DYE1) or two (DYE2) thiophene units as the π-bridge and (iii) a cyano-acrylic moiety as acceptor and anchoring part. Electrochemical impedance spectroscopy indicated that the injected electron lifetime is higher in the case of DYE2, probably due to the length of the π-spacer that, in combination with the alkyl chain on the carbazole unit, hampers the charge recombination with the electrolyte. Stability tests on TiO₂-sensitized films revealed that the di-anchoring remarkably slows down the desorption process, which conversely is evident for classic reference dyes. The highest power conversion efficiency reaches 5.01% in the case of DYE2 with a photovoltage of 0.70 V and a photocurrent of 10.52 mA cm^-2, substantially deriving from a broader absorption with respect to DYE1, as also confirmed by IPCE measurements. These results support the efforts aimed at the structural engineering of D(-π-A)₂ dyes to design new, more efficient and stable organic sensitizers.

Full text of DOI:10.1016/j.dyepig.2013.02.012

Accepted Manuscript
Anchoring stability and photovoltaic properties of new D(-π-A) dyes for dye-
2
sensitized solar cell applications
Roberto Grisorio, Luisa De Marco, Giovanni Allegretta, Roberto Giannuzzi, Gian
Paolo Suranna, Michele Manca, Piero Mastrorilli, Giuseppe Gigli
PII:
S0143-7208(13)00062-4
10.1016/j.dyepig.2013.02.012
DYPI 3859
DOI:
Reference:
To appear in: Dyes and Pigments
Received Date: 11 January 2013
Revised Date: 19 February 2013
Accepted Date: 20 February 2013
Please cite this article as: Grisorio R, De Marco L, Allegretta G, Giannuzzi R, Suranna GP, Manca M,
Mastrorilli P, Gigli G, Anchoring stability and photovoltaic properties of new D(-π-A) dyes for dye-
2
sensitized solar cell applications, Dyes and Pigments (2013), doi: 10.1016/j.dyepig.2013.02.012.
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

ACCEPTED MANUSCRIPT
Anchoring stability and photovoltaic properties of new D(-π-A)2 dyes
for dye-sensitized solar cell applications
Roberto Grisorio,a Luisa De Marco,b Giovanni Allegretta,a Roberto Giannuzzi,b Gian Paolo
Suranna,a,* Michele Manca,b Piero Mastrorilli,a,c Giuseppe Giglib,d,e
aDICATECh - Dipartimento di Ingegneria Civile, Ambientale, del Territorio, Edile e di Chimica,
Via Orabona, 4 I-70125 Bari, Italy. E-mail: surannag@poliba.it
bCenter for Biomolecular Nanotechnology (CBN) Fondazione Istituto Italiano di Tecnologia (IIT),
Via Barsanti 1, Arnesano, 73010, Italy.
cIstituto di Chimica dei Composti Organometallici, Area di Ricerca CNR di Firenze, Via Madonna
del Piano 10, 50019 Sesto Fiorentino (Firenze) Italy.
dNNL - Istituto Nanoscienze, CNR c/o distretto tecnologico Lecce, via Arnesano 16, 73100 Lecce,
Italy.
eDepartment of Mathematics and Physics “E. De Giorgi”, University of Salento, Campus
Universitario, via Monteroni, 73100 Lecce, Italy.
Abstract
This study deals with the synthesis and characterization of two new di-anchoring dyes for
applications in dye-sensitized solar cells. The materials were designed with a branched D(-π-A)2
structure containing i) a rigid alkyl-functionalized carbazole core as the donor part, ii) one (DYE1)
or two (DYE2) thiophene units as the π-bridge and iii) a cyano-acrylic moiety as acceptor and
anchoring part. Electrochemical impedance spectroscopy indicated that the injected electron
lifetime are higher in the case of DYE2, probably due to the length of the π-spacer that, in
1

ACCEPTED MANUSCRIPT
combination with the alkyl chain on the carbazole unit, hampers the charge recombination with the
electrolyte. Stability tests on TiO2-sensitized films revealed that the di-anchoring remarkably slows
down the desorption process, which conversely is evident for classic reference dyes. The highest
power conversion efficiency reaches 5.01 % in the case of DYE2 with a photovoltage of 0.70 V and
a photocurrent of 10.52 mA cm–2, substantially deriving from a broader absorption with respect to
DYE1, as also confirmed by IPCE measurements. These results support the efforts aimed at the
structural engineering of D(-π-A)2 dyes to design new, more efficient and stable organic sensitizers.
Keywords: DSSC, di-anchoring dye, anchoring stability.
1. Introduction
The concerted optimization of the four components (i.e. semiconducting oxide, dye, electrolyte,
counter-electrode) of a dye-sensitized solar cell (DSSC) can rightly be included amongst the
intriguing quests, that chemistry faces today as the multidisciplinary interface science devoted to
unravelling the behaviour of complex matter [1,2]. The continuing contributions from a significant
part of the whole scientific community ultimately aimed at allowing a further decisive efficiency
and long-term stability improvement needed to fulfill the promises of this technology. The
performances of DSSC are now beyond the second generation of solar cells, but the result of the
huge research would qualify them as third generation devices, which are expected to deliver electric
power in a larger scale at a lower price per Watt [3].
In this wider framework, the goal of obtaining more efficient and robust dyes is being carried out
since the first seminal report by Grätzel [4]. While this two-decade long race has not yielded
breakthrough improvement in terms of efficiency with respect to the well-known paradigmatic
ruthenium dyes N3 and N719 [5], it has nevertheless allowed a considerable understanding of the
2

ACCEPTED MANUSCRIPT
principles underlying the structure-property relationships of sensitizers, that led to unquestionable
durability improvement. Thousands of dyes have been tested to date, making their classification a
quite difficult task. A very coarse subdivision can, however, be made between metal complexes [6]
and synthetic fully-organic dyes [7]. There are several reasons that indicate how organic dyes might
have a considerable edge over noble metal complexes, among which their predicted lower
production cost, lower environmental impact and higher molar extinction coefficients. The degrees
of freedom provided by organic synthesis has granted enormous versatility to this class of
sensitizers, although a generally accepted design rule relies on the D-π-A concept. In this motif, an
electron rich group (D) is linked, through a π-conjugated spacer (π), to an electron acceptor group
(A) carrying an anchoring group, most typically a carboxyl group functionality allowing a strong
binding to the TiO2 mesoporous layer. While a convergence has been reached on the use of cyano-
acrylic acid, as acceptor, much of the current research is devoted to the choice of the suitable
donor/π-bridge to increase the DSSC performance.
In the search for novel dye design that could lead to amelioration of the D-π-A concept, it was
recently highlighted that the presence of only one anchoring group in the common organic dyes
could represent a limitation with respect to the Ru-based dyes carrying up to four anchoring groups
[8,9,10,11]. The multi-branched multi-anchoring dye idea was consequently proposed over the
traditional D-π-A architectural design. This concept was exploited for the design of cruciform [12],
spiro- [13] and H-shaped [14] architectures and has also been implemented for the preparation of
dyes for p-type DSSC devices [15,16,17]. The interest toward this design approach is justified by
several reasons. Primarily, the more extended π-system due to the presence of two π-bridges should
lead to an extension of the absorption to longer wavelengths, to a broadening of the absorption
profile and to an enhanced molar extinction coefficient with respect to those of D-π-A sensitizers,
all factors contributing to an improved light harvesting. Moreover, the presence of a further
electron-withdrawing unit (A) could decrease the HOMO-LUMO gap of the organic dye and,
3

ACCEPTED MANUSCRIPT
hence, prove beneficial for its photo-stability. Last but not least, less complex synthetic strategies
are required for the preparation of symmetrically branched organic molecules, with evident
reduction of the prospected production costs of the material.
However, it should be pointed out that the adoption of a di-anchoring architecture D(-π-A)2 does not
necessarily imply a performance enhancement if, for instance, the structural features of the dye do
not efficiently shield the TiO2 surface from the oxidized form of the redox shuttle [18], thereby
allowing unwanted recombinations, detrimental to the overall performances [19].
Our contribution in this field aims at gaining a deeper insight into D(-π-A)2 structures studying the
effect of the π-bridge length on dyes obtained binding an N-octyl-3,6-carbazole core [20] as donor
to cyano-acrylic acid groups as acceptors, through either thiophene or bithiophene spacers (Figure
1).
Figure 1. Structures of the synthesized molecules DYE1 and DYE2.
2. Results and discussion
2.1. Synthesis and characterization
The synthetic sequence for the obtainment of DYE1 and DYE2 is reported in Scheme 1. The
carbazole core was straightforwardly prepared from 3,6-dibromo-N-octyl-carbazole (1) that was
converted into the corresponding boronic ester (2) by lithiation with n-BuLi and subsequent
reaction with 2-isopropoxy-4,4,5,5-tetramethyl-1,3,2-dioxaborolane. Concerning the thiophene-
4

ACCEPTED MANUSCRIPT
based π-conjugated precursors needed for the subsequent step, the commercially available 5-
bromothiophene-2-carbaldeyde was reacted with tributyl(thiophen-2-yl)stannane in a Pd-catalyzed
Stille coupling to obtain the bithiophene derivative 4. Bromination of 4 with NBS permitted to
obtain the aldehyde derivative 5 needed for the following step. The electron-donating carbazole unit
was then bound to the two thiophene-based π-conjugated systems by a Suzuki-Miyaura coupling.
The reaction between 2 and two equivalents of 5-bromothiophene-2-carbaldeyde yielded the
corresponding dialdehyde 3. The synthesis of the target molecule DYE1 was achieved by the
Knoevenagel condensation of the dialdehyde 3 with cyano-acetic acid in the presence of ammonium
acetate. For the preparation of the second target molecule, a similar approach was followed: the
reaction between 2 and two equivalents of 5 allowed the obtainment of the dialdehyde 6 that, after
Knoevenagel condensation with cyano-acetic acid yielded DYE2. The two target molecules were
sufficiently soluble in THF, DMF, ethanol and DMSO to permit their characterization in solution as
well as their application in DSSC devices. The structures of DYE1-2 were confirmed (see
experimental part) by elemental analysis, 1H-NMR, ESI-HRMS as well as by IR analyses.
O
O
Br
Br
B
B
O
O
n-BuLi
N
C
N
O
B
2
O
C8H17
8H17
1
O
CHO
S
S
NCCH2C OO H
AcONH4/AcOH
CHO
Br
2
DYE1
C8H17
N
Pd(PPh3
)
4
S
Sn(n-Bu)3
S
3
CHO
Pd(PPh3
)
4
S
CHO
S
S
2
NCCH2COO H
AcONH4/AcOH
S
NBS
S
Br
CHO
C8H17
N
CHO
DYE2
S
S
Pd(PPh3
)
4
6
5
4
CHO
S
Scheme 1. Synthetic sequence followed for the preparation of DYE1 and DYE2.
5

ACCEPTED MANUSCRIPT
2.2. Optical characterization
The photophysical properties of DYE1 and DYE2 were investigated by UV-Vis and
photoluminescence (PL) spectroscopy. The UV-Vis absorption spectra of the two compounds
measured in THF, are shown in Figures 2 and 3. The presence of a bithiophene π-bridge between
the donor and the acceptor moieties in DYE2 leads to a substantial red-shift of both the absorption
maximum (470 nm) as well as of the absorption onset (557 nm) with respect to those exhibited by
the monothiophene bridged DYE1 (453 and 520 nm, respectively). The broader absorption profile
of DYE2 in the visible region was also accompanied by a higher molar extinction coefficient
(54000 M–1cm–1 at 470 nm) compared to DYE1 (33900 M–1cm–1 at 453 nm). The length of the π-
spacer also has a strong influence on the solution (THF) PL behaviour of the synthesised dyes: the
λem of DYE1 was recorded at 559 nm while that of DYE2 was considerably red shifted, falling at
610 nm.
When organic dyes are adsorbed onto the TiO2 layer, it is very likely that the free –COOH group is converted
into a titanium carboxylate (–COOTi) form, corresponding to a deprotonation of the carboxylic acid
functionality [21]. Most D-π-A organic dyes show a large absorption blue-shift upon binding onto TiO2,
which is detrimental in terms of light harvesting. Generally, in donor-acceptor architectures, the carboxylic
acid deprotonation leads to a blue-shift of the absorption maximum deriving from the attenuation of the
intramolecular charge transfer, due to the lower electron acceptor character of the carboxylate moiety with
respect to the carboxylic acid. To deepen this aspect, the UV-Vis spectra in THF of the dyes were also
recorded in the presence of an excess of a base (triethylamine) and of an acid (formic acid), as shown in
Figures 2 and 3. The absorption maxima for DYE1 and DYE2 in the presence of triethylamine
(THF/NEt3 = 4/1 v/v) were found at 430 nm and 432 nm respectively, remarkably blue-shifted with
respect to those recorded in THF, as a consequence of the complete deprotonation of the carboxylic
acid functionality [22]. Conversely, a remarkable broadening of the absorption spectra in the
6

ACCEPTED MANUSCRIPT
presence of formic acid (THF/HCOOH = 4/1 v/v) was observed for DYE1 and DYE2 with red-
shifted absorption maxima at 468 nm and 479 nm respectively.
THF
1,00
0,75
0,50
0,25
0,00
THF/HCOOH (4/1 v/v)
THF/NEt3 (4/1 v/v)
on TiO2
350
400
450
500
550
600
650
700
Wavelength (nm)
Figure 2. Normalized absorption spectra of DYE1 in THF, THF/NEt3 (4/1 v/v), THF/HCOOH (4/1
v/v) and on TiO2 film.
1,00
THF
THF/HCOOH (4/1 v/v)
THF/NEt3 (4/1 v/v)
0,75
0,50
0,25
0,00
on TiO2
350
400
450
500
550
600
650
700
Wavelength (nm)
Figure 3. Normalized absorption spectra of DYE2 in THF, THF/NEt3 (4/1 v/v), THF/HCOOH (4/1
v/v) and on TiO2 film.
Furthermore, when DYE1 and DYE2 are adsorbed to the TiO2 surface, their UV-vis spectra are
further blue-shifted compared to those in THF/NEt3, although in this case their spectral responses
are broadened due to intermolecular interactions between dye molecules on TiO2 film. The
7

ACCEPTED MANUSCRIPT
absorption maxima for DYE1 and DYE2 adsorbed onto 5 µm-thick TiO2 mesoporous film were, in
fact, found at 397 nm and 424 nm, respectively, blue-shifted by 56 and 46 nm with respect to those
recorded in THF. On the bases of the previously described solvatochromic behaviour, this
observation is indicative of the deprotonation of both carboxylic acid groups on TiO2 surface
[18,23].
To corroborate this assumption and since the anchoring modes of the dyes are related with the
interfacial electron injection, FT-IR measurements of the dyes absorbed on the TiO2 surface were
carried out. Figure 4 shows the IR spectra in the region 4000–400 cm–1 of DYE1- and DYE2-
sensitized TiO2 films and of bare TiO2.
4000
3500
3000
2500
2000
1500
1000
DYE1
DYE2
DYE1-sensitized TiO2
DYE2-sensitized TiO2
4000
3500
3000
2500
2000
1500
1000
Wavenumber [cm-1]
Figure 4. FT-IR spectra of DYE1 and DYE2 in KBr and adsorbed on TiO2 film.
The IR spectra of the sensitized films display characteristic narrow absorption bands at 2930 cm–1
and 2850 cm–1, which correspond to the antisymmetric and symmetric C–H stretching vibrations,
respectively, of the –CH2– moieties of the hydrocarbon chain; moreover, the band associated to the
8

ACCEPTED MANUSCRIPT
C≡N stretching of the cyano-acrylic acid group was clearly observed at 2216 cm–1. The two
characteristic bands centred at ~1590 and ~1385 cm–1, respectively, are associated with the COO–
antisymmetric and symmetric stretching vibrations of carboxylate groups coordinated with surface
titanium atoms; the lack of clear evidence for the C=O stretching band of free carboxyl groups at
~1710–1720 cm–1 and the difference in wavenumbers between the aforementioned carboxylate
bands (ꢀυ = ~ 205 cm–1) confirm that the dye acts as bidentate since both its carboxyl moieties are
bound to the TiO2 surface. These data confirm the absence of free carboxylic acid groups as well as
the bidentate adsorption of both dyes onto the TiO2 surface [9,18, 24].
2.3. Electrochemical properties
Cyclic voltammetry (CV) of sensitizers for DSSC applications provide crucial information on the
feasibility of the electron injection from the excited states of the dye to the TiO2 conduction band
and on the possibility of dye regeneration by means of the electrolyte redox couple. As shown in
Figure 5, the two dyes exhibited only one reversible oxidation event. The CV wave did not undergo
substantial modification after repeated scans, revealing that both DYE1 and DYE2 are
electrochemically stable under oxidative conditions. It is worth noting that the presence of a
thiophene unit in the π-bridge of DYE2 lowered its oxidation potential (corresponding to the onset
of the anodic event) of ~0.2 V with respect to DYE1. The evaluation of the HOMO energy levels
(see experimental part) led to values of –5.4 and –5.2 eV for DYE1 and DYE2, respectively. The
obtained values indicate that an efficient regeneration of the oxidised dyes by the iodide/triiodide
couple is thermodynamically feasible, lowering the possibility of a geminated recombination
between the oxidized dye and the photo-injected electrons in the TiO2-based anode. Noteworthy,
the relatively deep HOMO energy values of both dyes lead to a prediction of a high stability
towards oxidation, which constitutes an advantage for the prospected applications of these
materials.
9

ACCEPTED MANUSCRIPT
DYE1
DYE2
0.5 µA
-0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Voltage vs Ag/Ag+ (V)
Figure 5. Anodic CV curves of the dyes in THF solutions.
Since no cathodic behaviour of both DYE1 and DYE2 could be recorded by CV, their excited state
oxidation potentials (corresponding to the LUMO energy levels) were calculated by adding the
energy gap (estimated from the onset of the absorption spectra recorded in THF solutions) to
HOMO energy values. It can be noted how the obtained LUMO of both dyes (–2.9 eV) lays above
the TiO2 conduction band edge (~ –4.0 eV) warranting the necessary driving force for an efficient
electron transfer process.
2.4. Theoretical investigations
A key step for the accurate theoretical description of the electronic and thermodynamic properties
of an organic semiconductor is a reliable evaluation of its molecular geometry. In the specific case
of di-branched dyes, the identification of an accurate and reliable molecular modelling method for
the prediction of the equilibrium molecular geometry is mandatory. Hence, since the two thienyl
moieties bound to the carbazole unit can adopt a different orientation with respect to each other,
molecular geometries were optimized in vacuo, constraining the dihedral angles of the C–C bonds
10

ACCEPTED MANUSCRIPT
connecting the thienyl units to carbazole to fixed values and subsequently finding the minimum
energy structures. To minimize the computational load, the optimization was performed using the
semi-empirical (PM6) method. Since, due to the molecular structure, the length of the alkyl
substituent on the carbazole nitrogen atom is not expected to have an influence on the conformation
or on the HOMO and LUMO energy, it was chosen to replace the n-octyl with n-propyl groups.
Moreover, in the case of DYE2, the geometry optimization was carried out starting from a structure
in which the thiophene units of the bithiophene moiety were placed in the anti-orientation, as
proposed for similar compounds. These calculations [24] demonstrated that in the most stable
conformers the thienyl units are tilted towards the same side (as shown in Figure 1) although the
low torsional barrier (~2.9 kcal/mol) does not allow the freezing of the conformational equilibrium.
Starting from these conformations, the geometry of both dyes were fully optimized in vacuo until
convergence by density functional theory (DFT) calculations at the B3LYP/6-31G(d,p) level. The
molecular orbital energy of the dyes as well as their electron density distribution and their
optimized geometry were estimated by DFT calculations at the B3LYP/6-31G(d,p) level including
solvent effects (THF) using non-equilibrium implementation of the conductor-like polarizable
continuum model (C-PCM).
In the case of DYE1, the electron density of the HOMO state is distributed along the entire
molecule with a higher contribution of the electron-donor carbazole moiety, as evident from Figure
6. The presence of a further thiophene unit in the π-bridge of DYE2 spacing the electron-donor
from the electron acceptor leads to a decrease of the HOMO electron density on the cyano-acrylic
units that, in principle, could reduce the probability of the aforementioned charge recombination of
the oxidised dye with the TiO2 electron. Conversely, the LUMO electron density distributions of
both dyes are prevalently localized around the cyano-acrylic units, thereby favouring an efficient
electron transfer from the excited state of the dyes to the TiO2 conduction band edge. Although a
straightforward correspondence between experimental and theoretical HOMO-LUMO energy levels
11

ACCEPTED MANUSCRIPT
is not often observed for organic dyes, in the case of DYE1 and DYE2, the calculated and
electrochemical HOMO-LUMO energies are in rather good agreement (Table 1).
Figure 6. Electronic density distributions of the HOMO–1, HOMO, LUMO and LUMO+1 orbitals
of DYE1 and DYE2.
The calculation of the vertical S0 → Sn excitation energies at the TDDFT level of theory provides
more insight into the optical behaviour of the two materials. For a reliable description of the
transitions with non-negligible charge transfer character, such as those expected for our D(-π-A)2
structure, a judicious choice of the proper functional is crucial. In fact, use of a conventional
exchange-correlation functional such as a pure and hybrid functional would remarkably
underestimate the energy of the excited states with significant charge transfer character due to the
self-interaction error [21]. The Coulomb attenuating method CAM-B3LYP functional [22] was
therefore chosen to calculate both the vertical excitations energy and the oscillator strengths. This
method considers long range corrections, which are appropriate for the description of charge-
12

ACCEPTED MANUSCRIPT
transfer type excitations. Also, before describing the results, it should be highlighted that, although
the HOMO→LUMO energy transition is generally used as a measure of the excitation energy, this
might not be dominated by the HOMO→LUMO transition and, analogously, the charge injection
from the excited state of the dye into the semiconductor cannot always be described as a simple
transfer from the LUMO of the dye into the semiconductor’s conduction band.
dye
HOMO
LUMO
nm
f
configuration
–5.54
–2.72
423
390
457
438
1.28
1.01
1.77
1.50
H→L (0.59); H–1→L+1 (0.30)
H→L+1 (0.52); H–1→L (0.36)
H→L (0.51); H–1→L+1 (0.40)
H→L+1 (0.48); H–1→L (0.44)
DYE1
–5.28
–2.84
DYE2
Table 1. Theoretical HOMO and LUMO energy values in THF as determined at the C-PCM-
B3LYP/6-31g(d,p) level of theory for DYE1 and. Excited-state transition energies (wavelengths),
their electronic configurations and oscillator strength (f) determined in THF by TD-C-PCM-CAM-
B3LYP/6-31g(d,p) using the B3LYP/6-31g(d,p) geometries for DYE1 and DYE2.
As shown in Table 1, the TDDFT calculation gave two main transitions in the visible region for
both dyes, which are in good agreement with the experimental data, since DYE2 showed red-shifted
vertical excitation energies with higher oscillator strengths with respect to those exhibited by
DYE1. Furthermore, according to these results, the S0→S1 transition at 423 nm and 457 nm for
DYE1 and DYE2, respectively, not only correspond to the promotion of an electron from the
HOMO to the LUMO but is also associated with the promotion of an electron from the HOMO–1 to
the LUMO+1. These latter two molecular orbitals are also involved in the description of the S0→S2
transition at 390 nm and 438 nm for DYE1 and DYE2, respectively. Nevertheless, it is noteworthy
that while HOMO–1 electronic density is almost fully spread along the entire molecule both in
13

ACCEPTED MANUSCRIPT
DYE1 and in DYE2, the LUMO+1 electronic density is predominantly concentrated around the
cyano-acrylic groups. This electronic localization of the LUMO+1 should warrant an efficient
electron transfer also for the transitions involving this excited state. From the results reported up to
this point, it is apparent how the “branched” architecture approach positively affects the light
harvesting efficiency (by increasing the number of aromatic units present in the related molecules)
without depressing the dye key features: the electron injection ability, the feasibility of the oxidised
dye reduction by the electrolyte and their photo-stability.
2.5. Photovoltaic performances and stability
Having established the potential of the new sensitizers, these have been implemented in
photoelectrochemical solar cells devices and their performances investigated both in terms of J-V
characteristics and of electrochemical impedance spectroscopy (EIS) analysis. Two different
typologies of photoanodes were fabricated to this purpose: a first batch of 7 ꢁm-thick transparent
films made of 20 nm-sized TiO2 nanocrystals, and a second batch of 17 µm-thick photoelectrodes
comprising a 12 µm-thick transparent layer and a 5 µm-thick scattering layer constituted by 300
nm-sized light reflecting titania particles. The photovoltaic characteristics of these devices
(measured under standard illumination conditions, namely AM 1.5G at 100 mW cm–2) are
summarized in Table 2 and compared with similar devices employing D5 and N719 as reference
organic and ruthenium-based dyes; the J-V plots are reported in Figure 7A. The DYE1- and DYE2-
based devices made with the 7 µm-thick layer exhibited a short circuit photocurrent (Jsc) of 7.38 and
8.41 mA cm–2, open circuit voltages (Voc) of 0.723 and 0.715 V and fill factors (ff) of 0.69 and 0.68,
respectively, resulting in an efficiency (η) of 3.68 % and 4.09 % respectively. Not much higher
performances are obtained for the N719-based device (Jsc = 9.27 mA cm–2, Voc = 0.820 V, ff = 0.67,
η = 5.09 %) and for the D5-based device ( Jsc = 9.11 mA cm–2, Voc = 0.74 V, ff= 0.72, η = 4.85 %)
14

ACCEPTED MANUSCRIPT
fabricated and measured under the same conditions. Subsequently, the implementation of thicker
photoelectrodes with improved light harvesting capabilities led, as expected, to a significant
enhancement of the short circuit photocurrent and, hence, of the DSSC performances.
70
12
B
A
11
10
9
60
50
40
30
20
10
0
DYE1
DYE2
D5
8
7
6
D5 17 µm
5
DYE2 17 µm
DYE1 17 µm
D5 7 µm
DYE2 7 µm
DYE1 7 µm
4
3
2
1
0
350 400 450 500 550 600 650 700 750
0,0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
Wavelength (nm)
Bias Voltage (V)
Figure 7. A) J-V curve of DSSCs prepared with transparent photoelectrodes 7 ꢁm thick and opaque
double-layered photoanodes 17 ꢁm-thick. B) IPCE spectra of DSSCs prepared with double-layered
17 ꢁm-thick photoelectrodes.
Devices made with the 17 µm-thick photoanodes yielded, in fact, remarkably higher photocurrents
(9.36 and 10.52 mA cm–2 for DYE1 and DYE2, respectively) but substantially similar Voc (0.71 and
0.70 V) and ff values (0.67 and 0.68) corresponding overall conversion efficiencies of 4.47 % and
5.01 %, respectively. However, the device parameters of the reference dyes showed different
degrees of improvement for N719 and D5 upon increasing of the photoelectrode thickness.
Whereas, in the case of the 7 µm-thick photoelectrode, the ruthenium complex showed only 1.2 and
1.1 times higher photocurrent than those of the DYE1 and DYE2, in the case of thicker
photoelectrodes, the Jsc produced by N719 were found to be 1.5 and 1.3 times higher with respect to
15

ACCEPTED MANUSCRIPT
DYE1 and DYE2 respectively. It is reasonable to suppose that, due to their much higher extinction
coefficients, DYE1 and DYE2 tend to saturate incoming light within a relatively low thickness. As
a consequence, the gap in photocurrent respect to the ruthenium dye increases with the electrode
thickness. By contrast, the photocurrent produced by D5 using thicker photoelectrodes (10.06 mA
cm–2) was even lower than that of DYE2, probably due to the better light-harvesting efficiency of
the latter, supported by theoretical calculations (vide supra).
Dye
Jsc (mA cm–2)
7.38a ; 9.36b
8.41a ; 10.52b
7.89b
Voc (V)
ff
η
(%)
0.723a ; 0.712b
0.715a ; 0.701b
0.703b
0.69a ; 0.67b
0.68a ; 0.68b
0.66b
3.68a ; 4.47b
4.09a ; 5.01b
3.66b
DYE1
DYE2
DYE1c
DYE2c
D5
8.87b
0.700b
0.68b
4.22b
9.11a ; 10.06b
9.27a ; 13.91b
0.740a ; 0.732b
0.820a ; 0.804b
0.72a ; 0.72b
4.85a ; 5.34b
0.67a ; 0.70b
5.09a ; 7.83b
N719
a
b
c
Table 2. Solar cell performance values. TiO2 thickness = 7 ꢁm. TiO2 thickness = 17 ꢁm. uptake
performed in the presence of CDCA (10 mM).
Spectra of the monochromatic incident photon-to-current conversion efficiency (IPCE) for the
DSSCs based on the two here presented dyes are reported in Figure 7B. The IPCE spectrum of the
DYE2-sensitized cell is red shifted compared to that of DYE1, which is in good agreement with the
absorption behavior of the dyes absorbed on TiO2 film. The onset of the IPCE spectrum of DYE2
falls at 700 nm and a high conversion efficiency (60 %) was observed in the range 400-550 nm. In
contrast, the DYE1-based cell showed a blue-shifted onset of the IPCE spectrum (at 620 nm) and
exhibited a good plateau (>60 %) in a range comprised between 400 and 500 nm. Therefore, the
relatively lower photocurrent produced by DYE1, leading to its relatively lower efficiency, is
16

ACCEPTED MANUSCRIPT
clearly attributable to its narrow and blue-shifted spectral response compared to the one exhibited
by DYE2.These results pinpoint the bithiophene spacer present in DYE2 between the donor and the
acceptor moieties as an excellent π-conjugation system to harvest a larger fraction of the solar
spectrum. In fact, DYE2-based device gave a better higher short-circuit current density when
compared to analogous DYE1 which contains just one thienyl unit.
The electron lifetime in a DSSC is a fundamental parameter to assess the recombination dynamics
at the TiO2–dye–electrolyte interface [25]. If all the devices are fabricated and tested in the same
conditions, only the differences in the molecular structure of the sensitizers can influence the
electron lifetime. It has been calculated by an electrochemical impedance spectroscopy (EIS)
investigation carried out on the devices under 1 sun illumination. Figure 8A reports the Nyquist
plots obtained for devices based on DYE1, DYE2 and D5 at the VOC, the range explored being
from 100 kHz to 10 mHz. Figure 8 shows the plot of the apparent electron lifetime in DYE1, DYE2
and D5-based devices as obtained from EIS measurements performed under 1 sun illumination. EIS
spectra were analyzed through the well-known equivalent circuit [25]. By this approach, a
transmission line model is applied to the nanocrystalline TiO2 film. The electron lifetime was
calculated by the equation:
τ = Rct/Cꢁ
in which Rct and Cꢁ are the charge transport resistance at the TiO2/dye/electrolyte interface and the
chemical capacitance respectively. The apparent electron lifetimes reported in Figure 8B have been
plotted as a function of the corrected potential, as described in the literature [25]. The potential
correction is mandatory in order to account for the losses due to the total series resistance Rs, which
leads to a potential drop that is not associated with the displacement of the Fermi level.
17

ACCEPTED MANUSCRIPT
30
25
20
15
10
5
A
DYE 1
DYE 2
D5
0
10 20 30 40 50 60 70 80 90 100 110
Z' [Ω]
0,008
B
0,006
0,004
DYE1
DYE2
D5
0,002
0,55
0,60
0,65
0,70
0,75
0,80
0,85
Corrected potential (V)
Figure 8. (A) Nyquist plots obtained from EIS measurements at VOC under 1 sun illumination for
devices based on DYE1, DYE2, and D5. (B) Apparent electron lifetime as a function of the
corrected potential for DYE1, DYE2, and D5.
It is evident how the apparent electron lifetimes of DYE1 and DYE2 are remarkably shorter than
those of the organic reference D5, which suggests that injected electrons from DYE1 and DYE2
undergo an easier recombination than those injected from D5. Such recombination can occur with
the oxidized form of the redox species present in the electrolyte as well as with the oxidized form of
the dye generated after the electron transfer (geminate recombination). This drawback can either be
ascribed to the absence of a bulky triarylamine group that, in D5 keeps the electrolyte from the TiO2
18

ACCEPTED MANUSCRIPT
surface, or to the non-optimal HOMO and HOMO–1 electronic distributions of DYE1 and DYE2,
which, being involved in the main electronic transitions, could favor the geminate recombination.
Furthermore, an estimate of the dye loading for DYE1 (5.68 × 10–8 mol cm–2), DYE2 (5.63 × 10–8
mol cm–2) and D5 (6.73 × 10–8 mol cm–2) revealed that the amount of adsorbed DYE1 and DYE2 is
lower compared to that of D5. The double anchoring structure reasonably hampers the access to
adsorption sites with respect to the monobranched sensitizer, providing an explanation for the lower
efficiencies, ascribable to the increased probability of detrimental charge recombination processes.
Nevertheless, DYE2-sensitised photoelectrodes are characterised by a slightly longer apparent
electron lifetime with respect to those of DYE1, supporting a more favorable behavior in terms of
recombination dynamics, which might be associated with the presence of the bithiophene spacer in
DYE2.
Since the recombination of injected electrons from the semiconductor to the oxidized species of the
redox electrolyte might be circumvented by the use of suitable additives promoting a complete
coverage of the TiO2 surface, two devices were prepared using chenodeoxycholic acid (CDCA, 10
mM) as coadsorbant. The photovoltaic characteristics of the devices are reported in Table 2.
Unfortunately, no efficiency improvement could be observed, suggesting that the geometry of the
di-branched molecules hampers the efficient intercalation of the CDCA at the TiO2 surface. As a
consequence the additive competes with the dye molecules for the anchoring sites, slightly
degrading the performances.
Beside the quantum conversion efficiency, another fundamental issue to be addressed to qualify a
sensitizer as promising refers to the device stability. The 20-years durability is in fact a key
requirement for the commercialization of dye sensitized solar cells. Although many factors affect
the lifetime of DSSCs the anchoring stability of the adsorbed sensitizers is an issue that should be
regarded as critical. In particular, it is well known that the presence of even small traces of water
impairs long term stability of the cells since it facilitates dye desorption [26]. Hence, a robust and
19

ACCEPTED MANUSCRIPT
indissoluble anchoring on the TiO2 surface is also an indispensable feature of any “good” sensitizer.
To investigate this aspect, we also aimed at assessing the anchoring strength of DYE1 and DYE2 in
comparison with the reference molecules D5 and N719. Dye-sensitized nanocrystalline TiO2 films
were thus immersed in an acetonitrile solution containing a 5%w of water. The stability of dyes
adsorbed on the photoanodes was evaluated by carrying out UV-Vis absorption measurements at
different time intervals, as shown in Figure 9. The D5-sensitized photoelectrode exhibited a
remarkable drop of the optical density as a consequence of the water promoted carboxylate group
cleavage from the TiO2 surface bond. Remarkably, no substantial changes in the absorbance spectra
were observed for DYE1- and DYE2-sensitized films.
2,5
2,0
1,5
1,0
0,5
0,0
2,0
1,5
1,0
0,5
0,0
DYE1
t=0 h
t=4 h
D5
t=0 h
t=4 h
t=14 h
t=14 h
350 400
450
500 550
600
650 700
750
350 400 450 500 550 600 650 700 750
Wavelength (nm)
Wavelength (nm)
1,2
2,5
2,0
1,5
1,0
0,5
0,0
1,0
0,8
0,6
0,4
0,2
DYE2
t=0 h
t=4 h
t=14 h
N719
t=0 h
t=4 h
t=14 h
350 400
450 500
550 600
650 700 750
350 400
450
500
550 600
650
700
750
Wavelength (nm)
Wavelength (nm)
Figure 9. Absorption spectra of D5-, DYE1-, DYE2- and N719- sensitized nanocrystalline TiO2
films before and after 4 hours and 14 hours of desorbing treatment.
20

ACCEPTED MANUSCRIPT
This difference in behavior is clearly correlated to the exceptional robustness of their di-anchoring
mode as well as to their superior hydrophobicity imparted by the long pending n-octyl chain.
Figure 10 shows a picture of D5, DYE1, DYE2 and N719 adsorbed on TiO2 films before the
treatment in acetonitrile/water solution, then after 4 hours and 14 hours of immersion in solution.
As evident, the purplish color imparted by the N719 almost completely vanished after the treatment,
and this observation is accompanied by the almost complete disappearance of its UV-Vis peak as
the result of the detachment of dye molecules from the semiconductor surface. The desorption is
much more contained for the organic dyes, being lowest for DYE1 and DYE2 as evidenced by the
UV-vis measurements described above.
Figure 10. Picture of D5-, DYE1-, DYE2- and N719-sensitized nanocrystalline TiO2 films before
the treatment with the acetonitrile/water solution, after 4 hours and 14 hours of desorbing treatment.
3. Conclusions
21

ACCEPTED MANUSCRIPT
We have designed and synthesized new D(-π-A)2 organic molecules DYE1 and DYE2,
incorporating a thiophene or bithiophene unit as the π-bridge between the carbazole donor and the
cyano-acrylic acceptor. The absorption, energy levels and frontier orbitals of DYE1 and DYE2
were studied in detail. We found that the D(-π-A)2 organic dyes show an additional strong transition
in comparison to the common D-π-A dyes, resulting in a broadening of the absorption range.
Electrochemical impedance spectroscopy showed that the injected electron lifetime were increased
for DYE2 devices, that can be attributed to the introduction of a further thiophene unit in the π-
bridge with respect to DYE1. The highest power conversion efficiency reaches 5.01% in the case of
DYE2 with a photocurrent of 10.52 mA cm–2 and a photovoltage of 0.70 V, substantially deriving
from a broader absorption with respect to DYE1, as also confirmed by IPCE measurements.
Stability tests aimed at assessing the anchoring strength of DYE1 and DYE2 revealed their
exceptional robustness, deriving from the di-anchoring architecture, in comparison with the
common reference dyes. These results confirm and encourage the use of a dianchoring architecture
for the development of robust and efficient organic dyes and stimulate more research in the field
aimed at their optimization for practical application.
4. Experimental section
4.1. General remarks
Reactants were purchased from commercial sources and used without further purification. All
manipulations were carried out under inert nitrogen atmosphere using standard Schlenk techniques.
All solvents used were carefully dried and freshly distilled according to standard laboratory
practice. Flash chromatography was performed using a silica gel of 230–400 mesh.1H-NMR and
13C{1H}-NMR spectra were recorded on a Bruker Avance 700 MHz. The high resolution
electrospray mass spectrometry analyses were performed using a Bruker micro TOF QII mass
22

ACCEPTED MANUSCRIPT
spectrometer equipped with an electrospray ion source operated in negative ion mode. The sample
solutions (CH2Cl2/CH3OH) were introduced by continuous infusion at a flow rate of 180 ꢁL min−1
with the aid of a syringe pump. The instrument was operated with end-plate offset and capillary
voltages set to 500 V and 3500 V, respectively. The nebulizer pressure was 0.8 bar (N2), and the
drying gas (N2) flow rate was 7.0 L min−1. The capillary exit and skimmer 1 voltages were 90 V and
30 V, respectively. The drying gas temperature was set at 180°C. The calibration was carried out
with sodium formiate. Elemental analyses were obtained on a EuroVector CHNS EA3000
elemental analyser. FT-IR measurements were recorded on a JASCO FT/IR 4200 instrument. UV-
Vis spectra were recorded on a Jasco V-670 instrument and fluorescence spectra were obtained on a
Varian Cary Eclipse spectrofluorimeter. Cyclic voltammetry was carried out on a Metrohm Autolab
PGSTAT 302-N potentiostat. Measurements were carried at 25 °C in THF solution (~10–4 M)
containing tetrabutylammonium tetrafluoroborate (0.025 M) as supporting electrolyte with a scan
rate of 100 mVs–1. The potentials were measured versus Ag/Ag+ as the quasi-reference electrode.
Subsequently to each experiment, the potential of the Ag/Ag+ electrode was calibrated against the
ferrocene/ferrocenium (Fc/Fc+) redox couple. The HOMO energy values were estimated from the
onset potentials (Eoxonset) of the first oxidation event. After calibration of the measurements against
Fc/Fc+, the oxidation potential of which is assumed at + 4.8 eV below the vacuum level, the HOMO
energy level were calculated according to the following equations:
EHOMO (eV) = – [Eoxonset – E1/2(Fc/Fc+) + 4.8]
where E1/2(Fc/Fc+) is the half-wave potential of the Fc/Fc+ couple against the Ag/Ag+ electrode. All
theoretical calculations were carried out with the Gaussian09 program package. Analyses of the
ground-state structures for the molecules were carried out using density functional theory (DFT)
23

ACCEPTED MANUSCRIPT
and the B3LYP function [27] was used in conjunction with the 6-31G(d,p) basis set. Time-
dependent DFT (TDDFT) calculations performed to assess the excited-state transition energies.
4.2. Fabrication of DSSC and photovoltaic measurements
Fluorine-doped tin oxide (FTO, 10 ohm/sq., provided by SolaronixS.A.) glass plates were first
cleaned in a detergent solution using an ultrasonic bath for 15 min, and then rinsed with water and
ethanol. For the fabrication of photoelectrodes, two different titanium nanoxide pastes were
deposited by doctor blading: 18NR-T purchased by Dyesol, which gave transparent TiO2 films 7
ꢁm-thick, and D/SP paste provided by Solaronix, which was used as active-opaque overlayer for 17
ꢁm-thick double-layered photoanodes. The doctor blading procedure was repeated to obtain the
desired film thickness, finally measured using a profilometer (Tencor Alpha-Step 500 Surface
Profiler). The active area of photoanodes was 0.16 cm2. The electrodes coated with the TiO2 pastes
were gradually heated under an air flow and sintered at 450 °C for 30 min. The substrate
temperature was then allowed to slowly decrease. Once cooled down at about 80°C, the electrodes
were immersed into 0.2 mM solutions of dyes DYE1 and DYE2 in EtOH and kept for 12 h in dark
at room temperature. The reference photoanodes were prepared by dyeing into a solution 0.2 mM of
(bis(tetrabutylammonium)-cis-di(thiocyanato)-N,N′-bis (4-carboxylato-4′-carboxylic acid-2, 2-
bipyridine) ruthenium(II) (N719, provided by Solaronix S.A.) in a mixture of acetonitrile and tert-
butyl alcohol (v/v = 1:1) at room temperature for 14 h. 3-[5-(4-(Diphenylamino)styryl)-thiophen-2-
yl]-2-cyano-acrylic acid (D5) and the corresponding photoanode were prepared according to a
literature procedure [28]. The counter electrodes were prepared by sputtering a 50 nm Pt layer on a
hole drilled cleaned FTO plate. The two plates were faced and assembled by means of a gasket of
50ꢁm-thick Surlyn® foil (Dyesol Ltd) interposed between them. The redox electrolyte (0.1 M LiI,
0.05 M I2, 0.6 M 1-methyl-3-propylimidazolium iodide, and 0.5 M tert-butylpyridine in dry
acetonitrile) was vacuum-injected into the space between the electrodes. Photocurrent-voltage I-V
24

ACCEPTED MANUSCRIPT
measurements were performed using a Keithley unit (Model 2400 Source Meter). A Newport AM
1.5 Solar Simulator (Model 91160A equipped with a 300W Xenon Arc Lamp) serving as a light
source. The light intensity (or radiant power) was calibrated to 100 mW/cm2 using as reference a Si
solar cell. The incident photon-to-current conversion efficiency (IPCE) was measurement by a DC
method. IPCE measurements were carried out with a computerized setup consisting of a xenon arc
lamp (140 W, Newport, 67005) coupled to a monochromator (Cornerstore 260 Oriel 74125). Light
intensity was measured by a calibrated UV silicon photodetector (Oriel 71675) and the short circuit
currents of the DSSCs were measured by using an optical power/energy meter, dual channel
(Newport 2936-C). To evaluate the dye loading, transparent 5-µm thick photoelectrodes (1 cm × 1
cm in size) were stained with DYE1, DYE2 and D5, following the same procedures described
above. Subsequently, the substrates were thoroughly rinsed with the corresponding solvent, in order
to remove the fraction of dye that was not chemisorbed onto the TiO2 surface, and were eventually
placed in a KOH solution in DMF (0.05 M) for 48 h in order to achieve a complete desorption, as
testified by substrate discoloration. The evaluation of the dye concentration in the solvent, obtained
by UV-Vis measurements, allowed the calculation of the amount of the adsorbed molecules,
expressed in terms of moles of dye anchored per projected unit area of the photoelectrode.
Electrochemical impedance spectroscopy (EIS) was performed by an AUTOLAB PGSTAT 302N
(Eco Chemie B.V.) in a frequency range between 300 kHz–30 mHz. The impedance measurements
were carried out at different voltage biases under 1 sun illumination. The resulting impedance
spectra were fitted with ZView software (Scribner Associates).
4.3. Synthesis
4.3.1. 9-Octyl-3,6-bis(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-9H-carbazole (2)
25

ACCEPTED MANUSCRIPT
A solution of n-BuLi (1.6 M in hexanes, 8.2 mL, 13.17 mmol) was added dropwise to a solution of
1 [29] (1.92 g, 4.39 mmol) in THF (80 mL) kept at –80 °C. The obtained mixture was vigorously
stirred for 1 hour at –80 °C before the addition of 2-isopropoxy-4,4,5,5-tetramethyl-[1,3,2]-
dioxaborolane (2.45 g, 13.17 mmol) in one portion. The resulting solution was then allowed to
reach room temperature and to react overnight. After removing the solvent under vacuum, CH2Cl2
(50 mL) was added and the obtained solution was washed with water (3 × 50 mL) and dried over
Na2SO4. The solvent was removed under vacuum and the obtained crude product was purified by
flash column chromatography (SiO2, petroleum ether 40-60 °C/CH2Cl2 = 1/1 v/v) to give 2 (1.42 g,
1
61%) as a white solid. H-NMR (700 MHz, CDCl3): δ 8.68 (s, 2H), 7.87-7.56 (m, 4H), 4.36 (t, J =
7.0 Hz, 2H), 1.95-1.90 (m, 2H), 1.44-1.39 (m, 2H), 1.38-1.33 (m, 2H), 1.32-1.23 (m, 6H), 0.88 (t, J
13
= 7.0 Hz, 3H) ppm. C{1H} NMR (176 MHz, CDCl3): δ 144.6, 141.2, 132.1, 128.4, 124.3, 111.4,
83.9, 43.6, 31.7, 29.3, 29.1, 29.0, 27.2, 25.4, 22.6, 14.0 ppm.
4.3.2. 5,5'-(9-Octyl-9H-carbazole-3,6-diyl)dithiophene-2-carbaldehyde (3)
A mixture of 5-bromo-thiophene-2-carbaldehyde (0.52 g, 2.72 mmol), 2 (0.72 g, 1.35 mmol) and
Pd(PPh3)4 (23 mg, 0.02 mmol) in toluene (16 mL) and a 2.0 M K2CO3 aqueous solution (8 mL) was
refluxed overnight. After cooling the mixture down to room temperature, CH2Cl2 (40 mL) was
added and the obtained mixture was washed with water (3 × 40 mL). After solvent removal, the
crude product was purified by flash chromatography (SiO2, CH2Cl2) to afford 3 (0.55 g, 81%) as a
yellow solid.1H-NMR (700 MHz, CDCl3): δ 9.92 (s, 2H), 8.47 (d, J = 1.4 Hz, 2H), 7.85 (dd, J = 8.5,
1.4 Hz, 2H), 7.81 (d, J = 3.9 Hz, 2H), 7.51 (d, J = 3.9 Hz, 2H), 7.49 (d, J = 8.5 Hz, 2H), 4.36 (t, J =
7.0 Hz, 2H), 1.95-1.90 (m, 2H), 1.44-1.39 (m, 2H), 1.38-1.33 (m, 2H), 1.32-1.23 (m, 6H), 0.88 (t, J
13
= 7.0 Hz, 3H) ppm. C{1H}-NMR (176 MHz, CDCl3): δ 182.6, 155.7, 141.7, 141.6, 137.7, 125.2,
26

ACCEPTED MANUSCRIPT
124.9, 123.3, 123.2, 118.7, 109.8, 43.6, 31.7, 29.3, 29.1, 29.0, 27.2, 22.6, 14.0 ppm. Elemental
analysis for C30H29NO2S2: calcd C, 72.11; H, 5.85; N, 2.80; found C, 72.00; H, 5.80, N, 2.77.
4.3.3. 2,2'-Bithiophene-5-carbaldehyde (4)
A mixture of 5-bromo-thiophene-2-carbaldehyde (0.52 g, 2.72 mmol), tributyl(thiophen-2-
yl)stannane (1.02 g, 2.72 mmol) and Pd(PPh3)4 (23 mg, 0.02 mmol) in toluene (20 mL) was
refluxed overnight. After cooling the mixture down to room temperature, CH2Cl2 (40 mL) was
added and the obtained mixture was washed with water (3 × 40 mL). After solvent removal, the
crude product was purified by flash chromatography (SiO2, CH2Cl2) to afford 4 (%) as a pale
1
yellow solid. H-NMR (700 MHz, CDCl3): δ 9.70 (s, 1H), 7.55 (d, J = 4.0 Hz, 1H), 7.30-7.10 (m,
3H), 6.97 (d, J = 3.9 Hz, 1H) ppm. 13C{1H}-NMR (176 MHz, CDCl3): δ 182.7, 155.7, 141.9, 136.7,
135.1, 131.5, 126.7, 123.2, 113.8 ppm.
4.3.4. 5'-Bromo-2,2'-bithiophene-5-carbaldehyde (5)
N-bromosuccinimide (1.78 g, 10.0 mmol) was portionwise added to a solution of 4 (1.94 g, 10.0
mmol) in DMF (60 mL) kept at 0 °C. The system was then allowed to reach room temperature and
to react overnight. The reaction mixture was poured into water (500mL) and the resulting
precipitate was filtered, washed with water and purified by flash chromatography (SiO2, CH2Cl2) to
give 6 in 86% yield as a pale yellow solid. 1H-NMR (700 MHz, CDCl3): δ 9.89 (s, 1H), 7.68 (d, J =
4.0 Hz, 1H), 7.20 (d, J = 4.0 Hz, 1H), 7.13 (d, J = 3.9 Hz, 1H), 7.06 (d, J = 3.9 Hz, 1H) ppm.
13C{1H}-NMR (176 MHz, CDCl3): δ 182.4, 145.8, 142.1, 137.5, 137.1, 131.2, 126.2, 124.4, 114.2
ppm.
27

ACCEPTED MANUSCRIPT
4.3.5. 5',5''-(9-Octyl-9H-carbazole-3,6-diyl)di-2,2'-bithiophene-5-carbaldehyde (6)
A mixture of 5 (0.74 g, 2.72 mmol), 2 (0.72 g, 1.35 mmol) and Pd(PPh3)4 (23 mg, 0.02 mmol) in
toluene (16 mL) and a 2.0 M K2CO3 aqueous solution (8 mL) was refluxed overnight. After cooling
the mixture down to room temperature, CH2Cl2 (40 mL) was added and the obtained mixture was
washed with water (3 × 40 mL). After solvent removal, the crude product was purified by flash
1
chromatography (SiO2, CH2Cl2) to afford 6 (0.57 g, 64%) as an orange solid. H-NMR (700 MHz,
CDCl3): δ 9.90 (s, 2H), 8.39 (s, 2H), 7.78 (d, J = 8.8 Hz, 2H), 7.72 (d, J = 3.9 Hz, 2H), 7.46 (d, J =
8.8 Hz, 2H), 7.41 (d, J = 3.6 Hz, 2H), 7.36 (d, J = 3.6 Hz, 2H), 7.31 (d, J = 3.9 Hz, 2H), 4.36 (t, J =
7.0 Hz, 2H), 1.95-1.90 (m, 2H), 1.44-1.39 (m, 2H), 1.38-1.33 (m, 2H), 1.32-1.23 (m, 6H), 0.88 (t, J
= 7.0 Hz, 3H) ppm.13C{1H} NMR (176 MHz, CDCl3): δ 182.4, 147.6, 141.3, 140.9, 137.4, 134.0,
127.3, 125.2, 124.5, 123.6, 123.3, 123.2, 118.0, 109.6, 43.5, 31.8, 29.3, 29.1, 29.0, 27.3, 22.6, 14.0
ppm. Elemental analysis for C38H33NO2S4: calcd C, 68.74; H, 5.01; N, 2.11; found C, 68.71; H,
4.95, N, 2.00.
4.3.6. 3,3'-(5,5'-(9-Octyl-9H-carbazole-3,6-diyl)bis(thiophene-5,2-diyl))bis(2-cyanoacrylic acid)
(DYE1).
A mixture of 3 (150 mg, 0.30 mmol), cyanoacetic acid (76 mg, 0.90 mmol) and ammonium acetate
(8 mg, 0.10 mmol) in acetic acid (10 mL) was refluxed overnight. After cooling to room
temperature, the reaction mixture was poured into water (100 mL) obtaining a precipitate that was
filtered and repeatedly washed with water. The obtained crude product was purified by flash
column chromatography (SiO2, CH2Cl2/ethanol = 7/1 v/v) to afford DYE1 (130 mg, 68%) as a red
powder. 1H-NMR (700 MHz, CDCl3): δ 8.78 (s, 2H), 8.40 (s, 2H), 7.97 (d, J = 4.0 Hz, 2H), 7.90 (d,
J = 8.7 Hz, 2H), 7.80 (d, J = 4.0 Hz, 2H), 7.47 (d, J = 8.7 Hz, 2H), 4.35 (t, J = 6.9 Hz, 2H), 1.95-
1.90 (m, 2H), 1.44-1.39 (m, 2H), 1.38-1.33 (m, 2H), 1.32-1.23 (m, 6H), 0.88 (t, J = 7.0 Hz, 3H)
28