This document is the Accepted Manuscript version of a Published Work that appeared in final form in Energy Environ. Sci.,
2015, 8, 2010 DOI: 10.1039/c5ee00444f

Metal-free organic sensitizers with narrow absorption in
the visible for solar cells exceeding
10% efficiency
D. Joly; L. Pelleja; S. Narbey; F. Oswald; T. Meyer; Y. Kervella; P. Maldivi;
`
J. N. Clifford; E. Palomares; R. Demadrille
A novel family of six donor–acceptor type organic sensitizers for dye-sensitized solar cells (DSSCs) is reported. The dyes
have been designed to have outstanding light absorption properties in the visible range and being able to achieve high
photon-to-electrical current conversion for BIPV (building-integrated photovoltaic). Moreover, stability tests under illumination
at 1 Sun and 65 1C showed a great stability for some of the devices, with less than 6% decrease of power conversion
efficiency after 3000 hours. The differences in the performance of the six sensitizers under standard illumination
conditions can be correlated with the observed differences in the photo-induced transient photovoltage and in charge
extraction measurements. We report the use of one of the dyes for the fabrication of semi-transparent
solar modules showing an active area of 1400 cm2 and a power output of 10.5 W m-2.

Broader context
Tremendous progress has been achieved in the past decade in the effciency of dye-sensitized mesoscopic solar cells (DSSC), owing to the development of new metal-free organic
sensitizers. To increase their commercial attractiveness and to make them a realistic approach towards energy markets such as BIPV (building integrated photovoltaics), the new dyes
must be cheap, colorful, and effcient, and moreover they must show a high stability. In this paper, we report a family of donor–acceptor type metal-free photosensitizers with simple structures
for DSSCs. The precise molecular design of the dyes leads to a quite narrow light absorption in the visible range and a high photon-to-electrical current conversion in solar cells. These
organic sensitizers that can be prepared in few steps show power efficiencies over 10% when they are used with iodine-based liquid electrolytes. Stability tests and fabrication of
semitransparent solar modules
with an active area of 1400 cm2 demonstrate the potential of these organic dyes for large scale applications and mass production for BIPV.

1 Introduction
Among the novel emerging photovoltaic technologies suitable for BIPV,
dye-sensitized solar cells (DSSCs) have the potential of becoming a clean
and affordable source of renewable energy. Their high performance
combined with low-cost manufacturing processes and steadily improved
stability has drawn tremendous
a

Univ. Grenoble Alpes, INAC-SPRAM, F-38000 Grenoble, France
CNRS, Alpes, INAC-SPRAM, F-38000 Grenoble, France
c
CEA, INAC-SPRAM, F-38000 Grenoble, France. E-mail: renaud.demadrille@cea.fr
d
17 Rue des Martyrs, 38054 Grenoble Cedex 9, France
e
Institute of Chemical Research of Catalonia (ICIQ), Avenguda Pa¨sos Catalans, 16,
ı
Tarragona 43007, Spain. E-mail: epalomares@iciq.es
f
Solaronix SA, Rue de l’Ouriette 129, 1170 Aubonne, Switzerland
g
Univ. Grenoble Alpes, CEA, INAC-SCIB, F-38000 Grenoble, France
h
ICREA, Passeig Llu´ Companys, 23, Barcelona, E-08010, Spain
ıs
b

interest from both academic and industrial researchers. In these devices,
photons are absorbed by photoactive molecules that are anchored to the
surface of a mesoporous wide band gap semi- conductor oxide such as
TiO2,1a–c ZnO,1d or Zn2SnO4,1e, f which is deposited on a transparent
conductive oxide.1 The system is
usually completed with a liquid electrolyte containing a redox shuttle. In a
DSSC, the sensitizer is a crucial element, because it is responsible for the
absorption of photons and the generation of free charges after the injection
of electrons from the excited state of the dye into the semi-conducting oxide.
Furthermore the dye acts as the first ‘‘barrier’’ for back-electron transfer from
the photo-injected electrons to the oxidised electrolyte.
Presently, the dyes for DSSCs can be classified into three main categories:
(i) ruthenium-containing dyes such as polypyridyl complexes, (ii) metallated
porphyrins and (iii) fully organic dyes whose structure is based on a typical
‘‘electron donor–p-bridge– electron acceptor’’ (D–p–A) design.2
Ruthenium-dye based DSSCs can show high effciency (10–11%)
due to their wide absorption spectra ranging from the visible to

.

the near-infrared (NIR), but on the other hand, the use of toxic and scarce
ruthenium inhibits their application for mass production.
Ring-metalated porphyrin dyes were used for the first time
as sensitizers in DSSCs in 2007.3 Taking advantage of their large absorption
coefficients in the visible region associated with their Soret and Qbands, impressive progress has been made with this class of materials
leading to power conversion efficiencies (PCEs) up to 12–13% when a
cobalt-based electrolyte is employed.4 However, the multiple steps that are
necessary for their preparation, the low yield of porphyrin formation and
the
inherent difficulties in their purification indicate that they will not be
suitable candidates for use in DSSC technology in the near future. On
the other hand, fully organic sensitizers present many advantages as they
can be prepared in a few steps, they can be purified using simple methods,
and thanks to molecular engineering one can conveniently tune their energy
band gap to
modulate their light-harvesting capabilities.5a,b
To date, when combined with the classical iodide/tri-iodide
(I-/I3 -) electrolyte, only a few metal-free organic sensitizers are
capable of achieving over 9% photovoltaic effciency.5c–f The iodide/triiodide redox shuttle (I-/I3-) is the most widely used system, because it
has favorable penetration ability into the pores of the semiconductor
oxide film, and it gives rise to a very fast dye regeneration and relatively
slow recombination with injected photoelectrons.6 The main drawback
is the fact that the redox potential of (I-/I3-) couple limits the
photovoltage. However this system is until now the only redox couple
which has been proven to have long term stability.7
For the development of devices applicable in BIPV such as skylights
or façades, it now appears crucial to develop new efficient and robust sensitizers
with selected absorption in the visible.
In this paper we report a new family of D–p–A organic dyes
revealing a narrow absorption in the visible that can be combined
with (I-/I3-) based electrolytes for the development of highly efficient and
stable semi-transparent solar cells. These dyes are based on the molecular
structure of RK1, a robust organic dye that achieved 10.20% efficiency.5d
The performance is comparable with that of conventional N719,
which gave identical efficiency under the same experimental conditions.
Five different dyes were prepared and fully characterized by
electrochemical and spectroscopy techniques. The photovoltaic

Fig. 1 Synthetic routes and chemical structures of the organic sensitizers.

performances of these sensitizers in DSSC devices were inves- tigated
under standard A.M 1.5G simulated solar radiation. After optimization of the
electrode thickness and dyeing conditions, and using a volatile electrolyte,
all the dyes showed power con- version effciencies over 8.50% and four of
them had efficiencies ranging from 9.51% to 10.11% with current
densities ( Jsc)
spanning from 16.55 mA cm-2 to 18.82 mA cm-2.
When the volatile electrolyte was replaced by an ionic liquid based one,
effciencies up to 7.87% with Jsc up to 17.50 mA cm-2 were obtained. These
values are among the highest reported for organic sensitizers with a nonvolatile electrolyte. Herein, we discuss the difference in DSSC performances
utilizing these dyes by monitoring the electron transfer processes in complete
devices,
and using a full set of electrochemical and optical data combined with
theoretical modelling.

2. Results and discussion
2.1

Design and synthesis of the RK sensitizers

The chemical design of the sensitizers was inspired by the molecular
structure of RK1, an effcient orange dye that can be obtained in only five
steps with an overall yield of 50%. One should note that the development
of convenient and up-scalable synthetic methods is important for future
development of DSSCs at the industrial level. With the goal to tune the
absorption spectrum of RK1 and to improve its solubility in various nonhalogenated solvents, several strategies were followed in this work (Fig.
1).
First, alkoxy groups or alkyl substituents were introduced in the para
position of the triphenylamine (TPA) unit leading respectively to dyes
6ORK1 and 6RK1.
In addition, to confer better solubility, the introduction of long alkyl
and alkoxy groups on the electron-donating moiety is expected to
produce a red-shift in the absorption spectra8 and to diminish dye
aggregation and charge recombination,9 leading to an increase in the Jsc
value.10 Therefore, the linear octyl substituent of the thiophene unit in RK1
was replaced by a branched alkyl chain, i.e. 2-ethylhexyl, leading to dye
RK2. The presence of the branched 2-ethylhexyl chain is expected to
suppress the charge recombination between electrons in the

electrolyte and the TiO2 film more efficiently than the linear octyl
chain.11
In order to shift the position of the internal charge transfer absorption
band (ICT) in the visible range towards higher wavelengths and, as a
consequence to tune the color of the dyes, two approaches have been
investigated. First, the sulfur atom of the benzothiadiazole electron-accepting
unit was replaced by a selenium atom12 giving rise to the dye RKSe.
Second, the
rigidification of the backbone was carried out with the goal of improving
the delocalization of the p-electrons over the TPA and the p-conjugated
bridge.13 This was achieved by fusing the phenyl ring of the TPA with the
thiophene unit of the central unit leading to the dye RKF. The substitution
of the resulting indeno[1,2-b]- thiophene moiety with two octyl chains will
ensure good solubility for this dye.
The synthetic strategies developed to access the different dyes are
illustrated in Fig. 1. These involve various palladium- catalyzed crosscoupling reactions. Suzuki coupling was employed to prepare the TPA(alkyl)thiophene building blocks 3, 4, 5, and
6 whereas a Buchwald–Hartwig coupling reaction was used to
prepare compound 2, namely TPA-(4,4-dioctyl-4H-indeno[1,2-b]- thiophene).
Finally a Stille coupling reaction was chosen to access the six different aldehyde
intermediates starting from the stannyl- ated TPA-based precursors and the
brominated benzothiadiazole or benzoselenadiazole benzaldehyde reagents 7
and 8 respectively.
All the compounds bearing an aldehyde functional group, i.e. 9, 10,
11, 12, 13, and 14, were subsequently converted into the corresponding
dyes by Knoevenagel condensation with cyano- acrylic acid in the presence
of piperidine. RK2, 6ORK1, 6RK1, RKSe and RKF were obtained with
average to good overall yields ranging from 11% to 41%. All the
intermediates and the sensiti- zers were fully characterized; the synthetic
procedures and analytical data are reported in the ESI.†

.

2.2

Optical and electrochemical data and DFT calculations

The UV-vis absorption spectra of the five new dyes in CH2Cl2 solution
and on 2 mm thick TiO2 films are shown in Fig. 2 and compared to that of
RK1 which is our reference dye. The absorp- tion bands as well as their
extinction coefficients are summarized in Table 1.
As expected for the commonly used D–p–A structure, two sets of
absorption bands are observed. The first one is attri- buted to the p–p*
transitions of the aromatic rings being in conjugation and is located in
the UV region with an extension close to the frontier between the UV and
the visible regions.
The second one is associated with the intramolecular charge transfer
(ICT) absorption and is clearly located in the visible region. The
replacement of the linear alkyl chain on the thio- phene unit by a branched
one (RK2) has absolutely no influence on the position and the width of the
UV absorption band. The introduction of alkoxy or alkyl groups on the TPA
unit in the para position (6ORK1 and 6RK1) as well as by a Se atom
(RKSe) induces a slight bathochromic shift of the UV band of circa
15 nm in solution. The effect of the bridged unit resulting from the fusion
between the TPA and the thiophene unit in RKF is more pronounced. In
this case two absorption bands are observed

Fig. 2

Absorption spectra of RK dyes in CH2Cl2 (top), on TiO2 films (bottom).

in the UV and the lmax of the UV band is bathochromically shifted by 38 nm
compared to RK1.
As mentioned before, all of the synthesized dyes have absorption
maxima in the visible region (between 475 and
535 nm) originating from an intramolecular charge transfer.
The intensity and the position of the ICT bands are strongly dependent
on the modification of the chemical structure. The replacement of the
octyl group in RK1 by a 2-ethylhexyl group in RK2 results in a slight
hypsochromic and hypochromic effect on the ICT band. This may be
explained by the steric hindrance induced by the branched chain that
increases the torsion angle between the TPA and the thiophene unit.
This torsion impedes good delocalization of the p electrons
and blue-shifts the position of the ICT band. On the other hand, the
introduction of alkyl and alkoxy donating units on the TPA moiety
induces a red-shift of 6 nm and 14 nm respec- tively for 6RK1 and
6ORK1 leading to more reddish dyes. This shift is associated to a
decrease of the molar extinction coefficient attributed to the increased
proportion of the alkyl chains. A more pronounced bathochromic effect is
observed when the benzothiadiazole electron withdrawing unit is
replaced by a benzoselenadiazole unit in RKSe. This shift is attributed to
the higher electron accepting ability of the benzoselenadiazole unit as
confirmed by the better localization of the LUMO energy level on this
unit (see DFT calculations, ESI,† Fig. S3). Herein the decrease of the
molar absorption coefficient is more important which could be understood
by the low oscillator strength of the charge-transfer transition according to
the Franck–Condon
principle.9b
The most important bathochromic effect of the ICT band is
observed for RKF yielding a pink-violet dye. The indeno- thiophene
bridged unit plays a critical role in the enhanced absorption of RKF.
Indeed, the planarization of the TPA with the adjacent thiophene leads to
an increased delocalization of

Table 1 Optical and electrochemical data of the RK dyes

lmax
UVa [nm]
RK1
RK2
6RK1
6ORK1
RKSe
RKF

lmax
Visa [nm]

lema [nm]

ea
[M-1 cm-1]

D E opt b [eV]

E(S+/S)c
[eV]

E(0–0)d
[eV]

E(S+/S*)e [V]

LUMO f
[eV]

HOMO f
[eV]

D Eg [eV]

330
330
344
345
343
368

484
477
490
498
510
533

713
720
732
578
745
733

27 600
26 900
23 000
19 900
18 200
25 900

2.12
2.16
2.09
2.00
2.00
1.87

-5.62
-5.65
-5.53
-5.39
-5.64
-5.45

2.15
2.06
2.10
2.30
1.97
2.04

-3.47
-3.6
-3.44
-3.08
-3.67
-3.41

-3.22
-3.21
-3.22
-3.21
-3.42
-3.21

-5.23
-5.23
-5.12
-5.03
-5.23
-5.14

2.01
2.02
1.90
1.82
1.81
1.93

(364)
(371)
(363)
(364)
(364)
(412)

(471)
(472)
(484)
(495)
(505)
(553)

a

Measured in CH2Cl2, in parentheses measured onTiO2 (2 mm). b Measured on TiO2 (2 mm) from lonset. c All potentials were obtained during cyclic voltammetric
investigations in 0.2 M Bu4NPF6 in CH2Cl2. Platinum electrode diameter 1 mm, sweep rate 200 mV s-1. Potentials measured vs. Fc+/ Fc were converted to NHE by addition
of +0.69 V. d The zero–zero excitation energies, E(0–0), are estimated from the intercept of the normalized absorption and emission spectra. e The excited state oxidation
potentials were derived from the equation E(S+/S*) = E(S+/S) – E(0–0). f Calculation B3LYP. g D E = HOMO–LUMO from calculation.

the HOMO energy level as revealed by DFT calculations. The higher
electron-donating ability of this unit and the better delocalization of
electrons shift by ca. 50 nm the position of the ICT band.
When the dyes are absorbed on the TiO2 surface, they undergo a
red shift of their UV absorption band and a slight hypsochromic shift of
the visible absorption band is observed for all the dyes (from 4 to 13 nm)
except for RKF. With this dye we observed a red-shift of the visible
absorption band. This may indicate that the dyes have the propensity to
form H-aggregates onto the TiO2 surface whereas RKF tends to form Jaggregates owing to the two alkyl substituents attached to the spiro-carbon.
The electrochemical properties of the dyes were investigated by cyclic
voltammetry (CV) in dichloromethane with the goal to determine the
energy level positions of their frontier orbitals. Upon application of positive
potentials they showed two reversible oxidation waves corresponding to an
oxidation/re-reduction pro- cess (see ESI,† Fig. S1). Upon application of
negative potentials, the dye undergoes a quasi-reversible reduction/re-oxidation
process. As shown in Table 1, the LUMO energy levels that are determined
from the onset of the reduction waves are lying between -3.57 and
-3.78 eV for all the sensitizers. The lowest LUMO value was
found for the selenium-containing dye, reflecting the increased electronwithdrawing character of the benzoselenadiazole unit. All the LUMO
levels are higher than the conduction band edge (ECB) of titania, whose
commonly accepted values in the DSSC literature are found between -0.2
and -0.5 V vs. NHE, that is, between -4.25 and -3.95 eV vs. vacuum.14
The exact position of the ECB of TiO2 is difficult to determine since
it can be shifted depending on the acid functionalized dyes and additive
molecule’s dipole moment.15 Due to the presence of the benzoselenadiazole, the LUMO energy level of RKSe approaches that of the
conduction band of TiO2 which may limit the efficiency of electron
injection into the TiO2 film. For the other dyes the energy difference
between the LUMO level and the ECB of the semiconductor is high
enough (0.3 eV) to ensure sufficient thermodynamic driving force for the
injection of electrons from the excited state of the dyes into the oxide.16
As expected, the oxidation potentials of RK1, RK2 and RKSe are quite
similar since they possess the same TPA-thiophene unit. The increased
electron-donating ability of the substituted or fused TPA units slightly
raises the HOMO level of 6RK1, 6ORK1 and RKF. All the

HOMO energy levels of these sensitizers range from -5.37 to
-5.58 eV, indicating that the redox couple (I-/I3-) is an appro- priate
redox system for the regeneration of the oxidized dyes.
To determine the geometrical configuration and the electro- nic
structure of the dyes, molecular orbital calculations were performed by
using the hybrid B3LYP functional in combi- nation with the triple zeta
basis set.17 The good spatial separa- tion of the frontier molecular orbitals
was evidenced by DFT calculations (see ESI,† Fig. S2). Isodensity plots
of the HOMO and LUMO energy levels of the dyes show a directional
electron distribution, with the delocalization of the HOMOs on the
TPA group and the thiophene and the LUMOs predominantly delocalized
on the benzothiadiazole (or benzoselenadiazole) phenylvinylcyanoacetic
acid units. This directional electron distribution is consistent with the
dissymmetrisation of the core. However, one should note that in the case
of RKSe the LUMO energy is preferentially localized over the
benzoselenadiazole unit with a small contribution from the anchoring group.
Indeed the DFT calculations indicate that the contribution of the benzoselenadiazole to the LUMO of the molecule is 45.5% whereas the
contribution of benzothiadiazole is only 36.7%. This highlights the
enhanced electron-withdrawing strength of this unit, explain- ing the shift in
absorption that is observed for the ICT band. However the higher
localization of the LUMO level mainly on the benzoselenadiazole may be
detrimental to the injection process of electrons into the conduction band of
the TiO2.
2.3

Photovoltaic devices with a standard liquid electrolyte

Fig. 3 shows the action spectra of incident photon-to-current conversion
efficiency (IPCE) and the photocurrent–voltage (I–V) plots of DSSCs
fabricated with these dyes. The detailed photo- voltaic parameters of the
( Jsc), open-circuit
best performing cells, i.e., short circuit current
photovoltage (Voc), fill factor (FF), and the power conversion efficiency (Z),
measured under AM 1.5 solar light (1000 W m-2) are summarized in Table 2.
The devices were fabricated using screen printed TiO2 films as the
working electrode and platinized FTO glass was employed as the counter
electrode. The volatile electrolyte used consisted of 0.5 M 1-butyl-3-methylimidazolium iodide (BMII), 0.1 M lithium iodide, 0.05 M iodine and 0.5
M tert-butyl-pyridine in acetonitrile. In order to demonstrate the
reproducibility and the reliability of our devices, the fabrication of the solar
cells and the

Fig. 3
Left: IPCE curves of liquid electrolyte based DSSCs using optimized
parameters, right: I–V curves in the dark (dotted lines) and under irradiation (solid lines) of
liquid electrolyte based DSSCs using optimized fabrication parameters (aperture area 16
mm2).

.

measurements of their photovoltaic parameters were carried out
independently at ICIQ (aperture area of 0.16 cm2) and at Solaronix
(aperture area of 0.36 cm2). The losses of light reflection and
absorption by the conducting glass were not corrected. Many parameters
are known to influence the performances of the cells such as the TiO2
thickness.
In this work the best efficiencies were achieved using relatively
high thicknesses varying from 12 to 16 mm with an additional scattering
layer of 4 mm.
For the dyeing of the TiO2 electrodes, several tests using different
solvents were carried out. Indeed dyes can exhibit different interactions
with solvents which could cause changes of the physical and chemical
interactions between the dyes and the surface of the semiconductor.18 The
dyeing solutions also contain various amounts of chenodeoxycholic acid
(CDCA) because this additive is known to diminish the undesirable
formation of dye aggregates on the TiO2 surface,19 to decrease recombina
tion processes between TiO2 electrons and electrolyte20 and to
cause positive shifts of the surface potential of the TiO2 leading to an
enhancement of Voc.21 We found that the best results are obtained using
alcoholic solutions with a dye concentration of
0.2 to 0.5 mM and 2 to 5 mM of CDCA. Clearly, the use of
chlorinated solvents such as chloroform decreased the perfor- mances of
the cells (see ESI,† Table S1). The performance of solar cells
containing RK1 and various amounts of CDCA
Table 2

Photovoltaic parameters of liquid electrolyte based DSSCs using optimized parameters

Adsorbing solution
N719
RK1
RK2
6RK1
6ORK1
RKSe
RKF
a
d

(see ESI,† Table S2) suggests that the co-adsorbent plays a minor role
in the dye aggregation in our case, because Jsc values of the solar cells are
not deteriorated even without using CDCA. However we observed that Voc
increases with increasing concen- tration of CDCA. This suggests that the
presence of CDCA leads to a positive shift of the TiO2 vacuum level
and thus its CB
edge.21a Similar observations, i.e. a minor effect on Jsc and a strong
improvement in the photovoltage in the presence of CDCA, have also been
observed with ruthenium dyes by Neale and co-workers, who emphasized the
predominance of shift of TiO2 CB edge leading to overall higher
efficiency. The same mechanism seems to apply to RK dyes.
For optimized devices, RKSe and RK2 show the highest Z of
8.53% and 8.71% respectively, but more interestingly, three dyes
revealed Z over 9.50% and among them 6ORK1 overcomes the 10%
efficiency target. These outstanding performances are quite comparable to
those of RK1 (our reference dye) and N719 (a ruthenium complex that has
been widely used in the scientific
literature as a standard system).5d The combination of a rather
narrow absorption in the visible range with excellent photo- voltaic
performances, comparable to among the best for organic dyes,22 makes
6RK1, 6ORK1, and RKF very attractive for future developments and
applications in solar modules.
It should be underlined that quite high current densities are delivered
with these dyes and that the solar cells containing
6ORK1, RKSe and RKF show Jsc values ranging from 17.80 mA cm-2
up to 18.80 mA cm-2. For comparison purpose, DSSCs with similar
electrode thicknesses were fabricated, i.e. 13 + 4 mm, masked
0.16 cm2. RK2 and RKSe show the lowest Jsc in these conditions, whereas
6ORK1 and RK1 show the highest values. These differences cannot
be explained by the differences in the optical properties of the dyes because
6ORK1 and RKSe show very similar absorption spectra, and besides, they
show almost identical molar extinction coefficients.
The IPCE measurements of the cells confirm that the differences
in Jsc originate in the efficiency of conversion of the photons into
electrons. On the other hand, the dyes show quite similar Voc (0.75 and
0.76 V) except in the case of RKSe which shows a much lower Voc of
0.69 V. It is known that the

0.5
0.5
0.5
0.2
0.5
0.2
0.5
0.2
0.2
0.2
0.5
0.2
0.5
0.2

mM EtOH, 5 mM CDCA
mM EtOH, 5 mM CDCA
mM EtOH, 5 mM CDCA
mM MeOH, 2 mM CDCA
mM EtOH, 5 mM CDCA
mM MeOH, 2 mM CDCA
mM EtOH, 2 mM CDCA
mM ACN–tBuOH, 2 mM CDCA
mM ACN–tBuOH, 2 mM CDCA
mM ACN–tBuOH, 2 mM CDCA
mM EtOH, 5 mM CDCA
mM MeOH, 2 mM CDCA
mM EtOH, 5 mM CDCA
mM ACN–tBuOH, 2 mM CDCA

TiO2 [mm]
a

13 + 4
12.5 + 3.5b
13 + 4a
13 + 3.5b
13 + 4a
15 + 4b
13 + 4a
12 + 4b
13 + 4a
12 + 4b
13 + 4a
16 + 4b
13 + 4a
15 + 4b

Mask [mm2]
c

16
36d
16c
36d
16c
36d
16c
36d
16c
36d
16c
36d
16c
36d

Jsc [mA cm-2]

Voc [V]

FF [%]

Z [%]

17.91
20.21
18.26
19.50
14.92
15.96
16.76
15.96
17.81
17.87
13.43
18.37
16.55
18.82

0.80
0.72
0.76
0.71
0.75
0.71
0.76
0.74
0.76
0.75
0.69
0.67
0.75
0.71

71
70
74
72
78
73
76
70
75
70
77
68
77
72

10.19
10.25
10.20
10.00
8.71
8.22
9.67
8.25
10.11
9.41
7.14
8.53
9.51
9.69

TiO2 Dyesol (18NR-T 1 18NR-AO). b TiO2 Solaronix (HT/SP 1 R/SP). c Fabrication of the devices and measurements performed at ICIQ.
Fabrication of the devices and measurements performed at Solaronix.

LUMO energy level on this unit. This prevents significant
localization of the LUMO on the anchoring function and
consequently this may hinder good injection of electrons into the oxide
explaining the low charge density on these devices. All the other dyes show a
directional electron distribution, with the delocalization of the HOMO on
the triphenylamine group and the thiophene and a LUMO predominantly
delocalized on the BTD-phenylvinylcyanoacetic units. As a consequence
they pro- vide a better charge extraction for the devices.

.

Fig. 4 TPV curves (left) and CE curves (right) of liquid electrolyte based DSSCs
using optimized parameters, corresponding to the devices with a TiO2 thickness of 13
+ 4 mm.

Voc of a DSSC is determined by the difference between the quasiFermi-level of the oxide semiconductor and the redox potential of the
electrolyte. The quasi-Fermi-level of the oxide is correlated with its
conduction band edge (Ec) and electron density which is itself
dependent on the rate of recombination between semiconductor electrons
and the electrolyte. As a first approximation to understand the device
differences using the different sensitizers, charge extraction (CE) and
transient photo- voltage (TPV) measurements were carried out (see Fig. 4).
Charge extraction data show similar charge density (defined as the
number of charges stored at the solar cell at each different voltage) for
all the devices except for the RKSe devices. The TPV experiments reveal
that the electron lifetimes are quite similar for all the dyes except for
RKSe whose electron lifetime is one order of magnitude shorter compared
to the other set of sensi- tizers. It is interesting to perceive that the highest
Voc of 0.76 V is achieved with RK1, 6RK1 and 6ORK1, i.e. the dyes
with a prolonged lifetime for electrons stored in titania. Based on TPV
measurements we can conclude that the charge recombination in RKSe
based DSSC is much faster, which a priori explains the much lower Voc
(o0.70 V) and the lower performance observed under identical illumination
conditions.
This peculiar behaviour can be explained by its electronic structure.
Indeed, we found by DFT calculations (vide supra and in ESI,† Fig. S3)
that the replacement of the S atom of the
benzothiadiazole by a Se induces the spatial localisation of the

Table 3

RK2
6RK1
6ORK1
RKSe
RKF
a

For real use of DSSCs and future applications in BIPV, in addition to high
PCEs, long device lifetimes are required. Highly volatile electrolytes, such
ones, must be replaced because they are
as acetonitrile-based
inappropriate for stability measurements. During the last decade, ionic
liquids (IL) have been investigated as alternative solvents in DSSCs due to
their numerous appeal- ing properties for DSSC applications. Indeed
they can show wide electrochemical windows, a good thermal
stability, an extremely low volatility, and a high ionic conductivity
combined with the ability to dissolve organic and inorganic compounds.
Therefore we have fabricated and characterized solar cells with a solventfree ionic liquid electrolyte with the goal to obtain stable and efficient
devices.
Two types of devices were fabricated, using regular mesoporous TiO2 or a macrochannel TiO2 paste. After careful tuning of the
electrodes and optimization of the dyeing conditions, Z values of over
6.1% were achieved with all the dyes except for RKSe, the better, RKF,
reaching circa 7.9% (Table 3).
For all the dyes combined with ionic liquid electrolytes a
significant decrease of Voc is observed. The Voc of a DSC is determined
by the quasi-Fermi-level of the metal oxide semi- conductor, which is
correlated with its conduction band edge and the electron density.23a
The electron density is dependent on the rate of recombina- tion
between semiconductor electrons and oxidized electrolyte species. Electron
lifetimes measured using transient photo- voltage measurements (Fig. 5)
are found to be shorter (10 times
lower) than for the devices fabricated with a liquid electrolyte.

Photovoltaic parameters of ionic liquid electrolyte based DSSCs using optimized parameters

Adsorbing solution
RK1

2.4 Photovoltaic devices with an ionic liquid ‘‘robust’’
electrolyte

0.5
0.2
0.5
0.2
0.2
0.2
0.2
0.2
0.5
0.2
0.5
0.2

mM EtOH, 5 mM CDCA
mM MeOH, 2 mM CDCA
mM EtOH, 5 mM CDCA
mM MeOH, 2 mM CDCA
mM EtOH, 2 mM CDCA
mM ACN–tBuOH, 2 mM CDCA
mM ACN–tBuOH, 2 mM CDCA
mM ETOH–CHCl3, 2 mM CDCA
mM EtOH, 5 mM CDCA
mM ACN–tBuOH, 2 mM CDCA
mM EtOH, 5 mM CDCA
mM MeOH, 2 mM CDCA

TiO2 [mm]
13 + 4
4+4+
13 + 4
4+4+
13 + 4
4+4+
13 + 4
4+4+
13 + 4
4+4+
13 + 4
4+4+

Mask [mm2]
b

3.5a
3.5a
3.5a
3.5a
3.5a
3.5a

16
36c
16b
36c
16b
36c
16b
36c
16b
36c
16b
36c

Jsc [mA cm-2]

Voc [V]

FF [%]

Z [%]

15.77
15.40
12.04
13.49
11.86
14.49
14.18
15.05
11.86
12.94
15.24
17.50

0.70
0.67
0.68
0.63
0.70
0.65
0.69
0.67
0.64
0.63
0.69
0.67

70
69
74
72
61
69
70
64
71
71
69
67

7.35
7.36
6.10
6.08
5.13
6.48
6.84
6.47
5.43
5.75
7.27
7.87

TiO2 macrochannel Solaronix (HT/SP 1 MC/SP 1 R/SP). b Fabrication and measurements performed at ICIQ. c Fabrication and measurements performed at Solaronix.

Fig. 5 TPV curves (left) and I–V curves (right) of ionic electrolyte based DSSCs
using optimized parameters (electrode thickness 13 + 4 mm, aperture area 0.16
cm2).

This behaviour
has systematically
been reported for these
electrolytes.23b Therefore, the combined effect of the lower lying TiO2
conduction band and shorter electron lifetime explains the general trend
of lower Voc for these devices compared to DSSCs with the acetonitrile
electrolyte.
With 8 mm thick TiO2 films coated with a 3.5 mm thick
reflecting layer, RKF containing devices showed an impressive Jsc of 17.50
mA cm-2, a Voc of 0.67 V, and a FF of 67% yielding a Z of 7.87% at 1 Sun.
This performance is better than that of RK1 (our reference dye) and is
considerably better than most of the performances using ionic liquid
based electrolytes employing
(I-/I3 -) as a redox shuttle that are reported for organic sensitizers.24 This performance approaches the record efficiencies of
the best ruthenium-based dyes under the AM 1.5G full sunlight.25

.

2.5

Stability tests and solar module application

For any emerging photovoltaic technology, it is very critical to fulfil a
long-term stability criterion. Besides it is important to assess its potential
for large scale applications.26 However, for probing the stability of new
systems it is not practical to make a large solar panel and perform
outdoor evaluation over long
periods. Hence, we submitted our laboratory cells covered with a UV
absorbing polymer film to a 3000 h accelerated testing at
65 1C, in a solar simulator with a light intensity of 1000 W m-2.
In Fig. 6, the evolution of the PCE measured at different time intervals
under the irradiance of AM 1.5G is shown. After light

Fig. 6 Variation of the PCE of the solar cells measured under the irradiance
of AM 1.5G sunlight during successive full-sun visible-light
soaking (1 Sun 1000 W m-2) at 65 1C.

soaking, the photocurrent and power conversion efficiency of
some of the solar cells increases
This behaviour is related to the improved penetration of the ionic
liquid electrolyte through the mesoporous film and the
‘‘activation’’ of the entire electrode in depth.27 With the exception of 6ORK1- and RK2-based devices that showed a loss of
30% and 25%, respectively, of their initial effciency after 3000 h, all the other
solar cells reveal rather stable performances. RK1 and 6RK1 appear as the
most robust dyes since they maintain
95% of their initial performances in the same time interval. It should be
noticed that RK1 devices can stably perform for more than 9000 hours under
these harsh conditions with a loss of only
30% of the initial efficiency. From the literature, this stability performance
is the best ever achieved for an organic sensitizer (see ESI,† Fig. S4).
Owing to its high effciency and its outstanding stability, we identified
RK1 as a promising organic dye for the fabrication of a prototype of
large area (4500 cm2) photovoltaic modules (Fig. 7).
On FTO coated glass (35 cm x 50 cm), 13 single cells were
interconnected in series using a W-type design with an overall active area
of 1400 cm2 representing 80% of the total area (see Fig. 7). The
solution processing steps (i.e. printing of the electrodes and dyeing
procedure) were identical to those when the small area cells are
fabricated.
For the fabrication of the
module an ionic liquid electrolyte was used and introduced by
capillarity. In order to achieve a good semi-transparency in the visible and
to facilitate the penetration of the electrolyte, the thickness of the titania
electrode was kept at 2 mm without any scattering layer. The completed
solar cells show an aesthetic red-orange colour with a transmittance
over 50% between
620 nm and 1000 nm (see ESI,† Fig. S5).
The first prototype of this large-area module showed a current,
Isc, of 692 mA, a Voc of 8 V, and a FF of 26.5% leading to a power
output of 10.5 W m-2 when tested under 1 Sun
irradiation. When tested under 0.3 Sun the power output of the panel still
reached 7.3 W m-2. The low FF of this preliminary prototype can be
explained by the higher series resistance in this device configuration,
whereas the deteriorated current
density compared to our best solar cells is attributed to the
extremely thin titania layer that was employed. Further developments are
in progress to boost the performances of the
modules.

Fig. 7

Picture of the large area solar panels fabricated at Solaronix.

3. Conclusions
In summary, we have designed, synthesised and characterized a series of
structurally simple dissymmetric organic dyes for DSSCs. Four of the
new dyes have demonstrated power conver- sion efficiencies over 9.5% and
two of them have exceeded 10% effciency when combined with an iodinebased electrolyte. The replacement of the volatile electrolyte by an ionic
liquid yielded solar cells with power conversion efficiencies up to 7.9%
show- ing very promising stability of the performances under accelerated ageing conditions. Finally, the most robust organic dye, RK1, has
been successfully incorporated in a large area solar
panel (1750 cm2) exhibiting high transparency in the visible
range with a power output of 10.5 W m-2. This work highlights the
potential of metal-free organic dyes for future applications in buildingintegrated photovoltaics.

Acknowledgements
The authors acknowledge the European Union Research Executive Agency for
funding through the Adios-Ru research project. This project is funded by
the European Union Research Executive Agency (contract 315131) on the
Research for SMEs programme. E.P. acknowledges ICIQ and ICREA for
financial support. E.P. thanks the MICINN for projects CTQ2013-47183-R
and the Severo Ochoa Excellence Accreditation 2014.2018 (SEV-2013-0319).

.

Notes and references
1 (a) M. Gr¨ tzel, Nature, 2001, 414, 338; (b) B. O’Regan and M.
a
Gr¨ tzel, Nature, 1991, 353, 737; (c) T. W. Hamann and J. W.
a
Ondersma, Energy Environ. Sci., 2011, 4, 370; (d) N.
Memarian, I. Concina, A. Braga, S. M. Rozati, A. Vomiero
and G. Sberveglieri, Angew. Chem., Int. Ed.,
2011, 50, 12321; (e) D. Hwang, J.-S. Jin, H. Lee, H.-J. Kim, H.
Chung, D.-Y. Kim, S.-Y. Jang and D. Kim, Sci. Rep., 2014,
4, 7353; ( f ) S.-H. Choi, D. Hwang, D.-Y. Kim, Y. Kervella, P.
Maldivi, S.-Y. Jang, R. Demadrille and I.-D. Kim, Adv. Funct.
Mater., 2013, 23, 3146.
2 S. Ahmad, E. Guillen, L. Kavan, M. Gr¨ tzel and
a
M. K. Nazeeruddin, Energy Environ. Sci., 2013, 6, 3439.
3 W. M. Campbell, K. W. Jolley, P. Wagner, K. Wagner, P. J.
Walsh, K. C. Gordon, L. Schmidt-Mende, M. K. Nazeeruddin,
Q. Wang, M. Gr¨ tzel and D. L. Officer, J. Phys. Chem. C,
a
2007, 111, 11760.
4 (a) A. Yella, H.-W. Lee, H. N. Tsao, C. Yi, A. K. Chandiran, M.
K. Nazeeruddin, E. W.-G. Diau, C.-Y. Yeh, S. M.
a
Zakeeruddin and M. Gr¨ tzel, Science, 2011, 334, 629; (b) S.
Mathew, A. Yella, P. Gao, R. Humphry-Baker, B. F. E. Curchod, N.
Ashari-Astani, I. Tavernelli, U. Rothlisberger, M. K. Nazeeruddin and
M. Gr¨ tzel, Nat. Chem., 2014, 6, 242; (c) A. Yella, C.-L. Mai, S.
a
M. Zakeeruddin, S.-N. Chang, C.-H. Hsieh, C.-Y. Yeh and M.
Gr¨ tzel, Angew. Chem., Int. Ed.,
a
2014, 53, 2973.
5 (a) A. Dessi, M. Calamante, A. Mordini, M. Peruzzini, A.
Sinicropi, R. Basosi, F. Fabrizi de Biani, M. Taddei,

D. the photocurrent Carlo, G. Reginato and of
soaking,Colonna, A. Di and power conversion e ciency L. Zani,
Chem.Commun., 2014, 50, 13952; (b) G. Zhang, H. Bala, Y.
Cheng, D. Shi, X. Lv, Q. Yu and P. Wang, Chem. Commun.,
2009,
2198; (c) W. Zeng, Y. Cao, Y. Bai, Y. Wang, Y. Shi, M. Zhang, F.
Wang, C. Pan and P. Wang, Chem. Mater., 2010, 22, 1915; (d) S.
Ito, H. Miura, S. Uchida, M. Takata, K. Sumioka, P. Liska, P.
Comte, P. Pechy and M. Gr¨ tzel, Chem. Commun.,
a
2008, 5194; (e) D. Joly, L. Pelleja, S. Narbey, F. Oswald, J.
Chiron, J. N. Clifford, E. Palomares and R. Demadrille, Sci. Rep.,
2014, 4, 4033; ( f ) Z. Yao, H. Wu, Y. Ren, Y. Guo and P. Wang,
Energy Environ. Sci., 2015, 8, 1438.
6 Z. Yu, N. Vlachopoulos, M. Gorlov and L. Kloo, Dalton
Trans., 2011, 40, 10289.
7 (a) S. K. Pathak, A. Abate, T. Leijtens, D. J. Hollman, J.
Teuscher, L. Pazos, P. Docampo, U. Steiner and H. J. Snaith,
Adv. Energy Mater., 2014, 1301667; (b) M. K. Kashif, M. Nippe, N.
W. Duffy, C. M. Forsyth, C. J. Chang, J. R. Long, L. Spiccia and U.
Bach, Angew. Chem., Int. Ed., 2013,
52, 5527.
8 Z. Ning and H. Tian, Chem. Commun., 2009, 5483.
9 (a) Z. Ning, Y. Fu and H. Tian, Energy Environ. Sci., 2010,
3, 1170; (b) M. Xu, R. Li, N. Pootrakulchote, D. Shi, J. Guo, Z. Yi,
a
S. M. Zakeeruddin, M. Gr¨ tzel and P. Wang, J. Phys. Chem. C,
2008, 112, 19770.
10 (a) Y. Ooyama and Y. Harima, Eur. J. Org. Chem., 2009, 2903; (b) Y.
Ooyama and Y. Harima, ChemPhysChem, 2012, 13, 4032.
11 Q. Feng, W. Zhang, G. Zhou and Z.-S. Wang, Chem. – Asian J.,
2013, 8, 168.
12 M. Velusamy, K. R. J. Thomas, J. T. Lin, Y.-C. Hsu and
K.-C. Ho, Org. Lett., 2005, 7, 1899.
13 (a) K. Lim, C. Kim, J. Song, T. Yu, W. Lim, K. Song, P. Wang, N.
Zu and J. Ko, J. Phys. Chem. C, 2011, 115, 22640; (b) H.
Choi, J. K. Lee, K. H. Song, K. Song, S. O. Kang and J. Ko,
Tetrahedron, 2007, 63, 3115.
14 (a) M. Gr¨ tzel, Nature, 2001, 414, 338–344; (b) I. Chung,
a
B. Lee, J. He, R. P. H. Chang and M. G. Kanatzidis, Nature,
2012, 485, 486–489.
15 F. Ambrosio, N. Martsinovich and A. Troisi, J. Phys. Chem. C,
2012, 116, 2622–2629.
16 (a) A. Hagfeldt and M. Gr¨ tzel, Chem. Rev., 1995, 95, 49; (b) P.
a
Wang, S. M. Zakeeruddin, J.-E. Moser and M. Gr¨ tzel, J. Phys.
a
Chem. B, 2003, 107, 13280.
17 (a) A. D. Becke, J. Chem. Phys., 1993, 98, 5648; (b) SCM, ADF2010,
Vrije Universiteit, Amsterdam, The Netherlands, http://www.scm.com.
G. te Velde, F. M. Bickelhaupt, E. J. Baerends, C. F. Guerra, S. J. A.
Van Gisbergen, J. G. Snijders and T. Ziegler, J. Comput. Chem.,
2001, 22, 931–967.
18 (a) N. Robertson, Angew. Chem., Int. Ed., 2006, 45, 2338;
(b) C. Sakong, S. H. Kim, S. B. Yuk, J. W. Namgoong, S. W.
Park, M. J. Ko, D. H. Kim, K. S. Hong and J. P. Kim, Chem. –
Asian J., 2012, 7(60), 1817; (c) H. Ozawa, M. Awa, J. Ono and H.
Arakawa, Chem. – Asian J., 2012, 7, 156; (d) W. Ying, F. Guo,
J. Li, Q. Zhang, W. Wu, H. Tian and J. Hua, ACS Appl. Mater.
Interfaces, 2012, 4, 4215; (e) S. Cai, X. Hu, Z. Zhang, J. Su, X. Li,
A. Islam, L. Han and H. Tian, J. Mater. Chem. A, 2013, 1, 4763.

20
21

.

22
23

Int. Ed., 2009, 48, 2474.
T. Geiger, F. Nuesch, M. Gr¨ tzel and M. K. Nazeeruddin,
¨
a
Nanotechnology, 2008, 19, 424005.
25 (a) S. M. Zakeeruddin and M. Gr¨ tzel, Adv. Funct. Mater.,
a
2009, 19, 2187; (b) M. Gorlov and L. Kloo, Dalton Trans.,
A. C. Khazraji, S. Hotchandani, S. Das and P. V. Kamat,
2008, 2655; (c) D. Shi, N. Pootrakulchote, R. Z. Li, J. Guo, Y.
J. Phys. Chem. B, 1999, 103, 4693.
Wang, S. M. Zakeeruddin, M. Gratzel and P. Wang, J. Phys.
(a) S. Sakaguchi, S. S. Pandey, K. Okada, Y. Yamaguchi and S.
Hayase, Appl. Phys. Express, 2008, 1, 105001; (b) N. R. Neale,
Chem. C, 2008, 112, 17046.
a
N. Kopidakis, J. van de Lagemaat, M. Gr¨ tzel and A. J. Frank, J. 26 (a) A. Fakharuddin, R. Jose, T. M. Brown, F. FabregatPhys. Chem. B, 2005, 109, 23183.
Santiago and J. Bisquert, Energy Environ. Sci., 2014, 7, 3952; (b) A.
M. Liang and J. Chen, Chem. Soc. Rev., 2013, 42, 3453.
Hinsch, H. Brandt, W. Veurman, S. Hemming, M. Nittel, U.
(a) J. Idigoras, R. Tena-Zaerab and J. A. Anta, Phys. Chem.
Wurfel, P. Putyra, C. Lang-Koetz, M. Stabe, S. Beucker and K.
¨
Chem. Phys., 2014, 16, 21513; (b) F. Fabregat-Santiago, J.
Fichter, Sol. Energy Mater. Sol. Cells, 2009,
Bisquert, E. Palomares, L. Otero, D. B. Kuang, S. M.
93, 820.
Zakeeruddin and M. Gr¨ tzel, J. Phys. Chem. C, 2007, 111,
a
27 Z. Zhang, S. Ito, J.-E. Moser, S. M. Zakeeruddin and
6550–6560.
a
M. Gr¨ tzel, ChemPhysChem, 2009, 10, 1834.