Supramolecular light harvesting antennas to enhance absorption cross-section in dye-sensitized solar cells

Article abstract of DOI:10.1039/c1cc16066d

Doping a zinc porphyrin based-sensitizer with antenna molecules, axially held by metallo-supramolecular interactions, enhances the light-harvesting efficiency and the overall photo-conversion efficiency of the solar cells by about 30%.

Full text of DOI:10.1039/c1cc16066d

View Online / Journal Homepage / Table of Contents for this issue
This article is part of the
Artificial photosynthesis
web themed issue
Guest editor: Andrew Benniston
All articles in this issue will be gathered together
online at
www.rsc.org/photosynthesis

Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2012, 48, 675–677
www.rsc.org/chemcomm
COMMUNICATION
Supramolecular light harvesting antennas to enhance absorption
cross-section in dye-sensitized solar cellswz
Julien Warnan, Yann Pellegrin, Errol Blart and Fabrice Odobel*
Received 29th September 2011, Accepted 18th November 2011
DOI: 10.1039/c1cc16066d
Doping a zinc porphyrin based-sensitizer with antenna molecules,
axially held by metallo-supramolecular interactions, enhances the
light-harvesting efficiency and the overall photo-conversion efficiency
of the solar cells by about 30%.
dye and the sensitizer (energy acceptor) is necessarily large.
It depends on the relay concentration in the electrolyte and on
the TiO2 pore size (over nm). Inspired by the natural light
harvesting antenna of photosynthetic organisms, we reported
earlier a fourth approach, which consists in covalently tethering
several dyes with complementary absorption spectra to a
sensitizer.6 The potential of this strategy with covalent multi-
chromophoric assemblies was fully demonstrated.7,8 The intent
of this work is to extend this strategy to supramolecular
systems using the well-known pentacoordination preference
of zinc in porphyrins and phthalocyanines.9 The utilization of
supramolecular interactions in solar cells is scarce.10 Metallo-
supramolecular bonding is already at work in the natural
light harvesting antenna, since magnesium chlorophylls are
held by axial ligation on histidine residues of the protein
scaffold.9 If the covalent multichromophoric assembly is
replaced by a supramolecule, the potential time-consuming
and expensive synthesis of covalent arrays could be avoided. We
are interested in organic dyes because they are of potentially
lower-cost and display higher oscillator strength transitions than
most transition metal complexes. Moreover, the latter often
exhibit triplet excited-states which are not compatible with
The seminal paper of O’Reagan and Gratzel in 1991 featuring
¨
a dye-sensitized solar cell (DSSC) with a 7.1% photo-conversion
efficiency has brought about an explosion of academic and
industrial interests for this promising low-cost photovoltaic
technology.1 Light capture is the primary step and a fundamental
function in any photovoltaic device. A panchromatic sensitizer
is essential to achieve efficient solar cells in order to match
the cell’s response to the sun’s spectrum. Various approaches
for optimizing light management in a DSSC have been
reported. The first and most straightforward one is to design
a panchromatic sensitizer, such as the black dye.2 However,
the black dye contains ruthenium which is a noxious and
non-abundant element and the broad absorbance of this
complex is paid at the expense of a modest absorption
coefficient. A second approach is the co-sensitization concept,
where several different dyes with complementary absorption
spectra are attached to the titanium dioxide (TiO2) surface.3
However, the limited number of sites available on the TiO2
surface reduces the amount of dyes which can be loaded on the
electrodes. Moreover, the performances greatly depend on
the concentration ratio dye1/dye2 in the dyeing bath and the
adsorption duration; therefore the optimization can be very
cumbersome already for two dyes only. More recently a third
concept proposes to use an energy relay dye in solution
within the electrolyte, acting as an energy donor towards the
chemisorbed dye, and allowing a greater light harvesting.4
singlet to triplet energy transfers by a Forster mechanism.5
¨
However, the absorption bands of organic dyes are usually
narrow and this justifies the development of multichromo-
phoric arrays to cover a broad spectrum. In addition to this, the
use of high extinction coefficient dyes allows a correspondingly
thinner layer of TiO2 to effect an equivalent light harvesting,
which is highly desirable for electrolyte-free DSSCs.
We selected borondipyrromethene (BODIPY) and diketo-
pyrrolopyrrole (DPP) derivatives as light harvesters because
they exhibit high molar absorptivities in the 450–550 nm range
precisely where the ZnP absorption is the weakest, but where
solar energy is abundant (Scheme 1). This weak point is
known to be the limiting factor of most porphyrin sensitizers
in DSSCs.11 Although any zinc porphyrin sensitizer could have
been used to demonstrate the validity of the approach, we
chose the dyad D because it was readily available from a previous
study.8 The antennas were easily available in a few synthetic
steps by Sonogashira cross-coupling reaction (see ESIz). The
UV-Visible spectra of the antennas A1 and A2 displayed
absorbance across the 450–550 nm window, filling the absorption
gap of ZnP in this region (Fig. 1; Fig. S1, ESIz). All the antennas
have a significant emission overlap with the absorption spectrum
However, the Forster radius (r ) is very short if the overlap
¨
0
between the emission spectrum of the donor relay and
the absorption spectrum of the acceptor sensitizer is weak or
if the photoluminescence quantum efficiency of the relay is
low.5 Therefore a modest energy transfer quantum yield could
be expected, because the mean distance ‘‘d’’ between the relay
CEISAM, Universite´ de Nantes, CNRS, 2 rue de la Houssinie`re,
44322 Nantes cedex 3, France. E-mail: Fabrice.Odobel@univ-nantes.fr;
Fax: +33 25 112 5402; Tel: +33 25 112 5429
w This article is part of the ChemComm ‘Artificial photosynthesis’ web
themed issue.
z Electronic supplementary information (ESI) available. See DOI:
10.1039/c1cc16066d
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 675–677 675

Table 1 Photovoltaic performances of D in DSSCs in response to
100 mW cmꢀ2 illumination after cells were dipped in an antenna bath
for 0.5 h
Dye
Voc/mV
Jsc/mA cmꢀ2
ff (%)
Z (%)
D
D+A1
D+A2
545
565
535
9.25
9.67
9.82
72
72
73
3.64
3.94
3.81
TiO2 film thickness = 12.7 mm of nanoparticles + 4 mm scattering layer.
Scheme 1 Supramolecular assemblies between D and the antenna Ai.
Fig. 2 Photoaction spectra of D (black solid line), with A1 (red line),
1 (red dashed line), A2 (blue line) or 2 (blue dashed line) in solution
within the electrolyte.
absorbance of the electrodes between 450–550 nm, which
dominantly contribute to increase Jsc and therefore the overall
power conversion efficiency (Table 1).
It is worth mentioning that if the antenna itself binds to the
TiO2 electrode, the photo-conversion efficiencies of the corres-
ponding devices were very poor (Table S2, ESIz) underscoring
that high photocurrents cannot be produced independently of
the sensitizer. The photo-action spectra clearly show that the
IPCE in the 450–550 nm region is strongly enhanced when the
dyad is connected to the antenna (Fig. 2; Fig. S3, ESIz). This
directly reflects the higher absorbance and therefore the higher
LHE of the electrode in this region.
Fig. 1 Overlay of the absorption spectra recorded on 4 mm thick
TiO2 electrodes coated with D (black) and its association with A1 (red)
and A2 (blue) after two hours of dipping in THF.
of dyad D resulting in an efficient Forster energy transfer (Fig. S2,
¨
ESIz). 12.7 mm thick mesoporous TiO2 films plus a 4 mm
scattering layer prepared according to the literature procedure
were first coated with D in the presence of deoxycholic acid
(DCA) to avoid aggregation, then dipped into a solution of the
antenna (0.2 mM) in THF.8 The strong affinity of imidazole
for ZnP drives the formation of the trichromophoric array
D-Ai. The UV-visible absorption spectra of the supramolecular
association were initially recorded on thin TiO2 layers (4 mm)
without the scattering layer (Fig. 1). The broad spectral
coverage from 400 until 700 nm achieved with the trichromo-
phoric supramolecular arrays demonstrates the benefits of the
antenna to enhance the LHE, particularly in the 500 nm
region. These supramolecular assemblies appear black on the
TiO2 electrode, while the initial film (coated with D only) is
green. The authenticity of the supramolecular self-assembly of
the antenna onto the ZnP was confirmed by the usual red-shift
of the Soret (Dl = 4 nm) and Q (Dl = 10 nm) bands
evidencing the binding of the antenna through axial ligation
on the zinc atom.9 The cells were removed from the antenna
bath, washed with ethanol, dried, sealed with a platinised
counter electrode and filled with a classical electrolyte
(0.6 M 1,2-dimethyl-3-butylimidazolium iodide, 0.1 M LiI
and 0.05 M I2 in acetonitrile). A reference cell sensitized with
D was also prepared under the same conditions. The devices
containing the antenna perform significantly better than that
sensitized only with D because A1 and A2 enhance the
The mechanism for the increased photocurrent is attributed
to Forster energy transfer from the antenna to the dyad. The
¨
antenna absorbs light and, through energy transfer, funnels
the incident energy to ZnP and flows to the reddest absorbing
species (SQ) on the surface of TiO2, from which electron
injection occurs. To make sure that all the zinc porphyrin
sites were saturated, the assembling time was increased but
weaker performances were measured, mainly due to lower Jsc
(Table S3, ESIz). Although an increase of the IPCE was
observed between 450 and 550 nm (specific absorption of A1
and A2), the concomitant decrease of the signal at 420 and 670 nm
(dyad specific absorption) prevails (Fig. S4 and S5, ESIz). This
was attributed to dye desorption during the supramolecular
assembly step as confirmed when a photoanode was abandoned
in plain THF. After one hour, the THF bath turned to a green
hue and the corresponding solar cell displayed smaller Jsc and
Voc owing to the excellent solubility of D in this solvent and
probably to the fact that D bears a single carboxylic acid
anchoring group. In order to overcome these problems, the
solvent used for the supramolecular association was changed to
acetonitrile (ACN) in which D is insoluble. Rewardingly, perfor-
mances were increased to 4.10% with A1 (Table S4, ESIz),
c
676 Chem. Commun., 2012, 48, 675–677
This journal is The Royal Society of Chemistry 2012

Table 2 Photovoltaic performances of D assembled with antennas
present within the electrolyte and recorded under full sun
stable since they only slightly decrease (by less than 10%) over
5 days (Fig. S6, ESIz). We also note that upon ageing, the
performances of the cell with dyad D only decrease as
well, signifying that the degradation of the sensitizer could
be also responsible for the diminished photovoltaic efficiency
rather than the disruption of the supramolecular assembly. We
report herein the easy fabrication of panchromatic sensitizers by
using appropriate combinations of organic dyes associated
by supramolecular interactions. The use of antennas bound
to the chemisorbed sensitizer via a supramolecular interaction
broadens the absorbance of the sensitizers and shows undis-
putable superiority to antenna-free systems and to soluble
antennas displaying no affinity for the sensitizer (e.g. 1 and 2).
Moreover, this type of association is rather stable provided
that a slight amount of the antenna is introduced in the
electrolyte. This approach can be extended to other classes
of pigments since the functionalization with an imidazole moiety
can be straightforward in many cases. Even more importantly,
this concept can be extended to any dye based on zinc, magnesium
or ruthenium porphyrins or phthalocyanines, greatly improving
the light harvesting efficiency of this class of sensitizers which
recently proved to be highly competitive in the field of DSSC.11
This approach could furthermore find interest in solid-state dye
sensitized solar cells and organic photovoltaic (OPV), where
desorption of the chemisorbed dye is no longer an issue.
Dye
[A]/mmol Lꢀ1 Voc/mV Jsc/mA cmꢀ2 ff (%) Z (%)
D
D+A1
/
1
10
545
555
575
535
565
585
9.25
9.97
11.01
11.64
8.61
72
73
73
72
74
73
3.64
4.05
4.64
4.43
3.59
3.89
D+A2 sat.
D+1
D+2
10
10
9.09
confirming the importance of the solvent used for the supra-
molecular assembly and dipping time. However, after 5 days, a
significant drop of the photovoltaic performance was again
observed. This trend was monitored as well for the devices
built from THF (Table S5, ESIz). Because A1 and A2 are
slightly soluble in acetonitrile (the electrolyte solvent), an
equilibrium is established (TiO2–D–Ai 2 TiO2–D + Ai) in
the sealed cell and yields to a slight yet observable de-coordination
of the antennas from the zinc porphyrin partially damaging
the supramolecular assembly. An ‘‘orthogonal’’ solvent, where
D nor any antennas would be soluble, could circumvent this
problem. Another approach consists in solubilizing A1 or A2
directly into the liquid electrolyte, which will displace the
aforementioned equilibrium to the left and insures that no
desorption of D occurs. Thus, novel electrolytes were prepared
consisting of the traditional electrolyte into which was dissolved
A1 or A2. Several concentrations were investigated for A1: 1, 10
and 20 mM (electrolytes E1, E10 and E20 respectively). To
evaluate the effect of the supramolecular association over the
energy relay approach recently reported,5 we also tested two
other electrolytes containing BODIPY and DPP dyes lacking
the imidazole moiety (1 and 2, respectively see ESIz) at 10 mM
concentration (Table 2). First of all, when the antennas 1 and 2
were dissolved in the electrolyte, only a very weak contribution
of the latter to the IPCE was observed (dashed-line in Fig. 2),
highlighting the importance of formation of the supramolecular
assemblies for efficient energy transfer.
We acknowledge financial support from CNRS and the ANR
HABISOL (program Asyscol, no. ANR-08-HABISOL-002).
Notes and references
1 B. O’Regan and M. Gratzel, Nature, 1991, 353, 737–740.
¨
2 M. K. Nazeeruddin, P. Pechy, T. Renouard, S. M. Zakeeruddin,
R. Humphry-Baker, P. Comte, P. Liska, L. Cevey, E. Costa,
V. Shklover, L. Spiccia, G. B. Deacon, C. A. Bignozzi and
¨
M. Gratzel, J. Am. Chem. Soc., 2001, 123, 1613–1624.
3 Y. Chen, Z. Zeng, C. Li, W. Wang, X. Wang and B. Zhang, New J.
Chem., 2005, 29, 773–776; J.-H. Yum, S.-R. Jang, P. Walter,
T. Geiger, F. Nueesch, S. Kim, J. Ko, M. Gratzel and
¨
M. K. Nazeeruddin, Chem. Commun., 2007, 4680–4682.
4 B. E. Hardin, E. T. Hoke, P. B. Armstrong, J.-H. Yum, P. Comte,
T. Torres, J. M. J. Frechet, M. K. Nazeeruddin, M. Gratzel and
¨
The best performances were obtained for electrolytes contain-
ing 10 mM of A1 (E10) and A2 saturated (o10 mM), thanks to
an increase of Jsc by 25% and 20% respectively. Despite the fact
that A2 has a weaker molar extinction coefficient than A1, the
photocurrent generated was higher possibly due to this broader
absorption band permitting to collect more light. An increase of
the photoconversion efficiency was observed with E1 too (1 mM),
although to a lesser extent probably because of insufficient
amounts of available antenna. The quantity of dye D adsorbed
on TiO2 was estimated to be around 13 nmol per electrode by
dye desorption (detailed procedure in ESIz). Assuming that an
available volume of the electrolyte inside the cell is ca. 2 mL, there
is only 2 nmol of antenna in the electrolyte E1 before supra-
molecular assembly, entailing incomplete ZnP coordination with
the antenna. Electrolyte E10 is well-suited to saturate all ZnP
sites and keep a reasonable amount of antenna in solution to
minimize its decoordination. As regards E20, despite an increased
Voc, a major drop of the Jsc (ca. 1.5 mA cmꢀ2) is experienced
possibly by shielding of the available light for D and/or by
quenching the excited state via antenna aggregation.
M. D. McGehee, Nat. Photonics, 2009, 3, 406–411; K. Shankar,
X. Feng and C. A. Grimes, ACS Nano, 2009, 3, 788–794.
5 V. Balzani and F. Scandola, Supramolecular Photochemistry,
Ellis Horwood, Chischester, UK, 1991.
6 F. Odobel and H. Zabri, Inorg. Chem., 2005, 44, 5600–5611.
7 R. Amadelli, R. Argazzi, C. A. Bignozzi and F. Scandola, J. Am.
Chem. Soc., 1990, 112, 7099–7103; L. Chang Yeon and
T. H. Joseph, Langmuir, 2010, 26, 3760–3765; C. Siegers,
J. Hohl-Ebinger, B. Zimmermann, U. Wurfel, R. Muelhaupt,
¨
A. Hinsch and R. Haag, ChemPhysChem, 2007, 8, 1548–1556.
8 J. Warnan, F. Buchet, Y. Pellegrin, E. Blart and F. Odobel, Org.
Lett., 2011, 13, 3944–3947.
9 Y. Kobuke, Struct. Bonding, 121, 49–104.
10 N. K. Subbaiyan, C. A. Wijesinghe and F. D’Souza, J. Am. Chem.
Soc., 2009, 131, 14646–14647; A. Nomoto and Y. Kobuke, Chem.
Commun., 2002, 1104–1105; F. Silvestri, I. Lopez-Duarte, W. Seitz,
L. Beverina, M. V. Martinez-Diaz, T. J. Marks, D. M. Guldi,
G. A. Pagani and T. Torres, Chem. Commun., 2009, 4500–4502;
N. K. Subbaiyan, J. P. Hill, K. Ariga, S. Fukuzumi and
F. D’Souza, Chem. Commun., 2011, 47, 6003–6005.
11 M. G. Walter, A. B. Rudine and C. C. Wamser, J. Porphyrins
Phthalocyanines, 2010, 14, 759–792; Y.-C. Chang, C.-L. Wang,
T.-Y. Pan, S.-H. Hong, C.-M. Lan, H.-H. Kuo, C.-F. Lo,
H.-Y. Hsu, C.-Y. Lin and E. W.-G. Diau, Chem. Commun., 2011,
47, 8910–8912; A. Yella, H.-W. Lee, H. N. Tsao, C. Yi,
A. K. Chandiran, M. K. Nazeeruddin, E. W.-G. Diau, C.-Y. Yeh,
Very interestingly, when the antenna is present in the
electrolyte the photoconversion efficiencies remain relatively
S. M. Zakeeruddin and M. Gratzel, Science, 2011, 334, 629–634.
¨
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 675–677 677