Kinetic competition in liquid electrolyte and solid-state cyanine dye sensitized solar cells

Article abstract of DOI:10.1039/b703750c

The photovoltaic performance of liquid electrolyte and solid-state dye sensitized solar cells, employing a squarilium methoxy cyanide dye, are evaluated in terms of interfacial electron transfer kinetics. Dye adsorption to the metal oxide film resulted in a mixed population of aggregated and monomeric sensitizer dyes. Emission quenching data, coupled with transient absorption studies, indicate that efficient electron injection was only achieved by the monomeric dyes, with the aggregated dye population having an injection yield an order of magnitude lower. In liquid electrolyte devices, transient absorption studies indicate that photocurrent generation is further limited by slow kinetics of the regeneration of monomeric dye cations by the iodide/iodine redox couple. The regeneration dynamics are observed to be too slow (? 100 s) to compete effectively with the recombination of injected electrons with dye cations. In contrast, for solid-state devices employing the organic hole conductor spiro-OMeTAD, the regeneration dynamics are fast enough (? 1 s) to compete effectively with this recombination reaction, resulting in enhanced photocurrent generation. The Royal Society of Chemistry.

Full text of DOI:10.1039/b703750c

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/materials | Journal of Materials Chemistry
Kinetic competition in liquid electrolyte and solid-state cyanine dye
sensitized solar cells{{
Sergio Tatay,a Saif A. Haque,b Brian O’Regan,b James R. Durrant,b W. J. H. Verhees,c J. M. Kroon,c
A. Vidal-Ferran,de Pablo Gavin˜aa and Emilio Palomares*e
Received 13th March 2007, Accepted 4th May 2007
First published as an Advance Article on the web 18th May 2007
DOI: 10.1039/b703750c
The photovoltaic performance of liquid electrolyte and solid-state dye sensitized solar cells,
employing a squarilium methoxy cyanide dye, are evaluated in terms of interfacial electron
transfer kinetics. Dye adsorption to the metal oxide film resulted in a mixed population of
aggregated and monomeric sensitizer dyes. Emission quenching data, coupled with transient
absorption studies, indicate that efficient electron injection was only achieved by the monomeric
dyes, with the aggregated dye population having an injection yield an order of magnitude lower.
In liquid electrolyte devices, transient absorption studies indicate that photocurrent generation is
further limited by slow kinetics of the regeneration of monomeric dye cations by the iodide/iodine
redox couple. The regeneration dynamics are observed to be too slow (& 100 ms) to compete
effectively with the recombination of injected electrons with dye cations. In contrast, for solid-
state devices employing the organic hole conductor spiro-OMeTAD, the regeneration dynamics
are fast enough (% 1 ms) to compete effectively with this recombination reaction, resulting in
enhanced photocurrent generation.
molecular extinction coefficients in the visible are still major
Introduction
challenges for researchers.
Ruthenium complexes have attracted much interest in the area
Recently there has been an increasing interest in the
development of organic dyes,4 which appear to be also
of molecular solar cells due to their exceptional photophysical
properties.1 Dye sensitized solar cells (DSSCs) based on
promising due to their outstanding photophysical properties.
ruthenium polypiridine complexes such as cis-dithiocyanato-
Organic dyes present higher molar extinction coefficients when
bis(4,49-dicarboxy-2,29-bipyridine)ruthenium(II) (called N3)
compared with ruthenium complexes. This property allows
and
trithiocyanato(4,49,40-tricarboxy-2,29:69,20-terpyridine)
efficient light harvesting to be achieved with thinner meso-
porous metal oxide films. This advantage is particularly
important for DSSCs employing solid-state hole transport
materials, where efficient charge collection has only been
demonstrated for film thickness of the order of 2 mm.4 For this
reason organic dyes are an interesting approach for solid-state
dye sensitized solar cells. Several groups have focused on the
development of pure organic dyes4 such as coumarines,
merocyanines, perylenes, hemicyanines, indolines and polyene
dyes. Among them, indoline and thiophene dyes with an
outstanding efficiency of ca. 8%, are so far the most efficient
organic dyes for DSSCs.4a,4c However, many organic dyes
have a tendency of forming molecular aggregates on the
surface of the semiconductor nanoparticles, which it has been
suggested can result in a decrease in device efficiency. While
the formation of molecular aggregates has been intensively
studied and the low photocurrent has been usually assigned to
intermolcular quenching,5 rather little attention has been
placed on interfacial electron transfer kinetics of such organic
DSSC devices. For example, cyanine dyes have been exten-
sively explored as near infrared dyes for DSSCs5c,6 alongside
other optoelectronic applications such as optical data storage7
and xerography.8 In fact, during our investigation, similar
cyanine molecules were used as sensitizers for DSSCs.9
ruthenium(II) (black dye) have given the best light-to-energy
conversion efficiencies.2 Over the past decade, most studies
have focused on changes in the electrochemical and photo-
physical properties of the complexes affected by the substitu-
tion or replacement of the pyridine rings by electron donor or
aliphatic ligands, in order to control the electrochemical and
photophysical properties of the complex and, thereby, improve
the device performance.3 However, despite the broad metal-to-
ligand charge transfer (MLCT) absorption band of these
ruthenium complexes, which covers a wide spectral range from
the UV to the visible region of the solar spectra, the lack of
absorption in the near infrared region and the need for higher
aICMol-UV, Poligono Industrial La Coma s/n, Paterna, Spain
bCentre for Electronic Materials and Devices, Chemistry Department,
Imperial College, South Kensington, London, UK SW7 2AZ
cEnergy Research Centre of the Netherlands (ECN) Unit Solar Energy,
Westerduinwed 3, 1755 LE, Petten, The Netherlands
dCatalan Institution for Research and Advanced Studies (ICREA)
eInstitute of Chemical Research of Catalonia (ICIQ), Avgda, Pa¨ısos
Catalans 16, 43007 Tarragona, Spain. E-mail: epalomares@iciq.es;
Fax: +349770920220; Tel: +34977920241
{ The HTML version of this article has been enhanced with colour
images.
{ Electronic supplementary information (ESI) available: Molecular
structures and transient absorption kinetics. See DOI: 10.1039/
b703750c
In this paper, we report the synthesis and study of the
interfacial electron transfer dynamics for a cyanine dye
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 3037–3044 | 3037

View Article Online
1H NMR (CDCl3, 300 MHz): d 7.42 (d, J = 8.5 Hz, 1H),
derivative (MSqb) (Scheme 1) adsorbed onto a nanocrystalline
TiO2 film. The organic dye shows an absorption band edge
shifted to the red region (y700 nm), obtained by expanding
the p-conjugation system, and in addition the insertion of
methoxy groups on the molecular structure to achieve the
appropriate redox potentials to allow efficient charge
regeneration from the electrolyte. Our aim is not only to
study the effect of the dye structure upon the electron injection
yield but also on the dye regeneration dynamics. Our data
indicates that the efficiency of this regeneration reaction may
be a critical factor in limiting the efficiency in organic dye
sensitized solar cells.
6.83 (m, 1H), 6.80 (d, J = 2.6 Hz, 1H), 3.82 (s, 3H), 2.24 (s,
3H), 1.28 (s, 6HH).
HRMS (FAB): Calc. for C12H15NO 190.1232; found
190.1235.
N-(p-Carboxybenzyl)-5-methoxy-2,3,3-trimethyl-3H-indol
quaternary salt (2). A mixture of 1 (1.89 g, 10 mmol),
p-chloromethyl benzoic acid (1.76 g, 10.1 mmol) and
o-dichlorobenzene was refluxed under argon at 110u for 12 h.
The solvent was evaporated under reduced pressure and the
resulting solid was re-dissolved in the minimum amount of
acetonitrile. A black dense oil was obtained. Acetone was then
added until a precipitate appeared. The solid was then filtered
off and washed with acetone to yield the pure product as a
brown solid (1.4 g, 38%).
Experimental section
Materials and reagents
1H NMR (acetonitrile-d3, 300 MHz): d 7.92 (d, J = 8.1 Hz,
1H), 6.83 (s, 1H), 6.81 (d, J = 8.1 Hz, 1H), 3.82 (s, 3H), 2.24 (s,
3H), 1.28 (s, 6H).
All reagents were purchased from Sigma-Aldrich. The solvents
used were of HPLC degree and those for electrochemical mea-
surements were previously degassed with argon. Transparent
conducting glass (TCO, F-doped SnO2, 14V cm22) was
purchased from Hartford Glass Company Inc. and washed
with water and soap, ethanol and acetone successively under
sonication for at least 10 min prior to utilization.
HRMS (FAB): Calc. for C20H22NO3 324.1600; found
324.1601.
MSqb. Compound 2 (0.72 g, 2 mmol) and squaric acid
(0.115 g, 1 mmol) were heated in an mixture of butanol
(9 mL)–toluene (9 mL) with pyridine as a catalyst for 6 h. The
solvent was evaporated and acetonitrile was added until a
precipitate appeared. The precipitate was then filtered to
obtain the pure product without further purification (green
solid, 0.42 g, 58%). MP = 278 uC.
Synthesis of the squarilium methoxy cyanide dye
The squarilium methoxy cyanide dye was obtained following
previous literature reports.9a All reagents were commercially
available and used without further purification. A brief
description of the procedure is described as follows:
1H NMR (dmso-d6, 300 MHz): d 7.92 (d, J = 7.9 Hz, 4H),
7.29 (d, J = 7.9 Hz, 4H), 7.22 (d, J = 2.5 Hz, 2H), 7.19 (d, J =
8.8 Hz, 2H), 6.86 (dd, J = 8.8 and 2.5 Hz, 2H), 5.66 (s, 2H),
5.43 (br. s, 4H), 3.77 (s, 6H) and 1.71 (s, 12H). HRMS (FAB):
Calc. for C44H40N2O8 724.278; found 724.280.
5-Methoxy-2,3,3-trimethyl-3H-indol (1). 4-Methoxyphenyl-
hydrazine hydrochloride (4.37 g, 25 mmol) and 3-methyl-2-
butanone (6.5 mL, 60 mmol) were refluxed in CH3COOH
(80 mL) for 5 h. The red solution was neutralized with KOH at
0 uC. The mixture was then extracted with Et2O. The organic
phase was dried over Na2SO4 and the solvent was removed
under reduced pressure. The resulting oil was purified by
column chromatography (SiO2, Hexane–Et2O) to afford 1
(3,78 g, 80%) as a yellow solid.
Instruments
UV-Vis absorption spectra were recorded on a Shimadzu
spectrophotometer model 2101 PC. In the case of trans-
parent 4 mm methoxy cyanide dye sensitized TiO2 films
Scheme 1
3038 | J. Mater. Chem., 2007, 17, 3037–3044
This journal is ß The Royal Society of Chemistry 2007

View Article Online
(MSqb/TiO2), the blank was obtained using identical dye-
free mesoporous TiO2 films to correct the light scatter
phenomena.
Results and discussion
UV-Vis absorption spectra
Fluorescence emission and excitation spectra were measured
using an Aminco Bowman spectrofluorometer model with a
support for solid films. Fluorescence emission decays were
recorded using an Edinburgh Instrument picosecond pulsed
laser diode with a nominal wavelength of 635 nm (instrument
response 350 ps). Transient absorption measurements were
recorded as reported before.10 Cyclic voltammetry was
conducted in dry solvents containing 1 mM of the cyanide
dye and 0.1 M of dry tetrabutyl ammonium perchlorate as the
supporting electrolyte. Reference electrode Ag/AgCl; working
The absorption spectra of MSqb in solution and for the
MSqb/TiO2 film are shown in Fig. 1. It can be seen that
the MSqb has a well-defined spectrum with a maximum at
l = 655 nm and a shoulder at l = 600 nm when dry pure
acetonitrile (ACN) is used as a solvent. As reported before,5a
we assign this spectrum to the non-aggregated form of MSqb.
The use of other solvents such as dimethylformamide (DMF),
dimethyl sulfoxide (DMSO) or dichloromethane (DCM)
induces the formation of J- and H-aggregated dimers for
MSqb dye (data not shown).
electrode and counter electrode, Pt; scan rate 100 mV s21
;
The absorption spectrum of the MSqb sensitized TiO2 film
shows clear evidence for significant aggregation of the MSqb
dye. It is apparent that the absorption of the MSqb/TiO2 film
has an absorption peak at l = 600 nm unlike the MSqb dye in
ACN. This spectrum is indicative of the presence of H-type
MSqb aggregates. The residual peak at 650 nm is assigned to a
sub-population of monomeric dyes adsorbed onto the TiO2, as
evidenced by emission data shown below.
temperature 25 uC.
Squarilium methoxy cyanide dye sensitized TiO2 films
(MSqb/TiO2)
Nanocrystalline TiO2 and ZrO2 particles and the mesoporous
TiO2 and ZrO2 films were prepared as reported before.10
The 4 mm thick transparent TiO2 films were used for
characterizations such as spectroscopic and electrochemical
measurements. Double-layer mesoporous TiO2 films (8 mm)
were employed for photovoltaic measurements with an
additional layer of 4 mm thick scatter TiO2 particles (diameter
ca. 100 nm). The TiO2 films were calcined at 450 uC for 30 min.
The film thicknesses were measured using a surface profiler
model XP1 profilometer, Ambios Technology. The films
were cut into pieces of 1 cm2 with an active area of
0.160 cm2 and immersed in a dye solution for at least 5 h.
The dye solution (0.5 mM) was prepared in a mixed solvent of
acetonitrile–4-tert-butyl alcohol with a volume ratio of 1 : 1.
The electrolyte for MSqb liquid dye sensitized solar cells
(Msqb/DSSCs) comprised tetra-butyl ammonium iodide
0.6 M (2.216 g), LiI 0.1 M (0.134 g), I2 0.1 M (0.253 g), and
4-tert-butyl pyridine 0.5 M (0.676 g) in 10 mL of acetonitrile,
HPLC degree. This electrolyte was introduced through holes
drilled in the counter electrode, which were sealed immediately
with a microscope cover slide. For solid-state Msqb devices
(MSqb/s-DSSCs) we employed spiro-OMeTAD hole con-
ductor [chemical name: 2,29,7,79-tetrakis(N,N-di-p-methyloxy-
phenylamine)9,99spirobifluoreno]. Several salts and additives
such as tert-butyl pyridine, lithium bis-trifluorosulfinimid
[Li(CF3SO2)2N] and the chemical dopant tris(p-bromophenyl)
ammoniumyl hexachloroantimonate, [N(p-C6H4Br)3][SbCl6].
A thin layer of evaporated gold was used as the counter
electrode.
Luminescence and electrochemistry measurements
Emission data for the MSqb dye were obtained employing
non-injecting mesporous ZrO2 films. Absorption spectra for
MSqb/TiO2 and MSqb/ZrO2 were indistinguishable. The emis-
max
em
sion spectra of MSqb/ZrO2 films revealed a l
= 706 nm.
The corresponding excitation spectrum for this film exhibits a
well defined band at 650 nm, corresponding to the absorption
maximum of monomeric MSqb dyes, but
a negligible
contribution from the aggregated dye absorption band—
indicative of rapid non-radiative quenching of photogenerated
excited states in the dye aggregates.
The emission properties of the MSqb are summarized in
Table 1.
The transient luminescence studies confirmed efficient
electron injection for the MSqb sensitized TiO2 films, as
shown in Fig. 2.
Light-to-energy conversion measurements
The I–V measurements for liquid based MSqb/DSSCs were
realized using an OrielE 250 W Xenon lamp with the
appropriate OrielE filters to simulate the desired 1.5 AM solar
spectrum and a KeithleyE 2400 source meter. I–V measure-
ments for solid-state DSSCs have been carried out using a
SteuernagelE constant 575 solar simulator and a KeithleyE
2400 source meter. Calibration of this system has been
carried out by comparative measurements of DSSCs with a
SpectrolabE X-10 solar .
Fig. 1 The UV-Vis absorption spectra of the monomer as the
aggregated form of the MSqb (H-aggregate).
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 3037–3044 | 3039

View Article Online
Table 1 Physcical parameters for MSqb
Absorption/a.u.
Emission/a.u.
Emission lifetime/ns
Em (dye+/dye)a
MSqb/ACN
MSqb/ZrO2
MSqb/TiO2
a
600, 655
615,660
615, 660
670
700
706
t1 = 0.64 (0.25) t2 = 2.1 (0.75)
t1 = 0.23 (0.80) t2 = 0.95 (0.20)
t , IRFb
0.451
—
—
b
V vs. SCE. The emission lifetime was within the instrument response of the apparatus measured using a TiO2 film as a blank. The
excitation wavelength was 635 nm. IRF = Instrument response file.
This comparison shows the transient luminescence decay of
MSqb sensitized TiO2 and ZrO2 films, normalized to the same
number of absorbed photons. ZrO2 film was employed as a
control film, as the relatively high conduction band edge of
this material prevents electron injection, consistent with the
long excited lifetime observed for dyes adsorbed to this film. It
can be noted that when compared to the control ZrO2 films,
the MSqb luminescence is strongly quenched on the TiO2 film,
consistent with efficient electron injection. In addition, despite
the lower emission quantum yield of the MSqb/ZrO2 films,
when compared to acetonitrile solution, a double exponential
decay with a component of t1 = 0.71 ns and t2 = 1.2 ns was
measured, which is in agreement with MSqb in ACN solution
that also shows multi-exponential behaviour with t1 = 0.6 ns
and t2 = 2.1 ns (Table 1). According to the steady-state
emission spectra of the MSqb/ZrO2 film illustrated in Fig. 3,
we have assigned the emission decay to the non-aggregated
forms on the MSqb/ZrO2. We are aware that lower excitation
wavelengths (lex , 635 nm) will be desirable to further prove
our hypothesis by exciting the aggregates. Unfortunately, fast-
pulsed laser diodes between 550 to 620 nm were not available
in our instrument. Nonetheless, as illustrated elsewhere on the
manuscript, we demonstrated that electron injection mainly
occurs from monomers by using incident photon-to-current
conversion efficiency (IPCE) measurements.
E0-0 = 1.8 eV which is the energy difference between the
ground state oxidation potential [E0 (D+/D)] and the excited-
state oxidation potential [E0 (D*/D+)] of the MSqb dye.
From this numbers we can estimated the excited-state
oxidation potential to be y21.35 vs. SCE, which is more
negative than the energy level of the TiO2 conduction band
edge4b estimated to be located at y20.74 V vs. SCE suggesting
that electron injection should be possible.
An analysis of the electronic structure of MSqb in vacuo was
undertaken employing DFT calculations using the SpartanE
06 program package. A schematic representation of the
molecular orbitals is illustrated in Fig. 4. The calculated
energy for the highest occupied molecular orbital (HOMO) of
Msqb was 24.47 eV while for the lowest unoccupied molecular
orbital (LUMO) the energy value found was 22.26 eV. Hence,
as a result of these calculations, a HOMO–LUMO gap of
2.2 eV can be determined, which is in reasonable agreement
with the experimental value obtained for E0-0
.
Photovoltaic performances of MSqb/TiO2 solar cells
The photovoltaic performances of the different DSSCs based
on MSqb under AM 1.5 irradiation (100 mW cm22) are shown
in Table 2.
Despite the high emission quenching observed in the
luminescence experiments described above, the efficiencies of
the liquid MSqb based DSSCs are low. We investigate such
Cyclic voltammograms were performed to determine the
ground state oxidation potential for the MSqb dye. The
oxidation potential of MSqb was found to be 0.451 V vs. SCE.
From the MSqb/ZrO2 emission spectra we can consider that
Fig. 2 Time resolved single photon counting decays of emission
from MSqb sensitized on ZrO2 [black, (a)] and TiO2[grey, (b)]. The
acquisition time for both samples was 600 seconds, lex = 635 nm and
Fig. 3 The excitation in black (lem = 730 nm) and the singlet excited
emission spectra in grey (lex = 650 nm) under standard conditions for a
4 mm thick MSqb/ZrO2 film.
lem = 710 nm. Laser power at lex = 635 nm, 20 nJ cm22
.
3040 | J. Mater. Chem., 2007, 17, 3037–3044
This journal is ß The Royal Society of Chemistry 2007

View Article Online
Fig. 5 Photocurrent action spectrum obtained with a Msqb/TiO2
DSSC. The redox electrolyte comprised tetra-butylammonium iodide
0.6 M (2.216 g), LiI 0.1 M (0.134 g), I2 0.1 M (0.253 g), and 4-tert-butyl
pyridine 0.5 M (0.676 g) in acetonitrile.
Fig. 4 Optimised molecular structures of Msqb by DFT calculations
and the graphical representation of the HOMO (a) and LUMO (b)
orbitals.
issues further measuring incident photon-to-current conver-
sion efficiency (IPCE) spectra.
As can be observed in Fig. 5, the IPCE shows mainly
injection of electrons from the non-aggregated dyes of the
MSqb/TiO2 film (lmax = 670 nm). On the other hand, it is
well established that to achieve high electron injection
yields, the electron injection rate must be fast enough to
avoid kinetic competition with the de-excitation of the excited
states (Fig. 6).
Katoh et al. have studied the electron transfer kinetics in
TiO2 and ZnO sensitized films with coumarine dyes and they
have observed ultra-fast (,150 fs) kinetics for the electron
injection reaction.11 We have shown a high electron injection
yield from the non-aggregated MSqb into the CB (conductive
band) of TiO2. Moreover, the redox potential of the oxidized
dye should accept electrons from the I2 ions regenerating the
ground state of the dye (Fig. 6) as the energetic driving force
for the regeneration of the dye ground state is adequate. In
order to seek further, we decided to perform laser transient
Fig. 6 Schematic energy diagram for a MSqb DSSC with iodine/
iodide electrolyte. (1) Light excitation, (2) emission back electron
transfer reaction, (3) electron injection process, (4) electron recombina-
tion with the oxidized dye. kRC represents the kinetic rate for electron
recombination and kRG the kinetic rate for dye regeneration.
absorption studies to investigate: (1) the electron recombina-
tion dynamics between the light induced MSqb cation and the
injected electrons at the TiO2 surface and (2) the regeneration
reaction between the electrolyte and the oxidized MSqb dye in
the device.
Table 2 Photovoltaic performance of MSqb DSSCs
Isc/mA cm22
Voc/V
FF
g (%)
MSqb/DSSC
N719/DSSCa
MSqb/s-DSSCb
N719/s-DSSC
a
0.75
14.8
1.0
0.460
0.702
0.59
0.710
0.68
0.44
0.24
7.0
0.26
1.43
Laser transient absorption measurements
2.5
0.845
0.670
DSSCs employing N719 [chemical name cis-bis (thiocyanato)bis-
(2,29-bipiridine-4,49-dicarboxylato)ruthenium ditetrabutylammonium
salt] dye were employed for comparison purposes. The samples
were annealed at 80 uC for 1 min. The results are the average of
four devices. Isc = Current at short circuit, Voc = voltage at open
circuit, FF = fill factor, g = efficiency.
The primary reason to use the extended carboxylated p system
was to increase the UV-Vis band edge in the infrared region of
the spectra, but also to achieve a controlled separation of the
organic dye from the TiO2 surface. We have previously
demonstrated, using ruthenium based dyes, the relationship
b
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 3037–3044 | 3041

View Article Online
between electron recombination dynamics and the HOMO
distance of the anchored molecule.3a,12
cation electron recombination kinetics, and a slow phase,
assigned to the formation of I22, with an estimated half time of
t50% # 0.3 s for N719 (ESI Fig. 2S{). However, the transient
absorption data observed following excitation of the Msqb/
TiO2 film in the presence the redox electrolyte strongly indicate
The presence of the benzene ring as a spacer on the MSqb
˚
dye results in a y7 A separation of the charged species to the
nanoparticle surface. The DFT theoretical calculations
described before illustrate that either the HOMO or the
LUMO orbitals of MSqb are delocalised on the aromatic
backbone of the molecule, in contrast with the ruthenium dye
where the HOMO is situated mainly on the axial ligands and
the LUMO orbital on the carboxyl bipyridine which anchor
the dye to the TiO2 surface. The decay upon induced light
absorption was monitored following laser excitation of the dye
sensitized TiO2 film, in the absence of any applied bias or
redox active electrolyte. Fig. 7 illustrates the charge recombi-
2
a negligible yield of I2 (Fig. 8).
The absence of I22 yield in the kinetics can be understood in
terms of kinetic competition between the recombination
reaction [Fig. 6, reaction (4), kRC] and the electron regenera-
tion of the oxidized dye by the redox couple [Fig. 6, reaction
(5), kRG]. Under normal conditions of device operation for
N719 dye sensitized solar cells, kRG & kRC, avoiding the
wasteful electron recombination reaction (4) and therefore, for
liquid electrolyte devices, the charge recombination between
injected electrons and oxidized dye molecules will not be an
important loss mechanism limiting the device efficiency.
However, in our case kRG # kRC and hence the back electron
transfer reaction (4) limits the device performance. We have
recently suggested the presence of an intermediate mechanism
for dye sensitized solar cells, resulting in complex formation
between the dye cation and the electrolyte, potentially
necessary for successful regeneration dynamics.14 Although it
has been only demonstrated in the case of ruthenium based
DSSCs, we would like to stress that, in the case of organic
dyes, it is an important issue to take into account for the design
of organic DSSCs since the cation is only formed on the non-
aggregated dyes as demonstrated in this work. Further to
prove that the results described above are not mainly due to
the lack of enough driving force between the oxidized dye and
the redox electrolyte, we used an organic molecule as a hole
conductor that has lower oxidation potential than our liquid
redox electrolyte and we performed identical experiments.
Schmidt-Mende et al. have reported a striking 4% solid-state
solar cell using an indoline dye and spiro-OMeTAD as a
molecular hole conductor which approaches the efficiencies
obtained for such a dye when a liquid iodine/iodide electrolyte
is employed (g # 8% under 100 mW cm22 1.5 AM irradiation)
and improves the efficiency reported for ruthenium based
nation decay for MSqb on TiO2. The estimated half time, t50%
,
for MSqb/TiO2 is y0.1 ms.
This result is in good agreement with the recombination
versus HOMO distance dependence observed for the
Ru(dcbpy)2Cl2 complex (dcbpy = 4,49-dicarboxy-2,29-bipyr-
˚
idine) with a distance of y7.4 A, reported previously by
Clifford et al.12a (ESI, Fig 1S{). However, this value indicates
that the electron recombination reaction is y10 times
faster when compared with the ruthenium based dye N719
(chemical name: cis-bis(thiocyanato)bis(2,29-bipiridine-4,49-
dicarboxylato)ruthenium(II) ditetrabutylammonium salt). This
faster recombination cannot by itself explain the unexpected
low efficiency of liquid MSqb DSSCs, the 0.1 ms half time still
being slower than the regeneration half time observed for
N719 in typical acetonitrile based electrolytes (y1 ms).
This observation therefore raises the issue of whether the
regeneration dynamics in MSqb DSSCs may be significantly
slower than N719 based devices, potentially resulting in the
observed cells’ lower performance. We have previously demon-
strated13 that the presence of the I2/ I3 redox electrolyte
2
induces, upon light excitation of the dye sensitized TiO2 film, a
biphasic transient decay with a fast phase, assigned to the dye
Fig. 7 Transient absorption data monitoring the electron recombina-
tion dynamics for the MSqb/TiO2 film. The inset shows the transient
absorption spectra for the MSqb/TiO2 film monitored at 0.1 ms after
excitation at 650 nm.
Fig. 8 Transient absorption data monitoring at 750 nm for a MSqb/
TiO2 film (grey) and for a MSqb/TiO2 open-cell (black). The open-cell
electrolyte consisted of 0.1 M LiI, 0.05 M I2 in propylene carbonate.
3042 | J. Mater. Chem., 2007, 17, 3037–3044
This journal is ß The Royal Society of Chemistry 2007

View Article Online
electrolyte is used, resulting in a significant reduction in the
yield of the regeneration reaction and therefore low device
performance due to the fast electron recombination kinetics
for MSqb/TiO2 films. The laser transient spectroscopy data
shows negligible changes due to the presence of the electrolyte
on the Msqb recombination dynamics, indicating the lack of
2
I2 formation in contrast to what it has been observed for
other efficient dyes. However, an extensive interest in organic
dye sensitized solar cells is currently under way and the use
of solid hole conductor electrolytes where the regeneration
dynamics (polymer electrolytes, small organic molecules,
etc…) may result in significant improvement of device perfor-
mance. Thus, our studies of charge recombination dynamics
between the oxidized organic dye and the organic hole
conductor spiro-OMeTAD, show a remarkable yield of hole
transfer and hence, we expect that, for dyes which do not form
a stable intermediate with the liquid electrolyte, it will be
possible to achieve higher efficiencies for the solid-state version
of the dye sensitized solar cells similar to the devices based on a
liquid electrolyte.
Fig. 9 Transient absorption kinetics obtained for MSqb/TiO2/spiro-
OMETAD film upon excitation at lex = 650 nm (laser pulse intensity
25–50 mJ cm22) and monitoring at lprobe = 570 nm (black decay). As
control sample an MSqb/TiO2 was employed (grey decay).
Acknowledgements
solid-state DSSCs.15 Interested by these results and the fact
that the solid-state MSqb DSSCs showed relative higher
efficiencies for solid-state devices, we explored the kinetics for
dye regeneration using the organic hole conductor molecule
instead of the liquid redox electrolyte. Klug and coworkers
have shown previously that the hole transfer from the oxidized
dye to the organic hole conductor spiro-OMeTAD occurs on
the 100 ps timescale regenerating the dye’s original ground
state.16 Fig. 9 shows the decay for the spiro-OMeTAD cation
upon excitation of the MSqb/TiO2/spiro-OMeTAD film.
As can be seen, a long-lived positive signal is observed when
we probe at l = 570 nm upon excitation at l = 650 nm. This is
in clear contrast with the spectrum recorded for MSqb/TiO2
(Fig. 7 inset) where a broad negative feature can be observed
which we have assigned to the bleaching of the MSqb ground
state. Hence, the positive signal can only be due to the
formation of the spiro-OMeTAD cation. Unfortunately, the
hole injection kinetics are not resolved on the time scale
employed in our experiments. It is obvious that kRG & kRC
when an organic molecule hole conductor is employed as a
‘‘solid electrolyte’’, solving the ‘‘kinetic competition’’ between
reactions (4) and (5), despite the lower oxidation potential of
the organic hole conductor.
EP thanks the Spanish Ministerio de Educacio´n y Ciencia
(MEC) for the Ramo´n y Cajal Fellowship and the PROFIT
Project NanoSolar, Amparo Forneli for excellent technical
assistance on the preparation of devices and Aplicaciones
Te´cnicas de la Energ´ıa (ATERSA S.L.) for economical
support.
References
1 (a) N. Robertson, Angew. Chem., Int. Ed., 2006, 45, 2338; (b)
M. K. Nazeeruddin, C. Klein, P. Liska and M. Graetzel, Coord.
Chem. Rev., 2005, 249, 1460.
2 (a) M. Gratzel, Prog. Photovoltaics, 2006, 15, 429; (b) Y. Chiba,
A. Islam, R. Komiya, N. Koide and L. Han, Appl. Phys. Lett.,
2006, 88, 223505/1.
3 (a) S. A. Haque, S. Handa, K. Peter, E. Palomares, M. Thelakkat
and J. R. Durrant, Angew. Chem., Int. Ed., 2005, 44, 5740; (b)
C. Klein, M. K. Nazeeruddin, P. Liska, D. Di Censo, N. Hirata,
E. Palomares, J. R. Durrant and M. Graetzel, Inorg. Chem., 2005,
44, 178; (c) M. Yanagida, T. Yamaguchi, M. Kurashige, K. Hara,
R. Katoh, H. Sugiara and H. Arakawa, Inorg. Chem., 2003, 42,
7921; (d) N. Hirata, J.-J. Lagref, E. Palomares, J. R. Durrant,
K. K. Nazeerddin, M. Gratzel and D. Di Censo, Chem.–Eur. J.,
2004, 10, 595.
4 (a) S. Kim, J.-K. Lee, S.-O. Kang, J. Ko, J.-H. Yum, S. Fantacci,
F. de Angelis, D. Di Censo, M. K. Nazeeruddin and M. Gratzel,
J. Am. Chem. Soc., 2006, 128, 16701; (b) K. Hara, T. Sato,
R. Katoh, A. Furube, Y. Ohga, A. Shinpo, S. Suga, K. Sayama,
H. Sugihara and H. Arakawa, J. Phys. Chem. B, 2003, 107, 597; (c)
T. Horiuchi, H. Miura, K. Sumikoa and S. Uchida, J. Am. Chem.
Soc., 2004, 126, 12218; (d) D. P. Hagberg, T. Edvisson,
T. Marinado, G. Boschloo, A. Hagfeldt and L. Sun, Chem.
Commun., 2006, 2245.
5 (a) W. West and S. Pearce, J. Phys. Chem., 1965, 69, 1894; (b)
A. Ehret, L. Stuhl and M. T. Spitler, Electrochim. Acta, 2000, 45,
4553; (c) K. Takeshi, P. K. Sudeep and P. V. Kamat, J. Phys.
Chem. B, 2006, 110, 16169.
6 K. Sayama, S. Tsukagoshi, K. Hara, Y. Ohga, A. Shinpou, Y. Abe,
S. Suga and H. Arakawa, J. Phys. Chem. B, 2002, 106, 1363.
7 J. Li and F. Gan, Appl. Opt., 2000, 39, 3525.
8 A. R. Tameev, A. V. Vannikov and H. F. M. Schoo, Thin Solid
Films, 2004, 451–452, 109.
9 (a) X. Chen, J. Guo, X. Peng, M. Guo, Y. Xu, L. Shi, C. Liang,
L. Wang, Y. Gao, S. Sun and S. Cai, J. Photochem. Photobiol., A,
Conclusions
In summary, we directly addressed the role of the molecular
aggregates on the injection kinetics for a sensitized meso-
porous TiO2 film. We find that upon light excitation the
electron injection only occurs from the monomers and the
yield remains high (y95% of electrons are injected into the CB
of TiO2) despite the presence of H-aggregates on the film.
Moreover, the steady-state luminescence and the IPCE experi-
ments also showed that only the non-aggregated Msqb dyes
are directly involved in the electron injection dynamics. On
the other hand, kinetic competition between kRC and kRG is
observed to be important when an iodine/iodide standard
This journal is ß The Royal Society of Chemistry 2007
J. Mater. Chem., 2007, 17, 3037–3044 | 3043

View Article Online
2005, 171, 231; (b) S. Alex, U. Santhosh and S. Das, J. Photochem.
Photobiol., A, 2005, 172, 63.
S. A. Haque, T. Torres and J. R. Durrant, Chem. Commun.,
2004, 2112.
10 S. Hore, E. Palomares, H. Smit, N. J. Bakker, P. Comte, P. Liska,
K. R. Thampi, J. M. Kroon, A. Hinsch and J. R. Durrant,
J. Mater. Chem., 2005, 15, 412.
11 T. Yoahihara, R. Katoh, A. Furube, M. Murai, Y. Tamaki,
K. Hara, S. Murata, H. Arakawa and M. Tachiya, J. Phys. Chem.
B, 2004, 108, 2643.
12 (a) J. N. Clifford, E. Palomares, M. K. Nazeeruddin, M. Gratzel,
J. Nelson, X. Li, N. J. Long and J. R. Durrant, J. Am. Chem. Soc.,
2004, 126, 5225; (b) J. R. Durrant, S. A. Haque and E. Palomares,
Chem. Commun., 2006, 3279; (c) E. Palomares, M. V. Mart´ınez-D´ıaz,
13 A. N. M. Green, E. Palomares, S. A. Haque, J. M. Kroon and
J. R. Durrant, J. Phys. Chem. B, 2005, 109, 12525.
14 J. N. Clifford, E. Palomares, M. K. Nazeruddin, M. Gratzel,
W. R. Griffith and J. R. Durrant, J. Phys, Chem. C, 2007, 111,
6561–6567.
15 L. Schmidt-Mende, U. Bach, R. Humphry-Baker, T. Horiuchi,
H. Miura, S. Ito, S. Uchida and M. Gratzel, Adv. Mater., 2005, 17,
813.
16 U. Bach, Y. Tachibana, S. A. Haque, J.-E. Moser, J. R. Durrant,
M. Gratzel and D. R. Klug, J. Am. Chem. Soc., 1999, 121, 7445.
3044 | J. Mater. Chem., 2007, 17, 3037–3044
This journal is ß The Royal Society of Chemistry 2007