Flexible Dye-Sensitized Nanocrystalline
TiO2 Solar Cells
P.M. (Paul) Sommeling, Martin Späth, Jan Kroon, Ronald Kinderman, John van Roosmalen
ECN Solar Energy, P.O. Box 1, 1755 ZG Petten, The Netherlands,
tel. +31 224 564276, fax. +31 224 563214, email: firstname.lastname@example.org.
ABSTRACT: The Dye-sensitized nanocrystaline TiO2 solar cell (nc-DSC) developed by Grätzel  has the potential
to reach low costs in future outdoor power applications. In addition, due to its expected ease of production and possibilities to
adjust its appearance, the potential for application is very broad. When plastic foil is used as a substrate for the nc-DSC the
production and application possibilities are even bigger. Polymer foils are easier to handle in processing steps such as cutting of
larger entities into smaller individual modules. Moreover the processing of foil into complete flexible solar cells could be realized
by means of a continuous roll to roll proces. This implies that for the flexible nc-DSC a higher throughput in production could be
realized. Also special applications requiring a certain flexibility of the solar cell, such as smart cards, come into play. In this paper
the differences in cell performance and -stability between glass based nc-DSC's and polymer foil based nc-DSC's are studied. The
ITO coating on the polymer foil appears to be a critical factor with respect to the photoelectrode stability. TiO2 sintered at low
temperatures (150 °C) does not influence the cell stability in a negative way.
Keywords:TiO2 - 1: Plastic - 2: Stability - 3.
A new type of solar cell based on dye-sensitized
nanocrystalline titanium dioxide has been developed by M.
Grätzel and coworkers [1,2]. Remarkably high quantum
efficiencies have been reported for this type of solar cell (a
so called nc-DSC), with overall conversion efficiencies up
to 11 % . This fact , in combination with the expected
relatively easy and low cost manufacturing makes this new
technology an interesting alternative for existing solar cell
Realisation of stable efficiencies in the order of 10 % in
production, however , requires a lot of effort on the
research and development side. For this reason, the first
application of this type of solar cell will probably be one in
which only a low power output is required, since this is
easier to achieve. Less stringent efficiency requirements
leave room for increased flexibility in the manufacturing
process of these cells, i.e. the requirements that are put on
the materials used are less severe. This results in lower Figure 1: A 4 cell module of a flexible nc-DSC
manufacturing costs and more flexibility in materials
choice, opening the way for alternative production
processes. 2 WORKING PRINCIPLE
One alternative concept for the nc-DSC is based on the
introduction of polymer foils as a substrate instead of glass. nc-DSC’s are based on a wide bandgap semiconductor,
This opens the way to attractive roll to roll production usually TiO2, which is sensitized for visible light by a
processes and special applications in which flexibility of monolayer of adsorbed dye. The most frequently used dye
the solar cell is to some extent required. Moreover is, for reasons of best efficiencies, cis-(NCS)2bis(4,4’-
processing and handling of polymer foil devices in general dicarboxy-2,2’bipyridine)-ruthenium(II).
is easier than for the glass version especially when it comes The photoelectrode in such a device consists of a
to cutting of glass or drilling of holes in glass substrates. nanoporous TiO2 film (approx. 10µm thick) deposited on a
The aim of this study is to investigate the critical layer of transparant conducting oxide (TCO, usually
parameters regarding the stability of the nc-DSC's on SnO2:F) on glass (Fig.2). The counter electrode also
polymer foil in comparison to the glass version. A consists of TCO coated glass on which a small amount of
prototype of a flexible DSC was already demonstrated in platinum catalyst is deposited.
1998 and is shown in Fig.1. The aim of this paper is to
study the stability and performance of nc-DSC's on
polymer foils in comprison to nc-DSC's on glass.
certain period of time.
Figure 2: Working principle of a nc-DSC. After drying the stained photoelectrode is assembled with
the platinized counter electrode and a thermoplastic thin
counter electrode glass film in between (not on the active cell surface). The two
e- TCO electrodes are sealed together by heating up to 150°C and
I3- 3I- Pt applying pressure.
The cell is completed by filling with electrolyte through
small prefabricated holes in the counter electrode. The
I3- electrolyte is spread in the very small space between the
electrodes (approx. 10μm) by capillary forces. Finally, the
S* holes are sealed with a thermoplastic film and cover
light dye 3.2 nc-DSC's on polymer foil
TCO When a polymer foil is used as a substrate for nc-
photoelectrode glass DCS's instead of glass, the production process is different.
Polymer foils allow roll to roll production which is a means
to achieve high throughput. Moreover, alternative sealing
In a complete cell, photo- and counter electrode are technology can be used. ECN has developed a method for the
clamped together and the space between the electrodes and filling of a nc-DSC with electrolyte without the use of a
the voids between the TiO2 particles are filled with an filling hole, which is required in the full glass version .
electrolyte. This electrolyte consists of an organic solvent This strongly facilitates the manufacturing of nc-DSC's.
containing a redox couple, usually iodide/triiodide (I-/I3-). Some of the drawbacks of polymer foils are the fact that
A dye monolayer on a flat surface absorbs less than 1 % of polymers do not resist high temperatures and that these
the incoming light (one sun conditions). To obtain materials are permeable for water and oxygen. This has
reasonable efficiencies comparable to established solar cell immediate consequences for the solar cell processing
technologies, in the nc-DSC the surface area is enlarged by parameters and for the performance and stability. The main
a factor of 1000, by using nanoparticles of TiO2 with a consequences of using a polymer foil substrate instead of
diameter of approximately 10-20 nm. glass are the following:
The working principle of the nc-DSC is based on excitation
of the dye followed by fast electron injection into the 1. TCO
conduction band of the TiO2, leaving an oxidized dye SnO2:F cannot be applied to polymer foils because of high
molecule on the TiO2 surface. Injected electrons percolate processing temperatures; room temperature sputtered indium
through the TiO2 and are fed into the external circuit. At tin oxide (ITO) can be used instead.
the counter electrode, triiodide is reduced to iodide by 2. Sintering of TiO2.
metallic platinum under uptake of electrons from the In the glass based manufacturing process TiO2 is
external circuit: screenprinted using a paste which contains substantial
amounts of organic additives that have to be burned out at
I3- + 2e- --> 3I- 450°C. This procedure cannot be used for the manufacturing
of the polymer foil solar cell, because organic residues
Iodide is transported through the electrolyte towards the cannot be removed effectively at temperatures as low as
photoelectrode, where it reduces the oxidized dye. The dye 150°C. Aqueous TiO2 paste without organic surfactants
molecule is then ready for the next excitation/oxidation/ should be used. Sintering temperatures as low as 150°C are
reduction cycle. sufficient to produce mechanically stable TiO2 films.
3. MANUFACTURING OF nc-DSC's Platinum cannot be deposited using a thermal platinization
process with a platinum salt as precursor. Instead, Pt can be
3.1 Manufacturing of nc-DSC’s on glass deposited using a galvanic process or by room temperature
In this section the basic manufacturing technology for sputtering.
nc-DSC’s on glass substrates is described. The first process 4. Polymer foils are not gastight, oxygen and water vapor can
step consists of the preparation of the photo electrode by permeate through the polymer into the solar cell.
deposition of a layer of nanocrystalline titanium dioxide
onto the SnO2:F coated glass substrate. On an industrial 4. EXPERIMENTAL METHODS
scale this can be realized by means of screen printing of an
ink, containing a TiO2 colloidal suspension. This layer is 4.1 Cell preparation
dried and sintered in a furnace at 450-550 °C to eliminate Photo electrodes have been prepared both on glass
organic residues of the screen print medium and to substrates and on polymer foil, i.e. polyethyleneterephtalate
establish electrical contact between the TiO2 particles. (PET), by means of doctor blading . The titanium
After cooling down of the photoelectrode, it is stained by dioxide paste used in this process has been developed for
emerging the electrode in a solution of organic dye for a sintering at lower temperatures (150 °C) but has also been
applied to the electrodes on glass substrates which have
been sintered at 450 °C. Titanium dioxide pastes were
based on in house synthesized colloid. The latter material 16
was synthesized by hydrolysis of an organic titanium 14
precursor according to the procedure described in literature 12
The colloid was processed to produce a titanium dioxide
paste that can be deposited onto glass or polymer foil
Several sintering temperatures have been applied to the
photoelectrodes: 150 °C for the polymer foil electrodes,
150 °C and 450 °C for the glass electrodes. 0 0.1 0.2 0.3 0.4 0.5
After sintering for 30 minutes the photoelectrodes have V (V)
been immersed overnight in a solution of 3*10-4 M cis-
(NCS)2bis(4,4’-dicarboxy-2,2’bipyridine)-ruthenium(II) in Figure 3: IV characteristic of a nc-DSC on PET at 250 lux
Counter electrodes have been prepared by deposition of a At an illumination intensity of 250 lux, typical for indoor
thin layer of a 150 mM solution of H2PtCl6 in isopropanol conditions, the following cell parameters have been
and subsequent heating at 450°C for 30 minutes. determined: Voc: 0.48 V, Isc: 15μA/cm2, FF: 67 %. Taking
Temperature sensitive polymer foil counter electrodes have into account that the average power demand of a calculator is
been made by a galvanic platinisation process. 10 μW, a 5 cm2 sized polymer foil nc-DSC is able to produce
Cell assembly has been carried out according to the this power down to a light intensity of 100 lux. It is possible
procedure described in section 3.1. The electrolyte to apply polymer foils which resist high temperature
composition was as follows: 0.5 M LiI, 0.05 M I2 and 0.4 treatment. These materials however are temperature-stable
M tertiair-butylpyridine in methoxypropionitrile. because of the high content of polyaromatic systems which
are more or less intensely colored. This filters out an
4.2 Characterization and stability testing important part of the light and results in lower power output.
Cells have been characterized using fluorescent tube
lamps with a light intensity of 50-5000 lux to mimic indoor Cell stability:
artificial light conditions. IV characteristics have been In practice the power output of a nc-DSC on (ITO) coated
determined with a Keithley 2400 source meter. Ageing has polyethelenterephtalate (PET) foil decreases with time.
been carried out using a sulfur lamp. The irradiation Though cell stability is not yet established and subject of
intensity was approximately 2 sun. Cells have been under study for all nc-DSC's, the decrease in performance of the nc-
continuous illumination and IV characteristics of DSC on ITO-PET foil is substantially faster than on SnO2:F-
individual cells have been monitored with time. Light glass. The following factors could play a role:
intensities have been measured using a lux meter (Licor
188B integrating quantum radiometer). 1. Instead of SnO2:F, sputtered ITO is used as a TCO. This
specific material has not been applied earlier in nc-DSC's.
The suitability of the material for application as a substrate
5. RESULTS AND DISCUSSION for nc-DSC's is uncertain.
2. Effects of low temperature sintering on cell stability could
The items mentioned in section 3.2 will influence be related to insufficient contact between TiO2 particles,
the efficiency and the stablity of the solar cells. causing the nanoporous film to fall apart.
Cell efficiency: 3. The ITO/polymer foil itself could influence the solar cell.
TiO2 sintered at 150 °C results in less performing nc-DSC's Plasticizers or stabilizers could leach out of the polymer foil
than TiO2 sintered at 450 °C, though comparable efficiencies into the electrolyte or undesirable components could leach
have been found for low temperature and high temperature out of the ITO. These kinds of compounds could interfere
sintering of TiO2 in another study, with sintering times up to with the delicate cell chemistry. Second, oxygen and water
48 h . The open circuit voltage and fill factor are vapor can permeate through the polymer influencing the cell
unaffected but the short circuit current is decreased to about stability in a negative way. the latter problem could be
half its value for TiO2 sintered at 450 °C. However, the avoided by application of barrier coatings to the polymer foil.
polymer foil nc-DSC's exhibit a performance which is still 4. Galvanically deposited platinum on the counter electrode
enough to power indoor applications such as watches, could degrade during cell operation.
calculators or smart cards. A typical IV characteristic of a nc-
DSC on ITO-PET is given in figure 3. Efficiency and stability of nc-DSC's have been studied under
continuous illumination by monitoring IV characteristics
with time. Both glass and PET substrates have been applied,
also combined in one nc-DSC, and TiO2 films have been
sintered at several temperatures.
Table I gives an overview of the relevant manufacturing
conditions for the nc-DSC's that have been tested.
Cell no. Photoelectr. Counter electr. sinter T (°C)
1 glass polymer 450
2 glass polymer 150
3 glass glass 150
4 glass glass 450
5 polymer polymer 150
Figure 6: Fill factor vs time
From these results the following conclusions can be drawn:
Figures 4,5 and 6 represent the Jsc, Voc and FF with time for
the various nc-DSC's, all measured under illumination of 1. low temperature sintering of TiO2 results in lower initial
2000 lux. The cells have been aged under illumination of peformance but this performance is stable. The degradation
approximately 2 sun. of the polymer foil cells cannot be attributed to the low sinter
temperature of the TiO2.
2. The fact that the ITO-PET is in contact with the electrolyte
of the cell and serves as a sealing on one side does not affect
the cell stability. From this it can be derived that possible
leaching of disturbing compounds from the ITO-PET does
not play a role. More importantly, also permeation of water
and oxygen does not seem to play a major role in degradation
of the polymer nc-DSC's.
3. The application of galvanically deposited platinum on ITO
as a catalyst does not affect the cell stability in a negative
At this stage it can be concluded that the instability of the nc-
DSC's on polymer foil is related to the photoelectrode
exclusively. This could be due to the composition of the TiO2
Figure 4: Short cicuit current vs time paste, which is strongly acidic. The sputtered ITO on PET is
proven to be sensitive towards acidic media. However experi-
ments using alkaline TiO2 paste resulted in the same
Polymer foil cells have been dissembled after degradation. It
appeared that the TiO2 film peeled off of the foil. Moreover
the ITO coating of the photoelectrode had been subject to
degradation; in several cases parts of the ITO coating had
This leads to the conclusion that the properties of the transpa-
rant conducting oxide in combination with a polymer foil are
of crucial importance when it comes to manufacturing nc-
DSC's on polymer foil.
6 CONCLUDING REMARKS
Figure 5: Open circuit voltage vs time nc-DSC cells on polymer foil substrates can meet the
demands required by calculators, watches etc. under typical
From figures 4-6 it can be derived that nc-DSC's with indoor conditions. From the point of view of production
polymer photo-electrodes degraded within the time period of technology and special applications, flexible solar cells are
the measurements (3.5 months). As a reference nc-DSC's on an attractive alternative to the existing rigid systems. The
SnO2:F coated glass have been used. The reference cells have currently used substrate PET foil is a good candidate from
been stable within the time period of testing. Interestingly the the point of view of light transmission, though the
full glass cells containing 150 °C sintered TiO2 have been sputtered ITO coating causes stability problems in the solar
remarkably stable as well as hybrid cells, composed of a glass cells. The sintering of TiO2 at temperatures as low as 150
photo-electrode and an ITO-PET counter electrode. °C does not seem to have a negative influence on the
stability of the nc-DSC. This is an important advantage in
cell production processes. The permeation of water through
the polymer foil does not seem to be a factor influencing
 B.O’ Regan and M. Grätzel, Nature 353 (1991) 737
 M.K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-
Baker, E. Müller, P. Liska, N. Vlachopoulos and M.
Grätzel, J. Am. Chem. Soc. 115 (1993) 6382
 M.A. Green, K. Emery, K. Bücher, D.L. King and S.
Igari, Prog. Photovoltaics: Res. Appl. 6 (1998) 35-42
 Int. Pat. Method for manufacturing a liquid -containing
photovoltaic element and element manufactured according
to this method, A.C. Tip, M. Späth and P.M. Sommeling,
appl. No PCT/NL99/00370
 F. Pichot, S. Ferrere, R.J. Pitts, B.A. Gregg, J. of the
Electrochem. Society, 146 (11) 4324-4326 (1999)
This work has been carried out with financial support from
the ECN Cooperation Funding Programme.