Stabilization of the nanomorphology of polymer-fullerene "bulk heterojunction" blends using a novel polymerizable fullerene derivative

Article abstract of DOI:10.1039/b505361g

The morphological stabilization of donor-acceptor blends for bulk heterojunction solar cells can be achieved by cross-linking of the small molecular phase in the polymer matrix using a polymerizable fullerene derivative. In a comparative study the morphol

Full text of DOI:10.1039/b505361g

View Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/materials | Journal of Materials Chemistry
Stabilization of the nanomorphology of polymer–fullerene ‘‘bulk
heterojunction’’ blends using a novel polymerizable fullerene derivative
Martin Drees,*a Harald Hoppe,a Christoph Winder,a Helmut Neugebauer,a Niyazi S. Sariciftci,a
Wolfgang Schwinger,b Friedrich Scha¨ffler,b Christoph Topf,c Markus C. Scharber,c Zhengguo Zhud and
Russell Gaudianad
Received 15th April 2005, Accepted 12th October 2005
First published as an Advance Article on the web 26th October 2005
DOI: 10.1039/b505361g
The morphological stabilization of donor–acceptor blends for bulk heterojunction solar cells can
be achieved by cross-linking of the small molecular phase in the polymer matrix using a
polymerizable fullerene derivative. In a comparative study the morphology of polymer–fullerene
blend films was investigated using poly(3-hexylthiophene) (P3HT) as the polymer and C61-butyric
acid methyl ester (PCBM) or the newly synthesized polymerizable fullerene derivative, C61-butyric
acid glycidol ester, PCBG, as the acceptor molecule, respectively. Changes in the
nanomorphology due to heat treatment of the films were studied by means of atomic force
microscopy (AFM), transmission electron microscopy (TEM) and photoluminescence (PL)
studies. The polymerization process was monitored with infrared absorption studies. As
demonstrated by these comparative studies this newly synthesized fullerene gives considerable
stabilization of the solid state morphology in these blends. Such prevention of the long term, high
temperature instability of bulk heterojunction morphology displays an important route to
increase the operational stability of plastic solar cells in future applications.
The polymer–fullerene bulk heterojunction concept is based
Introduction
on an active layer formed from a blend of an electron donor
polymer and an acceptor fullerene material.14 The active layer
Organic photovoltaic devices have been studied intensively for
more than three decades due to their potential for low-cost,
can be produced in an easy fashion by casting films from a
light-weight, flexible solar cells.1,2 Many of the investigated
blend solution. In such a solid state blend the intimate mixing
organic semiconductors were based on small molecules
of the two phases enables an interface within a couple of
vacuum evaporated into thin films.3 The bilayer configuration
nanometres everywhere within the bulk, thus ‘‘bulk hetero-
used in many of the solar cell device architectures is based on
junction’’. Excitons created in either phase will encounter
the idea of exciton dissociation at donor–acceptor interfaces.
within their lifetime an interface to the opposite phase and
As such there is an electron transfer from the donor phase
dissociate there, resulting in charge carriers. Thus, the length
onto the acceptor phase with consecutive transport of the
scale of the phase separation of the two components within the
charges to the respective electrodes. A major breakthrough
bulk heterojunction shall be less than the exciton diffusion
in the field of donor–acceptor based systems for photovoltaic
length within the organic semiconductors, i.e. ca 10–20 nm.
applications was the discovery of an ultrafast charge transfer
Solar cell (AM1.5) power conversion efficiencies exceeding 3%
from conjugated polymers to the fullerene C60.4 In blends of
have been demonstrated using this approach.15
conjugated, semiconducting polymers with fullerenes the
An important parameter for the solar conversion efficiency
forward electron transfer takes place within approximately
of these blend devices is the nanomorphology of the active
45 femtoseconds.5 This charge transfer is the basis for efficient
layer.7 Close proximity of the phase boundary to the origin of
excitation on the polymer is detrimental,16,17 phase separation
photoinduced charge separation at the donor–acceptor
interfaces that is needed in photovoltaic devices. Several
into domains larger than the exciton diffusion length of ca
approaches towards efficient organic solar cells have been
10–20 nm has to be prevented. To obtain such intimate
investigated including polymer–fullerene blends,6–8 polymer–
mixing different solvents and post-production treatments were
polymer blends,9–11 and evaporated small molecules3,12,13 as
reported earlier.7,15 Once the nanomorphology is optimized
the active layer.
there is also a challenge to preserve the nanostructure of this.
For market implementation of organic solar cells the stability
aLinz Institute for Organic Solar Cells (LIOS), Johannes Kepler
University Linz, Altenbergerstr. 69, A-4040 Linz, Austria.
of the morphology is therefore a critical issue. Especially
diffusion of the low-molecular-weight components at
E-mail: martin.drees@jku.at
bInstitut fu¨r Halbleiter- und Festko¨rperphysik, Johannes Kepler
University Linz, Altenbergerstr. 69, A-4040 Linz, Austria
cKonarka Austria Forschungs und Entwicklungs GmbH, Gruberstr.
40-42, A-4010 Linz, Austria
elevated temperatures leading to phase separation needs to
be prevented. Here the fullerene molecules are the more critical
component in such bulk heterojunctions. Studies show that
fullerene molecules diffuse into large aggregates under elevated
temperatures leading to a demixing of the active layer.18,19
dKonarka Technologies, Inc., 100 Foot of John Street, Boott Mill South,
Third Floor, Suite 12 Lowell, MA 01852, USA
5158 | J. Mater. Chem., 2005, 15, 5158–5163
This journal is ß The Royal Society of Chemistry 2005

View Online
Several approaches have been made to stabilize the morpho-
logy of donor–acceptor blends for photovoltaic application. In
the so-called double-cable approach donor and acceptor are
covalently bound together.20,21 Also, fullerene-containing
polyester was used as a polymer acceptor in polymer–polymer
blends.22
Here we present
a novel approach to stabilize the
morphology of conjugated polymer–fullerene blend films by
introducing the polymerizable fullerene C61-butyric acid
glycidol ester (hereafter referred to as PCBG) as the fullerene
acceptor material. The polymerization reaction can be induced
by an initiator material (tris(pentafluorophenyl)borane) or by
heat, leading to a network of fullerene molecules throughout
the bulk of the active layer that prevents further diffusion
of fullerene molecules. The novelty of this approach is that
the PCBG molecule is very similar to the well-established
C61-butyric acid methyl ester (PCBM) and therefore all
methods to improve the morphology of polymer–fullerene
blends can be used. After optimum morphology realization
it can be fixed by polymerization of the PCBG molecules.
In a comparative study the morphology of polymer–
fullerene blend films was investigated using poly(3-hexyl-
thiophene) (P3HT) as the polymer and PCBM or the newly
synthesized PCBG as the acceptor molecule, respectively.
Changes in the nanomorphology due to heat treatment of the
films were studied by means of atomic force microscopy
(AFM), transmission electron microscopy (TEM) and photo-
luminescence (PL) studies. The results clearly demonstrate the
stabilization of the nanomorphology preventing diffusion of
the fullerene component. Fourier transform infrared spectro-
scopy (FTIR) was used to study the polymerization of PCBG
in films. In addition photovoltaic devices were fabricated and
compared in device performance using either PCBM or PCBG
as the electron acceptor.
Scheme 1 Synthesis of C61-butyric acid glycidol ester (PCBG).
The reaction was stirred at room temperature for 20 h and the
solvents were removed. The residue was purified by flash
chromatography (dichloromethane–hexanes, 1 : 4 ramping
up to 4 : 1) to give the pure PCBG (203 mg, 25% from the
carboxylic acid, 23% from PCBM): 1H NMR (CDCl3,
250 MHz): 7.94 (m, 2H), 7.60 (m, 3H), 4.47 (dd, 13.3, 3.0 Hz,
1H), 3.93 (dd, 12.3, 6.4 Hz, 1H), 3.22 (m, 1H), 2.91 (m, 3H),
2.67 (m, 3H), 2.22 (m, 2H); Exact mass calcd for C74H16O3Na
(M + Na): 975.0992, found: 975.0979.
Polymerization of PCBG
The polymerization of glycidyl ester can be initiated by the
addition of Lewis acid such as Et2O?BF3 (boron trifluoride
diethyl etherate).24 However, this causes PCBG to precipitate
from solution almost immediately making the solution
unusable for film coating. Thus a sterically more hindered,
organic-solvent-soluble Lewis acid tris(pentafluorophenyl)-
borane was used as a mild initiator. The polymerization
procedure of PCBG in solution was carried out using the
following procedure: PCBG (7 mg) was dissolved in dry
chloroform (500 mg) in a Schlenk tube. To this solution was
added tris(pentafluorophenyl)borane (0.07 mg in 0.5 mL dry
dichloromethane). The tube was sealed and the solution was
heated to 120 uC for 5 min resulting in brown precipitation of
the polymer or oligomer of the ring-opened product of PCBG.
Thin layer chromatography (TLC) of the solution showed only
traces of starting materials. The addition of mild acid initiator
or no acid initiator allows the polymerization to occur very
slowly in solution so that homogeneous films can be coated.
Further polymerization of PCBG in solid film can be realized
by heating. However, significant amount of PCBG remained
reacted even after the films were annealed at 120 uC for 5 min.
Experimental
Synthesis
All solvents were spectral grade unless otherwise noted.
Silica gel was obtained from VWR International. Anhydrous
1,2-dichlorobenzene, acetic acid, hydrochloric acid, carbon
disulfide, thionyl chloride, toluene, pyridine, glycidol, hexanes
and dichloromethane were purchased from Aldrich Chemical
Co., Inc.
The route to PCBG is outlined in Scheme 1. First
C61-butyric acid methyl ester (PCBM) was synthesized
according to literature procedure.23 PCBM (1.00 g, 1.1 mmol)
was dissolved in 1,2-dichlorbenzene and to this solution was
added acetic acid (100 mL) and concentrated aqueous
hydrochloric acid (40 mL). The mixture was heated to reflux
for 20 h. The solvents were removed in vacuo and residue was
suspended in methanol. The precipitates were collected by
filtration and dried in vacuum to afford the corresponding
carboxylic acid (0.94 g, 95%). The carboxylic acid was then
suspended in dry carbon disulfide (300 mL) under nitrogen
and thionyl chloride (300 mL) was added. The reaction was
heated to reflux for 18 h and the volatile components were
removed on a rotary evaporator. To the residue was added dry
toluene (100 mL), pyridine (4 mL) and then glycidol (200 mg).
Morphology studies
Regioregular P3HT was purchased from American Dye Source
(regio regularity .96%, molecular weight y36 000 g mol21
polydispersity 1.9). Blend films were prepared by spin casting
films from a 3% wt/vol P3HT–fullerene (1 : 2 by weight)
solution from dichlorobenzene onto glass substrates. The
surface morphology was then analysed with a Dimension
3100 AFM system (Digital Instruments, Santa Barbara, CA)
,
This journal is ß The Royal Society of Chemistry 2005
J. Mater. Chem., 2005, 15, 5158–5163 | 5159

View Online
in tapping mode. TEM specimen were prepared by spin casting
films onto glass substrates that were covered with a thin
layer of sodium metaphosphate (Victawet, SPI supplies,
Westchester, PA). Afterwards films were released from the
substrate by submerging the substrate in deionized water and
floating off the spincast polymer–fullerene layer. Films were
then lifted onto the TEM grids and dried in an N2 flow box
before measurements with a JEOL 2011 FasTEM high
resolution TEM in bright field mode. PL spectra were obtained
at room temperature under vacuum using a monochromator
in combination with a photodetector and an Ar-ion laser as
excitation source (476 nm, 40 mW). The heat treatment of the
films was done on a hot plate inside of an Argon glovebox.
FTIR studies
Fig. 1 AFM images of P3HT–PCBM (a–c) and P3HT–PCBG (d–f)
blend films spin cast on glass. The images show unheated films (a and
d), films annealed for 4 min at 140 uC (b and e), and films annealed for
1 h at 140 uC (c and f).
Attenuated total reflection infrared spectroscopy (ATR-FTIR)
was used to study the polymerization of PCBG in films.
Samples were prepared by dropcasting 1% wt/vol fullerene or
1% wt/vol polymer–fullerene–initiator (1 : 1 : 2.5 6 1024
)
micrometres in diameter. In contrast to that the PCBG
containing film barely shows any increase in surface roughness
compared to initial heating periods.
blend solutions from chloroform onto the Ge elements. For
the blend solutions the initiator was added 10 min before
dropcasting the films. Films were dried in vacuum for 1 h
before measurement. Annealing of the films was done on a hot
plate inside of an argon glovebox.
The AFM studies are complemented by TEM studies of the
films. While AFM can only give an image of the surface
topography of a film, TEM allows one to look into the bulk of
the film. In addition, selected areas can immediately be
tested for crystallinity by checking the diffraction image of
the electron beam. The aggregate formation already observed
in the AFM images of the long-heated P3HT–PCBM films can
also be seen in TEM images shown in Fig. 2. The dark square-
shaped borders in the images are from the Cu grid, which was
chosen to have a small grid-size for good support of the films.
The P3HT–PCBM blend film annealed at 140 uC for 1 h
shows large dendrite-like aggregates that have formed during
the heating process. They are several micrometres in diameter
and several tens of micrometres long. On the other hand
the P3HT–PCBG film heated under the same conditions
doesn’t show any observable aggregation. Electron scattering
experiments on the aggregates in the heated P3HT–PCBM
films show a diffraction pattern with distinct reflexes indicat-
ing crystallinity of the aggregates. The P3HT–PCBG films
don’t show any such signs of crystallinity even after long
heating periods.
Device preparation
Plastic solar cells were prepared as described elsewhere.15 In
the work presented here the photoactive materials P3HT and
PCBG or P3HT and PCBM were dissolved in chloroform
(3% wt/vol) in a concentration 1 : 2 by weight. The cross
linking of PCBG molecules in the active layer was induced
by adding small amounts of initiator tris(pentafluorophenyl)-
borane or by simply storing the sample for up to 1 h at about
140 uC. At this temperature P3HT–PCBM devices show a
strong morphology change while P3HT–PCBG films remain
clear indicating that a cross-linking takes place at this
temperature. Different metal electrodes where evaporated as
cathode including LiF/Al, Ca, Ca/Al. Devices were post treated
similar to the procedures in ref. 15 by annealing devices with
the metal top electrode on a hot plate in an Ar glovebox
for 4 min at 140 uC. Longer annealing times exceeding the
optimum post treatment time were chosen to study the
influence of elevated temperatures on the device performance.
Results
A. AFM and TEM studies
AFM is a powerful tool to study the nanomorphology of
organic blends in thin films. The surface topography usually
gives a good insight into the film formation ability and the
tendency of the components to phase separate. Fig. 1 shows a
series of AFM images of the P3HT–fullerene blend films. The
unheated blend films are both very flat and do not show any
coarse separation into different phases. Upon heating to 140 uC
for 4 min the surface roughness of both films increases. The
film containing PCBG is slightly rougher than the film
containing PCBM. When the heating time is extended to 1 h
the PCBM containing film shows large aggregates of several
Fig. 2 TEM images of a P3HT–PCBM film (left) and a P3HT–PCBG
film (right), both annealed at 140 uC for 1 h. While the PCBM has
phase separated into large crystals, the PCBG does not show any large
scale phase separation.
5160 | J. Mater. Chem., 2005, 15, 5158–5163
This journal is ß The Royal Society of Chemistry 2005

View Online
B. Photoluminescence measurements
The photoluminescence of blend films, which has to be
quenched by photoinduced electron transfer upon intimate
mixing of the donor and acceptor phases, can be used to study
the stability of the nanomorphology of donor–acceptor blends.
If there is (or will be) a large scale phase separation which
goes beyond the exciton diffusion length scale of 10–20 nm, the
luminescence will be recovered again as compared to the
intimate mixture. Fig. 3 shows that upon introducing PCBM
or PCBG as the acceptor in the films the PL is quenched by
two orders of magnitude compared to pristine polymer film.
Once these films are heat treated, the PL increases for both
PCBM and PCBG by an order of magnitude.
In addition to the magnitude also the shape of the PL
spectra changes. This can be seen in the normalized PL spectra
in Fig. 4. The pristine polymer film shows its maximum PL at
720 nm and a smaller peak at 665 nm. In contrast to this the
unheated blend films have their maximum PL at 640 nm and a
second, weaker peak at 700 nm.
Fig. 4 Normalized PL spectra of unheated and heated P3HT–PCBM
(a) and P3HT–PCBG (b) films. While the unheated blend films show
their maximum PL around 640 nm and a second peak at 700 nm, the
pristine polymer has its maximum PL at 720 nm and a smaller peak
around 667 nm. The heated P3HT–PCBM films show PL peaks that
shift with longer heating times from the unheated to the pristine
polymer peak positions. The heated P3HT–PCBG films show two
peaks of about the same magnitude, one of them close the maximum
PL of the unheated blends and one close to the maximum of the
pristine polymer film. The heating time does not influence the shape of
the PL in the case of P3HT–PCBG films.
Heat treatment of the P3HT–PCBM film results in a redshift
of the PL spectrum (compare Fig. 4a). Also the intensity of
the lower wavelength peak gets weaker than the second peak.
While the film heated for 4 min shows an intermediate state
of shifting from the unheated blend film PL to the pristine
polymer film PL, the film heated for 1 h exhibits a normalized
PL spectrum that closely matches the spectrum of the pristine
polymer film.
peak barely shifts compared to the unheated film. The second
peak redshifts to 710 nm and therefore is close to the position
of the stronger peak of the pristine polymer film. Both peaks of
the heated P3HT–PCBG spectra have the same intensity.
The PL spectra of the P3HT–PCBG blend films change in a
different way (compare Fig. 4b). First of all the change in the
shape of the spectrum is independent of the heating time. The
normalized spectrum of the film heated for 4 min closely
matches the one heated for 1 h. Also the lower wavelength
C. FTIR studies on the cross-linking
FTIR can be used to study the molecular structure of
materials. Epoxy rings typically have absorption features in
the region of 800 to 1000 wavenumbers where PCBM does not
show any strong features. These features should disappear
during the ring-opening polymerization. To demonstrate this,
a PCBG film dropcast from chloroform was measured using
ATR-FTIR technique. Afterwards the film was polymerized
by dropcasting a 0.05% wt/vol tris(pentafluorophenyl)borane
solution in chloroform on top of the fullerene film and then
remeasured. As reference a PCBM film was measured and
treated the same way. Fig. 5 shows the IR absorption spectra
of the fullerene films before addition of the initiator and of a
pristine initiator film (drop cast from a 0.05% wt/vol solution
in chloroform). In addition Fig. 5 shows the absorption
difference spectra of the fullerene films before and after the
addition of the initiator. The absorption difference spectrum of
the PCBG film shows two negative peaks at 910 and around
850 cm21 whereas the PCBM does not show any changes. This
shows that the epoxy ring can be detected in PCBG. Also its
disappearing during the polymerization reaction is observable.
The positive signal at 975 cm21 in the absorption difference
spectra stem from the initiator that was added in between
measurements. In addition it should be mentioned that the
polymerized PCBG film was completely insoluble whereas the
PCBM film was easily dissolved after the initiator treatment.
In a next step P3HT–PCBG–initiator blend films were
prepared with very low concentration of initiator (see
Fig. 3 Photoluminescence spectra of unheated and heated P3HT–
PCBM and P3HT–PCBG blend films. In the unheated blend films the
PL is quenched by two orders of magnitude compared to the pristine
polymer. In the heated films the PL increases slightly compared to the
unheated films.
This journal is ß The Royal Society of Chemistry 2005
J. Mater. Chem., 2005, 15, 5158–5163 | 5161

View Online
current (5.2 mA cm22) and the open circuit voltage (530 mV)
remain unchanged the FF is reduced from 0.621 to 0.3. P3HT–
PCBM devices that were prepared as references showed a
completely different degradation behaviour. The shape of the
current–voltage curve under illumination did not change. Only
the short circuit current decreased due to the demixing of
P3HT and PCBM.
Discussion
The AFM and the TEM studies both clearly show that the
blend films containing PCBG do not phase separate under the
prolonged heating. While the PCBM can easily diffuse when
the films are annealed at 140 uC and forms aggregates at the
surface of the film, this diffusion is not observed for PCBG.
Both studies show that PCBG stabilizes the morphology of the
blend films preventing phase separation of the polymer and
fullerene content.
Fig. 5 IR absorption spectrum of
a PCBG, a PCBM and a
tris(pentafluorophenyl)borane film (top) and IR absorption difference
spectra of PCBG and PCBM before and after addition of the initiator
(bottom).
It should be noted here that the annealing for AFM and
TEM studies was done without the metal top electrode. It has
been shown that confinement of the active layer by a metal
electrode can reduce the process of phase separation.25
Nevertheless the reduction in short circuit current in our
P3HT–PCBM devices clearly shows that the demixing of
polymer and fullerene also occurs with the metal electrode
in place.
experimental section for details). Higher initiator concentra-
tions lead to doping of the photoactive layer and are therefore
not suitable for photovoltaic device application. Films were
measured before and after 1 h annealing at 140 uC. For
comparison a pristine P3HT film was measured. Results are
shown in Fig. 6. While the epoxy ring feature at 850 cm21 is
overlapped by an absorption of the P3HT, the absorption at
910 cm21 is clearly observable. The peak intensity of the epoxy
ring absorption is slightly reduced after heating of the film.
The PL studies show about the same PL quenching in blends
with PCBG and PCBM compared to a pristine polymer film.
This shows the modification of PCBM to PCBG does not
influence the molecule’s ability as an electron acceptor. Also
even though the PL quenching in all heated films is weaker
than in the unheated films, Fig. 3 indicates that the PL
quenching is less influenced by long heating periods when
PCBG is used as the acceptor. Since the PL should increase
when polymer and fullerene separate into different phases the
lack of increase in the PL signal after 1 h heating compared to
the 4 min again shows the stability of the morphology.
There is a difference in the PL spectrum shape of a pristine
P3HT film as compared to the unheated blend film. This is due
D. Devices
P3HT–PCBG devices with efficiencies .2% (80 mW cm22
,
AM1.5 simulated) can be prepared. Devices using PCBG
have not been optimized. When heating devices to 140 uC for
several minutes (exceeding the post treatment annealing) the
active layer of the device showed no change in the film
morphology. However a strong degradation of the current
voltage curve is observed (Fig. 7). While the short circuit
Fig. 7 Current–voltage curves measured under illumination (AM
1.5 simulated, 80 mW cm22). Squares, P3HT–PCBG device after
production and post treatment, circles, P3HT–PCBG device after an
additional thermal treatment (10 min, 140 uC).
Fig. 6 IR absorption spectra of a P3HT–PCBG–initiator film before
(solid) and after (dotted) heating at 140 uC for 1 h. For reference a
pristine P3HT film is shown (dashed).
5162 | J. Mater. Chem., 2005, 15, 5158–5163
This journal is ß The Royal Society of Chemistry 2005

View Online
to the fact that the regioregular P3HT chains can organize
themselves into ordered structures in the pristine film.26 This
lamellar order is interrupted by the presence of the fullerene
molecules. In the case of PCBM, where the fullerene molecules
can easily move in the film at elevated temperatures, the
polymer chains can rearrange into their ordered structure and
the shape of the PL spectra shifts with longer heating periods
more and more from the disordered mixture PL in the unheated
films to the spectrum of pristine, ordered P3HT films.
In contrast to this the PL spectra of the P3HT–PCBG
films do not change shape with long heating periods. The
normalized PL of the film heated for 1 h exactly matches
the short-heated film. This shows that the PCBG fixes the
morphology of the film rapidly and does not allow any further
rearrangements in the film.
solar cell in future applications. The electrical stability of these
devices has to be achieved by improving the method of
crosslinking to minimize the incomplete reaction.
Acknowledgements
This work has been performed partially within the Christian
Doppler Society’s dedicated laboratory on plastic solar cells
co-funded by Konarka Corp. The TEM images were recorded
at the Technical Service Unit (TSE) of the Johannes Kepler
University.
References
1 Clean Electricity from Photovoltaics, ed. M. D. Archer, R. Hill,
Imperial College Press, London, 2001.
The IR studies show that PCBG can be polymerized with
significant amounts of initiator that are not suitable for
photovoltaic device operation. The epoxy ring absorption
features fully disappear showing the ring-opening polymeriza-
tion. Unfortunately the polymerization of PCBG is much less
efficient using small amounts of initiator and annealing. The
IR absorption spectra show significant amounts of epoxy rings
remaining even after 1 h of heat treatment at 140 uC.
2 Organic Photovoltaics: Concepts and Realization, ed. C. J. Brabec,
V. Dyakonov, J. Parisi, N. S. Sariciftci, Springer-Verlag, Berlin,
2003.
3 C. W. Tang, Appl. Phys. Lett., 1986, 48, 2, 183.
4 N. S. Sariciftci, L. Smilowitz, A. J. Heeger and F. Wudl, Science,
1992, 258, 1474.
5 C. J. Brabec, G. Zerza, G. Cerullo, S. De Silvestri, S. Luzzati,
J. C. Hummelen and N. S. Sariciftci, Chem. Phys. Lett., 2001, 340,
3–4, 232.
6 G. Yu, J. Gao, J. C. Hummelen, F. Wudl and A. J. Heeger,
Science, 1995, 270, 1789.
Even though AFM, TEM and PL studies show that the
morphology of P3HT–PCBG blend films is very stable, the IR
absorption studies show that the PCBG molecules do not all
polymerize into a network. A significant amount of epoxy
rings still remains in the film. It is likely that during heat
treatment small amounts of PCBG polymerize to very short
oligomer chains but these short chains are already enough to
suppress diffusion, therefore explaining the discrepancy
between stable morphology and remaining epoxy rings.
While the morphology of the P3HT–PCBG blend films is
stabilized by the polymerization reaction, the electrical charac-
teristics of the solar cells built from this blend show an unusual
degradation during the heat treatment. The origin of this
degradation process is not known. It could be related to the
incomplete polymerization of PCBG molecules. The shape of
the current–voltage curve is typical for a barrier blocking the
charge transport which is usually formed near the electron
accepting electrode. It is interesting to note that the short
circuit current and the open circuit voltage remain unchanged
after the thermal treatment. This indicates that the active layer
is not degraded and the film morphology does not change.
Most likely the epoxy groups may react with the evaporated
metal atoms forming an insulating layer at the metal–active
layer interface.
7 S. E. Shaheen, C. J. Brabec, N. S. Saricftci, F. Padinger,
T. Fromherz and J. C. Hummelen, Appl. Phys. Lett., 2001, 78, 841.
8 L. Chen, D. Godovsky, O. Ingana¨s, J. C. Hummelen, R. A. J. Janssens,
M. Svensson and M. R. Andersson, Adv. Mater., 2000, 12, 18, 1367.
9 J. J. M. Halls, C. A. Walsh, N. C. Greenham, E. A. Marseglia,
R. H. Friend, S. C. Moratti and A. B. Holmes, Nature, 1995, 376,
6540, 498.
10 G. Yu and A. J. Heeger, J. Appl. Phys., 1995, 78, 7, 4510.
11 M. Granstrom, K. Petritsch, A. C. Arias, A. Lux, M. R. Andersson
and R. H. Friend, Nature, 1998, 395, 257.
12 P. Peumans, A. Yakimov and S. R. Forrest, J. Appl. Phys., 2003,
93, 7, 3693.
13 B. Maennig, J. Drechsel, D. Gebeyehu, P. Simon, F. Kozlowski,
A. Werner, F. Li, S. Grundmann, S. Sonntag, M. Koch, K. Leo,
M. Pfeiffer, H. Hoppe, D. Meissner, N. S. Sariciftci, I. Riedel,
V. Dyakonov and J. Parisi, Appl. Phys. A: Solid Surf., 2004, 79, 1.
14 N. S. Sariciftci and A. J. Heeger, in Handbook of Conductive
Molecules and Polymers, John Wiley & Sons, New York, 1997,
vol. 1, ch. 8.
15 F. Padinger, R. S. Rittberger and N. S. Sariciftci, Adv. Funct.
Mater., 2003, 13, 2, 1.
16 J. J. M. Halls, K. Pichler, R. H. Friend, S. C. Moratti and
A. B. Holmes, Appl. Phys. Lett., 1996, 68, 3120.
17 D. Vacar, E. S. Maniloff, D. W. McBranch and A. J. Heeger, Phys.
Rev. B: Condens. Matter, 1997, 56, 4573.
18 H. Hoppe, M. Niggemann, C. Winder, J. Kraut, R. Hiesgen,
A. Hinsch, D. Meissner and N. S. Sariciftci, Adv. Funct. Mater.,
2004, 14, 1005.
19 X. Yang, J. K. J. Van Duren, R. A. J. Janssen, M. A. J. Michels
and J. Loos, Macromolecules, 2004, 37, 2151.
20 A. Cravino, G. Zerza, M. Maggini, S. Bucella, M. Svensson,
M. R. Andersson, H. Neugebauer and N. S. Sariciftci, Chem.
Commun., 2000, 2487.
Conclusions
The morphological stabilization of the bulk heterojunction
solar cells can be achieved by cross linking of the small
molecular phase using a polymerizable fullerene derivative. As
demonstrated by comparative nanomorphology studies using
combined AFM, TEM and photoluminescence experiments
this approach gives considerable stabilization of the solid state
morphology. Such prevention of the long term, high tempera-
ture instability of bulk heterojunction morphology displays an
important route to increase the operational stability of plastic
21 A. M. Ramos, M. T. Rispens, J. K. J. van Duren, J. C. Hummelen
and R. A. J. Janssen, J. Am. Chem. Soc., 2001, 123, 27, 6714.
22 M. G. Nava, S. Setayesh, A. Rameau, P. Masson and
J. F. Nierengarten, New J. Chem., 2002, 11, 1584.
23 J. C. Hummelen, B. W. Knight, F. LePeq, F. Wudl, J. Yao and
C. L. Wilkins, J. Org. Chem., 1995, 60, 532.
24 M. Miyamoto, Y. Saeki, C. W. Lee, Y. Kimura, H. Maeda and
K. Tsutsui, Macromolecules, 1997, 30, 6067.
25 P. Peumans, S. Uchida and S. R. Forrest, Nature, 2003, 425, 158.
26 T. A. Chen, X. Wu and R. D. Rieke, J. Am. Chem. Soc., 1995, 117,
233.
This journal is ß The Royal Society of Chemistry 2005
J. Mater. Chem., 2005, 15, 5158–5163 | 5163