Thermal stability of poly[2-methoxy-5-(2′-phenylethoxy)-1,4- phenylenevinylene] (MPE-PPV):Fullerene bulk heterojunction solar cells

Article abstract of DOI:10.1021/ma201911a

To improve the thermal stability of polymer:fullerene bulk heterojunction solar cells, a new polymer, poly[2-methoxy-5-(2′-phenylethoxy)-1,4- phenylenevinylene] (MPE-PPV), has been designed and synthesized, which showed an increased glass transition tempe

Full text of DOI:10.1021/ma201911a

ARTICLE
pubs.acs.org/Macromolecules
Thermal Stability of Poly[2-methoxy-5-(20-phenylethoxy)-1,4-
phenylenevinylene] (MPE-PPV):Fullerene Bulk Heterojunction Solar Cells
J. Vandenbergh,† B. Conings,† S. Bertho,† J. Kesters,† D. Spoltore,† S. Esiner,‡ J. Zhao,§ G. Van Assche,§
||
M. M. Wienk,‡ W. Maes,†, L. Lutsen,^ B. Van Mele,§ R. A. J. Janssen,‡ J. Manca,†,^ and
D. J. M. Vanderzande*,†,^
†Institute for Materials Research (IMO), Hasselt University, Universitaire Campus, Building D, B-3590 Diepenbeek, Belgium
‡Molecular Materials and Nanosystems, Eindhoven University of Technology, P.O. Box 513, 5600MB Eindhoven, The Netherlands
§Physical Chemistry and Polymer Science, Department of Materials and Chemistry, Vrije Universiteit Brussel, Pleinlaan 2, B-1050
Brussels, Belgium
Molecular Design and Synthesis, Katholieke Universiteit Leuven, Celestijnenlaan 200F, B-3001 Heverlee, Belgium
^Division IMOMEC, IMEC, Universitaire Campus, Wetenschapspark 1, B-3590 Diepenbeek, Belgium
S Supporting Information
b
ABSTRACT: To improve the thermal stability of polymer:fullerene bulk hetero-
junction solar cells, a new polymer, poly[2-methoxy-5-(20-phenylethoxy)-1,4-
phenylenevinylene] (MPE-PPV), has been designed and synthesized, which showed
an increased glass transition temperature (Tg) of 111 ꢀC. The thermal character-
istics and phase behavior of MPE-PPV:[6,6]-phenyl C61-butyric acid methyl ester
([60]PCBM) blends were investigated by means of modulated temperature differ-
ential scanning calorimetry and rapid heatingꢀcooling calorimetry. The thermal
stability of MPE-PPV:[60]PCBM solar cells was compared with devices based on the
reference MDMO-PPV material with a Tg of 45 ꢀC. Monitoring of the photo-
currentꢀvoltage characteristics at elevated temperatures revealed that the use of high-
Tg MPE-PPV resulted in a substantial improvement of the thermal stability of the
solar cells. Furthermore, a systematic transmission electron microscope study of the
active polymer:fullerene layer at elevated temperatures likewise demonstrated a more
stable morphology for the MPE-PPV:[60]PCBM blend. Both observations indicate that the use of high-Tg MPE-PPV as donor material
leads to a reduced free movement of the fullerene molecules within the active layer of the photovoltaic device. Finally, optimization of the
PPV:fullerene solar cells revealed that for both types of devices the use of [6,6]-phenyl C71-butyric acid methyl ester ([70]PCBM) resulted
in a substantial increase of current density and power conversion efficiency, up to 3.0% for MDMO-PPV:[70]PCBM and 2.3% for MPE-
PPV:[70]PCBM.
’ INTRODUCTION
which can cross-link and freeze in the ultimate morphology after a
thermal or UV light treatment.21 In this paper, a third approach is
further investigated, which is based on the use of a donor polymer
with a high glass transition temperature (Tg).22ꢀ26 In this context
the question arises how to adjust the chemical structure of the
polymer to induce a higher Tg, without jeopardizing the opto-
electronic characteristics of the polymer. Here we demonstrate that
by changing the side-chain substituents of the polymer, the Tg can
be deliberately increased. Poly[(2-methoxy-5-(30,70-dimethyl-
octyloxy))-1,4-phenylenevinylene] (MDMO-PPV) has a quite
low Tg of around 45 ꢀC due to the long flexible alkoxy side chains,
which cause a plasticizing effect on the PPV backbone.22 To reduce
the side-chain flexibility, a novel poly(p-phenylenevinylene) derivative,
poly[2-methoxy-5-(20-phenylethoxy)-1,4-phenylenevinylene]
In polymer:fullerene bulk heterojunction solar cells, the
nanoscale morphology of the active layer influences the electronic
properties of the devices to a great extent.1ꢀ4 Via proper sample
preparation, such as annealing conditions and the choice of proces-
sing solvents, the (nano)morphology of the donorꢀacceptor blend
can be optimized, leading to improved exciton dissociation and
charge transport.3ꢀ6 It was already demonstrated, however, that this
initial optimized morphology deteriorates rather fast (during storage
at elevated temperatures).7ꢀ17 Long-term annealing treatments
induce severe phase separation between the donor polymer and
the fullerene acceptor material, leading to a decreased interfacial area
and less efficient exciton dissociation.18ꢀ20 To stabilize the initial
(nano)morphology of a bulk heterojunction layer, several pathways
can be followed. A first approach is to use donorꢀacceptor diblock
copolymers to covalently fixate the morphology via the scale of the
block lengths.13,14 A second way to obtain a stable network is to
build in a certain amount of functional groups in the donor polymer
Received: August 22, 2011
Revised:
October 2, 2011
Published: October 17, 2011
r
2011 American Chemical Society
8470
dx.doi.org/10.1021/ma201911a Macromolecules 2011, 44, 8470–8478
|

Macromolecules
ARTICLE
(MPE-PPV), has been designed and synthesized. The bulky and
rigid phenyl side groups decrease the rotational freedom of the
polymer backbone, thereby increasing the Tg at which the amor-
phous polymer transfers from a rigid glass into soft flexible material.
The Tg of the polymer as well as the phase behavior in blends with
[6,6]-phenyl C61-butyric acid methyl ester ([60]PCBM) was
determined by means of advanced thermal analysis techniques, i.e.,
modulated temperature differential scanning calorimetry (MTDSC)
and rapid heatingꢀcooling calorimetry (RHC). The effect of
elevated temperatures on the bulk morphology of MPE-PPV:-
[60]PCBM blends and the photovoltaic performance of the
resulting devices was studied via transmission electron micro-
scopy (TEM) and via in-situ monitoring of the photovoltaic output
(i.e., the short circuit current, open circuit voltage, fill factor, and
efficiency). The results will be compared with reference MDMO-
PPV:[60]PCBM solar cells. Finally, the MPE-PPV:fullerene and
MDMO-PPV:fullerene devices were optimized for maximal photo-
voltaic output. Different spin-coating solvents were applied, and
both [60]PCBM and [6,6]-phenyl C71-butyric acid methyl ester
([70]PCBM) fullerene derivatives were used as acceptor materials.
and indium. About 0.5 mg samples were sealed in aluminum RHC
crucibles with a weight of less than 2 mg. First, the samples were heated
to 350 ꢀC for 0.1 min. Subsequently, they were quenched in the RHC
cell to ꢀ150 ꢀC and then reheated to 350 ꢀC at 500 K minꢀ1
.
Nonoptimized bulk heterojunction solar cells, combining either
poly[(2-methoxy-5-(30,70-dimethyloctyloxy))-1,4-phenylenevinylene]
(MDMO-PPV) with Tg ≈ 45 ꢀC (Covion, Mw = 716 kg/mol, PD = 6.5)
or poly[(2-methoxy-5-(20-phenylethoxy))-1,4-phenylenevinylene] (MPE-
PPV) with Tg ≈ 111 ꢀC (Mw = 123 kg/mol, PD = 2.3), as electron donor
materials with [60]PCBM (Solenne-BV) as acceptor material were made
according to the following preparation guidelines. The devices had an
ITO/PEDOTꢀPSS/polymer:[60]PCBM/Ca/Al structure. Each device
had an active area of 25.0 mm2. Indium tin oxide (ITO, 100 nm)-coated
glass plates were successively cleaned in a soap solution, demineralized
water, and acetone, each for 10 min in an ultrasonic bath. This was
followed by an additional cleaning step in boiling isopropanol for 10 min
and an UV/O3 treatment of 30 min. A 60 nm thick poly(3,4-ethylene-
dioxythiopheneꢀpolystyrenesulfonate (PEDOTꢀPSS (HC Starck,
Clevios P VP AI 4083)) layer was spin-coated on the clean glass/ITO
substrates. The substrates were dried for 20 min on a hot plate at 120 ꢀC.
The active layer, with thickness of 80 nm, consisting of a blend of the
respective PPV polymer and [60]PCBM in chlorobenzene, was spin-
coated on top of the PEDOTꢀPSS layer. The polymer:[60]PCBM ratio
for MDMO-PPV and MPE-PPV was 1:4 (this ratio was chosen because
it resulted in the best solar cell performance). The concentrations of the
polymer solutions were 0.5 wt % (weight percentage of polymer in
chlorobenzene solvent). To obtain complete dissolution, the polymer:-
[60]PCBM solutions were stirred overnight at 50 ꢀC. The solar cells
were completed by subsequently evaporating 20 nm of Ca on top of the
active layer, followed by 60 nm of Al. The currentꢀvoltage (IV)
characteristics were measured with a Newport class A solar simulator
(model 91195A) calibrated with a silicon solar cell to give an AM1.5
spectrum. The influence of thermal annealing on the photovoltaic
performance was measured in a setup that measures IV characteristics
at regular time intervals (illumination with a White LED LZ4-00CW10
(LedEngin)) while the samples were kept under continuous annealing in
a nitrogen atmosphere. In between the measurements, the samples were
kept in the dark. The active layer morphology was studied with a
transmission electron microscope (FEI Tecnai Spirit), operated at 120
kV. To prepare TEM specimens, the flotation technique was used to
remove the films from the glass surface via treatment with HF (40%).
The films floating on the surface of deionized water were subsequently
picked up by a 400 mesh copper grid.
Optimized bulk heterojunction solar cells with either MDMO-PPV or
MPE-PPV as electron donor materials and either [60]PCBM or
[70]PCBM (Solenne-BV) as acceptor material were made according
to the following preparation guidelines. The devices had an ITO/
PEDOTꢀPSS/polymer:fullerene/LiF/Al structure. The devices had
active areas of either 9 or 16 mm2. Indium tin oxide (ITO, 100 nm)-
coated glass plates were successively cleaned and rubbed in acetone,
ethanol, soap solution, demineralized water, and isopropanol, each for
15 min in an ultrasonic bath, followed by an UV/O3 treatment of 30 min.
A 60 nm thick PEDOT-PSS (HC Starck, Clevios P VP AI 4083) layer
was spin-coated on the clean glass/ITO substrates. Subsequently,
polymer:fullerene (1:4) active layers with different thicknesses were
spin-coated from different solvents on top of the PEDOT-PSS layer. The
concentrations of the solutions were 0.5 wt % (weight percentage of
polymer in solvent) for both polymers. To obtain complete dissolution,
the polymer:fullerene solutions were stirred overnight at 60 ꢀC. The
toluene solutions were stirred subsequently at 90 ꢀC for 1 h. The solar
cells were completed by subsequently evaporating 1 nm of LiF on top of
the active layer, followed by 100 nm of Al. The currentꢀvoltage (IV)
characteristics were measured with a Keithley 2400 SMU, under white
light illumination with UV(GG 385) and infrared (KG1) filtered light
’ EXPERIMENTAL SECTION
General. Unless stated otherwise, all reagents and chemicals were
obtained from commercial sources (Acros and Aldrich) and used
without further purification. 1,4-Dioxane was dried by distillation from
Na/benzophenone. NMR spectra were recorded with a Varian Inova
300 spectrometer using a 5 mm probe. Gas chromatography/mass
spectrometry (GC/MS) analyses were carried out with TSQ-70 and
Voyager mass spectrometers (Thermoquest); the capillary column was a
Chrompack Cpsil5CB or Cpsil8CB. Analytical size exclusion chroma-
tography (SEC) was performed using a Spectra series P100 (Spectra-
Physics) pump equipped with two mixed-B columns (10 μm, 0.75 cm ꢁ
30 cm, Polymer Laboratories) and a refractive index detector (Shodex)
at 40 ꢀC. THF was used as the eluent at a flow rate of 1.0 mL/min.
Molecular weight distributions are given relative to polystyrene stan-
dards. FT-IR spectra were collected with a Perkin-Elmer Spectrum
One FT-IR spectrometer (nominal resolution 4 cmꢀ1, summation of
16 scans). UVꢀvis spectroscopy was performed on a Varian Cary 500
UVꢀvisꢀNIR spectrophotometer (scan rate 600 nm/min). Samples for
thin-film FT-IR and UVꢀvis characterization were prepared by drop-
casting the precursor polymer from a CHCl3 solution (10 mg/mL) onto
NaCl or quartz disks.
The DSC and MTDSC measurements were performed on a TA
Instruments Q2000 DSC (Tzero DSC technique) with the MDSC
option, equipped with an RCS cooling accessory and purged with
nitrogen (50 mL minꢀ1). Baseline, heat capacity, and temperature were
calibrated with sapphire and indium. About 5 mg samples were sealed in
aluminum crucibles (Tzero, 40 μL). For the DSC measurements, the
samples were heated and cooled between ꢀ90 and 300 ꢀC at 10 K
minꢀ1. To reduce the sample preparation effects, the first cooling and
the second heating were used for discussion. Actually, such a methodol-
ogy can to a certain extent avoid the effects of casting solvents19,20 and
annealing conditions.6,21 Heating and cooling curves of subsequent
cycles coincide. The melting and crystallization peaks were characterized
by their peak temperatures. For the MTDSC measurements, the
modulation amplitude was 0.5 K with a period of 60 s. First, the samples
were kept at 300 ꢀC for 1 min. Subsequently, they were quenched in the
DSC cell at about 100 K minꢀ1 to ꢀ90 ꢀC and then reheated to 300 ꢀC
at 2.5 K minꢀ1. The RHC measurements were performed on a TA
Instruments RHC prototype (Tzero DSC technique) equipped with a
LNCS cooling accessory and purged with neon (10 mL minꢀ1).
Baseline, heat capacity, and temperature were calibrated with sapphire
8471
dx.doi.org/10.1021/ma201911a |Macromolecules 2011, 44, 8470–8478

Macromolecules
ARTICLE
Scheme 1. Synthesis of MPE-PPV via the Gilch Precursor Routea
a Reagents and conditions: (i) 2-bromoethylbenzene, NatBuO, NaI, EtOH (39% yield); (ii) p-CH2O, Ac2O, HCl (88% yield); (iii) 1: KtBuO, 1,4-
dioxane; 2: ΔT (56% yield).
from a tungsten halogen lamp with a maximum estimated light output
of 1000 W/m2. The solar simulator was calibrated with a silicon solar
reference cell. The currents reported were corrected for spectral
mismatch.8
Synthesis. 1-Methoxy-4-(20-phenylethoxy)benzene (1). In
a three-necked round-bottom flask, 4-methoxyphenol (32.0 g,
0.256 mol) and NatBuO (29.8 g, 0.310 mol) were dissolved in ethanol
(260 mL). The mixture was stirred for 1 h at ambient temperature
under a nitrogen atmosphere. Subsequently, 2-bromoethylbenzene
(52.1 g, 0.280 mol) was added dropwise to the mixture, and NaI
(1.0 g, 0.007 mol) was additionally added in one go. The mixture was
heated at reflux for 16 h under a nitrogen atmosphere. Water (200 mL)
was added, ethanol was evaporated, and the mixture was extracted with
CHCl3 (3 ꢁ 50 mL). The organic phase was washed with 10% NaOH
and subsequently dried over MgSO4. After filtration and evaporation of
the solvent, the crude product was obtained as a clear brown solid. The
product was purified by column chromatography (SiO2, eluent hexane/
chloroform 1/1) to afford white crystals. Yield: 39% (23.0 g). 1H NMR
(CDCl3): 7.35ꢀ7.18 (m, 5H, ArH), 6.82 (s, 4H, ArH), 4.10 (t, 2H,
OCH2, J = 6.1 Hz), 3.75 (s, 3H, OCH3), 3.07 (t, 2H, CH2ꢀAr, J = 5.9
Hz). 13C NMR (CDCl3): 151.7, 150.8, 138.1, 129.3, 128.5, 115.0, 114.1,
69.2, 56.0, 35.7. MS (EI, m/z): 228 [M+]. IR (NaCl, cmꢀ1): 2956, 2924,
2850, 1591, 1500, 1462, 1412, 1380, 1315, 1261, 1206, 1040, 867.
2,5-Bis(chloromethyl)-1-methoxy-4-(20-phenylethoxy)ben-
zene (2). 1-Methoxy-4-(20-phenylethoxy)benzene (1) (3.2 g, 0.014 mol)
and p-formaldehyde (1.2 g, 0.038 mol) were added to a three-necked
round-bottom flask, which was placed in an ice bath and under a nitrogen
atmosphere. After addition of HCl (37%) (9.0 g, 0.092 mol), acetic
anhydride (14.2 g, 0.140 mol) was added dropwise at such a rate that the
temperature of the mixture did not exceed 70 ꢀC. Subsequently, the
mixture was stirred for 6 h at 70 ꢀC under a nitrogen atmosphere.
Afterward, the mixture was cooled down to room temperature and
decanted into water (100 mL) to form a white precipitate. After filtration
of the mixture, the crude product was purified via recrystallization in
CHCl3 to obtain white crystals. Yield: 88% (4.0 g). 1H NMR (CDCl3):
7.38ꢀ7.18 (m, 5H, ArH), 6.88 (s, 2H, ArH), 4.60 (s, 2H, CH2Cl), 4.58 (s,
Figure 1. Thin film UVꢀvis spectra before (full line) and after (dashed
2H, CH2Cl), 4.20 (t, 2H, OCH2, J = 6.6 Hz), 3.82 (s, 3H, OCH3), 3.10 (t,
line) thermal elimination (a) and FT-IR spectrum (b) of MPE-PPV.
2H, CH2ꢀAr, J = 6.4 Hz). 13C NMR (CDCl3): 151.2, 150.30, 138.0,
129.3, 128.5, 127.5, 127.0, 126.5, 114.5, 113.0, 69.5, 56.2, 41.0, 35.8. MS
(EI, m/z): 324 [M+]. IR (NaCl, cmꢀ1): 2956, 2924, 2850, 1591, 1500,
polymer was redissolved in 1,4-dioxane (100 mL). The mixture was
refluxed for 16 h under a nitrogen atmosphere to ensure complete
formation of theconjugatedsystem. Subsequently, the mixture was cooled
down to room temperature, and the polymer was precipitated via addition
of cold methanol (100 mL). The mixture was filtered, and the red fibrous
polymer was collectedanddriedinvacuo. The residualfractions contained
only monomers and oligomers. Yield: 56% (218 mg). 1H NMR (CDCl3):
7.6ꢀ7.0 (br, 9H, ArH/CHdCH), 4.4ꢀ4.2 (br, 2H, OCH2), 4.0ꢀ3.8 (br,
3H, OCH3), 3.25ꢀ2.95 (br, 2H, CH2ꢀAr). 13C NMR (CDCl3): 151.5,
150.8, 138.2, 129.0, 128.5, 127.4, 127.0, 126.5, 123.5, 110.5, 109.1, 70.0,
56.1, 35.7. IR (NaCl, cmꢀ1): 2956, 2924, 2850, 1591, 1500, 1462, 1412,
1351, 1261, 1206, 1040, 965, 867. UVꢀvis (thin film): λmax = 506 nm.
SEC (THF): Mw = 123 ꢁ 103 g/mol (PD = 2.5).
1462, 1412, 1380, 1315, 1261, 1206, 1040, 867, 736, 704, 681.
Poly[(2-methoxy-5-(20-phenylethoxy))-1,4-phenyleneviny-
lene] (MPE-PPV) (3). In a typical Gilch procedure, 2,5-bis(chloromethyl)-
1-methoxy-4-(20-phenylethoxy)benzene (2) (0.5 g, 1.5 mmol) was dissolved
in dry 1,4-dioxane (152 mL), giving a concentration of 0.01 M. The mixture
was stirred at 25 ꢀC under a continuous flow of nitrogen. A KtBuO solution
(1.5 equiv, 0.87 M in 1,4-dioxane) was added dropwise over a time period of
15 min to the stirred monomer solution. The reaction was kept for 2 h at
25 ꢀC under a nitrogen atmosphere, and the mixture was subsequently
quenched in ice water (100 mL). The excess of base was neutralized with HCl
(1 M in H2O). The aqueous phase was extracted with CHCl3 (3 ꢁ 40 mL).
After combination of the organic phases, the solvent was evaporated and the
8472
dx.doi.org/10.1021/ma201911a |Macromolecules 2011, 44, 8470–8478

Macromolecules
ARTICLE
’ RESULTS AND DISCUSSION
Synthesis and Standard Characterization. PPV derivatives
are usually prepared via precursor routes, of which the Gilch
procedure (dehydrohalogenation polymerization) is the most
straightforward method.27 The desired MPE-PPV polymer was
prepared through a convenient three-step sequence (Scheme 1).
First, 1-methoxy-4-(20-phenylethoxy)benzene (1) was synthe-
sized using a Williamson ether reaction combining 4-methoxy-
phenol and 2-bromoethylbenzene. In a second step, the MPE
Gilch precursor monomer, bischloromethyl derivative 2, was
prepared via chloromethylation with p-formaldehyde and HCl
in acetic anhydride. Premonomer 2 was efficiently purified via
recrystallization in chloroform.
Polymerization was performed under a nitrogen atmosphere.
To this end, bis(chloromethyl) premonomer 2 was dissolved in
dry 1,4-dioxane (0.01 M), after which 1.5 equiv of KtBuO base
was added. After work-up, the orange sticky precursor polymer
was isolated via precipitation in cold methanol. The precursor
polymer was already partially conjugated at this stage due to the
small excess of base applied during polymerization. UVꢀvis
spectroscopy of the precursor polymer revealed an absorption
maximum of 468 nm (Figure 1a). The material was converted
further into a fully conjugated polymer via thermal elimination at
100 ꢀC in 1,4-dioxane. Upon heating, the chlorine groups were
eliminated to form fully conjugated MPE-PPV 3. After work-up
of the elimination reaction, the final polymer was isolated via
precipitation in cold methanol. Molecular weights were deter-
mined by size exclusion chromatography (SEC). The obtained
MPE-PPV had a Mw of 123 ꢁ 103 g/mol and a PD of 2.5. In the
UVꢀvis spectrum, the absorption band at 468 nm shifted to
506 nm and increased in absorptivity, indicating that a fully
conjugated system had developed. FT-IR spectroscopy revealed
the characteristic vibration frequencies of MPE-PPV (Figure 1b).
The signal at 965 cmꢀ1 originates from the trans-vinylene double
bonds of the conjugated polymer and is applied as a standard
criterion for the development of the conjugated PPV system.
Thermal Analysis. The thermal properties and phase behavior
of MPE-PPV:[60]PCBM blends were investigated by means of
MTDSC and RHC, which is a useful tool to investigate the glass
transition of semicrystalline samples due to its high scan rate (up
to 2000 K minꢀ1). Both techniques were used to investigate
crystallization, melting, and glass transition of the PPV:fullerene
blends, with an emphasis on the different component miscibil-
ities and possible phase separation. Differences in extent of phase
separation can account for possible deviating performances of the
corresponding polymer solar cells made from these blends.
In general, the initial morphology of deposited blend films is
the result of a kinetically frozen phase separation or crystal-
lization during solvent evaporation. Consequently, both thermo-
dynamic and kinetic parameters are responsible for the
morphology obtained.28ꢀ30 The phase behavior of such blends
with a well-defined thermal history is important to understand
and (ultimately) control morphology development, long-term
stability of the film morphology, and the photovoltaic perfor-
Figure 2. MTDSC thermograms showing apparent specific heat
capacity (cpapp) (a) and RHC thermograms showing heat flow (HF)
(b) during heating after quenching for MPE-PPV:[60]PCBM blends with
various fw[60]PCBM. All curves are shifted vertically for clarity.
Therefore, both the crystallization and melting in the blends can
be attributed to the fullerene component rather than the MPE-
PPV polymer. All samples werequenched atabout 100 K minꢀ1 to
reduce the crystallinity and enhance the visibility of the glass
transition. Although the crystallinity only decreased slightly due to
the high crystallization rate of [60]PCBM, clear glass transitions
could be seen for all the blends and pure components.
To study if phase separation occurred in MPE-PPV:[60]PCBM
blends, RHC was used to study the glass transitions.34,35 Figure 2b
shows the RHC heat flow (HF) curves for the blends. It can clearly
be seen that RHC can completely avoid crystallization of the
PPV:[60]PCBM blends due to its considerably enhanced cooling
rate, reaching a maximum of about 2000 K minꢀ1, compared to
regular DSC (about 100 K minꢀ1 without the use of liquid
nitrogen). No melting peaks can be distinguished for the
MPE-PPV:[60]PCBM blends, not even with high weight fractions
of [60]PCBM (fw[60]PCBM), except for pure PCBM, where the
melting of crystals formed during heating is observed.
mance of the corresponding solar cells.31ꢀ33
app
Figure 2a shows MTDSC apparent specific heat capacity (cp
)
curves of MPE-PPV:[60]PCBM blends over the full composition
range. From these curves, the crystallization and melting behavior
of these blends can be seen. The pure MPE-PPV sample is
amorphous; neither crystallization nor melting can be seen. Pure
[60]PCBM samples are semicrystalline, as reported before.33
Figure 3 summarizes the characteristic temperatures, melt crystal-
lization temperature (Tmc), cold crystallization temperature (Tcc),
melting temperature (Tm), and glass transition temperature (Tg),
as a function of fw[60]PCBM. It can be observed that with decreas-
[60]PCBM
ing fw
the crystallization of [60]PCBM slows down
8473
dx.doi.org/10.1021/ma201911a |Macromolecules 2011, 44, 8470–8478

Macromolecules
ARTICLE
(Tmc increases and Tcc decreases slightly) and less [60]PCBM
crystallizes (the melting peak area decreases). Below 70 wt % of
[60]PCBM, the samples do not show crystallization or melting
at the cooling and heating rates employed. With decreasing
fw[60]PCBM, the Tg decreases gradually from the Tg of pure
[60]PCBM (about 131.2 ꢀC) to the nearly as high Tg of neat
MPE-PPV (about 111.2 ꢀC). All blends show a single Tg in
between of those of the pure components.
In previous studies, it was demonstrated that MDMO-PPV:-
[60]PCBM blends show a clear double Tg, which is an obvious
indication of phase separation in the molten state.19,29,36 For
[60]PCBM contents from 70% to 90% double glass transitions
were noted, with temperatures corresponding to phases with
estimated compositions of 55 ( 5% and 97 ( 2%, respectively.
For the MPE-PPV:[60]PCBM blends, it was also supposed that
phase separation might occur in the molten state due to the
similar molecular structures. However, single Tg’s were observed
in our MTDSC and RHC experiments, which would indicate the
absence of phase separation in the molten state. Nevertheless,
due to the very close Tg’s of MPE-PPV and [60]PCBM, the gap
between them is only ∼20 ꢀC; it would be difficult to see two
glass transition steps of partially phase-separated phases, espe-
cially since one Tg transition can be as wide as 50 ꢀC. As shown
later in this paper, TEM images show the absence of a nanoscale
structure in the as-cast MPE-PPV:[60]PCBM blends, in contrast
to the nanoscale structure of as cast MDMO-PPV:[60]PCBM.
Device Characterization. Nonoptimized bulk heterojunction
solar cells, combining MDMO-PPV or MPE-PPV with [60]PCBM
(1:4 ratio), were made with a general ITO/PEDOTꢀPSS/
polymer:[60]PCBM/Ca/Al structure. Table 1 shows the initial
values of short circuit current, fill factor, open circuit voltage, and the
efficiency of the solar cells. The initial IV curves are depicted in
Figure 4. The initial efficiencies were in the range of 1.3% for both
PPV polymers. The MPE-PPV devices show similar Voc and
substantial higher Isc than those based on MDMO-PPV, indicating
potential for improved performance. These first devices were,
however, not optimized for maximum performance.
When these solar cells are to be used outdoors in full sunlight,
the devices must be able to withstand the temperatures as high as
75 ꢀC reached in such case. Previous studies already demonstrated
that the poor thermal stability of MDMO-PPV:[60]PCBM solar
cells resulted in a gradual decrease of the efficiency upon heat-
ing.22,23 The use of conjugated polymers with a Tg substantially
above 75 ꢀC as a donor material should render the amorphous
polymer:fullerene phase glassy at this maximum operation tem-
perature, drastically slowing down diffusion rates and basically
preventing the further growth of the [60]PCBM crystals, thus
overall resulting in a better thermal stability of the device. Hence, it
is expected that, although the initial efficiency values are about
the same for both devices, the MPE-PPV:[60]PCBM solar cells
will have a better resistance toward high temperatures than
MDMO-PPV:[60]PCBM-based solar cells.
To verify this hypothesis, both devices were annealed at two
different temperatures, either 90 or 110 ꢀC, in a dark nitrogen
atmosphere for 100 h. In addition, the thermally more stable
MPE-PPV:[60]PCBM solar cells were also annealed at an even
higher temperature of 130 ꢀC. These temperatures are all well-
above the Tg of the MDMO-PPV:[60]PCBM (1:4) blend, while
being below, near, and above the Tg of the MPE-PPV:[60]PCBM
(1:4) blend. During the annealing treatment, the IV character-
istics were monitored at specific time intervals. Figure 5 shows
the relative decay of the photovoltaic parameters of both solar
cells for the different annealing temperatures. In most cases the
open circuit voltage (Voc) is barely sensitive to the thermal
treatment performed over a long period of time. Thermal
annealing of the solar cells at all temperatures resulted in small
decreases in Voc of about 10ꢀ13% (Figure 5b and Table 1). This
is consistent with the fact that the Voc mainly depends on material
properties, i.e., the ionization potential of the donor polymer and
the electron affinity of the acceptor fullerene molecule. After 100 h
the short circuit current of the MPE-PPV solar cells showed a
decrease of 18% when annealed at 90 or 110 ꢀC and a decrease of
29% when annealed at 130 ꢀC. The MDMO-PPV-based solar
cells suffered more substantially from the annealing process.
Figure 3. Dependence of the characteristic temperatures Tmc, Tcc, Tm,
[60]PCBM
and Tg, measured by (MT)DSC, on the fw
for MPE-PPV:-
[60]PCBM blends.
Table 1. Values of Short Circuit Current (Isc), Open Circuit Voltage (Voc), Fill Factor (FF), and Power Conversion Efficiency (η)
for MPE-PPV:[60]PCBM and MDMO-PPV:[60]PCBM (1:4) Solar Cells, Spin-Coated from Chlorobenzene (Initial Values and
Values after Thermal Annealing)
polymer:[60]PCBM (1:4)
MPE-PPV:[60]PCBM
annealing
Isc (mA/cm2)
Voc (V)
FF
η (%)
0 h (25 ꢀC)
4.14
3.38
3.39
2.92
2.87
1.70
1.45
0.783
0.692
0.733
0.713
0.785
0.710
0.685
0.40
0.39
0.35
0.37
0.60
0.43
0.42
1.30
0.91
0.88
0.76
1.30
0.52
0.40
100 h at 90 ꢀC
100 h at 110 ꢀC
100 h at 130 ꢀC
0 h (25 ꢀC)
100 h at 90 ꢀC
100 h at 110 ꢀC
MDMO-PPV:[60]PCBM
8474
dx.doi.org/10.1021/ma201911a |Macromolecules 2011, 44, 8470–8478

Macromolecules
ARTICLE
Annealing at 90 ꢀC resulted in a decrease of 41% of the initial
short circuit current while annealing at 110 ꢀC resulted in a
decrease of 49%. After 100 h of annealing, the fill factor of the
MPE-PPV solar cells showed small decreases between 2 and 12%.
The fill factor of MDMO-PPV solar cells decreased 28ꢀ30%.
Finally, after 100 h, the overall efficiency, being proportional to
the short-circuit current and the fill factor, is much more stable
for MPE-PPV-based solar cells (a decrease of 30% after annealing at
90 ꢀC, a decrease of 32% after annealing at 110 ꢀC, and a decrease of
42% after annealing at 130 ꢀC) as compared to MDMO-PPV-based
solar cells (a decrease of 61% after annealing at 90 ꢀC and a decrease
of 69% after annealing at 110 ꢀC). These results are consistent with
earlier reported observations on a commercial high-Tg PPV
terpolymer.22,23 The importance of the relative position of the glass
transition of the conjugated polymer with respect to the temperature
at which the performance stability is evaluated is clearly seen.
The higher thermal stability of the MPE-PPV solar cells is
revealed not only in the photovoltaic parameters but also in the
bulk morphology of the active layer. The blend morphology
of both types of solar cells was studied with TEM (Figure 6).
For the MDMO-PPV:[60]PCBM blend, earlier TEM studies
showed pronounced [60]PCBM cluster formation upon anneal-
ing at 110 ꢀC.22,23 After 2 h of annealing at 110 ꢀC, all the
[60]PCBM had diffused out of the blend. Figure 6 shows large
[60]PCBM clusters (dark areas) formed very rapidly upon
annealing at 130 ꢀC.19 The selected area electron diffraction
(SAED) patterns (insets in Figure 6) indicated that the clusters
were groups of single [60]PCBM crystals. After only 0.5 h of
annealing at 130 ꢀC, the homogeneous MDMO-PPV:[60]PCBM
matrix disappeared. Most of the[60]PCBMhaddiffused out of the
blend and assembled in the [60]PCBM clusters. The same process
occurred upon annealing at 90 ꢀC, only slower (see Supporting
Information). After 20h of annealing at90 ꢀC, [60]PCBMcrystals
wereformed. Because of the smaller interfacialareabetween donor
Figure 4. Currentꢀvoltage characteristics obtained for MPE-PPV:-
[60]PCBM (gray circles) and MDMO-PPV:[60]PCBM (black squares)
(1:4) solar cells under AM 1.5 illumination.
Figure 5. Relative decay of short circuit current Isc (a), open circuit voltage Voc (b), fill factor (c), and power conversion efficiency (d) for MDMO-
PPV:[60]PCBM (1:4) solar cells (gray) and MPE-PPV:[60]PCBM (1:4) solar cells (black), during annealing at different temperatures.
8475
dx.doi.org/10.1021/ma201911a |Macromolecules 2011, 44, 8470–8478

Macromolecules
ARTICLE
Figure 6. Bright-Field TEM images of the active layer of MDMO-PPV:[60]PCBM (1:4) solar cells (aꢀd) and MPE-PPV:[60]PCBM (1:4) solar cells
(a0ꢀd0), spin-coated from chlorobenzene, after annealing at 130 ꢀC for 0 h (a, a0), 0.5 h (b, b0), 4 h (c, c0), and 20 h (d, d0) (scale bar: 2 μm). Insets: SAED
patterns of matrix (a, a0) and [60]PCBM clusters (b, d0).
Table 2. Values of Short Circuit Current (Isc), Open Circuit Voltage (Voc), Fill Factor (FF), and Power Conversion Efficiency (η)
for Different MPE-PPV:Fullerene (1:4) Solar Cells and a Reference MDMO-PPV:[70]PCBM (1:4) Solar Cell, Spin-Coated from
Different Solvents, at Room Temperaturea
polymer:fullerene (1:4)
processing solvent
Isc (mA/cm2)
Voc (V)
FF
η (%)
MPE-PPV:[60]PCBM
MPE-PPV:[60]PCBM
MPE-PPV:[60]PCBM
MPE-PPV:[70]PCBM
MPE-PPV:[70]PCBM
MPE-PPV:[70]PCBM
MDMO-PPV:[70]PCBM
chloroform
3.31
2.87
3.14
5.66
5.36
4.87
6.52
0.84
0.83
0.78
0.85
0.85
0.80
0.80
0.45
0.42
0.42
0.48
0.48
0.47
0.57
1.25
1.00
1.03
2.31
2.19
1.83
2.97
chlorobenzene
toluene
chloroform
chlorobenzene
toluene
1,2-dichlorobenzene
a The currents reported are corrected for spectral mismatch.
and acceptor material, the photocurrent output decreased, as was
demonstrated above. The TEM images of the active layer of
the MPE-PPV solar cells (Figure 6a0ꢀd0) showed a much more
stable morphology. The film had to be annealed for 20 h at the
highest temperature of 130 ꢀC before a [60]PCBM cluster
was detected. Because of the considerably lower Tg (45 ꢀC) of
MDMO-PPV, the [60]PCBM is able to crystallize at a higher rate
in MDMO-PPV:[60]PCBM active layers than in high-Tg MPE-PPV:-
[60]PCBM active layers. When the solar cell is subjected to
temperatures below the Tg of the applied polymer, the matrix is in
a glassy state and the fullerene molecules are restricted to move
freely. Thermal treatments near or above Tg cause the matrix to
devitrify and rendering it much easier for the fullerene molecules
to move and cluster.
As can be observed from the TEM images, the MPE-PPV:-
[60]PCBM solar cells do not exhibit an initial fine nanoscale
phase separation, as compared to MDMO-PPV:[60]PCBM
solar cells, in agreement with the single Tg observed for the
MPE-PPV:[60]PCBM. This correlates also with the quite low fill
factor of the obtained MPE-PPV based devices (FF = 0.40) com-
pared to MDMO-PPV based solar cells (FF = 0.60). To improve
this initial morphology, an optimization study was performed
whereby the active layers were spin-coated from either chloroform,
chlorobenzene, or toluene. Because of the varying solubility char-
acteristics of both the polymer and fullerene materials in the different
Figure 7. Currentꢀvoltage characteristics obtained for differentMPE-PPV:
fullerene (1:4) solar cells and a reference MDMO-PPV:[70]PCBM (1:4)
solar cell under AM 1.5 illumination (not corrected for spectral mismatch).
solvents, the initial morphology of the active layer will be influenced
to a large extent. In addition, not only [60]PCBM was used as an
electron acceptor but also the [70]PCBM derivative, which in some
solvents can crystallize more easily than [60]PCBM.37 The results
are shown in Table 2 and Figure 7.
8476
dx.doi.org/10.1021/ma201911a |Macromolecules 2011, 44, 8470–8478

Macromolecules
ARTICLE
From the IV data, it can be derived that the use of [70]PCBM
as the acceptor resulted in a substantial improvement of the
device performance, mainly due to an improved light absorption
of [70]PCBM, compared to [60]PCBM.38 Consequently, higher
current densities in the corresponding photovoltaic cells were
observed. This was already demonstrated previously for
MDMO-PPV:[70]PCBM solar cells, for which 1,2-dichloroben-
zene proved to be the best processing solvent.37 The IV
characteristics have been added to Table 2 for comparison.
Furthermore, the fill factor of MPE-PPV:[70]PCBM solar cells
could be increased to 0.48, indicating an improved initial nano-
scale phase-separated morphology. Chloroform was found to be
the best processing solvent for MPE-PPV:fullerene solar cells, and
the most efficient device displayed an efficiency of 2.3%, after
correction for spectral mismatch.8 The investigation of the
thermal stability of this optimized MPE-PPV:[70]PCBM solar
cell and the comparison with the results obtained for the initial
MPE-PPV:[60]PCBM devices will be the subject of a follow-
up study.
Science Policy (BELSPO) in the frame of network IAP P6/27,
initiated by the Belgian State Prime Minister’s Office, for their
financial support. We also thank the Institute for the Promotion
of Innovation by Science and Technology in Flanders (IWT) for
the financial support via the IWT-SBO project 060843 “Poly-
spec”. TA Instruments is acknowledged for support through
“Project RHC”. We also thank F. Piersimoni and J. D’Haen for
help with the experimental setup.
’ REFERENCES
(1) Brabec, C. J.; Gowrisanker, S.; Halls, J. J. M.; Laird, D.; Jia, S.;
Williams, S. P. Adv. Mater. 2010, 22, 3839–3856.
(2) Moulꢀe, A. J.; Meerholz, K. Adv. Funct. Mater. 2009, 19, 3028–3036.
(3) Shaheen, S. E.; Brabec, C. J.; Sariciftci, N. S.; Padinger, F.;
Fromherz, T.; Hummelen, J. C. Appl. Phys. Lett. 2001, 78, 841–843.
(4) Munters, T.; Martens, T.; Goris, L.; Vrindts, V.; Manca, J.;
Lutsen, L.; De Ceuninck, W.; Vanderzande, D.; De Schepper, L.; Gelan, J.;
Sariciftci, N. S.; Brabec, C. J. Thin Solid Films 2002, 403ꢀ404, 247–251.
(5) Ma, W.; Yang, C.; Gong, X.; Lee, K.; Heeger, A. J. Adv. Funct.
Mater. 2005, 15, 1617–1622.
(6) Padinger, F.; Fromherz, T.; Denk, P.; Brabec, C. J.; Zettner, J.;
Hierl, T.; Sariciftci, N. S. Synth. Met. 2001, 121, 1605–1606.
(7) During final redaction of this manuscript, a special issue
appeared on the 3th International Summit on OPV Stability: Krebs,
F. C., Ed. Sol. Energy Mater. Sol. Cells 2011, 95, 1251ꢀ1418.
(8) Kroon, J. M.; Wienk, M. M.; Verhees, W. J. H.; Hummelen, J. C.
Thin Solid Films 2002, 403ꢀ404, 223–228.
’ CONCLUSIONS
A novel PPV polymer, poly[2-methoxy-5-(20-phenylethoxy)-
1,4-phenylenevinylene] (MPE-PPV), with a Tg of 111 ꢀC, has
been synthesized by an efficient three-step Gilch protocol and has
been explored as donor material in bulk heterojunction solar
cells, demonstrating at the same time the engineering of Tg by
stiffening the side-chain structure. The thermal stabilities of the
nanomorphology of MDMO-PPV:fullerene and MPE-PPV:full-
erene blends were compared, and the solar cell efficiencies of
devices made from these blends were analyzed. Using [70]PCBM
instead of [60]PCBM as acceptor resulted in a 70% increase in
current density and device efficiency, with an optimized MPE-
PPV:[70]PCBM device efficiency of 2.3%. It can be concluded
that devices based on high-Tg MPE-PPV clearly outperform the
MDMO-PPV:[60]PCBM-based solar cells in terms of a (thermally)
more stable bulk (nano)morphology and a longer lifetime. For
the MPE-PPV:[60]PCBM active layer, the high-Tg polymer
matrix reduces the diffusion of the fullerene molecules, thereby
slowing down the detrimental clustering process and maintaining
a large interfacial area between donor and acceptor. It has hence
been shown that the use of a high-Tg polymer is an effective way
toward the development of more stable and efficient organic
photovoltaic devices.
(9) Krebs, F. C.; Carlꢀe, J. E.; Cruys-Bagger, N.; Andersen, M.;
Lilliedal, M. R.; Hammond, M. A.; Hvidt, S. Sol. Energy Mater. Sol. Cells
2005, 86, 499–516.
(10) Schuller, S.; Schilinsky, P.; Hauch, J.; Brabec, C. J. Appl. Phys. A:
Mater. Sci. Process. 2004, 79, 37–40.
(11) Neugebauer, H.; Brabec, C.; Hummelen, J. C.; Sariciftci, N. S.
Sol. Energy Mater. Sol. Cells 2000, 61, 35–42.
(12) Camaioni, N.; Ridolfi, G.; Casalbore-Miceli, G.; Possamai, G.;
Garlaschelli, L.; Maggini, M. Sol. Energy Mater. Sol. Cells 2003, 76,
107–113.
(13) Hoppe, H.; Sariciftci, N. S. J. Mater. Res. 2004, 19, 1924–1945.
(14) Neugebauer, H.; Brabec, C. J.; Hummelen, J. C.; Janssen,
R. A. J.; Sariciftci, N. S. Synth. Met. 1999, 102, 1002–1003.
(15) Krebs, F. C.; Spanggaard, H. Chem. Mater. 2005, 17, 5235–5237.
(16) Krebs, F. C.; Norrman, K. Prog. Photovoltaics 2007, 15, 697–712.
(17) Conings, B.; Bertho, S.; Vandewal, K.; Senes, A.; D’Haen, J.;
Manca, J.; Janssen, R. A. J. Appl. Phys. Lett. 2010, 96, 163301.
(18) Yang, X.; van Duren, J. K. J.; Rispens, M. R.; Hummelen, J. C.;
Janssen, R. A. J.; Michels, M. A. J.; Loos, J. Adv. Mater. 2004, 16, 802–806.
(19) Yang, X.; van Duren, J. K. J.; Janssen, R. A. J.; Michels, M. A. J.;
Loos, J. Macromolecules 2004, 37, 2151–2158.
’ ASSOCIATED CONTENT
(20) Yang, X.; Loos, J.; Veenstra, S. C.; Verhees, W. J. H.; Wienk, M.;
Kroon, J. M.; Michels, M. A. J.; Janssen, R. A. J. Nano Lett. 2005, 5,
579–583.
(21) Campo, B.; Oosterbaan, W.; Gilot, J.; Cleij, T.; Lutsen, L.;
Janssen, R. A. J.; Vanderzande, D. Proc. SPIE 2009, 7416, 74161G.
(22) Bertho, S.; Haeldermans, I.; Swinnen, A.; Moons, W.; Martens,
T.; Lutsen, L.; Vanderzande, D.; Manca, J.; Senes, A.; Bonfiglio, A. Sol.
Energy Mater. Sol. Cells 2007, 91, 385–389.
(23) Bertho, S.; Janssen, G.; Cleij, T. J.; Conings, B.; Moons, W.;
Gadisa, A.; D’Haen, J.; Goovaerts, E.; Lutsen, L.; Manca, J.; Vanderzande,
D. Sol. Energy Mater. Sol. Cells 2008, 92, 753–760.
(24) Deimede, V.; Kallitsis, J. K.; Pakula, T. J. Polym. Sci., Part A:
Polym. Chem. 2001, 39, 3168–3179.
S
Supporting Information. Bright-field TEM images of the
b
active layer of MDMO-PPV:[60]PCBM 1:4 solar cells and MPE-
PPV:[60]PCBM 1:4 solar cells, spin-coated from chlorobenzene,
after annealing at 90 ꢀC for 0, 0.5, 4, and 20 h. This material is
available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*Phone: +32 (0)11 26 83 21. Fax: +32 (0)11 26 83 01. E-mail:
dirk.vanderzande@uhasselt.be.
(25) Strukelj, M.; Papadimitrakopoulos, F.; Miller, T. M.; Rothberg,
L. J. Science 1995, 267, 1969–1972.
(26) Johansson, D. M.; Srdanov, G.; Yu, G.; Theander, M.; Inganas,
O.; Andersson, R. M. Macromolecules 2000, 33, 2525–2529.
(27) Gilch, H. G.; Wheelwright, W. L. J. Polym. Sci. 1966, 4, 1337–1349.
’ ACKNOWLEDGMENT
The authors gratefully acknowledge the Fund for Scientific
Research-Flanders (FWO) (PhD grants for J.V. and S.B., the
project G.0091.07, postdoctoral grant for W.M.) and the Belgian
8477
dx.doi.org/10.1021/ma201911a |Macromolecules 2011, 44, 8470–8478

Macromolecules
ARTICLE
(28) Thompson, B. C.; Frꢀechet, J. M. J. Angew. Chem., Int. Ed. 2008,
47, 58–77.
(29) Yang, X.; Loos, J. Macromolecules 2007, 40, 1353–1362.
(30) Moulꢀe, A. J.; Meerholz, K. Adv. Mater. 2008, 20, 240–245.
(31) Mueller, C.; Ferenczi, T. A. M.; Campoy-Quiles, M.; Frost,
J. M.; Bradley, D. D. C.; Smith, P.; Stingelin-Stutzmann, N.; Nelson, J.
Adv. Mater. 2008, 20, 3510–3515.
(32) Kim, J. Y.; Frisbie, C. D. J. Phys. Chem. C 2008, 112, 17726–17736.
(33) Zhao, J.; Swinnen, A.; Van Assche, G.; Manca, J.; Vanderzande,
D.; Van Mele, B. J. Phys. Chem. B 2009, 113, 1587–1591.
(34) Danley, R. L.; Caulfield, P. A.; Aubuchon, S. R. Am. Lab. 2008,
40, 9–11.
(35) Miltner, H. E.; Grossiord, N.; Lu, K.; Loos, J.; Koning, C. E.;
Van Mele, B. Macromolecules 2008, 41, 5753–5762.
(36) Zhao, J.; Bertho, S.; Vandenbergh, J.; Van Assche, G.; Manca, J.;
Vanderzande, D.; Cleij, T.; Lutsen, L.; Van Mele, B. Phys. Chem. Chem.
Phys. 2011, 13, 12285–12292.
(37) Wienk, M. M.; Kroon, J. M.; Verhees, W. J. H.; Knol, J.;
Hummelen, J. C.; van Hal, P. A.; Janssen, R. A. J. Angew. Chem., Int.
Ed. 2003, 42, 3371–3375.
(38) Arbogast, J. W.; Foote, C. S. J. Am. Chem. Soc. 1991, 113,
8886–8889.
8478
dx.doi.org/10.1021/ma201911a |Macromolecules 2011, 44, 8470–8478