Beyond PCBM: Methoxylated 1,4-bisbenzyl[60]fullerene adducts for efficient organic solar cells

Article abstract of DOI:10.1039/c5ta07688a

Organic solar cells have been based mostly on conjugated polymers and the classic fullerene derivative PCBM and are characterized by modest open circuit voltages (V_oc). Increasing V_oc requires fullerene acceptors with higher LUMOs th

Full text of DOI:10.1039/c5ta07688a

View Article Online
View Journal
Journal of
Materials Chemistry A
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: S. Huang, G.
Zhang, N. S. Knutson, M. T. Fontana, R. C. Huber, A. S. Ferreira, S. Tolbert, B. J. Schwartz and Y. Rubin, J.
Mater. Chem. A, 2015, DOI: 10.1039/C5TA07688A.
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the Ethical guidelines still
apply. In no event shall the Royal Society of Chemistry be held
responsible for any errors or omissions in this Accepted Manuscript
or any consequences arising from the use of any information it
contains.
www.rsc.org/materialsA

Please do not adjust margins
Journal of Materials Chemistry A
Page 1 of 8
View Article Online
DOI: 10.1039/C5TA07688A
Journal of Materials Chemistry A
ARTICLE
Beyond PCBM: Methoxylated 1,4-Bisbenzyl [60]Fullerene Adducts
for Efficient Organic Solar Cells
Received 00th January 20xx,
Accepted 00th January 20xx
Shaohua Huang,
‡
a Guangye Zhang,
‡
a Nicholas S. Knutson,a Matthew T. Fontana,a Rachel C.
Huber,a Amy S. Ferreira,a Sarah H. Tolbert,
*
ab Benjamin J. Schwartz
*
a and Yves Rubin
*
a
DOI: 10.1039/x0xx00000x
Organic solar cells have been based mostly on conjugated polymers and the classic fullerene derivative PCBM and are
characterized by modest open circuit voltages (Voc). Increasing Voc requires fullerene acceptors with higher LUMOs than
PCBM. To date, most fullerene derivatives synthesized for this purpose either do not achieve the high photocurrent
afforded by PCBM or show relatively poor compatibility with the next-generation low bandgap conjugated polymers used
in high-efficiency organic solar cells. Here, we report the facile synthesis of methoxylated 1,4-bisbenzyl fullerene adducts
and their application as efficient electron acceptors in conjugated polymer-based solar cells. The methoxy groups are
found to be essential to increasing the LUMO levels, and accordingly the Voc, of the devices compared to the parent 1,4-
www.rsc.org/
bisbenzyl fullerene, and more importantly, to PCBM. The best fullerene 1,4-bisadduct provides a ~20% enhancement in
power conversion efficiency over PCBM when used with the classic crystalline polymer P3HT. When used in combination
with a higher-performance low bandgap polymer, PTB7, the bisadduct both increases the device open-circuit voltage and
maintains the high photocurrent provided by the more traditional PCBM. We also examine 10 different 1,4-fullerene
bisadducts and show that the photovoltaic device performance is strongly influenced by the number and relative position
of the methoxy substituents on the benzyl addends: moving a single methoxy substituent by one position on the benzyl
rings can change the device efficiency by over a factor of 2.
and/or the use of new device architectures.10,11 In contrast,
progress on the design and synthesis of novel fullerene
acceptors for highꢀefficiency organic photovoltaics has been
much less rapid. Most of the highest performing devices9,12 still
utilize the classic fullerene derivative [6,6]ꢀphenylꢀCꢀ61ꢀbutyric
acid methyl ester (PCBM), synthesized more than twenty years
ago,13 or its expensive C70 analogue, PC71BM.14
Introduction
Organic solar cells have received great attention in recent years
as a potential alternative to silicon solar cells because of their
ability to be inexpensively solutionꢀprocessed and because they
can be lightweight and flexible.1–5 The key component of an
organicꢀbased photovoltaic system is its active layer, wherein a
The PCE of a solar cell is proportional to the product of the
shortꢀcircuit current (Jsc), openꢀcircuit voltage (Voc) and fill
factor (FF). Thus, one strategy for improving device efficiency
is to increase the Voc, which is directly related to the difference
between the energy of the highest occupied molecular orbital
(HOMO) of the polymer donor and that of the lowest
unoccupied molecular orbital (LUMO) of the fullerene
acceptor. As long as the LUMO of the fullerene is lower than
that of the polymer by an amount sufficient to promote charge
separation, raising the fullerene LUMO level should increase
the Voc and thus the PCE.
p
ꢀtype conjugated polymer and an nꢀtype acceptor material mix
to form a bicontinuous interpenetrating network that is referred
to as a bulk heterojunction (BHJ).6,7 Fullerene derivatives have
been used extensively as the acceptors in BHJ solar cells thanks
to their high electron affinities and electron mobilities.8 The
power conversion efficiency (PCE) of polymer/fullerene BHJ
solar cells can be as high as 10.8%,9 with most of the recent
advances coming from the design of new polymer donors
To this end, several research groups have synthesized
fullerene derivatives where two or more double bonds of the
fullerene cage are saturated.15–24 Although Voc has been
demonstrated to increase with this method, devices based on
most of these new fullerenes fail to maintain high Jsc and/or FF
,
and as a result, the overall device efficiency suffers.19,25–27 This
is because altering the chemical structure of a fullerene addend
can have detrimental effects on device performance for two
reasons: first, the size and pattern of the addends can
This journal is © The Royal Society of Chemistry 20xx
J. Mater. Chem. A, 201X, 00, 1-3 | 1
Please do not adjust margins

Please do not adjust margins
Journal of Materials Chemistry A
Page 2 of 8
View Article Online
DOI: 10.1039/C5TA07688A
ARTICLE
Journal of Materials Chemistry A
dramatically influence the electronic coupling between adjacent that the relative position and number of methoxy groups can
fullerenes through steric and/or electronic effects, causing a have a dramatic effect on the performance of these fullerene
significant decrease in local electron mobility;27 and second, derivatives in polymerꢀbased solar cells.
changes in the fullerene addends can alter the degree of phase
Table 1 Synthesis of symmetric and asymmetric 1,4ꢀbisbenzyl fullerene C60 adducts 1a–
separation from the polymer, sometimes caused by a reduction
k.
in fullerene crystallinity due to packing constraints or mixtures
of isomers. In cases where fullerene crystallization drives
phase separation, this can dramatically change the morphology
of the bulk heterojunction.27
There are a few select fullerene derivatives, such as the
bisadduct of PCBM (bisꢀPCBM)28, indene C60 bisadduct
(ICBA)15 and its C70 congeners,29 dihydronaphthylꢀbased C60
bisadduct (NCBA)30 and di(4ꢀmethylꢀphenyl)methanoꢀC60
bisadduct (DMPCBA),31 that have shown good photovoltaic
performance when used in combination with the classic
crystalline semiconducting polymer poly(3ꢀhexylthiopheneꢀ2,5ꢀ
diyl) (P3HT). However, when these fullerene derivatives are
used in combination with less crystalline stateꢀofꢀtheꢀart lowꢀ
bandgap polymers, the performance of the solar cells is
typically low, with greatly decreased Jsc and FF 25,26,32–34
Therefore, despite the aforementioned efforts to increase the
oc, PCBM (and its C70 analogue) are still the most successful
.
V
fullerene acceptors for highꢀperformance polymerꢀbased
photovoltaics to date.
In order to design new fullerene acceptors for high
performance polymerꢀbased solar cells, the following factors
need to be considered: 1) The LUMO level of the fullerene
derivative should be carefully tuned so that when paired with
the polymer of choice, an ideal energy level offset between the
fullerene donor and polymer acceptor is attained. Although still
under debate,34,35 the generally accepted range of this offset is
about 0.3 eV, depending on the materials.26,36,37 2) Sizeꢀsuitable
addends are needed to assist close contacts between fullerene
balls, thereby facilitating favorable electronic coupling to
facilitate charge transport within the fullerene domains.27,38 3)
Derivatives must possess good solubility in organic solvents for
solution processing, and must form reasonable BHJ structures
with a variety of polymers. In this paper, we thus present the
synthesis of a series of new fullerene derivatives that satisfy all
of these requirements. We find that when carefully designed,
fullerenes with higher LUMOs can be prepared that produce
devices with higher
Voc’s – without significant loss of Jsc or FF
– and thus higher power conversion efficiencies.
Our new derivatives are methoxylated 1,4ꢀbisbenzyl
fullerene adducts (MeOꢀBBF, Table 1), i.e. two addends are
located at the “para” positions of a sixꢀmembered ring on the
fullerene cage.39 These derivatives have a smaller
πꢀconjugated
system with reduced symmetry40–43 resulting in a slightly higher
LUMO level than the corresponding 1,2ꢀfullerene bisadducts.
The LUMO level, side chain nature and solubility of the 1,4ꢀ
fullerene bisadducts can be further tuned by altering each
addend independently.39 The molecules we focus on bear
electronꢀdonating methoxy group(s) on the benzyl ring(s).
Since these electronꢀrich methoxy substituents are not
conjugated with the fullerene πꢀsystem, they would be expected
to have negligible electronic interaction. However, we find
through experiments that are supported by DFT calculations
We have investigated the performance of these fullerenes in
combination with both P3HT and the lowꢀbandgap polymer
PTB7.44 The best PCE based on P3HT:1ꢀ(2’,5’ꢀdimethoꢀ
xybenzyl)ꢀ4ꢀ(2”,6”ꢀdimethoxybenzyl)[60]fullerene bisadduct
(
1i) is 4.1%, which is a > 20% enhancement relative to devices
made from P3HT:PCBM.
Furthermore, devices based on PTB7:1e show a Voc of 0.83
V, a respectable FF (53%) and a Jsc (12.3 mA/cm2) that is
higher than that for devices based on PTB7:PCBM, resulting in
a PCE of 5.4%. Perhaps more importantly, our results clearly
2 | J. Mater. Chem. A, 201X, 00, 1-8
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Journal of Materials Chemistry A
Page 3 of 8
View Article Online
DOI: 10.1039/C5TA07688A
Journal of Materials Chemistry A
ARTICLE
show that the precise nature and degree of substitution of the All of the films were then thermally annealed at 150 ºC for 20
methoxy group(s), even on a single benzyl addend, greatly min on a hot plate under an argon atmosphere. The thickness
influences both the Voc and PCE of the BHJ photovoltaic of the polymer:fullerene layers were ~160
–180 nm, as
devices. Moreover, the higher conformational flexibility of the measured with a Dektak 150 Stylus Surface Profiler.
benzyl vs. an aryl substituent in these 1,4ꢀbisadducts appears to
play a significant role in obtaining high PCE values.39,43
PTB7:fullerene solutions were prepared by dissolving solid
PTB7 (Solarmer Energy Inc.) and solid fullerene derivatives in
a mixture of 95% chlorobenzene (CB)/5% 1,8ꢀdiiodooctane
(DIO) v/v with a polymer:fullerene weight ratio of 1:1.5 for the
fullerene bisadducts and 1:1.34 for PCBM (the change in
weight ratio accounts for molecular weight differences to
ensure that all polymer:fullerene blends were equimolar). The
concentration was 10 mg/mL with respect to PTB7. The
solutions were stirred at 55 ºC overnight prior to being spun at
1000 rpm for 60 s onto PEDOT:PSSꢀcovered substrates. The
films were then transferred to the antechamber of the glovebox
and held under vacuum for ~1 hr. Pure methanol was then spun
onto the films at a speed of 2500 rpm for 40 s to remove
residual DIO. No further treatments were performed after this
step before the deposition of the metal cathode. Cathode
deposition consisted of ~10 nm of Ca evaporated at a rate of
~0.5 Å/s followed by 70 nm of Al at rates below 1 Å/s. The
resulting device active areas were 7.2 mm2.
Experimental
Synthesis
Our ability to synthesize the MeOꢀBBF bisadducts is a direct
result of the ease of alkylating the C60 dianion.45–50 As shown in
2–
Table 1, C60 can be generated readily in dry degassed PhCN
when treated with hydroquinone and base (SI).51 By adding a
large excess of a substituted benzyl bromide to a dark red
solution of C602–, we produced the symmetric 1,4ꢀdibenzyl C60
bisadducts in relatively good yields (Table 1).
The synthesis of asymmetric 1,4ꢀdibenzyl C60 bisadducts
involves a stepwise alkylation procedure (SI). As shown in
2–
Table 1, C60 also can be reduced to C60 with
n
ꢀPrS–Cs+
generated in situ through the reaction of
nꢀpropanethiol and
Cs2CO3 in DMSO.52 C60 reacts with substituted benzyl
chlorides to provide monoadducts 2a,b or 2e in 50ꢀ60% yields.
The lower reactivity of benzyl chlorides relative to benzyl
bromides towards SN2 is likely at the origin of the different
outcomes of these two reaction conditions.49 The subsequent
benzylation of deprotonated monoadducts 2a,b or 2e with a
2–
Current density–applied voltage (J–V) measurements were
performed in an argon atmosphere using a Keithley 2400
source meter. A xenon arc lamp and 1ꢀsunꢀcalibrated AMꢀ1.5
filter were used as the excitation source.
External quantum efficiency (EQE) measurements were
taken using a homeꢀbuilt setꢀup that has been detailed in
previous publications by our group.53
benzyl bromide provides the asymmetric 1,4ꢀbisadducts 1j–k
,
which bear two different addends.
Films for nonꢀdevice measurements were prepared using
identical procedures to those described above but without
deposition of a top electrode.
The synthesis details of each fullerene derivative and their
NMR, massꢀspectrometry and cyclic voltammetry characterꢀ
istics can be found in the Supporting Information (SI).
Structural and Optical Characterization
Photovoltaic Device and Active Layer Fabrication Procedures
UVꢀVisible absorption spectra were collected using a Perkinꢀ
Elmer Lambda 25 UV/Vis Spectrophotometer.
We fabricated polymer:fullerene BHJ solar cells by starting
with prepatterned tinꢀdoped indium oxide (ITO)ꢀcoated
substrates (TFD Inc.) and cleaning them by successive
sonication in detergent solution, deionized water, acetone and
isopropanol for ~10 min each. After drying in vacuum for at
least an hour, we treated the ITO substrates with an air plasma
The 2ꢀD grazing incidence wide angle Xꢀray scattering
(GIWAXS) experiments were performed at the Stanford
Synchrotron Radiation Lightsource on beamline 11ꢀ3 using a
wavelength of 0.9742 Å. Figure 3c in the next section was
obtained by radiallyꢀintegrating the full 2ꢀD diffractograms.
Each data curve in Fig. 3c is the average of at least three
different samples prepared using the same conditions. The 2ꢀD
images were collected on a plate with the detector 400 mm
away from the center of the measured sample. The beam spot
had a width of ~150 ꢁm. A helium chamber was used to reduce
the noise. The software package WxDiff was used to analyze
the GIWAXS data.
(200 mTorr, 15 min).
A thin layer of poly(ethyleneꢀ
dioxythiophene):poly(styrenesulfonic acid) (PEDOT:PSS)
(CleviosTM P VP AI 4083) was then spinꢀcoated onto the clean
substrates in air at 5000 rpm for 20 s, and the PEDOT:PSSꢀ
covered substrate was then baked at 150 ºC for 20 min in air.
P3HT:fullerene blend solutions were prepared by dissolving
solid P3HT (Rieke Metal Inc. P100) and solid fullerene
derivatives in oꢀdichlorobenzene (oꢀDCB) with a weight ratio
Results and discussion
of 1:0.8. The concentration was 20 mg/mL with respect to
P3HT. The solutions were stirred at 55 ºC overnight on a hot
plate in a nitrogen glovebox before being cooled down to room
Material properties
temperature. The solutions were spun onto the PEDOT:PSSꢀ Fullerene derivatives 1a–k all have good solubility in common
covered substrates at 1160 rpm for 20 s. The active layers were organic solvents for solar cell fabrication, such as CHCl3, CS2,
still wet when the samples were taken off the spinꢀcoater. CB and
oꢀDCB. The products were characterized by mass
1
Without covering or any other form of solvent annealing, the spectrometry, HꢀNMR and 13CꢀNMR (SI) spectroscopy. The
wet films became dry in the nitrogen glovebox after ~2 min. 1HꢀNMR spectra show that the peaks of methylene groups are
This journal is © The Royal Society of Chemistry 20xx
J. Name., 201X, 00, 1-8 | 3
Please do not adjust margins

Please do not adjust margins
Journal of Materials Chemistry A
Page 4 of 8
View Article Online
DOI: 10.1039/C5TA07688A
ARTICLE
split into AB quartets for the symmetrical 1,4ꢀbisadducts 1a–f
Journal of Materials Chemistry A
,
and mostly split into four doublets for unsymmetrical 1g–k (see
SI), which implies the 1,4ꢀaddition pattern. Single crystals of 1c
and 1e were obtained through slow diffusion of EtOH into a
CS2 solution, providing the Xꢀray structures shown in Fig. 1.
Their structures show a number of short intermolecular C–C
contact distances between fullerene carbons, as shown in Fig.
1b,e; these short packing distances should favor electron
mobility through increased intermolecular orbital interactions.
The shortest contacts are 3.301 Å and 3.028 Å for 1c and 1e
respectively. Despite the different shortꢀcontact distances, both
structures have similar 2ꢀD layered structure and an
,
a
interpenetrating network of methoxylated benzyl groups. The
opposite side of the interpenetrated network of methoxylated
Fig. 2 DFT (B3LYP/6ꢀ31G(d)) calculated HOMO (red dash) and LUMO (blue dash)
energies and selected LUMOs from experimental CV data (green triangle) for PCBM
and symmetrically substituted 1,4ꢀbisbenzyl [60]fullerene adducts (i.e. R1=R2 in Table
1) where the compound legend after PCBM and 1,4ꢀbisbenzyl indicates the relative
position(s) of methoxy group(s) on the benzyl substituents.
The green triangles in Fig. 2 show the electrochemical
properties of several selected MeOꢀBBFs. These cyclic
voltammetry (CV) measurements show that the LUMO levels
of our MeOꢀBBFs are higher than that of PCBM by ~0.05–0.09
eV, depending on the position of the methoxy substituent.
Methoxy substitution at the 2ꢀposition results in a slightly
higher LUMO level than substitution at the 3ꢀ or 4ꢀpositions.
DFT calculations (B3LYP/6ꢀ31G(d)) of the HOMO and LUMO
levels of MeOꢀBBFs 1a–f also show the same trend in
increasing HOMO and LUMO energies as the relative
placement of the methoxy group changes from the 2ꢀ to the 3ꢀ
and 4ꢀpositions (Fig. 2), although the absolute values do not
agree with experiment, as expected for a calculation of this
type. Furthermore, increasing the number of methoxy
substituents also steadily raises both the HOMO and LUMO
energies. This effect of substitution can be explained by the
WheelerꢀHouk model:54–56 since there can be only negligible
overlap between the πꢀsystems of the benzyl substituents and
the fullerene, the interaction between them is primarily
electrostatic. Thus, proximal oxygen lone pairs can increase
electron density on the fullerene πꢀsystem. A methoxy group at
the 2ꢀposition has its lone pairs closest to the fullerene πꢀ
system, while methoxy groups at the 3ꢀ or 4ꢀpositions show less
interaction as the average distance increases. Similarly,
increasing the number of methoxy groups from one (1a–c) to
two (1d–f or 1,4ꢀbis[2,6ꢀdimethoxybenzyl]) or three (1,4ꢀ
bis[2,3,6ꢀtrimethoxybenzyl]) results in easier ionization and
makes reduction become more difficult (Fig. 2).
Fig. 1 a,d) ORTEP representations of the single crystal structures for 1,4ꢀbisadducts 1c
and 1e, respectively. b,e) Packing modes and intermolecular C–C contacts shorter than
van der Waals distances (≤ 0.05 Å) for 1c and 1e, respectively. c,f) Packing structures
for 1c down the crystallographic aꢀaxis, and 1e down the crystallographic bꢀaxis,
respectively. Both are 2ꢀD layered structures. Hydrogens and CS2 cocrystallization
solvent molecules are omitted for clarity.
Photovoltaic device performance and structural characterization
To examine the performance of the MeOꢀBBF 1a–k in BHJ
solar cells, we first blended our new fullerene derivatives with
benzyl groups is a fullerene bilayer displaying a number of
fullereneꢀfullerene closeꢀcontacts (Fig. 1c,f).
4 | J. Mater. Chem. A, 201X, 00, 1-8
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Journal of Materials Chemistry A
Page 5 of 8
View Article Online
DOI: 10.1039/C5TA07688A
Journal of Materials Chemistry A
ARTICLE
P3HT and fabricated photovoltaic devices with a structure All active layers were thermally annealed at 150 °C for 20 min
glass/ITO/PEDOT:PSS (30 nm)/1:0.8 polymer:fullerene (~160ꢀ prior to deposition of the cathode, and J–V curves were
180 nm)/Ca(10 nm)/Al (70 nm). All of the performance
measured under AM 1.5G illumination.
Figure 3a and Table summarize the performance of
2
P3HT:MeOꢀBBFꢀbased devices where the benzyl rings
connected to the fullerene ball have substituents with a single
side group (1a
substituents on even one of the benzyl rings resulted in either
lower FFs (P3HT:1k) or lower sc’s (P3HT:1j) than benzyl
–1c, 1j, 1k). We find that nonꢀmethoxy
J
groups with methoxy substitution; the PCE’s for devices based
on nonꢀmethoxylated 1,4ꢀbisadducts were in the relatively low
range of 2.2 to 2.4%. In contrast, methoxy substitution led to
higherꢀperforming devices (PCE’s ranging from 2.3 to 3.1% for
P3HT:1a) depending on the exact position at which the
methoxy groups were substituted.
Table 2. Summary of Photovoltaic Device Parameters
BHJ
Voc
Jsc
FF
PCE
(mV)
(mA/cm2)
8.2 ± 0.2
7.6 ± 0.3
6.9 ± 0.2
6.1 ± 0.2
7.8 ± 0.1
8.5 ± 0.2
6.0 ± 0.1
8.0 ± 0.2
8.4 ± 0.5
8.3 ± 0.3
5.7 ± 0.3
7.2 ± 0.1
12.1 ± 0.2
12.3 ± 0.2
(%)
(%)
P3HT:PCBM
P3HT:1a
P3HT:1b
P3HT:1c
P3HT:1d
P3HT:1e
P3HT:1f
613 ± 2
716 ± 5
640 ± 4
679 ± 2
680 ± 1
715 ± 1
588 ± 5
720 ± 1
740 ± 1
771 ± 2
704 ± 1
667 ± 4
760 ± 1
825 ± 9
66.6 ± 0.4
57.6 ± 1.1
59.4 ± 0.6
55.4 ± 1.6
59.7 ± 1.1
66.3 ± 0.7
55.4 ± 1.3
57.5 ± 0.1
55.0 ± 0.1
60.1 ± 0.8
56.1 ± 1.5
49.2 ± 1.1
64.4 ± 0.1
53.3 ± 0.1
3.4 ± 0.1
3.1 ± 0.1
2.6 ± 0.1
2.3 ± 0.1
3.2 ± 0.1
3.9 ± 0.1
2.0 ± 0.1
3.3 ± 0.1
3.4 ± 0.2
3.9 ± 0.2
2.2 ± 0.1
2.4 ± 0.1
5.9 ± 0.3
5.4 ± 0.1
P3HT:1g
P3HT:1h
P3HT:1i
P3HT:1j
P3HT:1k
PTB7:PCBM
PTB7:1e
To understand why such subtle variations in the substitution
pattern of our 1,4ꢀbisadducts led to such widely varying device
performance, we studied the morphology of the solar cell active
layers using GIWAXS. The experiments were performed at the
Stanford Synchrotron Radiation Light Source on beamline
11−3 with a wavelength of 0.9742 Å. For these experiments,
we focused on three selected polymer:fullerene systems:
P3HT:PCBM, P3HT:1b and P3HT:1j (Fig. 3c).
In pure films, P3HT orients with the chains edgeꢀon to the
substrate,57–60 and based on the relative inꢀplane and outꢀof
plane scattering intensities, we see that this preferred chain
orientation is maintained upon addition of either MeOꢀBBF or
PCBM (see SI). However, both the characteristic fullerene
diffraction observed at ~1.4 Å–1 and the crystallinity of the
P3HT (as measured by the intensity of the (200) peak) are
smaller when 1b or 1j are used in the active layer compared to
when PCBM is used. Lower crystallinity materials should have
Fig. 3 (a) and (b): Current density versus applied bias for photovoltaic devices based on
P3HT:MeOꢀBBFs where each of the benzyl rings in the MeOꢀBBFs are substituted with
one side group (a) and two methoxy groups (b). The JꢀV curve of
a standard
a poorer carrier mobility, which could explain the lower Jsc and
P3HT:PCBMꢀbased control device is plotted in (b) as the black curve/squares. The error
bars show 1 standard deviation for measurements over at least 6 independent devices.
(c): Example of radiallyꢀintegrated 2ꢀD GIWAXS intensities for three P3HT:fullerene
active layers processed on silicon substrates.
FF of the photovoltaic devices based on these active layers.60
Another notable difference is that for P3HT:1j, the fullerene
peak is shifted towards lower
increased spacing between fullerenes. Fullerene 1j contains a
bulky ꢀbutyl group substituted on one benzyl ring, which likely
Q, which corresponds to an
comparisons and conclusions we draw are based on the study of
devices with compositionꢀ and thicknessꢀmatched active layers.
t
hinders close packing of the fullerene molecules. Consequently,
This journal is © The Royal Society of Chemistry 20xx
J. Name., 201X, 00, 1-8 | 5
Please do not adjust margins

Please do not adjust margins
Journal of Materials Chemistry A
Page 6 of 8
View Article Online
DOI: 10.1039/C5TA07688A
ARTICLE
this increase in interꢀfullerene spacing leads to a decreased Examples of the
Journal of Materials Chemistry A
J–
V
curves for these devices under AM1.5G
electronic coupling between fullerenes and therefore to a illumination are plotted in Fig. 3b.
decrease in carrier mobility.27 This is also consistent with 1j
The most striking result of Fig. 3b and Table 2 is that 1,4ꢀ
having the lowest sc among all the MEOꢀBBFs. bisadduct fullerenes with bismethoxyꢀsubstituted benzyl rings
J
Figure 3a and Table 2 also show that when the methoxy lead to improved solar cell efficiency; in fact, most of the PCEs
group is at the 2ꢀposition of the benzyl group (1a), the reach or surpass those of P3HT:PCBM. Both P3HT:1e and
corresponding device has a greater PCE (3.1%) compared to the P3HT:1i show a ~20% enhancement in PCE compared to
3ꢀposition (2.6%, 1b), which in turn is greater than with the 4ꢀ P3HT:PCBM (black squares in Fig. 3), and P3HT:1d, P3HT:1g
position (2.3%, 1c). The changes in
devices track the changes in experimental (CV) and DFTꢀ devices. P3HT:1f is clearly an exception, having the lowest
calculated LUMO levels of MeOꢀBBFs 1a–f (Fig. 2). PCE of the group, and we are currently investigating the
Although electrostatic interactions can explain the changes morphology of this active layer to understand why.
oc, the variation in the Jsc of the devices cannot be easily Compared to the monomethoxy MeOꢀBBFs 1a
explained by the electrochemical or computational results. fabricated using 1d 1i (except for 1f) show both higher Jsc and
There are two potential reasons for the sensitivity of the oc. We measured nideal values for these devices (Table 3), and
photocurrent to the substitution position. First, the position of found the general trend that devices based on 1d 1i have more
the methoxy group can affect the fullereneꢀtoꢀfullerene contact ideal charge carrier recombination than those based on 1a 1c
Voc (Table 2) of these three and P3HT:1h all have comparable PCEs to P3HT:PCBM
in
V
–1c, devices
–
V
–
–
.
distance and thus the electronic coupling and local carrier This indicates that the nanoscale BHJ morphology is improved
mobility. Second, the way the fullerene interacts with the by the addition of the second methoxy group on the benzyl
polymer could be altered by the structure of the benzyl side rings of the MeOꢀBBFs. This morphology improvement also
chain, which would alter the resultant BHJ morphology. likely contributes to the increase in both
Finally, the propensity of the fullerene to crystallize could alter Figure 3b and Table 2 also reinforce the observation that the
the polymer/fullerene phase separation, again changing the position of the methoxy group(s) has a significant effect on
overall BHJ morphology. device performance. The dependence of Voc on methoxy
Jsc and Voc.
To investigate this, we measured the diode ideality factor, position can be summarized as follows: placing methoxy
nideal, for each of the MeOꢀBBFꢀbased devices by fitting their groups at the 2ꢀ or 2,6ꢀpositions increases Voc, whereas placing
dark J–V curves (Table 3); nideal provides an indicator of the methoxy groups at the 3ꢀ or 5ꢀ positions lowers the Voc. In fact,
charge carrier recombination mechanism.61,62 Table 3 in the SI fullerenes with two methoxy groups at the 3ꢀ (or 5ꢀ) position,
shows that devices with P3HT:1a have the most ideal (i.e., such as 1d and 1e, show lower Voc than those with fewer, such
closest to bimolecular rather than trapꢀdominated) as 1g
recombination among fullerene derivatives 1a , suggesting 5ꢀ positions, as with fullerene 1f, the resultant device shows the
that the variations in Jsc predominantly reflect changes in BHJ lowest Voc. With this trend in mind, we then synthesized the
, 1h and 1i. When both methoxy groups are at the 3ꢀ and
–
c
morphology.
MeOꢀBBF derivative with methoxy groups at the 2,6ꢀpositions
of each benzyl ring, which in principle should be the bestꢀ
performing derivative. Unfortunately, this compound was not
soluble enough in the solvents needed for device fabrication, so
we could not test it in an organic photovoltaic device.
Table 3. Summary of Photovoltaic Device Parameters
BHJ
nideal
Rseires
Rshunt
(×105 ꢂꢀcm2)
2.0 ± 0.1
29 ± 7
(ꢂꢀcm2)
3.2 ± 0.1
4.4 ± 0.4
3.5 ± 0.2
6.0 ± 0.6
4.0 ± 0.3
3.3 ± 0.1
3.8 ± 1.5
4.2 ± 0.8
3.2 ± 0.1
3.3 ± 0.1
6.0 ± 0.9
3.8 ± 0.6
1.5 ± 0.1
3.7 ± 0.4
P3HT:PCBM
P3HT:1a
P3HT:1b
P3HT:1c
P3HT:1d
P3HT:1e
P3HT:1f
1.34 ± 0.03
1.40 ± 0.02
1.80 ± 0.15
1.44 ± 0.03
1.39 ± 0.03
1.31 ± 0.01
1.63 ± 0.04
1.27 ± 0.01
1.41 ± 0.02
1.28 ± 0.03
1.44 ± 0.04
1.51 ± 0.02
1.37 ± 0.01
1.37 ± 0.01
1.2 ± 0.6
14 ± 0.8
8.0 ± 0.2
20 ± 6
3.5 ± 0.7
37 ± 3
P3HT:1g
P3HT:1h
P3HT:1i
7.5 ± 0.4
43 ± 17
16 ± 4
Fig.
4 a) Current density versus applied bias for photovoltaic devices based on
P3HT:1j
PTB7:PCBM (open black square) and PTB7:1e (red circle). The error bars show 1
standard deviation for measurements over at least 6 independent devices. b) UVꢀvisible
absorption spectra for the same active layers used in (a).
P3HT:1k
PTB7:PCBM
PTB7:1e
42 ± 12
0.1 ± 0.01
0.7 ± 0.2
After establishing that alkyl group substitution on the
benzyl ring was inferior to methoxy group substitution for solar
cell performance, we next turned to study the effects of the
number and position of the methoxy substituents. To this end,
we synthesized fullerene derivatives 1d–1i (Table 1), and
fabricated photovoltaic devices from those derivatives. The
performance parameters are again summarized in Table 2.
As mentioned in the introduction, the most successful highꢀ
LUMO fullerenes studied to date (e.g., ICBA and its C70
analog) show poor compatibility with modern low bandgap
pushꢀpull polymers. To see if we could break this trend with
our MeOꢀBBF derivatives, we investigated the compatibility of
one of our best derivatives (1e) with the highꢀperformance lowꢀ
6 | J. Mater. Chem. A, 201X, 00, 1-8
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Please do not adjust margins
Journal of Materials Chemistry A
Page 7 of 8
View Article Online
DOI: 10.1039/C5TA07688A
Journal of Materials Chemistry A
ARTICLE
bandgap polymer PTB7 (Fig. 4). We employed the same device the National Science Foundation under Grants CHEꢀ1112569
structure, with the active layer consisting of PTB7 and 1e at a and CBETꢀ 1510353 (optical, device studies, and fullerene
polymer:fullerene weight ratio of 1:1.5 with a thickness of ~90 synthesis). Portions of this research were carried out at the
nm. These active layers were used asꢀcast, without thermal Stanford Synchrotron Radiation Lightsource, a Directorate of
annealing prior to cathode deposition. For comparison, we also SLAC National Accelerator Laboratory and an Office of
fabricated PTB7:PCBM control devices, taking care to keep the Science User Facility operated for the U.S. Department of
two sets of devices thicknessꢀ and compositionꢀmatched, as Energy Office of Science by Stanford University.
shown by UVꢀVis absorption (Fig. 4b).
Figure 4a and the bottom portion of Table 2 summarize the
References
device performance results. Clearly, fullerene 1e
, which
showed the best performance when blended with P3HT, also
demonstrates excellent compatibility with PTB7. PTB7:1e
devices show a higher Voc and a similarly high Jsc as
PTB7:PCBM devices. The overall device efficiency of
PTB7:1e is slightly lower than PTB7:PCBM, however, due to a
slightly lower FF that may be due to less efficient
crystallization of derivative 1e compared to PCBM, changing
the resultant BHJ morphology.
1
2
C. Deibel and V. Dyakonov, Rep. Prog. Phys, 2010, 73,
096401.
M. Jørgensen, J. E. Carlé, R. R. Søndergaard, M. Lauritzen,
N. A. DagnæsꢀHansen, S. L. Byskov, T. R. Andersen, T. T.
LarsenꢀOlsen, A. P. L. Böttiger, B. Andreasen, L. Fu, L. Zuo,
Y. Liu, E. Bundgaard, X. Zhan, H. Chen and F. C. Krebs, Sol.
Energy Mater. Sol. Cells, 2013, 119, 84–93.
F. C. Krebs, N. Espinosa, M. Hösel, R. R. Søndergaard and
M. Jørgensen, Adv. Mater., 2014, 26, 29–39.
R. R. Søndergaard, M. Hösel and F. C. Krebs, J. Polym. Sci.
Part B Polym. Phys., 2013, 51, 16–34.
M. C. Scharber and N. S. Sariciftci, Prog. Polym. Sci., 2013,
38, 1929–1940.
G. Yu, J. Gao, J. C. Hummelen, F. Wudl and A. J. Heeger,
Science (80ꢀ. )., 1995, 270, 1789–1791.
3
4
5
6
Conclusions
In summary, we have synthesized a series of methoxylated 1,4ꢀ
[60]fullerene bisadducts, 1a–k, which have higher LUMO
levels than PCBM. We evaluated their performance in BHJ
solar cells and compared the resulting BHJ morphologies to
those of the workhorse fullerene PCBM. Our best fullerene
derivatives show more than 20% enhancement in device
efficiency when combined with P3HT, largely due to the
improved Voc resulting from the higher LUMO of the 1,4ꢀbisꢀ
7
8
A. J. Heeger, Adv. Mater., 2014, 26, 10–27.
C.ꢀZ. Li, H.ꢀL. Yip and A. K.ꢀY. Jen, J. Mater. Chem., 2012,
22, 4161.
Y. Liu, J. Zhao, Z. Li, C. Mu, W. Ma, H. Hu, K. Jiang, H.
Lin, H. Ade and H. Yan, Nat. Commun., 2014, 5, 5293.
9
10 H. Zhou, L. Yang and W. You, Macromolecules, 2012.
substituted fullerenes. We found that adding methoxy groups to 11 I. Etxebarria, J. Ajuria and R. Pacios, Org. Electron., 2015,
19, 34–60.
the benzyl rings increases the device performance and that the
number and position of these groups has a dramatic effect on
the solar cell efficiency due to morphological changes. Unlike
12 Z. He, C. Zhong, S. Su, M. Xu, H. Wu and Y. Cao, Nat.
Photonics, 2012, 6, 591–595.
13 J. C. Hummelen, B. W. Knight, F. Lepeq, F. Wudl, J. Yao
and C. L. Wilkins, J. Org. Chem., 1995, 60, 532–538.
previouslyꢀstudied
fullerene
bisadducts,
which
are
outperformed by PCBM when combined into BHJs with low 14 M. M. Wienk, J. M. Kroon, W. J. H. Verhees, J. Knol, J. C.
Hummelen, P. a van Hal and R. A. J. Janssen, Angew. Chem.
Int. Ed. Engl., 2003, 42, 3371–5.
15 Y. He, H. Chen, J. Hou and Y. Li, J. Am. Chem. Soc., 2010,
132, 1377–1382.
16 Y. He and Y. Li, Phys. Chem. Chem. Phys., 2011, 13, 1970–
83.
bandgap polymers, our 1,4ꢀbisadduct 1e demonstrated good
compatibility with one of the best performing red polymers,
PTB7. We are currently working on a deeper understanding of
both the device physics and activeꢀlayer morphologies for these
fullerene derivatives in polymer solar cells to help guide further
synthesis and increase device performance. We believe that 1,4ꢀ
bisbenzyl fullerene bisadducts in general and future C70 analogs
of these molecules are promising candidates for replacing
PCBM to improve the performance of conjugated polymerꢀ
based solar cells.
17 X. Meng, Q. Xu, W. Zhang, Z. Tan, Y. Li, Z. Zhang, L.
Jiang, C. Shu and C. Wang, ACS Appl. Mater. Interfaces
2012, 4, 5966–5973.
,
18 C. Zhang, S. Chen, Z. Xiao, Q. Zuo and L. Ding, Org. Lett.
2012, 14, 1508–1511.
,
19 B. Kim, J. Lee, J. H. Seo, F. Wudl, S. H. Park and C. Yang, J.
Mater. Chem., 2012, 22, 22958.
20 Y. He, M. Shao, K. Xiao, S. C. Smith and K. Hong, Sol.
Energy Mater. Sol. Cells, 2013, 118, 171–178.
Acknowledgements
21 S. Chen, G. Ye, Z. Xiao and L. Ding, J. Mater. Chem. A
2013, 1, 5562–5566.
,
We thank Prof. Kendall N. Houk for his insightful suggestions
on the DFT calculations. This work was supported by the
Center for Molecularly Engineered Energy Materials (MEEM),
an Energy Frontier Research Center (EFRC) funded by the U.S.
Department of Energy (DOE), Office of Science, Office of
Basic Energy Sciences (BES) under Contract Number DEꢀ
AC06ꢀ76RLOꢀ1830 (device studies, Xꢀray diffraction,
manuscript preparation). Additional support was provided by
22 D. He, C. Zuo, S. Chen, Z. Xiao and L. Ding, Phys. Chem.
Chem. Phys., 2014, 16, 7205–7208.
23 D. He, X. Du, Z. Xiao and L. Ding, Org. Lett., 2014, 16, 612–
615.
24 G. Ye, S. Chen, Z. Xiao, Q. Zuo, Q. Wei and L. Ding, J.
Mater. Chem., 2012, 22, 22374–22377.
25 J.ꢀH. Huang, Y.ꢀS. Hsiao, E. Richard, C.ꢀC. Chen, P. Chen,
G. Li, C.ꢀW. Chu and Y. Yang, Appl. Phys. Lett., 2013, 103,
043304.
This journal is © The Royal Society of Chemistry 20xx
J. Name., 201X, 00, 1-8 | 7
Please do not adjust margins

Please do not adjust margins
Journal of Materials Chemistry A
Page 8 of 8
View Article Online
DOI: 10.1039/C5TA07688A
ARTICLE
Journal of Materials Chemistry A
26 M. A. Faist, S. Shoaee, S. Tuladhar, G. F. A. Dibb, S. Foster, 52 R. Subramanian, P. Boulas, M. N. Vijayashree, F. D’Souza,
W. Gong, T. Kirchartz, D. D. C. Bradley, J. R. Durrant and J.
Nelson, Adv. Energy Mater., 2013, 3, 744–752.
M. T. Jones and K. M. Kadish, J. Chem. Soc. Chem.
Commun., 1994, 1847.
27 J. C. Aguirre, C. Arntsen, S. Hernandez, R. Huber, A. M. 53 S. A. Hawks, J. C. Aguirre, L. T. Schelhas, R. J. Thompson,
Nardes, M. Halim, D. Kilbride, N. Kopidakis, S. H. Tolbert,
B. J. Schwartz and N. Daniel, Adv. Funct. Mater., 2013, 24,
784–792.
R. C. Huber, A. S. Ferreira, G. Zhang, A. A. Herzing, S. H.
Tolbert and B. J. Schwartz, J. Phys. Chem. C, 2014, 118,
17413–17425.
28 M. Lenes, G.ꢀJ. A. H. Wetzelaer, F. B. Kooistra, S. C. 54 S. E. Wheeler and K. N. Houk, J. Am. Chem. Soc., 2008, 130,
Veenstra, J. C. Hummelen and P. W. M. Blom, Adv. Mater.
2008, 20, 2116–2119.
,
10854–10855.
55 S. E. Wheeler, J. Am. Chem. Soc., 2011, 133, 10262–10274.
29 Y. He, G. Zhao, B. Peng and Y. Li, Adv. Funct. Mater., 2010, 56 For a different explanation of the effect of substitution on
20, 3383–3389.
PCBM derivatives, see: F. B. Kooistra, J. Knol, F.
Kastenberg, L. M. Popescu, W. J. H. Verhees, J. M. Kroon
and J. C. Hummelen, Org. Lett., 2007, 9, 551–554.
30 X. Meng, G. Zhao, Q. Xu, Z. Tan, Z. Zhang, L. Jiang, C. Shu,
C. Wang and Y. Li, Adv. Funct. Mater., 2014, 24, 158–163.
31 Y.ꢀJ. Cheng, M.ꢀH. Liao, C.ꢀY. Chang, W.ꢀS. Kao, C.ꢀE. Wu 57 H. Sirringhaus, P. J. Brown, R. H. Friend, M. M. Nielsen, K.
and C.ꢀS. Hsu, Chem. Mater., 2011, 23, 4056–4062.
32 N. C. Miller, S. Sweetnam, E. T. Hoke, R. Gysel, C. E.
Miller, J. A. Bartelt, X. Xie, M. F. Toney and M. D.
McGehee, Nano Lett., 2012, 12, 1566–70.
33 D. Di Nuzzo, G.ꢀJ. A. H. Wetzelaer, R. K. M. Bouwer, V. S.
Gevaerts, S. C. J. Meskers, J. C. Hummelen, P. W. M. Blom
and R. A. J. Janssen, Adv. Energy Mater., 2013, 3, 85–94.
Bechgaard, B. M. W. LangeveldꢀVoss, A. J. H. Spiering, R.
A. J. Janssen, E. W. Meijer, P. Herwig and D. M. de Leeuw,
Nature, 1999, 401, 685–688.
58 M. Sanyal, B. SchmidtꢀHansberg, M. F. G. Klein, A.
Colsmann, C. Munuera, A. Vorobiev, U. Lemmer, W.
Schabel, H. Dosch and E. Barrena, Adv. Energy Mater., 2011,
1, 363–367.
34 S. Shoaee, S. Subramaniyan, H. Xin, C. Keiderling, P. S. 59 Y. Kim, S. Cook, S. M. Tuladhar, S. A. Choulis, J. Nelson, J.
Tuladhar, F. Jamieson, S. A. Jenekhe and J. R. Durrant, Adv.
Funct. Mater., 2013, 23, 3286–3298.
R. Durrant, D. D. C. Bradley, M. Giles, I. McCulloch, C.ꢀS.
Ha and M. Ree, Nat. Mater., 2006, 5, 197–203.
35 A. A. Bakulin, A. Rao, V. G. Pavelyev, P. H. M. van 60 G. Zhang, R. C. Huber, A. S. Ferreira, S. D. Boyd, C. K.
Loosdrecht, M. S. Pshenichnikov, D. Niedzialek, J. Cornil, D.
Beljonne and R. H. Friend, Science, 2012, 335, 1340–4.
Luscombe, S. H. Tolbert and B. J. Schwartz, J. Phys. Chem.
, 2014, 118, 18424–18435.
C
36 C. J. Brabec, C. Winder, N. S. Sariciftci, J. C. Hummelen, A. 61 G. A. H. Wetzelaer, M. Kuik, M. Lenes and P. W. M. Blom,
Dhanabalan, P. A. van Hal and R. A. J. Janssen, Adv. Funct.
Mater., 2002, 12, 709–712.
Appl. Phys. Lett., 2011, 99, 153506.
62 R. A. Street, A. Krakaris and S. R. Cowan, Adv. Funct.
37 W. Li, W. S. C. Roelofs, M. M. Wienk and R. A. J. Janssen,
J. Am. Chem. Soc., 2012, 134, 13787–95.
Mater., 2012, 22, 4608–4619.
38 A. B. Sieval, N. D. Treat, D. Rozema, J. C. Hummelen and N.
Stingelin, Chem. Commun., 2015, 51, 8126–8129.
39 Y. Matsuo, H. Oyama, I. Soga, T. Okamoto, H. Tanaka, A.
Saeki, S. Seki and E. Nakamura, Chem. Asian J., 2013, 8,
121–128.
40 Y. Matsuo, A. Iwashita, Y. Abe, C.ꢀZ. Li, K. Matsuo, M.
Hashiguchi and E. Nakamura, J. Am. Chem. Soc., 2008, 130,
15429–15436.
41 Y. Matsuo, Y. Sato, T. Niinomi, I. Soga, H. Tanaka and E.
Nakamura, J. Am. Chem. Soc., 2009, 131, 16048–16050.
42 H. Tanaka, Y. Abe, Y. Matsuo, J. Kawai, I. Soga, Y. Sato and
E. Nakamura, Adv. Mater., 2012, 24, 3521–3525.
43 A. Varotto, N. D. Treat, J. Jo, C. G. Shuttle, N. A. Batara, F.
G. Brunetti, J. H. Seo, M. L. Chabinyc, C. J. Hawker, A. J.
Heeger and F. Wudl, Angew. Chem. Int. Ed., 2011, 50, 5166–
5169.
44 Y. Liang, Z. Xu, J. Xia, S.ꢀT. Tsai, Y. Wu, G. Li, C. Ray and
L. Yu, Adv. Mater., 2010, 22, 1–4.
45 R. Subramanian, K. M. Kadish, M. N. Vijayashree, X. Gao,
M. T. Jones, M. D. Miller, K. L. Krause, T. Suenobu and S.
Fukuzumi, J. Phys. Chem., 1996, 100, 16327–16335.
46 E. Allard, L. Rivitre, J. Delaunay, D. Dubois and J. Cousseau,
Tetrahedron Lett., 1999, 40, 7223–7226.
47 F. Cheng, Y. Murata and K. Komatsu, Org. Lett., 2002, 4,
2541–2544.
48 E. Allard, J. Delaunay and J. Cousseau, Org. Lett., 2003, 5,
2239–2242.
49 E. Allard, F. Cheng, S. Chopin, J. Delaunay, D. Rondeau and
J. Cousseau, New J. Chem., 2003, 27, 188–192.
50 J. Cousseau, E. Allard and S. Chopin, Comptes Rendus
Chim., 2006, 9, 1051–1057.
51 M. Toganoh, K. Suzuki, R. Udagawa, A. Hirai, M. Sawamura
and E. Nakamura, Org. Biomol. Chemsitry, 2003, 1, 2604–
2611.
8 | J. Mater. Chem. A, 201X, 00, 1-8
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins