1,4-fullerene derivatives: Tuning the properties of the electron transporting layer in bulk-heterojunction solar cells

Article abstract of DOI:10.1002/anie.201100029

Tune me up: The increasing number of new donor materials for organic solar cells requires compatible electron acceptors. A series of 1,4-fullerene adducts with tunable chemical, electronic, and material properties is introduced to effectively influence the photovoltaic characteristics of solar cells. Copyright

Full text of DOI:10.1002/anie.201100029

Communications
DOI: 10.1002/anie.201100029
Organic Solar Cells
1,4-Fullerene Derivatives: Tuning the Properties of the Electron
Transporting Layer in Bulk-Heterojunction Solar Cells**
Alessandro Varotto, Neil D. Treat, Jang Jo, Christopher G. Shuttle, Nicolas A. Batara,
Fulvio G. Brunetti, Jung Hwa Seo, Michael L. Chabinyc, Craig J. Hawker, Alan J. Heeger, and
Fred Wudl*
Light and flexible photovoltaic devices based on organic
materials are extensively studied as an alternative to expen-
sive and fragile silicon-based solar cells.[1] The efficiency of
these devices is rapidly increasing with the most recent power
conversion efficiency (PCE) of greater than 8% bringing
them closer to commercial viability.[2] Further improvements
are needed and can be achieved by optimizing the ratio
between donor and acceptor, modifying the electronic
properties of the materials,[3] and optimizing the morphology
of the resulting bulk heterojunction (BHJ).[4]
A direct way to increase the efficiency is to lower the band
gap of the donor material in order to absorb a greater fraction
of the solar spectrum.[5] This concept has been explored
through the synthesis of a series of new low band gap
polymers, which exhibit a decreased band gap as a result of
lowering the lowest unoccupied molecular orbital (LUMO).[6]
An emerging challenge is the need for electronically compat-
ible acceptors with sufficiently low LUMO levels so that
charge separation is efficiently promoted. Optimum misci-
bility of a specific polymer and fullerene combination to
create the optimum degree of phase separation is also a
feature to take into account. Typically, the approach used to
test new donor materials is to fabricate devices using PC61BM
([6,6]-phenyl-C61-butyric acid methyl ester), a well-studied
benchmark acceptor. Only a limited number of other full-
erene derivatives have been successfully employed.[1,7]
Although these fullerene derivatives bear different functional
groups, they are related by the positioning of the substituents
on carbons 1 and 2 of a six-membered ring. From literature
studies it is apparent that even subtle modification of the
nature and especially position of substituents can drastically
alter the electronic properties of fullerene derivatives. For
example, PC71BM has reduced symmetry that increases
visible light absorption and has a positive effect on current
generation in polymer BHJ cells.[8]
Here we report the synthesis of a series of novel fullerene
derivatives functionalized through the “1,4” position and
their use in organic photovoltaics (OPVs). Features that
distinguish this class of fullerene derivatives are: 1) straight-
forward synthesis which includes versatility of functionaliza-
tion with different substrates starting from the same material
(fullerenol, Scheme 1); 2) tunable LUMO energy by append-
ing electron-donating or electron-withdrawing groups;
3) lower symmetry which decreases the optical gap and
produces an increased absorption in the visible (the extinction
coefficient at 480 nm of a 1,4-adduct is approximately 8 times
larger than that of PCBM) ; 4) tunable solubility which
influences the morphology of the BHJ. Like PC71BM, 1,4-
addends have increased light absorption at ca. 500 nm
(Figure 1) with the advantage that all the C60 derivatives
[*] Dr. A. Varotto,[+] N. D. Treat,[+] Dr. C. G. Shuttle, N. A. Batara,
Dr. F. G. Brunetti, Prof. M. L. Chabinyc, Prof. C. J. Hawker,
Prof. F. Wudl
California NanoSystems Institute and
the Material Research Laboratory
University of California Santa Barbara, CA, 93106 (USA)
E-mail: wudl@chem.ucsb.edu
Homepage: http://www.mc-cam.ucsb.edu/
Dr. J. Jo, Dr. J. H. Seo, Prof. A. J. Heeger
Department of Physics
University of California Santa Barbara, CA, 93106 (USA)
[+] Both authors contributed equally to this work.
[**] Partial support of this work (A.V., N.A.B., F.G.B., F.W.) was provided
by MC-CAM IRP7. Additional support (C.S., N.D.T., M.L.C.) was
provided by the Center for Energy Efficient Material, an Energy
Frontier Research Center funded by the U.S. Department of Energy,
Office of Science, Office of Basic Energy Sciences under Award
Number DE-SC0001009.
Scheme 1. Synthesis of the derivatives from fullerenols. a) Pentafluoro-
benzene, TsOH, 708C, 12 h. b) Substituted benzene (trialkyloxyben-
zene, 2-ethylhexylbenzene, phenylvaleric acid methylester, aniline, ani-
sole), TsOH, 808C, 12 h.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201100029.
5166
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 5166 –5169

presented here are the same kind of molecules and not a
mixture of isomers.[8]
In virtue of all these points we decided to investigate the
chemistry to synthesize a large set of 1,4 fullerene derivatives
(Schemes 1 and 2) with different LUMO energy levels and
different solubility. Then, in order to prove the concept of
on the fullerene, which is consistent with C2 symmetry) as well
as mass spectrometry. The UV/Vis spectrophotometry of
these addends displayed an increased and wider absorption
compared to traditional fullerenes, with an onset at approx-
imately 500 nm and extinction coefficient of ca.
5000 molÀ1 cmÀ1 (Figure 1). The energy of the LUMO was
Figure 1. UV/Vis absorption of PEHOB (gray; see Scheme 2) and
PCBM (black) in 1,2-dichlorobenzene normalized at 350 nm.
estimated by the first peak of the reduction measured by
cyclic voltammetry and that of the HOMO by subtracting the
value of the optical gap in solution. These values were
compared to those obtained by UV photoelectron spectros-
copy (UPS), in which the energy of the HOMO was
calculated from the binding energy in the solid state and
that of the LUMO was estimated using the optical gap of the
films on glass. Despite the large difference in the absolute
values using these techniques, the relative values are con-
sistent. A summary of the LUMO energy values for the 1,4-
fullerene adducts synthesized here and for PCBM can be
found in Table 1 (calculation detailed in the Supporting
Information). Compared to the PCBM analogues described
by Hummelen et al.,[3] the 1,4-fullerene addends with sub-
stituents directly appended appear to affect their redox
potentials to a greater degree. For instance, one pentafluoro-
phenyl group on PCBM raises the redox potential from
À1.08 mV (PCBM) to À1.042 mV (net change 0.038 V),
whereas PF8 is increased from À1.02 to À0.910 mV (net
change 0.11 V; see Table 1 and reference [3]).
Scheme 2. Structure of some of the 1,4-addends synthesized and
studied. R=propyloxy or hexyloxy.
tunability of their properties, we fabricated BHJ solar cells
employing some of these derivatives. We observed a signifi-
cant change in the photovoltaic (PV) characteristics of the
devices that could be predicted from the specific properties of
each fullerene (for instance increased open circuit voltage
using higher LUMO fullerenes). In addition to study the
chemistry of C60, we also tested the reaction on C70, where the
1H NMR spectrum displayed three proton signals of exactly
the same intensity corresponding to the alcohol group. We
concluded that three isomers formed during this reaction,
which is consistent with other C70 functionalization such as
the synthesis of PC71BM.[7] To the best of our knowledge,
there is only one example of 1,4-fullerene addend used in
solar cells in a p-i-n layered architecture different from the
classical BHJ.[9] Nevertheless, in this example the fullerene
derivative bears the same functional groups on both carbons
and therefore it is more symmetrical. Additionally, it is
employed in a non-classical BHJ architecture using an
insoluble benzoporphyrin as the organic donor, instead of a
polymer.
Table 1: Reduction potential of addends calculated from cyclic voltam-
metry (1,2-dichlorobenzene, scan rate 50 mVsÀ1).
Addend
E1
E2
LUMO [eV][a]
LUMO [eV][b]
red
red
PFOH
PF8
ANP
PEHOB
PCVM
PTHOB
PCBM
À0.94
À0.91
À0.99
À1.06
À1.02
À1.15
À1.05
À1.46
À1.37
À1.53
À1.50
À1.48
À1.59
À1.47
À3.86
À3.91
À3.81
À3.74
À3.76
À3.65
À3.74
–
The synthesis of the novel 1,4-adducts reported herein was
achieved by Wang and co-workers by reaction of 1,4-full-
erenol with a variety of different aromatic substituents.[10] The
procedure developed for the synthesis of the addends was to
react the 1,4-aryl fullerenol with a substituted benzene, used
as the solvent of the reaction, and in the presence of p-toluene
sulfonic acid (TsOH).[10a] The identity of the 1,4-fullerene
À4.75
–
À4.30
À4.30
À4.24
À4.30
[b] Calculated from the onset of the first reduction potential against
ferrocene measured from CV. [b] Estimated from the onset of the optical
density added to the HOMO value calculated from UPS.
1
addends was confirmed by H- and 13C-NMR spectroscopy
(the latter displayed more than 40 signals for the sp2 carbons
Angew. Chem. Int. Ed. 2011, 50, 5166 –5169
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5167

Communications
As anticipated above, we decided to focus our attention
synthetic conditions we only observed formation of an N-
bonded addend directly bound to the C-4 on C60 (ANP). In
this case, since the nitrogen is more electronegative than
carbon, it acted as an electron-withdrawing group and
lowered the LUMO by approximately 50 mV. The yield was
much higher than for previous derivatives (approximately
40%) and the synthesis is interesting because it showed a
on tuning various properties of the fullerenes by appending
different functional groups to the 1,4-fullerenol. In order to
study the reactivity we initially allowed the aryl fullerenol to
react with anisole and we observed the formation of the para
isomer. The appending of the methoxy group did not alter
significantly the LUMO energy, compared to other 1,2-
addends such as PCBM. This was in agreement with a
previous report, where only one alkyloxy group in proximity
of PCBM did not alter the LUMO energy.[3] Nevertheless, the
HOMO level was higher as the band gap was smaller, which
was confirmed quantitatively by UPS. Due to the poor
solubility, this derivative was not employed to fabricate
photovoltaic devices. In order to make this compound
suitable for solution processing, we appended a 2-ethylhex-
yloxy group yielding PEHOB. The synthetic procedure was
analogous to that of the anisole derivative but using
2-ethylhexyloxy benzene. Although the overall yield was
low (ca. 10%), more than 60% of fullerenol, as well as the
2-ethylhexyloxy benzene, could be recovered and re-used.
In order to increase the LUMO energy of the fullerene
derivative, we chose to append the stronger electron-donating
group 1,2,3-trialkyloxy benzene (propyloxy benzene yielded
PTPOB and hexyloxy benzene PTHOB; see Scheme 2). The
tri-alkyloxybenzenes were synthesized according to literature
procedures and used in multiple reactions as the solvent.[11]
Surprisingly, despite the steric hindrance caused by the
alkyloxy groups, we only observed the formation of the
product with the benzene appended in meta to the central
alkoxy carbon. Both 1H (Figure 2) and 13C NMR spectroscopy
confirmed the presence of two distinct protons and five
carbon atoms. The LUMO energy of PTHOB/PTPOB was 50
to 100 mV higher than that of PCBM, depending whether
measured in solution or in the solid phase. In addition to a
smaller band gap as other 1,4 derivatives, the trialkyloxy
derivatives delivered a higher VOC (detailed below) and
tunable miscibility with the polymer by varying the alkyl
chain length. The yield was approximately 20% and, once
again, both fullerenol and alkyloxy benzene could be re-used.
We then attempted to create a derivative with a further
increased LUMO by reacting fullerenol with aniline. The
aniline could act as electron-donating group, but under these
À
simple and straightforward way to create a C N bond directly
on the fullerene.
To stabilize the LUMO energy, the fullerenol derivative
was allowed to react with pentafluoro benzene, a very strong
electron-withdrawing group. The synthesis was performed
using a para-trifluoromethyl-substituted fullerenol as shown
in Scheme 1 (PFOH). The LUMO energy is greatly decreased
by more than 200 mV and the HOMO is similarly lowered.
This material should be compatible with very low LUMO
polymers which cannot transfer charge to PCBM or other
acceptors with similar energy levels. Moreover, due to their
low surface energy, highly fluorinated compounds tend to
self-aggregate on the top of blended films[12] which is
favorable for the specific device architecture of OPVs.
BHJ solar cells were also fabricated using regioregular
poly(3-hexylthiophene) (rr-P3HT) as the electron-donating
polymer and PTPOB and PTHOB as the electron-accepting
1,4-fullerene adducts. This series was chosen because it allows
us to demonstrate both the tunability of the electronic
properties (higher LUMO level) and relative degree of
phase separation (using different alkyl chain lengths). As
anticipated above, this series of materials possesses a higher
LUMO and broader absorption as well as tunable solubility
with no effect on the energy levels. The ability to separate the
relative degree of phase separation from the electronic
properties is of particular interest for future studies. These
devices were fabricated using a solvent annealing technique
(detailed in the Supporting Information). For the fabrication
of the devices we used conditions appropriate for P3HT/
PCBM BHJ solar cells, but there might be processing
conditions that could further optimize the performance of
1,4-addends. Given the different properties of each fullerene,
the solar cells would require a large degree of optimization
that is beyond the scope of this report. All current–voltage
characteristics for the fabricated devices containing the
different 1,4-fullerene adducts can be viewed in the Support-
ing Information. A parallel study was also conducted using
PCBM, which has a lower LUMO and does not absorb in the
visible. As reported in Figure 3, devices of P3HT:PTPOB
combined at a 1:1 w/w ratio and annealed at 1658C for 10 min
yielded devices with a median efficiency of 2.3%. Similarly,
devices of P3HT:PTHOB were fabricated with a 1:1.1 w/w
ratio (larger fraction of fullerene to compensate for larger
weight of insulating alkyl chains) and found to yield devices
with a median efficiency of 1.2%, after annealing at 1658C for
30 min. Increasing the length of alkyl chain from propyl to
hexyl (without changing the molar ratio of fullerene to
polymer) yielded devices with a lower short circuit current
density and fill factor (Figure 3). It is hypothesized that this is
due to non-optimal phase separation within the device. These
materials provide a means to test important hypotheses with
regards to the relationship of the solubility of fullerenes in the
Figure 2. 1H NMR spectrum of PTHOB. R=hexyloxy.
5168
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 5166 –5169

Keywords: flexible photovoltaics · fullerenes ·
.
organic electronics · polymers · solar cells
[1] a) B. C. Thompson, J. M. J. Frꢀchet, Angew. Chem. 2008, 120,
62 – 82; Angew. Chem. Int. Ed. 2008, 47, 58 – 77; b) A. W. Hains,
Z. Liang, M. A. Woodhouse, B. A. Gregg, Chem. Rev. 2010, 110,
6689 – 6735; c) N. R. Armstrong, W. Wang, D. M. Alloway, D.
Placencia, E. Ratcliff, M. Brumbach, Macromol. Rapid
Commun. 2009, 30, 717 – 731; d) J. L. Delgado, P.-A. Bouit, S.
Filippone, M. A. Herranz, N. Martꢁn, Chem. Commun. 2010, 46,
4853 – 4865; e) J. Jang, S. Nam, D. S. Chung, S. H. Kim, W. M.
Yun, C. E. Park, Adv. Funct. Mater. 2010, 20, 2543 – 2736, f) H.
Bolink, J. Coronado, E. Forment-Aliaga, A. Lenes, M. La Rosa,
A. Filippone, N. Martin, J. Mater. Chem. 2011, 21, 1382 – 1386;
g) M. Hallermann, E. Da Como, J. Feldmann, M. Izquierdo, S.
Filippone, N. Martin, S. Juchter, E. von Hauff, Appl. Phys. Lett.
2010, 97, 023301.
Figure 3. J–V curves of solar cells fabricated with PTPOB/P3HT in a
1:1 ratio and PTHOB/P3HT in a 1.1:1 ratio.
[2] The OSA Direct Newsletter, 2010.
polymer to photoconversion efficiency by providing a series
of derivatives with similar electronic levels, but varying
solubility. This relationship of solubility to efficiency has
been previously detailed with a number of different 1,2-
fullerene derivatives.[13] Note that devices made from PCBM
as the electron acceptor at a ratio of 1:0.8 were found to give
median efficiencies of 2.8%. The most notable improvement
was that the alkoxylated 1,4-derivatives yielded approxi-
mately 70 mV higher VOC, consistent with the observation of
higher LUMO energy. However, the increased absorption
had little improvement on the short circuit current density,
which is attributed to non-optimal active layer morphology
and may be more important for polymers that do not strongly
absorb light with wavelengths ranging from 400–500 nm.
In conclusion we explored the syntheses of novel fullerene
addends based on 1,4-addition, tuned the electronic, chemical
and physical properties of the derivatives, and investigated
their use in BHJ solar cells. More specifically, we described an
approach to increase or decrease the LUMO energy of the
1,4-addends and how its value influences the VOC of the
device. The 1,4-derivatives are less symmetrical than previous
reported fullerenes and therefore display a stronger absorp-
tion in the visible. Finally, the effect of altering the solubility
of the fullerene in the polymer has been explored and will be
subject of future reports.
[3] F. B. Kooistra, J. Knol, F. Kastenberg, L. M. Popescu, W. J. H.
Verhees, J. M. Kroon, J. C. Hummelen, Org. Lett. 2007, 9, 551 –
554.
[4] a) J. K. Lee, W. L. Ma, C. J. Brabec, J. Yuen, J. S. Moon, J. Y.
Kim, K. Lee, G. C. Bazan, A. J. Heeger, J. Am. Chem. Soc. 2008,
130, 3619 – 3623; b) J. Peet, J. Y. Kim, N. E. Coates, W. L. Ma, D.
Moses, A. J. Heeger, G. C. Bazan, Nat. Mater. 2007, 6, 497 – 500;
c) R. D. Kennedy, A. L. Ayzner, D. D. Wanger, C. T. Day, M.
Halim, S. I. Khan, S. H. Tolbert, B. J. Schwartz, Y. Rubin, J. Am.
Chem. Soc. 2008, 130, 17290 – 17292.
[5] a) J. Y. Kim, K. Lee, N. E. Coates, D. Moses, T.-Q. Nguyen, M.
Dante, A. J. Heeger, Science 2007, 317, 222 – 225; b) G. C. Welch,
G. C. Bazan, J. Am. Chem. Soc. 2011, 133, 4632 – 4644; c) C. H.
Woo, P. M. Beaujuge, T. W. Holcombe, O. P. Lee, J. M. J.
Frechet, J. Am. Chem. Soc. 2010, 132, 15547 – 15549; d) J. Hou,
H.-Y. Chen, S. Zhang, R. I. Chen, Y. Yang, W. Wu, G. Li, J. Am.
Chem. Soc. 2009, 131, 15586 – 15587.
[6] a) E. Bundgaard, F. C. Krebs, Sol. Energy Mater. Sol. Cells 2007,
91, 954 – 985; b) Z. Zhu, D. Waller, C. J. Brabec, New Con-
struction of Low-Bandgap Conducting Polymers, Wiley-VCH,
Weinheim, 2009; c) C. Li, M. Liu, N. G. Pschirer, M. Baum-
garten, K. Mꢂllen, Chem. Rev. 2010, 110, 6817 – 6855.
[7] a) Y. He, H.-Y. Chen, J. Hou, Y. Li, J. Am. Chem. Soc. 2010, 132,
1377 – 1382; b) S. A. Backer, K. Sivula, D. F. Kavulak, J. M. J.
Frꢀchet, Chem. Mater. 2007, 19, 2927 – 2929; c) M. Lenes, G.-
J. A. H. Wetzelaer, F. B. Kooistra, S. C. Veenstra, J. C. Humme-
len, P. W. M. Blom, Adv. Mater. 2008, 20, 2116 – 2119.
[8] M. M. Wienk, J. M. Kroon, W. J. H. Verhees, J. Knol, J. C.
Hummelen, P. A. van Hal, R. A. J. Janssen, Angew. Chem. 2003,
115, 3493 – 3497; Angew. Chem. Int. Ed. 2003, 42, 3371 – 3375.
[9] Y. Matsuo, Y. Sato, T. Niinomi, I. Soga, H. Tanaka, E. Nakamura,
J. Am. Chem. Soc. 2009, 131, 16048 – 16050.
Experimental Section
Chemicals were purchased from Sigma–Aldrich, Acros Organics, SES
Research, Fisher Scientific, Merck and used as received without
further purification unless otherwise stated. 1H and 13C NMR spectra
were obtained on a Bruker 500 MHz spectrometer and referenced to
the solvent peak. Mass spectrometry was performed by the UC Santa
Barbara Mass Spectrometry Laboratory. UV/Vis spectra were
recorded on an Agilent 8453 spectrophotometer using 1,2-dichloro-
benzene (ODCB) solution of fullerene in 1 cm quartz cuvettes at
room temperature. The electrochemical measurements were carried
out with a Princeton Applied Research Model 263A Potentiostat/
Galvanostat employing Ag/AgCl as reference electrode, a platinum
wire as a counter electrode, and an internal ferrocene/ferrocenium
standard. Solutions of fullerenes were in dry ODCB containing 0.1m
tetrabutylammoniumhexafluorophosphate as the supporting electro-
lyte.
[10] a) G.-W. Wang, Y.-M. Lu, Z.-X. Chen, Org. Lett. 2009, 11, 1507 –
1510; b) Z.-X. Chen, G.-W. Wang, J. Org. Chem. 2005, 70, 2380 –
2383.
[11] M. J. Jeong, J. H. Park, C. Lee, J. Y. Chang, Org. Lett. 2006, 8,
2221 – 2224.
[12] Q. Wei, T. Nishizawa, K. Tajima, K. Hashimoto, Adv. Mater.
2008, 20, 2211 – 2216.
[13] P. A. Troshin, H. Hoppe, J. Renz, M. Egginger, J. Y. Mayorova,
A. E. Goryachev, A. S. Peregudov, R. N. Lyubovskaya, G.
Gobsch, N. S. Sariciftci, V. F. Razumov, Adv. Funct. Mater.
2009, 19, 779 – 788.
Received: January 4, 2011
Published online: April 19, 2011
Angew. Chem. Int. Ed. 2011, 50, 5166 –5169
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5169