FeCl3-mediated synthesis of fullerenyl esters as low-LUMO acceptors for organic photovoltaic devices

Article abstract of DOI:10.1021/ol301186u

C₆₀ reacted with aromatic and aliphatic carboxylic acids in the presence of inexpensive FeCl₃ at room temperature to produce hydroxyfullerenyl esters C₆₀(OCOR)(OH) in up to 68% isolated yield. The hydroxyl group was utilized in functional group transformations to obtain a diester derivative C₆₀(OCOAr)(OCOPh) (Ar = 2,6-xylyl) and a siloxyl derivative C₆₀(OCOAr)(OSiMe₃). The diester and siloxyl derivatives were found to possess low-lying LUMO levels were utilized in organic photovoltaic devices showing 1.3% power conversion efficiency.

Full text of DOI:10.1021/ol301186u

ORGANIC
LETTERS
2012
Vol. 14, No. 13
3276–3279
FeCl3-Mediated Synthesis of Fullerenyl
Esters as Low-LUMO Acceptors for
Organic Photovoltaic Devices
Masahiko Hashiguchi,*,† Naoki Obata,‡,§ Masashi Maruyama,‡ Kee Sheng Yeo,‡
Takao Ueno,† Tomohiko Ikebe,† Isao Takahashi,† and Yutaka Matsuo*,‡
Fullerene Development Group, Performance Products Division, Mitsubishi Chemical
Corporation, 1-1 Kurosakishiroishi, Yahatanishi-ku, Kitakyushu 806-0004, Japan,
Department of Chemistry, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo
113-0033, Japan, and Mitsubishi Chemical Group Science and Technology Research
Center, Inc., 1000 Kamoshida, Aoba-ku, Yokohama, Kanagawa, 227-8502, Japan
6206530@cc.m-kagaku.co.jp; matsuo@chem.s.u-tokyo.ac.jp
Received May 1, 2012
ABSTRACT
C60 reacted with aromatic and aliphatic carboxylic acids in the presence of inexpensive FeCl3 at room temperature to produce hydroxyfullerenyl
esters C60(OCOR)(OH) in up to 68% isolated yield. The hydroxyl group was utilized in functional group transformations to obtain a diester
derivative C60(OCOAr)(OCOPh) (Ar = 2,6-xylyl) and a siloxyl derivative C60(OCOAr)(OSiMe3). The diester and siloxyl derivatives were found to
possess low-lying LUMO levels were utilized in organic photovoltaic devices showing 1.3% power conversion efficiency.
Development of fullerene derivatives having various
LUMO levels is important for organic photovoltaic
(OPV) applications.1 Although obtaining high-LUMO
fullerenes is relatively straightforward, since the addition
of carbon functional groups to fullerene generally raises
LUMO levels,2 the approach toward low-LUMO fuller-
enes is rather limited.3
Fullerenyl esters4 are considered to be suitable for OPV
application as low-LUMO acceptors: these compounds
bear electron-withdrawing ester moieties and exhibit high
solubility in organic solvents. For fullerenyl esters, only the
synthesis of acetoxylated fullerenes and C60-fused lactones
have been reported in the literature.4,5 These compounds have
(3) (a) Brabec, C. J.; Cravino, A.; Meissner, D.; Sariciftci, N. S.;
Fromherz, T.; Rispens, M. T.; Sanchez, L.; Hummelen, J. C. Adv. Funct.
Mater. 2001, 11, 374–380. (b) Brabec, C. J.; Cravino, A.; Meissner, D.;
Sariciftci, N. S.; Rispens, M. T.; Sanchez, L.; Hummelen, J. C.; Fromherz,
T. Thin Solid Films 2002, 368, 403–404. (c) Creegan, K. M.; Robbins, J. L.;
Robbins, W. K.; Millar, J. M.; Sherwood, R. D.; Tindall, P. J.; Cox, D. M.;
Smith, A. B., III; McCauley, J. P., Jr.; Jones, D. R.; Gallagher, R. T. J. Am.
Chem. Soc. 1992, 114, 1103–1105. (d) Keshavarz-K, M.; Knight, B.;
Srdanov, G.; Wudl, F. J. Am. Chem. Soc. 1995, 117, 11371–11372. (e)
Keshavarz-K, M.; Knight, B.; Haddon, R. C.; Wudl, F. Tetrahedron 1996,
52, 5149–5159.
(4) (a) Zhdankin, V. V.; Hanson, K. J.; Koposov, A. E.; Blomquist,
E.; Tykwinski, R. R. Mendeleev Commun. 2001, 11, 51–52. (b) Bernstein,
R.; Foote, C. S. Tetrahedron Lett. 1998, 39, 7051–7054. (c) Chen, Z.-X.;
Wang, G.-W. J. Org. Chem. 2005, 70, 2380–2383. (d) Wang, G.-W.; Li,
F.-B.; Zhang, T.-H. Org. Lett. 2006, 8, 1355–1358. (e) Li, F.-B.; Liu,
T.-X.; Huang, Y.-S.; Wang, G.-W. J. Org. Chem. 2009, 74, 7743–7749.
(f) Petrie, S.; Javahery, G.; Wincel, H.; Wang, J.; Bohme, D. K. Int. J.
Mass Spectrom. Ion Processes 1994, 138, 187–208.
† Fullerene Development Group, Performance Products Division,
Mitsubishi Chemical Corporation.
‡ Department of Chemistry, The University of Tokyo.
§ Mitsubishi Chemical Group Science and Technology Research Center,
Inc.
(1) Brabec, C.; Dyakonov, V.; Scherf, U. Organic Photovoltaics;
Wiley-VCH: Weinheim, 2008.
(2) (a) Kooistra, F. B.; Knol, J.; Kastenberg, F.; Popescu, L. M.;
Verhees, W. J. H.; Kroon, J. M.; Hummelen, J. C. Org. Lett. 2007, 9,
551–554. (b) Lenes, M.; Wetzelaer, G.-J. A. H.; Kooistra, F. B.;
Veenstra, S.-C.; Hummelen, J. C.; Blom, P. W. M. Adv. Mater. 2008,
20, 2116–2119. (c) He, Y.; Chen, H.-Y.; Hou, J.; Li, Y. J. Am. Chem. Soc.
2010, 132, 1377–1382. (d) Matsuo, Y.; Sato, Y.; Niinomi, T.; Soga, I.;
Tanaka, H.; Nakamura, E. J. Am. Chem. Soc. 2009, 131, 16048–16050.
(e) Zhang, Y.; Matsuo, Y.; Li, C.-Z.; Tanaka, H.; Nakamura, E. J. Am.
Chem. Soc. 2011, 133, 8086–8089. (f) Matsuo, Y.; Hatano, J.; Kuwabara,
T.; Takahashi, K. Appl. Phys. Lett. 2012, 100, 063303.
r
10.1021/ol301186u
Published on Web 06/12/2012
2012 American Chemical Society

been synthesized through one-pot reactions with (diacetoxy)-
iodobenzene,4a Mn(OAc) 2H2O, or Pb(OAc)4.4cꢀe However,
3
Table 1. Reaction Conditions and Yields for the Reaction of C60
with Carboxylic Acids 1aꢀn in the Presence of FeCl3
a general synthetic methodology to obtain fullerenyl esters
with a broad substrate scope has not been realized so far.
Herein we report facile one-pot syntheses of hydroxy-
fullerenyl esters C60(OCOR)(OH) by using versatile car-
boxylic acids and inexpensive iron trichloride (FeCl3). This
reaction proceeds under mild conditions and possesses
substrate diversity catering to both aromatic and aliphatic
carboxylic acids.
a
Treatment of C60 with 2,6-dimethylbenzoic acid (1a, 10
equiv) in the presence of FeCl3 (20 equiv) in 1,1,2,2-tetra-
chloroethane (TCE) at 25 °C for 4 h produced 1-hydroxy-
2-{2,6-dimethylphenylcarbonyloxy}[60]fullerene (hydroxyl-
fullerenyl 2,6-dimethylbenzoate), C60(OCOC6H3Me2-2,6)
(2a), in 68% isolated yield (Table 1, entry 1). Aromatic
organic solvents such as toluene and xylene, often used in
the synthesis of fullerene derivatives, were found to be
unsuitable for this reaction. The use of toluene and xylene
causes a side reaction, hydroarylation,6 which is mediated
by Lewis acidic FeCl3. This will in turn lead to a decrease in
the chemical yield of the desired product. Therefore, TCE
was chosen instead as the solvent for this reaction, since
it did not have any aromatic component, and it could
dissolve the starting material C60. This reaction afforded
almost quantitative conversion of C60 upon stirring at rt
for 4ꢀ8 h. As shown in Table 1, 2a was obtained as the
main product. Possible side products in this reaction were
di- and multiadducts. Although numerous side products
were formed due to the inevitable formation of regiochem-
ical isomers of di- and multiadducts, the absolute amount
obtained for each side product was very small. The draw-
back of this reaction was the need for a large excess of
substrates and FeCl3; maximum selectivity was obtained
when 10 equiv of carboxylic acid and 20 equiv of FeCl3
were used (Table S4). Nevertheless, this reaction is highly
applicable to industrial production because of the relative
low cost of both substrate and iron trichloride as compared
to C60. The reaction mechanism is still unclear, but we
propose that the mechanism involves a cationic fullerene as
a key intermediate (Scheme S1).6d
Product 2a can be purified with either silica-gel column
chromatography orHPLC separation. Purified compound
2a was characterized with 1H and 13C NMR, IR, electro-
spray ionization (ESI)-mass spectroscopic analysis, and
elemental analysis. The structure of 2a was determined
unambiguously by X-ray crystallographic analysis of its
single crystal, which was obtained by slow evaporation of a
CS2 solution of 2a (Figure 1). It was elucidated from the
X-ray structure that the arylcarbonyloxyl group and the
hydroxyl group were added to C60 in a 1,2-addition
pattern.
a All reactions were performed under a nitrogen atmosphere with a
molar ratio of C60/FeCl3/1 = 1:20:10. b Conversion of C60 and selectivity
of product were determined by HPLC analysis. c Isolated yield. d No
reaction.
(6) (a) Olah, G. A.; Bucsi, I.; Lambert, C.; Aniszfeld, R.; Trivedi,
N. J.; Sensharma, D. K.; Prakash, G. K. S. J. Am. Chem. Soc. 1991, 113,
9387–9388. (b) Iwashita, A.; Matsuo, Y.; Nakamura, E. Angew. Chem.,
Int. Ed. 2007, 46, 3513–3517. (c) Hoke, S. H., II; Molstad, J.; Payne,
G. L.; Kahr, B.; Ben-Amotz, D.; Cooks, R. G. Rapid Commun., Mass
Spectrosc. 1991, 5, 472–474. (d) Hashiguchi, M.; Watanabe, K.; Matsuo,
Y. Org. Biomol. Chem. 2011, 9, 6417–6421.
(5) (a) Irngartinger, H.; Weber, A. Tetrahedron Lett. 1997, 38, 2075–
2076. (b) Tuktarov, A. R.; Akhmetov, A. R.; Pudas, M.; Ibragimov,
A. G.; Dzhemilev, U. M. Tetrahedron Lett. 2008, 49, 808–810. (c) Wang,
G.-W.; Lu, Y.-M.; Chen, Z.-X. Org. Lett. 2009, 11, 1507–1510.
(d) Irngartinger, H.; Weber, A.; Escher, T. Eur. J. Org. Chem. 2000,
1647–1651.
Org. Lett., Vol. 14, No. 13, 2012
3277

used. Wang et al. had previously reported an Fe(ClO4)3-
mediatedadditionof aldehydesand esterstofullerene.8 We
attempted the reaction using other Lewis acids, oxidants,
and iron complexes (AlCl3, RuCl3, BF3 Et2O, CuCl2,
3
FeCl2, ZnCl2, Fe(acac)3, and Fe(ClO4)3) in place of FeCl3,
but we also could not obtain the desired fullerenyl esters
from their respective carboxylic acids. Therefore, this
reaction was found to be specific for FeCl3 and carboxylic
acids only.
Next, we performed functional group transformations
on the hydroxyl groups of 2a. Electrochemicalstudies of 2a
showed its irreversible reduction/oxidation process due to
the presence of the hydroxyl group. Masking the hydrogen
atom of the hydroxyl group is thus necessary to obtain an
electrochemically stable fullerene-based electron acceptor
for application in OPV devices. Syntheses of fullerene
ethers and esters from fullerenols is generally straight-
forward.5b Treatment of fullerenol 2a with benzoic acid
chloride (5 equiv), in the presence of diethyl amine (5 equiv)
and dimethylaminopyridine (DMAP, 0.1 equiv) in pyridine
at 0 °C for 4 h, produced a diester derivative 3 in 70% yield
(Scheme 1). A siloxyl derivative 4 was also obtained in 87%
yield via the reaction of fullerenol 2a with trimethylsilyl
chloride (20 equiv), in the presence of DMAP (0.1 equiv) in
pyridine at 25 °C for 24 h. Compounds 3 and 4 were
Figure 1. Crystal structure of 2a 0.5CS2. (a) ORTEP drawing
with thermal ellipsoids at 30% probability. (b) Ball-and-stick
model.
3
The scope and limitation of the reaction is illustrated in
Table 1. The reaction was applicable to both aromatic and
aliphatic carboxylic acids, though the latter type gave
lower yields. Nonsubstituted benzoic acid (1b) showed
similar reactivity to 1a (conversion of C60 = 95%), but
comparatively low selectivity, to produce the correspond-
ing fullerenyl ester 2b (14% yield, entry 2). We postulate
that the poor selectivity was caused by the coordination of
FeCl3 to the carbonyl group of the substrate. Another
potential reason for this low yield would be the side
reaction between a possible cationic fullerene intermediate
(vide infra) and nonsubstituted aryl groups. There had
been many reports on reactions of fullerene cations with
aryl groups in literature.5c,6a,6d,7 Monomethylated benzoic
acids 1c, 1d, 1e, gave corresponding fullerenyl esters 2c, 2d,
2e, in chemical yields of 24ꢀ28%. For dimethylated
benzoic acids, 2,4-dimethylbenzoic acid (1f) afforded a
30% yielding reaction, while 3,5-dimethylbenzoic acid
(1g) gave a comparatively lower yield of 22%. 2-Methoxy-
benzoic acid (1h) and 2,6-dimethoxybenzoic acid (1i) did
not produce any fullerenyl ester, most likely due to co-
ordination of FeCl3 to the methoxy groups. 2-Chloroben-
zoic acid (1j) gave the second best result (35% yield), but
2,6-dichlorobenzoic acid (1k) did not yield any fullerenyl
ester at all. Aliphatic carboxylic acids 1l, 1m, 1n reacted
with C60 under the same conditions to produce fullerenyl
esters 2l, 2m, 2n, respectively. Among these three sub-
strates, pivalic acid (1n) gave the lowest yield. The reason
for this result is unclear, but we consider the electronic
nature of the substrates also affects the yield.
1
characterized via H NMR, 13C NMR, IR, ESI-MS, ele-
mental analysis, and X-ray crystallography.
Scheme 1. Synthesis of Compounds 3 and 4
Electrochemical studies were performed on 3 and 4 to
determine the LUMO levels of the compounds (Table 2).
Both compounds 3 and 4 exhibited reversible reductive
waves, as observed in the measurements. The first reduc-
tion potential for both 3 and 4 in 1,2-dichlorobenzene was
ꢀ1.11 and ꢀ1.13 V vs Fc/Fcþ. The LUMO level of 3 was
formally calculated to be ꢀ3.69 eV. This value is the same
as that for the LUMO level of C60 (ꢀ3.69 eV, estimated
Attempts to use aldehydes such as benzaldehyde and
esters such as ethyl benzoate, in place of carboxylic acids,
were unsuccessful when similar reaction conditions were
(7) (a) Avent, A. G.; Birkett, P. R.; Crane, J. D.; Darwish, A. D.;
Langley, G. J.; Kroto, H. W.; Taylor, R.; Walton, D. R. M. J. Chem.
Soc., Chem. Commun. 1994, 1463–1464. (b) Birkett, P. R.; Avent, A. G.;
Darwish, A. D.; Hahn, I.; Kroto, H. W.; Langley, G. J.; O’Loughlin, J.;
Taylor, R.; Walton, D. R. M. J. Chem. Soc., Perkin Trans. 2 1997, 1121–
1126. (c) Taylor, R. Synlett 2000, 776–793. (d) Kitagawa, T.; Takeuchi,
K. Bull. Chem. Soc. Jpn. 2001, 74, 785–800. (e) Kitagawa, T.; Lee, Y.;
Hanamura, M.; Sakamoto, H.; Konno, H.; Takeuchi, K.; Komatsu, K.
Chem. Commun. 2002, 3062–3063. (f) Zhang, Y.; Matsuo, Y.; Nakamura,
E. Org. Lett. 2011, 13, 6058–6061.
(8) (a) Li, F.-B.; Liu, T.-X.; Wang, G.-W. J. Org. Chem. 2008, 73,
6417–6420. (b) Li, F.-B.; Liu, T.-X.; You, X.; Wang, G.-W. Org. Lett.
2010, 12, 3258–3261. (c) Li, F.-B.; You, X.; Wang, G.-W. Org. Lett.
2010, 12, 4896–4899. (d) Li, F.-B.; You, X.; Liu, T.-X.; Wang, G.-W.
Org. Lett. 2012, 14, 1800–1803. (e) Li, F.-B.; Zhu, S.-E.; You, X.; Wang,
G.-W. Chin. Sci. Bull. 2012, 57, in press [doi: 10.1007/s11434-012-5103-5].
3278
Org. Lett., Vol. 14, No. 13, 2012

from electrochemical measurement in 1,2-dichlorobenzene)
but lower than that of PCBM (ꢀ3.61 eV). The two electron-
withdrawing groups attached to each fullerene molecule
confer large electron affinities and low LUMO levels to
products 3 and 4. These compounds are considered to be
rare; they possess low-LUMO levels similar to C60, while at
the same time they exhibit a much higher solubility than
C60. DFT calculation (B3LYP/6-311þG** level, Support-
ing Information) revealed that the LUMO of 3 and 4 are
localized on the fullerene core, not on the carbonyl groups.
1 and 2. Due to the recent push toward using low band-
gap polymers as electron-donating materials in OPVs,
electron-accepting materials with low-lying LUMO levels,
such ascompounds 3 and 4, are expected toserve as diverse
materials in the photovoltaic research field, playing
a strategic role in the future enhancement of OPV
performance.
Table 3. OPV Device Performance for Compounds 3 and 4a
VOC
JSC
(mA/cm2)
PCE
(%)
device
compd
donor
(V)
FF
Table 2. Reduction Potentials and LUMO Levels for 3 and 4 in
Comparison with C60 and PCBMa
1
2
3
4
4
3
4
3
PBDTTT
PBDTTT
P3HT
0.74
0.69
0.46
0.46
5.4
5.0
2.2
2.1
0.34
0.32
0.39
0.29
1.3
1.1
E1/2red (V vs Fc/Fcþ)
0.40
0.28
P3HT
LUMO level
compd
E1
E2
E3
(eV)b
a Information for the device configurations are stored in the Sup-
porting Information.
C60
ꢀ1.11
ꢀ1.11
ꢀ1.13
ꢀ1.19
ꢀ1.49
ꢀ1.49
ꢀ1.50
ꢀ1.56
ꢀ1.93
ꢀ1.93
ꢀ1.94
ꢀ2.05
ꢀ3.69
ꢀ3.69
ꢀ3.67
ꢀ3.61
3
4
In summary, we have demonstrated a facile synthetic
methodology to introduce two electron-withdrawing groups
to fullerene, to obtain fullerene electron acceptors that
have low LUMO levels. The synthesis relies on FeCl3-
mediated introduction of carbonyloxyl groups by using
carboxylic acids. Inexpensive FeCl3 is used for this reac-
tion, and the reaction proceeds under mild conditions.
Although there are some limitations on the carboxylic acid
substrates due to the use of Lewis acidic FeCl3, it has been
observed that both aromatic and aliphatic carboxylic acids
are compatible with this reaction. Fullerenyl esters such as
compounds 3 and 4 are representative of soluble low-
LUMO electron acceptors for application in solution-
processed OPV devices. The use of low-LUMO acceptors
provides the opportunity to employ a wider range of donor
materials, especially those with low LUMO levels. Further
molecular design and optimization of the organic compo-
nents (i.e., phenyl groups) of the fullerenyl esters sre
underway to achieve high efficiencies in OPV devices.
PCBM
a Potentials in V vs a ferrocene/ferrocenium (Fc/Fcþ) couple were
measured by cyclic voltammetry in a 1,2-Cl2C6H4 solution containing
Bu4NþPF6ꢀ as a supporting electrolyte at 25 °Cwithascanrateof0.1Vsꢀ1
.
Glassy carbon, platinum wire, and Ag/Agþ electrodes were used as the
working, counter, and reference electrodes, respectively. b Estimated
using the following equation: LUMO level = ꢀ(4.8 þ E1/2red1) eV.9
Compounds 3 and 4 were both examined for suitability
in OPV application, via the fabrication of bulk heterojunc-
tion solar cells with poly(3-hexylthiophene) (P3HT) or
poly(4,8-bis-octoxybenzo(1,2-b:4,5-b)dithiophene-2,6-diyl-
alt-(dodecyl-thieno(3,4-b)thiophene-2-carboxylate)-2,6-diyl)
(PBDTTT)10 as the donor material in the active layer. We
summarized the device performances in Table 3. Device
1 showed respectable performance with an open-circuit
voltage (VOC) of 0.74 V, a short-circuit current density
(JSC) of 5.4 mA/cm2, and a fill factor (FF) of 0.34.
Although the power conversion efficiency (PCE, 1.3%)
was lower than that of standard OPV devices,2e the results
contained valuable information. The low-lying LUMO
(ꢀ3.55 eV) and HOMO (ꢀ5.12 eV) levels of the low
band-gap polymer, PBDTTT, are compatible with the
low-lying LUMO levels of compounds 3 and 4. This results
in a higher VOC as compared to when using P3HT. Also,
PBDTTT absorbs in the near-IR region of the solar
spectrum, contributing to the higher JSC seen in devices
Acknowledgment. We thank Prof. Eiichi Nakamura
(Univ. of Tokyo) for technical consultation. This research
is partially supported by the Japan Society for the Promo-
tion of Science (JSPS) through its “Funding Program for
Next Generation World-Leading Researchers”.
Supporting Information Available. Synthetic proce-
dures, HPLC charts, NMR spectra, CV data, molecular
orbitals, crystallographic data for 2a, 3, and 4 (CIF), and
OPV performance. This material is available free of
charge via the Internet at http://pubs.acs.org.
(9) Matsuo, Y.; Iwashita, A.; Abe, Y.; Li, C.-Z.; Matsuo, K.;
Hashiguchi, M.; Nakamura, E. J. Am. Chem. Soc. 2008, 130, 15429–
15436.
(10) (a) Liang, Y.; Wu, Y.; Feng, D.; Tsai, S.-T.; Son, H.-J.; Li, G.;
Yu, L. J. Am. Chem. Soc. 2009, 131, 56–57. (b) Hou, J.; Chen, H.-Y.;
Zhang, S.; Chen, R. I.; Yang, Y.; Wu, Y.; Li, G. J. Am. Chem. Soc. 2009,
131, 15586–15587.
The authors declare no competing financial interest.
Org. Lett., Vol. 14, No. 13, 2012
3279