Molecular engineering of C60-based conjugated oligomer ensembles: Modulating the competition between photoinduced energy and electron transfer processes

Article abstract of DOI:10.1021/jo0108313

A series of novel and soluble C₆₀-(π-conjugated oligomer) dyads were synthesized, starting from suitably functionalized oligomer precursors (i.e., dihexyloxynaphthalene, dihexyloxynaphthalene-thiophene, and dihexyloxybenzene-thiophene). A syste

Full text of DOI:10.1021/jo0108313

J . Org. Chem. 2002, 67, 1141-1152
1141
Molecu la r En gin eer in g of C60-Ba sed Con ju ga ted Oligom er
En sem bles: Mod u la tin g th e Com p etition betw een P h otoin d u ced
En er gy a n d Electr on Tr a n sfer P r ocesses
Dirk M. Guldi,* Chuping Luo, and Angela Swartz
Radiation Laboratory, University of Notre Dame, Indiana 46556
Rafael Go´mez, J ose´ L. Segura, and Nazario Mart´ın*
Departamento de Quı´mica Orga´nica, Facultad de Quı´mica, Universidad Complutense,
E-28040 Madrid, Spain
Christoph Brabec and N. Serdar Sariciftci*
Christian Doppler Laboratory for Plastic Solar Cells, Linz Institute for Organic Solar Cells (LIOS),
Physical Chemistry, J ohannes Kepler University Linz, A-4040 Linz, Austria
nazmar@quim.ucm.es
Received August 16, 2001
A series of novel and soluble C60-(π-conjugated oligomer) dyads were synthesized, starting from
suitably functionalized oligomer precursors (i.e., dihexyloxynaphthalene, dihexyloxynaphthalene-
thiophene, and dihexyloxybenzene-thiophene). A systematic change in the nature of the oligomeric
component allowed (i) tailoring the light absorption of the chromophore by shifting the ground-
state absorption from the ultraviolet to the visible region and (ii) varying the oxidation potential
of the donor. The resulting electro- and photoactive dyads were examined by electrochemical and
photophysical means. In general, both singlet-singlet energy transfer and intramolecular electron
transfer were found to take place and, most importantly, to compete with each other in the overall
deactivation of the photoexcited oligomer. The selection of polar solvents in combination with the
dihexyloxybenzene-thiophene donor shifted the reactivity from an all energy (1a ; dihexyloxynaph-
thalene) to an all electron-transfer scenario (1d , dihexyloxybenzene-thiophene). Encouraged by
the favorable electron-transfer properties of dyad 1d , we prepared photodiodes by embedding 1d
between asymmetric metal contacts, which showed external monochromatic efficiencies (IPCE) close
to 10% at the maximum absorption of the molecule.
In tr od u ction
Current trends in polymer chemistry/physics are rap-
idly moving toward smaller devices with high integration
densities. This development, that is, shifting from micro-
to nanoelectronics, implies molecular electronics/photo-
nics that exhibit full functionality on the smallest realiz-
able scale. Guided by the concepts of semiconductor
technology, the design of molecules with electronic/
photonic functions is one viable route toward organic
photovoltaic applications8-10 and molecular electronics.11-15
Oligomeric architectures have assumed paramount
importance as a diverse platform for conducting system-
atic chemistry. Importantly, owing to the wide-range
significance of these low-molecular mass building blocks,
considerable attention has focused on employing them
as references for their larger and more complex polymeric
analogues.1 Major efforts have been invested into their
examinations, for example, as model systems to (i) clarify
the relationships between substitution pattern, solid-
state structure and emission properties and other elec-
trooptical properties of conjugated materials,2 (ii) apply
quantum-chemical approaches to the determination of
optical phenomena and transport properties,3 (iii) ratio-
nalize the electrochemical behavior,4 and (iv) investigate
the photophysical properties of conjugated polymers.5 In
many aspects, well-defined monodisperse oligomers have
now advanced into strong fields on their own.6,7
(6) Segura, J . L.; Mart´ın, N. J . Mater. Chem. 2000, 10, 2403.
(7) Mart´ın, R. E.; Diederich, F. Angew. Chem., Int. Ed. 1999, 38,
1350.
(8) (a) Sariciftci, N. S.; Smilowitz, L.; Heeger, A. J .; Wudl, F. Science
1992, 258, 1474. (b) Sariciftci, N. S. Prog. Quantum Electron. 1995,
19, 131. (c) Sariciftci, N. S.; Heeger, A. J . In Handbook of Organic
Conductive Molecules and Polymers; Nalwa, H. S., Ed.; Wiley: New
York, 1996. (c) Brabec, C. J .; Sariciftci, N. S.; Hummelen, J . C. Adv.
Funct. Mater. 2001, 11, 15.
(9) Yu, G.; Gao, Y.; Hummelen, J . C.; Wudl, F.; Heeger, A. J . Science
1995, 270, 1789.
(10) (a) Granstro¨m, M.; Petritsch, K.; Arias, A. C.; Lux, A.; Ander-
sson, M. R.; Friend, M. R. Nature 1998, 395, 257. (b) Gao, J .; Yu, G.;
Heeger, A. J . Adv. Mater. 1998, 10, 692. (c) van Hal, P. A.; Christiaans,
M. P. T.; Wienk, M. M.; Kroon, J . M.; J anssen, R. A. J . J . Phys. Chem.
B 1999, 103, 4352. (d) Huynh, W. U.; Peng, X.; Alivisatos, A. P. Adv.
Mater. 1999, 11, 923. (e) Halls, J . J . M.; Arias, A. C.; Mackenzie, J . D.;
Wu, W.; Inbasekaran, M.; Woo, E. P.; Friend, R. H. Adv. Mater. 2000,
12, 498. (f) Wallace, G. G.; Dastoor, P. C.; Officer, D. L.; Too, C. O.
Chem. Innovation 2000, 30(1), 14.
(1) Van Hutten, P. F.; Krasnikov, V. V.; Hadziioannou, G. Acc. Chem.
Res. 1999, 32, 257.
(2) Garnier, F. Acc. Chem. Res. 1999, 32, 209.
(3) (a) Bre´das, J . L. Adv. Mater. 1995, 7, 263. (b) Bre´das, J . L.;
Cornil, J .; Beljonne, D.; Dos Santos, D. A.; Shuai, Z. Acc. Chem. Res.
1999, 32, 267.
(4) Heinze, J .; Tschuncky P. Electronic materials: the oligomer
approach; Mu¨llen, K., Wegner, G., Eds.; Wiley-VCH: Weinheim, 1998;
pp 479-513.
(11) Carter, F. L. Molecular Electronic Devices; Marcel Dekker: New
York, 1987; Vol. II.
(5) Ba¨ssler, H.; Schweitzer, B. Acc. Chem. Res. 1999, 32, 173.
10.1021/jo0108313 CCC: $22.00 © 2002 American Chemical Society
Published on Web 01/31/2002

1142 J . Org. Chem., Vol. 67, No. 4, 2002
Guldi et al.
As far as the performance of organic photovoltaic
devices (i.e., a conjugated polymer/fullerene blend) is
concerned, the overriding principle is an intermolecular
electron transfer from the photoexcited polymer to the
electron accepting fullerene. A major shortcoming is the
tendency, especially that of pristine C60, to phase-
separate and subsequently to crystallize. This imposes
important limitations on the solubility of C60 within a
conjugated polymer matrix.
Uniformity and high quality of conjugated polymer/
fullerene thin films are, however, essential requisites for
device fabrication. Thus, different synthetic strategies
have been refined to overcome the problem of phase
separation. In the earlier years, the design of soluble
fullerene derivatives was pursued to secure control over
the morphology of the interpenetrating networks: Homo-
geneous and stable blends, containing more than 80 wt
% fullerenes were manufactured by using functionalized
fullerene derivatives rather than pristine fullerene ma-
terials.10 Succeeding this area, conjugated-polymer struc-
tures, bearing covalently linked fullerene moieties (i.e.,
in-chain versus on-chain), of different complexity emerged
as viable alternatives for fabricating stable bicontinuous
networks. More recently, the benefits of monodisperse
conjugated-oligomers, eminently easier to handle and to
characterize, have been acclaimed through the develop-
ment of C60-(π-conjugated oligomer) donor-acceptor
arrays.
In recognition of this importance, several donor-
acceptor arrays16,17 (i.e., π-conjugated oligomer/fullerene)
have been probed in time-resolved and steady-state
photolytic experiments. One of the most striking features
of C60-(π-conjugated oligomer) systems is that their
excited-state dynamics are dominated by a competition
between rapid energy and electron-transfer reactions.
The outcome of this contest depends largely on the nature
of the π-conjugated oligomeric system, as well as on the
number of repeating units in the oligomer. Importantly,
contributions stemming from an energy transfer lower
the photovoltaic conversion efficiency.
To understand better the relationship between electron
and energy transfer in C60-(π-conjugated oligomer)
ensembles, we have developed several new synthetic
model systems (1a -d ). In particular, different arylene
units ranging from naphthalene and benzene to thiophene
have been incorporated into a π-conjugated backbone as
singlet excited-state electron donors. A covalent linkage
juxtaposes these donors to a fullerene moiety as the
acceptor. With these new building blocks, the ground-
state as well as the excited-state properties of the
resulting electroactive dyads were systematically fine-
tuned. The incentives are 3-fold: First, to impact the
overall thermodynamics (i.e., facilitate or retard intra-
molecular electron-transfer processes); second, to influ-
ence the competition between the dominant singlet-
singlet energy transfer and electron-transfer events (i.e.,
shifting the deactivation of the photoexcited chromophore
from an energy to an electron-transfer scenario); and
third, to exploit the photovoltaic properties of these new
donor-acceptor structures.
Resu lts a n d Discu ssion
Syn th esis. Among the suitable protocols for the func-
tionalization of fullerenes, the 1,3-dipolar cycloaddition
of azomethine ylides to C60 is an important methodology
for the production of stable fullerene derivatives.18 To
date, a broad spectrum of intriguing fulleropyrrolidine
derivatives, bearing electro- and/or photoactive addends,
have been synthesized via this procedure.19 Therefore,
we targeted at the synthesis of conjugated oligomers with
an aldehyde group, enabling us to adapt this synthetic
methodology for the production of C60-(π-conjugated
oligomer) dyads.
Scheme 1 summarizes the synthetic pathways and
reaction conditions, which were selected en route to the
required conjugated systems. Specifically, Wittig reaction
of aldehydes 3b20 and commercially available 8b with bis-
[(triphenylphosphonium)methyl]naphthalene 220 and bis-
[(triphenylphosphonium)methyl]benzene 1421a afforded
the corresponding dibromosubstituted compounds 5,20
10,21b and 16.
As the reaction of semistabilized triphenylphospho-
nium benzylides with aldehydes affords mixtures of (Z)-
and (E)-alkenes nonstereospecifically,22 5, 10, and 16
were obtained as mixtures of trans-trans isomers en-
riched with the corresponding cis-trans and cis-cis
analogues. This required isomerization of the reaction
mixture to the all-trans isomers via reflux in xylene,
containing a catalytic trace of iodine. The enhanced
solubility of these compounds, stemming mainly from the
flexible side chains, allowed monitoring of this aspect by
3
high-resolution NMR. In particular, a J value of 16 Hz
of the vinylic protons relates to the trans-trans configu-
ration. Moreover, the 1H NMR spectra recorded after
completion of the isomerization reaction is much simpler
than that of the starting isomeric mixture. A similar
simplification of the 13C NMR spectra was observed after
isomerization.
The Rosenmund von Braun reaction23 of the dibromo
derivatives 5, 10, and 16, namely, the treatment with
copper cyanide in dry N,N-dimethylformamide in the
presence of a catalytic amount of sodium iodide, yielded
(12) Carter, F. L.; Siatkowski, R. E.; Wohltjen, H. Molecular
Electronic Devices; North-Holland: Amsterdam, 1988.
(13) Aviram, A. Molecular ElectronicssScience and Technology;
United Engineering Trustees: New York, 1989.
(14) Hong, F. T. Molecular ElectronicssBiosensors and Biocomput-
ers; Plenum Press: New York, 1989.
(15) J ortner, J .; Ratner, M. In Molecular Electronics; Blackwell
Science Ltd: Oxford, 1997.
(16) Segura, J . L.; Go´mez, R.; Mart´ın, N.; Luo, C.; Guldi, D. M.
Chem. Commun. 2000, 701.
(17) (a) Nierengarten, J .-F.; Eckert, J .-F.; Nicoud, J .-F.; Ouali, L.;
Krasnikov, V.; Hadziioannou, G. Chem. Commun. 1999, 617. (b) Eckert,
J .-F.; Nicoud, J .-F.; Nierengarten, J .-F.; Liu, S. G.; Armaroli, N.; Ouali,
L.; Krasnikov, V.; Hadziioannou, G. J . Am. Chem. Soc. 2000, 122, 7467.
(c) Armaroli, N.; Barigelleti, F.; Ceroni, P.; Eckert, J .-F.; Nicoud, J .-
F.; Nierengarten, J .-F. Chem. Commun. 2000, 599. (d) Yamashiro, T.;
Aso, Y.; Otsubo, T.; Tang, H.; Harima, Y.; Yamashita, K. Chem. Lett.
1999, 443. (e) Knorr, S.; Grupp, A.; Mehring, M.; Grube, G.; Effen-
berger, F. J . Chem. Phys. 1999, 110, 3502. (f) Peeters, E.; van Hal, P.
A.; Knol, J .; Brabec, C. J .; Sariciftci, N. S.; Hummelen, J . C.; J anssen,
R. A. J . J . Phys. Chem. B 2000, 104, 10174. (g) Martini, I.; Ma, B.; da
Ros, T.; Helgeson, R.; Wudl, F.; Schwartz B. J . Chem. Phys. Lett. 2000,
327, 253.
(18) Prato, M.; Maggini, M. Acc. Chem. Res. 1998, 31, 519.
(19) (a) Mart´ın, N.; Sa´nchez, L.; Illescas, B.; Pe´rez, I. Chem. Rev.
1998, 98, 2527. (b) Imahori, H.; Sakata, Y. Adv. Mater. 1997, 9, 537.
(c) Imahori, H.; Sakata, Y. Eur. J . Org. Chem. 1999, 2445.
(20) Segura, J . L.; Mart´ın, N.; Hanack, M. Eur. J . Org. Chem. 1999,
643.
(21) (a) Pang, Y.; Li, J .; Hu, B.; Karasz, E. Macromolecules 1999,
32, 3946. (b) Go´mez, R.; Segura, J . L.; Mart´ın, N. J . Org. Chem. 2000,
65, 7501.
(22) (a) J ohnson, A. W. Ylides and Imines of Phosphorous; J ohn
Wiley & Sons: New York, 1993; Chapters 3 and 9. (b) Wennerstro¨m,
O.; Raston, J .; Sundahl, M.; Tanner, D. Chem. Scr. 1987, 27, 567.
(23) Ellis, G. P.; Romney-Alexander, T. M. Chem. Rev. 1987, 87, 779.

C60-Based Conjugated Oligomer Ensembles
J . Org. Chem., Vol. 67, No. 4, 2002 1143
Sch em e 1
a mixture of the monocyano (6, 11, and 17) and dicyano
derivatives (12 and 19). These were subsequently sepa-
rated by column chromatography. Selective reduction of
the nitrile groups using diisobutylaluminum hydride24
yielded monaldehydes 7, 13, and 18 almost quantita-
tively. Although the direct formylation of the bromod-
erivatives, by using the Bouveault method,25 was also
employed, the two-step reaction sequence involving the
Rosenmund von Braun reaction, followed by treatment
with DIBAL-H (see Scheme 1), proved to be more
efficient. Two factors are responsible for this trend: (i)
the global yield was higher and (ii) the cyano derivatives
were easier to separate than the aldehydes.
Unsubstituted π-conjugated systems 4,20 9, and 15,
which are reference compounds for the electrochemical
and photophysical studies, were synthesized by a similar
Wittig reaction involving the bis(triphenylphosphonium)
salts 2 and 14, formylnaphthalene 3a ,20 and formyl-
thiophene 8a , respectively (Scheme 1). Iodine-catalyzed
thermal isomerization was also deemed necessary to
obtain pure all-trans isomers.
addition to 2-bromo-6-formyl-1,5-dihexyloxynaphthalene
(3b),20 which was employed in the synthesis of dyad 1a .
Treating monoaldehydes 3b, 7, 13, and 18 with C60 and
excess sarcosine in refluxing toluene for 24 h yielded
dyads 1a -d (Scheme 2). Due to the presence of the long
alkoxy substituents, these dyads are highly soluble in
common organic solvents, and complete spectroscopic
characterization was achieved.
1
In CDCl3 solutions, the H NMR spectra of the com-
pounds investigated reveal that the stereochemistry of
the vinylenic double bonds remains unaffected after the
cycloaddition reaction. The fulleropyrrolidine signals
appear, for instance, at around 5.1 ppm (CH) as a singlet
and a set of two doublets (CH2) with coupling constants
around J ) 9.8 Hz, integrating for one hydrogen. On the
other hand, the 13C NMR spectra show the signals
originating from the conjugated oligomer moiety as well
as from the fullerene system. Specifically, in addition to
the characteristic fullerene’s sp2 region another four
signals, which correspond to the sp3 carbon atoms of the
fulleropyrrolidine ring, were noted between 70 and 80
ppm and further complemented by those of the hexyloxy
chains between 70 and 15 ppm.
The monoaldehydes 7, 13, and 18 were selected as
starting materials for the preparation of dyads 1a -d , in
Electr och em istr y. The electrochemical features of
dyads 1a -d were probed by cyclic voltammetry at room
temperature. The redox potentials are collected in Table
(24) Marshall, J . A.; Andersen, W. H.; Schrider, J . W. J . Org. Chem.
1970, 35, 858.
(25) Meiler, D.; Semmelhack, M. F. Org. Synth. 1986, 64, 114.

1144 J . Org. Chem., Vol. 67, No. 4, 2002
Guldi et al.
Sch em e 2
Ta ble 1. Red ox P oten tia ls a n d UV-Vis Ma xim a of
the first three one-electron reduction steps of the fullerene
cores. These reduction potential values are shifted to
more negative values relative to that of pristine [60]-
fullerene. The underlying cathodic shift stems from the
saturation of a double bond of the C60 core, which,
accordingly, raises the lowest unoccupied molecular
orbital (LUMO) energy of the resulting fullerene deriva-
tive.27 A closer inspection of the data reveals that the first
reduction potential values of these dyads are all similar.
Furthermore, they compare quite well with that of the
fulleropyrrolidine reference (see Figure S1).
F u ller en e-Ba sed Dya d s 1 a n d Refer en ce Com p ou n d s
a
b
compd
E1
E1
E2
E3
E4
pc
λ
max (nm)
pa
pc
pc
pc
20a
1a
4
1b
9
1c
15
1d
21
1.41
1.49
1.31
1.44
1.20
1.30
1.01
1.08
1.32
328
328
413
422
412
432
412
422
322
-0.62
-0.66
-0.65
-1.09
-1.04
-1.02
-1.62
-1.66
-1.69
-1.94
-0.65
-0.71
-1.07
-1.14
-1.64
-1.75
a
b
The redox behavior of the conjugated π-systems shows
an increase of the donor strength, when going from
dihexyloxynaphthalene (20a : Eox ) 1.41 V) to the tri-
meric analogue (4: Eox ) 1.31 V), and also upon replacing
the peripheral naphthalene units in the trimer by the
π-excedent thiophene ring (9: Eox ) 1.20 V). The best
pa: anodic peak. pc: cathodic peak. In dichloromethane
solutions using Bu4NClO4 (0.3 mg L-1, SCE reference electrode
and glassy carbon as working electrode).
1 and compared with those of the unsubstituted π-con-
jugated (4, 9, 15, and 2020) and fulleropyrrolidine (21)
references.26
As a general feature, dyads 1a -d give rise to three
(27) (a) Suzuki, T.; Maruyama, T.; Akasaba, T.; Ando, W.; Koba-
yashi, K.; Nagase, S. J . Am. Chem. Soc. 1994, 116, 1359. (b) Chlistouff,
J .; Cliffel, D.; Bard, A. J . In Handbook of Organic Conductive Molecules
and Polymers; Nalwa, N. S., Ed.; J ohn Wiley & Sons: New York, 1997;
Vol. 1, Chapter 7. (c) Echegoyen, L.; Echegoyen, L. E. Acc. Chem. Res.
1998, 31, 593.
quasireversible one-electron reduction waves that reflect
(26) Williams, R. M.; Zwier, J . M.; Verhoeven, J . W. J . Am. Chem.
Soc. 1995, 117, 4093.

C60-Based Conjugated Oligomer Ensembles
J . Org. Chem., Vol. 67, No. 4, 2002 1145
F igu r e 2. UV-vis spectra of dyad 1d along with references
15, 16, and 21 measured in dichloromethane.
F igu r e 1. Transient absorption spectrum (visible-near-
infrared part) recorded 20 ps (dashed line) and 5000 ps (solid
line) upon flash photolysis of reference 9 (2.0 × 10-5 M) at
355 nm in deoxygenated toluene, indicating the oligomer
singlet-singlet (λmax ) 850 nm) and triplet-triplet features
(λmax ) 580 nm), respectively.
nm) are governed by the disappearance of the singlet-
singlet absorption, the simultaneous formation of new
maxima was noted in the blue region (Figure 1; 5000 ps
time delay). Particularly important is that the time
constant of the near-infrared-decay (9: 6.0 × 108 s-1) is
identical with that of the visible-growth (9: 5.7 × 108
s-1). From this, we conclude that a spin-forbidden inter-
system crossing (ISC) to the energetically lower-lying
triplet excited-state prevails.29
Dyads. Steady-State Tech n iqu es. Absor ption Spec-
tr a . The absorption spectra of dyads 1a -d in dichlo-
romethane solutions reveal significant deviations from
a simple superposition of the spectra of the individual
chromophores (i.e., C60 and the conjugated π-systems).
As a representative example, the UV-vis spectrum of
dyad 1d is shown in Figure 2, together with that of
references 15, 16, and 21. Unequivocally, the maximum
observed at 412 nm (15) differs from those of 1d (422
nm) and 16 (420 nm). A similar lack of superimposition
was found in toluene, THF and benzonitrile solutions.
This further supports the notion (i.e., see electrochem-
istry) that electronic perturbation exists between the
individual chromophores in their ground-state configu-
ration.
Em ission Sp ectr a . To shed light on the extent of the
electronic interaction between the conjugated π-systems
and the fullerene fragment in the excited state, emission
measurements were carried out with dyads 1a -d in
solvents of different polarity and compared to those of
model compounds 4, 9, 15, and 20b.
In general, the strong emission of the reference com-
pounds (i.e., between 54% and 72%) is quenched in all
dyads investigated with values usually between 0.03%
and 0.5% (Table 2). It is important to note that, despite
the weak fluorescence, the emission patterns of the
conjugated π-systems remain unchanged and are not
impacted noticeably by the presence of the attached
fullerene cores.
results, in terms of oxidation, were observed for the
π-conjugated system, in which the central naphthalene
moiety is replaced by a phenylene unit (15: Eox ) 1.01
V). This trend is in good agreement with previous studies
on conjugated oligomers containing naphthalene, thio-
phene, and phenylene units.26,28 Interestingly, the pres-
ence of the large naphthalene system induces some steric
interactions that presumably lead to a decrease in the
conjugation of these systems. In turn, this appears as a
sound rationale for the less positive oxidation potential
value observed upon replacing the dihexyloxynaphtha-
lene system by the less sterically demanding dihexhyl-
oxybenzene analogue.
Similar trends were noted for the irreversible oxidation
processes in the corresponding dyads. However, the
potential values are slightly anodically shifted relative
to the parent, unsubstituted oligomers. From this shift,
we conclude that small, but notable, electronic interac-
tions prevail between the donating π-conjugated system
and the electron acceptor C60.
P h otop h ysica l P r op er ties. Oligom er Refer en ces.
Oligomer references 4, 9, 15, and 20b are all strong
fluorophores (Figure S2 and Table S1, Supporting Infor-
mation), which renders them convenient probes for the
photophysical measurements with fullerene-containing
dyads (1a -d ).
The singlet excited states of 4, 9, 15, and 20b display,
besides their strong emissive features in the 400-550 nm
range, characteristic singlet-singlet transitions in the
far-visible/near-infrared (Table S1, Supporting Informa-
tion). For example, photoexcitation of 9 led to the
instantaneous formation of a strong maximum in the
800-900 nm region (Figure 1; 20 ps time delay). We
ascribe this to the singlet excited-state absorption fea-
tures of the conjugated π-system. Typically, these tran-
sients are short-lived with decay dynamics that obey a
monoexponential rate law and rates on the order of 109
s-1 (Table S1, Supporting Information). While the ab-
sorption changes in the near-infrared region (800-900
By inspecting the fluorescence data in depth a number
of important trends can be appreciated, which shall be
(29) The spectral features of the one-electron oxidized products were
characterized by pulse radiolytically oxidation experiments. This
technique was chosen since it is known to be one of the most powerful
tools to investigate reactive intermediates, evolving from the selective
one-electron transfer, for example, to [CH2Cl2]•+ and •OOCH2Cl/•-
OOCHCl2 peroxyl radicals in oxygenated dichloromethane solutions.
Under aerobic conditions, pulse radiolysis of the tested references led
to differential absorption spectra, which are characterized by new
maxima in the 600-700 nm region.
(28) (a) Do¨ttinger, S. E.; Hohloch, M.; Segura, J . L.; Steinhuber, E.;
Hanack, M.; Tompert, A.; Oelkrug, D. Adv. Mater. 1997, 9, 233. (b)
Hohloch, M.; Segura, J . L.; Do¨ttinger, S. E.; Hohnholz, D.; Steinhuber,
E.; Spreitzer, H.; Hanack, M. Synth. Met. 1997, 84, 319.

1146 J . Org. Chem., Vol. 67, No. 4, 2002
Guldi et al.
Ta ble 2. P h otop h ysica l Da ta of F u ller en e-Ba sed Dya d s 1
1a 1b
compd
21
1c
1d
fluorescence
maxima
(nm)
Φfluoresc
oligomer
(toluene)
Φfluoresc
oligomer
(THF)
Φfluoresc
fullerene
(toluene)
Φfluoresc
fullerene
(THF)
715a
715a
715b
715a
715b
715a
715b
715a
715b
715b
5.4 × 10-3 c
4.4 × 10-3 c
4.98 × 10-4
4.59 × 10-4
e
2.7 × 10-3 d
4.1 × 10-3 d
4.68 × 10-4 d
3.56 × 10-4 d
2.9 × 1011
7.6 × 10-4 d
5.6 × 10-4 d
6.0 × 10-4
2.9 × 10-4
5.4 × 1011
5.4 × 10-4 d
3.6 × 10-4 d
6.0 × 10-4 d
0.31 × 10-4 d
5.4 × 1011
1.24 ns
6.0 × 10-4
6.0 × 10-4
kenergy
(s-1
)
(toluene)
ISC
(toluene)
ISC
(BZCN)
τfluoresc
fullerene
(BZCN)
triplet
1.35 ns
1.39 ns
1.38 ns
1.2 ns
1.2 ns
1.32 ns
1.21 ns
1.21 ns
1.21 ns
1.22 ns
1.29 ns
1.39 ns
0.98 ns
700
700
700
700
700
maxima
(nm)
Φtriplet
fullerene
(toluene)
Φtriplet
fullerene
(THF)
0.98
0.95
0.95
0.83
0.75
0.72
0.64
0.59
0.57
0.98
0.45
0.33
0.98
radical pair
radical pair
Φtriplet
fullerene
(BZCN)
a
b
d
In toluene. In THF. c Excitation wavelength 334 nm. Excitation wavelength 410 nm. e No reference available.
discussed in the following. Even in nonpolar toluene the
oligomer emission in dyads 1a -d is quenched by up to 3
orders of magnitude. Instead, the familiar fullerene
fluorescence spectrum appears in the near-infrared with
a strong *0-0 emission at 715 nm,30 despite almost
exclusive excitation of the oligomer moiety at 410 nm (see
Figure 3a). For example, the absorption ratio at 410 nm
between the fullerene and oligomer moieties in dyads 1c
and 1d are 1.5/8.5 and 0.7/9.3, respectively, based on the
absorbancies of the components at 410 nm.31,32
is given in Figure 4, displaying the spectra of trimer 4
and dyad 1b in different solvents. It should be noted that
the maxima further resemble the ground-state transi-
tions. Taking this evidence into account we reach the
conclusion that the origin of excited-state energy is that
of the conjugated π-systems. Furthermore, it allows to
attribute the underlying mechanism: An efficient in-
tramolecular transfer of singlet excited state energy from
the photoexcited oligomer to the covalently linked fullerene
governs the photophysics of the C60-based arrays.
To unravel, and likewise to test, the mechanism of
producing the fullerene emission, a decisive excitation
spectrum of the 715 nm emission was taken. In fact, the
excitation spectra of dyads 1a-d were reasonable matches
of those recorded for the reference models. An illustration
C60-(π-conjugated oligomer)
9
hν8
C60-1*(π-conjugated oligomer) f
1*(C60)-(π-conjugated oligomer) (1)
(30) On a different note, it should be considered that, at least, in
THF and benzonitrile the fullerene singlet excited state is still powerful
enough to guarantee activation of an intramolecular electron-transfer
event. Thus, a scenario, in which the singlet-singlet transfer, as it
dominates the deactivation in toluene, is followed by an intramolecular
electron transfer evolving from the fullerene singlet excited state can
principally not be ruled out. It remains, however, questionable at this
point, if a 95% emission quenching, as observed in the 1,4-dihexyloxy-
benzene/thiophene based dyad and triad, is an accord with the small
-∆GCS° value (+0.14 in THF and +0.27 in benzonitrile).
From the quantum yields (Φ) of the fluorescence and
the lifetime (τ) of photoexcited 4, 9, 15, and 20b (X), the
rate constant (k) for the intramolecular singlet-singlet
energy transfer in the investigated dyads (Y) can be
calculated according to the following expression:
k ) [Φ(X) - Φ(Y)]/[τ(X) Φ(Y)]
(2)
(31) Conclusive evidence for the fullerene emission came from
experiments that make use of the heavy ion effect for a controlled
acceleration of the ISC process between the fullerene singlet and triplet
excited states. Addition of ethyl iodide, a heavy atom provider, led to
a complete cancellation of this fullerene fluorescence and, instead, the
characteristic 824 nm fullerene phosphorescence was detected.32
(32) (a) Zeng, Y.; Biczok, L.; Linschitz, H. J . Phys. Chem. 1992, 96,
5237. (b) Guldi, D. M.; Asmus, K.-D. J . Phys. Chem. A 1997, 101, 1472.
(c) Guldi, D. M.; Maggini, M. Gaz. Chim. Ital. 1997, 127, 779.
The rate constants are listed in Table 3.33 Remarkably,
the extremely fast rates (1011-1012 s-1), as they were
(33) In this context, we like to emphasize that our previously
reported values are considered to be too small, a fact that stems from
the time resolution of our picosecond apparatus. The latter allowed
us only to detect the final fraction of the energy transfer event.

C60-Based Conjugated Oligomer Ensembles
J . Org. Chem., Vol. 67, No. 4, 2002 1147
Ta ble 3. Dr ivin g F or ce Dep en d en ce (-∆GCR°; -∆GCS°)
a n d Th er m od yn a m ic P a r a m eter s (λ; ∆GCSq) for
In tr a m olecu la r Electr on -Tr a n sfer Even ts in
F u ller en e-Ba sed Dya d s 1
compd
1a
1b
1c
1d
radius donor; R+ (Å)
donor-acceptor separation;
2.45
8.2
11.35
16.0
8.5
12.8
7.5
11.1
R
D-A (Å)
-∆GCR
-∆GCR
-∆GCR
-∆GCS
oligomer
°
°
°
a (toluene) (eV)
a (THF) (eV)
2.89
2.03
1.75
0.65
2.46
2.04
1.90
0.43
2.29
2.02
1.86
1.74
0.62
1.62
1.49
0.68
a (BZCN) (eV)
b (toluene) (eV)
°
-∆GCS
-∆GCS
-∆GCS
°
°
°
b (THF) (eV)
1.51
1.79
-1.13
0.85
0.99
-0.70
1.05
1.17
-0.53
1.08
1.21
-0.26
b (BZCN) (eV)
b (toluene) (eV)
fullerene
-∆GCS
°
°
b (THF) (eV)
-0.27
0.01
0.39
1.43
1.47
-0.28
-0.14
0.34
0.85
0.87
-0.10
0.02
0.34
0.85
0.86
0.06
0.14
0.27
0.34
0.82
0.84
0.09
-∆GCS
b (BZCN) (eV)
λ
λ
λ
c (toluene) (eV)
c (THF) (eV)
c (BZCN) (eV)
q d
∆GCS (toluene) (eV)
0.046
0.006
oligomer
∆GCSq d (THF) [eV]
0.001
0.017
0
0.012
0.027
0.03
0.05
q d
∆GCS (BZCN) (eV)
0.004
fullerene
q d
∆GCS (THF) (eV)
0.14
0.09
q d
∆GCS (BZCN) (eV)
0.21
a
Determined from the following relations:
-∆GCR° ) E1/2(D•+/D) - E1/2(A/A•-) + ∆GS
e2
1
1
1
1
1
1
1
∆GS
)
+
-
-
+
(
[
)
(
)
]
4πꢀ0 2R+ 2R- RD-A ꢀS
2R+ 2R- ꢀR
E1/2(D•+/D) ) oxidation potential of donor (see Table 1)
F igu r e 3. Emission spectra of dyads 1c and 1d in toluene
(a) and THF (b) with matching absorption at the 410 nm
excitation wavelength (i.e., OD410nm ) 0.5).
E
1/2(A/A•-) ) reduction potential of acceptor (see Table 1)
R+ ) radius donor (see the top of this table)
RD-A ) donor-acceptor separation
(see the top of this table)
R- ) radius acceptor (4.4 Å)
ꢀS ) solvent dielectric constant (toluene ) 2.39;
THF ) 7.6; benzonitrile ) 24.8)
ꢀR ) solvent dielectric constant for electrochemical
measurements (9.412)
Determined from the following relation:
-∆GCS° ) ∆E0-0 - (-∆GCR
∆E0-0 ) excited-state energy of the chromophore
b
)
c Determined from the following relations:
λ ) λi + λS
F igu r e 4. Excitation spectra (normalized to exhibit matching
signals around 400 nm) of dyad 1b in toluene (dotted line) and
THF (solid line) and reference 4 in toluene (solid line),
monitoring the fullerene emission at 715 nm and oligomer
emission at 450 nm, respectively.
e2
1
1
1
1
1
ꢀS
λS )
+
-
+
2
(
[
) (n
)
]
4πꢀ0 2R+ 2R- RD-A
λi ) internal reorganization energy (0.3 eV)
λS ) solvent reorganization energy
determined via this relation, infer a strong coupling
between the two moieties, a hypothesis that appears
reasonable bearing in mind the appreciable perturbation
seen in the ground-state absorption features (vide supra).
By employing reference 21 as a standard, with match-
ing absorption at the excitation wavelength, the fullerene
fluorescence quantum yields were deduced for the dif-
ferent donor-acceptor ensembles. In toluene, Φ values
nearly approach that measured for the reference 21,
namely, 6.0 × 10-4, and confirm the efficient transfer of
singlet excited-state energy.
n ) solvent refractive index (toluene ) 1.496;
THF ) 1.407; benzonitrile ) 1.528)
d
determined from the following relation:
(∆GCS + λ)2
q
∆GCS
)
4λ
∆Gq ) Gibbs activation energy
Next, we shift the focus to a more polar solvent (i.e.,
THF). To facilitate the discussion and interpretation, we

1148 J . Org. Chem., Vol. 67, No. 4, 2002
Guldi et al.
group dyads 1a -d into two different categories. The first
class comprises the dihexyloxynaphthalene system (20,
1a ) and its trimeric analogue (4, 1b), both displaying high
oxidation potentials, while the lower potentials of di-
hexyloxynaphthalene-thiophene (9, 1c) and dihexyloxy-
benzene-thiophene (15, 1d ) lead us to deal with them
separately.
Similar to the trends seen for toluene, the emission of
the conjugated π-system in dyads 1a -d in THF is
strongly quenched compared to the emission of the
corresponding references, with quantum yields that are
somewhat smaller than in toluene. By contrast, the
fullerene fluorescence quantum yields deviate substan-
tially from this tendency (see Figure 3b). Considering the
quantum yields in the first group, the THF values diverge
moderately by ∼10% and ∼22% from the corresponding
toluene values for the monomeric (1a ) and trimeric
dihexyloxynaphthalene (1b), respectively. Substantial
effects were, however, seen in the second category. In
particular, the differences between the toluene and THF
samples were 57% (dihexyloxynaphthalene-thiophene,
1c) and 95% (dihexyloxybenzene-thiophene, 1d ). These
values testify to the presence of yet another second
pathway, competing with the energy transfer in the
overall deactivation of the photoexcited chromophore. At
first glance, a highly exothermic electron-transfer sce-
nario could be considered as a conceivable rationale for
the absence of the fullerene emission.
F igu r e 5. Transient absorption spectrum (visible-near-
infrared part) recorded 20 ps (dashed line) and 5000 ps (solid
line) upon flash photolysis of dyad 1d (∼5.0 × 10-5 M) at 355
nm in deoxygenated toluene, indicating the fullerene singlet-
singlet (λmax ) 900 nm) and triplet-triplet features (λmax
)
700 nm), respectively.
to the triplet excited state and finally to the singlet
ground state.
The picture is quite different for the dihexyloxyben-
zene-thiophene-based dyad (1d ). In fact, the moderate
oxidation potential of the dihexyloxybenzene-thiophene
donor shifts the energy of the charge-separated state in
both THF and benzonitrile markedly below that of the
fullerene singlet excited state. Hence, it is likely that
charge separation plays the dominant role, an assump-
tion, which is consistent with the fluorescence data. The
latter showed a complete abolishment of the fullerene
fluorescence in THF and benzonitrile (Table 3). The
energy levels, as extracted from Tables 1 and 3, are
shown in Scheme S1a (toluene) and S1b (benzonitrile)
(Supporting Information) to illustrate the different re-
laxation pathways of the photoexcited states.
C60-(π-conjugated oligomer)
9
hν8
C60-1*(π-conjugated oligomer) f
(C60•-)-(π-conjugated oligomer•+) (3)
Another plausible alternative implies the following
sequential energy and electron-transfer events:
C60-(π-conjugated oligomer)
9
hν8
C60-1*(π-conjugated oligomer) f
1*(C60)-(π-conjugated oligomer) f
(C60•-)-(π-conjugated oligomer •+) (4)
Tim e-Resolved Tech n iqu es. The conclusion of the
emission studies along with the thermodynamic calcula-
tions presumes that an intramolecular electron-transfer
scenario competes with an intramolecular energy transfer
led us to characterize the fullerene-based hybrids by
time-resolved transient absorption spectroscopy in dif-
ferent solvents.
Tr a n sien t Sp ectr oscop y in Non p ola r Tolu en e. In
nonpolar toluene, the fluorescence experiments reveal the
same reactivity for all dyads, which led us to exemplify
the transient features by describing dyad 1d in detail.
Transient absorption changes, recorded upon 18 ps
laser excitation of the investigated dyads, are superim-
posable with those recorded for the fullerene model (21),
despite the predominant excitation of the conjugated
π-systems. The absorption ratiosfullerene versus
oligomersat 355 nm in dyad 1d is, for example 3.3/6.7.
In Figure 5 the differential changes for 1d are shown
after a 20 ps time delay. Instead of the singlet-singlet
absorption of the oligomer units (see for illustration
Figure 1), broad maxima, which correspond to the
fullerene singlet-singlet absorption, were found around
900 nm. Similar spectra were noted for dyads 1a -c in
toluene.
Th er m od yn a m ic Con sid er a tion s. Looking at the
emission quenching and, more importantly, at the vary-
ing donor strength of 4, 9, 15, and 20b, we gathered that
an intramolecular electron transfer, evolving from the
excited chromophore to the fullerene moiety, may com-
pete with an energy transfer. As a test to validate the
above conclusion, the driving forces (-∆GCS°), associated
with an intramolecular charge separation, were calcu-
lated in accordance with the approximations given in
Table 3.34
The model gives rise to large -∆GCS° values in toluene,
THF, and benzonitrile (Table 3) for 1a -c. Considering,
however, the energy of the fullerene singlet excited state
of 1.76 eV, relative to those of the charge-separated state,
it is clear that the largest energy gap opens between the
two singlet excited states (i.e., oligomer and fullerene).
Consequently, the energy-transfer route is the thermo-
dynamically favored pathway. From this we can assume
that after the completion of the singlet-singlet energy
transfer the fullerene singlet excited state can only decay
Time profiles, taken at various wavelengths, indicate
that the singlet excited state is formed instantaneously
and in a single step. The formation dynamics are masked
(34) Atkins, P. W.; MacDermott, A. J . J . Chem. Educ. 1982, 59, 359.

C60-Based Conjugated Oligomer Ensembles
J . Org. Chem., Vol. 67, No. 4, 2002 1149
F igu r e 6. Transient absorption spectrum (visible-near-
infrared part) recorded 50 ns upon flash photolysis of dyad
1d (2.0 × 10-5 M) at 355 nm in deoxygenated toluene (dashed
line) and deoxygenated benzonitrile solutions (solid line),
indicating the fullerene triplet-triplet (λmax at 360 and 700
nm) and the charge-separated radical pair features (λmax at
680 and 1000 nm), respectively.
F igu r e 7. Transient absorption spectrum (visible-near-
infrared part) recorded 20 ps upon flash photolysis of dyad 1c
(dashed line) and 1d (solid line) (2.0 × 10-5 M) at 355 nm in
deoxygenated benzonitrile, indicating the oligomer π-radical
cation features (λmax around 700).
(20 ps) after irradiating 1c with a picosecond laser pulse.
Charge separation was inferred on the basis of the
instantaneously produced 700 nm maximum, while the
presence of the 900 nm absorption and an ISC rate of
7.1 × 108 s-1 are clear attributes connected with the
energy transfer product.
by the apparatus’s resolution (18 ps). This allowed only
estimation of the intramolecular rate constant for the
energy transfer process (k g 5.0 × 1010 s-1), which
appears credible in light of the rates derived from the
fluorescence data. Hereafter, an ISC is the predominant
deactivation of the fullerene singlet excited states. The
accordingly formed triplet excited state reveal charac-
teristic peaks around 700 nm (Figure 5; 5000 ps time
delay).
In line with the proposed energy-transfer mechanism
(eq 1), the differential absorption changes, recorded
immediately after an 8 ns pulse (Figure 6), showed the
same spectral features of the fullerene triplet excited
state as observed at the end of the picosecond experi-
ments. Precisely, two maxima located at 360 and 700 nm
and a low-energy shoulder around 800 nm were observed.
Thus, our analysis clearly shows the photosensitization
effect of the conjugated π-system chromophore, acting as
an antenna system and transmitting its excited energy
to the covalently attached fullerene moieties. The overall
triplet quantum yields (Table 2), as measured with the
comparative method (i.e., using 21 as a standard), in
toluene vary between 0.64 and 0.98 and are slightly lower
In sharp contrast to the dihexyloxybenzene-thiophene
(1d ) and dihexyloxynaphthalene-thiophene (1c) systems
are the picosecond changes of 1b and 1a . They are
overshadowed by the dominance of the fullerene photo-
physics (similar to Figure 5), namely, instantaneous
formation of the singlet excited state followed by the
subsequent ISC. A plausible explanation for this observa-
tion is the unfavorably raised energies of the charge-
separated state (1.90 eV), which in the dihexyloxynaph-
thalene systems even passed that of the fullerene singlet
(Table 3).
Full spectral characterization of the charge-separated
state was accomplished directly with the help of comple-
mentary nanosecond measurements. For example, pho-
tolysis of a benzonitrile solution of 1d led to maxima at
1000 and 680 nm (Figure 6). Based on a spectral
comparison, we ascribe the former band to the fullerene
π-radical anion, while the latter reveals the broad
characteristics of the dihexyloxybenzene-thiophene π-rad-
ical cation. In accordance with these results, we propose
the occurrence of a photoinduced electron transfer, evolv-
ing from the singlet excited state of the chromophore (i.e.,
conjugated π-system) to the electron-accepting fullerene
(eq 3).
than the one found for the reference compound (Φtriplet
0.98).
)
Tr a n sien t Sp ectr oscop y in P ola r Ben zon itr ile.
The choice of this strongly polar solvent led to a different
reactivity. In particular, the dihexyloxybenzene-thio-
phene-based dyad (1d ) gave rise, following the 18 ps laser
pulse, to the immediate occurrence of a 690 nm band, as
shown in Figure 7. In accordance with the pulse radiolytic
oxidation experiments, we assign this transient absorp-
tion to the π-radical cation of the oligomer. The fast grow-
in (e18 ps) supports the direct mechanism (eq 3), rather
than the sequential mechanism (eq 4). This reactivity
reflects the emission studies (vide supra), implying that
the energy of the charge-separated state (1.49 eV) is
below that of the fullerene singlet excited state.
Raising the oxidation potential (i.e., dihexyloxynaph-
thalene-thiophene; 1c) and, in turn, bringing the energy
of the charge-separated state (1.74 eV) close to that of
the fullerene singlet led to the detection of two photo-
products. This is illustrated in Figure 7, which displays
the transient absorption changes recorded immediately
More complicated is the situation encountered upon
raising the oxidation potential of the donor, especially
in the dihexyloxynaphthalene-based donor-acceptor en-
sembles. In these cases, the strong triplet-triplet absorp-
tion (ꢀ700nm ∼16 100 M-1 cm-1) dominates both the pico-
and nanosecond spectra. Consequently, to obtain direct
evidence in support of the weaker absorbing radical pair
(ꢀ1000nm ∼8000 M-1 cm-1) in 1a and 1b, subtraction of the
triplet absorption features from, for example, the nano-
second spectrum became necessary. The resulting dif-
ferential absorption changes reveal a near-infrared maxi-
mum at 1000 nm and a visible-doublet at 660 and 800
nm, which are characteristics of the reduced fullerene
moiety and the oxidized dihexyloxynaphthalene system
4 (see above), respectively.35

1150 J . Org. Chem., Vol. 67, No. 4, 2002
Guldi et al.
F igu r e 8. I/V characteristics of a ITO/PEDOT/1d /LiF/al diode
under 80 mW/cm2 white light (open squares) from a solar
simulator and in the dark (open circles).
F igu r e 9. Spectrally resolved photocurrent, converted to
incident photon to converted electron efficiency (IPCE) of
photodiodes from 1d .
Both fingerprints allowed us to determine the lifetime
of the associated charge-separated state. The decay
curves were well fitted by a single-exponential decay
component. In particular, lifetimes that are on the order
of a microsecond were derived from the decays of the
oxidized donor and reduced acceptor absorption at 680
and 1000 nm, respectively (see Figure S3, Supporting
Information). Additional proof for an intramolecular
charge recombination stems from a set of experiments,
conducted with different laser power and different dyad
concentrations. In particular, decreasing the radical-pair
concentration in various increments by up to 65% led to
a maximal change of the radical pair lifetime. In other
words, a contribution, stemming from an intermolecular
charge recombination process, is assumed to be insig-
nificant. In general, the lifetimes of the charge-separated
states can be correlated with the solvent polarity. In the
case of 1d (0.71 µs (THF), 1.37 µs (benzonitrile)) the
lifetimes clearly increase as the solvent polarity in-
creases.
in magnitude. The high quality of the diode further
confirms that 1d exhibits sufficiently good film forming
properties when processed from solution. This allows to
cast thin films with a closed surface. Under irradiation
with AM 1.5 solar simulated light with 80 mW/cm2, an
open circuit voltage of 0.67 V, a short circuit current of
0.75 mA/cm2, and a fill factor <0.3 were measured. Taken
all these parameters into concert, we conclude an overall
white light power efficiency of e0.2%.
In addition, the results stemming from the spectral
photocurrent measurements (Figure 9) are encouraging.
A monochromatic peak efficiency of ∼10% (IPCE) was
measured at 440 nm, directly proving the contribution
of the oligomer (see Figure 6) to the photocurrent.
Between 500 and 700 nm, the IPCE drops below 4%. In
this energy region the photocurrent generation is, how-
ever, due to a photoinduced hole transfer of the weakly
absorbing fullerene to the covalently linked oligomer unit.
Su m m a r y
Devices. Since an optimized electron transfer is an
imperative requisite for devising a high-performance
photovoltaic device, we probed the photovoltaic properties
of dyad 1d , which in solution shows an all-electron-
transfer behavior when sandwiched between asymmetric
electrodes and forming a quasi ohmic contact. In prin-
ciple, this approach gives rise to a structure similar to
that of metal-insulator-metal (MIM) diodes. Devices
that are based on such an electrode configuration are
expected to lead to an asymmetric current-to-voltage
behavior, independent of the rectifying properties of the
molecule. A reverse bias direction (i.e., electron injection
into the oligomer and hole injection into the fullerene),
for example, for 1d is energetically unfavorable. Con-
versely, a forward bias resulting in an electron injection
into the fullerene and hole injection into the oligomer is
energetically favorable.
Typical current-to-voltage characteristics of 1d , under
illumination and also in the dark, are given in Figure 8.
The dark I/V curve is symmetric in the region between
+0.5 and -0.5 V, which is, in part, due to a small shunt
resistivity of the diode leading to an ohmic contribution.
A diode turn on behavior is observed at voltages higher
than 0.7 V, with rectification ratios of more than 2 orders
The synthesis of fulleropyrrolidines, covalently linked
to fluorescent π-conjugated systems bearing long alkyl
chains, has been carried out from suitably functionalized
oligomers by 1,3-dipolar cycloaddition reaction of azome-
thine ylides to C60.
Cyclic voltammetry reveals the presence of three one-
electron reduction steps, occurring all at the fullerene
electron acceptor. On the anodic side, only a single one-
electron oxidation wave is observed, which relates to the
oxidation of the oligomeric unit. This oxidation wave is
shifted to less positive values when replacing the periph-
eral naphthalene units in the trimeric system (1b) by
thiophene rings (1c). Additionally, substitution of the
central naphthalene unit by a phenylene moiety (1d )
further increases the electron donor ability.
By controlling the structure of fluorescent π-conjugated
systems, the outcome of an ultra rapid intramolecular
deactivation of a photoexcited oligomer core was success-
fully shifted from an all energy transfer to an all electron-
transfer scenario. In particular, the different oxidation
potentials of, for example, dihexyloxynaphthalene (i.e.,
monomeric and trimeric form), dihexyloxybenzene/thio-
phene and dihexyloxynaphthalene/thiophene moieties
helped to direct the competition between these two
transfer channels. Hereby, the relative position of the
charge-separated state in reference to the fullerene
(35) This subtraction procedure hampered a meaningful analysis
of the charge recombination kinetics.

C60-Based Conjugated Oligomer Ensembles
J . Org. Chem., Vol. 67, No. 4, 2002 1151
MHz for 13C spectrometer. Chemical shifts are given as δ
values (internal standard: TMS). 2-Bromo-5-formylthiophene
(8b), 2-formylthiophene (8a ), 1,5-dihydroxynaphthalene, 1,4-
hydroquinone, sarcosine, and [60]fullerene are commercially
available and were used without further purification. Some
of the functionalized π-conjugated systems (5,20 10,21b 13,21b
and 1421a) and reference compounds (4,20 20,20 and 2126) were
obtained by following previously described synthetic proce-
dures. Tetrahydrofuran and dimethylformamide were dried
with sodium and calcium hydride, respectively, while chloro-
form and dichloromethane were distilled from CaCl2.
1,3-Dip ola r Cycloa d d ition Rea ction s. Syn th esis of
Dya d s. Gen er a l P r oced u r e. A mixture of 0.05 mmol of
monoaldehyde, 38 mg (0.05 mmol) of [60]fullerene, and 0.25
mmol of N-methylglycine (sarcosine) was dissolved in 30 mL
of toluene, and the mixture was refluxed for 24 h. After this
time, the reaction was allowed to reach room temperature, and
then the solvent was partially vacuum-evaporated and poured
on a silica gel column. The black solid obtained after chroma-
tography (cyclohexane/toluene) was further purified by repeti-
tive centrifugation using methanol and diethyl ether to yield
the corresponding dyads as black solids.
singlet excited state determines the nature of the pho-
toproduct: Only the energetically lowest lying state is
populated in high yields.
Photovoltaic devices confirm the efficient photoinduced
charge generation within one of the investigated donor-
acceptor ensembles (1d ). Spectrally resolved photocurrent
measurements revealed contributions from both the
oligomer and the fullerene moiety to the photocurrent
generation. Moreover, photovoltaic devices showed white
light efficiencies up to ∼0.2%. This is one of the highest
values ever reported for a solution processed single
component organic solar cell. In fact, charge generation,
as well as electrical current rectification, may indeed be
cooperatively realized within this molecular donor-
acceptor ensemble, thereby combining both essential
features for photovoltaics within a single molecule.
Currently, we are directing our efforts toward optimiz-
ing parameters such as the thickness and morphology of
the active layer, which are expected to lead to further
improved and higher photocurrent efficiencies.
N-Meth yl-2′-(6-br om o-1,5-d ih exyloxy-2-n a p h th yl)p yr -
r olid in o[3′,4′:1,2][60]fu ller en e (1a ). By following the above
general procedure and using monoaldehyde 3b28 as the start-
ing material, 16 mg (27%) of dyad 1a was obtained: 1H NMR
(CDCl3, 300 MHz) δ 8.20 (d, 1H, J ) 8.9 Hz), 7.97 (d, 1H, J )
8.9 Hz), 7.71 (d, 1H, J ) 9.0 Hz), 7.58 (d, 1H, J ) 9.0 Hz),
5.57 (s, 1H), 5.04 (d, 1H, J ) 9.4 Hz), 4.32 (d, 1H, J ) 9.4 Hz),
4.16-4.07 (m, 4H), 2.77 (s, 3H), 2.07-1.90 (m, 4H), 1.58-1.37
(m, 12H), 0.96 (t, 6H); 13C NMR (CDCl3, 75 MHz) δ 156.28,
155.47, 154.33, 154.03, 153.22, 152.75, 147.20, 147.17, 146.50,
146.42, 146.19, 146.11, 146.05, 146.00, 145.96, 145.86, 145.84,
145.61, 145.46, 145.41, 145.29, 145.25, 145.19, 145.11, 145.03,
142.98, 142.90, 142.54, 142.50, 142.45, 142.42, 142.18, 142.16,
142.04, 142.01, 141.91, 141.29, 141.67, 141.62, 141.49, 140.10,
140.02, 139.72, 139.54, 136.78, 136.43, 135.73, 134.85, 130.80,
130.15, 128.79, 127.88, 126.27, 119.74, 118.55, 117.46, 113.83,
77.21, 76.65, 76.37, 76.04, 74.42, 69.92, 31.74, 31.68, 30.41,
30.21, 25.96, 25.64, 22.75, 22.57, 14.17, 14.05; FTIR (KBr,
cm-1) 2920, 2850, 2774, 1620, 1581, 1427, 1407, 1361, 1331,
1177, 1024, 526; MS m/z (APCI) (rel intensity) 1184 (M+, 100),
1105 (28).
P yr r olid in o[3′,4′:1,2][60]fu ller en e 1b. By following the
above general procedure and using monoaldehyde 7 as the
starting material, 58 mg (63%) of dyad 1b was obtained: 1H
NMR (CDCl3, 300 MHz) δ 8.10 (d, 2H, J ) 9.0 Hz), 7.94 (d,
2H, J ) 8.8 Hz), 7.97-7.89 (m, 6H), 7.81-7.74 (m, 4H), 7.59
(d, 2H, J ) 8.8 Hz), 5.54 (s, 1H), 4.98 (d, 1H, J ) 9.6 Hz), 4.27
(d, 1H, J ) 9.6 Hz), 4.19-4.14 (m, 2H), 4.10-4.03 (m, 10H),
2.74 (s 3H), 1.98 (t, 12H), 1.59 (s, 12H), 1.35 (s, 24H), 0.88 (t,
18H); 13C NMR (CDCl3, 75 MHz) δ 155.62, 154.59, 154.22,
153.37, 147.27, 146.55, 146.21, 146.09, 145.94, 145.74, 145.55,
145.28, 145.21, 144.56, 144.44, 142.99, 142.59, 142.27, 142.13,
141.78, 141.60, 140.11, 139.83, 139.17, 136.85, 136.56, 135.88,
134.94, 130.44-113.55 (54C), 74.59, 70.06, 69.57, 40.15, 31.86,
31.83, 31.79, 31.75, 30.55, 30.30, 26.14, 26.10, 25.76, 22.86,
22.74, 22.66, 14.26, 14.08; FTIR (KBr, cm-1) 2922, 2853, 1462,
1406, 1340, 1176, 1042, 816, 526; MS m/z (ESI) (rel intensity)
1886 (M+ + 23, 100). Anal. Calcd for C133H100BrNO6: C, 84.60;
H, 5.34; N, 0.74. Found: C, 82.36; H, 5.47, N, 0.94.
Exp er im en ta l Section
Picosecond laser flash photolysis experiments were carried
out with 355-nm laser pulses from a mode-locked laser system
(pulse width 18 ps, 2-3 mJ /pulse). The white continuum
picosecond probe pulse was generated by passing the funda-
mental output through a D2O/H2O solution. The excitation and
the probe were fed to a spectrograph with fiberoptic cables and
were analyzed with a dual diode array detector interfaced with
a
computer. The nanosecond laser flash photolysis were
performed with laser pulses from a nitrogen laser system
(337.1 nm, 8 ns pulse width, 1 mJ /pulse). A typical experiment
consisted of 5-10 replicate pulses per measurement.
Fluorescence lifetimes were measured with a laser strope
fluorescence lifetime spectrometer with 337 nm laser pulses
from a nitrogen laser fiber coupled to a lens-based T-formal
sample compartment equipped with a stroboscopic detector.
Fluorescence spectra were measured in methylcyclohexane
solutions containing fullerene forming clear, noncracking
glasses in liquid nitrogen. In the case of any other solvents,
the experiments were performed at room temperature. A 570
nm long-pass filter in the emission path was used in order to
eliminate the interference from the solvent and stray light for
recording the fullerene fluorescence. No corrections were
performed for the fluorescence, but long integration times (20
s) and low increments (0.1 nm) were applied. The slits were 2
and 8 nm, and each spectrum was an average of at least five
individual scans.
Devices were fabricated by spin coating 1d from a 1.2%
toluene/chlorobenzene (3:2) solution until a thickness of ap-
proximately 60 nm on top of an ITO/PEDOT:PSS substrate
was achieved. A LiF/Al (6 Å/100 nm) layer was thermally
deposited as a top electrode in a two-step evaporation process
through a shadow mask to define a device area of 3.3 mm2.
Thus, the insertion of the LiF layer ensures better contacts
with low contact resistivities.36,37
Cyclic voltammograms were recorded on a potentiostat/
galvanostat in a conventional three-compartment cell equipped
with a software electrochemical analysis by using a GCE
(glassy carbon) as working electrode, SCE as reference elec-
trode, Bu4NClO4 as supporting electrolyte, a toluene-aceto-
nitrile solvent mixture (v/v 4:1), and at a scan rate of 200 mV/
s. All melting points were measured with a melting point
apparatus and are uncorrected.
P yr r olid in o[3′,4′:1,2][60]fu ller en e (1c). By following the
above general procedure and using monoaldehyde 13 as the
starting material, 26 mg (39%) of dyad 1c was obtained: 1H
NMR (CDCl3, 300 MHz) δ 7.86 (d, 1H, J ) 9.0 Hz), 7.84 (d,
1H, J ) 9.0 Hz), 7.70 (d, 2H, J ) 9.0 Hz), 7.63 (m, 1H), 7.58
(d, 2H, J trans ) 16.4 Hz), 7.39-7.33 (m, 1H), 7.26 (d, 2H, J trans
) 16.4 Hz), 7.07 (d, 2H, J ) 3.9 Hz), 5.15 (s, 1H), 5.01 (d, 1H,
J ) 9.7 Hz), 4.26 (d, 1H, J ) 9.7 Hz), 3.99 (t, 4H), 2.86 (s, 3H),
1.86 (q, 4H), 1.31 (m, 12H), 0.85 (t, 6H); 13C NMR (CDCl3, 75
MHz) δ 161.39, 154.19, 153.21, 150.64, 147.35, 147.34, 147.33,
146.75, 146.35, 146.34, 146.23, 146.20, 146.15, 146.12, 145.97,
145.95, 145.79, 145.61, 145.57, 145.49, 145.45, 145.38, 145.32,
145.29, 145.28, 145.19, 144.72, 144.59, 144.36, 144.35, 143.15,
142.98, 142.72, 142.61, 142.59, 142.22, 142.18, 142.17, 142.13,
FTIR spectra were recorded either as KBr pellets. 13C and
1H NMR spectra were recorded with a 300 MHz for H and 75
1
(36) J abbour, G. E.; Kippelen, B.; Armstrong, N. R.; Peyghambarian,
N. Appl. Phys. Lett. 1998, 73, 1185.
(37) Hung, L. S.; Tang, C. W.; Mason, M. G. Appl. Phys. Lett. 1997,
70, 152.

1152 J . Org. Chem., Vol. 67, No. 4, 2002
Guldi et al.
142.07, 142.06, 141.97, 141.93, 141.91, 141.69, 141.61, 140.19,
139.93, 139.76, 138.12, 135.85, 135.57, 130.17, 129.34, 129.02,
128.21, 127.18, 126.64, 125.45, 124.85, 123.71, 123.09, 122.93,
120.42, 118.99, 118.87, 114.51, 107.09, 77.19, 76.32, 75.98,
68.68, 50.87, 40.39, 31.86, 31.76, 30.44, 30.42, 26.18, 26.06,
22.77, 22.68, 14.24, 14.08; FTIR (KBr, cm-1) 2930, 2864, 2216,
1500, 1440, 1224, 526; MS m/z (ESI) (rel intensity) 1345 (M+
+ 1, 100), 685 (40). Anal. Calcd for C98H44N2O2S2: C, 87.48;
H, 3.30; N, 2.08; S, 4.77. Found: C, 85.82; H, 3.93, N, 1.75; S,
4.57.
P yr r olid in o[3′,4′:1,2][60]fu ller en e (1d ). By following the
above general procedure and using monoaldehyde 18 as the
starting material, 19 mg (27%) of dyad 1d was obtained: 1H
NMR (CDCl3, 300 MHz) δ 7.25 (d, 1H, J ) 3.6 Hz), 7.24 (d,
2H, J trans ) 16.3 Hz), 7. 04 (d, 2H, J trans ) 16.3 Hz), 7.14-7.13
(m, 1H), 7.01-6.97 (m, 2H), 6.94 (d, 1H, J ) 3.6 Hz), 6.77 (d,
1H, J ) 3.9 Hz), 5.15 (s, 1H), 4.92 (d, 1H, J ) 9.6 Hz), 4.18 (d,
1H, J ) 9.6 Hz), 3.94 (m, 4H), 2.87 (s, 3H), 1.76 (q, 4H), 1.32
(s, 12H), 0.85 (t, 6H); 13C NMR (CDCl3, 75 MHz) δ 155.84,
153.92, 153.16, 153.14, 151.08, 151.06, 147.33, 146.87, 146.35,
146.32, 146.30, 146.22, 146.18, 146.15, 146.01, 145.96, 145.95,
145.79, 145.58, 145.54, 145.52, 145.46, 145.36, 145.31, 145.27,
145.18, 144.94, 144.71, 144.62, 144.37, 143.16, 142.97, 142.71,
142.60, 142.58, 142.26, 142.21, 142.17, 142.15, 142.08, 142.06,
141.98, 141.95, 141.68, 141.61, 140.17, 139.92, 139.78, 139.58,
137.06, 136.63, 135.85, 135.61, 130.45, 128.73, 126.55, 126.03,
125.79, 125.20, 123.89, 123.70, 122.40, 121.54, 110.89, 110.78,
110.55, 79.57, 77.10, 70.03, 69.49, 68.81, 40.35, 31.57, 29.34,
25.86, 22.62, 14.08; FTIR (KBr, cm-1) 2932, 2860, 2788, 1497,
1461, 1420, 1334, 1202, 960, 526. Anal. Calcd for C93H42
-
BrNO2S2: C, 82.78; H, 3.14; N, 1.04; S, 4.75. Found: C, 82.12;
H, 3.38, N, 1.02; S, 4.59.
Ack n ow led gm en t. This work has been partially
supported by the DGESIC of Spain (Project PB98-0818),
the European Comission (Contract J OR3CT980206),
and the Office of Basic Energy Sciences of the U.S.
Department of Energy (contribution no. NDRL-4358
from the Notre Dame Radiation Laboratory). We are
also indebted to Centro de Espectroscopia de la UCM.
Su p p or tin g In for m a tion Ava ila ble: Synthetic Proce-
dures, cyclic voltammograms and some additional photophysi-
cal information. This material is available free of charge via
the Internet at http://pubs.acs.org.
J O0108313