Controlling short- and long-range electron transfer processes in molecular dyads and triads

Article abstract of DOI:10.1002/chem.200204494

Novel π-extended tetrathia-fulvalene (exTTF)-based donor acceptor hybrids - dyads and triads - have been synthesized following a multistep synthetic procedure. Cyclic voltammetry and absorption spectroscopy, conducted in room temperature solutions, reveal

Full text of DOI:10.1002/chem.200204494

FULL PAPER
Controlling Short- and Long-Range Electron Transfer Processes in Molecular
Dyads and Triads
Luis Sa¬nchez,[a] Ignacio Pe¬rez,[a] Nazario MartÌn,*[a] and Dirk M. Guldi*[b]
Dedicated to Professor Klaus-Dieter Asmus on the occasion of his 65th birthday
Abstract: Novel p-extended tetrathia-
fulvalene (exTTF)-based donor accept-
or hybrids–dyads and triads–have
been synthesized following a multistep
synthetic procedure. Cyclic voltammetry
and absorption spectroscopy, conducted
in room temperature solutions, reveal
features that are identical to the sum of
the separate donor and acceptor moi-
eties. Steady-state and time-resolved
photolytic techniques confirm that upon
photoexcitation of the fullerene chro-
mophore, rapid (1.25 Â 1010 sÀ1) and ef-
ficient (67%) charge separation leads to
long-lived, charge-separated radical
pairs. Typical lifetimes for the dyad
ensembles range between 54 and
460 ns, with the longer values found in
more polar solvents. This indicates that
the dynamics are located in the ×normal
region× of the Marcus curve. In the
triads, subsequent charge shifts trans-
form the adjacent radical pair into the
distant radical pair, for which we deter-
mined lifetimes of up to 111 ms in
DMF–values never previously accom-
plished in molecular triads. In the final
charge-separated state, large donor-ac-
ceptor separation (center-to-center dis-
tances: ꢀ30 ä) minimizes the coupling
between reduced acceptor and oxidized
donor. Analysis of the charge recombi-
nation kinetics shows that a stepwise
mechanism accounts for the unusually
long lifetimes.
Keywords: charge separation
cycloaddition donor acceptor
systems ¥ sulfur heterocycles
¥
¥
The continuing quest for new donor-acceptor ensembles is
driven by several goals, namely faster and more efficient
charge separation, slower charge recombination, minimizing
the loss of excited state energy, and exploring simpler systems
with fewer components.
So far, porphyrins represent the most frequently used
excited state donor in combination with C60 (see, as a
representative example, dyad 1), generating large and com-
plex, yet highly ordered, molecular and supramolecular
Introduction
Owing to the importance and complexity of the natural
photosynthetic apparatus, the study thereof requires the use
of suitably simple models. Therefore, the design and synthesis
of molecular architectures–artificial photosynthetic antenna
and reaction centers–emerges as an interdisciplinary strategy
for devising integrated, multicomponent model systems that
transmit and process solar energy.[1]
Among the many photo- and redox-active donor and/or
acceptor building blocks explored to construct artificial
5]
photosynthetic models,[2 C60 carries great potential as a
three-dimensional, spherical electron acceptor. This is be-
cause of the small reorganization energy of C60 in electron
transfer reactions[6] and its remarkable chemical reactivity.[7]
Thus, C60 renders itself an excellent candidate for the
preparation of photo- and redox-active systems.[8]
¬
¬
[a] Prof.Dr. N. MartÌn, Dr. L. Sanchez, Lcdo. I. Perez
¬
Departamento de QuÌmica Organica I
Facultad de QuÌmica, Universidad Complutense, 28040 Madrid (Spain)
E-mail: nazmar@quim.ucm.es
[b] Dr.habil. D. M. Guldi
entities with specific functions. At small donor acceptor
separations, however, relatively short lifetimes have been
reported for radical ion pairs that fall within the picosecond
time scale. Longer lifetimes have been accomplished upon
increasing the relative separation. To cite an instance when
Radiation Laboratory
University of Notre Dame, Indiana 46556 (USA)
E-mail: guldi.1@nd.edu
Supporting information for this article is avalaible on the WWW under
http://www.chemeurj.org or from the authors.
Chem. Eur. J. 2003, 9, 2457 2468
DOI: 10.1002/chem.200204494
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2457

N. MartÌn, D. M. Guldi et al.
FULL PAPER
larger spacers (i.e., non photo- and electro-
active moieties) were used to link donor
and acceptor, lifetimes of several hundreds
of nanoseconds have been recorded.
A viable alternative implies the incorpo-
ration of additional donor moieties, form-
ing triads/tetrads/etc. The basic notion is to
achieve long-distance charge-separated
states using a sequence of several short-
range electron transfer reactions along a
well-designed redox gradient. Spectacular
radical pair lifetimes–close to seconds–
have been found.[6c]
The mechanism of charge recombination
has a considerable impact on the rate
constant. In principle, either a stepwise,
that is, overcoming a thermally activated
barrier (i.e., bona fide intermediate state),
or a concerted pathway, that is, electronic
tunneling by means of super exchange (i.e.,
virtual intermediate state) can occur.[9] Two
factors control/alter the mechanism: 1)
energy gap and 2) electronic coupling
element. Importantly, slower recombina-
tion rates are expected for the stepwise
scenario.
Recent reports indicate that aromatiza-
tion of an oxidized donor moiety can also
have a notable impact on the improvement
of light-induced charge-separation.[10] An
illustration is given in the case where a
tetrathiafulvalene (TTF) donor has been
Abstract in Spanish: Nuevos sistemas da-
dor-aceptor -dÌadas y trÌadas- derivados de
tetratÌafulvaleno p-extendido (exTTF) han
sido obtenidos mediante un procedimiento
¬
sintetico en varios pasos. La voltamperome-
attached in close proximity to C60.[11] In the resulting C60-TTF
dyads (2, 3), charge-separated radical ion pairs are formed
whose lifetimes are in the nanosecond range. Clearly, the gain
of aromaticity of the 1,3-dithiolium cation generated upon
oxidation is central to this effect.
¬
trÌa cÌclica y espectroscopia de absorcion, realizadas en
disoluciones a temperatura ambiente, muestran caracterÌsticas
que corresponden a la suma de ambos fragmentos dador y
aceptor por separado. Las tecnicas fotolÌticas en estado
estacionario y a tiempos definidos confirman que la fotoexci-
¬
Further advances in the stabilization concept were based on
the use of p-extended tetrathiafulvalenes (exTTF).[12] We
have previously reported the synthesis of different C60-exTTF
dyads (4 6).[13] Pico- and nanosecond transient absorption
measurements reveal that the instantaneously formed full-
erene singlet excited state transforms rapidly into the charge-
separated radical pair. Remarkably, the lifetimes of the
charge-separated states in C60-exTTF are in the range of
several hundreds of nanoseconds in deoxygenated benzoni-
trile and slightly lower in the less polar CH2Cl2.[13c] For
comparison, the parent TTF dyads (2) or zinc tetraphenyl-
porphyrin dyads (1) yielded radical pairs having lifetimes of
up to a few nanoseconds.
¬
¬
¬
tacion del cromoforo de fullereno conduce a una rapida
À1
(1.25 Â 1010 s ) y eficiente (67%) separacion de carga con
¬
¬
formacion de pares radical con tiempos de vida largos. Los
tiempos de vida tÌpicos para las dÌadas oscilan entre 54 y 460 ns,
con los valores mas elevados encontrados en los disolventes
mas polares, indicando que la dinamica se localiza en la
¬
¬
¬
¬
™region normal∫ de la curva de Marcus. En las trÌadas, el
posterior desplazamiento de carga transforma el par radical
adyacente en un par radical a larga distancia, para el cual se ha
determinado un tiempo de vida de 111 ms en DMF, valores estos
que no habÌan sido alcanzados hasta ahora en trÌadas
moleculares. El analisis de la cinetica de recombinacion de
carga indica que los elevados tiempos de vida serÌan debidos a
un mecanismo por etapas.
¬
¬
¬
Herein we apply the concept of gain of aromaticity and
planarity upon oxidation of the exTTF donor to a newly
2458
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 2457 2468

Electron Transfer Processes in Molecular Dyads and Triads
2457 2468
designed series of different exTTF-containing dyads and
triads. Details on the synthetic, electrochemical, and photo-
physical work will be presented, highlighting the noteworthy
impact that a chemical spacer [C60-exTTF (18, 23)], or a
second exTTF unit [C60-exTTF1-exTTF2 (25, 26)], exert on the
improvement of light-induced charge separation. Particular
emphasis is placed on attaining incoherent sequential charge
recombination by designing components that result in small
driving forces. The lifetimes for intramolecular charge sepa-
ration, determined in 25 and 26, are by far the longest values
ever reported in molecular triads.
benzoic acid (16) in the presence of dimethylaminopyridine
(DMAP) and N,N'-dicyclohexylcarbodiimide (DCC), to af-
ford diformyl-substituted donors 19a, b, which were further
used for the preparation of triads 25 and 26.
To obtain dyad 18, anthraflavic acid was monoalkylated
with hexyl bromide and potassium carbonate, which yielded
quinone 8.[16] The remaining hydroxy group in 8 was protected
by reaction with tert-butylchlorodiphenylsilane leading to
compound 9. Olefination of quinone 9 with phosphonate ester
11 and butyllithium gave 12. Deprotection of the donor
compound 12 by treatment with tetrabutylammonium fluo-
ride gave the monohydroxy-substituted exTTF 13. Finally, an
esterification reaction of 13 with 4-formylbenzoic acid (16)
afforded compound 17, which then reacted by 1,3-dipolar
cycloaddition with C60 and sarcosine to yield dyad 18 in 61%
yield.
A similar synthetic strategy was chosen for the preparation
of dyad 23 and also triads 25 and 26 (Scheme 2). In particular,
2-hydroxymethyl-9,10-bis(1,3-dithiol-2-ylidene)-9,10-dihydro-
anthracene (20)[17a] was converted with 4-formylbenzoic acid
(16) in the presence of DMAP and DCC to the formyl-
containing ester (22). Subsequent reaction with C60 and
sarcosine (N-methylglicine) in refluxing toluene yielded 23
in 61% yield.
Results and Discussion
Synthesis: Target molecules 18, 23, 25, and 26 were prepared
by a 1,3-dipolar cycloaddition to C60 of azomethine ylides
generated in situ.[14]
The synthetic route towards dyad 18 is summarized in
Scheme 1. The preparation of p-extended-TTF electron
donors (15a, b) requires the previous protection of the
hydroxy groups of the starting anthraflavic acid (2,6-dihy-
droxyanthraquinone) (7) in the form of bissilylated deriva-
tives (10). This was carried out by the reaction of 7 with tert-
butylchlorodiphenylsilane in the presence of imidazole. Sub-
sequent twofold reaction of protected quinone 10 with a
phosphorus-stabilized carbanion, generated under basic con-
ditions (n-butyllithium), gave p-extended-TTFs 14a, b. Com-
pounds 14a, b were deprotected by treatment with tetrabu-
tylammonium fluoride to yield the respective diols 15a, b,
similar to that reported by Bryce and Marshallay.[15] Com-
pounds 15a, b, which were sparingly soluble in common
organic solvents, reacted with two equivalents of 4-formyl-
The synthetic workup enroute to triads 25 and 26 starts
from 2-hydroxymethyl substituted exTTF 20. In this case, 20
was transformed into a Wittig reagent (21) by treating it with
triphenylphosphane hydrobromide (PHPh3 BrÀ) in refluxing
‡
toluene.[17b] A Wittig olefination reaction with bisaldehydes
19a, b in a 1:1 stoichiometric ratio gave bis-p-extended TTFs
24a, b in moderate yields (43 and 38%, respectively). Finally,
a cycloaddition reaction of the azomethyne ylides generated
in situ from exTTF-exTTF-CHO (24a, b) and sarcosine to C60
Scheme 1. a) Hexylbromide, K2CO3, DMF; b) tBuPh2SiCl, imidazole, DMF; c) nBuLi, THF, À788C ; d)nBu4NF, THF; e) 4-carboxybenzaldehyde (16),
DCC, DMAP, CH2Cl2; f) C60, sarcosine, toluene, D.
Chem. Eur. J. 2003, 9, 2457 2468
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2459

N. MartÌn, D. M. Guldi et al.
FULL PAPER
Scheme 2. a) PPh3HBr, toluene; b) 4-carboxybenzaldehyde (16), DCC, DMAP, CH2Cl2; c) C60, sarcosine, toluene, D d) 19a, b, tBuOK, toluene, D.
yielded triads 25 and 26 as stable
brown-yellowish solids in 43 and
51% yields, respectively.
1
The H NMR spectra of electro-
active dyads and triads 18, 23, 25, and
26 show, in addition to the aromatic
signals, resonance signals of the
À
N Me group at d ˆ 2.9 ppm, and
the pyrrolidine protons at around
d ˆ 5.01 5.07 (d, J ˆ 9.5 Hz, 1H),
4.30 4.37 (d, J ˆ 9.5 Hz, 1H), and
5.01 5.10 ppm (s, 1H), in good
agreement with other N-methylful-
leropyrrolidine derivatives.[14] A re-
stricted rotation of phenyl substitu-
ents on the pyrrolidine ring has been
described for different phenylfuller-
opyrrolidine derivatives.[18] We have
also observed this dynamic effect in
the dyads and triads prepared, which
showed broad signals for the ortho
Figure 1. 13CNMR spectrum for dyad 23 in CDCl3.
aromatic protons close to the fullerene surface at d ˆ 8.0 ppm.
The large number of 13C NMR signals seen in CDCl3:CS2
mixtures confirm the lack of symmetry in these fulleropyrro-
lidine derivatives–see Experimental Section and Figure 1.
The UV/Vis spectra of 18, 23, 25, and 26 show absorption
bands at approximately 432 nm, which clearly correspond to
the features of p-extended TTF.[12] This absorption hides the
typically weak absorption band of fullerene derivatives close
to 430 nm, and is responsible for the yellow color that these
dyads and triads show in solution.
Table 1. Redox potentials of 4a, 18, 19a, 19b, 22, 23, 25, 26, and 27 at room
temperature.[a]
Compound E1
E2
E3
E4
red
Eox
red
red
red
4a
18
19a
19b
22
23
25
26
À 0.66
À 1.01
À 1.08
À 1.67
À 1.62
À 1.96 0.46 (2eÀ)
À 2.01 0.41 (2eÀ)
À 0.70
À 1.51[b]
À 1.43[b]
0.47 (0.45)[c] (2eÀ)
0.58 (0.56)[c] (2eÀ)
0.46 (2eÀ)
À 0.69
À 0.70
À 0.69
À 0.60
À 1.07
À 1.08
À 1.07
À 1.00
À 1.61
À 1.61
À 1.61
À 1.52
À 2.01 0.39 (2eÀ)
À 2.01 0.41 (br) (4eÀ)
À 1.99 0.43 (br) (4eÀ)
À 2.04
C60
27[d]
0.45 (2eÀ)
Electrochemistry: The CV data of the studied compounds are
collected in Table 1 along with those of C60 and the parent
unsubstituted 9,10-bis(1,3-dithiol-2-ylidene)-9,10-dihydroan-
thracene (27).[12]
[a] Experimental conditions: V versus SCE, GCE as working electrode,
Bu4NClO4 (0.1m) as supporting electrolyte, ODCB/MeCN (4/1) as solvent,
100 mVsÀ1 scan rate. [b] Corresponding to the aldehyde group. [c] In
CH2Cl2. [d] 27: 9,10-bis(1,3-dithiol-2-ylidene)-9,10-dihydroanthracene, in
2460
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 2457 2468

Electron Transfer Processes in Molecular Dyads and Triads
2457 2468
Table 2. Fluorescence and singlet excited state features of photoexcited
4a, 18, 23, 25, 26, and 28 at room temperature.
The dyads and triads prepared exhibited an amphoteric
redox behavior. Thus, all compounds (18, 23, 25, 26) showed
the presence of four quasireversible reduction waves, resem-
bling the trend found for the parent C60. These reduction
potentials are shifted to more negative values (compared to
C60) as a consequence of the partial loss of p conjugation in
the C60 core. This raises the LUMO energy of the resulting
modified fullerene.[19]
On the oxidation side, dyads 18 and 23 show the presence of
an electrochemically quasirreversible oxidation wave involv-
ing a two-electron process to form the dication, similar to that
observed for the parent unsubstituted exTTF 27.[12] The
slightly stronger electron donor ability of 23 can be accounted
for by the presence of the electron releasing -O-CH2- group
on the aromatic hydrocarbon skeleton. Triads 25 and 26
exhibit a broad quasirreversible oxidation wave to form the
tetracation species. Triad 25, bearing two identical p-extended
TTFs (R'' ˆ H), shows the broad oxidation peak centered at
0.41 V. Similarly, triad 26, which bears two different p-
extended TTFs (R'' ˆ SMe) (see Scheme 2), shows a broad
oxidation peak centered at 0.43 V involving the oxidation of
both exTTF units. This overlap necessitates determining the
oxidation potentials for the respective exTTF units in
references 19a and 19b.[20]
Solvent
t singlet
t fluorescence [ns]
F fluorescence
excited state [ns]
28
4a
23
toluene
THF
benzonitrile
1.35
1.38
1.39
1.5
1.5
1.41
6.0 Â 10À4
6.0 Â 10À4
5.9 Â 10À4
toluene
CH2Cl2
benzonitrile
0.18
0.13
0.08
0.21
0.4 Â 10À4
0.18 Â 10À4
0.1 Â 10À4
toluene
THF
CH2Cl2
benzonitrile
0.57
0.49
0.47
0.38
0.61
0.50
0.45
0.36
3.1 Â 10À4
2.0 Â 10À4
2.1 Â 10À4
1.9 Â 10À4
18
25
toluene
THF
CH2Cl2
benzonitrile
0.45
0.33
0.25
0.15
0.41
0.29
0.21
2.1 Â 10À4
1.8 Â 10À4
1.7 Â 10À4
1.1 Â 10À4
toluene
THF
CH2Cl2
benzonitrile
DMF
0.34
0.21
0.18
0.15
0.11
0.36
0.21
0.18
26
toluene
THF
CH2Cl2
benzonitrile
DMF
0.38
0.24
0.24
0.18
0.12
0.39
0.25
0.23
Photophysical measurements: C60-extendedTTF dyads: At
first, a series of C60-exTTF dyads (4a, 18, and 23) were
considered to examine the short- and long-range interactions
between the two redox-reactive centers at different donor
acceptor separations and compositions (see Figure S2 in the
Supporting Information).
Preliminary insight into conceivable implications was taken
from steady-state fluorescence experiments. The data were set
in relation to a fullerene reference (N-methylpyrrolidi-
no[3',4':1,2][60]fulllerene, 28), which lacks the exTTF donor.
Owing to the extensive absorption features of exTTFs in the
ultraviolet and visible region–which exhibit maxima around
360 and 430 nm–337 nm was chosen as the excitation wave-
length to ensure predominant excitation of C60. The absorp-
tion ratio between C60 and exTTF at 337 nm is approximately
9:1, which still necessitated correction of the fluorescence
quantum yields.
Fulleropyrrolidine 28 reveals nearly solvent-independent
fluorescence quantum yields (F) of 6.0 Â 10À4 and fluores-
cence lifetimes (t) of 1.5 ns (Table 2). The only observable
that changes in room temperature emission measurements is
the maximum, which undergoes a moderate red-shift of
around 5 nm when moving from toluene towards more polar
environments.
In dyads 4a, 18, and 23, a notable quenching of the fullerene
fluorescence is seen (Table 2), which, in toluene, reaches a
factor of 15 relative to 28. Upon modifying the solvent
polarity from toluene (e ˆ 2.38) to THF (e ˆ 7.6) then to
dichloromethane (e ˆ 9.08) and benzonitrile (e ˆ 24.8) a
gradual intensification of the quenching is discernible. An
illustration is given in Figure 2 for reference 28 and dyad 18.
We postulate that the underlying solvent dependence is due to
an intramolecular electron transfer between the exTTF donor
Figure 2. Emission spectra of fulleropyrrolidine 28 in toluene (solid line)
and dyad 18 in different solvents (see labels for assignment) with matching
absorption at the 337 nm excitation wavelength, OD337 nm ˆ 0.5.
.
À
and the photoexcited fullerene (1.76 eV) to yield C60
(exTTF) [Eq. (1)].
-
.
‡
hn
.
.
‡
1
C60-(exTTF) ! *C60-(exTTF) ! C60 À-(exTTF)
(1)
The suggestion of a ™through-bond∫ transfer is in sound
agreement with the rigid nature of the intervening spacers,
which prohibit major conformational rearrangements in the
ground and excited state. In fact, theoretical modeling of 4a,
18, and 23 (i.e., PM3, see Figure S2 in the Supporting
Information) confirms the structural rigidity, ensuring the
position of both C60 and exTTF in well-defined locations.
Extra support for the consideration of an intramolecular
electron transfer scenario stems from the thermodynamic
driving force determination. This is documented in detail in
the Supporting Information.
Chem. Eur. J. 2003, 9, 2457 2468
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2461

N. MartÌn, D. M. Guldi et al.
FULL PAPER
In dyads 4a, 18, and 23 two structural variables are
modified to shed light on effects evolving from 1) the
donor acceptor separation and 2) the oxidation strength of
the electron donor on the reactivity of photoexcited C60. In
line with these alterations, the fullerene fluorescence increas-
es in parallel with larger donor acceptor separations (i.e., 4a
compared to 23) and higher oxidation potentials of the donor
(i.e., 23 compared to 18). To illustrate this, the quantum yields
in toluene range from around 0.4 Â 10À4 (4a) to 3.1 Â 10À4
(23). Similar trends were derived for the quantum yields in
more polar THF, dichloromethane, and benzonitrile.
and 700 nm, accompanied by a low-energy shoulder at
800 nm.[23]
Detecting the instantaneous appearance (i.e., 18 ps) of the
880 nm absorption affirms the successful C60 excitation in
dyads 4a, 18, and 23 similar to that shown in Figure 3. In-
stead of seeing, however, the slow ISCdynamics, as 28
exhibits, the singlet singlet absorption decays in the presence
An independent probe for the magnitude of electron
transfer quenching is fluorescence lifetime experiments. For
this we recorded the 715 nm maximum of the fluorescence in
fulleropyrrolidines. Importantly, fluorescence lifetime experi-
ments allow segregating contributions of exTTF from those of
C60. In general, the fluorescence decay curves were best fitted
by a single exponential decay law. The rate constants (k ˆ 1/t)
agree quite well with those determined indirectly by applying
the approximation given in Equation (2).
Figure 3. Differential absorption spectra obtained upon picosecond flash
photolysis (355 nm) of ꢀ1.0 Â 10À5 m solutions of fulleropyrrolidine 28 in
nitrogen-saturated toluene with a time delay of 50 ps.
k ˆ 1/t(4a, 18, 23) ˆ [F(28) À F(4a, 18, 23)]/[t(28)F(4a, 18, 23)
]
(2)
Equation (2) relates the fluorescence quantum yields in the
of exTTF donors with accelerated dynamics. The singlet
excited state lifetimes, as determined from an average of first-
order fits of the time-absorption profiles at various wave-
length (850 950 nm), are listed in Table 2. Spectroscopically,
the transient absorption changes, taken after the completion
of the decay, bear no resemblance with the C60 triplet excited
state. In particular, the new transients reveal strong maxima at
ꢀ665 nm, which match those of the one-electron oxidized
exTTF radical cations.[24] The spectral differences, namely, of
the C60 triplet and that of charge-separated radical pair, are
illustrated in Figures 4 and 5, respectively.
donor acceptor ensemble F(4a, 18, 23) to that of fulleropyrro-
[21]
lidine 28, F(28), and its lifetime, t(28)
.
To exploit the interactions between photoexcited exTTF
and ground-state C60, the former was intentionally excited at
430 nm. The relative absorption ratio between C60 and exTTF
at 430 nm is 1:9. Although exTTF references emit around
475 nm, which correlates to a singlet excited state energy of
ꢀ2.6 eV, the donor emission remains nearly unaffected in 4a,
18, and 23. Only a marginal quenching of 10% is noted, for
example, in toluene. This suggests that exTTFs, once photo-
excited, fail to contribute notably to the overall electron and
energy transfer reactivity in 4a, 18, and 23. A similar
conclusion has been reached for weaker absorbing tetrathia-
fulvalenes (TTF) in C60-TTF donor-acceptor systems.[22]
In summary, steady-state and time-resolved fluorescence
measurements testify that a solvent-dependent and rapid
electron transfer decay of the C60 singlet excited state prevails
in 4a, 18, and 23. To shed light on the nature of the product
evolving from this intramolecular deactivation, complemen-
tary transient absorption measurements were necessary (i.e.,
with picosecond to millisecond time-resolution). Following
the time evolution of the characteristic singlet excited state
features of C60, for instance, is a convenient means of
identifying spectral features of the resulting photo-products
and to determine absolute rate constants for the intramolec-
ular decay.
Figure 4. Differential absorption spectra obtained upon nanosecond flash
photolysis (337 nm) of ꢀ1.0 Â 10À5 m solutions of fulleropyrrolidine 28 in
nitrogen-saturated toluene with a time delay of 50 ns.
Up front, the well-known excited state properties of
fulleropyrrolidines (28) will be discussed, since they emerge
as important reference points for the interpretation of the
features expected in dyads 4a, 18, and 23. The singlet excited
An important observation is that the singlet excited state
lifetimes match quantitatively those values derived in the
fluorescence experiments–steady-state and time-resolved.
Overall three trends are established for the fullerene singlet
excited state features: The underlying decays decrease as a
state, displaying
a distinctive singlet singlet transition
around 880 nm, undergoes a quantitative intersystem crossing
(ISC) with a rate of 5 Â 108 sÀ1.[23] The ISCprocess yields the
long-lived triplet manifold, for which maxima are noted at 360
2462
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 2457 2468

Electron Transfer Processes in Molecular Dyads and Triads
2457 2468
Table 3. Radical cation characteristics, radical pair lifetimes and radical pair
quantum yields in 4a, 18, 23, 25, and 26 at room temperature.
Solvent
Radical
cation
Lifetime Lifetime Quantum Quantum
(t)
(t)
yield (F) yield (F)
absorption radical
radical
radical
radical
pair
[nm]
pair [ns] pair [ms] pair
4a toluene
CH2Cl2
benzonitrile
660
71
204
0.39
0.1
23
18
25
toluene
THF
CH2Cl2
benzonitrile
54
95
115
310
0.15
0.17
0.2
660
665
0.15
toluene
THF
CH2Cl2
benzonitrile
150
310
322
460
0.13
0.5
0.67
0.52
Figure 5. Differential absorption spectra obtained upon nanosecond flash
photolysis (337 nm) of ꢀ1.0 Â 10À5 m solutions of dyad 18 in nitrogen-
saturated benzonitrile with a time delay of 50 ns.
toluene
THF
130
413
12.1
0.16
0.025
function of 1) solvent polarity (4a, 18, and 23), 2) donor
acceptor separation (4a and 23) and 3) donor strength of the
exTTF moiety (23 and 18). All these tendencies may be
rationalized with respect to the free energy changes associ-
ated with an intramolecular electron transfer between the
fullerene singlet excited state, the electron acceptor, and the
exTTF ground state, the electron donor.
CH2Cl2
benzonitrile
DMF
665/660
680/660
662
708
54.2
93.1
0.28
0.22
0.053
0.04
26
toluene
THF
CH2Cl2
benzonitrile
DMF
105
389
18.9
85.1
0.24
0.28
582
664
0.28
0.22
0.14
0.11
111
To examine the charge-recombination dynamics, the same
dyad solutions were excited with a 6 ns laser pulse. In this
context, the spectral fingerprints of the fullerene p-radical
anion (1000 nm; e ꢀ 10000 MÀ1 cmÀ1)[23] and that of the
exTTF p-radical cation (665 nm; e ꢀ 25000 MÀ1 cmÀ1)[24]
(Figures 5 and 6, respectively), as seen immediately after the
ns laser pulse, are useful probes. It is an important fact that the
decay of both probes resemble each other and give rise to
kinetics that obey to a clean unimolecular rate law. The rates
and quantum yields are listed in Table 3.[25]
scaffolds, thus rendering the reverse reduction process to the
neutral molecule more difficult, both parameters provide
additional stabilization forces for the radical pair.
Extending the donor-acceptor separation from 9.5 ä (4a)
to 14.0 ä (23), while keeping the thermodynamics unchanged,
led to a surprising observation. Instead of the expected large
increase in radical pair lifetime, as observed in C60 porphyrin
systems, only a moderate (66%) improvement was seen.[26]
This leads us to assume that the stabilization forces, as they
stem from the gain of aromaticity/planarity, impose much
stronger bearings on the radical pair at shorter separations as
in dyad 4a. At larger separations, for example in dyad 23, the
electronic coupling (V) becomes the dominant factor con-
trolling the kinetics.
C60-extendedTTF-extendedTTF triads[27]: We note the follow-
ing structural features: Dyad 18 and triads 25 and 26 bear the
same C60-exTTF primary building block. In this context, it is
vital to realize that the composition of this subunit controls
the deactivation of the fullerene singlet excited state, while
the presence of a second exTTF donor or fullerene acceptor is,
to a large degree, irrelevant to the initial charge-separation
event. The fluorescence lifetimes of 25 are practically
identical to those listed for dyad 18 (Table 2). This supports
the notion that the photoexcited fullerene in 18, and 25 indeed
deactivates in a similar manner.
Regarding the picosecond transient absorption measure-
ments, immediately, after the 18 ps laser excitation of 25, the
strong singlet-singlet absorption of the fullerene (lmax at
900 nm) was found (Figure 7).[21] Despite the presence of two
exTTFs, this once again confirms the successful formation of
the fullerene singlet excited state. Similar to 4a, 18, and 23
Figure 6. Differential absorption spectra obtained upon nanosecond flash
photolysis (337 nm) of ꢀ1.0 Â 10À5 m solutions of dyad 18 in nitrogen-
saturated benzonitrile with a time delay of 50 ns.
A central issue is the stabilization of the charge-separated
radical pair as a function of donor acceptor separations.
Strong effects were noted in 4a. This observation is consistent
with our recent report on fullerene derivatives (i.e., [6-5]-
open and [6-6]-closed methanofullerenes)[13c] linked directly
to a series of exTTF donors. Since upon one-electron
oxidation exTTF moieties turn into aromatic and also planar
Chem. Eur. J. 2003, 9, 2457 2468
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2463

N. MartÌn, D. M. Guldi et al.
FULL PAPER
Figure 7. Differential absorption spectra obtained upon picosecond flash
photolysis (355 nm) of ꢀ1.0 Â 10À5 m solutions of triad 25 in nitrogen-
saturated benzonitrile with a time delay of 50 ps.
(see above) the fullerene singlet excited state is subject to a
solvent-dependent decay. Also in the case of 25, the product
of the initial decay is the charge-separated radical pair, that is,
.
.
‡
À
[28]
C60 and (exTTF)
.
In light of the two exTTF electron
donors, which support an additional intramolecular charge-
.
‡
shift reaction (ÀDGCS ˆ 0.06 eV), that is, (exTTF)1
!
.
‡
(exTTF)2 , we analyzed the picosecond kinetics in depth
[Eq. (3)]. All attempts to monitor this charge-shift failed,
indicating that this intramolecular reaction constitutes either
only a minor component and/or takes place with dynamics
that are out of our assessable time domain (i.e., 18 5000 ps).
Figure 8. Time-absorption profiles at 1000 nm monitoring the fullerene
.
.
.
À
‡
radical anion decay dynamics of a) C60 À-(exTTF)1 -(exTTF)2 and b) C60
-
.
‡
(exTTF)1-(exTTF)2 in triad 25 in nitrogen-saturated DMF.
hn
1
C60-(exTTF)1-(exTTF)2 ! *C60-(exTTF)1-(exTTF)2
.
.
(3)
‡
! C60 À-(exTTF)1 -(exTTF)2
In a detailed pulse radiolytic work,[24] we demonstrated that
lmax is impacted by the substitution pattern of the exTTF
derivative. In fact, the maximum for the parent exTTF p-
radical cation is found at 660 nm, which resembles that of the
distant radical pair.
The transient absorption changes, recorded 50 ns after 6 ns
laser excitation, reveal the attributes for the one-electron
oxidized exTTF (lmax ꢀ 665 nm) and that of the one-electron
reduced fullerene (lmax ꢀ 1000 nm)–compare Figure 5 and 6.
The decay dynamics of the radical pair absorption, as typically
recorded on the nanosecond/microsecond timescale, gives rise
to a two-component decay (Figure 8, DMF). The faster segment
reveals a lifetime of several hundred nanoseconds, while the
slower segment lies in the range of several tens of microseconds,
see Table 3. Both decay components were best fitted by first-
order kinetics, confirming intramolecular reactions.[29]
Determining the quantum yields of charge-separation (see
.
.
‡
Table 3) for the two radical pairs, namely, C60 À-(exTTF)1
-
.
.
‡
(exTTF)2 and C60 À-(exTTF)1-(exTTF)2 led to the conclu-
sion that the distant and long-lived radical pair is being
formed in minor yields. The thermodynamics estimated for 25
unveil that a charge-shift reaction between the two adjacent
.
.
‡
‡
exTTF moieties (i.e., (exTTF)1 ! (exTTF)2 ) has a small
driving force, À0.06 eV. Consequently, the formation of the
distant radical pair can only be rationalized in terms of an
existing equilibrium. Applying the energy difference between
the two states, we estimate quantum yields of ꢀ15% relative
to that of the adjacent radical pair, which is in satisfying
agreement with the experimental values.[30]
Interestingly, not only are the decay rates of the short-lived
species identical with those of dyad 18, but also their spectral
features (i.e., lmax ꢀ 665 nm) are superimposable. From these
kinetic and spectroscopic arguments we must infer that the
.
À
origin of this product is the adjacent radical pair, C60
-
.
‡
(exTTF)1 -(exTTF)2.
Raising the oxidation potential of the exTTF moiety
neighboring the fullerene leads to a number of important
consequences. First, it clearly decreases the driving force for
the initial electron transfer starting from the photoexcited
fullerene in triad 26. Second, it effects the charge-shift
reaction towards the distant radical pair (ÀDGCS ˆ 0.17 eV).
The first outcome should have no implications, since the free
energy changes associated with the initial intramolecular
electron transfer are sufficiently exothermic that increasing
the oxidation potential (i.e., comparing 25 with 26) still
guarantees activation of the electron transfer process. By
By analyzing the exTTF radical cation absorption we noted
some differences between the maximum of the short-lived and
long-lived species of about 5 nm (i.e., 665 ! 660 nm). This can
be reasonably correlated with the changes in the chemical
substitution of the two disparate exTTF donors, namely,
(exTTF)1 and (exTTF)2. Consequently, we postulate that the
.
À
long-lived product is the distant radical pair, namely, C60
-
.
‡
(exTTF)1-(exTTF)2 [Eq. (4)].
.
.
.
.
‡
‡
C60 À-(exTTF)1 -(exTTF)2 ! C60 À-(exTTF)1-(exTTF)2
(4)
2464
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 2457 2468

Electron Transfer Processes in Molecular Dyads and Triads
2457 2468
contrast, altering the charge-shift driving forces is expected to
have a more subtle impact, namely, to facilitate mediation of
the positive charge to the remote end of the linear triad, the
energy gap dependence, as varied by the solvent dependence,
leads to the conclusion that both charge-recombination
processes are also in the normal region of the Marcus-
parabola (i.e., ÀDGCR < l). Due to the smaller energy gap
(ÀDGCR) in 25, charge-recombination within the adjacent
pair is slowed-down by about 10% relative to that 26.
In line with the concept of modulating the redox gradient to
control the distant radical pair, the quantum yield of the latter
increased up to 50%. Interestingly, although the energies of
.
.
‡
distant C60 À-(exTTF)1-(exTTF)2 radical pair.
As far as the emission studies are concerned, the fullerene
fluorescence lifetime is surely lengthened in 26 relative to 25.
This reflects the smaller free energy changes (ÀDGCS) in 26.
From the transient absorption measurements, a parallel trend
was measured for the singlet excited state lifetimes. In
general, the deactivation rates differ between these two C60-
exTTF-exTTF triads by around 20%.
The differential absorption changes, recorded at the end of
the charge-separation process (200 400 ps), disclosed dissim-
ilar transient species in the two different triads (25 and 26).
While we registered a lmax at 665 nm for 25, 26 has its strongest
absorption at 680 nm (Figure 9a). As stated above, our
independent pulse-radiolysis work[24] led to a maximum for
a S À Me derivative (resembling that in 26) at 675 nm. Based
.
.
‡
the C60 À-(exTTF)1-(exTTF)2 radical pair in 25 and 26 are
identical, notable changes are seen in the corresponding
lifetimes. In particular, lowering the energy of the distant
radical pair below that of the adjacent one slows-down the
charge-recombination. In DMF, 93.1 ms should be compared
to 111 ms for 25 and 26, respectively.
Considering donor acceptor separations of about 30 ä,
electronic coupling within the distant radical pair is weak and
will not change notably between triad 25 and 26.[32]A temper-
ature dependence of the intramolecular charge recombination
rates gave further insight into the thermally activated barrier.
The Arrhenius plot (Figure 10) for triad 26 in deoxygenated
Figure 10. Arrhenius plot of intramolecular charge recombination in triad
26 in deoxygenated benzonitrile.
benzonitrile gives rise to a good linear relationship between ln
kCR and 1/T. The activation barrier (Ea), as determined from
the slope, is 0.16 eV, which agrees with the difference in
.
.
.
À
‡
energy level between C60 À-(exTTF)1-(exTTF)2 and C60
-
.
‡
(exTTF)1 -(exTTF)2. Given the good agreement between
activation barrier (Ea) and energy gap (ÀDGCS), we conclude
.
.
‡
that C60 À-(exTTF)1-(exTTF)2 deactivates by means of a
stepwise electron transfer mechanism (see Equation (5))
rather than a concerted mechanism (see Equation (6).)[33]
Figure 9. Differential absorption spectra obtained upon nanosecond flash
photolysis (337 nm) of ꢀ1.0 Â 10À5 m solutions of triad 26 in nitrogen-
.
.
‡
saturated benzonitrile with a time delay of a) 50 ns, C60 À-(exTTF)1
-
.
.
.
.
‡
‡
C60 À-(exTTF)1-(exTTF)2 ! C60 À-(exTTF)1 -(exTTF)2
(5)
(6)
.
.
‡
(exTTF)2 and b) 2000 ns C60 À-(exTTF)1-(exTTF)2
.
! C60-(exTTF)1-(exTTF)2
.
.
‡
C60 À-(exTTF)1-(exTTF)2 ! C60-(exTTF)1-(exTTF)2
on this spectral similarity, we conclude the successful gen-
eration of the adjacent radical pair.[31]
Again, two major decays were noted in triad 26, which
support the presence of both radical pairs, namely, adjacent
(lmax at 680 nm, Figure 9a) and distant (lmax at 660 nm,
Figure 9b). These represent a fast step occurring on a
timescale of several hundred nanoseconds and a slower one,
which yields lifetimes that typically vary between 18.9 and
111 ms in THF and DMF, respectively. Inspection of the
Conclusions
We have described the multistep synthesis of novel electro-
active dyads (18 and 23) and triads (25 and 26) by 1,3-dipolar
cycloaddition reactions of azomethine ylides generated in
situ, endowed with the strong electron donor exTTFs, to C60.
Chem. Eur. J. 2003, 9, 2457 2468
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2465

N. MartÌn, D. M. Guldi et al.
FULL PAPER
3J1(H,H) ˆ 8.4 Hz, 4J2(H,H) ˆ 2.4 Hz, 1H), 6.76 (dd, 3J1(H,H) ˆ 8.4 Hz,
4J2(H,H) ˆ 2.4 Hz, 1H), 6.30 (d, 3J(H,H) ˆ 12 Hz, 4H), 5.10 (s, 1H), 5.07 (d,
3J(H,H) ˆ 9.6 Hz, 1H), 4.37 (d, 3J(H,H) ˆ 9.6 Hz, 1H), 4.01 (t, 3J(H,H) ˆ
7.5, 2H), 2.89 (s, 3H), 1.83 (q, 3J(H,H) ˆ 7.7 Hz, 2H), 1.51 1.39 (m, 6H),
0.91 ppm (m, 3H); 13C NMR (125 MHz, CDCl3:CS2, 258C): d ˆ 164.4,
157.2, 152.7, 152.3, 147.3, 146.3, 145.9, 145.7, 145.6, 145.2, 144.7, 144.3, 143.1,
143.0, 142.6, 142.1, 141.7, 140.2, 137.0, 136.0, 135.6, 133.1, 130.6, 130.0, 129.5,
128.1, 126.0, 125.8, 118.5, 118.2, 117.3, 112.2, 110.9, 83.1, 69.9, 69.0, 68.2, 61.0,
40.0, 31.8, 29.9, 25.9, 22.8, 14.2 ppm; FTIR (KBr) 125 MHz: nÄ ˆ 2921, 2850,
2777, 1733, 1600, 1543, 1506, 1460, 1410, 1243, 1196, 1101, 1062, 1016, 635,
525 cmÀ1; UV/Vis (CH2Cl2): lmax (log e) ˆ 254 (4.84), 276 (4.77), 310 (4.56),
In order to carry out a systematic study, different rigid
chemical spacers and substitution patterns on the donor
moiety have been used for the dyads, while the same chemical
spacer was used to evaluate the effect that a second donor unit
has on the electrochemical and photophysical properties in
these rigid ensembles. It is worth mentioning that the syn-
thesis of the reported compounds is rather complex and, thus,
compound 26 requires twenty synthetic steps for its prepara-
tion.
Donor acceptor distances between 9.5 and 15.3 ä prevent
substantial charge transfer character in the ground state, as,
for example, confirmed in cyclic voltammetry and absorption
spectroscopy. Only upon photoexcitation of the fullerene
chromophore rapid and efficient charge transfer is seen,
‡
432 (4.30) nm; MS (MALDI-TOF): m/z (%) 1376 (100) [M ].
Dyad C60-exTTF (23): 61% yield; m.p. > 3008C(decomp); 1H NMR
(300 MHz, CDCl3, 258C): d ˆ 8.16 (d, 3J(H,H) ˆ 8.4 Hz, 2H), 8.10 (d,
4
3J(H,H) ˆ 8.0 Hz, 1H), 7.90 (brs, 2H), 7.76 (d, J(H,H) ˆ 1.8 Hz, 1H), 7.65
(m, 2H), 7.34 (dd, 3J1(H,H) ˆ 8.0 Hz, 4J1(H,H) ˆ 1.8 Hz, 1H), 7.28 (m, 2H),
6.28 (m, 4H), 5.43 (s, 2H), 5.02 (d, 3J(H,H) ˆ 9.5 Hz, 1H), 5.01 (s, 1H), 4.30
(d, 3J(H,H) ˆ 9.5 Hz 1H), 2.82 ppm (s, 3H); 13CNMR (75 MHz,
CDCl3:CS2, 258C): d ˆ 165.9, 156.7, 153.8, 152.6, 152.3, 146.4, 146.3, 146.2,
146.1, 146.0, 145.8, 145.7, 145.6, 145.5, 145.4, 145.3, 145.2, 145.1, 145.0, 144.6,
144.5, 144.3, 144.2, 143.0, 142.9, 142.6, 142.5, 142.4, 142.3, 142.2, 142.1, 142.0,
141.9, 141.8, 141.6, 141.4, 140.1, 139.8, 139.5, 136.8, 136.3, 135.8, 135.6, 135.5,
135.2, 135.1, 135.0, 133.4, 130.2, 130.1, 129.9, 129.2, 128.9, 128.1, 125.9, 125.7,
125.2, 125.1, 124.9, 124.8, 124.6, 121.7, 117.2, 117.1, 117.0, 83.0, 76.8, 69.9, 67.6,
66.7, 39.9 ppm; FTIR (KBr): nÄ ˆ 1744, 1541, 1507, 1451, 1424, 1384, 1365,
1284, 1265, 1231, 1186, 1033, 762, 744, 576, 526 cmÀ1; UV/Vis (CH2Cl2): lmax
(log e) ˆ 254 (4.84), 276 (4.77), 310 (4.56), 432 (4.30) nm; MS (MALDI-
.
.
.
‡
which yields C60 À-(exTTF) (4a, 18, 23) and C60 À-(ex-
.
‡
TTF) -(exTTF) (25, 26). For dyads 4a, 18, and 23 charge
recombination follows directly and regenerates quantitatively
the singlet ground state. In the triads, on the other hand,
sequential electron transfer by means of the transient adjacent
radical pair [C60 À-(exTTF) -(exTTF)] transforms the full-
erene singlet excited state ultimately into the distant radical
pair [C60 À-(exTTF)-(exTTF) ], for which we determined
lifetimes of up to 111 ms in DMF. These values have never
previously been accomplished in molecular triads. Notable
couplings and small energy gaps between the adjacent and
distant radical pair in dyads 25 and 26 lead to a stepwise
charge recombination mechanism (Scheme 3).
.
.
‡
.
.
‡
‡
TOF): m/z (%) 1290 (100) [M ].
Triad C60-exTTF1-exTTF1 (25): 43% yield, m.p. > 3008C(decomp);
3
1H NMR (300 MHz, CDCl3:CS2, 258C): d ˆ 8.30 (d, J(H,H) ˆ 9 Hz, 2H),
8.20 (d, 3J(H,H) ˆ 9 Hz, 2H), 8.05 (brs, 2H), 7.85 (s, 1H), 7.78 7.58 (m,
8H), 7.48 (dd, 3J1(H,H) ˆ 6 Hz, 4J1(H,H) ˆ 1.2 Hz, 1H), 7.32 7.24 (m, 5H),
7.17 (m, 2H), 6.34 (m, 8H), 5.10 (s, 1H), 5.07 (d, 3J(H,H) ˆ 8.2 Hz, 1H) 4.36
(d, 3J(H,H) ˆ 8.2 Hz, 1H), 2.89 ppm (s, 3H); 13CNMR (125 MHz,
CDCl3:CS2, 258C): d ˆ 164.0, 155.6, 153.4, 152.6, 152.2, 148.4, 147.1, 146.1,
145.9, 145.7, 145.4, 145.1, 144.5, 144.2, 142.9, 142.7, 142.5, 142.4, 141.9, 141.8,
141.5, 140.0, 139.8, 139.4, 137.3, 136.8, 136.6, 136.3, 135.8, 135.4, 135.0, 134.1,
132.9, 131.9, 130.5, 130.1, 129.6, 129.3, 128.0, 127.2, 126.3, 125.9, 125.6, 125.2,
124.8, 124.5, 123.1, 121.8, 120.4, 118.6, 118.0, 117.1, 82.9, 69.8, 68.8,
39.8 ppm; FTIR (KBr): nÄ ˆ 2927, 1733, 1718, 1701, 1683, 1652, 1602, 1558,
1541, 1508, 1488, 1257, 1176, 1112, 1068, 1014, 800, 704, 526 cmÀ1; UV/Vis
(CH2Cl2): lmax (log e) ˆ 252 (4.74), 278 (4.72), 330 (4.51), 431 (4.23) nm; MS
Experimental Section
Synthesis of fulleropyrrolidine dyads and triads: general procedure: A
solution of C60 (0.138 mmol), the corresponding aldehyde (0.166 mmol for
17 and 0.064 mmol for 22, 24a and 24b) and N-methylglycine (0.5 mmol) in
toluene (100 mL) was refluxed for 16 h. Then the solvent was removed
under reduced pressure and the crude material was carefully chromato-
graphed on a silica gel column using chloroform as eluent. Further
purification was accomplished by repetitive precipitation and centrifuga-
tion by using methanol as solvent.
Dyad C60-exTTF (18): 61% yield; m.p. > 3008C(decomp); 1H NMR
(300 MHz, CDCl3:CS2, 258C): d ˆ 8.30 (d, 3J(H,H) ˆ 8.7 Hz, 2H), 8.01
(brs, 2H), 7.66 (d, 3J(H,H) ˆ 8.4 Hz, 1H), 7.52 (d, 3J(H,H) ˆ 8.4 Hz, 1H),
7.48 (d, 4J(H,H) ˆ 2.4 Hz, 1H), 7.15 (d, 4J(H,H) ˆ 2.4 Hz, 1H), 7.10 (dd,
‡
(MALDI-TOF): m/z (%): 1801 (100) [M ].
Triad C60-exTTF1-exTTF2 (26): 51% yield, m.p. > 3008C(decomp);
1H NMR (300 MHz, CDCl3:CS2, 258C): d ˆ 8.32 (d, 3J(H,H) ˆ 9.5 Hz,
2H), 8.23 (d, 3J(H,H) ˆ 8.4 Hz, 2H), 7.92 (brs, 2H), 7.78 7.66 (m, 6H), 7.60
(d, 3J(H,H) ˆ 8.4 Hz), 7.57 7.43 (m, 2H), 7.32 (m, 2H), 6.35 (m, 4H), 5.07
(s, 1H), 5.04 (d, 3J(H,H) ˆ 8.8 Hz, 1H) 4.33 (d, 3J(H,H) ˆ 8.8 Hz, 1H), 2.85
(s, 3H), 2.42 ppm (s, 12H); 13C NMR (125 MHz, CDCl3:CS2, 258C): d ˆ
164.0, 155.8, 153.1, 152.7, 152.5, 148.2, 147.0, 146.4, 145.8, 145.6, 145.4, 144.6,
Scheme 3. Possible charge-recombination mechanisms for the systems discussed.
2466
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 2457 2468

Electron Transfer Processes in Molecular Dyads and Triads
2457 2468
¬
144.4, 144.3, 143.3, 143.1, 142.7, 142.6, 142.5, 142.2, 142.2, 142.1, 142.0, 141.9,
141.7, 141.6, 140.2, 139.9, 136.2, 136.1, 136.0, 135.9, 135.6, 135.5, 135.2, 132.5,
132.3, 132.2, 131.5, 130.8, 130.7, 130.6, 130.5, 129.6, 129.5, 128.8, 126.5, 126.3,
126.0, 125.9, 125.4, 124.9, 124.7, 124.4, 123.2, 122.0, 121.9, 119.4, 119.3, 118.6,
118.3, 117.3, 82.8, 69.8, 68.8, 39.9, 19.1 ppm; FTIR (KBr): nÄ ˆ 2927, 1733,
1718, 1701, 1683, 1652, 1602, 1558, 1541, 1508, 1488, 1257, 1176, 1112, 1068,
1014, 800, 704, 526 cmÀ1; UV/Vis (CH2Cl2): lmax (log e) ˆ 252 (4.74), 278
(4.72), 330 (4.51), 431 (4.23) nm; MS (MALDI-TOF): m/z (%): 1985 (100)
MartÌn, A. Anton, J. GarÌn, J. Orduna, J. Org. Chem. 2000, 65, 1978
1983.
[11] For a recent review on TTF, see: J. L. Segura, N. MartÌn, Angew.
Chem. 2001, 113, 1416 1455; Angew. Chem. Int. Ed. 2001, 40, 1372
1409.
¬
[12] a) N. MartÌn, L. Sanchez, C. Seoane, E. OrtÌ, P. M. Viruela, R. Viruela,
J. Org. Chem. 1998, 63, 1268 1279; b) M. R. Bryce, A. J. Moore, M.
Hasan, G. J. Ashwell, A. T. Fraser, W. Clegg, A. I. Hursthouse, M. B.
Karaulov, Angew. Chem. 1990, 102, 1493 1495; Angew. Chem. Int.
Ed. Engl. 1990, 29, 1450 1452; c) Y. Yamashita, Y. Kobayashi, T.
Miyashi, Angew. Chem. 1989, 101, 1090 1091; Angew. Chem. Int. Ed.
‡
[M ].
¬
¬
Engl. 1989, 28, 1052 1053; d) N. MartÌn, I. Perez, L. Sanchez, C.
Seoane, J. Org. Chem. 1997, 62, 870 877.
Acknowledgements
¬
¬
[13] a) N. MartÌn, I. Perez, L. Sanchez, C. Seoane, J. Org. Chem. 1997, 62,
5690 5695; b) M. A. Herranz, N. MartÌn, Org. Lett. 1999, 1, 2005
We are indebted to MCYT of Spain (Project BQU2002 00855) and the
Office of Basic Energy Sciences of the U.S. Department of Energy
(contribution No. NDRL-4444 from the Notre Dame Radiation Labora-
tory) for financial support.
¬
2007; c) N. MartÌn, L. Sanchez, D. M. Guldi, Chem. Commun. 2000,
113 114.
[14] a) M. Maggini, G. Scorrano, M. Prato, J. Am. Chem. Soc. 1993, 115,
9798 9799; b) M. Prato, M. Maggini, Acc. Chem. Res. 1998, 31, 519
526.
[15] G. J. Marshallay, M. R. Bryce, J. Org. Chem. 1994, 59, 6847 6849.
¬
[16] R. Gomez, PhD thesis, Universidad Complutense de Madrid (Spain),
[1] a) J. Deisenhofer, O. Epp, K. Miki, R. Huber, H. Michel, J. Mol. Biol.
1984, 180, 385; b) W. Rettig, Angew. Chem. 1986, 98, 969 986; Angew.
Chem. Int. Ed. Engl. 1986, 25, 971 988; c) T. J. Meyer, Acc. Chem.
Res. 1989, 22, 163 170.
2001.
[17] a) Compound 20 was obtained in
a three-step procedure from
commercially available 2-hydroxymethyl-anthraquinone. See: S. Gon-
¬
zalez, N. MartÌn, D. M. Guldi, J. Org. Chem. 2003, 68, 779 791; b) I.
[2] a) J. S. Connolly, J. R. Bolton in Photoinduced Electron Transfer (Eds.:
M. A. Fox, M. Chanon), Elsevier, Amsterdam, 1988, Part D, pp. 303
393; b) M. R. Wasielewski in Photoinduced Electron Transfer (Eds.:
M. A. Fox, M. Chanon), Elsevier, Amsterdam, 1988, Part A, pp. 161
206.
¬
Perez, PhD thesis, Universidad Complutense de Madrid (Spain), 2003.
¬
[18] a) B. Illescas, M. A. MartÌnez-Grau, M. L. Torres, J. Fernandez-
Gadea, N. MartÌn, Tetrahedron Lett. 2002, 43, 4133 4136; b) T. Gu,
C. Bourgogne, J. F. Nierengarten, Tetrahedron Lett. 2001, 42, 7249
¬
7252; c) P. de la Cruz, A. de la Hoz, L. A. Font, F. Langa, M. C. Perez-
[3] a) M. R. Wasielewski, Chem. Rev. 1992, 92, 435 461; b) M. Bixon, J.
Fajer, G. Feher, J. H. Freed, D. Gamliel, A. J. Hoff, H. Levanon, D.
Mˆbius, J. R. Norris, R. Nechushtai, A. Scherz, J. L. Sessler, D. H. A.
Stehlik, Isr. J. Chem. 1992, 32, 369; c) D. Gust, T. A. Moore, A. L.
Moore, Acc. Chem. Res. 1993, 26, 198 205; d) H. Kurreck, M. Huber,
Angew. Chem. 1995, 107, 929 947; Angew. Chem. Int. Ed. Engl. 1995,
34, 849–866; e) D. Gust, T. A. Moore in The Porphyrin Handbook,
Vol. 8 (Eds.: K. M. Kadish, R. Smith, R. Guilard), Academic Press,
San Diego, CA, 2000, pp. 153 190; f) D. Gust, T. A. Moore, A. L.
Moore, Acc. Chem. Res. 2001, 34, 40 48; g) D. M. Guldi, N. MartÌn, J.
Mat. Chem. 2002, 12, 1978 1992.
Rodriguez, Tetrahedron Lett. 1998, 39, 6053 6056.
[19] a) T. Suzuki, Y. Maruyama, T. Akasaba, W. Ando, K. Kobayashi, S.
Nagase, J. Am. Chem. Soc. 1994, 116, 1359 1363; b) L. Echegoyen,
L. E. Echegoyen, Acc. Chem. Res. 1998, 31, 593 601.
[20] The aforementioned redox potentials indicate the absence of signifi-
cant electronic interaction between the electroactive chromophores in
18, 23, 25, and 26.
[21] We would like to point out that the missing data points in Table 2
could not be determined within the 100 ps time-resolution of the
apparatus. The complementary picosecond-resolved transient absorp-
tion measurements infer indeed lifetimes that are either close to or
within the 100 ps apparatus time-resolution.
[4] a) J.-C. Chambron, S. Chardon-Noblat, A. Harriman, V. Heitz, J. P.
Sauvage, Pure Appl. Chem. 1993, 65, 2343; b) A. Harriman, J.-P.
Sauvage, Chem. Soc. Rev. 1996, 26, 41 48; c) M.-J. Blanco, M. C.
[22] G. Kodis, P. A. Liddell, L. de la Garza, A. L. Moore, T. A. Moore, D.
Gust, J. Mater. Chem. 2002, 2100 2108.
¬
Jimenez, J.-C. Chambron, V. Heitz, M. Linke, J.-P. Sauvage, Chem.
[23] D. M. Guldi, M. Prato, Acc. Chem. Res. 2000, 33, 695 703.
Soc. Rev. 1999, 28, 293 306; d) V. Balzani, A. Juris, M. Venturi, S.
Campagna, S. Serroni, Chem. Rev. 1996, 96, 759 834; e) Electron
Transfer in Chemistry (Ed.: V. Balzani), Wiley-VCH, Weinheim, 2001.
[5] a) M. N. Paddon-Row, Acc. Chem. Res. 1994, 27, 18 25; b) J. W.
Verhoeven in Electron Transfer–From Isolated Molecules to Bio-
molecules, Part 1, (Eds.: J. Jortner, M. Bixon), Wiley, New York, 1999,
pp. 603 644; c) K. Maruyama, A. Osuka, Pure Appl. Chem. 1990, 62,
1511; d) K. Maruyama, A. Osuka, N. Mataga, Pure Appl. Chem. 1994,
66, 867; e) A. Osuka, N. Mataga, T. Okada, Pure Appl. Chem. 1997, 69,
796 802; f) L. Sun, L. Hammarstrˆm, B. äkermark, S. Styring, Chem.
Soc. Rev. 2001, 30, 36 49.
¬
[24] D. M. Guldi, L. Sanchez, N. MartÌn, J. Phys. Chem. B 2001, 105, 7139
7144.
[25] The charge-recombination rates are smaller in polar than in nonpolar
solvents. This relationship suggests that the rate constants are in the
normal region of the Marcus parabola (ÀDGCR < l). Considering the
structural changes that the one-electron oxidation of exTTFs causes,
the reorganization energy of the radical pair is assumed to be quite
large. In addition, the ÀDGCR values are insufficiently large as to
surpass the reorganization energy in 4a, 18, and 23 (ÀDGCR ꢀ l),
which would shift the charge-recombination into the inverted region
of the Marcus parabola (ÀDGCR > l).
[6] a) D. M. Guldi, K.-D. Asmus, J. Am. Chem. Soc. 1997, 119, 5744
5745; b) H. Imahori, K. Hagiwara, T. Akiyama, M. Aoki, S. Taniguchi,
T. Okada, M. Shirakawa, Y. Sakata, Chem. Phys. Lett. 1996, 263, 545
550; c) D. M. Guldi, Chem. Soc. Rev. 2002, 31, 22 36.
[26] Interestingly, the lifetime of 23, which is in the range of several
hundred nanoseconds, is virtually comparable to those found in C60
donor dyads with a similar spacer (see reference [6c]).
[27] The reactivity scheme of 4a, 18, and 23–comparing the steady-state
and time-resolved fluorescence experiments–fails to reveal substan-
tial differences. Therefore, only time-resolved fluorescence lifetime
measurements will be discussed for triads 25 and 26. In addition, the
presence of two exTTF moieties imposes notable difficulties on
convoluting the individual contributions in the steady-state emission
experiments.
[28] The intramolecular decay rates in triad 26 do not differ notably from
those of dyad 18.
[29] Additional proof for an intramolecular charge-recombination stems
from a set of experiments, conducted with different laser power and
different dyad concentrations: Decreasing the radical-pair concen-
[7] a) Fullerenes: From Synthesis to Optoelectronic Properties (Eds.:
D. M. Guldi, N. MartÌn), Kluwer, Dordrecht, 2002; b) Fullerenes and
Related Structures, Vol. 199 (Ed.: A. Hirsch), Springer, Berlin, 1999.
¬
¬
[8] a) N. MartÌn, L. Sanchez, B. Illescas, I. Perez, I. Chem. Rev. 1998, 98,
2527 2548; b) H. Imahori, Y. Sakata, Adv. Mater. 1997, 9, 537 546;
c) H. Imahori, Y. Sakata, Eur. J. Org. Chem. 1999, 64, 2445 2457.
[9] a) P. Barbara, T. J. Meyer, M. A. Ratner, J. Phys. Chem. 1996, 100,
13148 13168; b) P. Piotrowiak, Chem. Soc. Rev. 1999, 28, 143 150;
c) C. C. Moser, J. M. Keske, K. Warncke, R. S. Farid, P. L. Dutton,
Nature 1992, 355, 796 802.
¬
[10] a) N. MartÌn, L. Sanchez, M. A. Herranz, D. M. Guldi, J. Phys.
¬
Chem. A. 2000, 104, 4648 4657; b) D. M. Guldi, S. Gonzalez, N.
Chem. Eur. J. 2003, 9, 2457 2468 www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2467

N. MartÌn, D. M. Guldi et al.
FULL PAPER
tration in various increments by up to 65% failed to lead to any
notable changes in the charge-recombination kinetics.
[30] Determined by means of DGCS ˆ ÀRTlnk for equilibrium conditions
in Equation (4).
derived from the intercept of the Arrhenius plot (see Figure 10).
Therefore, the only reasonable rationale for this trend implies a
sequential electron transfer ([Eq. (5)]), that is a thermally activated
.
.
‡
process, via the intermediate C60 À-(exTTF)1 -(exTTF)2 rather than
electron tunneling by means of superexchange ([Eq. (6)]).
[33] The smaller energy gap in triad 25 (0.06 eV) expedites the charge
recombination.
[31] Spectroscopically, the 20 nm difference between the two oxidized
.
.
‡
‡
species, that is, (exTTF)1 and (exTTF)2 opened the possibility for
us to successfully analyze the charge-shift rates, by following the band
grow in at 660 nm. The rate constants are 7.7 Â 108 sÀ1 (toluene), 8.2 Â
108 sÀ1 (THF), 9.3 Â 108 sÀ1 (benzonitrile) and 1.2 Â 109 sÀ1 (DMF).
[32] Further support for this assumption comes from the very slow charge
Received: October 10, 2002
Revised: December 16, 2002 [F4494]
recombination rate under barrier-less conditions 3.4 Â 107 sÀ1
, as
2468
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 2457 2468