A metallofullerene electron donor that powers an efficient spin flip in a linear electron donor-acceptor conjugate

Article abstract of DOI:10.1021/ja403763e

The dream target of artificial photosynthesis is the realization of long-lived radical ion pair states that power catalytic centers and, consequently, the production of solar fuels. Notably, magnetic field effects, especially internal magnetic field effects, are rarely employed in this context. Here, we report on a linear Lu₃N@I_h-C₈₀-PDI electron donor-acceptor conjugate, in which the presence of the Lu₃N cluster exerts an appreciable electron nuclear hyperfine coupling on the charge transfer dynamics. As such, a fairly efficient radical ion pair intersystem crossing converts the initially formed singlet radical ion pair state, ¹[(Lu₃N@I_h-C₈₀)^?+- PDI^?-], to the corresponding triplet radical ion pair state, ³[(Lu₃N@I_h-C₈₀)^?+- PDI^?-]. Most notably, the radical ion pair state lifetime of the latter is nearly 1000 times longer than that of the former.

Full text of DOI:10.1021/ja403763e

Article
pubs.acs.org/JACS
A Metallofullerene Electron Donor that Powers an Efficient Spin Flip
in a Linear Electron Donor−Acceptor Conjugate
Marc Rudolf,¶ Lai Feng,*,†,‡ Zdenek Slanina,‡ Takeshi Akasaka,*,‡,§,∥,⊥ Shigeru Nagase,□
and Dirk M. Guldi*,¶
†Jiangsu Key Laboratory of Thin Films and School of Energy, Soochow University, Suzhou 215006, China
‡Life Science Center of Tsukuba Advanced Research Alliance, University of Tsukuba, Tsukuba 305-8577, Japan
§College of Materials Science and Engineering, Huazhong University of Science and Technology, Wuhan 430074, China
∥Foundation for Advancement of International Science, Tsukuba 305-0821, Japan
⊥Department of Chemistry, Tokyo Gakugei University, Tokyo 305-0821, Japan
¶Department of Chemistry and Pharmacy and Interdisciplinary Center for Molecular Materials, Friedrich-Alexander-Universitat
̈
Erlangen-Nurnberg, 91058 Erlangen, Germany
̈
□Fukui Institute for Fundamental Chemistry, University of Kyoto, Kyoto 606-8103, Japan
S
* Supporting Information
ABSTRACT: The dream target of artificial photosynthesis is the
realization of long-lived radical ion pair states that power catalytic
centers and, consequently, the production of solar fuels. Notably,
magnetic field effects, especially internal magnetic field effects, are
rarely employed in this context. Here, we report on a linear Lu3N@Ih-
C80−PDI electron donor−acceptor conjugate, in which the presence of
the Lu3N cluster exerts an appreciable electron nuclear hyperfine
coupling on the charge transfer dynamics. As such, a fairly efficient
radical ion pair intersystem crossing converts the initially formed
singlet radical ion pair state, 1[(Lu3N@Ih-C80)•+−PDI•−], to the
corresponding triplet radical ion pair state, 3[(Lu3N@Ih-C80)•+−
PDI•−]. Most notably, the radical ion pair state lifetime of the latter
is nearly 1000 times longer than that of the former.
been emerging as novel electron acceptors or even as electron
INTRODUCTION
■
donors in a variety of photofunctional ensembles taking
advantage of their unique redox properties.5−8 Specifically,
Lu3N@C80 and La2@C80 both exhibit much lower oxidation
potentials than C60 and, therefore, exhibit much better electron-
donating property.9 Nevertheless, only a few examples, in which
electron acceptors, such as 1,6,7,12-tetrachloro-3,4,9,10-peryle-
nediimide (PDI) or 11,11,12,12-tetracyano-9,10-anthra-p-qui-
nodimethane (TCNQ), are linked to endohedral metal-
lofullerenes via a flexible σ-spacer or a pyrrolidine ring, were
reported.5f,6d Related computational studies documented that
in these electron donor−acceptor conjugates the two photo-
and redox-active constituents tend to approach each other quite
closely due to strong van der Waals interactions and/or the
short covalent linkages. However, such close-contact geo-
metries are unfavorable for long-lived charge separation events
as revealed by photophysical investigations. As such, a structural
or geometrical modification of the electron donor−acceptor
conjugate shall be taken into consideration.
In recent years, a large number of covalently linked electron
donor−acceptor conjugates have been developed to mimic
photosynthesis and to realize optoelectronic devices.1,2
Comprehensive studies have been performed to address the
mechanistic details regarding charge transfer events in these
electron donor−acceptor conjugates and the fate of the
subsequently formed radical ion pair states. It has been
demonstrated that the magnitude of charge transfer and the
lifetimes of radical ion pair states depend on a series of
parameters, including the nature of redox-active constituents,
the length and properties of the linker, the topology and
relative orientation of the individual components, etc.
Importantly, the control over the electron donor−acceptor
geometry has been considered as the most versatile means to
impact the performance of electron donor−acceptor con-
jugates, especially in terms of radical ion pair state lifetimes.
Fullerenes C60 and C70 have been widely employed as
electron acceptor due to their remarkable reduction properties
and their small reorganization energies in charge transfer
reactions.3,4 Recently, several endohedral metallofullerenes,
such as M3N@C80, M2@C80, La@C82, Li+@C60, etc., have
Received: April 16, 2013
Published: June 27, 2013
© 2013 American Chemical Society
11165
dx.doi.org/10.1021/ja403763e | J. Am. Chem. Soc. 2013, 135, 11165−11174

Journal of the American Chemical Society
Article
Scheme 1. Structures of References Used in This Study (up), and Synthesis Scheme of Conjugates Lu3N@Ih-C80−PDI (6) and
a
C60−PDI (7) (down):
a
Reagents and conditions: i) Lu3N@C80, NaOMe, pyridine, o-DCB, 80-85 °C, then HPLC, 42%; ii) C60, NaOMe, pyridine, o-DCB, 70-75 °C; iii) o-
DCB, 150 °C, then HPLC, 73%.
Magnetic field effects have shown that photoexcitation of
some rigid electron donor−acceptor conjugates leads to
surprisingly long-lived radical ion pair states. It is especially
pertinent for linear or rod-like electron donor−acceptor
conjugates that are susceptible to changes exerted by either
internal or external magnetic fields.10a,b It is important in this
context that photoexcitation may generate two types of radical
ion pair states, namely singlet and triplet radical ion pair states.
Their interconversion is typically modulated by contributions
from spin rotation, spin−orbit coupling, the hyperfine
interaction of the unpaired electron with strongly coupled
nuclei, or differences in the electron g-factors of the two radical
ions.10c Charge recombination of triplet radical ion pair states
to form the singlet ground state is a spin-forbidden process and,
in turn, potentially slower than that seen in singlet radical ion
pair states. Its rate depends on the rate of spin conversion and
is determined by a number of factors (vide supra). As such, the
distance between the individual radical ions and the nature of
the spacer have decisive roles.
dissecting the influence of a linear geometry on the charge
transfer properties, in general, and of the Lu3N cluster on the
interconversion in radical ion pair states, in particular. To this
end, an arsenal of physicochemical tools have been employed to
characterize this linear Lu3N@Ih-C80−PDI conjugate in the
ground, reduced, oxidized, and excited states. Additional
computational studies shed light on the geometry effect on
the ground-state electronic configuration of Lu3N@Ih-C80−
PDI, which might be relevant to the charge transfer properties.
RESULTS AND DISCUSSION
■
The synthesis of Lu3N@Ih-C80−PDI (6) was carried out by a
[1+2]-cycloaddition reaction of a diazo compound that was
generated in situ by a Bamford−Stevens reaction between
tosylhydrazone 5 and sodium methoxide, see Scheme 1. To this
end, the synthesis of the target tosylhydrazone 5 bearing PDI
was designed starting with the 1,6,7,12-tetrachloroperylene-
3,4,9,10-tetracarboxylic dianhydride 1 and the commercially
available 4-aminobenzoacetone as well as n-docylamine
(Scheme S1, Supporting Information [SI]). Condensation of
3 with p-tosylhydrazide in 1,2-dichloroethane afforded
tosylhydrazone 5. A closer look at the HPLC revealed the
formation of 6 as major product in a yield of 42%, when a large
excess 5 was employed to react with Lu3N@Ih-C80. C60−PDI
(7) was synthesized in a similar way (Scheme S2, SI) and was
Herein, we report a linear electron donor−acceptor
conjugate, in which Lu3N@Ih-C80 has been incorporated as
electron donor, while perylenediimide serves as electron
acceptor. Specifically, a PDI moiety is covalently linked to a
Lu3N@Ih-C80 in a 6,6-fulleroid fashion via a phenyl group. Full-
fledged assays have been performed with particular emphasis on
11166
dx.doi.org/10.1021/ja403763e | J. Am. Chem. Soc. 2013, 135, 11165−11174

Journal of the American Chemical Society
Article
used as reference (vide infra). A comparison showed that the
reaction of Lu3N@Ih-C80 proceeds more slowly than that of
C60, reflecting the lower reactivity of Lu3N@Ih-C80 relative to
that of C60.
The purities of the conjugates Lu3N@Ih-C80−PDI (6) and
C60−PDI (7) were checked by HPLC (Figures S3 and S4, SI),
and their compositions were confirmed via MALDI-TOF mass
spectroscopy. Specifically, in the mass spectra of 6 and 7, each
shows a single molecular ion peak at 2298 m/z or 1518 m/z,
respectively. Also, their isotopic distributions agree well with
the theoretical calculations, confirming their overall composi-
tion as Lu3C124H36O4N3Cl4 or C104H36O4N2Cl4. (Figure S5,
SI). Furthermore, the structures of these conjugates were
verified by means of 1H, 13C, and a series of 2D NMR
experiments (i.e., COSY, DEPT, HMQC, and HMBC).
1
Accordingly, all the H signals and most 13C signals can be
unambiguously assigned (see SI). Particularly, the HMBC
NMR experiment could demonstrate structural information
more directly: as highlighted in Figure 1, the long-distance H−
Figure 2. Differential pulse voltammograms of Lu3N@Ih-C80−PDI (6)
(up) and C60−PDI (7) (down) in o-dichlorobenzene (0.05 M (n-
Bu)4NPF6 with ferrocene as internal standard, scan rate: 20 mV s−1).
dichlorobenzene with 0.05 M tetra-n-butylammonium hexa-
fluorophosphate (TBAPF6) as a supporting electrolyte and
using ferrocene as an internal standard. Lu3N@Ih-C80−PDI (6)
reveals four one-electron reductions at −0.85, −1.08, −1.41,
and −1.90 V as well as two one-electron oxidations at +0.55
and +1.07 V in the electrochemical window between −2.0 and
1.3 V (see Table 1). The first and the second reductions of
Lu3N@Ih-C80−PDI (6), which are fully reversible as confirmed
by CV, agree well with those of PDI-ref (3) (Figure S19, SI)
and, in turn, are assigned to the PDI moiety. On the other
hand, other reductions and oxidations correspond nicely to
those of Lu3N@C80-ref (9), reflecting the redox activity of
Lu3N@C80 in Lu3N@Ih-C80−PDI (6). The fact that the first
oxidation and the first reduction of Lu3N@Ih-C80−PDI (6) are
very similar to those seen for the references suggests the lack of
appreciable intramolecular interactions between Lu3N@Ih-C80
and PDI and/or any electronic perturbation. As for C60−PDI
(7), the redox data are comparable to what has been discussed
for Lu3N@Ih-C80−PDI (6). Specifically, its DPV reveals three
reductions at −0.85, −0.11, and −1.51 V as well as one
oxidation +1.11 V. Notably, the second reduction is a two-
electron process and involves the simultaneous reduction of
PDI and C60. Other reductions and oxidation are one-electron
processes and are either centered on PDI or on C60. Similar
results were also derived from CV.
Next we performed computational studies to shed light onto
the geometrical and electronic features of the conjugates
Lu3N@Ih-C80−PDI (6) and C60−PDI (7). All of the structures
were optimized at the M06-2X11/3-21G12∼SDD13 level using
Gaussian 09 program package.14 As for Lu3N@Ih-C80−PDI (6),
three conformers, in which the Lu3N clusters are differently
oriented, have been taken into consideration. Their orbital
analysis and the relative energies were calculated at the same or
higher level and are listed in Table 2. Owing to the fact that all
the conformers of 6 show different relative energies, a hindered
rotation of the Lu3N cluster can be proposed. In fact, the most
favorable conformer is that with the Lu3N cluster crossing the
plane of the bridgehead carbons and the spiro carbon in an
angle of 45°. The center-to-center and the edge-to-edge
Figure 1. HMBC spectrum of Lu3N@Ih-C80−PDI (6) with observed
couplings between CH3/bridgehead carbons as well as CH3/spiro
carbon.
C couplings between methyl protons and the bridgehead
carbons as well as the spiro carbon were clearly observed after
320 scans, thus confirming the overall structure of 6 as
presented in Scheme 1.
In addition, conjugates Lu3N@Ih-C80−PDI (6) and C60−PDI
(7) were characterized electrochemically by means of differ-
ential pulse voltammetry (DPV) and cyclic voltammetry (CV)
(Table 1, Figures 2 and S17 and S18 in SI). All of the
measurements were carried out at room temperature in o-
a
Table 1. Redox Potentials of Conjugates 6 and 7 and
References
2Eox
1Eox
1Ered
2Ered
3Ered
4Ered
6
7
3
9
1.07
0.55
1.11
−0.85
−0.85
−0.84
−1.46
−1.08
−1.11
−1.08
−1.94
−1.41
−1.51
−1.90
b
b
0.55
a
All the potentials, in volts, were measured relative to the Fc0/+ couple
b
by means of DPV. Two-electron reduction process.
11167
dx.doi.org/10.1021/ja403763e | J. Am. Chem. Soc. 2013, 135, 11165−11174

Journal of the American Chemical Society
Article
oxidation of Lu3N@C80 relative to C60 and the comparable
reductions of PDI in 6 and 7. When compared to the folded
conformer of 10, 6 features slightly higher LUMO/HOMO
energy levels. From the latter we conclude for this linearly
conjugated Lu3N@Ih-C80 and PDI the lack of intramolecular
interactions in the ground state.
Table 2. HOMO/LUMO Energies (eV) of Conjugates 6, 7,
10 at the M06-2X/3-21G∼SDD Level as well as the Relative
Energies (kcal/mol) of 6 with Differently Rotated Lu3N
Cluster
c,d
6
7
10
a
rotation angle
90°
45°
0°
The absorption spectra of the as-prepared conjugates Lu3N@
Ih-C80−PDI (6) and C60−PDI (7) are shown in Figure 3b. A
close inspection reveals that the absorptions of 6 and 7 are
simply the superimposition of the spectral features of the
individual constituents, that is, PDI, on one hand, and Lu3N@
Ih-C80 and C60, on the other hand, lacking any notable
perturbations or additional charge-transfer transitions. In
general, conjugates 6 and 7 display intense absorptions at
489 and 523 nm, respectively, which relate to PDI-centered
transitions according to the reference spectrum of PDI-ref (3).
Besides, Lu3N@Ih-C80−PDI (6) gives rise to strong absorptions
below 400 nm and broad bands at around 410 and 680 nm,
which are due to Lu3N@Ih-C80, while C60−PDI (7) reveals
weaker bands at 330 and 430 nm, which are due to C60. The
large electron donor−acceptor separation of 6.1 Å is likely to be
responsible for the lack of interactions in the electronic ground
state. The latter also prevents detectable interactions in
similarly spaced electron donor−acceptor conjugates.15
LUMO
HOMO
ΔE
−3.29
−6.84
2.026
1.973
−3.28
−6.75
0
−3.28
−6.78
0.336
1.650
−3.25
−7.37
−
−3.40
−6.91
b
ΔE
0
−
a
Refers to the angle between Lu3N cluster and the coplanar
bridgehead carbons and spiro carbon. Relative energies (kcal/mol)
at the M06-2X/6-31G*∼SDD level. Folded conformer of 10. Data
from reference 5f.
b
c
d
distances between the two chromophores in 6 are 15.3 and 6.1
Å, while in 7 they are 14.7 and 6.1 Å, respectively. Notably,
these distances are substantially larger than those previously
reported for the folded conformer of Lu3N@Ih-C80−PDI (10)
(see Scheme 1), indicating a longer pathway of photoinduced
electron-transfer. Orbital analyses of 6 and 7 suggest that their
LUMOs are mainly localized on the PDI moiety while their
HOMOs are mainly localized on the fullerenes (see Figure 3a).
In particular, as for 6, the contribution of the Lu3N cluster to
the HOMO is almost negligible, predicting an electron-transfer
induced oxidation that is centered on the cage of Lu3N@Ih-C80.
By comparison, the LUMO energies of 6 and 7 are very close to
each other, whereas the calculated HOMO energy of 6 is 0.62
eV higher than that of 7. Overall, the computational results
agree well with the electrochemical data, suggesting the easier
Different are the interactions in the excited state. For
example, fluorescence assays with Lu3N@Ih-C80−PDI (6) and
C60−PDI (7) reveal a rather strong quenching of the PDI
fluorescence. On one hand, PDI-ref (3) reveals a high solvent-
independent fluorescence quantum yield of 0.91. On the other
hand, the fluorescence quantum yields for 7 were 0.007
(toluene), 0.009 (chlorobenzene), and 0.010 (benzonitrile),
Figure 3. (a) LUMO/HOMO distributions of Lu3N@Ih-C80−PDI (6) (left panel) and C60−PDI (7) (right panel) at the M06-2X/3-21G∼SDD
level; (b) Absorption spectra of 6 (εmax: 61010), 7 (εmax: 47050), PDI-ref (3) (εmax: 44770), and Lu3N@C80-ref (9) in toluene at 298 K.
11168
dx.doi.org/10.1021/ja403763e | J. Am. Chem. Soc. 2013, 135, 11165−11174

Journal of the American Chemical Society
Article
Figure 4. (a) Differential absorption spectra (visible and near-infrared) obtained upon femtosecond flash photolysis (530 nm) of Lu3N@Ih-C80−PDI
(6) (10−5 M) in argon-saturated chlorobenzene with several time delays between 0 and 6750 ps at room temperature. (b) Time-absorption profiles
of the spectra shown in (a) at 565 and 918 nm monitoring the charge separation and the charge recombination. (c) Differential absorption spectra
(visible and near-infrared) obtained upon nanosecond flash photolysis (532 nm) of 6 (10−5 M) in argon-saturated chlorobenzene with several time
delays between 25 and 150 ns at room temperature. (d) Time-absorption profiles of the spectra shown in (c) at 680 and 760 nm monitoring the
charge recombination.
while for 6 they were 0.0010 (toluene), 0.0009 (chloroben-
zene), 0.0005 (benzonitrile), and 0.0011 (DMF). Notably, the
weak fluorescent features of C60-ref (11) and Lu3N@C80-ref
(9) with quantum yields on the order of 10−4 or less hampered
any meaningful quantitative analyses.
As for C60-ref (11), differential absorption changes evolve
immediately after the 387 nm laser pulse that are characterized
by maxima at 440 and 515 nm and broad absorption that spans
from 700 to 1300 nm. The C60 singlet excited state features
(1.8 eV) decay somewhat slower (i.e., 1.5 0.1 ns) than those
of Lu3N@Ih-C80 to afford the energetically lower-lying triplet
excited state. In complementary nanosecond experiments, the
triplet excited state of C60 (1.5 eV) with a transient maximum at
720 nm gives rise to a lifetime of up to 20 μs when molecular
oxygen is absent. Otherwise, a diffusion-controlled deactivation
(1010 M−1 s−1) of the C60 triplet excited state sets in and
produces singlet oxygen.
In contrast to the 387 nm laser excitation of Lu3N@C80-ref
(9) and C60-ref (11), PDI-ref (3) reveals upon excitation at 530
nm differential absorption changes that include transient
maxima at 700, 775, 800, 845, 890, and 970 nm as well as
transient minima at 490 and 515 nm. These features relate to
PDI singlet−singlet transitions, which decay with 4.0 0.2 ns.
It is, however, the ground state rather than the triplet excited
state (1.2 eV) that is nearly exclusively populated due to an
inefficient intersystem crossing in 3. Complementary nano-
second excitation at 532 nm further corroborates the spin-
allowed ground state recovery, that is, the lack of an appreciable
transient (not shown).
Further insights into the excited state deactivation in the
reference chromophores 9, 11, 3, and the conjugates 6, 7, in
general, and into the corresponding photoproducts, in
particular, came from transient absorption measurements
following femtosecond and nanosecond excitation. We
interpret the differential absorption changes upon 387 nm
excitation of 9 as population of the Lu3N@Ih-C80 singlet excited
state (1.78 eV), which features characteristic absorption
maxima at 460, 500, and 580 nm and a broad absorptions
that spans from 700 to 1500 nm and that maximizes around
1300 nm (see Figure S21 in SI). The latter undergoes, however,
a fast intersystem crossing, 20 10 ps, to the triplet manifold
due to the presence of the Lu3N cluster. The newly developing
bands at 460 and 580 nm and a broad band that spans from 700
up to 1600 nm with maxima at 830 and 1260 nm reflects the
diagnostic signature of the Lu3N@Ih-C80 triplet excited state
(1.4 eV), from which a lifetime of 300 10 ns has been derived
in nanosecond experiments upon 355 nm excitation.
11169
dx.doi.org/10.1021/ja403763e | J. Am. Chem. Soc. 2013, 135, 11165−11174

Journal of the American Chemical Society
Article
Figure 5. (a) Differential absorption spectra (visible and near-infrared) obtained upon femtosecond flash photolysis (530 nm) of Lu3N@Ih-C80−PDI
(6) (10−5 M) in argon-saturated benzonitrile with several time delays between 0 and 6750 ps at room temperature. (b) Time-absorption profiles of
the spectra shown in (a) at 490 and 925 nm monitoring the charge separation and the charge recombination. (c) Differential absorption spectra
(visible and near-infrared) obtained upon nanosecond flash photolysis (532 nm) of 6 (10−5 M) in argon-saturated benzonitrile with several time
delays between 50 and 300 ns at room temperature. (d) Time-absorption profiles of the spectra shown in (c) at 560 and 760 nm monitoring the
charge recombination.
A study of Lu3N@Ih-C80−PDI (6) reveals, commencing with
the conclusion of the 530 nm excitation, distinct differential
absorption changes in the visible as well as in the near-infrared.
Specifically, transient maxima at 710, 755, 795, 845, 910, and
990 nm as well as transient minima at 490 and 515 nm are
discernible. In line with the reference experiments with PDI-ref
(3), we assign these changes to the PDI singlet exited state.
Instead of seeing, however, the slow intersystem crossing, the
PDI singlet excited state decays ultra-rapidly with lifetimes of
2.2 0.1 ps (toluene), 2.2 0.1 ps (chlorobenzene), 2.2 0.1
PDI•− radical ion pair state originates from a localized singlet
excited state, its spin state is likely that of a singlet. The
quantum efficiencies of the (Lu3N@Ih-C80)•+−PDI•− formation
are close to unity (>99.9%) in any of the given solvents, as
determined by correlating the rate constants for intrinsic
excited state deactivation with that of the charge transfer, and
its energy ranges from 1.42 to 1.10 eV when going from toluene
to DMF. In toluene, chlorobenzene, and THF, the (Lu3N@Ih-
C80)•+−PDI•− singlet radical ion pair state decays via a mono-
exponential rate expression to the PDI triplet excited state. In
ps (THF), 2.3
0.1 ps (benzonitrile), and 2.7
0.1 ps
particular, lifetimes of 160
20 ps (toluene), 610
40 ps
(DMF). Simultaneously with the latter decay, new transitions
grow-in in the visible and the near-infrared. In fact, the
spectrum as shown in Figures 4 and 5 bears close resemblance
in the visible, namely maxima around 680, 765, 925, and 1025
nm, with the one-electron-reduced PDI π-radical anion
(PDI•−) shown in Figure S22 in SI. In the near-infrared, a
broad near-infrared tail is seen that correlates well with the one-
electron oxidized Lu3N@Ih-C80 π-radical cation [(Lu3N@Ih-
C80)•+]. Thus, we conclude that an energetically low-lying
radical ion pair state in the form of (Lu3N@Ih-C80)•+−PDI•−
evolves from an intramolecular charge transfer between the
electron-donating Lu3N@Ih-C80 and the electron-accepting PDI
singlet excited state. Considering that the (Lu3N@Ih-C80)•+−
(chlorobenzene), and 1100 200 ps (THF) were derived from
the decay of the metastable (Lu3N@Ih-C80)•+−PDI•− singlet
radical ion pair state fingerprints as well as the growth of the
characteristic PDI triplet excited state marker at 560 nm. The
(Lu3N@Ih-C80)•+−PDI•− singlet radical ion pair state is also
metastable in benzonitrile (40 8 ps) and DMF (25 5 ps).
As far as the product of charge recombination is concerned, we
note that in benzonitrile likewise the features of the PDI triplet
excited state evolve, which are, however, completely lacking in
DMF.16
Next, we took a closer look at the correlation between charge
transfer rates (i.e., charge separation and charge recombination)
and free energy changes for the underlying reaction. The free
11170
dx.doi.org/10.1021/ja403763e | J. Am. Chem. Soc. 2013, 135, 11165−11174

Journal of the American Chemical Society
Article
Figure 6. Energy level diagrams of Lu3N@Ih-C80−PDI (6) in toluene and chlorobenzene, benzonitrile, and DMF reflecting the different pathways of
electron transfer.
energy changes were determined from the oxidation potential
of Lu3N@Ih-C80, the reduction potential of PDI, and the
solvent correction term based on the dielectric continuum
model, together with the energy level of the PDI singlet excited
state. To this end, a parabolic dependence, as illustrated in
Figure S23 in SI, with parameters of the reorganization energy
(0.84 eV) and electronic coupling (59.9 cm−1) is found for
Lu3N@Ih-C80−PDI (6). This particular finding is of great value,
since it is the first quantitative manifestation that corroborates
the small reorganization energy of fullerene electron donors,
that is Lu3N@Ih-C80, in photoinduced electron transfer
reactions.
intermolecular electron transfer. From the long-lived features
3
of the [(Lu3N@Ih-C80)•+−PDI•−] triplet radical ion pair state
we calculated the overall quantum yields as 0.4 0.2, 4.3 0.5,
5.5 0.5, and 5.2 0.5% in chlorobenzene, THF, benzonitrile,
and DMF, respectively.17
In a nutshell, photoexcited Lu3N@Ih-C80−PDI (6) deacti-
vates through the following sequence of events (see Figure 6).
First, the PDI-centered singlet excited state (2.32 eV)
deactivates via an exergonic intramolecular charge transfer
from the electron-donating Lu3N@Ih-C80 to the PDI singlet
1
excited state to yield the [(Lu3N@Ih-C80)•+−PDI•−] singlet
radical ion pair state. The free energy changes range from
To fully explore the charge recombination mechanism we
performed complementary nanosecond experiments with
Lu3N@Ih-C80−PDI (6) and 532 nm excitation (see Figures 4
and 5). In line with the femtosecond experiments, the product
of charge recombination differs as the solvent polarity is varied.
On one hand, toluene, chlorobenzene, and THF clearly favor
the PDI triplet excited state formation. Additional proof came
from adding molecular oxygen in different concentrations,
which resulted in diffusion-controlled formation of singlet
oxygen (vide infra). The rate constants are (1.1 0.2) × 109,
(1.0 0.2) × 109, and (1.3 0.2) × 109 M−1 s−1 in toluene,
chlorobenzene, and THF, respectively. On the other hand,
charge recombination to the PDI triplet excited state plays only
a minor role in benzonitrile, while no spectroscopic evidence is
seen at all for the PDI triplet excited state in DMF. In Lu3N@
Ih-C80−PDI (6), the difference between these scenarios is a
radical ion pair state energy of around 1.2 eVan energy that
relates well to that of the PDI triplet excited state. Nevertheless,
we noticed persistent features of the (Lu3N@Ih-C80)•+−PDI•−
radical ion pair state in all of the tested solvents on the
nanosecond time regime with lifetimes of 102 5 ns (toluene),
−0.90 to −1.22 eV in toluene and DMF, respectively. Second,
1
the [(Lu3N@Ih-C80)•+−PDI•−] singlet radical ion pair state
undergoes charge recombination that differs as the solvent
polarity is changed. On one hand, toluene and chlorobenzene
favor the PDI triplet excited state formation, which is in line
with the free energy changes for charge recombination to the
PDI triplet excited state of −0.22 and −0.02 eV in toluene and
chlorobenzene, respectively. On the other hand, charge
recombination to populate the PDI triplet excited state plays
only a minor role in benzonitrile, while no spectroscopic
evidence for the PDI triplet excited state is seen at all in DMF.
Increasing the solvent polarity to benzonitrile and DMF, the
energy of the radical ion pair state drops below that of the
energy of the PDI triplet excited state (1.2 eV). In fact, the
dielectric continuum model suggests that the PDI triplet
excited state formation in benzonitrile and DMF should be
slightly endergonic, namely, energetically uphill by +0.09 and
+0.10 eV, respectively. Instead, we note that the (Lu3N@Ih-
C80)•+−PDI•− radical ion pair state decays without, however,
changing its spectroscopic features. As such, we describe this
deactivation to a singlet radical ion pair state to triplet radical
ion pair state conversion, which is facilitated by the Lu3N
cluster. The spin change of the radical ion pair must be in
competition with charge recombination to the PDI triplet state.
Both spin states of the (Lu3N@Ih-C80)•+−PDI•− radical ion
pair convert irreversibly to the PDI triplet excited state in
toluene and chlorobenzene. Interesting is the scenario in
benzonitrilehere, the radical ion pair states lie close in energy
to that of the PDI triplet excited state and set up a
thermodynamic equilibrium.18 Third, the 3[(Lu3N@Ih-
C80)•+−PDI•−] triplet radical ion pair decays via quantitative
regeneration of the ground state. Finally, the PDI triplet excited
state, for which singlet oxygen quantum yields of 0.43, 0.80,
0.17, and 0.08 were determined in toluene, chlorobenzene,
THF, and benzonitrile, respectively, decays likewise to the
41
4 ns (chlorobenzene), 70
5 ns (THF), 44
4 ns
(benzonitrile), and 58
4 ns (DMF). Taking the
aforementioned into concert, we hypothesize a different spin
nature to be responsible for the short-lived and long-lived
(Lu3N@Ih-C80)•+−PDI•− radical ion pair state. On the femto-
and picosecond time scale, it is the singlet radical ion pair state,
while the triplet radical ion pair state is persistent on the
nanosecond time scale. Importantly, the decay curves were well
fit by a single exponential decay component throughout the
experimental time scale in the absence of molecular oxygen. In
the presence of different concentrations of molecular oxygen,
an activation controlled decay(2.2 4.4) × 108 M−1 s−1of
the one-electron-reduced PDI π-radical anion sets in in THF
•−
and produces superoxide radical anion O2
via an
11171
dx.doi.org/10.1021/ja403763e | J. Am. Chem. Soc. 2013, 135, 11165−11174

Journal of the American Chemical Society
Article
ground state. Notable is the sharp drop in quantum yields in
the latter two solvents when compared to the earlier two.
Moreover, in DMF no appreciable singlet oxygen emission is
detected. This again backs up our notion of a change in excited
state deactivation, namely triplet excited state formation versus
triplet radical ion pair state before the ground state is reinstated.
The 530 nm excitation of C60−PDI (7) results also in the
exclusive formation of the PDI singlet excited state. In
particular, transient maxima at 710, 760, 800, 845, 910, and
990 nm as well as transient minima at 490 and 515 nm that are
shown in Figure S24 (SI) are formed instantaneously. While
the presence of C60 fails to influence the formation of the PDI
singlet excited state, it exerts, however, a notable impact on its
lifetime. Via a rapid decay with underlying lifetimes of 4.7 0.1
ps (toluene), 4.5 0.1 ps (chlorobenzene), and 4.4 0.1 ps
(benzonitrile) rather broad features in the 800−1100 nm range
develop in the near-infrared region of the spectrum. In the
visible region, which is initially dominated by the ground state
bleaching of PDI, an additional band is discernible at 710 nm.
Overall, a good spectral resemblance with the features seen
during the photolysis study with C60-ref (11) confirms that an
intramolecular exergonic transduction of singlet excited state
energy is operative to form in C60−PDI (7) the C60 singlet
excited state. As in the reference experiments with 11, the
deactivation of the C60 singlet excited state in 7 is dominated by
intersystem crossing (1.5 0.1 ns) to the energetically lower
lying triplet excited state. In this regard, it is important to note
that the decay kinetics of the 880 nm transition coincide with
the growth kinetics of the 710 nm transition. In other words,
the singlet excited state decay matches the triplet excited state
growth with kinetics that are hardly faster (1.4 0.1 ns) than
the inherent intersystem crossing dynamics in 11. Interestingly,
we did not find the characteristic PDI triplet excited state
featurea strong triplet−triplet transition at 560 nmon the
time scale of our femtosecond setup, that is, 8 ns. From this we
infer that the C60 triplet excited state (1.5 eV), once formed,
does not undergo a thermodynamically allowed transfer of
triplet excited state energy to PDI (1.2 eV) on the femto- to
picosecond time scale. However, in complementary nano-
second experimentsfollowing 532 nm excitationwe
CONCLUSIONS
■
In conclusion, a linear Lu3N@Ih-C80−PDI electron donor−
acceptor conjugate has been designed, synthesized, and
characterized. In the ground state, no appreciable electronic
interactions attest the efficient decoupling of the electron-
donating Lu3N@Ih-C80 from the electron-accepting PDI. In the
excited state, the latter interact, however, and an intramolecular
electron transfer commences with the photoexcitation of PDI.
1
The initially formed singlet radical ion pair state, [(Lu3N@Ih-
C80)•+−PDI•−], undergoes radical ion pair intersystem crossing
induced by electron nuclear hyperfine coupling within the
3
radicals to produce the triplet radical ion pair state, [(Lu3N@
Ih-C80)•+−PDI•−]. Key to this interconversion is certainly the
presence of the Lu3N cluster. The accordingly formed radical
ion pairs recombine much faster from the singlet manifold than
from the triplet manifold with lifetimes as long as 102 ns
(THF) and as short as 28 ps (DMF), respectively. Overall, the
lifetime of the radical ion pair state increases about 1000-fold.
EXPERIMENTAL SECTION
■
Spectroscopy. All NMR spectra were recorded respectively on a
Bruker AC 300 spectrometer or Bruker AV 500 spectrometer with a
CryoProbe system, locked on deuterated solvents and referenced to
the solvent peak. The 1D (1H, 13C, and DEPT135) and 2D
experiments (COSY, HMQC, HMBC) were performed by means of
standard experimental procedures of the Bruker library. Absorption
spectra of all samples were recorded in toluene with a Shimadzu UV-
3150 spectrometer using a quartz cell and 1-nm resolution. Matrix-
assisted laser desorption-ionization time-of flight (MALDI-TOF) mass
spectra were recorded with a Bruker BIFLEX-III mass spectrometer
using 1,1,4,4-tetraphenyl-1,3-butadiene as the matrix. The measure-
ments were performed in both positive and negative ion modes.
Steady-State Emission. The spectra were recorded on a
FluoroMax 3 fluorometer (vis detection) and on a Fluorolog
spectrometer (NIR detection). Both spectrometers were built by
HORIBA JobinYvon. The measurements were carried out at room
temperature.
Time-Resolved Absorption. Femtosecond transient absorption
studies were performed with 387 and 530 nm laser pulses (1 kHz, 150
fs pulse width) from an amplified Ti:Sapphire laser system (Clark-
MXR, Inc.), the laser energy was 200 nJ. Nanosecond laser flash
photolysis experiments were performed with 355 and 532 nm laser
pulses from a Quanta-Ray CDR Nd:YAG system (6 ns pulse width) in
a front face excitation geometry.
noticed a C60 triplet excited state decay of 730
50 ns,
Time-Resolved Emission. Fluorescence lifetimes were measured
which is appreciably faster than what is seen for 11 (20 μs). A
closer look at the transient spectra, which develop during the
later stages of the decay, reveals the signatures of the PDI
triplet excited state. The latter decays with kinetics of 1.2 0.1
μs. With the decay being independent of the nature, that is,
either a C60 or a PDI triplet excited state, we note for 7 singlet
oxygen quantum yields of 0.47, 0.76, 0.15, and 0.61 in toluene,
chlorobenzene, THF, and benzonitrile, respectively.
In short, once photoexcited, C60−PDI (7) deactivates via the
following cascade of events. The PDI singlet excited state (2.32
eV) decays through an exergonic intramolecular energy transfer
to afford the C60 singlet excited state (1.8 eV). This process is
then followed by a C60-centered intersystem crossing, which
affords the corresponding triplet excited state (1.5 eV). Instead
of a direct recovery of the ground state, the triplet excited state
energy is funneled back to the PDI (1.2 eV) via a
thermodynamically allowed intramolecular triplet excited state
energy transfer and, as such, decays indirectly to the ground
state.
by using a Fluorolog (HORIBA Jobin Yvon).
Electrochemistry. Differential pulse voltammetry (DPV) and
cyclic voltammetry (CV) were carried out in o-DCB using a BAS CW-
50 instrument. A conventional three-electrode cell consisting of a
platinum working electrode, a platinum counterelectrode, and a
saturated calomel reference electrode (SCE) was used for both
measurements. Supporting electrolyte was 0.05 M (n-Bu)4NPF6. All
potentials were recorded against an SCE reference electrode and
corrected against Fc/Fc+. DPV and CV were measured at a scan rate of
20 and 100 mV s−1, respectively.
Materials. All chemicals were of reagent grade and purchased from
Wako. Lu3N@Ih-C80 (>99%) was purchased from Luna Co.
Preparative and analysis HPLC were performed on semi-preparative
Buckyprep column (ø 10 mm × 100 mm, Cosmosil), and analysis
Buckyprep column (ø 4.6 mm × 100 mm, Cosmosil), respectively.
Toluene was used as eluent.
Synthesis of Lu3N@Ih-C80−PDI (6). Tosylhydrazone 5 (4.5 mg,
4.6 μmol) and NaOMe (0.8 mg, 15 μmol) were dissolved in pyridine
(0.9 mL) and stirred for 40 min at 70 °C under Ar. Then, Lu3N@C80
(1.5 mg, 1 μmol) in 3 mL o-DCB was added in. The mixture was
stirred at 80 °C for 5 h under Ar. The reaction mixture was separated
by HPLC (Buckyprep column, toluene/acetonitrile (15/1)); the
11172
dx.doi.org/10.1021/ja403763e | J. Am. Chem. Soc. 2013, 135, 11165−11174

Journal of the American Chemical Society
Article
second fraction is the dyad 6 (Lu3N@C80−PDI). 6 was further
purified by Buckyprep column. Yield: ∼0.7 mg, 42% based on
Notes
The authors declare no competing financial interest.
1
consumed Lu3N@C80; H NMR (500 MHz, CDCl3): δ = 8.74 (s,
2H), 8.71 (s, 2H), 8.34 (d, 2J = 7.9 Hz, 2H), 7.52 (d, 2J = 8.1 Hz, 2H),
3
ACKNOWLEDGMENTS
4.22 (t, J = 7.5 Hz, 2H), 2.51 (s, 3H), 1.75 (m, 2H), 1.44 (m, 2H),
■
1.36 (m, 2H), 1.30−1.22 (m, br, 14H), 0.88 ppm (t, 3J = 7.1 Hz, 3H);
13C NMR (125 MHz, CDCl3): δ = 162.87 (CO, imide), 162.65
(CO, imide), 152.03, 151.81, 150.61, 150.50, 149.28, 148.72,
148.62, 148.55, 148.24, 148.06, 148.03, 147.55, 146.99, 146.27, 146.25,
145.94, 145.49, 145.42, 145.37, 145.35, 145.30, 145.11, 145.06, 144.83,
144.79, 144.70, 144.64, 144.04, 143.85, 143.71, 143.66, 143.43, 143.32,
143.27, 143.12, 142.97, 142.93, 142.70, 142.57, 142.34, 142.30, 142.04,
142.00, 141.47, 141.15, 140.94, 140.82, 140.79, 140.80, 140.33, 140.12,
140.08, 139.72, 139.69, 139.57, 138.45, 138.39, 136.10, 136.07, 135.89,
135.87, 135.70, 135.53, 135.39, 135.22, 134.96, 134.94, 134.86, 134.53,
133.77 (CH of PDI), 133.43 (CH of PDI), 132.03, 131.93, 129.59,
129.49 (ph), 129.30 (ph), 129.02, 128.87, 128.21, 127.27, 127.03,
126.56, 126.30, 125.90, 125.58, 124.08, 123.86, 123.71, 123.57, 98.59,
98.52, 95.82, 95.75, 48.13 (spiro C), 41.47, 32.37, 30.40 (CH3), 30.10,
30.08, 30.06, 29.99, 29.81 (overlapped), 28.56, 27.53, 23.15, 14.60
(CH3) ppm; MALDI-TOF MS (positive mode, TPB as matrix): m/z:
calcd for Lu3C124H36O4N3Cl4: 2297.97 (100% intensity); found: 2298
[M]+.
This work was supported in part by a Grant-in-Aid for Scientific
research on Innovative Areas (No. 20108001, “pi-Space”); a
Grant-in-Aid for Scientific Research (A) (No. 20245006) and
(B) (No. 24350019); The Next Generation Super Computing
Project (Nanoscience Project) and Specially Promoted
Research (No. 22000009); Nanotechnology Support Project
and Grants-in-Aid for Scientific research on Priority Area (Nos.
20036008, 20038007); and Specially Promoted Research from
the Ministry of Education, Culture, Sports, Science, and
Technology of Japan; and The Strategic Japanese−Spanish
Cooperative Program funded by JST and MICINN (fullsol@r
PLE2009-0039; Endosol@r PIB2010JP00196). F.L. Thanks for
NNSF of China (No. 21241004), NSF of Jiangsu province of
China (No. BK2012611), and a Project Funded by the Priority
Academic Program Development of Jiangsu Higher Education
Institutions (PAPD). Financial support by the Deutsche
Forschungsgemeinschaft through grant no. GU 517/14-1 is
gratefully acknowledged.
Synthesis of C60−PDI (7). Tosylhydrazone 5 (5.0 mg, 5.1 μmol)
and NaOMe (0.8 mg, 15 μmol) were dissolved in pyridine (0.35 mL)
and stirred for 30 min at 65 °C under Ar. Then C60 (3.7 mg, 5.1 μmol)
in 3 mL of o-DCB was added in. The mixture was stirred at 75 °C for
23 h under Ar. The reaction mixture was separated by HPLC
(Buckyprep column, toluene); the second fraction contains the
mixture of (5,6)- and (6,6)-monoadducts. By recycling the mixture on
Buckyprep column, the (5,6)- and (6,6)-isomers (8 and 7) can be
isolated from each other. (5,6)-Isomer (8) can be converted to (6,6)-
isomer (7) by a thermal treatment (150−160 °C in o-DCB for 5−6 h).
REFERENCES
■
(1) (a) Imahori, H.; Sakata, Y. Adv. Mater. 1997, 9, 537. (b) Imahori,
H.; Sakata, Y. Eur. J. Org. Chem. 1999, 2445. (c) Guldi, D. M. Chem.
Commun. 2000, 321. (d) Guldi, D. M.; Prato, M. Acc. Chem. Res. 2000,
33, 695.
(2) (a) Gust, D.; Moore, T. A.; Moore, A. L. Acc. Chem. Res. 2001,
34, 40. (b) Bottari, G.; de la Torre, G.; Guldi, D. M.; Torres, T. Chem.
Rev. 2010, 110, 6768.
1
Total Yield: 1.8 mg, 73% based on consumed C60; dyad 7: H NMR
2
(500 MHz, CDCl3): δ = 8.75 (s, 2H), 8.72 (s, 2H), 8.21 (d, J = 8.0
Hz, 2 H), 7.54 (d, 2J = 8.0 Hz, 2 H), 4.22 (t, 3J = 7.5 Hz, 2 H), 2.62 (s,
(3) (a) Echegoyen, L.; Echegoyen, L. E. Acc. Chem. Res. 1998, 31,
593. (b) Fukuzumi, S. Res. Chem. Intermed. 1997, 23, 519.
(4) (a) Kadish, K. M., Ruoff, R. M., Eds. Fullerenes: Chemistry, Physics
and Technology; Wiley-Interscience: New York, 2000. (b) Guldi, D. M.
Chem. Soc. Rev. 2002, 31, 22. (c) Figueira-Duarte, T. M.; Gegout, A.;
Nierengarten, J. F. Chem. Commun. 2007, 43, 109. (d) Delgado, J. L.;
Bouit, P.-A.; Filippone, S.; Herranz, M. A.; Martín, N. Chem. Commun.
2010, 46, 4853.
(5) (a) Pinzon, J. R.; Plonska-Brzezinska, M. E.; Cardona, C. M.;
́
Athans, A. J.; Gayathri, S. S.; Guldi, D. M.; Herranz, M. A.; Martín, N.;
Torres, T.; Echegoyen, L. Angew. Chem., Int. Ed. 2008, 47, 4173.
(b) Pinzon, J. R.; Gasca, D. C.; Sankaranarayanan, S. G.; Bottari, G.;
́
3
3 H), 1.75 (m, 2H), 1.43−1.26 (m, 18 H), 0.88 (t, J = 7.0 Hz, 3H)
ppm; 13C NMR (125 MHz, CDCl3): δ = 162.45 (CO, imide),
162.21 (CO, imide), 148.41, 147.84, 145.29, 145.16, 145.10, 145.08,
145.06, 145.01, 144.78, 144.66, 144.45, 144.36, 143.96, 143.73, 143.70,
143.10, 143.08, 142.98, 142.97, 142.94, 142.37, 142.24, 142.16, 142.12,
140.96, 140.82, 140.33, 138.06, 137.72, 135.61, 135.43, 134.01, 133.31
(CH of PDI), 132.98 (CH of PDI), 131.94 (ph), 131.59, 131.49,
129.11, 128.88 (ph), 128.44, 123.64, 123.40, 123.27, 123.18, 80.59,
80.56, 46.97 (spiro C), 41.02, 31.93, 29.66, 29.64, 29.61, 29.55, 29.37,
29.36, 28.11, 27.08, 22.71, 22.67 (CH3), 14.16 (CH3) ppm; MALDI-
TOF MS (negative mode, TPB as matrix): m/z: calcd for
C104H36O4N2Cl4: 1518.14 (100% intensity); found: 1518 [M]−.
Synthesis of Tosylhydrazone 5. See the SI.
Torres, T.; Guldi, D. M.; Echegoyen, L. J. Am. Chem. Soc. 2009, 131,
7727. (c) Ross, R. B.; Cardona, C. M.; Guldi, D. M.;
Sankaranarayanan, S. G.; Reese, M. O.; Kopidakis, N.; Peet, J.;
Walker, B.; Bazan, G. C.; Van Keuren, E.; Holloway, B. C.; Drees, M.
Theoretical Calculations. The calculations were carried out using
the hybrid density functional theory (DFT) at the M06-2X level11 as
implemented in the Gaussian 09 software package.14 The SDD basis
set13 with the relativistic effective core potential was employed for Lu,
3-21G basis set12 for C, H, O, N, and 3-21G* for Cl.
́
Nat. Mater. 2009, 8, 208. (d) Wolfrum, S.; Pinzon, J. R.; Molina-
Ontoria, A.; Gouloumis, A.; Martín, N.; Echegoyen, L.; Guldi, D. M.
Chem. Commun. 2011, 47, 2270. (e) Grimm, B.; Schornbaum, J.;
Cardona, C. M.; van Paauwe, J. D.; Boyd, P. D. W.; Guldi, D. M.
Chem. Sci. 2011, 2, 1530. (f) Feng, L.; Rudolf, M.; Wolfrum, S.;
Troeger, A.; Slanina, Z.; Akasaka, T.; Nagase, S.; Martín, N.; Ameri, T.;
Brabec, C. J.; Guldi, D. M. J. Am. Chem. Soc. 2012, 134, 12190.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental details including the reaction schemes, synthesis
procedures, HPLC profiles, mass spectra, CV, DPV curves, 1D
and 2D NMR spectra, NMR chemical shifts, DFT-optimized
structures of 6 and 7, differential absorption spectra. This
material is available free of charge via the Internet at http://
pubs.acs.org.
́
(6) (a) Takano, Y.; Herranz, A.; Martín, N.; Radhakrishnan, S. G.;
Guldi, D. M.; Tsuchiya, T.; Nagase, S.; Akasaka, T. J. Am. Chem. Soc.
2010, 132, 8048. (b) Guldi, D. M.; Feng, L.; Radhakrishnan, S. G.;
Nikawa, H.; Yamada, M.; Mizorogi, N.; Tsuchiya, T.; Akasaka, T.;
́
Nagase, S.; Herranz, A.; Martín, N. J. Am. Chem. Soc. 2010, 132, 9078.
(c) Feng, L.; Radhakrishnan, S. G.; Mizorogi, N.; Slanina, Z.; Nikawa,
H.; Tsuchiya, T.; Akasaka, T.; Nagase, S.; Martín, N.; Guldi, D. M. J.
Am. Chem. Soc. 2011, 133, 7608. (d) Takano, Y.; Obuchi, S.; Mizorogi,
AUTHOR INFORMATION
Corresponding Author
fenglai@suda.edu.cn (L.F.); akasaka@tara.tsukuba.ac.jp (T.A.);
dirk.guldi@chemie.uni.erlangen.de (D.M.G.).
■
́
N.; García, R.; Herranz, M. A.; Rudolf, M.; Guldi, D. M.; Martín, N.;
Nagase, S.; Akasaka, T. J. Am. Chem. Soc. 2012, 134, 19401.
(7) (a) Feng, L.; Slanina, Z.; Sato, S.; Yoza, K.; Tsuchiya, T.;
Mizorogi, N.; Akasaka, T.; Nagase, S.; Martin, N.; Guldi, D. M. Angew.
11173
dx.doi.org/10.1021/ja403763e | J. Am. Chem. Soc. 2013, 135, 11165−11174

Journal of the American Chemical Society
Article
Chem. 2011, 123, 6031; Angew. Chem., Int. Ed. 2011, 50, 5909.
́
(b) Takano, Y.; Obuchi, S.; Mizorogi, N.; García, R.; Herranz, M. A.;
Rudolf, M.; Wolfrum, S.; Guldi, D. M.; Martín, N.; Nagase, S.;
Akasaka, T. J. Am. Chem. Soc. 2012, 134, 16103.
(8) (a) Ohkubo, K.; Kawashima, Y.; Fukuzumi, S. Chem. Commun.
2012, 48, 4314. (b) Ohkubo, K.; Kawashima, Y.; Fukuzumi, S. J. Phys.
Chem. A 2012, 116, 8942.
(9) Chaur, M. N.; Melin, F.; Ortiz, A.; Echegoyen, L. Angew. Chem.,
Int. Ed. 2009, 48, 7514.
(10) (a) Weiss, E. A.; Chernick, E. T.; Wasielewski, M. R. J. Am.
Chem. Soc. 2004, 126, 2326. (b) Colvin, M. T.; Ricks, A. B.; Scott, A.
M.; Smeigh, A. L.; Carmieli, R.; Miura, T.; Wasielewski, M. R. J. Am.
Chem. Soc. 2011, 133, 1240. (c) Roth, H. Photochem. Photobiol. Sci.
2008, 8, 540.
(11) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215.
(12) (a) Binkley, J. S.; Pople, J. A.; Hehre, W. J. J. Am. Chem. Soc.
1980, 102, 939. (b) Gordon, M. S.; Binkley, J. S.; Pople, J. A.; Pietro,
W. J.; Hehre, W. J. J. Am. Chem. Soc. 1982, 104, 2797.
(13) Cao, X. Y.; Dolg, M. J. Mol. Struct. (THEOCHEM) 2002, 581,
139.
(14) Frisch, M. J. et al. Gaussian 09, Revision A. 02; Gaussian Inc.:
Wallingford, CT, 2009.
(15) (a) Shibano, Y.; Umeyama, T.; Matano, Y.; Tkachenko, N. V.;
Lemmetyinen, H.; Imahori, H. Org. Lett. 2006, 8, 4425. (b) Schuster,
D.; Li, K.; Guldi, D. M.; Palkar, A.; Echegoyen, L.; Stanisky, C.; Cross,
R. J.; Niemi, M.; Tkachenko, N. V.; Lemmetyinen, H. J. Am. Chem. Soc.
2007, 129, 15973.
(16) In benzonitrile, from a subsequent growth of the (Lu3N@Ih-
C80)•+−PDI•− radical ion pair state (800 300 ps), we gather that the
PDI triplet excited state is in thermodynamic equilibrium with the
latter.
(17) These quantum yields were determined by taking the values for
the ZnP triplet excited state (Φ = 0.73, ε = 5300 M−1 cm−1 at 745
nm), the H2P triplet (Φ = 0.7, ε = 6000 M−1 cm−1 at 780 nm), and the
C60 triplet excited state (Φ = 1, ε = 19500 M−1 cm−1 at 750 nm) in
toluene from the literature. On the contrary, the values for the one-
electron reduced PDI π-radical anion (ε = 80000 10000 M−1 cm−1
at 760 nm) were derived from our own spectroelectrochemical
investigations.
(18) In THF the mechanism is considered to be between the
deactivation pathways of the less polar solvents and benzonitrile.
11174
dx.doi.org/10.1021/ja403763e | J. Am. Chem. Soc. 2013, 135, 11165−11174