Electron transfer within self-assembling cyclic tetramers using chlorophyll-based donor-acceptor building blocks

Article abstract of DOI:10.1021/ja211329k

The synthesis and photoinduced charge transfer properties of a series of Chl-based donor-acceptor triad building blocks that self-assemble into cyclic tetramers are reported. Chlorophyll a was converted into zinc methyl 3-ethylpyrochlorophyllide a (Chl) and then further modified at its 20-position to covalently attach a pyromellitimide (PI) acceptor bearing a pyridine ligand and one or two naphthalene-1,8:4,5-bis(dicarboximide) (NDI) secondary electron acceptors to give Chl-PI-NDI and Chl-PI-NDI₂. The pyridine ligand within each ambident triad enables intermolecular Chl metal-ligand coordination in dry toluene, which results in the formation of cyclic tetramers in solution, as determined using small- and wide-angle X-ray scattering at a synchrotron source. Femtosecond and nanosecond transient absorption spectroscopy of the monomers in toluene-1% pyridine and the cyclic tetramers in toluene shows that the selective photoexcitation of Chl results in intramolecular electron transfer from ^1*Chl to PI to form Chl^+?-PI ^-?-NDI and Chl^+?-PI^-?-NDI ₂. This initial charge separation is followed by a rapid charge shift from PI^-? to NDI and subsequent charge recombination of Chl^+?-PI-NDI^-? and Chl^+?-PI- (NDI)NDI^-? on a 5-30 ns time scale. Charge recombination in the Chl-PI-NDI₂ cyclic tetramer (τ_CR = 30 ± 1 ns in toluene) is slower by a factor of 3 relative to the monomeric building blocks (τ_CR = 10 ± 1 ns in toluene-1% pyridine). This indicates that the self-assembly of these building blocks into the cyclic tetramers alters their structures in a way that lengthens their charge separation lifetimes, which is an advantageous strategy for artificial photosynthetic systems.

Full text of DOI:10.1021/ja211329k

Article
pubs.acs.org/JACS
Electron Transfer within Self-Assembling Cyclic Tetramers Using
Chlorophyll-Based Donor−Acceptor Building Blocks
Victoria L. Gunderson, Amanda L. Smeigh, Chul Hoon Kim, Dick T. Co, and Michael R. Wasielewski*
Department of Chemistry and Argonne-Northwestern Solar Energy Research (ANSER) Center, Northwestern University, Evanston
Illinois 60208-3113, United States
S
* Supporting Information
ABSTRACT: The synthesis and photoinduced charge transfer
properties of a series of Chl-based donor−acceptor triad building
blocks that self-assemble into cyclic tetramers are reported.
Chlorophyll a was converted into zinc methyl 3-ethylpyrochlor-
ophyllide a (Chl) and then further modified at its 20-position to
covalently attach a pyromellitimide (PI) acceptor bearing a
pyridine ligand and one or two naphthalene-1,8:4,5-bis-
(dicarboximide) (NDI) secondary electron acceptors to give
Chl−PI−NDI and Chl−PI−NDI2. The pyridine ligand within
each ambident triad enables intermolecular Chl metal−ligand coordination in dry toluene, which results in the formation of cyclic
tetramers in solution, as determined using small- and wide-angle X-ray scattering at a synchrotron source. Femtosecond and
nanosecond transient absorption spectroscopy of the monomers in toluene−1% pyridine and the cyclic tetramers in toluene
shows that the selective photoexcitation of Chl results in intramolecular electron transfer from 1*Chl to PI to form Chl+•−PI−•−
NDI and Chl+•−PI−•−NDI2. This initial charge separation is followed by a rapid charge shift from PI−• to NDI and subsequent
charge recombination of Chl+•−PI−NDI−• and Chl+•−PI−(NDI)NDI−• on a 5−30 ns time scale. Charge recombination in the
Chl−PI−NDI2 cyclic tetramer (τCR = 30 1 ns in toluene) is slower by a factor of 3 relative to the monomeric building blocks
(τCR = 10 1 ns in toluene−1% pyridine). This indicates that the self-assembly of these building blocks into the cyclic tetramers
alters their structures in a way that lengthens their charge separation lifetimes, which is an advantageous strategy for artificial
photosynthetic systems.
Semisynthetic chlorophyll-based systems have been pre-
viously shown to undergo efficient photoinduced electron
transfer.3 Recently, we have functionalized zinc methyl 3-
ethylpyrochlorophyllide a (Chl) at the 20-position to generate
a series of donor−acceptor dyads having either a naphthalene-
1,8:4,5-bis(dicarboximide) (NDI) or a perylene-3,4:9,10-bis-
(dicarboximide) (PDI) as the electron acceptor.4 Upon Chl
photoexcitation, ultrafast electron transfer from Chl to NDI or
INTRODUCTION
■
Photophysical studies of biomimetic systems have provided key
insights into the mechanisms of energy and electron transfer in
natural photosynthesis that are necessary for the development
of well-ordered supramolecular systems for artificial photosyn-
thesis. Specifically, understanding how to design molecular
building blocks that self-organize into architectures to improve
and/or promote light-driven functionality is important for
converting solar energy into electricity or fuels.1 A wide variety
of chromophores, including those found in natural photo-
synthetic proteins, have been utilized as self-assembling
molecular building blocks that both collect light energy and
carry out charge separation.1a,2 The tunable electronic and
redox properties of natural chlorophyll make it a valuable
model for probing electron transfer reactions. Chlorophylls
absorb light across a broad range of wavelengths and act as both
energy and electron donors and acceptors. Various substituents
on the chlorin core give rise to a diverse set of naturally
occurring pigments that can be extracted from photosynthetic
organisms, isolated, then synthetically modified. For example,
most naturally occurring chlorophyll pigments are oxidatively
unstable largely because of the β-keto ester in their isocyclic E
ring. As a consequence, decarboxylation of the β-keto ester to
give the ketone greatly increases their stability without
significantly changing their electronic properties.
PDI occurs in both polar (τCS =13−19 ps) and nonpolar (τCS
=
12−25 ps) solvents. In contrast, charge recombination in polar
solvents occurs on a picosecond time scale (τCR = 25−43 ps),
while it slows to a nanosecond time scale in nonpolar solvents
(τCR = 1.2−3.3 ns). Although these systems demonstrate nearly
unity charge separation quantum yields, their rapid charge
recombination in both polar and nonpolar solvents does not
allow for applications in artificial photosynthesis, where charge
migration and/or chemistry on a slower diffusional time scale
are required.
To slow charge recombination, extending the stepwise
charge transport array is an advantageous strategy. Following
the initial photoinitiated charge separation, natural reaction
center proteins use multiple electron transfer steps to increase
Received: December 3, 2011
Published: February 13, 2012
© 2012 American Chemical Society
4363
dx.doi.org/10.1021/ja211329k | J. Am. Chem. Soc. 2012, 134, 4363−4372

Journal of the American Chemical Society
Article
Scheme 1. Donor−Acceptor Triad Structures
the radical ion pair distances, thereby diminishing the electronic
coupling between them and slowing charge recombination.5
Extension of donor−acceptor dyads to donor−acceptor−
acceptor triads has been used to mimic this natural
phenomenon.1d,6 However, this strategy does not always yield
optimal systems. For example, a series of molecular dyads and
triads have been studied in which zinc methyl 131-
desoxopyrochlorophyllide a served as the electron donor and
was connected through the 31-position to pyromellitimide (PI),
NDI, or a serial pair of PI−NDI acceptors.7 Electron transfer
was observed in both dyads and the triad. However, it was
shown that the PI one-electron reduction potential is sensitive
to the imide substituents. This led to charge separation and
recombination lifetimes in the triad that were more rapid
relative to the Chl−PI dyad, 2 versus 25 ps and 140 versus 360
ps, respectively, in toluene. Thus, new ways to connect these
components that do not perturb the inherent donor and
acceptor redox potentials as well as the chromophore excited
state energies are necessary to realize long charge separation
lifetimes.
The synthesis of such systems often demands considerable
effort, so that numerous studies have been directed toward self-
assembling large supramolecular structures.1b,8 Chl’s naturally
form supramolecular assemblies using metal−ligand coordina-
tion as well as hydrogen bonding, π−π interactions, and
hydrophobic effects within the protein environment.9 For
example, Chl’s with modifications at the 3-, 8-, and 17-carbon
positions have replicated bacteriochlorophyll self-assembly
associated with the peripheral antenna complexes of green
sulfur bacteria.10 Additionally, modifications to Chl at the 20-
carbon position have also proven useful for producing building
blocks leading to discrete self-assembled structures.11 Specifi-
cally, the incorporation of pyridine ligands is a facile strategy for
the generation of discrete supramolecular Chl and porphyrin
systems, where substitution of a 4-pyridinyl group at the 20-Chl
position (or the 5-position of a porphyrin) has led to the
assembly of cyclic tetramers (molecular boxes) composed of
four self-coordinated Chl (or porphyrin) units, when the
monomers are dissolved in nonligating solvents, such as dry
toluene.11a,12 The formation of these supramolecular structures
enables new energy transfer pathways to form. However,
studies of self-assembled, semisynthetic, Chl-based, electron
transfer systems have been limited.13
To better determine how self-assembly affects photoinduced
charge separation in Chl-based systems, we have synthesized a
series of Chl-based donor−acceptor triads (Scheme 1) that self-
assemble into cyclic tetramers. Covalent attachment of NDI
derivatives to the PI central benzene ring limits perturbation of
the PI one-electron reduction potential. This occurs primarily
because the large dihedral angle between the attached aromatic
group and the PI benzene ring limits transmission of the π
substituent effect, and allows for efficient and long-lived charge
separation. The incorporation of a pyridine ligand on the
donor−acceptor molecular building block further facilitates self-
assembly through metal−ligand coordination in toluene. Our
results demonstrate that charge separation is stabilized in the
self-assembled cyclic arrays relative to that in the corresponding
monomers.
EXPERIMENTAL SECTION
■
Synthesis. The synthesis of the Chl−PI−NDI series is described
briefly below and outlined explicitly in the Supporting Information.
Chl a was extracted from the cyanobacterium Arthrospira platensis
strain Pacifica obtained from Cyanotech and converted into methyl
pyropheophorbide a.14 Selective hydrogenation of the double bond at
the 31-position yielded methyl 3-ethyl-pyropheophorbide a, which was
then brominated at the 20-position.15 Methyl 20-bromo-3-ethyl-
pyropheophorbide a underwent a series of reactions to yield zinc 2′-
ethylhexyl 3-ethyl-20-(p-aminophenyl)-pyropheophorbide a.4 This
4364
dx.doi.org/10.1021/ja211329k | J. Am. Chem. Soc. 2012, 134, 4363−4372

Journal of the American Chemical Society
Article
Figure 1. Normalized steady-state absorption spectra in (a) toluene−1% pyridine and (b) toluene.
was further reacted in a series of steps, which included imide
condensation reactions with a variety of pyromellitic dianhydride and
naphthalene-1,8:4,5-tetracarboxydianhydride derivatives, to yield Chl−
PI−NDI, and Chl−PI−NDI2. The synthetic intermediates and final
products were characterized by 1H NMR, MALDI-TOF, MS-ESI, and
UV−vis spectroscopy.
Electrochemistry. Electrochemical studies were carried out using
a CH Instruments Model 622 electrochemical workstation. The
measurements were performed in dichloromethane (DCM) contain-
ing 0.1 M tetra-n-butylammonium hexafluorophosphate (TBAPF6),
and the ferrocene/ferrocenium redox couple (Fc/Fc+, 0.475 V vs SCE
in DCM) was used as an internal reference. A 1.0 mm diameter
platinum disk electrode, platinum wire counter electrode, and Ag/
AgOx reference electrode were employed. Sample concentrations were
1 mM. All electrochemical measurements were performed under a dry
nitrogen atmosphere. TBAPF6 was purchased from Aldrich and
recrystallized twice from ethyl acetate prior to use.
X-ray Scattering. X-ray scattering measurements were performed
using the undulator beamline 12-ID at the Advanced Photon Source
(APS) at Argonne National Laboratory. The X-ray scattering
instrument utilized a double crystal Si(111) monochromator and a
two-dimensional mosaic CCD detector.16 The X-ray wavelength was
set at λ = 0.62 Å, and the distance between the sample and detector
was adjusted to achieve scattering measured across the 0.007 Å−1 < q <
0.027 Å−1 region, where q = (4π/λ) sin θ, λ is the X-ray scattering
wavelength, and 2θ is the scattering angle. A quartz flow cell was used
as the sample container. All samples were filtered through 200 nm
PTFE filters (Whatman) prior to measurements. The scattering
intensity was averaged over 800 measurements, and the solvent
scattering was subtracted from the sample spectrum.
Samples for nanosecond TA spectroscopy were placed in a 1-cm
path length cuvette and excited with 7 ns laser pulses generated using
the frequency-tripled output of a Continuum 8000 Nd:YAG laser to
pump a Continuum Panther OPO. The laser pulse energy was 1.2 mJ
at 416 nm. The excitation pulse was focused to a 5 mm diameter spot
and matched to the diameter of the probe pulse generated using a
xenon flashlamp (EG&G Electro-Optics FX-200). The signal was
detected using a monochromator (HORIBA Triax 180) coupled to a
photomultiplier tube (Hamamatsu R926) with high voltage applied
only to four dynodes. The total instrument response time was 7 ns and
was determined primarily by the laser pulse duration.
Picosecond time-resolved fluorescence (TRF) measurements were
made using a streak camera system (Hamamatsu C4780). The light
source was a lab-built cavity-dumped Kerr lens mode-locked
Ti:sapphire oscillator pumped by a commercial Nd:YVO4 laser
(Spectra-Physics, Millennia V). The center wavelength was 830 nm,
and the energy of the 25 fs output pulses was about 30 nJ at a 820 kHz
repetition rate. Pump pulses at 415 nm were generated by second
harmonic generation in a 200-μm thick LBO (lithium triborate)
crystal. A parabolic mirror was used to focus the excitation beam into
the sample and the subsequent fluorescence was collected in a back
scattering geometry using the same parabolic mirror. Magic angle
detection was used to avoid polarization effects. The excitation beam
was focused into a 1-cm path length cuvette using a singlet lens. The
fluorescence from the sample was collected and focused into a
monochromator by a lens pair. The IRF was 20 and 350 ps (fwhm) in
1 and 20 ns time windows, respectively. All data were acquired in
single photon counting mode using the Hamamatsu HPD-TA
software.
Optical Spectroscopy. Steady-state absorption spectroscopy was
performed using a Shimadzu 1800 UV/vis spectrophotometer. A
single-photon-counting fluorimeter (Photon Technology Interna-
tional) was used for emission experiments. Fluorescence measure-
ments were performed at room temperature in a 1 cm quartz cuvette
with excitation/emission geometries at right angles. All solvents were
spectroscopic grade and used as is, except for tetrahydrofuran (THF)
and toluene, which were further purified by passing them twice
through alumina (GlassContour) immediately prior to use.
Femtosecond transient absorption (TA) measurements were made
using a Ti:sapphire laser system detailed previously.17 The instrument
response function (IRF) for the pump−probe experiments was 150 fs.
Typically 5 s of averaging was used to obtain the transient spectrum at
a given delay time. Samples were photoexcited with 416 nm, 110 fs, 1.0
μJ laser pulses focused to a 200 μm spot in a 2 mm path length quartz
cuvette. The optical density at the pump wavelength was kept between
0.5 and 0.7. Analysis of the kinetic data was performed at multiple
wavelengths using a Levenberg−Marquardt nonlinear least-squares fit
to a general sum-of-exponentials function convoluted with a Gaussian
instrument response function. In addition, the three-dimensional data
set of ΔA versus wavelength and time from 440 to 800 nm and 0−6 ns
was analyzed using single value decomposition and global fitting to
obtain the decay-associated spectra using Surface Xplorer-Pro.18
RESULTS AND DISCUSSION
■
Steady-State Spectroscopy. The ground state absorption
spectra of the Chl-based series in toluene−1% pyridine are
shown in Figure 1a. The pyridine coordinates to the zinc metal
center of the Chl to disrupt self-assembly and yield the
spectrum of the monomer. The spectra were normalized at
their respective 431 nm B-band (Soret band) maxima. The
absorption features observed to the red of the Soret band are
also characteristic of the Chl chromophore, where the Qy(0,0)
band maxima for all molecules occur at 658 nm and exhibit
only slight extinction coefficient variations between molecules
in the series. The sharp absorption features observed at 363 and
383 nm in both Chl−PI−NDI and Chl−PI−NDI2 result from
the NDI acceptor. The extinction coefficient at these
wavelengths increases by a factor of 2 upon increasing the
number of NDI molecules from Chl−PI−NDI to Chl−PI−
NDI2. For both triads, substitution of the 20-phenyl-Chl
macrocycle with PI and NDI results in no discernible 20-
phenyl-Chl spectral distortions, which indicates that the
electronic coupling between these molecules is weak and that
4365
dx.doi.org/10.1021/ja211329k | J. Am. Chem. Soc. 2012, 134, 4363−4372

Journal of the American Chemical Society
Article
Figure 2. Left: Atomic pair distribution functions (PDFs) obtained from the X-ray scattering data (black) compared with those of the model
monomer (red) and model cyclic tetramer (blue) for (a) Chl−PI−NDI and (b) Chl−PI−NDI2. Right: Structures of cyclic tetramers. The individual
donor and acceptors are indicated: Chl (green), PI and spacers (blue), and NDI (red).
these systems are not conjugated supermolecules, but rather as
donor−acceptor−acceptor triads. The steady-state absorption
spectra of these molecules in dry toluene are shown in Figure
1b. Relative to the absorption spectra of the molecules in
toluene−1% pyridine, no significant changes to the spectral
features result from the absence of pyridine in solution. The
λmax of the Chl Soret band shifts slightly to 429 nm, and the Qy
(0,0) maximum shifts to 657 nm and broadens slightly on its
red edge while the NDI peaks do not shift.
When Chl-based systems self-assemble into supramolecular
structures, the close proximity of the Chl’s usually results in
exciton coupling of the Chl transition moments that produces
significant spectral shifts of the ground state absorption
spectra.11b,19 These shifts can be used to determine the degree
of Chl association and in some favorable cases provide
structural information as well. However, in many supra-
molecular Chl systems, the Chl absorption bands remain
unperturbed.11a,20 The nearly negligible spectral shifts in the
Chl−PI−NDI and Chl−PI−NDI2 ground state absorption
spectra in toluene versus toluene−1% pyridine preclude the use
of these spectra to study the self-assembly process. Thus,
characterizing the self-assembled species requires more
advanced techniques.
angle scattering region, q < 0.2 Å−1, the data were fit to the
Guinier relationship:
2
2
I(q) = I(0) exp(−q R /3)
g
(1)
where, I(0) is the forward scattering amplitude and Rg is the
radius of gyration of the assembly. A linear Guinier fit over the
data range 0.012 Å−2 < q2 < 0.02 Å−2 is shown in Figures S1
and S2 (insets), which indicates the formation of discrete
molecular aggregates well-approximated by spheres.21 A least-
squares fit to this region reveals that Rg = 17.4 and 15.5 Å for
Chl−PI−NDI and Chl−PI−NDI2, respectively. A comparison
of the experimental scattering data with scattering patterns
computed for a variety of model structures for Chl−PI−NDI
and Chl−PI−NDI2 aggregates over the range 0.1 Å−1 < q < 0.5
Å−1 are given in Figures S3 and S4, respectively. The
corresponding experimental data for monomeric Chl−PI−
NDI and Chl−PI−NDI2 in toluene−1% pyridine could not be
successfully obtained by subtraction of the solvent scattering,
presumably as a result of the relatively small size of the
monomers.
The linear Guinier fit allows an analysis using atomic pair
distribution functions (PDFs) by directly comparing reciprocal-
space scattering patterns and real-space molecular models. The
experimental PDFs for Chl−PI−NDI and Chl−PI−NDI2 were
obtained using the X-ray scattering fitting program GNOM22
and model PDFs were generated using the same method from
MM+ geometry-optimized models.23 Figure 2 compares the
experimental PDFs of Chl−PI−NDI and Chl−PI−NDI2 in
toluene to those of the corresponding monomer and cyclic
tetramer structural models. The number of major atom pair
distances and their values determined from the data correlate
Characterization of Self-Assembled Structures Using
X-ray Scattering. Self-assembly of the Chl−PI−NDI and
Chl−PI−NDI2 building blocks in toluene was probed by small-
and wide-angle X-ray scattering (SAXS/WAXS) experiments
performed at the Advanced Photon Source at Argonne
National Laboratory. The radially averaged scattering intensities
for Chl−PI−NDI and Chl−PI−NDI2 are shown in Supporting
Information Figures S1 and S2. In the low-resolution small-
4366
dx.doi.org/10.1021/ja211329k | J. Am. Chem. Soc. 2012, 134, 4363−4372

Journal of the American Chemical Society
Article
well with the cyclic tetramer models, while the observed
discrepancies in PDF peak intensities, that is, “amplitude
damping”, result from specific solvent interactions with the
structures.24 The Chl−PI−NDI data (Figure 2a) shows
significant peak intensities at 17.9, 26.0, and 32.8 Å, with a
weaker peak at 39.9 Å, while the corresponding peaks in the
static energy-minimized cyclic tetramer model are 19.3, 28.8,
and 37.3 Å, with a weaker peak at 44.6 Å. Figure 2a shows that
the corresponding monomeric model has no features beyond
20 Å, while alternative oligomeric linear models exhibit a series
of features that extend out to 70 Å (Figure S5). Also, forcing
the NDI groups to adopt a conformation in which they reside
inside the cyclic tetramer does not model the data well (Figure
S6). While the cyclic tetramer structure provides the best
model for the experimental data, the static model does not
capture the dynamic nature of the structure. As distance
increases, the linewidths of the experimental PDF features
increase as do their deviation from the distances predicted by
the static model structure. This most likely reflects both the
flexibility of the cyclic structure and the conformational
dispersity of Chl, PI, and NDI relative to one another as a
result of rotations about their single-bond linkages and the
ability of the pyridine to ligate the Chl zinc atom from either of
the two nonequivalent faces of the Chl macrocycle.
photogenerated ion pairs (Chl+•−PI−•−NDI and Chl+•−PI−•−
NDI2) above the ground state, ΔGIP:28
2
⎛
⎞
e
1
1
2
ΔG = E − E
−
+ e
+
⎜
⎝
⎟
⎠
IP
Ox
Red
r
ε
2r
2r
DA S
D
A
⎛
⎞
1
ε
1
×
−
⎜
⎝
⎟
⎠
ε
SP
(2)
S
where rDA is the donor−acceptor distance, rD and rA are the
respective ionic radii of the donor and acceptor, εS is the static
dielectric constant of the nonpolar solvent, εSP is the static
dielectric constant of a high polarity reference, and e is the
charge of an electron. A value of rDA = 12.1 Å was determined
for the Chl−PI distance using a geometry optimized MM+
calculation in Hyperchem23 by estimating the distance from the
Chl metal center to the center of the PI benzene ring. The ionic
radii for the donor and acceptor were each approximated as
r
DA/2.29 Using the experimental redox potentials measured in
DCM, ΔGIP = 1.80 eV for Chl+•−PI−•−NDI and Chl+•−PI−•−
NDI2 in toluene.
The free energies for the initial charge separation (ΔGCS1
)
and recombination (ΔGCR1) were then determined using eqs 3
and 4, respectively,
ΔG
ΔG
= ΔG − E
(3)
(4)
CS1
IP
S
= −ΔG
CR1
IP
The corresponding experimental PDF for Chl−PI−NDI2
shows peaks at 20.1, 26.1, 36.3, 42.1, 48.3, and 54.6 Å, while the
model shows peaks at 18.2, 27.1, 35.4, 42.7, 47.1, and 54.2 Å.
There is a small peak at 31.8 Å in the experimental data that is
not captured in the model; however, the breadth and flat top of
the 27.1 Å peak in the model PDF are indicative of two
unresolved peaks. The experimental Chl−PI−NDI2 PDF
features are narrower, better resolved, and the predicted
distances more closely match those of the cyclic tetramer
model relative to what is observed for Chl−PI−NDI. Once
again, comparisons of the experimental PDF data with PDFs
generated from both monomer and linear oligomer models
(Figure S7) show that these models describe the data poorly.
The data indicate that the Chl−PI−NDI2 cyclic tetramer has
less conformational dispersity than does Chl−PI−NDI. The
additional NDI groups restrict internal rotations within the
Chl−PI−NDI2 cyclic tetramer and essentially provide walls for
the box that metal−ligand coordination creates, which is
evident in the energy-minimized model shown in Figure 2b.
Charge Transfer Energetics. The one-electron reduction
of PI to PI−• was observed at −0.76 V versus SCE, while the
one-electron oxidation of 20-phenyl-Chl occurs at 0.63 V versus
SCE. PI has been shown to exhibit significant variations in its
reduction potential upon substitution at the imide position.7b,25
However, much smaller potential shifts are observed upon
increased substitution of PI on its central benzene core.26 The
one-electron reduction of NDI within Chl−PI−NDI and Chl−
PI−NDI2 occurs at −0.51 V versus SCE; thus, the redox
potentials for Chl, PI, and NDI all agree well with previous
measurements.4,27 In a nonpolar solvent (e.g., toluene), the
Weller equation can be used to estimate the energy of the initial
where the lowest excited singlet state energies (ES) for Chl
were determined by averaging the energies of lowest energy
absorption and the highest energy fluorescence maxima of 20-
phenyl-Chl, so that ES = 1.89 eV. The resulting values show
that the free energy for the reactions 1*Chl−PI−NDI →
1
Chl+•−PI−•−NDI and *Chl−PI−NDI2 → Chl+•−PI−•−NDI2
are both ΔGCS1 = −0.09 eV and ΔGCR1 = −1.80 eV,
respectively.
For the triad systems, the free energies of the charge shift
reactions, Chl+•−PI−•−NDI → Chl+•−PI−NDI−• and Chl+•−
PI−•−NDI → Chl+•−PI−(NDI)NDI−• were determined from
the difference in redox potentials of the acceptors and the
change in Coulomb energy needed to separate the charges
farther apart.6b Therefore,
ΔG
= ΔG
+ E (PI) − E (NDI)
IP2
IP1
Red
Red
2
⎛
⎞
e
1
1
r
IP2
−
+
⎜
⎟
⎠
ε
r
⎝
(5)
S
IP1
where ΔGIP1 and ΔGIP2 are the free energies for the formation
of the initial ion pair (IP1) state (Chl+•−PI−•−NDI and
Chl+•−PI−•−NDI2) and the final ion pair (IP2) state (Chl+•−
PI−NDI−• and Chl+•−PI−(NDI)NDI−•), respectively. As
above, the experimental redox potentials for the primary and
secondary acceptors were used, and MM+ geometry optimized
structures were used to determine the necessary ion pair
distances (rIP1 = 12.1 Å and rIP2 = 14.4 Å). The calculated free
energies of the charge shift reactions for both Chl−PI−NDI
and Chl−PI−NDI2 are ΔGCS2 = ΔGIP2 − ΔGIP1 = −0.17 eV,
while ΔGCR2 = −ΔGIP2 = −1.63 eV for the distal radical ion
pairs.
The free energies for the two charge separation steps offer an
interesting contrast to those of the linear dyads/triad studied
previously,7a which use donors and acceptors that are very
similar to those used in the perpendicular triads presented here.
4367
dx.doi.org/10.1021/ja211329k | J. Am. Chem. Soc. 2012, 134, 4363−4372

Journal of the American Chemical Society
Article
1
The major difference between the linear and perpendicular
series resides in the connectivity between the redox
components. The perpendicular orientation between Chl and
the secondary NDI acceptor in the compounds described here
results in shorter distance changes between Chl+•−PI−•−NDI
and Chl+•−PI−NDI−• relative to those in the linear systems.
This decreases the Coulomb energy penalty (last term in eq 5)
required to move the opposite charges farther apart in the
charge shift reaction, Chl+•−PI−•−NDI → Chl+•−PI−NDI−•.
Since through-bond electron transfer usually dominates in
covalent donor−acceptor molecules, the development of
systems with long through-bond pathways and short through-
space pathways offers the possibility of minimizing the
Coulomb energy penalty without sacrificing the long charge
separation lifetimes that are advantageous for artificial photo-
synthesis.
416 nm, 25 fs pulses, *Chl was shown to decay with τ = 4.11
0.03 ns, while analogous experiments in toluene reveal a τ =
3.29 0.02 ns lifetime.
When the Chl-based triads are dissolved in toluene−1%
pyridine, the pyridine coordinates to the zinc metal center to
disrupt self-assembly. Therefore, it is possible to probe the
charge transfer dynamics of the monomeric molecular building
blocks in a solvent with nearly the same dielectric constant as
the solvent in which the monomers self-assemble into the cyclic
tetramers. FsTA spectra for Chl−PI−NDI in toluene−1%
pyridine are shown in Figure 4a. Photoexcitation of the Chl
Soret band generates 1*Chl, which at early times shows spectral
features similar to those of 20-phenyl-Chl (Figure 3). However,
as time evolves, a sharp absorption band centered at 726 nm
appears, which is characteristic of PI−•.7a Observation of the
spectral changes resulting from the oxidation of 1*Chl to Chl+•
is difficult due to their very similar absorption spectra.4,7a Chl+•
has a weak feature at 820 nm32 that cannot be observed in these
experiments because the cutoff filters needed to eliminate the
residual 832 nm laser line block wavelengths near 820 nm from
the probe beam.
Charge Transfer Dynamics. Previous photophysical
studies on self-assembled Chl-based structures have been
directed largely toward understanding Chl−Chl excitation
energy transfer (EET) in constructs having multiple identical
Chl’s.1a,30 The closest Chl−Chl distances in the self-assembled
cyclic tetramers studied here are approximately 19 Å and are
The disappearance of the spectral features due to PI−• are
accompanied by formation of an absorption band at 475 nm,
which is characteristic of NDI−•,27 and has been seen in the
analogous Chl−NDI dyad.4,7a Analysis of the fsTA results using
singular value decomposition (SVD) and global fitting18 yields
the decay-associated spectra given in Figure 4b. The resulting
time constants, τCS1 and τCS2, for the two charge separation
thus within a reasonable range to observe Forster-type EET
̈
between the Chl’s. Using the Forster equation and
̈
experimental inputs (see Supporting Information), the
predicted EET rate constant between the Chl’s in the cyclic
tetramers is 2 × 1010 s−1 (τEET = 50 ps). As is seen below,
1
electron transfer from *Chl to the PI acceptor is faster in all
steps, *Chl−PI−NDI → Chl+•−PI−•−NDI and Chl+•−PI−•−
1
cases and significant Chl−Chl EET is not observed. The
femtosecond transient absorption (fsTA) kinetics are all laser
power independent, which indicates that singlet−singlet
annihilation does not occur in these systems by EET (Figure
S8).31
FsTA measurements on the 20-phenyl-Chl electron donor
were performed in toluene−1% pyridine using 416 nm, 110 fs
excitation pulses (Figure 3). Photoexcitation generates the Chl
NDI → Chl+•−PI−NDI−•, respectively, were determined from
the global fitting parameters and the quantum yield of NDI−• as
follows. The time constant for the primary charge separation
(τCS1) is obtained directly from the time constant of the decay-
associated spectrum given in Figure 4b, green trace, which
shows the disappearance of 1*Chl accompanied by the
appearance of PI−•. Given the 3−4 ns *Chl lifetime in 20-
1
phenyl-Chl and its corresponding 8−20 ps lifetimes in the
triads, the initial charge separation step (formation of PI−•)
proceeds with essentially unity quantum yield. To obtain the
time constant for the secondary charge shift reaction from PI−•
to NDI, the quantum yield of NDI−• formation is needed. As
reported earlier,4 photoexcitation of the corresponding Chl−
NDI dyad results in a unity quantum yield of Chl+•−NDI−•,
which produces a transient absorption spectrum for which
|ΔA475nm/ΔA658nm| = 0.47. The value of the ΔA658nm bleach
serves as an internal standard to gauge how much ground state
Chl is depleted by the laser pulse. The corresponding ratio
observed for Chl+•−PI−NDI−• divided by 0.47 gives the
NDI−• quantum yield (Table 1). The time constant (τPI) for
the disappearance of PI−• is that of the decay-associated
spectrum in Figure 4b, red trace. The rate constant for PI−•
Figure 3. Femtosecond TA spectra of 20-phenyl-Chl in toluene−1%
pyridine excited at 416 nm. Inset shows transient kinetics at 667 nm.
disappearance (kPI = 1/τPI) is assumed to equal kCS2 + kCR1
,
where kCS2 is the rate constant for charge shift from PI−• to
NDI and kCR1 is the rate constant for charge recombination of
Chl+•−PI−•−NDI. Thus, ϕNDI = kCS2/(kCS2 + kCR1) = kCS2τPI,
S2 state, which rapidly undergoes internal conversion from S2 to
S1 within the 150-fs IRF of the experiment.4,20,31d The ground
state bleach of the Qy(0,0) band is observed at 658 nm with
stimulated emission contributing to the asymmetry of the red
edge of the bleach along with a small negative vibronic feature
at 725 nm. The broad positive features throughout the
spectrum are attributed to 1*Chl absorption, and are
punctuated by ΔA due to the ground state bleach. The lifetime
or since τCS2 = 1/kCS2, τCS2 = τPI/ϕNDI and thus τCR1
=
ϕNDIτCS2/(1 − ϕNDI). This analysis is used to obtain τCS1, τCS2
,
−•
τCR1, and ϕNDI for each triad in both toluene−1% pyridine and
toluene (Table 1).
FsTA spectroscopy measurements were also performed on
Chl−PI−NDI in toluene (Figure 5a), and yield transient
spectra similar to those of Chl−PI−NDI in toluene−1%
pyridine. The charge separation and recombination time
1
of *Chl was independently determined using time-resolved
fluorescence (TRF) spectroscopy. After photoexcitation with
4368
dx.doi.org/10.1021/ja211329k | J. Am. Chem. Soc. 2012, 134, 4363−4372

Journal of the American Chemical Society
Article
Figure 4. Femtosecond TA spectra and decay-associated spectra of (a and b) Chl−PI−NDI and (c and d) Chl−PI−NDI2 in toluene−1% pyridine
excited at 416 nm.
Table 1. Charge Separation and Recombination Lifetimes and NDI−• Quantum Yields
−•
compound
solvent
τCS1 (ps)
τCS2 (ps)
τCR1 (ps)
τCR2 (ns)
ϕNDI
Chl−PI−NDI
Chl−PI−NDI
Chl−PI−NDI2
Chl−PI−NDI2
tol−1% pyr
Tol
19
15
15
8
1
1
1
1
148 10
518 20
217 10
827 36
299 10
778 30
300 10
1050 36
5−10
5−10
10
0.67 0.05
0.60 0.05
0.58 0.05
0.56 0.05
tol−1% pyr
Tol
1
1
30
constants derived using the analysis described above are
summarized in Table 1. Charge recombination of Chl+•−PI−
NDI−• in both toluene−1% pyridine and toluene occurs with
τCR2 = 5−10 ns and is difficult to assess more accurately given
the 6 ns limit of the pump−probe delay time of the fsTA
apparatus and the 7 ns instrument response function of the
nanosecond TA (nsTA) apparatus.
reasons, in addition to the fact that the distances over which
through-bond electron transfer occurs within the Chl−PI−NDI
and Chl−PI−NDI2 building blocks does not change when they
form the cyclic tetramers, the total nuclear reorganization
energy should not change appreciably. Thus, the observed
differences in τCS1, τCS2, and τCR1 between the Chl−PI−NDI
monomer and the cyclic tetramer most likely result from
changes in the electronic coupling matrix elements for charge
transfer. While this explanation is consistent with the SAXS/
WAXS data, which indicate that cyclic tetramer formation
restricts torsional motions about the single bonds linking the
donor and acceptors, the scattering data do not provide
sufficient structural resolution to identify the relevant dihedral
angle changes in the donor−acceptor linkages. Nevertheless,
changing torsional conformation distributions in donor−
bridge−acceptor molecules has previously been shown to
exert significant control over electron transfer rates.34 It is
important to note, however, that restricting the distribution of
conformations around the bonds linking the donor and
acceptors by assembling supramolecular structures does not
necessarily result in a set of conformations that favor charge
transfer. Depending on which conformations are important for
controlling the electronic coupling matrix elements for charge
transfer, self-assembly may result in conformations either
favorable or unfavorable for facile charge transfer. Careful
Comparing the Chl−PI−NDI monomer (toluene−1%
pyridine) and cyclic tetramer (toluene) data, the initial charge
separation (τCS1) occurs slightly faster in the tetramer. In
contrast, the secondary charge shift (τCS2) slows by about a
factor of 3 in the tetramer, yet the quantum yield of Chl+•−PI−
NDI−• decreases by only about 10%. This is consistent with the
fact that the time constant for charge recombination (τCR1) of
Chl+•−PI−•−NDI also slows in the cyclic tetramer to about the
same degree as does τCS2 for the charge shift reaction. Electron
transfer theory shows that the rate of charge transfer depends
on the free energy of reaction, the total nuclear reorganization
energy, and the electronic coupling matrix element for the
reaction.33 The free energies of reaction for CS1 and CS2 do
not change appreciably in going from toluene−1% pyridine to
toluene because the dielectric constants of these two media are
nearly identical and a single pyridine from either the solvent
(toluene−1% pyridine) or from the pyridine attached to PI
(toluene) is coordinated to the zinc atom of Chl. For the same
4369
dx.doi.org/10.1021/ja211329k | J. Am. Chem. Soc. 2012, 134, 4363−4372

Journal of the American Chemical Society
Article
Figure 5. Femtosecond TA spectra and decay-associated spectra of (a and b) Chl−PI−NDI, and (c and d) Chl−PI−NDI2 in toluene excited at 416
nm.
molecular design is therefore necessary to optimize the
supramolecular structure to favor the desired charge transfer
rate.
show that the Chl−PI−NDI2 cyclic tetramer has less structural
dispersity than does the corresponding Chl−PI−NDI cyclic
tetramer, it is reasonable that greater restriction of torsional
angles between the donor and acceptors in the Chl−PI−NDI2
cyclic tetramer should lead to a larger difference between the
monomer and tetramer charge separation rates. This difference
Photoexcitation of Chl−PI−NDI2 exhibits representative
1*Chl absorption, stimulated emission, and ground state bleach
features at 1.6 ps followed within tens of picoseconds by the
appearance of the characteristic PI−• absorption band at 726
nm (Figure 4c). The decay-associated spectra show that this
PI−• absorption band decays with τ = 126 4 ps (Figure 4d)
and is accompanied by the formation of the 480-nm NDI−•
band27 that, in turn, decays on a nanosecond time scale (Figure
S9a). NsTA spectroscopy shows that the 480-nm NDI−•
absorption decays with biexponential kinetics following photo-
excitation of Chl−PI−NDI2 with 416 nm, 7 ns pulses (Figure
S9b). The first decay component (τCR2 = 10 ns) is assigned to
Chl+•−PI−(NDI)NDI−• charge recombination while the
second oxygen-sensitive, residual long-lived component (τ >
300 ns) is assigned to 3*Chl formed by radical pair intersystem
crossing35 of 1(Chl+•−PI−(NDI)NDI−•) to 3(Chl+•−PI−
(NDI)NDI−•) followed by charge recombination.
The corresponding fsTA and nsTA spectra of Chl−PI−NDI2
in toluene are shown in Figure 5c,d and Figures S9c,d,
respectively. The observed trends in τCS1 and τCS2 in comparing
the Chl−PI−NDI2 monomer and cyclic tetramer are similar to
those observed for Chl−PI−NDI (Table 1), but are
accentuated somewhat, so that τCS1 is shorter by about a factor
of 2 and τCS2 is longer by about a factor of 4 in the cyclic
tetramer. Once again, there is very little change in the Chl+•−
PI−(NDI)NDI−• quantum yield between the monomer and
the cyclic tetramer, indicating that the charge recombination
time constant (τCR1) of the initial Chl+•−PI−•−NDI2 ion pair
also slows commensurately. Given that the SAXS/WAXS data
is also enhanced in the charge recombination rates (τCR2
)
measured by nsTA for the Chl−PI−NDI2 monomer and cyclic
tetramer. The value of τCR2 is a factor of 3 slower in the cyclic
tetramer and provides a potentially useful strategy for extending
charge separation lifetimes by using self-assembly to control the
electronic coupling matrix elements for charge transfer.
CONCLUSIONS
■
Donor−acceptor distances and orientations, their electronic
interaction strengths, and free energies of reaction all
significantly affect charge transfer rates and efficiencies. The
studies described here show how supramolecular assembly can
influence photoinduced charge transfer dynamics. Self-assembly
of the Chl−PI−NDI and Chl−PI−NDI2 triads into cyclic
tetramers increases their overall charge separation lifetimes and
suggests that controlling supramolecular structure can be
beneficial for tailoring charge transfer dynamics. The
fundamental molecular level understanding gained in such
systems provides key design guidance for the fabrication of
artificial photosynthetic systems that use self-assembly to
generate well-defined supramolecular structures to enhance
their photophysical and photochemical functionality.
4370
dx.doi.org/10.1021/ja211329k | J. Am. Chem. Soc. 2012, 134, 4363−4372

Journal of the American Chemical Society
Article
Tamiaki, H. Tetrahedron Lett. 2008, 49, 4113−4115. (c) Huber, V.;
ASSOCIATED CONTENT
* Supporting Information
Experimental details including synthesis, absorption spectra,
and additional SAXS/WAXS results and transient absorption
spectra and kinetics. This material is available free of charge via
the Internet at http://pubs.acs.org.
■
Sengupta, S.; Wurthner, F. Chem.Eur. J. 2008, 14, 7791−7807.
̈
S
(11) (a) Kelley, R. F.; Goldsmith, R. H.; Wasielewski, M. R. J. Am.
Chem. Soc. 2007, 129, 6384−6385. (b) Kelley, R. F.; Lee, S. J.; Wilson,
T. M.; Nakamura, Y.; Tiede, D. M.; Osuka, A.; Hupp, J. T.;
Wasielewski, M. R. J. Am. Chem. Soc. 2008, 130, 4277−4284.
(12) (a) Jensen, R. A.; Kelley, R. F.; Lee, S. J.; Wasielewski, M. R.;
Hupp, J. T.; Tiede, D. M. Chem. Commun. 2008, 1886−1888.
(b) Yatskou, M. M.; Koehorst, R. B. M.; Donker, H.; Schaafsma, T. J. J.
Phys. Chem. A 2001, 105, 11425−11431. (c) Kobuke, Y. Eur. J. Inorg.
Chem. 2006, 2006, 2333−2351. (d) Hwang, I.-W.; Kamada, T.; Ahn,
T. K.; Ko, D. M.; Nakamura, T.; Tsuda, A.; Osuka, A.; Kim, D. J. Am.
Chem. Soc. 2004, 126, 16187−16187.
AUTHOR INFORMATION
Corresponding Author
m-wasielewski@northwestern.edu
■
Notes
(13) (a) Bucks, R. R.; Boxer, S. G. J. Am. Chem. Soc. 1982, 104, 340−
343. (b) Bucks, R. R.; Netzel, T. L.; Fujita, I.; Boxer, S. G. J. Phys.
Chem. 1982, 86, 1947−1955.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
(14) Smith, K. M.; Goff, D. A.; Simpson, D. J. J. Am. Chem. Soc. 1985,
107, 4946−4954.
This work was supported by the Chemical Sciences, Geo-
sciences, and Biosciences Division, Office of Basic Energy
Sciences, DOE under grant no. DE-FG02-99ER14999. D.T.C.
and C.H.K. thank the Initiative for Sustainability and Energy at
Northwestern for support.
(15) Kenner, G. W.; McCombie, S. W.; Smith, K. M. J. Chem Soc.,
Perkin Trans. 1 1973, 2517−2523.
(16) Seifert, S.; Winans, R. E.; Tiede, D. M.; Thiyagarajan, P. J. Appl.
Crystallogr. 2000, 33, 782−784.
(17) Samuel, A. P. S; Co, D. T.; Stern, C. L.; Wasielewski, M. R. J.
Am. Chem. Soc. 2010, 132, 8813−8815.
REFERENCES
■
(18) Surface Xplorer, Ultrafast Systems LLC, Sarasota, FL 34234,
USA.
(1) (a) Wasielewski, M. R. Acc. Chem. Res. 2009, 42, 1910−1921.
(b) Wasielewski, M. R. J. Org. Chem. 2006, 71, 5051−5066.
(c) Magnuson, A.; Anderlund, M.; Johansson, O.; Lindblad, P.;
(19) Katz, J. J.; Shipman, L. L.; Cotton, T. M.; Janson, T. R. In The
Porphyrins; Dolphin, D., Ed.; Academic Press: New York, 1978; Vol. V.
(20) Gunderson, V. L.; Mickley Conron, S. M.; Wasielewski, M. R.
Chem. Commun. 2010, 46, 401−403.
(21) Svergun, D. I.; Koch, M. H. J. Rep. Prog. Phys. 2003, 66, 1735−
1782.
(22) Svergun, D. I. J. Appl. Crystallogr. 1992, 25, 495−503.
(23) Hypercube, Inc., 1115 NW 4th Street, Gainesville, Florida
32601, USA, 1994.
Lomoth, R.; Polivka, T.; Ott, S.; Stensjo, K.; Styring, S.; Sundstrom, V.;
̈
̈
Hammarstrom, L. Acc. Chem. Res. 2009, 42, 1899−1909. (d) Gust, D.;
̈
Moore, T. A.; Moore, A. L. Acc. Chem. Res. 2009, 42, 1890−1898.
(e) McConnell, I.; Li, G.; Brudvig, G. W. Chem. Biol. 2010, 17, 434−
447. (f) Kalyanasundaram, K.; Graetzel, M. Curr. Opin. Biotechnol.
2010, 21, 298−310.
(2) (a) Wurthner, F. Chem. Commun. 2004, 1564−1579. (b) Chen,
̈
Z.; Lohr, A.; Saha-Moeller, C. R.; Wuerthner, F. Chem. Soc. Rev. 2009,
38, 564−584. (c) Aratani, N.; Kim, D.; Osuka, A. Acc. Chem. Res. 2009,
42, 1922−1934.
(24) (a) O’Donnell, J. L.; Zuo, X.; Goshe, A. J.; Sarkisov, L.; Snurr, R.
Q.; Hupp, J. T.; Tiede, D. M. J. Am. Chem. Soc. 2007, 129, 1578−1585.
(b) Tiede, D. M.; Zhang, R.; Chen, L. X.; Yu, L.; Lindsey, J. S. J. Am.
Chem. Soc. 2004, 126, 14054−14062.
(3) (a) Wasielewski, M. R. In Chlorophylls; Scheer, H., Ed.; CRC
Press: Boca Raton, FL, 1991; pp 269−286; (b) Regehly, M.; Ermilov,
E. A.; Helmreich, M.; Hirsch, A.; Jux, N.; Roder, B. J. Phys. Chem. B
2007, 111, 998−1006. (c) Shinoda, S.; Tsukube, H.; Nishimura, Y.;
Yamazaki, I.; Osuka, A. Tetrahedron 1997, 53, 13657−13666.
(d) Tauber, A. Y.; Helaja, J.; Kilpelainen, I.; Hynninen, P. H. Acta
Chem. Scand. 1997, 51, 88−93.
(4) Kelley, R. F.; Tauber, M. J.; Wasielewski, M. R. J. Am. Chem. Soc.
2006, 128, 4779−4791.
(5) Blankenship, R. E. Molecular Mechanisms of Photosynthesis;
Blackwell Science: Oxford, 2002.
(6) (a) Gust, D.; Moore, T. A.; Moore, A. L. In Mimicking Bacterial
Photosynthesis; Collings, A. F., Critchley, C., Eds.; Wiley-VCH Verlag
GmbH & Co. KGaA: Weinheim, Germany, 2006. (b) Greenfield, S.
R.; Svec, W. A.; Gosztola, D.; Wasielewski, M. R. J. Am. Chem. Soc.
1996, 118, 6767−6777. (c) Bullock, J. E.; Carmieli, R.; Mickley, S. M.;
Vura-Weis, J.; Wasielewski, M. R. J. Am. Chem. Soc. 2009, 131, 11919−
11929.
(25) Viehbeck, A.; Goldberg, M. J.; Kovoc, C. A. J. Electrochem. Soc.
1990, 137, 1460−1466.
̈
(26) Ahrens, M. J.; Sinks, L. E.; Rybtchinski, B.; Liu, W.; Jones, B. A.;
Giaimo, J. M.; Gusev, A. V.; Goshe, A. J.; Tiede, D. M.; Wasielewski,
M. R. J. Am. Chem. Soc. 2004, 126, 8284−8294.
(27) Gosztola, D.; Niemczyk, M. P.; Svec, W.; Lukas, A. S.;
Wasielewski, M. R. J. Phys. Chem. A 2000, 104, 6545−6551.
(28) Weller, A. Z. Phys. Chem 1982, 133, 93−98.
(29) Basilevsky, M. V.; Rostov, I. V.; Newton, M. D. Chem. Phys.
1998, 232, 189−199.
(30) (a) Miyatake, T.; Tamiaki, H. Coord. Chem. Rev. 2010, 254,
2593−2602. (b) Balaban, T. S.; Bhise, A. D.; Bringmann, G.; Burck, J.;
̈
Chappaz-Gillot, C.; Eichhofer, A.; Fenske, D.; Gotz, D. C. G.; Knauer,
̈
̈
M.; Mizoguchi, T.; Mossinger, D.; Rosner, H.; Roussel, C.; Schraut,
M.; Tamiaki, H.; Vanthuyne, N. J. Am. Chem. Soc. 2009, 131, 14480−
̈
̈
14492.
(31) (a) Larsen, J.; Bruggemann, B.; Polivka, T.; Sunderstrom, V.;
Akesson, E.; Sly, J.; Crossley, M. J. J. Phys. Chem. A 2005, 109, 10654−
10662. (b) Bruggemann, B.; Herek, J. L.; Sunderstrom, V.; Pullerits,
T.; May, V. J. Phys. Chem. B 2001, 105, 11391−11394. (c) Valkunas,
L.; Trinkunas, G.; Liuolia, V.; Grondelle, R. v. Biophys. J. 1995, 69,
1117−1129. (d) Gunderson, V. L.; Wilson, T. M.; Wasielewski, M. R.
J. Phys. Chem. C 2009, 113, 11936−11942.
(7) (a) Wiederrecht, G. P.; Niemczyk, M. P.; Svec, W. A.;
Wasielewski, M. R. J. Am. Chem. Soc. 1996, 118, 81−88.
(b) Wiederrecht, G. P.; Svec, W. A.; Niemczyk, M. P.; Wasielewski,
M. R. J. Phys. Chem. 1995, 99, 8918−8926. (c) Wiederrecht, G. P.;
Svec, W. A.; Wasielewski, M. R. In Ultrafast Phenomena IX; Barbara, P.
F., Knox, W. H., Eds.; Springer-Verlag: Berlin, 1994; Vol. 60, pp 452−
453.
(8) Beletskaya, I.; Tyurin, V. S.; Tsivadze, A. Y.; Guilard, R.; Stern, C.
Chem. Rev. 2009, 109, 1659−1713.
(9) Balaban, T.; Tamiaki, H.; Holzwarth, A. In Supermolecular Dye
(32) Borg, D. C.; Fajer, J.; Felton, R. H.; Dolphin, D. Proc. Natl. Acad.
Sci. U.S.A. 1970, 67, 813−820.
(33) (a) Marcus, R. A. J. Chem. Phys. 1956, 24, 966−978. (b) Levich,
V. G.; Dogonadze, R. R. Collect. Czech. Chem. Commun. 1961, 26,
193−214. (c) Marcus, R. A. J. Chem. Phys. 1965, 43, 679−701.
(d) Hopfield, J. J. Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 3640−3644.
(e) Jortner, J. J. Chem. Phys. 1976, 64, 4860−4867.
Chemistry; Wurthner, F., Ed.; Springer: Berlin/Heidelberg, 2005; Vol.
̈
258, pp 585−585.
(10) (a) Sasaki, S.-i.; Omoda, M.; Tamiaki, H. Photochem. Photobiol.,
A 2004, 162, 307−315. (b) Sasaki, S.-i.; Mizutani, K.; Kunieda, M.;
4371
dx.doi.org/10.1021/ja211329k | J. Am. Chem. Soc. 2012, 134, 4363−4372

Journal of the American Chemical Society
Article
(34) (a) Davis, W. B.; Ratner, M. A.; Wasielewski, M. R. J. Am. Chem.
Soc. 2001, 123, 7877−7886. (b) Weiss, E. A.; Tauber, M. J.; Kelley, R.
F.; Ahrens, M. J.; Ratner, M. A.; Wasielewski, M. R. J. Am. Chem. Soc.
2005, 127, 11842−11850.
(35) (a) Buckley, C. D.; Hunter, D. A.; Hore, P. J.; McLauchlan, K.
A. Chem. Phys. Lett. 1987, 135, 307−312. (b) Till, U.; Hore, P. J. Mol.
Phys. 1997, 90, 289−296.
4372
dx.doi.org/10.1021/ja211329k | J. Am. Chem. Soc. 2012, 134, 4363−4372