Making a Molecular Wire: Charge and Spin Transport through para-Phenylene Oligomers

Article abstract of DOI:10.1021/ja0398215

Functional molecular wires are essential for the development of molecular electronics. Charge transport through molecules occurs primarily by means of two mechanisms, coherent superexchange and incoherent charge hopping. Rates of charge transport through

Full text of DOI:10.1021/ja0398215

Published on Web 04/01/2004
Making a Molecular Wire: Charge and Spin Transport through
para-Phenylene Oligomers
Emily A. Weiss, Michael J. Ahrens, Louise E. Sinks, Alexey V. Gusev,
Mark A. Ratner,* and Michael R. Wasielewski*
Contribution from the Center for Nanofabrication and Molecular Self-Assembly,
Department of Chemistry, Northwestern UniVersity, EVanston, Illinois 60208-3113
Received November 26, 2003; E-mail: wasielew@chem.northwestern.edu
Abstract: Functional molecular wires are essential for the development of molecular electronics. Charge
transport through molecules occurs primarily by means of two mechanisms, coherent superexchange and
incoherent charge hopping. Rates of charge transport through molecules in which superexchange dominates
decrease approximately exponentially with distance, which precludes using these molecules as effective
molecular wires. In contrast, charge transport rates through molecules in which incoherent charge hopping
prevails should display nearly distance independent, wirelike behavior. We are now able to determine how
each mechanism contributes to the overall charge transport characteristics of a donor-bridge-acceptor
(D-B-A) system, where D ) phenothiazine (PTZ), B ) p-oligophenylene, and A ) perylene-3,4:9,10-
bis(dicarboximide) (PDI), by measuring the interaction between two unpaired spins within the system’s
charge separated state via magnetic field effects on the yield of radical pair and triplet recombination product.
Introduction
The D-B-A system uses a series of p-phenylene (Phn)
oligomers, where n ) 1-5, to link a phenothiazine (PTZ)
A molecular wire is best understood as a molecular bridge
that can move charge rapidly and efficiently over many chemical
bond lengths. A major challenge in current chemistry lies in
finding molecules that exhibit long-distance charge-transport
mechanisms operating with efficiencies found widely in nature.
In particular, the photosynthetic reaction center protein rapidly
moves electrons with near unity quantum efficiency across a
bilayer lipid membrane using several redox cofactors,1 thus
serving as a model for developing synthetic, biomimetic
analogues for several applications including solar energy
conversion, molecular electronics, and photonic materials.
In assembling electron donor-bridge-acceptor (D-B-A)
systems for molecular electronics,2 one must design the system
to adopt the most efficient charge-transport mechanism possible,
one which maintains this efficiency as the bridge is lengthened.
Long-distance charge transfer (CT) is intrinsically a nonadiabatic
process3,4 in which the CT rate is dictated by some combination
of strongly distance-dependent coherent transport and weakly
distance-dependent incoherent charge hopping. Optimizing the
molecular structure to accentuate the latter process is key to
producing wirelike behavior in molecules. To achieve this goal
one must (1) isolate the contributions of each mechanism to
CT, (2) find the link between these contributions and the energy
levels of the system, and (3) choose redox components (donors,
bridges, and acceptors) that drive the system toward incoherent
behavior at long distances. Here, we present results on a
D-B-A system that addresses these issues directly.
electron donor to a perylene-3,4:9,10-bis(dicarboximide) (PDI)5
electron acceptor, Figure 1. Selective photoexcitation of PDI
within PTZ-Phn-PDI results in charge separation to produce a
spin-coherent singlet radical ion pair (RP), 1(PTZ+•-Phn-
PDI-•), which subsequently undergoes radical pair intersystem
3
crossing (RP-ISC) to yield (PTZ+•-Phn-PDI-•). The triplet
RP then recombines to give almost exclusively the lowest
excited triplet state of PDI (3*PDI). The RP-ISC mechanism
is well-known to account for triplet state formation within
photosynthetic reaction centers and in selected biomimetic
systems.6-12 The rate of decay of coherent charge and spin
transfer is monitored as the bridge length is increased by directly
measuring the distance dependence of the spin-spin exchange
interaction within the RP.11,12 Measurements of the overall
charge recombination rates at long distances show that charge
transport is dominated by incoherent hopping, that is, wire-
like behavior. Thus, the charge recombination mechanism
changes dramatically from coherent superexchange to incoherent
hopping as the bridge is lengthened. A simple model based on
(5) Van der Boom, T.; Hayes, R. T.; Zhao, Y.; Bushard, P. J.; Weiss, E. A.;
Wasielewski, M. R. J. Am. Chem. Soc. 2002, 124, 9582.
(6) Dutton, P. L.; Leigh, J. S.; Seibert, M. Biochem. Biophys. Res. Commun.
1972, 46, 406.
(7) Thurnauer, M. C.; Katz, J. J.; Norris, J. R. Proc. Natl. Acad. Sci. U.S.A.
1975, 72, 3270.
(8) Hoff, A. J. Photochem. Photobiol. 1986, 43, 727.
(9) Hasharoni, K.; Levanon, H.; Greenfield, S. R.; Gosztola, D. J.; Svec, W.
A.; Wasielewski, M. R. J. Am. Chem. Soc. 1995, 117, 8055-8056.
(10) Carbonera, D.; Di Valentin, M.; Corvaja, C.; Agostini, G.; Giacometti, G.;
Liddell, P. A.; Kuciauskas, D.; Moore, A. L.; Moore, T. A.; Gust, D. J.
Am. Chem. Soc. 1998, 120, 4398-4405.
(11) Lukas, A. S.; Bushard, P. J.; Weiss, E. A.; Wasielewski, M. R. J. Am.
Chem. Soc. 2003, 125, 3921.
(12) Weiss, E. A.; Ratner, M. A.; Wasielewski, M. R. J. Phys. Chem. A 2003,
107, 3639.
(1) Parson, W. W. Photosynth. Res. 2003, 76, 81-92.
(2) Jortner, J.; Ratner, M. A. Molecular Electronics; Blackwell: London, 1997.
(3) Jortner, J. J. Phys. Chem. 1976, 64, 4860.
(4) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1986, 811, 265.
9
10.1021/ja0398215 CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 5577-5584
5577

A R T I C L E S
Weiss et al.
spins of the RP.17-19,21-23 Therefore, the magnitude of the
magnetic interaction and its behavior as a function of donor-
acceptor distance should precisely mirror that of VDA
.
The McConnell model for superexchange24 predicts an
approximately exponential dependence of VDA on the donor-
acceptor distance, rDA, an assertion consistently verified by
experimental data.25-27 As the length of a bridge increases and
the superexchange interaction becomes small, the rate constant
for charge transfer may be dominated by the incoherent term.
Incoherent or sequential charge transfer involves real intermedi-
ate states that couple to internal nuclear motions of the bridge
and the surrounding medium and are therefore energetically
accessible.28 For long distances, the incoherent, wirelike channel
generally becomes more efficient than the coherent one.29,30
Measuring the Degree of Coherent Transport using Spin-
Spin Interactions. The magnetic interaction between the spins
S1 and S2 for paramagnetic centers 1 and 2 is written in the
form suggested by Heisenberg, Dirac, and Van Vleck31
HEX ) -2J‚S1‚S2
(1)
where J is positive if the spins are parallel and negative if they
are antiparallel. For two spin one-half particles, the eigenvalues
of HEX for S ) S1 + S2 give
Figure 1. (A) Chemical structure of PTZ-Phn-PDI, where n ) 1 for
compound 1, n ) 2 for compound 2, etc. (B) DFT (B3LYP, 6-31G**)
energy-minimized structures for 1 and 3 with their peripheral alkyl groups
and hydrogen atoms removed.
E(S) - E(S - 1) ) -2J‚S
ES-T ) E(1) - E(0) ) -2J
(2)
(3)
The singlet-triplet (S-T) splitting, ES-T, within the RP is
therefore given by the phenomenological parameter, 2J, the
magnitude of the indirect exchange interaction.
The total spin Hamiltonian for radical pairs in solution is
given by32
the relative energies of the RP states involved in the charge
recombination process shows that charge injection into the
bridge leading to wirelike transport requires a near-resonant
interaction between the state in which the donor is oxidized
and that in which the bridge is oxidized.
Superexchange: Coherent Charge Transport. Super-
exchange is thought to be an important mechanism for efficient
electron transfer within the photosynthetic reaction center13,14
and has been studied in various biomimetic systems.15,16 The
term was first used by Kramers17 and later by Anderson18,19 to
describe the indirect exchange coupling of unpaired spins via
orbitals having paired spins, which acquire paramagnetic
character through mixing with charge-transfer excited-state
configurations.19 In the context of electron transfer, super-
exchange is the virtual mediation of charge transport from donor
to acceptor via electronically well-separated bridge orbitals.
HST ) âB (g S + g S ) +
a S ‚ I +
a S ‚ I + H
2k k EX
∑
∑
k
0
1
1
2
2
1i
i
i
(4)
where â is the Bohr magneton, B0 is the applied magnetic field,
g1 and g2 are the electronic g-factors for each radical, S1 and S2
are electron spin operators for the two radicals within the radical
pair, Ii and Ik are nuclear spin operators, a1i and a2k are the
isotropic hyperfine coupling constants of nucleus i with radical
1 and nucleus k with radical 2. For organic radicals such as
those studied here, the small differences in g-factors, included
in the first term of eq 4, contribute to singlet-triplet mixing
only at field strengths of several Tesla, which are not relevant
here. Anisotropic exchange interactions and hyperfine couplings,
The rates of nonadiabatic electron-transfer reactions, kET
,
depend critically on the electronic coupling VDA, whose
magnitude gives the effective interaction energy between the
relevant orbitals on the donor and acceptor.3,20 When the charge-
transport process originates from a state in which the redox
centers are also paramagnetic, for example, charge recombina-
tion from an RP, the electronic coupling that dictates CT from
the RP to energetically proximate electronic states is also that
which facilitates the magnetic interaction between the unpaired
(21) Yamashita, J.; Kondo, J. Phys. ReV. 1958, 109, 730.
(22) Miller, J. S.; Epstein, A. J.; Reiff, W. M. Acc. Chem. Res. 1988, 21, 114.
(23) Feher, O. In Tunneling Conference; Chance, B., Devault, D., Frauenfelder,
H., Marcus, R. A., Schreiffer, J. R., Sutin, N., Eds.; Academic Press: New
York, 1979; pp 729-743.
(24) McConnell, H. M. J. Chem. Phys. 1961, 35, 508.
(25) Closs, G. L.; Piotrowiak, P. J.; MacInnis, J. M.; Fleming, G. R. J. Am.
Chem. Soc. 1988, 110, 2652.
(26) Roest, M. R.; Oliver, A. M.; Paddon-Row: M. N.; Verhoeven, J. W. J.
Phys. Chem. A 1997, 101, 4867.
(27) Paddon-Row: M. N.; Oliver, A. M.; Warman, J. M.; Smit, K. J.; Haas, M.
P.; Oevering, H.; Verhoeven, J. W. J. Phys. Chem. A 1988, 92, 6958.
(28) Nitzan, A. Annu. ReV. Phys. Chem. 2001, 52, 681.
(29) Davis, W. B.; Wasielewski, M. R.; Ratner, M. A.; Mujica, V.; Nitzan, A.
J. Phys. Chem. A 1997, 101, 6158.
(30) Davis, W. B.; Svec, W. A.; Ratner, M. A.; Wasielewski, M. R. Nature
1998, 396, 60.
(13) Marcus, R. A. Chem. Phys. Lett. 1987, 133, 471.
(14) Ogrodnik, A.; Michel-Beyerle, M.-E. Z. Naturforsch. 1989, 44a, 763.
(15) Kilsa, K.; Kajanus, J.; Macpherson, A. N.; Martensson, J.; Albisson, B. J.
Am. Chem. Soc. 2001, 123, 3069.
(16) Lukas, A. S.; Bushard, P. J.; Wasielewski, M. R. J. Phys. Chem. A 2002,
106, 2074-2082.
(17) Kramers, H. A. Physica 1934, 1.
(18) Anderson, P. W. Phys. ReV. 1950, 79, 350.
(19) Anderson, P. W. Phys. ReV. 1959, 115, 2.
(31) Heisenberg, W. Z. Phys. 1926, 38, 411.
(32) Steiner, U. E.; Ulrich, T. Chem. ReV. 1989, 89, 9.
(20) Marcus, R. A. J. Chem. Phys. 1965, 43, 679-701.
9
5578 J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004

Making a Molecular Wire: Charge and Spin Transport
A R T I C L E S
M-1) for nanosecond transient absorption (TA) and between 0.3 and
0.5 at 420 nm for femtosecond TA. Redox potentials for N-phenyl-
phenothiazine,33 PDI,34 and p-phenylene oligomers35 have been obtained
previously. Redox potentials obtained from these studies are listed in
Table S1. Steady-state fluorescence measurements were performed on
1-5 and model compounds 6-10 and quantum yields of 1*PDI
fluorescence were calculated against a Rhodamine 640 standard.
Quantum yields of 6-10 were all unity. The phenothiazine cation
spectrum measured in acetonitrile33 is characterized by a visible
absorption at 510 nm, which unfortunately is undetectable in our TA
experiment because of the strength of the ground-state bleach of PDI
in this spectral region.
Femtosecond transient absorption measurements were made either
using the 420-nm frequency-doubled output from a regeneratively
amplified titanium sapphire laser system operating at 2 kHz or the 400-
nm frequency-doubled output from a regeneratively amplified titanium
sapphire laser system operating at 1 kHz as the excitation pulse. A
white light continuum probe pulse was generated by focusing the IR
fundamental into a 1-mm sapphire disk.36 Detection with a CCD
spectrograph has previously been described.36 The samples were
irradiated with 0.5-1.0 µJ per pulse focused to a 200-µm spot. Samples
were placed in a 2-mm path length glass cuvette and stirred using a
motorized wire stirrer to prevent thermal lensing and sample degrada-
tion. The total instrument response time for the pump-probe experi-
ments was 150 fs. Transient absorption kinetics were fit to a sum of
exponentials with a Gaussian instrument function using Levenberg-
Marquardt least-squares fitting. Femtosecond transient absorption kinetic
traces for compounds 1-3 can be found in Figure S1.
Nanosecond transient absorption measurements were made using the
frequency-tripled output of a Continuum 8000 Nd:YAG laser to pump
a Continuum Panther OPO. Samples were placed in a 10-mm path
length quartz cuvette equipped with a vacuum adapter and subjected
to five freeze-pump-thaw degassing cycles prior to transient absorp-
tion measurements. The probe light in the nanosecond experiment was
generated using a xenon flashlamp (EG&G Electrooptics FX-200) and
detected using a photomultiplier tube with high voltage applied to only
four dynodes (Hamamatsu R928). The total instrument response time
is 7 ns and is determined primarily by the laser pulse duration.
Nanosecond kinetic traces for 2, 4, and 5 were recorded over a range
of 1 µs, while those for 3 were over a range of 2 µs. Between 50 and
100 shots were averaged at each field strength with a LeCroy 9384
digital oscilloscope and sent to a microcomputer, which calculated the
∆A. Nanosecond kinetic traces for compounds 2-5 can be found in
Figure S2. For the magnetic field effect experiment, the sample cuvette
was placed between the poles of a Walker Scientific HV-4W
electromagnet powered by a Walker Magnion HS-735 power supply.
The field strength was measured by a Lakeshore 450 gaussmeter with
a Hall effect probe. Both the electromagnet and the gaussmeter were
interfaced with the data collection computer, allowing measurement
and control of the magnetic field to (1 × 10-5 T during data
acquisition. The magnetic field was varied by a constant increment
(either 0.2 mT, 0.3 mT, 0.5 mT, 1 mT, or 5 mT depending on desired
resolution). Because of the length of the sample runs (>4 h), a small
amount of sample degradation was observed, resulting in a decrease
in the triplet yield at zero field, ∆A(B ) 0), during the experiments.
To compensate for this, the magnetic field was reset to B ) 0 mT
every five kinetic traces for increments of 5 mT, every three kinetic
traces for increments of 1, 0.5, and 0.3 mT, and every two kinetic
traces for increments of 0.2 mT and ∆A(B ) 0) was plotted and fit
with a polynomial or series of polynomials. These functions were used
Figure 2. Schematic of radical ion pair energy levels as a function of
magnetic field.
as well as the magnetic dipole-dipole interaction, are neglected
because the measurements are performed in solution. Also, it
is assumed that nuclei associated structurally with a given radical
couple only with the electron spin within that radical.
Immediately following rapid, nonadiabatic charge separation,
the correlated electron spins are in a singlet configuration. This
pure state is, in general, not an eigenstate of HST, as the weakly
coupled electron spins are free to precess independently around
the resultant of their respective local fields (mainly because of
electron-nuclear hyperfine interactions) and the external applied
field. After times usually in the range of a few nanoseconds,
RP-ISC results in formation of a triplet spin configuration.
When hyperfine and exchange interactions are isotropic and
spin-spin coupling is weak, each of the three zero-field triplet
states of the radical pair will be nearly degenerate with the
singlet and will be populated with equal probability at room
temperature. If the spin-spin exchange interaction within the
radical ion pair is nonzero, the triplet manifold is not initially
degenerate with the singlet, but rather separated from the singlet
by an energy 2J, Figure 2.
Application of a magnetic field results in Zeeman splitting
of the triplet sublevels, which at high fields can be described
by the T0 and T(1 states, Figure 2. In the high field limit,
population of the RP triplet state occurs exclusively by S-T0
mixing, while T-1 and T+1 remain unpopulated. Electron
paramagnetic resonance (EPR) measurements on closely related
compounds have confirmed that the triplet levels of the radical
pair are higher in energy than the singlet state as would be
expected from net antiferromagnetic exchange.9 When the
Zeeman energy from the applied field equals that of the S-T
splitting, the low energy triplet state, T-1, crosses the singlet,
and the RP-ISC rate is maximized, which produces a resonance
in the triplet yield at B2J. An increase in the rate of triplet
formation at resonance implies that the RP decay rate also
increases. One can therefore monitor the RP population as a
function of applied magnetic field and obtain a plot with a
minimum at B2J to obtain 2J as well.
Experimental Section
The synthesis and characterization of compounds 1-10 can be found
in the Supporting Information. Characterization was performed with a
Gemini 300 MHz, Varian 400 MHz, or INOVA 500 MHz NMR and
a PE BioSystems MALDI-TOF mass spectrometer. All solvents were
spectrophotometric grade or distilled prior to use.
Ground-state absorption measurements were made on a Shimadzu
(UV-1601) spectrophotometer. The optical density of all samples was
maintained between 0.7 and 1.0 at 532 nm (ꢀPDI, 550 nm ) 46,000 cm-1
(33) Daub, J.; Engl, R.; Kurzawa, J.; Miller, S. E.; Schneider, S.; Stockmann,
A.; Wasielewski, M. R. J. Phys. Chem. A 2001, 105, 5655.
(34) Gosztola, D.; Niemczyk, M. P.; Svec, W. B.; Lukas, A. S.; Wasielewski,
M. R. J. Phys. Chem. A 2000, 104, 6545.
(35) Meerholz, K.; Heinze, J. Electrochim. Acta 1996, 41, 1839.
(36) Giaimo, J. M.; Gusev, A. V.; Wasielewski, M. R. J. Am. Chem. Soc. 2002,
124, 8530.
9
J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004 5579

A R T I C L E S
Weiss et al.
to calculate the relative triplet yield or RP yield as a function of applied
field strength. The relative triplet yield is thus
∆A(B)
∆A(B ) 0)
T
T0
)
with an analogous expression for the RP yield. The results presented
are an average of three or more experiments conducted on separate
days with freshly prepared samples in spectrophotometric or freshly
distilled ACS grade toluene.
Results and Discussion
Synthesis of PTZ-Phn-PDI. The synthesis of each mol-
ecule followed a strategy in which the phenothiazine (PTZ)
donor or the perylenediimide (PDI) acceptor were functionalized
with a segment of the oligophenyl bridge followed by coupling
of the donor- and acceptor-bearing segments. This approach
avoids the solubility issues that are typically associated with
(Phn) oligomers. The known 1,7-dibromoperylene-3,4:9,10-
tetracarboxydianhydride was reacted with 3,5-di-tert-butylphenol
(3 equiv) in refluxing DMF in the presence of Cs2CO3 (2 equiv)
for 4 h. Precipitation from cold glacial acetic acid gave 1,7-
diphenoxyperylene-3,4:9,10-tetracarboxydianhydride (PDA). The
dianhydride is converted to the monoanhydride/monoimide by
reacting PDA with 2-ethylhexylamine in refluxing pyridine for
2 h, followed by chromatographic separation of the monoimide
(PIA) from the diimide and unreacted dianhydride, the latter
two of which can be recycled.
The synthesis of PTZ-Phn-PDI where n ) 1 and 2 begins
by coupling 4-nitrobromobenzene or 4′-nitrobromobiphenyl to
phenothiazine using Pd-catalyzed amination of the bromoarene.
This yields the PTZ-Phn-NO2 (n ) 1, 2) intermediates, both
of which were subjected to SnCl2/HCl reduction to afford the
corresponding PTZ-Phn-NH2 (n ) 1, 2) intermediates.
Finally, condensation of the amines with PIA in refluxing
pyridine/imidazole produced 1 and 2. The synthesis of PTZ-
Ph3-PDI, begins with the Pd-catalyzed amination of 4′,4-
dibromobiphenyl with phenothiazine using a 4-fold excess of
the dibrominated starting material to avoid disubstitution. This
reaction could be scaled up easily to produce large quantities
of the PTZ-Ph2-Br intermediate, which proved to be useful
for later reactions. Suzuki coupling of this intermediate with
4-aminophenylboronic ester gave PTZ-Ph3-NH2. Condensa-
tion of PTZ-Ph3-NH2 with PIA in refluxing DMF (with a
catalytic amount of ZnOAc2) produced 3. The synthesis of
PTZ-Ph4-PDI starts by reacting an excess of 4,4′-dibromo-
biphenyl with 4-aminophenylboronic ester to give NH2-Ph3-
Br. This was readily converted to the boronic ester using Pd-
catalyzed cross-coupling with pinacolborane to give NH2-Ph3-
DB. Because of solubility limitations, it was necessary to couple
the fourth phenyl to phenothiazine first. 4-Trimethylsilyl-
bromobenzene was aminated with phenothiazine using Pd-
catalyzed cross-coupling followed by removal of the TMS with
ICl to yield PTZ-Ph-I. Suzuki coupling of this material with
the NH2-Ph3-DB intermediate produced the PTZ-Ph4-NH2,
which was condensed with PIA in refluxing DMF (catalytic
ZnOAc2) to yield 4. PTZ-Ph5-PDI also required synthesis
of the oligophenylene bridge from both “directions”. First,
NH2-Ph3-Br was condensed with PIA by refluxing in pyridine
to give PDI-Ph3-Br. The intermediate PTZ-Ph2-Br was
easily converted to the boronic ester by Pd-catalyzed cross-
Figure 3. Energy levels for the electronic states relevant to the electron-
transfer pathways for 1-5.
coupling with pinacolborane in basic p-dioxane to give PTZ-
Ph2-DB. Final Suzuki coupling of PDI-Ph3-Br with PTZ-
Ph2-DB gave 5. All synthetic details and characterization
for compounds 1-10 can be found in the Supporting Informa-
tion.
Molecular Structures and Energy Levels. Energy-mini-
mized structures of 1-5, Figure 1, and their corresponding
radical ions were calculated using density functional theory
(DFT) employing Becke’s three parameter hybrid functional
using Lee, Yang, and Parr correlation functional (B3LYP)37,38
and a 6-31G** basis set (please see Supporting Information for
details).39 The “pucker angle” of phenothiazine, calculated to
be 147.70°, is in very good agreement with the value obtained
from the X-ray structure, 146.11°.33 PTZ and PDI are nearly
perpendicular to their respective nearest-neighbor bridge units
and the “twist” angle between bridge units is ∼35°. The effective
distances between PTZ+• and PDI-• within PTZ+•-Phn-PDI-•,
as well as those between Phn+• and PDI-• within PTZ-Phn
-
+•
PDI-•, Table S2, are measured from the centroid of the unpaired
+•
spin distributions of PDI-• to those of PTZ+• and Phn
,
respectively. The spin distribution of each radical ion was
obtained using the energy-minimized geometry of the radical
ion, then subtracting the â spin density from the R spin density
on the diagonal of the calculated spin density matrixes given
by unrestricted Hartree-Fock molecular orbital calculations
using the AM1 model.40
The energy levels for the low-lying PTZ+•-Phn-PDI-• and
PTZ-Phn+•-PDI-• radical ion pair states within 1-5 are shown
in Figure 3. The corresponding energies for PTZ+•-Phn
-
-•
PDI are all very high (g3.0 eV) and consequently will not be
considered explicitly for reasons presented below. These ener-
gies were estimated using Weller’s expression on the basis of
the Born dielectric continuum model41 of the solvent to
determine the energy of formation of an ion pair in a solvent of
arbitrary polarity:
e2
DAꢀs
1
1
1
1
EIP ) Eox - Ered
-
+ e2
+
-
(5)
(
)(
)
r
2r1 2r2 ꢀs ꢀsp
(37) Becke, A. D. J. Chem. Phys. 1993, 98, 1372.
(38) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
(39) Jaguar 3.5; Schrodinger, Inc.: Portland, Oregon, 1998.
(40) The semiempirical PM3 method implemented in HyperChem(TM) (Hy-
percube, Inc., 1115 NW 1114th Street, Gainesville, FL 32601).
(41) Weller, A. Z. Phys. Chem. 1982, 133, 93.
9
5580 J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004

Making a Molecular Wire: Charge and Spin Transport
A R T I C L E S
Figure 4. (A) Normalized ground-state absorption spectra of 1-5. The
fluorescence maxima all occur at 575 nm and the fluorescence quantum
yields for 1-5 are, respectively, 0.03, 0.09, 0.41, 0.83, 0.89; (B) Transient
absorption spectra of 1 following a 400 nm, 130 fs laser excitation pulse.
The features in these spectra are indicative of the charge-separated state of
Figure 5. (A) Logarithmic plot of the charge separation rate constant, kCS
vs donor-acceptor distance, rDA. The best fit line through the data points
for n ) 1-3 gives R2 ) 0.98, â ) 0.46 Å-1, and k0 ) 5 × 1012 s-1. (B)
Logarithmic plot of the charge recombination rate constant, kCR vs donor-
acceptor distance, rDA. The best fit line through the data points for n )
the system: PTZ+•-Ph-PDI-•
.
1-3 gives R2 ) 0.99, â ) 0.67 Å-1, and k0 ) 5 × 1012 s-1
.
where Eox and Ered are, respectively, the oxidation and reduction
potentials of the donor and acceptor in a high polarity solvent
with dielectric constant ꢀsp (ꢀsp ) 25 for butyronitrile used here),
e is the charge of the electron, r1 and r2 are the ionic radii of
the radical ions, rDA is the donor-acceptor distance, and ꢀs is
the static dielectric constant of the solvent in which the
spectroscopy is performed (ꢀs ) 2.38 for toluene used here).
Because of the well-known limitations of this treatment, the
values of EIP are accurate to no better than (0.1 eV. All
information used to obtain EIP is given in the Supporting
Information.
centered at 720 nm which overlaps the PDI-• feature, the decay
of the stimulated emission band is the more accurate measure
of charge separation rate constant within the PTZ-Phn-PDI
systems. Since the spectroscopic signature of PTZ+• cannot be
observed because of the strong bleach of the PDI ground state,
one must consider the possibility of charge separation resulting
directly in PTZ-Phn+•-PDI-• rather than PTZ+•-Phn-PDI-•.
Model compounds Phn-PDI, where n ) 1-5, were synthesized
1
to test this idea. The fluorescence quantum yield of *PDI in
each molecule was unity, in contrast to the PTZ-Phn-PDI
systems. Therefore, fluorescence of PDI is only quenched in
the presence of the PTZ donor. Magnetic field effect measure-
ments on triplet formation via the RP-ISC mechanism in PTZ-
Phn-PDI discussed below also strongly support this conclusion.
Charge-Transport Dynamics. The distance dependence of
the charge separation rate constants for 1-5 is monitored by
the decay of the 620 nm stimulated emission feature of 1*PDI,
Figure 5A. For 1-4 these rate constants, kCS, decay exponen-
tially with donor-acceptor distance so that
Spectroscopy. The ground-state spectra of compounds 1-5,
Figure 4A, have an absorption maximum at 550 nm, which is
very similar to that of PDI alone (ꢀ ) 46 000 M-1 cm-1).5
Additionally, there is a feature between 300 and 350 nm because
of Phn that red shifts and grows in intensity relative to the PDI
peaks as n increases. Spectroelectrochemical measurements
show that the ground-state absorption spectrum of PDI-• is
characterized by a strong absorption at 720 nm (ꢀ ) 79 800
M-1 cm-1).5 The transient absorption spectra of 1-5 following
photoexcitation of PDI with 400 nm, 130 fs laser pulses display
very similar features, having a broad absorption centered at
715-725 nm, a strong bleach of the ground state of PDI at 550
nm, and a band centered at 620 nm because of the stimulated
emission from the lowest excited singlet state of PDI, (1*PDI),
kCS ) k0e-â(r-r )
(6)
0
where k0 is the rate constant at the contact distance r0, and â )
0.46 Å-1. This strongly exponential dependence indicates that
electron transfer from PTZ to 1*PDI is dominated by the super-
1
Figure 4B. As *PDI has a broad, positive absorption band
9
J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004 5581

A R T I C L E S
Weiss et al.
exchange mechanism. Studies of the rates of energy transfer
from Ru(bpy)32+ to Os(bpy)32+ through p-oligophenylenes via
a Dexter-type (superexchange) mechanism show a similar
monitored at 480 nm, occurred with nearly the same rate
constant as PDI-• decay in compounds 2-4. In 5, the rise of
the triplet is obscured by a strong negative feature due to 1*PDI
fluorescence, which makes the kinetic trace difficult to fit. Third,
the RP yield of compounds 3-5, (see Figure 6B for data on 4)
reaches a minimum at B ) B2J, where RP-ISC is most efficient.
Therefore, the RP is decaying most efficiently when the most
triplet RP is being produced, indicating that the triplet recom-
bination pathway is significantly faster than the singlet pathway.
The reason for the dominance of the triplet pathway is that the
free energies for singlet charge recombination in 1-5, where
-∆GCR ) 2.0-2.3 eV, are much deeper within the Marcus
inverted region of the rate versus free energy of reaction
profile,20 that is, |∆G| > λ, where λ is the total reorganization
energy for charge recombination (∼0.6 eV for these systems,
please see the Supporting Information), than the corresponding
free energies for triplet charge recombination, where -∆GCR
) 0.8-1.1 eV.
exponential distance dependence with â ) 0.32 Å-1 42
The
.
energies of the electronic states with positive charge residing
on the bridge in these molecules, PTZ-Phn+•-PDI-•, are high
enough so that these states remain unpopulated virtual states.
The data in Figure 3 show that only in 4 and 5 is the energy
gap between the initial photoexcited state and the bridge RP
small. The rate constant for 5 clearly does not fall on the line
in Figure 5A, so that electron transfer through Ph5 may be on
the threshold of a change in electron-transfer mechanism to
incoherent hopping.
The rate constants for charge recombination in 1-5 are
obtained from the decay rate of PDI-•, Figure 5B, kinetic traces
in Figures S1 and S2. Charge recombination within compounds
1-3 is strongly exponential, and the plot of log(kCR) versus
rDA yields â ) 0.67 Å-1. The rate constants for charge
recombination in 4 and 5, however, do not lie along the linear
curve established by 1-3, and, in fact, actually increase with
increasing bridge length. For charge recombination within these
compounds, a switch in mechanism from superexchange to
thermally activated hopping is postulated as the bridge is
lengthened.
Magnetic Field Effects. Figure 6A, B shows a plot of 3*PDI
yield following RP recombination within 2 (A) and 3 (B) as a
function of applied magnetic field, while Figure 6C, D shows
a similar plot of the RP yield for 4 (C) and 5 (D). The maximum
of the triplet yield plot (and minimum of the RP yield plot)
marks the field at which the S and T-1 levels of the RP cross,
resulting in maximum triplet production. This field then
corresponds to the energy 2J, the zero-field splitting of the
singlet and triplet RP levels because of magnetic superexchange
coupling. Figure 7 shows a clear exponential decrease of 2J
with rDA for 2-5, providing strong evidence that the charges
recombining through the RP-ISC mechanism are localized on
the donor and acceptor and not somewhere on the bridge. The
value of 2J cannot be measured for compound 1 because the
RP does not live long enough for RP-ISC to produce adequate
yields of 3*PDI to observe.
The singlet and triplet RP states are either stabilized or
destabilized through one- and two-step virtual charge transfers
via coupling of the orbitals on the paramagnetic centers to the
bridge orbitals and to each other. The total perturbation to each
RP state, ∆ES or ∆ET for the singlet and triplet, respectively, is
a sum of pairwise interactions between the RP state and the
state to which it couples via charge transfer. By simple second-
order perturbation theory,
A change in charge recombination mechanism from super-
exchange to hopping relies on efficient charge injection into
+
the Phn bridge such that the PTZ+•-Phn-•-PDI or PTZ-Phn
-
PDI-• states are real intermediates. The energies of PTZ+•-
Phn-•-PDI for n ) 1-5 all are g3.0 eV, so that electron
injection onto the bridge during the charge recombination
reaction cannot occur. On the other hand, as the bridge
lengthens, the calculated energy of PTZ-Phn+•-PDI-• decreases
significantly because of the increased conjugation length, as can
be seen in the decrease in oligophenylene band gap energy from
7.3 eV for n ) 1 to 3.3 eV for n ) 5, as calculated with time-
dependent DFT (B3LYP hybrid functional, 6-31G** basis set)
using the Q-Chem 2.1 software package.43 The accompanying
increased ease of oxidation of the oligo-p-phenylene, coupled
with the increase in the energy of PTZ+•-Phn-PDI-• largely
because of Coulomb destabilization, Figure 3, leads to a near
resonance of PTZ-Phn+•-PDI-• with PTZ+•-Phn-PDI-• for
n ) 4 and 5. The slight increase in charge recombination rate
observed for 4 and 5, relative to 3, may be due to the
combination of the increasing electronic interaction between
PTZ+•-Phn-PDI-• and PTZ-Phn+•-PDI-• as the energy gap
between them becomes smaller and the decreasing internal
reorganization energy associated with putting a charge on a
longer bridge.
2J ) ∆ES - ∆ET )
2
2
| ΨRP|VRP-n|Ψn |
ERP - En - λ
| ΨRP|VRP-n|Ψn |
∑
ERP - En - λ
n
-
(7)
∑
[
]
[
]
n
S
T
where the indicated matrix elements couple the singlet and triplet
RP states to states n, ERP and En are energies of these states,
respectively, and λ is the total nuclear reorganization energy of
the charge-transfer reaction.19 In principle, eq 7 can be used to
determine the matrix elements for these interactions given that
2J has been determined experimentally for 2-5, En can be
obtained from the optical absorption and emission spectra of
the excited states, the Weller equation can be used to estimate
The RP states, PTZ+•-Phn-PDI-•, may of course recombine
either from the singlet or triplet RP state, but several observa-
tions suggest that the rate of charge recombination from the
triplet RP, 3(PTZ+•-Phn-PDI-•) f PTZ-Phn-3*PDI, is much
faster than the corresponding rate from the singlet RP to the
singlet ground state, 1(PTZ+•-Phn-PDI-•) f PTZ-Phn-PDI,
Figure 3. First, the recombination rates of compounds 2-5 are
markedly single-exponential despite the fact that there are two
possible recombination pathways. Second, 3*PDI formation,
E
RP, and the dielectric continuum model of solvation and
electronic structure methods can be used to calculate λ.
However, not all contributing states are observable spectroscopi-
cally, and the errors involved in estimating the energy denomi-
nators in eq 7 are far greater than the measured values of 2J,
(42) Schlicke, B.; Belser, P.; De Cola, L.; Sabbioni, E.; Balzani, V. J. Am. Chem.
Soc. 1999, 121, 4207-4214.
(43) Q-Chem 2.1; Q.-C. I.: Pittsburgh, PA, 1998.
9
5582 J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004

Making a Molecular Wire: Charge and Spin Transport
A R T I C L E S
Figure 6. (A) Plot of the relative yield of 3*PDI vs magnetic field strength for PTZ-Ph2-PDI (2J ) 170 mT, A) and PTZ-Ph3-PDI (2J ) 31 mT, (B)
plot of the relative radical ion pair yield vs magnetic field strength for PTZ-Ph4-PDI (2J ) 6.4 mT, C) and PTZ-Ph5-PDI (2J ) 1.5 mT, D).
(σ-σ, σ-π, π-π) may be contributing to the S-T splitting,
but 2J clearly depends exponentially on rDA within this series
(â ) 0.37 Å-1). Previous work in calculating 2J has shown44
that assigning different distance dependencies to different CT
pathways in hopes of distinguishing the respective orbitals
involved does not improve the fit to experimental data over using
a single exponential. Although we may not now be able to
deconvolute the separate contributions of various CT pathways
to the total S-T splitting, the 2J measurement is, in itself,
directly related to donor-acceptor coupling, which is also
inherently a sum of many electron-transfer processes. The
exponential decrease of 2J with rDA parallels the approximately
exponential decay of the coherent contribution to the various
ET processes. An increase in CR rate for 4 and 5 must therefore
Figure 7. Logarithmic plot of the spin-spin exchange interaction, 2J vs
r
DA. The best fit line through the data points for n ) 2-5 gives R2 ) 0.99
be due to the dominance of an incoherent CT channel, namely,
a hopping process. As stated previously, however, the magnitude
of 2J depends only on contributions from virtual excitations
from the charge-separated state, so the path that the electron
takes to get from donor to acceptor (whether it be real
occupation of the bridge or superexchange) is actually irrelevant
in determining 2J. It is true that, as the bridge is lengthened,
the changing energy level configuration promotes increased
mixing with the bridge-occupied state, but at the cost of mixing
with other configurations, and thus, as the distance between the
and a slope of -0.37 Å-1
.
which are about 10-5-10-7 eV. In addition, for singlet-triplet
splittings of such small magnitude, there may be significant
contributions from direct exchange terms that we are not
considering here.21,22,44 This is clearly a case where experiment
is far more accurate than quantitative theory at this time.
As in studies concerning couplings between paramagnetic
metal centers45-47 and biradicals,48-51 multiple CT pathways
(44) Weihe, H.; Guedel, H. U. J. Am. Chem. Soc. 1997, 119, 6539.
(45) Pardo, E.; Faus, J.; Julve, M.; Lloret, F.; Munoz, M. C.; Cano, J.;
Ottenwaelder, X.; Journaux, Y.; Carrasco, R.; Blay, G.; Fernandez, I.; Ruiz-
Garcia, R. J. Am. Chem. Soc. 2003, 125, 10770-10771.
(48) Closs, G. L.; Forbes, M. D. E.; Piotrowiak, P. J. J. Am. Chem. Soc. 1992,
114, 3285.
(49) Staerk, H.; Kuehnle, W.; Treichel, R.; Weller, A. Chem. Phys. Lett. 1985,
118, 19-24.
(46) Brunold, T. C.; Gamelin, D. R.; Solomon, E. I. J. Am. Chem. Soc. 2000,
122, 8511-8523.
(50) Staerk, H.; Busmann, H.-G.; Kuhnle, W.; Treichel, R. J. Phys. Chem. 1991,
95, 1906.
(47) McCleverty, J. A.; Ward, M. D. Acc. Chem. Res. 1998, 31, 842-851.
(51) Shultz, D. A. Comments Inorg. Chem. 2002, 23, 1-21.
9
J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004 5583

A R T I C L E S
Weiss et al.
two spins decreases, so does the overall magnitude of their
magnetic interaction.
first time the relative contributions of both the coherent
superexchange and incoherent hopping mechanisms to the
overall charge transport process in a conjugated bridge molecule.
Conclusions
Acknowledgment. M.R.W. acknowledges support by the
Division of Chemical Sciences, Office of Basic Energy Sciences,
U.S. Department of Energy under grant no. DE-FG02-99ER-
14999. M.A.R. acknowledges support from MRSEC sponsored
at Northwestern by the National Science Foundation, and the
MURI/DURINT program of the DOD. E.A.W. acknowledges
a Fellowship from the Link Energy Foundation. We thank A.
Burin, Igor Kurnikov, and A. Troisi for useful discussions.
The oligomeric p-phenylene bridge acts as a molecular wire
for the charge recombination reaction of PTZ+•-Phn-PDI-•
when n g 4. Charge recombination can occur via two distinct
pathways involving RP intermediates with singlet and triplet
spin configurations. Direct measurement of the singlet-triplet
splitting, 2J, within the RP, using the magnetic field dependence
of the RP and triplet product yields, provides quantitative
evidence that the magnetic superexchange coupling, and there-
fore the probability of a coherent electron transfer mechanism,
decreases approximately exponentially as the bridge lengthens.
Concurrently, the near resonant interaction of PTZ+•-Phn-
PDI-• and PTZ-Phn+•-PDI-•, when n ) 4 and 5, makes
incoherent hopping a viable charge-transfer mechanism. These
two effects combine to ensure the dominance of wirelike charge
transport for 4 and 5. Our results allow us to observe for the
Supporting Information Available: Experimental details
regarding the synthesis and characterization of molecules used
in this study, as well as additional transient kinetic and
electrochemical data (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
JA0398215
9
5584 J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004