Conformational gating of long distance electron transfer through wire-like bridges in donor-bridge-acceptor molecules

Article abstract of DOI:10.1021/ja010330z

A series of five donor-bridge-acceptor (DBA) molecules in which the donor is tetracene, the acceptor is pyromellitimide, and the bridge molecules are oligo-p-phenylenevinylenes (OPV) of increasing length has been shown to undergo electron transfer (ET) by means of two mechanisms. When the bridge is short, strongly distance dependent superexchange dynamics dominates, whereas when the bridge is longer, bridge-assisted hopping dynamics prevails. The latter mechanism results in relatively soft distance dependence for ET in which the OPV oligomers act effectively as molecular wires. We now report studies on the critical influence that bridge dynamics have on electron transfer through these oligomers. The temperature dependence of the charge separation (CS) rates in all five molecules does not appear to obey the predictions of standard ET theories based upon the Condon approximation. All five molecules show behavior consistent with CS being "gated" by torsional motion between the tetracene donor and the first bridge phenyl ring. This is based on the near equivalence of the CS activation energies measured for all five molecules with the frequency of a known vibrational mode in 5-phenyltetracene. In the molecule containing a trans-stilbene bridge, a competition occurs between the tetracene-phenyl torsional motion and one that occurs between the vinyl group and the phenyls linked to it. This results in complex temperature-dependent CS that exhibits both activated and negatively activated regimes. The charge recombination (CR) reactions within the molecules which have the two shortest bridges, namely phenyl and trans-stilbene, show a weaker dependence on these molecular motions. The three molecules with the longest bridges all display complex temperature dependencies in both their rates of CS and CR, most likely because of the complex torsional motions, which arise from the multiple phenyl-vinyl linkages. The data show that long-distance electron transfer and therefore wire-like behavior within conjugated bridge molecules depend critically on these low-frequency torsional motions. Molecular device designs that utilize such bridges will need to address these issues.

Full text of DOI:10.1021/ja010330z

J. Am. Chem. Soc. 2001, 123, 7877-7886
7877
Conformational Gating of Long Distance Electron Transfer through
Wire-like Bridges in Donor-Bridge-Acceptor Molecules
William B. Davis, Mark A. Ratner, and Michael R. Wasielewski*
Contribution from the Department of Chemistry and Center for Nanofabrication and
Molecular Self-Assembly, Northwestern UniVersity, EVanston, Illinois 60208-3113
ReceiVed February 6, 2001
Abstract: A series of five donor-bridge-acceptor (DBA) molecules in which the donor is tetracene, the
acceptor is pyromellitimide, and the bridge molecules are oligo-p-phenylenevinylenes (OPV) of increasing
length has been shown to undergo electron transfer (ET) by means of two mechanisms. When the bridge is
short, strongly distance dependent superexchange dynamics dominates, whereas when the bridge is longer,
bridge-assisted hopping dynamics prevails. The latter mechanism results in relatively soft distance dependence
for ET in which the OPV oligomers act effectively as molecular wires. We now report studies on the critical
influence that bridge dynamics have on electron transfer through these oligomers. The temperature dependence
of the charge separation (CS) rates in all five molecules does not appear to obey the predictions of standard
ET theories based upon the Condon approximation. All five molecules show behavior consistent with CS
being “gated” by torsional motion between the tetracene donor and the first bridge phenyl ring. This is based
on the near equivalence of the CS activation energies measured for all five molecules with the frequency of
a known vibrational mode in 5-phenyltetracene. In the molecule containing a trans-stilbene bridge, a competition
occurs between the tetracene-phenyl torsional motion and one that occurs between the vinyl group and the
phenyls linked to it. This results in complex temperature-dependent CS that exhibits both activated and negatively
activated regimes. The charge recombination (CR) reactions within the molecules which have the two shortest
bridges, namely phenyl and trans-stilbene, show a weaker dependence on these molecular motions. The three
molecules with the longest bridges all display complex temperature dependencies in both their rates of CS and
CR, most likely because of the complex torsional motions, which arise from the multiple phenyl-vinyl linkages.
The data show that long-distance electron transfer and therefore wire-like behavior within conjugated bridge
molecules depend critically on these low-frequency torsional motions. Molecular device designs that utilize
such bridges will need to address these issues.
Introduction
Sn
∞
2π
1
2
kET
)
|HDA
|
exp[-S]
∑
Progress toward a fundamental understanding of electron
transfer (ET) reactions has come from an intricate tapestry of
theoretical tools and techniques, new methods in synthetic
chemistry, and the development of spectroscopic apparatus
capable of spanning the time range from femtoseconds to hours.
Early theoretical treatments of electron transfer come from the
work of Marcus1,2 and Hush,3,4 which itself is an extension of
transition state theory. Later theories, such as those developed
by Jortner5 and Fischer and Van Duyne6 (henceforth referred
to as semiclassical ET theory), incorporate the quantum me-
chanical nature of high-frequency solvent and intramolecular
vibrational modes that are coupled to the ET reaction coordinate.
From the work of several groups,5-8 the rate of thermal,
nonadiabatic ET between a donor and acceptor assuming a single
high-frequency vibration can be written as
p
4πλk T
n!
x
n)0
B
-(∆G° + λ + npω)2
exp
(1)
[
]
4λkBT
In eq 1, kB is the Boltzmann constant, T is temperature, ∆G° is
the free energy change of the electron-transfer reaction, ω is
the frequency of the coupled quantum mechanical vibrational
mode, and λ is the reorganization energy for all classical modes.
In general, λ will contain contributions from both solvent
polarization, λs, and low-frequency intramolecular vibrations,
Low. S is the Huang-Rhys factor, mathematically expressed
as S ) λi/pω, where λi is the reorganization energy for the
quantum mechanical vibrational mode.
λ
When both electron tunneling and vibrational motion occur
on the same time scale, a breakdown of the Condon approxima-
tion can occur, which implies that HDA within eq 1 becomes
dependent upon the nuclear coordinates. Under these circum-
stances, the electronic motion can be completely controlled, or
“gated”,9-11 by the nuclear motions within the electron donor-
(1) Marcus, R. A. J. Chem. Phys. 1956, 24, 966-978.
(2) Marcus, R. A. Annu. ReV. Phys. Chem. 1964, 15, 155-196.
(3) Hush, N. S. J. Chem. Phys. 1958, 28, 962-972.
(4) Hush, N. S. Trans. Faraday Soc. 1961, 57, 577-580.
(5) Jortner, J. J. Chem. Phys. 1976, 64, 4860-4867.
(6) Fischer, S. F.; van Duyne, R. P. Chem. Phys. 1977, 26, 9-16.
(7) DeVault, D. Quantum Mechanical Tunneling in Biological Systems;
Cambridge University Press: Cambridge, 1984.
(9) Hoffman, B. M.; Ratner, M. A. J. Am. Chem. Soc. 1987, 109, 6237-
6243.
(8) Hopfield, J. J. Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 3640-3644.
10.1021/ja010330z CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/19/2001

7878 J. Am. Chem. Soc., Vol. 123, No. 32, 2001
DaVis et al.
N-1
acceptor molecule. Expressions describing the temperature and
free energy dependencies of kET must be reevaluated under these
circumstances.12 Equation 1, which is derived for the high-
temperature limit, nevertheless predicts an Arrhenius-like tem-
perature dependence for kET. In particular, when reactions occur
at high temperature in the Marcus normal region, where the
driving force is smaller than the total reorganization energy,
Marcus theory is often adequate.13-15 However, for reactions
occurring in the Marcus inverted region, the influence of high-
frequency vibrational modes5,6,16,17 can lead to ET rate constants
with very weak temperature dependence, or even complete
temperature independence.7,15,18-20 Non-Arrhenius temperature
dependence can also be found in ET systems in which slow
solvent relaxation controls the ET reaction.21-29 This is espe-
cially pronounced in glass forming organic solvents and proteins,
where the temperature dependence of the ET rate constants will
change below the glass transition temperature of the medium.30-36
V
DBVBA VBB
HDA
)
(2)
(
)
ωDB ωDB
In eq 2, VDB and VBA are the electronic coupling strengths of
the donor and bridge and bridge and acceptor, respectively, VBB
is the electronic coupling between bridge sites, ωDB is the energy
gap between the relevant states of the donor and bridge, and N
is the number of identical bridge sites. Equation 2 is approximate
and does not take into account nonnearest neighbor interactions
and multiple pathways.42 Torsional motions will primarily affect
the electronic coupling between adjacent electronic sites, and
will impart an angular dependence to Vij (i,j nearest neighbors)
of the form
Vij ) Vij0 cos(Φ)
(3)
where Φ is the torsional angle and Vij0 is the electronic coupling
between sites i and j at Φ ) 0°. If Φ is temperature dependent,
this will impart a temperature dependence to HDA that may
override the normal Arrhenius-like temperature dependence of
kET expected from semiclassical ET theory. For example, the
Condon approximation will fail when conformations with small
Boltzmann populations have anomalously large HDA values.
The above discussion of the temperature dependence of HDA
assumes that the ET system under consideration falls into a
superexchange regime. In particular, for eq 2 to hold, the
requirements of second-order perturbation theory must be met
by a particular DBA molecule. Several theoretical treatments
have now been developed to describe DBA systems in which
ET does not occur via the superexchange mechanism.45-50 In
particular, the most important situation in which the superex-
change mechanism breaks down occurs when the energy gap
between donor and bridge (or acceptor and bridge) is of the
same order of magnitude as the electronic coupling between
the donor and the bridge, VDB. When the energy gap separating
electron donor and bridge becomes small, the rate-limiting step
of the ET reaction becomes thermal promotion of the electron
from the donor to the bridge. This promotion occurs primarily
through the coupled vibrational modes, and the observed
activation energy is approximately the energy gap separating
donor and bridge. The implication of this result is that the bridge
becomes a real intermediate state with electronic population
residing on it for finite times. This contrasts with superexchange,
where the bridge serves as a tunneling barrier separating the
donor and acceptor, and electronic population does not usually
reside on it for any reasonable time. In this paper, we describe
the temperature dependence of both photoinduced charge
separation (CS) and thermal charge recombination (CR) in the
DBA series, 1-5,51 which undergoes a change in electron-
transfer mechanism from superexchange to bridge localization
Of all the vibrational modes available to conjugated organic
donor-bridge-acceptor (DBA) molecules, those associated with
donor-bridge and bridge-acceptor torsional motions are
expected to contribute the most to variation of HDA. In addition,
any other torsional motions within D, B, or A will also
contribute. From the McConnell analysis of electronic coupling
in organic radical exchange reactions,37 HDA can be ap-
proximated in the following form38-44
(10) Bechtold, R.; Kuehn, C.; Lepre, C.; Isied, S. S. Nature 1986, 322,
286-288.
(11) Brunschwig, B. S.; Sutin, N. J. Am. Chem. Soc. 1989, 111, 7454-
7465.
(12) Beratan, D. N.; Hopfield, J. J. J. Chem. Phys. 1984, 81, 5753-
5759.
(13) Kroon, J.; Oevering, H.; Verhoeven, J. W.; Warman, J. M.; Oliver,
A. M.; Paddon-Row: M. N. J. Phys. Chem. 1993, 97, 5065-5069.
(14) O’Driscoll, E.; Simon, J. D.; Peters, K. S. J. Am. Chem. Soc. 1990,
112.
(15) Liang, N.; Miller, J. R.; Closs, G. L. J. Am. Chem. Soc. 1989, 111,
8740-8741.
(16) Bixon, M.; Jortner, J. J. Phys. Chem. 1991, 95, 1941-1944.
(17) Kestner, N. R. J. Phys. Chem. 1980, 84, 1270-1275.
(18) Rehm, D.; Weller, A. Isr. J. Chem. 1970, 8, 259-271.
(19) Gunner, M. R.; Dutton, P. L. J. Phys. Chem. 1989, 93, 3400-3412.
(20) Smit, K. J.; Warman, J. M.; de Haas, M. P.; Paddon-Row, M. N.
Chem. Phys. Lett. 1988, 152, 177-182.
(21) Zusman, L. D. Chem. Phys. 1980, 49, 295-304.
(22) McGuire, M.; McLendon, G. J. Phys. Chem. 1986, 90, 2549-2551.
(23) Sumi, H.; Marcus, R. A. J. Chem. Phys. 1986, 84.
(24) Rips, I.; Jortner, J. Chem. Phys. Lett. 1987, 133, 411-414.
(25) Akesson, E.; Johnson, A. E.; Levinger, N. E.; Walker, G. C.; de
Bruil, T. P.; Barbara, P. F. J. Chem. Phys. 1992, 96, 7859-7862.
(26) Grote, R. F.; Hynes, J. T. J. Chem. Phys. 1981, 74, 4465-4475.
(27) Heitele, H.; Michel-Beyerle, M. E.; Finckh, P. Chem. Phys. Lett.
1987, 138, 237-243.
(28) Rempel, U.; von Maltzan, B.; von Borczyskowski, C. Z. Phys. Chem.
1991, 170, 107-116.
(29) Wiederrecht, G. P.; Svec, W. A.; Wasielewski, M. R. J. Phys. Chem.
B 1999, 103, 1386-1389.
(30) Harrison, R. J.; Pearce, B.; Beddard, G. S.; Cowan, J. A.; Sanders,
K. M. Chem. Phys. 1987, 116, 429-448.
(31) Chen, P.; Danielson, E.; Meyer, T. J. J. Phys. Chem. 1988, 92,
3708-3711.
(42) Beratan, D. N.; Betts, J.; Onuchic, J. N. Science 1991, 252, 1285-
1288.
(32) Gaines, G. L.; O’Neill, M. P.; Svec, W. A.; Niemczyk, M. P.;
Wasielewski, M. R. J. Am. Chem. Soc. 1991, 113, 719-721.
(33) Harriman, A.; Heitz, V.; Ebersole, M.; van Willigen, H. J. Phys.
Chem. 1994, 98, 4982-4989.
(43) Hsu, C. P.; Marcus, R. A. J. Chem. Phys. 1997, 106, 584-598.
(44) Onipko, A. Chem. Phys. Lett. 1998, 292, 267-272.
(45) Skourtis, S.; Mukamel, S. Chem. Phys. 1995, 197, 367-388.
(46) Davis, W. B.; Wasielewski, M. R.; Ratner, M. A.; Mujica, V.;
Nitzan, A. J. Phys. Chem. A 1997, 101, 6158-6164.
(47) Segal, D.; Nitzan, A.; Davis, W. B.; Wasielewski, M. R.; Ratner,
M. A. J. Phys. Chem. B 2000, 104, 3817-3829.
(48) Okada, A.; Chernyak, V.; Mukamel, S. J. Phys. Chem. A 1998, 102,
1241-1251.
(34) Hoffman, B. M.; Natan, M. J.; Nocek, J. M.; Wallin, S. A. Struct.
Bonding 1991, 75, 85-108.
(35) Hoffman, B. M.; Ratner, M. A. Inorg. Chim. Acta 1996, 243, 233-
278.
(36) Agmon, N.; Rabinovich, S. J. Chem. Phys. 1992, 97, 7270-7286.
(37) McConnell, H. M. J. Chem. Phys. 1961, 35, 508-515.
(38) Ratner, M. A. J. Phys. Chem. 1990, 94, 4877-4883.
(39) Evenson, J. W.; Karplus, M. Science 1993, 262, 1247-1249.
(40) Meade, T. J. Metal Ions in Biological Systems; Sigel, A., Sigel, H.,
Eds.; Marcel Dekker: New York, 1996; Vol. 33, p 453.
(41) Miller, J. R.; Beitz, J. V. J. Chem. Phys. 1981, 74, 6746-6756.
(49) Felts, A. K.; Pollard, W. T.; Friesner, R. A. J. Phys. Chem. 1995,
99, 2929-2940.
(50) Pollard, W. T.; Felts, A. K.; Friesner, R. A. AdV. Chem. Phys. 1996,
93, 77-134.
(51) Davis, W. B.; Svec, W. A.; Ratner, M. A.; Wasielewski, M. R.
Nature 1998, 396, 60-63.

Conformational Gating of Long Distance Electron Transfer
J. Am. Chem. Soc., Vol. 123, No. 32, 2001 7879
Chart 1
Figure 1. Transient absorption spectra of 1 in MTHF, obtained at 298
(s) and 170 K (- - -). These spectra were taken 10.5 and 40 ps,
respectively, after the arrival of a 150 fs, 483 nm pump pulse at the
sample. The data were normalized to a value of 1.0 at 720 nm in both
plots to accentuate the differences between the two spectra.
dynamics as the distance between the donor and acceptor is
increased. This results in a wire-like, soft distance dependence
of the electron-transfer rates. This series of molecules employs
a tetracene (TET) electron donor, a pyromellitimide (PI) electron
acceptor, and oligo-p-phenylenevinylene (OPV) bridges. At
room temperature, when the two shortest bridges, benzene and
trans-stilbene, are used, the electron-transfer rates show a steep
distance dependence in 2-methyltetrahydrofuran (MTHF). How-
ever, when OPV oligomers containing either three, four, or five
phenyl rings are used as the bridges, the rates of charge
separation are faster than those measured through the two shorter
bridges (benzene or trans-stilbene) and the charge separation
rates show a very weak distance dependence. In the studies
reported here, the temperature dependence of the charge
separation (kCS) and charge recombination (kCR) rates of 1-5
in the glass forming solvent MTHF were studied from room
temperature down to 77 K, which is well below the 110 K glass
transition temperature of MTHF. In addition, these rate constants
for 3 were measured down to 8 K. The motivation of this work
is to compare and contrast the temperature dependence of the
ET rates in the two different regimes, superexchange vs bridge
hopping. We find that the temperature dependence of kCS can
be assigned to a breakdown of the Condon approximation in
all five molecules. In particular, we find evidence for gating of
the donor to acceptor electron transfer by specific vibrational
modes, which control HDA via torsional motion of the donor
relative to the bridge, as well as torsional motion between
adjacent bridge sites. These molecular motions can have a
profound effect on the ability of conjugated molecules to
transport charge for long distances.
Figure 2. Plots of log kCS T1/2 vs T-1 for 1 and 2 in MTHF.
720 nm is due to the formation of PI-.52 These spectra show
that lowering the temperature results in relatively minor changes.
The most notable change is a sharpening of the first vibronic
band of PI- at 660 nm. The raw rate constants for CS and CR
within 1-5 in MTHF at all the temperatures measured are listed
in the Supporting Information. Given the functional form of eq
1, the CS rate constants are assumed to be of the form kCS
)
A′ T-1/2 exp(-Ea/kBT). The temperature dependence data for 1
1/2
and 2 are plotted as log(k
T
) vs T-1 for CS in Figure 2. The
CS
CS rate constants for 1 were fit to a straight line using linear
least-squares regression to yield Ea ) 349 ( 12 cm-1 and A′ )
1.28((0.16) × 1013 s-1 K1/2 with an acceptable quality of fit
(R ) 0.991). In contrast, fitting a plot of log kCS vs T-1 to a
straight line (not shown) yields an R value of only 0.952,
indicating that the Arrhenius prefactor most likely has a T-1/2
temperature dependence associated with it as predicted by
eq 1.
While the CS rate constants for 1 strictly increase with
increasing temperature, the CS rate constants for 2 show a
turnover from activated behavior at low temperatures to a
negatively activated regime above T ) 210 K, Figure 2. To see
Results
Transient absorption spectra of 1 following selective excita-
tion of TET with a 150 fs, 483 nm pump pulse at 298 (10.5 ps)
and 170 K (40 ps) are shown in Figure 1. The intense band at
(52) Gosztola, D.; Niemczyk, M. P.; Svec, W. A.; Lukas, A. S.;
Wasielewski, M. R. J. Phys Chem. A 2000, 104, 6545-6551.

7880 J. Am. Chem. Soc., Vol. 123, No. 32, 2001
DaVis et al.
Table 1. Arrhenius Prefactor and Activation Energy for the CS
Rates of 1-5 in MTHF
molecule
temp range
Ea (cm-1
)
A′ (s-1
)
R
1
2
3
4
5
120 K e T e 230 K 349 ( 12 1.28((0.16) × 1013 0.991
110 K e T e 210 K 366 ( 27 9.69((2.48) × 1012 0.986
120 K e T e 230 K 353 ( 11 4.20((0.39) × 1013 0.996
120 K e T e 210 K 373 ( 18 3.33((0.50) × 1013 0.991
120 K e T e 298 K 349 ( 13 1.99((0.21) × 1013 0.993
Figure 4. Plots of log kCS T1/2 vs T-1 for 3, 4, and 5 in MTHF.
ECR/kBT) to the CR data for 1 and 2 over the temperature ranges
within which the data display linear activated behavior are
presented in Table 2. The values of Ea are somewhat larger for
CR than those exhibited by the CS reactions. A comparison of
the CS and CR rate constants in 1 (Figures 1 and 2) shows that
at high temperatures the electron transfer occurs with nearly
the same rate in both the forward and reverse directions, but at
low temperatures (T < 190 K) the rate of CR slows down with
respect to the CS rate. In molecule 2, the CR rate constants are
always less than those of CS and the two sets of data do not
appear to be correlated in any way.
The temperature dependencies of the CS and CR rate
constants in 3-5 were also measured in MTHF. The data for 3
extend from room temperature down to 8 K. The data for 4
and 5 only extend from room temperature to 120 K because
these molecules are photochemically labile in the MTHF glass.
Figure 4 shows a plot of the CS data obtained from molecules
3-5. All three molecules show similar behavior above the Tg
of MTHF (∼110 K), namely a region of temperature indepen-
dence at high temperatures, followed by a region of thermal
activation at lower temperatures. At T e 120 K the CS rates
for 3 become temperature independent.
The three sets of CS data in Figure 4 were fit to the equation
kCS ) A′ T-1/2 exp(-Ea/kBT), over the temperature range in
which linear behavior was observed using a linear least-squares
fitting routine to obtain Ea and A′, Table 1. As the bridge gets
longer, A′ decreases and Ea appears to stay the same. Even more
striking is the similarity between these activation energies and
those measured for 1 and 2 in the same temperature region.
The activation energies measured using data obtained between
120 and 210 K are the same for all five DBA molecules within
experimental error.
Figure 3. Plots of log kCR T1/2 vs T-1 for 1 and 2 in MTHF.
if the temperature dependence for the CS rate constants in 2 in
the low-temperature activated region between 110 and 210 K
can be described by semiclassical ET theory, these data were
plotted as log(kCST1/2) vs T-1, and the data fit to a straight line
(not shown). The R-value of this fit, R ) 0.977, was not as
good as that for a linear fit to the CS data for 1. The values of
the parameters Ea and A′ from the fit to the CS data of 2 have
the values 366((27) cm-1 and 9.69((2.48) × 1012 s-1 K1/2
,
respectively. Thus the measured values of Ea and A′ from the
fits to the CS data are very similar for 1 and 2, and are
summarized in Table 1. In addition, Figure 2 shows that the
magnitudes of the CS rate constants for 1 and 2 differ by only
about a factor of 2 over the temperature range 210 K g T g
110 K. The line drawn through the temperature dependence data
for 2 shown in Figure 2 is a nonlinear least-squares fit to a
kinetic model that will be described below.
A comparison of the temperature dependencies of the
corresponding CR rate constants for 1 and 2 is shown in Figure
3. The data for 1 and 2 are distinct from one another, both in
magnitude and in thermal behavior. For 1, the CR rates are
essentially temperature independent at high temperatures, then
decrease with decreasing temperature below T ∼ 230 K. In
contrast, the CR rates measured for 2 appear to be thermally
activated from room temperature down to ∼140 K, after which
they show little if any temperature dependence. The results of
a linear least-squares fit of the function kCR ) A′ T-1/2 exp(-
The temperature dependencies of the CR rate constants in
3-5 are also very similar to one another, Figure 5. Molecules
3 and 4 exhibit monoexponential CR kinetics near room
Table 2. Arrhenius Prefactor and Activation Energy for the CR Rates of 1-5 in MTHF
molecule
component
temp range
Ea (cm-1
)
A′ (s-1
)
R
1
2
3
120 K e T e 250 K
140 K e T e 298 K
100 K e T e 230 K
79 K e T e 200 K
120 K e T e 210 K
120 K e T e 190 K
160 K e T e 250 K
120 K e T e 170 K
549 ( 19
460 ( 8
3.45((0.52) × 1013
2.15((0.13) × 1012
1.12((0.11) × 1011
3.79((0.70) × 1011
1.30 ((0.21) × 1011
2.12 ((0.52) × 1012
3.59 ((0.60) × 1010
4.86 ((1.11) × 1012
0.994
0.998
0.994
0.986
0.993
0.994
0.962
0.995
slow
fast
slow
fast
slow
fast
331 ( 11
291 ( 17
446 ( 19
507 ( 28
210 ( 24
505 ( 26
4
5

Conformational Gating of Long Distance Electron Transfer
J. Am. Chem. Soc., Vol. 123, No. 32, 2001 7881
slow CR rates of 3-5. The fit of this functional form to the
CR data of 3-5 is good in general, except for the fast rates in
3 and the slow rates in 5. For each of the three compounds, the
fast and slow CR rates do not have Arrhenius prefactors or
activation energies that match within the uncertainty of the fits.
Therefore, it appears as if the fast and slow CR rates arise from
two different processes occurring in all three molecules.
Discussion
The CS and CR reactions within molecules 1-5 exhibit
complex temperature dependencies. As one might expect, the
degree of complexity increases as the number of torsional
motions within these DBA molecules increases. These data will
be analyzed in a stepwise fashion beginning with the molecules
having the two shortest bridges. A precise description would
require analyzing the motion of the photogenerated wave packet
on the full multidimensional geometric space of the molecule.
A reasonable picture, more in the tradition of chemical state
language, will prove more helpful, and we adopt this approach
in describing the decay dynamics associated with the CS and
CR processes.
Molecules 1 and 2. Apart from the explicit temperature
dependence of kET expressed in eq 1, ∆G and λ are also
temperature dependent. The standard assumption for aromatic
organic π donors and acceptors is that an ET reaction will couple
to a high-frequency vibrational mode of about 1500 cm-1, the
frequency of a CdC stretch, and that λi has a value close to 0.3
eV.53 The electronic coupling matrix element HDA is usually
assumed to be temperature independent (the Condon ap-
proximation).
Weller derived the following equation for the free energy of
an ion-pair state, based upon the dielectric continuum model
of solvation:54
e2
ꢀsRDA
1
1
1
1
∆GIP0 ) EOX - ERED
-
+ e2
+
-
Figure 5. Plots of log kCR T1/2 vs T-1 for 3, 4, and 5 in MTHF: (A)
slow kinetic component and (B) fast kinetic component.
(
)(
)
2RD 2RA ꢀs ꢀs′
(4)
temperature, which become biexponential at lower temperatures.
The turnover from mono- to biexponential CR kinetics occurs
at 210 K in 3 and at 190 K in 4. As the temperature decreases,
the fast/slow amplitude ratio of the two components increases,
and at low temperatures, the faster component is completely
dominant. In 5, biexponential kinetics are already seen at room
temperature, and the fast/slow amplitude ratio increases with
decreasing temperature, similar to what is observed for 3 and
4. Figure 5 shows the dependencies of the fast and slow CR
rate constants of 3-5 on temperature. For 3 and 4, the
amplitudes of the slow CR components, Figure 5A, display a
region of temperature independence at high temperatures, as
well as a region of activated behavior at temperatures below
200 K. The slow CR rates in 5 increase as the temperature
decreases from room temperature to 250 K, after which these
rates strictly decrease with decreasing temperature. The ampli-
tudes of the fast CR components for 3 and 4 in Figure 5B are
thermally activated over their entire temperature range. The fast
CR rates in 5 increase as the temperature is lowered from room
temperature to 250 K, show a region of temperature indepen-
dence from 250 to 170 K, then decrease with decreasing
temperature.
where EOX and ERED were defined above, e is the electronic
charge, ꢀs is the static dielectric constant of the solvent in which
the rate constants are measured, ꢀs′ is the dielectric constant of
the solvent in which EOX and ERED are measured, RDA is the
donor-to-acceptor center-to-center distance, and RD and RA are
the spherical radii of the donor and acceptor, respectively. The
free energies for the CS and CR reactions are calculated using
∆GCS° ) -E0-0 + ∆GIP°
∆GCR° ) -∆GIP°
(5)
where E0-0 is the energy of the lowest excited singlet state of
the photoexcited donor. These equations assume that the zero
of the energy scale corresponds to the energy of the ground-
state molecule. Equation 5 indicates that the temperature
dependence of both ∆GCS and ∆GCR arises from the corre-
sponding dependence of ∆GIP. In turn, the temperature depen-
dence of ∆GIP is principally a function of the temperature
dependence of ꢀs because the redox potentials within the Weller
treatment are determined in a polar solvent with dielectric
constant ꢀs′ at room temperature. While the utility of the Weller
treatment has been questioned for solvents of low polarity, the
The fast and slow CR rates in 3-5 were all fit to the equation
kCR ) A′ T-1/2 exp(-Ea/kBT) in the temperature regions where
they show thermally activated behavior. The results of these
least-squares fits are listed in Table 2, along with the temperature
regions over which activated behavior is seen in the fast and
(53) Closs, G. L.; Calcaterra, L. T.; Green, N. J.; Penfield, K. W.; Miller,
J. R. J. Phys. Chem. 1986, 90, 3673-3683.
(54) Weller, A. Z. Phys. Chem. 1982, 133, 93-98.

7882 J. Am. Chem. Soc., Vol. 123, No. 32, 2001
DaVis et al.
Table 3. ∆GCS° and λs at 298 K Calculated for Molecules 1-5
Using Eqs 4-6
molecule
RDA (Å)
-∆GCS° (eV)
λs (eV)
1
2
3
4
5
11.1
17.7
24.3
30.9
38.0
0.81
0.74
0.71
0.69
0.67
0.83
1.0
1.1
1.1
1.2
use of eq 4 to estimate ∆GIP in solvents of moderate polarity
yields reasonable results.55
Again drawing upon dielectric continuum theory,1,2 Marcus
showed that λs can be approximated from the following
equation:
1
1
1
1
1
λs ) e2
+
-
-
(6)
(
)(
)
2RD 2RA RDA ꢀ∞ ꢀs
Figure 6. Plots of log kCS T1/2 vs T-1 for 1 and 2 in MTHF using rate
constants predicted by eq 1 using temperature-dependent values of ∆GCS
and λ.
where ꢀ∞ is the optical dielectric constant of the solvent and
the remaining parameters are defined above. The temperature
dependence of λs is due to that of ꢀ∞ and ꢀs. The utility of using
continuum models for predicting the temperature dependence
of λs in highly dipolar media has been questioned recently.56
These deviations may be diminished somewhat in solvents of
more modest polarity like MTHF. As an approximation eqs 4-6
are used to calculate ∆GCS° and λs for molecules 1-5 at 298
K, which are listed in Table 3. The values of RD ) 4.9 Å and
RA ) 3.4 Å, as well as the values of RDA listed in Table 3,
were obtained from MM+ optimized geometries.57 For MTHF,
ꢀ∞ ) 1.97 and ꢀs ) 6.97 at 298 K.58 The one-electron oxidation
and reduction potentials, EOX and ERED, for TET and PI are
0.77 and -0.79 V vs SCE, respectively, in butyronitrile (ꢀs′ )
24.858). The experimental temperature dependence of ꢀs for
MTHF at T g 140 K is accurately described by the equation
ꢀs(T) ) 3.1 + 860/(T - 73), while at lower temperatures, after
MTHF has entered into a glassy state, ꢀs ) 2.6.59 The
temperature dependence of ꢀ∞ is assumed to vary linearly from
1.97 at 298 K to 15.2 at 77 K (see Supporting Information).
The values of ꢀ∞ around and below the Tg of MTHF are suspect
because of the large changes in solvent properties that occur at
the glass transition. However, they will be used here with the
caveat that any data that are calculated at those temperatures
are considered highly approximate.
Equation 1, along with eqs 4-6 can be used to estimate the
value of HDA for 1 and 2 using the measured values of kET at
298 K along with the other parameters listed above. Doing so
gives HDA values of 50 and 48 cm-1 for 1 and 2, respectively.
Assuming that HDA is temperature independent, as well as using
the temperature-dependent values of ∆GCS° and λs, eq 1 predicts
the CS rate constant temperature dependencies for 1 and 2
shown in Figure 6. The predicted CS rate constants for both 1
and 2 are only weakly temperature dependent, and do not
reproduce the experimentally determined temperature depend-
encies.
Figure 7. Tetracene-bridge and intrabridge torsions and vibrations
within 1-5.
that the observed temperature dependence results from a
breakdown of the Condon approximation. There are two
common structural features present in both 1 and 2. The first is
the attachment of a phenyl group to the tetracene donor, and
the second is the attachment of a phenyl group to the pyro-
mellitimide acceptor. Since there are other donor-bridge-
acceptor molecules that incorporate a phenyl bound to pyro-
mellitimide and obey semiclassical ET theory,60,61 a molecular
motion associated with this linkage is probably not responsible
for Condon breakdown. It is therefore useful to focus on the
tetracene-phenyl linkage.
In the ground state of 5-phenyltetracene (Ph-TET), the
dihedral angle, Φ, between the tetracene π system and the
phenyl π system is reasonably large, ∼69° according to MM+
calculations. This accounts for the reasonably weak electronic
coupling between tetracene and the phenyl group seen in the
absorption spectra of 5-(p-tolyl)tetracene.51 In the S1 state of
Ph-TET, Φ ∼ 60°,62 which is a rather modest change in
dihedral angle from the ground state. Since π-π electronic
coupling depends on cos Φ, the electronic coupling between
the tetracene chromophore and the phenyl ring is expected to
Since the rate constants for CS within 1 and 2 are large and
the CS process is weakly activated for T e 210 K, it is possible
(55) Oevering, H.; Paddon-Row, M. N.; Heppener, M.; Oliver, A. M.;
Cotsaris, E.; Verhoeven, J. W.; Hush, N. S. J. Am. Chem. Soc. 1987, 109,
3258-3269.
(56) Vath, P.; Zimmt, M. B.; Matyushov, D. V.; Voth, G. A. J. Phys.
Chem. B 1999, 103, 9130-9140.
(57) Optimized geometries of 1-5 were obtained using the MM+ force
field within Hyperchem 5.01a (Hypercube, Waterloo, Ont.).
(58) CRC Handbook of Chemistry and Physics, 70th ed.; CRC Press:
Boca Raton, FL, 1989.
(60) Osuka, A.; Nagata, T.; Naruyama, K.; Mataga, N.; Asahi, T.;
Yamazaki, I.; Nishimura, Y. Chem. Phys. Lett 1991, 185, 88-94.
(61) Wiederrecht, G. P.; Svec, W. A.; Wasielewski, M. R. J. Phys. Chem.
B 1999, 103, 1386-1389.
(59) Furutsuka, T.; Imura, T.; Kojima, T.; Kawabe, K. Osaka UniV. Eng.
Rep. 1974, 24, 367.

Conformational Gating of Long Distance Electron Transfer
J. Am. Chem. Soc., Vol. 123, No. 32, 2001 7883
be similar in the S0 and S1 states of tetracene. There is a torsional
barrier in the S1 state of Ph-TET that impedes the tetracene
and phenyl group from reorienting into a more coplanar
geometry. The physical origin of this barrier is steric hindrance
between the hydrogens at the 4 and 6 positions of tetracene
and the hydrogens at the ortho positions of the phenyl ring,
Figure 7A. The attachment of a phenyl group to the 5-position
of tetracene results in a significant enhancement of a 296 cm-1
out-of-plane vibration of tetracene.62 This vibrational mode is
thought to lower the torsional barrier in the S1 state of Ph-
TET by displacing the tetracene hydrogen atoms enough to
allow the phenyl group to rotate into a more coplanar config-
uration. Thus, the following reaction scheme is proposed for 1:
group, which causes distortions from planarity.63-66 The flex-
ibility of this torsional mode, coupled with its low zero point
frequency, ∼8 cm-1, leads to strong coupling of this mode to
other vibrational modes in trans-stilbene. The normal mode
displacements of this stilbene torsional mode are shown in
Figure 7B. Even in the lowest vibrational level, the potential
of this torsional mode is predicted to have turning points at
(22°, indicating there are large deviations from planarity in
trans-stilbene even at low temperatures.63 Since electronic
coupling through a trans-stilbene bridge will approximately scale
like cos Θ, where Θ is the dihedral angle between the phenyl
and vinyl groups, decreasing the temperature will result in an
increase in the electronic coupling. Therefore, this torsional
motion is expected to exhibit a negative activation energy, which
is in fact found for the high-temperature CS rates in 2.
Assuming that the tetracene-bridge and stilbene torsional
modes lead to a break down of the Condon approximation, the
following kinetic scheme is proposed to explain the experimental
CS rates in molecule 2.
Since the donor and acceptor are in an unfavorable geometry
for CS when Φ is 60°, we can assume that k1 . k2. Also, we
assume that k3 . k1, k-1 because of strong donor-bridge
electronic overlap in a planar geometry. Under these conditions,
the observed CS rate in 1 will be k1, which is the rate of torsional
motion between the tetracene donor and phenyl bridge. This
kinetic scheme and accompanying assumptions about the
magnitudes of the rate constants contained within are able to
explain why the dominant thermal contribution to the temper-
ature dependence of kCS in 1 appears to be due to the tetracene-
phenyl torsional rearrangement. The kinetic schemes used to
analyze the results depend on equilibria between a small number
of conformations. While distributions of conformations un-
doubtedly exist, our data do not exhibit distributed kinetics. This
is most likely a consequence of narrow conformational distribu-
tions.
The rate constants k4 and k-4 describe the forward and backward
isomerization rates for the tetracene-phenyl torsional mode,
k5 is the isomerization rate of the stilbene torsional mode, and
k6 is the rate at which the electron moves from donor to acceptor,
once the molecule has assumed the optimum geometry for ET
(assumed to be Φ ∼ 0°, Θ ∼ 0°). The rate of the donor-to-
acceptor ET reaction in the initial excited-state geometry, k7, is
assumed to be negligible as compared to k4 because of weak
electronic coupling between the donor and acceptor in the
(Φ ∼ 60°, Θ * 0°) geometry. Using the steady-state ap-
proximation, the rate of formation of 1D*BA (Φ ∼ 0°,
Θ ∼ 0°), kplanar, is given by
Lowering the torsional barrier between the tetracene donor
and the phenyl or stilbenyl bridge allows more efficient
electronic coupling between the donor and acceptor in 1 and 2.
The torsional barrier is high at low temperatures because the
296 cm-1 vibrational mode is inactivated, resulting in relatively
weak electronic coupling between the donor and acceptor, and
hence a slower CS rate. Raising the temperature activates the
out-of-plane vibrational mode, lowering the torsional barrier and
allowing the donor and bridge to assume a more favorable
geometry for charge transport. It is intriguing that the frequency
of this mode, 296 cm-1 in the gas phase, is very similar to the
activation energy measured for the CS rates of 1 at all
temperatures measured, and the activation energy of kCS of 2 at
T e 210 K.
k5k4
k5k4
kplanar
)
)
(9)
k5 + k-4
k5 + (k4/K4)
where K4 is the equilibrium constant between the two tetracene-
phenyl conformers. If kplanar . k6, the rate of formation of the
ion-pair state will be dominated by k6 and the observed ET
dynamics should follow the temperature behavior predicted by
semiclassical ET theory. However, if kplanar e k6, kplanar will be
observed experimentally. Since the temperature dependence of
the CS rates in 2 does not appear to follow standard theories of
ET, it is assumed that the inequality kplanar e k6 holds. Therefore
the torsional dynamics of 2 control the observed CS rates,
kobs ) kplanar. Because k4 has a positive activation energy and
k5 has a negative activation energy, at high temperatures k4 .
k5 and kobs ∼ K4k5. At low temperatures, the opposite limit
should be seen: k5 . k4 and kobs ∼ k4. Therefore, at high
temperatures kobs should be an increasing function with decreas-
While the thermal activation of this tetracene vibrational mode
might explain the low-temperature data for 1 and 2, it cannot
account for the high-temperature turnover in kCS seen experi-
mentally for 2. Again, the chemical structures of 1 and 2 can
be examined for clues as to what is causing the differences in
the CS dynamics of these two donor-acceptor molecules. Both
molecules have the tetracene-phenyl and phenyl-pyromellit-
imide linkages, but only 2 has an intervening vinyl group
between the donor and acceptor. In the ground state, trans-
stilbene has a very anharmonic torsional mode about its vinyl
(63) Suzuki, T.; Mikami, N.; Ito, M. J. Phys. Chem. 1986, 90, 6431-
6440.
(64) Syage, J. A.; Felker, P. M.; Zewail, A. H. J. Chem. Phys. 1984, 81,
4685-4705.
(65) Haller, K.; Chiang, W.-Y.; del Rosario, A.; Laane, J. J. Mol. Struct.
1996, 379, 19-23.
(62) Lo¨hmannsro¨ben, H.-G.; Bahatt, D.; Even, U. J. Phys. Chem. 1990,
94, 4025-4029.
(66) Spangler, L. H.; van Zee, R.; Zwier, T. S. J. Phys. Chem. 1987, 91,
2781-2786.

7884 J. Am. Chem. Soc., Vol. 123, No. 32, 2001
DaVis et al.
ing temperature due to the behavior of k5. As the temperature
decreases further, the rates should turn over and kobs decrease
with decreasing temperature because of k4. Assuming k4 )
A′4T-1/2 exp(-Ea4/kBT) and k5 ) A′5T-1/2 exp(-Ea5/kBT), eq
9 can be used to fit the experimental data, Figure 2. For
simplicity, K4 was given a value of 1, and the best fit to the
experimental data gave the following values for the other four
The weak temperature dependence for CR observed at high
temperatures suggests that k9 . k8 in that regime. As the
temperature decreases, eq 1 predicts that k9 should decrease.
At the same time, k8 should increase due to the negatively
activated nature of this process. Thus, at low temperatures the
observed CR rate constant is k10. The kinetic scheme in eq 10
describes the experimental CR rates if k8 . k10 . k-8 at all
temperatures and k9 . k8 at high temperatures. Thus, it appears
that torsional motion between the twisted and planar donor-
bridge states does not kinetically control the CR dynamics of
1. However, on a cautious note, the equilibrium constant of the
donor-bridge torsional motion could change once CS has
occurred and CR commences depending on how strongly the
torsional mode will couple with other vibrational modes once
tetracene becomes a radical cation. Unfortunately, no experi-
mental data concerning this torsional-vibrational coupling in
the 5-phenyltetracene radical cation, or any other related
systems, are available.
model parameters: A′4 ) 2.57((2.04) × 1013 s-1 K1/2, Ea4
)
)
447 ( 94 cm-1, A′5 ) 5.53((1.45) × 1010 s-1 K1/2, and Ea5
-460 ( 58 cm-1. Varying K4 did not lead to an appreciable
improvement in ø2 for the fit, so only the data obtained with
K4 ) 1 will be discussed. The values of Ea4 and A′4 are similar
to the Arrhenius parameters obtained previously from the CS
rates of 2 from 210 K g T g 110 K, as well as those obtained
for 1. The activation energy of k5 is indeed negative.
The assignment of the negative activation energy to a single
coupled, higher frequency, vibrational mode in trans-stilbene
is complicated by the extreme anharmonicity of the torsional
mode. For instance, in the ground-state vibrational spectrum of
trans-stilbene there are a large number of overtone and
combination bands, which should not have significant population
normally. Out of all the low-frequency vibrational modes in
trans-stilbene, there is one that shows a large coupling to the
torsional motion of the phenyl-vinyl linkage with a frequency
of 413 cm-1. The assignment of this vibrational mode is still
somewhat ambiguous,63,64 but it has been tentatively assigned
to an in-plane bending motion of the phenyl rings about the
vinyl group. The qualitative behavior of this Ag symmetric mode
is shown in Figure 7C.65 This vibrational mode may couple
strongly to the torsional mode of stilbene because as the phenyl
rings move toward the vinyl group, there should be steric
repulsion between the ortho hydrogens on the phenyl rings and
the vinylic hydrogens, Figure 7D. Thermal activation of this
symmetric mode might push the phenyl rings out of plane to
minimize repulsion between the hydrogens.
Just as in molecule 1, there may be competition between CR
and torsional relaxation in 2. The data in Figure 3 show that
for 298 K > T > 150 K, the CR reaction exhibits a single
positive activation energy, EaCR ) 450 cm-1. On the basis of
the fact that EaCR is positive, the negatively activated phenyl-
vinyl torsional motion, which must be very rapid throughout
this temperature range, has little or no influence on the CR rates.
Thus, the same arguments developed for CR in 1 using eq 10
can be applied to CR in 2. As a result, invoking the donor-
bridge torsional motion is not necessary to explain the temper-
ature dependence of CR for 298 K > T > 150 K.
At temperatures below the Tg of MTHF, the CR rates become
temperature independent. This behavior has been seen before
in protein ET complexes. Perhaps the most famous example is
that of the cytochrome-photosynthetic reaction center complex
from the organism Chromatium Vinosum.67,68 In this protein
complex, the rate constant for ET from the cytochrome to the
oxidized reaction center is nearly temperature independent below
100 K, while above 100 K it is thermally activated. This
behavior can be described by semiclassical ET theory, which
assigns the temperature dependence seen at high temperatures
to an activated barrier crossing, while the ET rates are dominated
by nuclear tunneling at low temperatures.5,6 The similarity
between the qualitative behavior seen in the CR rates of 2 with
the predictions of semiclassical ET theory indicates that the CR
dynamics in 2 probably follow the Condon approximation at
all the temperatures studied.
Since the CS rates are controlled primarily by intramolecular
torsions, the CR rates may be sensitive to these molecular
motions as well. Since CS in 1 occurs at an optimized,
nonequilibrium geometry near the top of the tetracene-phenyl
torsional barrier, there must be kinetic competition between CR
back down to the ground state and geometrical relaxation of
the donor and bridge to a twisted geometry in the ion-pair state.
A kinetic scheme that describes these dynamics is outlined as
follows:
Molecules 3-5. In the previous section it was argued that
changes in the tetracene-phenyl torsional angle are responsible
for the temperature dependence of kCS in 1 and 2. This same
torsional motion could control the CS rates of 3-5 at temper-
atures above the Tg of MTHF as well. In our earlier work on
1-5 at 298 K we observed a superexchange ET mechanism in
1 and 2 and a hopping ET mechanism in 3-5.51 It is useful to
examine whether the kinetic scheme proposed earlier to account
for the CS data in 3-5 at 298 K can also be used to describe
the temperature dependence data presented here:51
In eq 10, it is assumed that the relaxed geometry between the
donor and bridge in the ion-pair state is located at Φ ∼ 60°.
Rate constant k9 is should be larger than k10 at all temperatures
because the electronic coupling between D+ and B in the planar
ion-pair state is larger than that in the twisted ion-pair state.
Thus, at all temperatures the observed rate constant for CR will
depend principally on competition between k8 and k9. If k9 .
k8, the observed CR rate constant will be k9, which should have
a weak temperature dependence due to the strong electronic
coupling between nearly coplanar D+ and B. On the other hand,
if k8 . k9 and k8 . k-8, the observed CR rate constant is k10.
(67) DeVault, D.; Chance, B. Biophys. J. 1966, 6, 825.
(68) DeVault, D.; Parkes, J. H.; Chance, B. Nature 1967, 215, 642-
644.

Conformational Gating of Long Distance Electron Transfer
J. Am. Chem. Soc., Vol. 123, No. 32, 2001 7885
The steady-state solution to this equation gives the following
prediction for the observed ET rate constant, kobs
the CS reaction, the energy gap between the 1D*BA and D+B-A
states dictates whether the superexchange and hopping electron
transport mechanisms dominate. Likewise, the energy gap
between the D+BA- and DB+A- states controls the hole
hopping CR dynamics. The energy gap between these two ion-
pair states is expected to change dramatically with temperature
in 3-5 because of structural changes in the OPV bridges. All
three bridges are expected to have the same out-of-plane
phenyl-vinyl torsional motion as the stilbene bridge in 2;
however, now there are multiple vinyl groups in each bridge.
As the number of rotors increases in these bridges, their potential
surfaces should become increasingly complex with multiple
barriers and extrema. In fact, numerous torsional conformations
of the bridge coexist at high temperatures, and the experimental
CR data reflect the ensemble average. As the temperature
decreases, the bridges are expected to become more planar
similar to the behavior of trans-stilbene.
The nonplanar conformations of the OPV bridges at room
temperature should keep the ion-pair energy gap relatively large,
and turn off hole hopping dynamics. In 3 and 4, this results in
monoexponential CR decays at high temperatures because
k16 . k14 in eq 13. As the temperature decreases, the OPV
bridges should become more planar, and hence the energy gap
between D+BA- and DB+A- should decrease. Due to the
complex torsional landscape in these bridges, some molecules
should be in a more planar geometry, while others may get
trapped in twisted local minima, thus 3 and 4 exhibit biexpo-
nential kinetics. As the temperature is decreased further, k14 is
expected to increase because of the decreasing donor-bridge
energy gap, while k15 remains fast at all temperatures because
of strong electronic coupling between B+ and A-. Thus, k14
will dominate the observed CR rates in 3 and 4 at low
temperatures and the observed CR decays will return to
monoexponential kinetics.
While the CR rate constants for 3 and 4 display the same
qualitative behavior in Figure 5, the rate constants for 5 differ.
Even at room temperature the fast decay component dominates
the CR dynamics. The fast rates in 5 increase dramatically as
the temperature is lowered from 298 to 250 K, and then level
out before decreasing as the temperature decreases further. Over
the temperature range 298 K g T g 200 K, the fast CR rate
constants in 5 are larger than all the CR rate constants in 3 and
4. This behavior most likely arises from the D+BA- and DB+A-
states becoming resonant in 5 near room temperature. From eq
4, at 298 K the energy of the D+BA- ion-pair state is 1.88 eV,
while that of the corresponding DB+A- ion-pair state is 1.91
eV. The energy difference between these two ion-pair states at
298 K is only 0.03 eV, or 242 cm-1. A similar analysis yields
D+BA- - DB+A- energy gaps of 0.16 eV (1300 cm-1) in 3
and 0.08 eV (650 cm-1) in 4. A small change in the planarity
of the OPV bridge in 5, induced by a small decrease in the
temperature, is probably sufficient to bring the two ion-pair
states D+BA- and DB+A- into resonance. This would account
for the strong biexponential kinetics for CR seen at room
temperature in 5. Because the ion-pair state energy gaps are
larger in 3 and 4, biexponential CR kinetics are not observed
in these compounds until the temperature is decreased. Reso-
nance between the D+BA- and DB+A- states in 5 would also
account for the fast and slow CR rates not behaving like those
seen in 3 and 4.
k12k11
k12k11
kobs ) k13
+
=
(12)
k12 + k-11 k12 + k-11
The last equality in eq 12 is assumed to hold for molecules
3-5, where the direct donor-to-acceptor charge-transfer rate,
k13, is assumed to be small based on the trend established by
the CS rates of 1 and 2. When k12 . k-11 in eq 12, kobs ∼ k11,
and when k12 , k-11, kobs ∼ K11k12, where K11 is the equilibrium
constant for reaction 11. Since k12 describes the charge shift
reaction between the bridge and acceptor, it may have a weak
temperature dependence since charge transfer from the bridge
to the acceptor is strongly exergonic. The rate constants k11 and
k-11 describe the rates of promotion of an electron from the
donor onto the bridge and vice versa. These two rate constants
should be strongly influenced by torsional motion between the
tetracene donor and first phenyl attached to it. At high
temperatures this motion should be rapid, and both k11 and k-11
may be larger than k12. Therefore, the observed rate of charge
transport will be dominated by k12, which is expected to possess
a weak temperature dependence. This may explain the observed
weak temperature dependence of the CS rate constants of 3-5
at high temperatures. As the temperature decreases, k11 and k-11
decrease, k12 . k11 and k-11, and the observed CS rate should
be dominated by k11. Thus, the kinetics of the donor-bridge
torsional motion should be the rate-limiting step at lower
temperatures, consistent with the similarity of the activation
energy of CS in all five ET molecules at temperatures between
200 and 120 K.
It is uncertain if the rates of CS in 4 and 5 become
temperature independent at low temperatures. At temperatures
at or below Tg for MTHF, slowing solvent dynamics will
strongly affect the observed rate constants. However, for 3 data
were obtained at lower temperatures and the rates do indeed
become temperature independent. Just as in trans-stilbene, the
OPV oligomers are expected to have anharmonic torsional
potentials associated with each of the vinyl linkages. Analyzing
this multiple rotor problem in detail is beyond the scope of this
paper. However, some qualitative arguments about their behavior
can be made. As the temperature decreases, the geometry of
the OPV bridge should become more planar overall because
the large-amplitude torsional motions about the vinyl linkages
diminish. Once MTHF enters into a glassy state, these motions
should be nearly absent in the molecule. As the bridge becomes
more planar, the energy of its LUMO is expected to decrease.
Because the bridge LUMO and the donor LUMO are nearly
degenerate at room temperature in 3-5,52 enhanced conjugation
in the bridge at low temperatures could serve to bring the donor
and bridge LUMOs into resonance. This would cause the
temperature dependence of the CS rates in 3-5 to become
activationless at low temperatures, as is observed experimentally
in 3.
Turning now to the CR rates in 3-5, the most notable feature
of the data is the observed biexponential kinetics. In our previous
work on these molecules,51 it was argued that the following
kinetic scheme described the CR dynamics at room temperature:
Conclusions
Molecules 3-5 display hole-hopping dynamics during the CR
reaction, as well as electron hopping during the CS reaction. In
Long conjugated molecules appear at first glance to be ideally
suited for incorporation into donor-bridge-acceptor assemblies

7886 J. Am. Chem. Soc., Vol. 123, No. 32, 2001
DaVis et al.
instrument function is 180 fs. The concentrations of 1-5 used for
transient absorption measurements were approximately 100 µM in 2
mm path length cells. Both pump beams and the probe beam were 200
mm in diameter with the 483 nm pump beam having an energy of 0.5
µJ at the sample.
designed to carry out long distance charge transport. However,
an important aspect of these long π bridges that is often
neglected is their propensity for significant conformational
flexibility due to low-frequency vibrational modes. The results
presented here show that low-frequency vibrations of the
tetracene donor and the OPV oligomers exert significant control
over the ET rates in 1-5, which results in conformational gating
of the ET reaction. In addition, these results serve as a warning
about neglecting bridge dynamics when analyzing both the
distance and temperature dependence of ET reactions. Other
conjugated bridges,69-71 such as poly(phenylene), poly(thiophene),
and poly(phenylene-ethynylene), are also strong candidate
systems for investigating the role of conformational gating in
photoinduced ET reactions. In addition, many of these concepts
can be explored further by preparing molecules in which
complex gating motions are restricted by the structure of the
donor-bridge-acceptor molecule.
For variable temperature experiments down to 77 K, a liquid nitrogen
cooled optical dewar (Janis Research VPS-100) was utilized, while for
experiments down to 8 K a liquid helium optical dewar (Janis Research
STVP-100) was utilized. The MTHF solvent used for each experiment
was freshly distilled and subjected to 3 freeze-pump-thaw cycles
immediately before use. All samples were prepared under a nitrogen
atmosphere to minimize the presence of oxygen. Each spectroscopic
sample was placed in a sealed, homemade, 2 mm path length optical
cell constructed from two transparent quartz windows and a 2 mm wide
Teflon spacer. After seating the sample holder in the dewar, the dewar
was evacuated to ∼10-3 Torr of pressure. The sample temperature was
maintained to within (0.5 K by a Lake Shore Cryotronics Model DRC-
82C temperature controller. The sample temperature was lowered from
room temperature down to the solvent glass transition temperature
(∼110-120 K for MTHF) by 10 to 20 deg increments. The sample
was allowed to equilibrate for 15-20 min at each new temperature
before data collection. After reaching the glass transition temperature,
the sample temperature was cooled all the way down to the temperature
of liquid nitrogen (∼77 K), or liquid helium (4.2 K), and warmed back
up to the glass transition temperature of the solvent in 10 deg
increments. Again, the sample was allowed to equilibrate at each new
temperature for a minimum of 15 min before data collection.
Experimental Section
Compounds 1-5 were synthesized using methods described in the
Supporting Information. The structures of these compounds were
1
confirmed using H NMR and mass spectral data.
Cyclic voltammetry (CV) and bulk electrolysis were carried out using
a potentiostat and voltage ramp programmer (Princeton Applied
Research, Models 173 and 175) and a standard three-electrode
arrangement. CV measurements used platinum working and auxiliary
electrodes as well as a saturated sodium calomel reference electrode
(SSCE). All electrochemical measurements were carried out in N2-
purged butyronitrile with 0.1 M Bu4NClO4 as the supporting electrolyte.
The scan rate for CV measurements was typically 50-100 mV/s.
Ferrocene was used as an internal redox standard for all measurements.
Ground state optical absorption measurements were recorded using
a computer controlled spectrophotometer (Shimadzu, Model 1601).
Femtosecond transient absorption measurements were obtained with
150 fs, 483 nm pump pulses and a 130 fs white light continuum probe
pulse produced by an amplified-Ti:sapphire laser system with an optical
parametric amplifier (OPA) as described previously.72 The total
Acknowledgment. This work was supported by the Division
of Chemical Sciences, Office of Basic Energy Sciences, U.S.
Department of Energy under grant no. DE-FG02-99ER14999
(M.R.W.) and by the NSF (M.A.R.). W.B.D. wishes to thank
the Link Foundation for a fellowship. The authors wish to thank
Mr. Walter A. Svec of Argonne National Laboratory for his
assistance in preparing 1-5.
Supporting Information Available: Synthetic and spec-
troscopic details, as well as temperature dependence of ꢀ∞ (PDF).
This material is available free of charge via the Internet at
http://pubs.acs.org.
(69) Helms, A.; Heiler, D.; McLendon, G. J. Am. Chem. Soc. 1992, 114,
6227-6238.
(70) Osuka, A.; Maruyama, K.; Mataga, N.; Asahi, T.; Yamazaki, I.;
Tamai, N. J. Am. Chem. Soc. 1990, 112, 4958-4959.
(71) Finckh, P.; Heitele, H.; Michel-Beyerle, M. E. J. Phys. Chem. 1988,
92, 6584-6590.
JA010330Z
(72) Lukas, A. S.; Miller, S. E.; Wasielewski, M. R. J. Phys. Chem. B
2000, 104, 931-940.