Femtosecond optical switching of electron transport direction in branched donor-acceptor arrays

Article abstract of DOI:10.1021/jp993774e

The branched donor-acceptor triad 4 and tetrad 5 were synthesized to study the possibility of controlling the direction of electron transfer in a divergent array of electron acceptors using ultrafast laser pulses. Compounds 4 and 5 employ 1,3,5-triaminobenzene as the central branch point. In 4, a4-(N-piperidinyl)-1,8-naphthaleneimide electron donor (ANI) was attached to the 1 position and two electron acceptors, 1,8:4,5-naphthalenediimide, NI, and pyromellitimide, PI, were attached to the 3 and 5 positions of the central benzene ring. In 5, the terminal end of the PI acceptor is functionalized with an additional NI molecule. Selective excitation of ANI in 4 and 5 with 400 nm, 130 fs laser pulses results in exclusive formation of NI^--ANI⁺-PI and NI^--ANI⁺-PI-NI, respectively, with τ = 115 ps, and a quantum yield of 0.99. Excitation of NI^- with 480 nm, 130 fs laser pulses produces the excited doublet state *NI^-, which transfers an electron to PI on the second branch of the benzene ring with a time constant of τ = 600 fs in 4 and τ = 750 fs in 5. The overall quantum yields for the two step process are 0.44 and 0.36 in 4 and 5, respectively. The resultant state NI-ANI⁺-PI^- in 4 undergoes a charge shift reaction returning to the initial ion pair state NI^--ANI⁺-PI with τ = 400 ps, while the corresponding state NI-ANI⁺-PI^--NI in 5 undergoes a charge shift with τ = 200 ps to yield NI-ANI⁺-PI-NI^-, which in turn exhibits a 2000 ps lifetime. These results show that once the electron has been switched to the branch containing the thermodynamically uphill PI acceptor, the electron will cascade down that branch to other acceptors that are more easily reduced. Model compounds 1-3, which were synthesized to aid in characterization of the switching dynamics, are also discussed. Photochemical control of directional charge transport may make it possible to design networks of organic molecules for information processing.

Full text of DOI:10.1021/jp993774e

J. Phys. Chem. B 2000, 104, 931-940
931
Femtosecond Optical Switching of Electron Transport Direction in Branched
Donor-Acceptor Arrays
Aaron S. Lukas, Scott E. Miller, and Michael R. Wasielewski*
Department of Chemistry, Northwestern UniVersity, EVanston, Illinois 60208-3113
ReceiVed: October 25, 1999
The branched donor-acceptor triad 4 and tetrad 5 were synthesized to study the possibility of controlling the
direction of electron transfer in a divergent array of electron acceptors using ultrafast laser pulses. Compounds
4 and 5 employ 1,3,5-triaminobenzene as the central branch point. In 4, a 4-(N-piperidinyl)-1,8-naphthaleneimide
electron donor (ANI) was attached to the 1 position and two electron acceptors, 1,8:4,5-naphthalenediimide,
NI, and pyromellitimide, PI, were attached to the 3 and 5 positions of the central benzene ring. In 5, the
terminal end of the PI acceptor is functionalized with an additional NI molecule. Selective excitation of ANI
in 4 and 5 with 400 nm, 130 fs laser pulses results in exclusive formation of NI--ANI+-PI and NI--
ANI+-PI-NI, respectively, with τ ) 115 ps, and a quantum yield of 0.99. Excitation of NI- with 480 nm,
130 fs laser pulses produces the excited doublet state *NI-, which transfers an electron to PI on the second
branch of the benzene ring with a time constant of τ ) 600 fs in 4 and τ ) 750 fs in 5. The overall quantum
yields for the two step process are 0.44 and 0.36 in 4 and 5, respectively. The resultant state NI-ANI+-PI-
in 4 undergoes a charge shift reaction returning to the initial ion pair state NI--ANI+-PI with τ ) 400 ps,
while the corresponding state NI-ANI+-PI--NI in 5 undergoes a charge shift with τ ) 200 ps to yield
NI-ANI+-PI-NI-, which in turn exhibits a 2000 ps lifetime. These results show that once the electron has
been switched to the branch containing the thermodynamically uphill PI acceptor, the electron will cascade
down that branch to other acceptors that are more easily reduced. Model compounds 1-3, which were
synthesized to aid in characterization of the switching dynamics, are also discussed. Photochemical control
of directional charge transport may make it possible to design networks of organic molecules for information
processing.
Introduction
molecules can occur on the femtosecond time scale. Thus, it
may be possible to use them to construct molecular analogues
Branched architectures have been employed for some time
in inorganic semiconductors used in optical waveguides and all-
optical switching devices.1,2 However, these systems rely
principally on nonlinear optical phenomena and changes in
refractive index of the semiconductors to carry out optical
switching. In contrast, chemists have focused on the design of
isolated complexes and single molecules as potential switching
devices. The preparation and study of molecules for use as
molecule-sized switching devices is an active area of research.3-37
Generally, approaches to molecule-sized switches have em-
ployed excitation photoisomerization,10,18,25,30,33-36 energy trans-
fer,8 electron transfer,9,16,17,19,20,23,24,27,35,38-41 changes in chiral-
ity,21,29,42 single-molecule photophysics,22,28,32 photomechanical
behavior,26 and changes in spin state31 as a means of switching.
Only recently have semiconductor design strategies been applied
to organic single molecule electronic devices. The fact that many
of the properties of bulk electronics do not apply to the
molecular regime leads us to explore the fundamental issues
that arise at this architectural length scale. One of these issues
involves controlling the flow of electrons through branch points
between molecules, which is a necessary prerequisite for the
design of networks. The wide versatility of organic electron
transfer systems makes them excellent candidates for these
studies. Energy and electron transfer processes within single
of electronic devices that respond equally rapidly.
Previous work in our laboratory has demonstrated the ability
to control the populations of radical ion intermediates within a
linear donor-acceptor(1)-acceptor(2) array using femtosecond
laser pulses.39 In this paper we report one- and two-pulse
picosecond electron transfer dynamics in two branched donor-
acceptor systems, triad 4 and tetrad 5, which employ a 1,3,5-
triaminobenzene ring as a branch point. Molecules 1-3 serve
as reference compounds that aid in the elucidation of the electron
transfer processes within 4 and 5. Excitation of the 4-(N-
piperidinyl)naphthalene-1,8-imide (ANI) electron donor with an
initial femtosecond laser pulse leads to selective one-electron
reduction of the 1,8:4,5-naphthalenediimide (NI) acceptor, which
forms one of the branches. Subsequent selective excitation of
the corresponding radical anion, NI- with a second femtosecond
laser pulse results in electron transfer from *NI- to the
pyromellitimide (PI) acceptor on the adjacent branch. Thus,
switching of the charge separated ion pair from one branch to
the other is accomplished by application of the second femto-
second laser pulse. Elaboration of each of these branches with
other acceptors that can accept the electron from either NI- or
PI- via a thermal electron transfer reaction can provide extended
pathways for movement of electrons along each of these
branches. The electron transfer dynamics of tetrad 5 illustrate
this behavior. These results open up the possibility of designing
networks of organic molecules in which electron movement
* To whom correspondence should be addressed. E-mail:
wasielew@chem.nwu.edu.
10.1021/jp993774e CCC: $19.00 © 2000 American Chemical Society
Published on Web 01/12/2000

932 J. Phys. Chem. B, Vol. 104, No. 5, 2000
Lukas et al.
CHART 1
through the network can be controlled by the appropriate
sequence of ultrafast laser pulses.
possesses the ANI electron donor and two amines for elaboration
of the two electron accepting branches. Reaction of 11 with an
excess of of N-(n-octyl)-naphthalene-1,8-dicarboxyanhydride-
4,5-dicarboximide or N-(n-octyl)-benzene-1,2-dicarboxyanhy-
dride-4,5-dicarboximide yields the single-acceptor monoamines
12 or 13, respectively. Further reaction of monoamine 13 with
an excess of N-(n-octyl)-naphthalene-1,8-dicarboxyanhydride-
4,5-dicarboximide yields triad 4. Similarly, reaction of mono-
amine 12 with an excess of anhydride 8 yields tetrad 5.
Results and Discussion
Synthesis. The synthesis of compounds 1-5 (Chart 1) follows
procedures similar to those developed previously43 and is
outlined in Scheme 1. 4-Bromonaphthalene-1,8-dicarboxyan-
hydride is condensed with a mole of 1,3-phenylenediamine or
1,3,5-triaminobenzene to produce the corresponding 3-ami-
nophenylcarboximide, 6, or 3,5-diaminophenylcarboximide, 10.
Nucleophilic displacement of the 4-bromo group on the
naphthalene in 6 and 10 is carried out using an excess of
piperidine to yield the corresponding 4-(N-piperidinyl) deriva-
tives, 7 and 11, respectively. The 4-(N-piperidinyl)naphthalene-
1,8-dicarboximide serves as the photoexcited electron donor in
1-5. Condensation of 6 with N-(n-octyl)-naphthalene-1,8-
dicarboxyanhydride-4,5-dicarboximide or N-(n-octyl)-benzene-
1,2-dicarboxy-anhydride-4,5-dicarboximide yields the corre-
sponding single-donor-single-acceptor reference compounds 1
or 2, respectively. The triad reference compound 3 is prepared
by condensation of 6 with anhydride 8, which contains both
the PI and NI electron acceptors. Compound 8 is prepared by
condensation of the known N-amino-N′-(n-octyl)naphthalene-
1,8:4,5-dicarboximide44 with a mole of pyromellitic dianhydride.
The branched triad and tetrad, 4 and 5, respectively, use 1,3,5-
triaminobenzene, 9, as the central branch point. Condensation
of 9 with one mole of 4-bromonaphthalene-1,8-dicarboxyan-
hydride yields diamine 10. Nucleophilic displacement of the
4-bromo group with excess piperidine in NMP yields 11, which
Single-Pulse Electron Transfer Dynamics. The photophys-
ics of the 4-(N-piperidinyl)naphthalene-1,8-imide (ANI) chro-
mophore have been characterized in detail earlier.43 Briefly, the
ground-state optical spectrum of ANI exhibits a broad absorption
at 400 nm, which possesses about 70% charge-transfer char-
acter.38,43 This charge transfer state decays radiatively with an
emission maximum at 500 nm in toluene. The fluorescence
1
quantum yield and lifetime of *ANI in toluene are 0.91 and
8.5 ns, respectively. 1*ANI readily donates or accepts electrons
in the presence of nearby electron acceptors or donors. The one-
electron oxidation and reduction potentials of ANI are 1.20 V
(vs SCE) and -1.41 V, respectively, while the one-electron
reduction potentials of PI and NI are -0.79 V and -0.53 V,
respectively. The energy levels of the ion pair states for
molecules 1-5 were determined using the 2.80 eV energy of
1*ANI, the redox potentials of ANI and the naphthalenediimide
(NI) and pyromellitimide (PI) acceptors, as well as the Cou-
lombic interaction energy between ANI+ and NI- or PI- by
methods given previously.43 The energy levels of the ion pair
states in molecules 1-3 are given in Figure 1, while those for
4 and 5 are presented in Figures 8 and 9.

Optical Switching of Electron Transport Direction
J. Phys. Chem. B, Vol. 104, No. 5, 2000 933
SCHEME 1
Figure 2. Single pulse transient absorption spectra of 1-3 in toluene
at t ) 2 ns after excitation with 400 nm, 130 fs laser pulses. Note that
in 3 the spectral signatures of both NI- and PI- are present.
spectra of the charge separated states, ANI+-NI- in 1 and
ANI+-PI- in 2, which exhibit characteristic optical absorptions
for NI- at 480 and 610 nm45 and for PI- at 720 nm.46 The
transient kinetics in Figure 3 show that ANI+-NI- in 1 and
ANI+-PI- in 2, form with τ ) 115 ps and τ ) 920 ps,
respectively. This suggests that electron transfer constitutes the
major deactivation pathway competing with fluorescence from
1
1*ANI, and that based on the 8.5 ns lifetime of *ANI, the
quantum yield for production of ANI+-NI- in 1 is 0.99, while
that for ANI+-PI- in 2 is 0.90. Charge recombination of
ANI+-NI- within 1 occurs with a time constant of τ ) 150 ns
and ∆G ) -2.4 eV to produce the ground state of 1, while
recombination of ANI+-PI- within 2 occurs with a time
constant of τ ) 10 ns and ∆G ) -0.6 eV to produce primarily
3*ANI. The free energy of the reaction ANI+-NI- f ANI-
NI lies strongly in the Marcus inverted region47 and is thus
Figure 1. Energy level diagrams for (A) 1 and 2 and (B) 3, all in
toluene.
Transient absorption measurements on 1 and 2 show that
direct excitation of the ANI chromophore using 130 fs, 400
1
nm laser pulses results in electron transfer from *ANI to the
adjacent NI and PI acceptors. Figure 2 shows the transient

934 J. Phys. Chem. B, Vol. 104, No. 5, 2000
Lukas et al.
such as toluene, the additional Coulombic energy required to
move the positive and negative charges further apart in the
reaction ANI+-PI--NI f ANI+-PI-NI- cancels most of the
redox potential difference between PI and NI that thermody-
namically favors the formation of ANI+-PI-NI-. Calculations
using the spectroscopic/electrochemical method of Greenfield
et al. show that the ANI+-PI-NI- state lies approximately 0.08
eV below that of ANI+-PI--NI.43 Thus, these ion pair states
are effectively isoenergetic. The similarity between the 28 ns
and 45 ns time constants for charge recombination of ANI-
PI--NI and ANI-PI-NI- within 3 suggests that significant
populations of both PI- and NI- are present during the charge
recombination process.
Figure 3. Transient absorption kinetics for 1 (s) at 480 nm and 2
(- -) at 720 nm in toluene following excitation with a single 400 nm,
130 fs, laser pulse. Inset: transient absorption kinetics for 3 at 480 nm
(s) and 720 nm (- -) in toluene following excitation with a single
400 nm, 130 fs laser pulse. Nonlinear least-squares fits to the data are
also shown.
In the molecular switch prototypes, 4 and 5, the donor and
acceptor groups are spaced evenly about the phenyl junction.
Molecular mechanics calculations yield structures for 4 and 5,
Figure 4, which show that steric interactions result in similar,
substantial dihedral angles between the π systems of ANI, NI,
and PI relative to that of the trisubstituted benzene, as well as
between the PI directly bonded to NI within 5.49 These strong
deviations from an all planar structure substantially reduce the
electronic coupling between the various electron donors and
acceptors within 4 and 5. Distances of approximately 14 Å and
four similar chemical bonds separate the ANI electron donor
from either the NI or PI acceptors attached to the benzene branch
point.49 Thus, the electronic coupling between the donor and
each acceptor should be similar, and the electron transfer rates
3
substantially slower than the reaction ANI+-PI- f *ANI-
PI, which lies in the normal region.
When an N-N bond is used to directly link PI and NI, the
one-electron reduction potential of PI is shifted more positive
by approximately 0.1V.48 This makes the free energy for charge
separation to PI 0.1 eV more negative. Thus, direct attachment
of NI to PI in 3 results in a somewhat faster charge separation
1
reaction *ANI-PI-NI f ANI+-PI--NI with τ ) 700 ps.
The yield of ANI+-PI--NI likewise increases to 0.92. The
time constant for the charge shift reaction ANI+-PI--NI f
ANI+-PI-NI- is τ ) 200 ps. The transient absorption spectrum
of the charge separated intermediate within 3 is shown in Figure
2, while the kinetics for charge separation detected at 480 and
720 nm are shown in the inset to Figure 3. The overall yield
for the two step charge separation is also 0.92, as the charge
recombination reaction ANI+-PI--NI f 3*ANI-PI-NI with
τ ) 10 ns cannot compete kinetically with charge separation to
NI. However, the transient absorption spectrum of 3 shown in
Figure 2 indicates that PI- and NI- are in equilibrium. While
NI is 0.26 V easier to reduce than is PI, in low polarity media
1
from *ANI to either PI or NI in these systems should be
determined primarily by their free energies of reaction. Electron
transfer from 1*ANI to NI is energetically favored over that to
PI by approximately 0.26 eV. Consequently, excitation of ANI
in 4 and 5 with 400 nm laser pulses results exclusively in
formation of ANI+-NI- with τ ) 115 ps and a quantum yield
of 0.99. The charge separation time constants and the yields of
the NI--ANI+-PI and NI--ANI+-PI-NI ion pair states in
4 and 5, respectively, are identical to those observed for
compound 1. There is no spectral evidence for electron transfer
to PI in either molecule, as seen in the inset to Figure 5. The
Figure 4. Structures of 4 and 5 derived from molecular mechanics calculations.

Optical Switching of Electron Transport Direction
J. Phys. Chem. B, Vol. 104, No. 5, 2000 935
Figure 5. Two pump transient absorption spectra of 1 and 4 in toluene.
Inset: single pulse transient absorption spectra of 4 and 5 at t ) 2 ns
after excitation with 400 nm, 130 fs laser pulses.
Figure 8. Energy level diagram for compound 4 in toluene.
Figure 6. Two pulse transient absorption kinetics for 4. Inset: The
formation of 1*ANI at 525 nm (s), and PI- at 720 nm (- -) in
compound 4.
Figure 9. Energy level diagram for compound 5 in toluene.
that may transfer an electron to nearby electron acceptors or to
solvent.50-60 The distinct 480 nm absorption band of NI- allows
for its selective excitation to its lowest excited state *NI- from
which deactivation occurs via competing routes. Excitation of
1 in toluene with a 400 nm laser pulse at t ) 0 ns initiates the
formation of ANI+-NI- (τ ) 115 ps) as in the single-pulse
experiment. Excitation of NI- with a 480 nm laser pulse at t )
2 ns following the 400 nm pulse yields the excited radical anion,
ANI+-*NI-. Depletion of the NI- ground state occurs within
the 180 fs instrument function, and is observed as a bleach at
610 nm.39 Deactivation of *NI- occurs rapidly via charge
1
recombination, reforming *ANI-NI with τ ) 500 fs. This is
1
indicated by the appearance of the *ANI stimulated emission
near 525 nm, Figure 5. There is no spectroscopic evidence for
charge transfer to nearby solvent molecules. The population of
1
1*ANI-NI formed via the reaction ANI+-*NI- f *ANI-
NI subsequently reacts to yield the charge separated state,
ANI+-NI-, again with τ ) 115 ps, monitored by the decays
of the bleaches at 525 and 610 nm. The data are consistent with
a quantum yield near unity for the reaction ANI+-*NI-
f
1*ANI-NI because the lifetime of *NI- itself is 240 ps.61
Excitation of the branched molecule 4 with a single 400 nm,
130 fs laser pulse at t ) 0 produces NI-1*ANI-PI, which
undergoes electron transfer to yield NI--ANI+-PI. The
formation and decay of this ion pair state occurs with the same
kinetics observed in 1. Subsequent excitation of NI- within
NI--ANI+-PI with 480 nm, 130 fs laser pulses at t ) 2 ns
results in the appearance of both the 525 nm stimulated emission
Figure 7. (A) Two pulse transient absorption kinetics for 5. (B) Kinetic
modeling for compound 5 as discussed in the text.
charge recombination reactions of NI--ANI+-PI and NI--
ANI+-PI-NI occur with τ ) 150 ns, and once again agree
very well with that determined for 1. Thus, using a single 400
nm excitation pulse, both 4 and 5 exhibit exclusive transfer of
an electron to the branch that contains the NI acceptor alone.
Two Pulse Electron Transfer Dynamics. Photoexcitation
of radical ions leads to the formation of excited doublet states
1
band from *ANI and the absorption band at 720 nm due to
PI-, Figure 5. This indicates that *NI--ANI+-PI undergoes
deactivation via two competitive pathways. *NI--ANI+-PI
undergoes charge recombination back to the NI-1*ANI-PI

936 J. Phys. Chem. B, Vol. 104, No. 5, 2000
Lukas et al.
excited state similar to the dynamics in dyad 1, or a charge
shift to form NI-ANI+-PI-. While the calculated energy levels
of the NI-1*ANI-PI and NI-ANI+-PI- states are similar,
Figure 8, the electronic coupling between each of them and
*NI--ANI+-PI may be different. Nevertheless, the observed
rates of these competing reactions are similar. The inset to Figure
6 shows that biexponential kinetics are observed for the reactions
*NI--ANI+-PI f NI-1*ANI-PI and *NI--ANI+-PI f
NI-ANI+-PI-. For the charge shift reaction the major kinetic
component occurs with τ ) 600 fs, while a smaller amplitude
τ ) 6 ps component also appears. The fast component is
assigned to the electron transfer reaction, while the second minor
component may be due to either solvent or vibrational relaxation
processes.48 The free energy for each reaction is approximately
1.2 eV, and is near the top of the Marcus parabola (vide infra).47
Thus, the large driving forces for these reactions are most likely
responsible for their ultrafast rates. Small differences in
electronic coupling between the reactant and product states
involved in the two deactivation pathways most likely have a
minor effect on the observed rates. The quantum yield for the
charge shift reaction *NI--ANI+-PI f NI-ANI+-PI- cal-
culated using the measured rate constants for the two competing
deactivation routes is 0.45. Given a 0.99 quantum yield for the
formation of NI--ANI+-PI, the overall quantum yield of NI-
ANI+-PI- formation is 0.44.
of τ ) 225 ps, after which the dynamics switch over to a much
slower process having a time constant of τ ) 2200 ps. The
amplitudes for the two components are 1:2, respectively, and
are dependent upon the equilibrium constant determined by k1
and k-1. A steady-state ensues upon establishment of the
equilibrium between NI-ANI+-PI--NI and NI-ANI+-PI-
NI- with the kinetics at longer times determined by k2. The
same argument can be applied to the kinetics observed at 610
nm. However, the amplitude of the fast decay component
increases relative to that of the slower process to approximately
2:1, respectively. This can be attributed to the additional rate
constant for the forward electron transfer reaction NI-1*ANI-
PI-NI f NI--ANI+-PI-NI, which contributes to the fast
kinetics. The best simulation of the experimentally observed
decay kinetics for the 610 and 720 nm signals is obtained with
k1 ) 5.0 × 109 s-1 (τ ) 200 ps) and k2 ) 5 × 108 s-1 (τ ) 2
ns). The 28 ns and 45 ns time constants for the return of NI-
ANI+-PI--NI and NI-ANI-PI-NI-, respectively, to ground
state indicated in Figure 9 are assumed to be the same as the
those for the corresponding processes within 3. Thus, the results
show that extension of the number of electron acceptors forming
an energetically downhill electron transfer cascade can result
in movement of the electron away from the branch point. Further
adjustments to both the free energies of reaction and the
electronic coupling between the donor and acceptors within these
branched molecular switches will be needed to optimize the
quantum yields for electron transfer down the branch accessed
using the second laser pulse. For example, changing the excited-
state energy of the electron donor group is one way to
accomplish this. We are thus studying alternatives to ANI as
the electron donor.
The subsequent decay of PI- within NI-ANI+-PI- can
occur via two possible routes: charge shift back NI, or direct
charge recombination with ANI+. Charge recombination within
NI-ANI+-PI- should occur in τ ) 10 ns by analogy with 2.
Thus, the observed decay of τ ) 400 ps is due entirely to charge
shift to NI. The population of the NI--ANI+-PI state prior to
the switching event induced by the second laser pulse is fully
restored, as evidenced by the complete recovery of the 610 nm
signal to base line, Figure 6.
Summary and Conclusions
These experiments demonstrate that it is possible to control
the direction of electron flow within a branched donor-acceptor
array on the femtosecond time scale. In triad 4, the initial ion
pair forms over a distance of 14 Å with a quantum yield of
0.99. Excitation of the reduced acceptor group switches the
system over to the second ion pair with a yield of 44% in
toluene. The second state of the switch decays back to the initial
ion pair state and has a lifetime of 400 ps. While the yield of
the switching decreases in tetrad 5 to 36%, the kinetic evidence
suggests that the lifetime of second state of the switch is
increased by about an order of magnitude.
The study of molecular arrays is paramount to developing a
fundamental understanding of the stepwise electron transfer
processes, which someday may prove useful for processing data.
While these studies are most easily done in a fluid solvent, future
studies must eventually adapt these concepts to the solid state
and to surfaces, where the electron transfer dynamics may differ
from those observed in solution. In addition, future research
will emphasize improving the yields for switching the ion pairs,
extension of the electron transfer pathways along each branch,
and strategies for the development of electron transfer networks
based on photon controlled branching.
The structure of compound 5 is very similar to 4 except that
a second NI has been directly linked to PI through an N-N
bond. As discussed for compound 3, this stabilizes PI- by
approximately 0.1 V and significantly affects the switching
dynamics. The time constant for the reaction *NI--ANI+-
PI-NI f NI-ANI+-PI--NI increases to τ ) 750 fs, and the
overall yield of the charge-transfer process also decreases to
0.36. These data suggest that the deactivation of *NI- is actually
slightly in the Marcus inverted region, as an increase in the
driving force decreases the reaction rate somewhat. The transient
absorption kinetics at 525, 610, and 720 nm for compound 5,
are shown in Figure 7A. In compound 5 both the 610 and 720
nm transient absorption changes display biexponential kinetics.
The kinetic analysis is further complicated by the inability to
distinguish between the spectra of the two NI- ion containing
intermediates. Comparing the electron transfer pathways detailed
in Figures 8 and 9, while a single process determines the lifetime
of the NI-ANI+-PI- ion pair in compound 4, there are now
two possible routes for the decay of NI-ANI+-PI--NI in 5.
The dynamics are governed by the rate constants labeled k1,
k-1, and k2 in Figure 9. The rate constant k1 is known to be
(200 ps)-1 from the single pulse studies on 3. However, the
rate constants k-1, and k2 cannot be determined directly.
Therefore, kinetic simulations were performed to estimate k-1
and k2.62 The rate constant k2 for charge shift from the PI-
containing branch back to the branch with NI alone is likely to
decrease from compound 4 to 5 because the free energy of
reaction for the charge-transfer back to the first branch decreases.
The results are shown in Figure 7B. Biexponential kinetics are
observed at 720 nm: an initial fast decay with a time constant
Experimental Section
Syntheses. Proton nuclear magnetic resonance spectra were
recorded on a Gemini-300 NMR spectrometer using TMS as
an internal standard. Laser desorption mass spectra were
obtained with a Kratos MALDI III spectrometer using 2-hy-
droxy-1-naphthoic acid as a matrix. All solvents and regents
were used as received. Column chromatography was performed
using Merck silica gel 60.

Optical Switching of Electron Transport Direction
J. Phys. Chem. B, Vol. 104, No. 5, 2000 937
N-(3-aminophenyl)-4-bromonaphthalene-1,8-dicarboxi-
mide, 6. 4-Bromonaphthalene-1,8-dicarboxylic anhydride (5.0
g, 0.018 mol) was dissolved in 300 mL 95% ethanol. 1,3-
Phenylenediamine (2.0 g, 0.0185 mol) was added and the
solution was bubbled with nitrogen for 10 min. The resulting
solution was refluxed for 2 h, cooled to 0 °C for 24 h, and then
filtered on a Bu¨chner funnel. The powder was rinsed with
ethanol, and dried to yield 6, 3.22 g, 49%. C18H11N2O2Br: mass
spectrum (m/e) 368.3 (calcd 367.2), 1H NMR (δ in CDCl3) 8.69
(d, J ) 7.2 Hz, 1H, 8-naphthyl), 8.625 (d, J ) 8.55 Hz, 1H,
6-naphthyl), 8.45 (d, J ) 7.85 Hz, 1H, 2-naphthyl), 8.073 (d, J
) 7.85 Hz, 1H, 3-naphthyl), 7.88 (d of d, J ) 8.5 Hz, J ) 7.3
Hz, 1H, 7-naphthyl), 7.324 (t, J ) 7.9 Hz, 1H, 5-phenyl), 6.88
(d, J ) 8.1 Hz, 1H, 6-phenyl), 6.69 (d, J ) 7.78 Hz, 1H,
4-phenyl), 6.616 (m, 1H, 2-phenyl), 1.9 (broad singlet, 2H,
3-phenylamine).
8.05 Hz, 1H, 3-naphthyl), 3.75 (t, J ) 7.3 Hz, 2H, N-methylene),
3.28 (m, 4H, piperidine), 1.9 (m, 4H, piperidine), 1.74 (m, 2H,
piperidine), 1.25 (broad singlet, 12H, octyl tail), 0.87 (t, J )
6.3 Hz, 3H, octyl tail methyl).
N-octyl-N′-(N′′-(1,2-Dicarboximide)-4,5-dicarboxyanhy-
dride)-naphthalene-1,2:4,5-dicarboximide, 8. Pyromellitic di-
anhydride (1.0 g, 9.2 mmol) and N-amino-N′-octylnaphthalene-
1,8:4,5-dicarboximide44 (0.755 g, 1.92 mmol) were dissolved
in 15 mL of DMF and heated to 140° for 24 h under a nitrogen
atmosphere. The solvent was stripped and the residue was
purified using silica gel chromatography. Elution with 10%
methanol in methylene chloride resulted in 8 as a sticky solid,
0.955 g, 85% yield. C32H23N3O8: no mass spectrum available.
1H NMR (δ in DMSO) 8.786 (s, 4H, naphthalene diimide),
8.475 (s, 2H, pyromellitc monoanhydride mono-imide), 4.06
(m, 2H, N-methylene), 1.676 (m, 2H, octyl tail), 1.26 (broad
singlet, 10H, octyl tail), 0.859 (t, J ) 6.3 Hz, 3H, octyl tail
methyl).
N-(3-aminophenyl)-4-(N-piperidyl)naphthalene-1,8-dicar-
boximide, 7. Compound 6 (1.0 g, 2.7 mmol) was dissolved in
50 mL of N-methylpyrrolidinone (NMP) along with piperidine
(2.3 g, 0.027 mol). The solution was refluxed for 2 h under a
nitrogen atmosphere. The solvent was stripped on a roughing
pump, and the resulting residue was purified using silica gel
chromatography. Elution with 10% acetone in methylene
chloride v/v, yielded 7 as a yellow powder, 0.55 g, 55%.
C23H21N3O2. Mass spectrum (m/e) 371.6 (calcd 371.4), 1H NMR
(δ in CDCl3) 8.598 (d, J ) 7.28 Hz, 1H, 8-naphthyl), 8.522 (d,
J ) 8.18 Hz, 2-naphthyl), 8.425 (d, J ) 8.5 Hz, 1H, 6-naphthyl),
7.692 (d of d, J ) 8.46 Hz, J ) 8.44 Hz, 1H, 7-naphthyl), 7.30
(t, J ) 7.94 Hz, 1H, 5-phenyl), 7.194 (d, J ) 8.13 Hz, 1H,
3-naphthyl), 6.76 (d, J ) 7.8 Hz, 1H, 6-phenyl), 6.682 (d, J )
7.8 Hz, 1H, 4-phenyl), 6.608 (t, J ) 2.0 Hz, 1H, 2-phenyl),
3.78 (broad singlet, 2H, phenylamine), 3.25 (m, 4H, piperidine),
1.88 (m, 4H, piperidine), 1.71 (m, 2H, piperidine).
Compound 3. Compounds 7 (0.060 g, 0.16 mmol) and 8
(0.224 g, 0.387 mmol) were dissolved in 10 mL of NMP and
heated at 150° for 48 h under a nitrogen atmosphere. This
solution was cooled, taken up in 30 mL of chloroform, and
washed with water (3 × 1000 mL). The organic layer was dried
over anhydrous magnesium sulfate, and stripped of solvent. The
residue was then adsorbed onto silica gel and purified using
silica gel chromatography. The product was eluted with 5%
acetone in methylene chloride (v/v), and the solvent stripped
to yield 3 as a yellow powder, 52 mg, 14%. C55H42N6O10: mass
1
spectrum 947.5 (calcd 947.0), H NMR (δ in CDCl3) 8.85 (s,
4H, naphthalene ring), 8.63 (d, J ) 7.32 Hz, 1H, 8-naphthyl),
8.56 (d, J ) 8.09 Hz, 1H, 2-naphthyl), 8.56 (s, 2H, pyromel-
litimide ring), 8.46 (d, J ) 8.5 Hz, 1H, 6-naphthyl), 7.72 (m,
3H, 7-naphthyl, 6-phenyl, 4-phenyl), 7.63 (m, 1H, 2-phenyl),
7.45 (m, 1H, 5-phenyl), 7.22 (d, J ) 8.26 Hz, 1H, 3-naphthyl),
4.23 (t, J ) 7.57 Hz, 2H, N-methylene), 3.275 (m, 4H, piperidine
ring), 1.92 (m, 4H, piperidine ring), 1.76 (m, 2H, piperidine
ring), 1.25 (broad singlet, 12H, octyl tail), 0.89 (t, J ) 7.0 Hz,
octyl tail methyl).
Compound 1. Compound 7 (0.06 g, 0.162 mmol) was
dissolved in 20 mL of pyridine. N-(n-octyl)-naphthalene-1,8-
dicarboxyanhydride-4,5-dicarboximide43 (0.064 g 0.169 mmol)
was added, and the solution was refluxed for 24 h under a
nitrogen atmosphere. The pyridine was stripped on a rotary
evaporator and the residue was purified using silica gel
chromatography. Elution with 5% acetone in methylene
chloride v/v yielded 1 as a yellow powder, 0.052 g, 43.8%.
C45H40N4O6: mass spectrum (m/e) 731.5 (calcd 732.8), 1H NMR
(δ in CDCl3) 8.807 (d, J ) 3.3 Hz, 4H, naphthalene ring), 8.625
(d, J ) 9.0 Hz, 1H, 8-naphthyl), 8.551 (d, J ) 8.1 Hz, 1H,
2-naphthyl), 8.431 (d, J ) 5.7 Hz,1H, 6-naphthyl), 7.72 (m,
2H, 7-naphthyl, 5-phenyl), 7.485 (m, 2H, 4-phenyl, 6-phenyl),
7.372 (s, 1H, 2-phenyl), 7.204 (d, J ) 8.4 Hz, 1H, 3-naphthyl),
4.2 (t, 2H, N-methylene), 3.26 (m, 4H, piperidine), 1.9 (m, 4H,
piperidine), 1.75 (m, 2H, piperidine), 1.28 (broad singlet, 12H,
octyl tail), 0.88 (t, J ) 6.6 Hz, 3H, octyl tail methyl).
1,3,5-Triaminobenzene, 9. 3,5-Dinitroaniline (1.592 g, 8.7
mmol) was dissolved in 50 mL of ethanol along with 1.5 g 5%
Pd on activated carbon. This solution was deoxygenated by
bubbling with nitrogen and heated to reflux. Hydrazine mono-
hydrate (2.06 g, 41.2 mmol) was added dropwise to the refluxing
solution. The reaction was heated an additional 15 min, then
filtered through Celite 454. The product was eluted from the
Celite using 150 mL of methanol. The product and the unreacted
hydrazine monohydrate were codistilled with 20 mL of DMF
on a rotary evaporator. The sticky black product was used
immediately without further characterization to avoid rapid
decomposition.
Compound 2. Compound 7 (0.103 g, 0.028 mmol) was
dissolved in 25 mL of NMP along with N-(n-octyl)-benzene-
1,2-dicarboxyanhydride-4,5-dicarboximide (0.214 g, 0.65 mmol).
The solution was refluxed for 48 h under a nitrogen atmosphere.
The solvent was stripped using a roughing pump. The residue
was dissolved in methylene chloride and applied to a silica gel
column. Elution with 4% acetone in methylene chloride v/v
yielded 2 as a yellow powder, 0.05 g, 26.4%. C41H38N4O6: mass
N-(3,5-Diaminophenyl)-4-bromonaphthalene-1,8-dicar-
boximide, 10. Compound 9 (1.07 g, 8.7 mmol) was dissolved
in 150 mL of ethanol. 4-Bromo-1,8-naphthalic dianhydride (1.5
g, 5.4 mmol) was added, and the solution was refluxed under
nitrogen for 3.5 h. The reaction was then cooled to 0 °C for at
least 4 h and filtered using a Bu¨chner funnel. The crude product
was adsorbed onto silica gel using methylene chloride and eluted
from a silica gel column using 50% acetone in methylene
chloride v/v. The solvent was stripped to yield 10, 0.867 g, 42%.
1
spectrum 682.6 (calcd 682.8), H NMR (δ in CDCl3) 8.62 (d,
1
J ) 7.32 Hz, 1H, 8-naphthyl), 8.545 (d, J ) 8.13 Hz, 1H,
2-naphthyl), 8.452 (d, J ) 8.44 Hz, 1H, 6-naphthyl), 8.367 (s,
2H, pyromellitimide), 7.715 (d of d, J ) 8.4 Hz, J ) 7.3 Hz,
1H, 7-naphthyl), 7.68 (m, 2H, 4-phenyl, 6-phenyl), 7.592 (t, J
) 1.5 Hz, 1H, 2-phenyl), 7.4 (m, 1H, 5-phenyl), 7.215 (d, J )
C18H12N3O2Br: mass spectrum 348.3 (calcd 348.3), H NMR
(δ in DMSO) 8.61 (d, J ) 7.47 Hz, 1H, 6-naphthyl), 8.584 (d,
J ) 8.56 Hz, 1H, 8-naphthyl), 8.373 (d, J ) 7.85 Hz, 1H,
3-naphthyl), 8.251 (d, J ) 7.89 Hz, 1H, 2-naphthyl), 8.025 (d
of d, J ) 7.37 Hz, J ) 7.29 Hz, 1H, 7-naphthyl), 5.89 (s, 1H,

938 J. Phys. Chem. B, Vol. 104, No. 5, 2000
Lukas et al.
4-phenyl), 5.72 (s, 2H, 2,6-phenyl), 3.48 (broad singlet, 4H,
3,5-phenylamine).
Compound 4. Compound 13 (0.126 g, 0.181 mmol) and
N-(n-octyl)-naphthalene-1,8-dicarboxyanhydride-4,5-dicarbox-
imide (0.07 g, 0.184 mmol) were dissolved in 10 mL of NMP.
The solution was deoxygenated by bubbling with nitrogen and
refluxed under a nitrogen atmosphere for 12 h. The reacted
solution was taken up in 30 mL of chloroform and washed with
2 × 800 mL water. The chloroform solution was dried over
anhydrous MgSO4, and the chloroform was stripped on a rotary
evaporator. The residue was chromatographed on silica gel, and
elution with 5% acetone in methylene chloride v/v yielded 4 as
a yellow powder, 0.052 g, 27%. C63H58N6O10: mass spectrum
N-(3,5-Diaminophenyl)-4-(N-piperidyl)-naphthalene-1,8-
dicarboximide, 11. Compound 10 (1.8 g (4.7 mmol) and
piperidine (2.9 g, 34 mmol) were dissolved in 50 mL of NMP
and refluxed 3-4 h under a nitrogen atmosphere. The solvent
was removed using a roughing pump, and the sticky residue
was taken up in a minimum amount of methylene chloride. The
dissolved residue was added dropwise to 50 mL of petroleum
ether and centrifuged 15 min. The powdery precipitate was
purified using silica gel column chromatography, eluting with
10% acetone in methylene chloride to yield 11 as a yellow
powder, 0.482 g, 26.5%. C23H22N4O2: mass spectrum (m/e)
387.4 (calcd 386.5), 1H NMR (δ in CDCl3) 8.604 (d, J ) 6.69
Hz, 1H, 8-naphthyl), 8.528 (d, J ) 8.1 Hz, 1H, 2-naphthyl),
8.426 (d, J ) 8.5 Hz, 1H, 6-naphthyl), 7.695 (d of d, J ) 7.31
Hz, J ) 8.42 Hz, 1H, 7-naphthyl), 7.197 (d, J ) 8.22 Hz, 1H,
3-naphthyl), 6.10 (m, 1H, 4-phenyl), 6.06 (m, 2H, 2-phenyl,
6-phenyl), 3.67 (broad singlet, 4H, 3,5-phenylamines), 3.242
(m, 4H, piperidine), 1.90 (m, 4H, piperidine), 1.73 (m, 2H,
piperidine).
1
1060.0 (calcd 1059.2), H NMR (δ in CDCl3) 8.815 (d of d, J
) 7.65 Hz, J ) 7.62 Hz, 4H, naphthalene ring), 8.64 (d, J )
7.44 Hz, 1H, 8-naphthyl), 8.56 (d, J ) 8.14 Hz, 1H, 2-naphthyl),
8.49 (d, J ) 8.3 Hz, 1H, 6-naphthyl), 8.38 (s, 2H, pyromelli-
timide), 7.914 (t, J ) 1.87 Hz, 1H, phenyl ring), 7.86 (t, J )
1.91 Hz, 1H, phenyl ring), 7.73 (d of d, J ) 8.3 Hz, J ) 8.42
Hz, 1H, 7-naphthyl), 7.47 (t, J ) 1.83 Hz, 1H, phenyl), 7.23
(d, J ) 8.4 Hz, 1H, 3-naphthyl), 4.21 (J ) 7.4 Hz, 2H,
N-methylene), 3.75 (t, J ) 7.3 Hz, 2H, N-methylene), 3.295
(m, 4H, piperidine), 1.93 (m, 4H, piperidine), 1.25 (broad singlet,
24H, octyl tails), 0.87 (m, 6H, octyl tail methyls).
Compound 12. Compound 11 (0.116 g, 0.297 mmol) was
dissolved in 15 mL of DMF and heated to 130°. N-(n-octyl)-
naphthalene-1,8-dicarboxyanhydride-4,5-dicarboximide (0.12 g,
0.316 mmol) was dissolved in 10 mL of DMF and added
dropwise to the hot solution over the course of 1 h. The resulting
solution was heated overnight under a nitrogen atmosphere. The
DMF was then stripped, and the residue was taken up in
methylene chloride and chromatographed on silica gel. Elution
with 10% acetone in methylene chloride v/v yielded 12 as a
yellow powder, 0.11 g, 50%. C45H41N5O6: mass spectrum (m/
Compound 5. Compounds 12 (0.071 g, 0.095 mmol) and 8
(0.124 g, 0.215 mmol) were dissolved in 5 mL of NMP and
heated to 180° for 28 h under a nitrogen atmosphere. The
solution was cooled, taken up in 30 mL of chloroform, and
washed with water (2 × 1000 mL). The organic layer was dried
over magnesium sulfate, and the solvent was stripped. The
resulting residue was taken up in methylene chloride and
adsorbed onto silica gel. Column chromatography and elution
with 5% acetone in methylene chloride (v/v) gave 5 as a yellow
powder, 30 mg, 24%. C77H62N8O14: mass spectrum (m/e)
1
1
e) 746.9 (calcd 747.4), H NMR (δ in CDCl3) 8.76 (d, J )
1324.6 (calcd 1323.4), H NMR (δ in CDCl3) 8.825 (m, 8H,
4.48 Hz, 4H, naphthalene ring), 8.58 (d, J ) 7.33 Hz, 1H,
8-naphthyl), 8.50 (d, J ) 8.1 Hz, 1H, 2-naphthyl), 8.37 (d, J )
8.2 Hz, 1H, 6-naphthyl), 7.65 (d of d, J ) 8.4 Hz, J ) 7.5 Hz,
1H, 7-naphthyl), 7.16 (d, J ) 8.26 Hz, 1H, 3-naphthyl), 6.69
(m, 2H, 4-phenyl, 6-phenyl), 6.54 (m, 1H, 2-phenyl), 4.18 (m,
2H, N-methylene), 3.22 (m, 4H, piperidine ring), 1.88 (m, 4H,
piperidine ring), 1.7 (m, 2H, piperidine ring), 1.25 (broad singlet,
12H, octyl tail), 0.87 (t, J ) 4.0 Hz, 3H, octyl tail methyl).
naphthalene rings), 8.65 (d, J ) 7.12 Hz, 1H, 8-naphthyl), 8.58
(s, 2H, pyromellitimide ring), 8.57 (d, J ) 8.14 Hz, 1H,
2-naphthyl), 8.44 (d, J ) 9.03 Hz, 1H, 6-naphthyl), 7.95 (t, J
) 1.83 Hz, 1H, phenyl spacer), 7.88 (t, J ) 1.87, 1H, phenyl
spacer), 7.72 (d of d, J ) 8.5 Hz, J ) 7.85 Hz, 1H, 7-naphthyl),
7.50 (t, J ) 1.83 Hz, 1H, phenyl spacer), 7.215 (d, J ) 8.14
Hz, 1H, 3-naphthyl), 4.21 (t, J ) 7.41 Hz, 4H, N-methylenes),
3.26 (m, 4H, piperidine ring), 1.91 (m, 4H, piperidine ring),
1.76 (m, 2H, piperidine ring), 1.25 (broad singlet, 24 H, octyl
tails), 0.88 (t, J ) 7.69 Hz, 6H, octyl tail methyls).
Electrochemistry. Cyclic voltammetry was performed at a
25 µm Pt disk electrode in butyronitrile containing 0.2M tetra-
n-butylammonium hexfluorophosphate. All measurements were
made vs a Ag/AgxO reference electrode, while a Pt wire served
as the counter electrode. Potentials were further referenced using
a ferrocene internal standard.
Compound 13. Compound 11 (0.412 g, 1.1 mmol) was
dissolved in 20 mL of DMF and heated to 130°. N-(n-octyl)-
benzene-1,2-dicarboxyanhydride-4,5-dicarboximide (0.358 g, 1.1
mmol) was dissolved in 15 mL of DMF and added dropwise to
the hot solution over the course of 30 min. The reaction was
heated under an atmosphere of nitrogen an additional 12 h. The
solvent was stripped, and the resulting residue was taken up in
a minimum amount of methylene chloride and precipitated by
slowly dropping the methylene chloride solution into petroleum
ether. The powdery precipitate was purified using silica gel
column chromatography. Elution with 10% acetone in methylene
chloride yielded 13 as a yellow powder (0.3 g, 39%).
C41H39N5O6: mass spectrum (m/e) 698.9 (calcd 697.3), 1H NMR
(δ in CDCl3) 8.458 (d, J ) 6.39 Hz, 1H, 8-naphthyl), 8.381 (d,
J ) 8.13 Hz, 1H, 2-naphthyl), 8.279 (d, J ) 7.53 Hz, 1H,
6-naphthyl), 8.184 (s, 2H, pyromellitimide), 7.549 (d of d, J )
7.31 Hz, J ) 8.3 Hz, 1H, 7-naphthyl), 7.052 (d, J ) 8.14 Hz,
1H, 3-naphthyl), 6.78 (t, J ) 1.93 Hz, 1H, 2-phenyl), 6.731 (t,
J ) 1.69 Hz, 1H, 5-phenyl), 6.523 (t, J ) 1.87 Hz, 1H,
4-phenyl), 3.9 (broad singlet, 2H, 5-phenylamine), 3.60 (t, J )
7.33 Hz, 2H, N-methylene), 3.115 (m, 4H, piperidine), 1.75 (m,
4H, piperidine), 1.58 (m, 2H, piperidine), 1.12 (broad singlet,
12H, octyl tail), 0.72 (t, J ) 4.15 Hz, 3H, octyl tail methyl).
Spectroscopy. UV-vis absorption measurements were made
on a Shimadzu spectrometer (UV160). Steady state fluorescence
measurements were made on a single photon counting fluorim-
eter (PTI). For fluorescence measurements the sample absorption
at the excitation wavelength was kept within 5% of 0.1 to avoid
reabsorption artifacts. The previously determined fluorescence
quantum yield of ANI (Φ ) 0.91 in toluene43) was used as a
standard for all quantum yield measurements.
The apparatus for nano- and microsecond transient absorption
measurements has been described elsewhere.43 Femtosecond
transient absorption measurements were obtained with a pump
wavelengths of 400 nm, 480 nm, or both using a Ti:sapphire
laser system. A Spectra-Physics Millenium frequency-doubled
CW Nd:YAG laser is used to pump a Coherent MIRA
Ti:sapphire oscillator. The 107 fs fwhm laser pulses from the

Optical Switching of Electron Transport Direction
J. Phys. Chem. B, Vol. 104, No. 5, 2000 939
oscillator are stretched to about 200 ps using a four-pass,
reflective, single-grating pulse stretcher and are used to seed a
homemade Ti:sapphire regenerative amplifier based on a 5 mm
path length Ti:sapphire rod and a Medox two-step Pockels cell
and driver. The amplifier is pumped by a homemade intracavity
frequency-doubled (Type II LBO), Q-switched Nd:YAG laser
(Quantronix 116E) operating at 2 kHz (3.5 mJ/pulse at 532 nm).
The amplified Ti:sapphire pulse is recompressed to ∼130 fs by
a four-pass, reflective, single-grating pulse compressor. The
pulse energy is ∼350 µJ/pulse following compression. Four
percent of the recompressed beam is split using a quartz disk,
λ/2 waveplate-polarizer combination to adjust its intensity, and
focused with a 100 mm focal length (fl) lens into a 2 mm thick
sapphire disk. A stable white light continuum is generated by
carefully adjusting the intensity of the pump beam to slightly
above threshold. Continuum intensity fluctuations are typically
under 5%. The white light beam is then split into probe and
reference beams.
magic angle (54.7°) with respect to the probe beam to avoid
anisotropic effects. The probe and reference beams are vertically
displaced, recollimated, and focused into a computer-controlled
monochromator (SPEX 270M) with a 75 mm fl lens. The two
beams are spatially separated after the exit slit and detected with
matched silicon photodiodes. The amplifier outputs were
measured with gated integrators and digitized with a 12 bit A/D
board within a personal computer. The total instrument function
is 180 fs.
Each shot pair (pump on, pump off) results in a single ∆A
measurement. The reference pulse, in addition to being used
for the ∆A calculation, is used to discriminate against probe
pulses of abnormal intensity. Data sets are typically the average
of 10 scans, with each point in the scan being an average of
two hundred laser shots. Samples for the transient absorption
experiments have optical densities at the excitation wavelength
of ∼0.5 (concentration < 10-4 M) for 2 mm path length cells.
Samples are stirred during the experiment using a wire stirrer,
preventing thermal lensing and sample degradation.
A second quartz disk is used to split off another 4% of the
residual 800 nm light to generate another stable white light
continuum pulse in the same manner as described above. This
second white light pulse is used to seed a two stage optical
parametric amplifier (OPA).63 The remaining intense 800 nm
beam is telescoped to a beam size of 1 mm and frequency
doubled in a 2 mm thick type I LBO crystal, providing 110 µJ
pulses at 400 nm. The 400 nm beam is split using a λ/2
waveplate-polarizer pair, with approximately 10 µJ diverted
to pump the first stage of the OPA. The 400 nm beam and the
white light continuum are focused with separate 30 mm fl lenses
into a 2 mm type II â-Barium Borate (BBO) crystal. Spatial
and temporal overlap of the beams is achieved by carefully
adjusting a variable delay line and pointing of the beams.
Energies of 0.3-0.4 µJ/pulse in the signal beam are achieved
in the first stage. Amplification of parametric light from the
first stage is achieved by overlapping the beam from the first
stage with the remaining 95 µJ of 400 nm light in a 2 mm type
I BBO crystal. The amplified parametric pulse (signal only, idler
is rejected using cutoff filters) is typically 9-10 µJ/pulse, with
a wavelength tuning range of 460-740 nm. The residual 400
nm light following parametric generation, about 20 µJ/pulse, is
recollimated with a 500 mm fl lens. The parametric beam is
collimated with a 400 mm fl lens and sent down a 1 m delay
line which introduces a delay between it and the 400 nm beam.
A dichroic mirror is used to spatially overlap the two pump
beams before they are both sent down a 1.3 m variable delay
line fitted with a retroreflector. Delays of 6 ns between the pump
and probe are possible with 6.6 fs resolution.
Acknowledgment. This work was supported by the National
Science Foundation (Grant CHE-9732840).
References and Notes
(1) Tajima, K. Jpn. J. Appl. Phys. 1993, 32, L1746-L1749.
(2) Nakamura, S.; Tajima, K.; Sugimoto, Y. Appl. Phys. Lett. 1994,
65, 283-285.
(3) Aviram, A.; Ratner, M. A. Chem. Phys. Lett. 1974, 29, 277-83.
(4) Hopfield, J. J.; Onuchic, J. N.; Beratan, D. N. Science 1988, 241,
817.
(5) Parthenopoulos, D. A.; Rentzepis, P. M. Science 1989, 245, 843-
5.
(6) Garnier, F.; Horowitz, G.; Peng, X.; Fichou, D. AdV. Mater. 1990,
2, 592-4.
(7) Waldeck, D. H.; Beratan, D. N. Science 1993, 261, 576-7.
(8) Wagner, R. W.; Lindsey, J. S. J. Am. Chem. Soc. 1994, 116, 9759-
60.
(9) Wagner, R. W.; Lindsey, J. S.; Seth, J.; Palaniappan, V.; Bocian,
D. F. J. Am. Chem. Soc. 1996, 118, 3996-7.
(10) Willner, I.; Willner, B. AdV. Mater. 1997, 9, 351-355.
(11) Arrhenius, T. S.; Blanchard-Desce, M.; Dvolaitzky, M.; Lehn, J.
M.; Malthete, J. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 5355-9.
(12) Carter, F. L. Molecular Electronic DeVices I; Marcel Dekker: New
York, 1982.
(13) Carter, F. L. Molecular Electronics II; Marcel Dekker: New York,
1987.
(14) Carter, F. L.; Siatkowski, R. E.; Wohltjen, H. Molecular Electronic
DeVices; North-Holland: Amsterdam, 1988.
(15) Pearson, D. L.; Schumm, J. S.; Tour, J. M. Macromolecules 1994,
27, 7, 2348-50.
(16) Tolbert, L. M. Acc. Chem. Res. 1992, 25, 561-8.
(17) DelMedico, A.; Fielder, S. S.; Lever, A. B. P.; Pietro, W. J. Inorg.
Chem. 1995, 34, 1507-13.
(18) Birge, R. R. Sci. Am. 1995, 272, 90-5.
(19) Tamai, N.; Saika, T.; Shimidzu, T. J. Phys. Chem. 1996, 100, 4689.
(20) Kenny, P. W.; Miller, L. L. J. Chem. Soc., Chem. Commun. 1988,
84.
For single pump excitation experiments the OPA beam is
blocked and a mechanical chopper synchronized to one-half the
repetition rate of the laser is used to block every other 400 nm
pulse. However, when two pump pulses are used to excite the
sample the first pulse occurs at 400 nm, and is not chopped.
The second laser pulse is tuned to 480 nm and is chopped at
one-half the repetition rate of the laser. This excitation scheme
makes the probed absorption signal sensitive only to changes
in state population induced by the second laser pulse.39 The
arrival time of the probe pulse is fixed, while the arrival of the
two pump pulses can be varied relative to it and one another.
(21) Huck, N. P. M.; Jager, W. F.; de Lang, B.; Feringa, B. L. Science
1996, 273, 1686-1688.
(22) Moerner, W. E.; Plakhotnik, T.; Irngartinger, T.; Croci, M.; Palm,
V.; Wild, U. P. J. Phys. Chem. 1994, 98, 7382-9.
(23) O’Neil, M. P.; Niemczyk, M. P.; Svec, W. A.; Gosztola, D.; Gaines,
G. L. I.; Wasielewski, M. R. Science 1992, 257, 63-5.
(24) de Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Huxley,
A. J. M.; McCoy, C. P.; Rademacher, J. T.; Rice, T. E. Chem. ReV. 1997,
97, 1515-1566.
(25) Archut, A.; Vogtle, F.; De Cola, L.; Azzellini, G. C.; Balzani, V.;
Ramanujam, P. S.; Berg, R. H. Chem. Eur. J. 1998, 4, 699-706.
(26) Balzani, V.; Gomez-Lopez, M.; Stoddart, J. F. Acc. Chem. Res.
1998, 31, 405-414.
(27) Beyeler, A.; Belser, P.; De Cola, L. Angew. Chem., Int. Ed. Engl.
1998, 36, 22779-2781.
(28) Dickson, R. M.; Cubitt, A. B.; Tsien, R. Y.; Moerner, W. E. Nature
(London) 1997, 388, 355-358.
(29) Feringa, B. L.; Huck, N. M. P.; Schoevars, A. M. AdV. Mater. 1996,
8, 681-684.
The pump beams are focused onto the sample using a 400
mm fl lens. The probe beam is focused separately using a 400
mm fl lens. The pump energies are attenuated using two λ/2
waveplate-polarizer combinations, one for each beam. Typical
energies of each pump pulse are 1-1.5 µJ/pulse incident on
the sample. The polarization of the pump beams are set at the

940 J. Phys. Chem. B, Vol. 104, No. 5, 2000
Lukas et al.
(30) Gobbi, L.; Seiler, P.; Diederich, F. Angew. Chem., Int. Ed. Engl.
1999, 38, 674-678.
(47) Marcus, R. A. J. Chem. Phys. 1956, 24, 966.
(48) Wiederrecht, G. P.; Niemczyk, M. P.; Svec, W. A.; Wasielewski,
M. R. J. Am. Chem. Soc. 1996, 118, 81-8.
(49) Ion pair distances were estimated from structures calculated using
a modified MM2 force field and AM1 MO calculations performed within
Hyperchem (Hypercube, Waterloo, Ontario).
(50) Fox, M. A. Chem. ReV. 1979, 79, 253-73.
(51) Scaiano, J. C.; Johnston, L. J.; McGimpsey, W. G.; Weir, D. Acc.
Chem. Res. 1988, 21, 22.
(52) Eriksen, J. In Photoinduced Electron Transfer; Fox, M. A., Chanon,
M., Eds.; Elsevier: Amsterdam, 1988; Vol. A, pp 391-408.
(53) Chervonnyi, V. S.; Sinyakov, G. N.; Gadonas, R.; Krasauskas, V.;
Pelakauskas, A.; Maslov, V. G.; Prokof’eva, T. P. Khim. Fiz. 1990, 9, 340-
6.
(54) Filatov, I. V.; Chirvonyi, V. S.; Sinyakov, G. N. Opt. Spektrosk.
1994, 77, 386-8.
(55) Filatov, I. V.; Chirvony, V. S.; Sinyakov, G. N. Proc. SPIE-Int.
Soc. Opt. Eng. 1995, 2370, 95-8.
(56) Ishida, A.; Fukui, M.; Ogawa, H.; Tojo, S.; Majima, T.; Takamuku,
S. J. Phys. Chem. 1995, 99, 10808-14.
(57) Fujita, M.; Ishida, A.; Majima, T.; Takamuku, S. J. Phys. Chem.
1996, 100, 5382-7.
(58) Majima, T.; Fukui, M.; Ishida, A.; Takamuku, S. J. Phys. Chem.
1996, 100, 8913-19.
(59) Gumy, J.-C.; Vauthey, E. J. Phys. Chem. A 1997, 101, 8575-8580.
(60) Cook, A. R.; Curtiss, L. A.; Miller, J. R. J. Am. Chem. Soc. 1997,
119, 5729-5734.
(31) Gu, Z. Z.; Sato, O.; Iyoda, T.; Hashimoto, K.; Fujishima, A. Chem.
Mater. 1997, 9, 1092-1097.
(32) Kulzer, F.; Matzke, R.; Braeuchle, C.; Basche, T. J. Phys. Chem.
A 1999, 103, 2408-2411.
(33) Owrutsky, J. C.; Nelson, H. H.; Baronavski, A. P.; Kim, O. K.;
Tsivgoulis, G. M.; Gilat, S. L.; Lehn, J. M. Chem. Phys. Lett. 1998, 293,
555-563.
(34) Schoevaars, A. M.; Kruizinga, W.; Zijlstra, R. W. J.; Veldman,
N.; Spek, A. L.; Feringa, B. L. J. Org. Chem. 1997, 62, 4943-4948.
(35) Tsivgoulis, G. M.; Lehn, J. M. AdV. Mater. 1997, 9, 39-42.
(36) Weber, C.; Rustemeyer, F.; Duerr, H. AdV. Mater. 1998, 10, 1348-
1351.
(37) Willner, I.; Doron, A.; Katz, E. J. Phys. Org. Chem. 1998, 11, 546-
560.
(38) de Silva, A. P.; Gunaratne, H. Q. N.; Habib-Jiwan, J. L.; McCoy,
C. P.; Rice, T. E.; Soumillion, J. P. Angew. Chem., Int. Ed. Engl. 1995, 34,
1728-1731.
(39) Debreczeny, M. P.; Svec, W. A.; Marsh, E. M.; Wasielewski, M.
R. J. Am. Chem. Soc. 1996, 118, 8174-8175.
(40) Debreczeny, M. P.; Svec, W. A.; Wasielewski, M. R. Science 1996,
274, 584-587.
(41) Gosztola, D.; Niemczyk, M. P.; Wasielewski, M. R. J. Am. Chem.
Soc. 1998, 120, 5118-5119.
(42) Feringa, B. L.; Schoevaars, A. M.; Jager, W. F.; De Lange, B.;
Huck, N. P. M. Enantiomer 1996, 1, 325-335.
(43) Greenfield, S. R.; Svec, W. A.; Gosztola, D.; Wasielewski, M. R.
J. Am. Chem. Soc. 1996, 118, 6767-6777.
(61) Gosztola, D.; Wasielewski, M. R. Unpublished results.
(62) The first-order system of differential equations was solved using
the fourth order Runge-Kutta method within Mathcad 7.0 Professional
(MathSoft, Inc., USA).
(44) Debreczeny, M. P.; Svec, W. A.; Wasielewski, M. R. New J. Chem.
1996, 20, 815-828.
(45) Viehbeck, A.; Goldberg, M. J.; Kovac, C. A. J. Electrochem. Soc.
1990, 137, 1460-1466.
(46) Osuka, A.; Nagata, T.; Maruyama, K.; Mataga, N.; Asahi, T.;
Yamazaki, I.; Nishimura, Y. Chem. Phys. Lett. 1991, 185, 88-94.
(63) Greenfield, S. R.; Wasielewski, M. R. Opt. Lett. 1995, 20, 1394-
6.