Photochromic control of photoinduced electron transfer. Molecular double-throw switch

Article abstract of DOI:10.1021/ja044128i

A molecular double-throw switch that employs a photochromic moiety to direct photoinduced electron transfer from an excited state donor down either of two pathways has been prepared. The molecular triad consists of a free base porphyrin (P) linked to both a C₆₀ electron acceptor and a dihydroindolizine (DHI) photochrome. Excitation of the porphyrin moiety of DHI-P-C₆₀ results in photoinduced electron transfer with a time constant of 2.3 ns to give the DHI-P^.+-C₆₀^.- charge-separated state with a quantum yield of 82%. UV (366 nm) light photoisomerizes the DHI moiety to the betaine (BT) form, which has a higher reduction potential than DHI. Excitation of the porphyrin of BT-P-C₆₀ is followed by photoinduced electron transfer with a time constant of 56 ps to produce BT^.--P^.+-C₆₀ in 99% yield. Isomerization of BT-P-C₆₀ back to DHI-P-C₆₀ may be achieved with visible light, or thermally. Thus, photoinduced charge separation originating from the porphyrin is reversibly directed down either of two different pathways by photoisomerization of the dihydroindolizine. The switch may be cycled many times.

Full text of DOI:10.1021/ja044128i

Published on Web 02/04/2005
Photochromic Control of Photoinduced Electron Transfer.
Molecular Double-Throw Switch
Stephen D. Straight, Joakim Andre´asson, Gerdenis Kodis, Ana L. Moore,*
Thomas A. Moore,* and Devens Gust*
Contribution from the Department of Chemistry and Biochemistry, Center for the Study of Early
EVents in Photosynthesis, Arizona State UniVersity, Tempe, Arizona 85287
Received September 27, 2004; E-mail: gust@asu.edu
Abstract: A molecular double-throw switch that employs a photochromic moiety to direct photoinduced
electron transfer from an excited state donor down either of two pathways has been prepared. The molecular
triad consists of a free base porphyrin (P) linked to both a C60 electron acceptor and a dihydroindolizine
(DHI) photochrome. Excitation of the porphyrin moiety of DHI-P-C60 results in photoinduced electron
transfer with a time constant of 2.3 ns to give the DHI-P•+-C60•- charge-separated state with a quantum
yield of 82%. UV (366 nm) light photoisomerizes the DHI moiety to the betaine (BT) form, which has a
higher reduction potential than DHI. Excitation of the porphyrin of BT-P-C60 is followed by photoinduced
electron transfer with a time constant of 56 ps to produce BT•--P•+-C60 in 99% yield. Isomerization of
BT-P-C60 back to DHI-P-C60 may be achieved with visible light, or thermally. Thus, photoinduced charge
separation originating from the porphyrin is reversibly directed down either of two different pathways by
photoisomerization of the dihydroindolizine. The switch may be cycled many times.
Introduction
electrical analogue of these molecular switches is the simple
single-throw switch, in which current is passed in the on
Photoinduced electron transfer is the basis of photosynthetic
conversion of light energy into useful electrochemical potential
and is important in a variety of technological applications. The
promise of using such a process in molecular-scale optoelec-
tronics has prompted researchers to develop ways to control
photoinduced electron transfer using light or other inputs.
Photochromic molecules, which can be isomerized between two
thermally stable molecular structures using light, are good
candidates for light-activated control units in molecule-based
devices of various types.1-6 Recently, we have demonstrated
on and off switching of photoinduced electron transfer in
molecular dyads and triads by using photochromes to reversibly
quench excited singlet states via energy transfer,7,8 reversibly
alter reduction potentials of electron acceptors,9 and reversibly
change oxidation potentials of electron donor moieties.10 The
position, and the circuit is interrupted in the off position. In the
present work, we have investigated a method for using light
excitation to direct electron transfer from an excited singlet state
of a donor down either of two different pathways. Such molec-
ular switches are analogues of the electrical double-throw switch,
which connects an input to either of two different circuits. In
other approaches to bidirectional photoinduced electron transfer,
Wasielewski and co-workers have reported a system in which
the initial photoinduced electron-transfer always occurs along
the same pathway, but a second femtosecond laser pulse initiates
a charge-shift reaction that moves the electron to another
pathway.11 The direction of a photocurrent has recently been
controlled by selective excitation of either of two chromophores
in separate molecules tethered to a gold surface by polypep-
tides.12 The approach used here employs a photochromic dihy-
droindolizine (DHI) covalently linked to a porphyrin (P), which
also bears a fullerene moiety (C60) to form triad 1 (Chart 1).
Since our first report of photoinduced electron transfer in a cova-
lently linked porphyrin-fullerene dyad,13 many researchers have
demonstrated photoinduced electron transfer from the first ex-
cited singlet state of a porphyrin to an attached fullerene. Such
transfer is characterized by rapid charge separation and relatively
slow charge recombination. Also, we have recently reported the
preparation and study of DHI-P dyad molecular switch 2.9 In
(1) Katz, E.; Willner, B.; Willner, I. Biosens. Bioelectron. 1997, 12, 703-
719.
(2) Mitchell, R. H.; Ward, T. R.; Chen, Y.; Wang, Y.; Weerawarna, S. A.;
Dibble, P. W.; Marsella, M. J.; Almutairi, A.; Wang, Z. Q. J. Am. Chem.
Soc. 2003, 125, 2974-2988.
(3) Willner, I.; Doron, A.; Katz, E. J. Phys. Org. Chem. 1998, 11, 546-560.
(4) Fernandez-Acebes, A.; Lehn, J.-M. ChemistrysEur. J. 1999, 5, 3285-
3292.
(5) Norsten, T. B.; Branda, N. R. J. Am. Chem. Soc. 2001, 123, 1784-1785.
(6) Walz, J.; Ulrich, K.; Port, H.; Wolf, H. C.; Wonner, J.; Effenberger, F.
Chem. Phys. Lett. 1993, 213, 321-324.
(7) Liddell, P. A.; Kodis, G.; Moore, A. L.; Moore, T. A.; Gust, D. J. Am.
Chem. Soc. 2002, 124, 7668-7669.
(8) Bahr, J. L.; Kodis, G.; de la Garza, L.; Lin, S.; Moore, A. L.; Moore, T.
A.; Gust, D. J. Am. Chem. Soc. 2001, 123, 7124-7133.
(9) Terazono, Y.; Kodis, G.; Andre´asson, J.; Jeong, G.; Brune, A.; Hartmann,
T.; Du¨rr, H.; Moore, A. L.; Moore, T. A.; Gust, D. J. Phys. Chem. B 2004,
108, 1812-1814.
(10) Liddell, P. A.; Kodis, G.; Andre´asson, J.; de la Garza, L.; Bandyopadhyay,
S.; Mitchell, R. H.; Moore, T. A.; Moore, A. L.; Gust, D. J. Am. Chem.
Soc. 2004, 126, 4803-4811.
(11) Lukas, A. S.; Miller, S. E.; Wasielewski, M. R. J. Phys. Chem. B 2000,
104, 931-940.
(12) Yasutomi, S.; Morita, T.; Imanishi, Y.; Kimura, S. Science 2004, 304,
1944-1947.
(13) Liddell, P. A.; Sumida, J. P.; Macpherson, A. N.; Noss, L.; Seely, G. R.;
Clark, K. N.; Moore, A. L.; Moore, T. A.; Gust, D. Photochem. Photobiol.
1994, 60, 537-541.
9
10.1021/ja044128i CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 2717-2724
2717

A R T I C L E S
Chart 1
Straight et al.
Chart 2
the closed, spirocyclic dihydroindolizine form 2c (Chart 1),
excitation of the porphyrin moiety in the 2-methyltetrahydro-
furan solution is followed by normal deactivation of the excited
singlet state via the usual photophysical processes of internal
conversion, intersystem crossing, and fluorescence (τ ) 11.5
ns). The porphyrin excited state is not perturbed by the attached
dihydroindolizine. Irradiation of DHI-P with UV (366 nm) light
photoisomerizes the DHI to the open, betaine form BT (2o,
analogous to 1o), which is much more easily reduced than the
DHI form. Excitation of the porphyrin moiety of BT-P is
followed by rapid (τ ) 50 ps) photoinduced electron transfer
to the BT, yielding a short-lived BT•--P•+ charge-separated
state. Therefore, we reasoned that if a fullerene was covalently
linked to the porphyrin of a molecule like 2 in such a way that
photoinduced electron transfer from the porphyrin first excited
singlet state to the C60 occurred much more slowly than electron
transfer to BT, a photochromically controlled molecular double-
throw switch could be prepared. A recent report14 suggested
that the linkage shown in 1 should provide a suitable transfer
rate, and this has now been verified by study of model P-C60
dyad 9. Indeed, the double-throw switch does perform as
envisioned, as described next.
Chart 3
Results
Synthesis. Synthesis began with the preparation of previously
known porphyrin 4 from 3 (Chart 2).15 The carboxylic acid
group of 4 was protected as the t-butyl ester (5), and the amino
group was coupled to pyridazine 4-carboxylate using 1-[3-
(dimethylamino)propyl]-3-ethylcarbodiimide and 4-(dimethyl-
amino)pyridine to obtain 6. This precursor was allowed to react
with spiro[fluorene-9,3′-[3H]pyrazole]-4′,5′-dicarbonitrile under
UV irradiation to yield DHI-P dyad 8 (Chart 3) and an isomer
resulting from reaction at the second pyridazine nitrogen, which
were separated chromatographically. Treatment of 8 with
triethylsilane and trifluoroacetic acid removed the protecting
group from the carboxylic acid. Meanwhile, fullerene precursor
7 was deprotected by treatment with boron tribromide. The
fullerene amine and porphyrin acid were coupled using 1,3-
dicyclohexylcarbodiimide to give triad 1. The model porphyrin-
fullerene dyad 9 was prepared using similar methods. The
details and characterization data are given in the Experimental
Procedures.
Photochemistry. We will first describe the photochemistry
of model P-C60 dyad 9 and model DHI-P 2 and then consider
the results for triad 1.
P-C60 Dyad 9. The absorption spectrum of 9 in the
2-methyltetrahydrofuran solution is shown in Figure 1. Por-
phyrin absorption maxima are observed at 418 (Soret), 482, 514,
547, 593, and 648 nm. The fullerene absorbs very weakly
throughout the visible, with its longest wavelength absorption
(14) Tkachenko, N. V.; Vehmanen, V.; Nikkanen, J.; Yamada, H.; Imahori, H.;
Fukuzumi, S.; Lemmetyinen, H. Chem. Phys. Lett. 2002, 366, 245-252.
(15) Liddell, P. A.; Kodis, G.; de la Garza, L.; Bahr, J. L.; Moore, A. L.; Moore,
T. A.; Gust, D. HelV. Chim. Acta 2001, 84, 2765-2783.
9
2718 J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005

Photochromic Control of Photoinduced Electron Transfer
A R T I C L E S
Figure 2. Transient absorption data for P-C60 dyad 9 in 2-methyltetrahy-
drofuran following excitation at 650 nm. (a) Rise and decay of the transient
absorption at 1010 nm of the fullerene radical anion of P•+-C60•- (O) and
fit to two exponential components with lifetimes of 2.0 and 4 ns (s). The
prompt rise occurs with the excitation pulse and is ascribed to the formation
of excited singlet states. The inset shows the spectrum taken 3.0 ns after
excitation. (b) Decay-associated spectra for the same solution and measured
in the visible region. The two components have lifetimes of 2.0 ns (s) and
4 ns (- - - -). The spectral shapes near 650 nm are somewhat distorted due
to the excitation pulse.
Figure 1. Absorption spectrum of P-C60 dyad 9 (s) in 2-methyltetrahy-
drofuran and fluorescence spectra of 9 (O) and model porphyrin 3 (0) with
the same absorbance at the 590 nm excitation wavelength.
band at 704 nm. These absorption bands are virtually unchanged
from those of model porphyrins such as 3 and model fullerenes
such as 7.
The fluorescence spectra of dyad 9 and model porphyrin 3
in 2-methyltetrahydrofuran, both solutions having the same
absorbance at the excitation wavelength of 590 nm, are also
shown in Figure 1. The emission spectra of 9 and 3 are of
virtually identical shape and are typical of free base porphyrins
of this type, with maxima at 650 and 720 nm. However, the
emission from the dyad is quenched relative to 3 by a factor of
about 6. This quenching is ascribed to photoinduced electron
transfer from the porphyrin to the fullerene to generate a P•+-
with characteristic absorption of the porphyrin radical cation
around 700 nm, corresponds to the decay of P•+-C60•-. (The
short time window available in the experiment limits the
precision of this time constant.)
DHI-P Dyad 2. The properties of model dyad 2 have been
reported previously9 and are summarized here for use in later
discussions. The absorption spectrum of DHI-P (2c) in 2-meth-
yltetrahydrofuran in the visible region is characteristic of an
unperturbed tetra-arylporphyrin with maxima at 419, 483, 515,
550, 593, and 650 nm. The emission from DHI-1P is also
unperturbed, with maxima at 655 and 721 nm and an excited-
state lifetime of 11.5 ns. Thus, the DHI moiety has no significant
effect upon the porphyrin excited singlet state.
When 2 is irradiated at 366 nm, DHI-P dyad 2c is
photoisomerized to the BT-P dyad (designated 2o). The
absorption spectrum of 2o shows the broad betaine absorption
underlying the porphyrin Q-bands. Porphyrin emission is
observed, but it is very strongly quenched relative to 2c.
Transient absorption experiments showed that the quenching
was due to photoinduced electron transfer from the porphyrin
first excited singlet state to the betaine to yield a short-lived
BT•--P•+ charge-separated state. The rate constant for photo-
induced electron transfer is 2.0 × 1010 s-1 (τ ) 50 ps, Φ )
1.0), and the rate constant for recombination is 3.4 × 1011 s-1
(τ ) 2.9 ps). The BT-P isomer 2o may be isomerized back to
the DHI form 2c by irradiation at wavelengths g590 nm (Φ
∼0.02),9 or thermally.
•-
C60 charge-separated state.
To more fully investigate the formation of the charge-
separated state, time-resolved fluorescence experiments were
undertaken in 2-methyltetrahydrofuran. A solution (∼1 × 10-5
M) of 3 was excited at 590 nm, and emission was measured at
720 nm. The data were fitted by a single-exponential decay with
a lifetime of 10.1 ns (ø2 ) 1.12). The data for dyad 9 required
two exponentials, with time constants of 2.0 ns (99.4%) and
11.0 ns (0.6%) (ø2 ) 1.09). The low-amplitude component with
the 11.0 ns lifetime is ascribed to an impurity. Thus, the time-
resolved data confirm the substantial quenching of the porphyrin
first excited singlet state by the attached fullerene.
The charge separation process was also examined using
transient absorption spectroscopy. Figure 2 shows the results
of excitation of the porphyrin moiety of 9 with ∼100 fs laser
pulses at 650 nm. The inset in Figure 2a shows the transient
absorption spectrum in the 1000 nm region taken 3.0 ns after
the laser pulse. The transient has a maximum at about 1010
nm that is characteristic of the fullerene radical anion of the
DHI-P-C60 Triad 1. With the results for 2 and 9 as
background, we turn our attention to DHI-P-C60 triad 1. Figure
3 shows the absorption spectra of 1 in 2-methyltetrahydrofuran.
The DHI-P-C60 form of the triad (1c), wherein the dihydroin-
dolizine moiety is in the spirocyclic form, resembles the
spectrum of model porphyrin 3 in the visible region, with bands
at 419 (Soret), 483, 515, 550, 594, and 649 nm. As is the case
for dyad 9, the fullerene moiety absorbs throughout the visible,
with its long wavelength maximum at 703 nm. Thus, the DHI
and fullerene moieties do not perturb the shape of the porphyrin
spectrum. After irradiation with UV light at 366 nm, the
P•+-C60 state. The kinetics at this wavelength, also shown
•-
in Figure 2a, was fitted with two exponential components
following the laser pulse: a 2.0 ns rising component and a 4 ns
decay. The 2.0 ns component is ascribed to formation of P•+-
•-
C60 and the 4 ns component to decay of this species. This
interpretation is consistent with the decay-associated spectra
obtained in the visible region and shown in Figure 2b. The
spectrum with the 2.0 ns time constant indicates decay of the
porphyrin first excited singlet state and concurrent formation
of the P•+-C60 charge-separated state. The 4 ns component,
•-
9
J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005 2719

A R T I C L E S
Straight et al.
Figure 4. (a) Transient absorption results at 1000 nm for triad 1 following
excitation of a 2-methyltetrahydrofuran solution at 650 nm with ∼100 fs
laser pulses. Data are for closed, DHI-P-C60 triad 1c (9) and open,
BT-P-C60 triad 1o (O). During the experiments on 1o, the sample was
continuously irradiated at 366 nm to prevent any net photoisomerization to
the closed form. The solid lines represent kinetic fits to the data. Please
note the break in the time axis. (b) Transient absorption spectra taken 1.0
ns after excitation for the closed, DHI-P-C60 triad (- - - -) and 25 ps after
excitation for the open, BT-P-C60 triad (s).
Figure 3. Absorption spectra of a sample of 1 in the DHI-P-C60 closed
form 1c (- - - -) and after conversion to a photostationary state consisting
mainly of the BT-P-C60 open form 1o (s). The solvent was 2-methyltet-
rahydrofuran.
spectrum changes as shown in Figure 3. This is due to
photoisomerization of the dihydroindolizine to yield the triad
in the betaine form 1o. The BT-P-C60 triad has absorption
maxima at 419, 485, 517, 552, 589, and 649 nm, plus the
fullerene absorption mentioned previously. The change in
appearance is due to the grow-in of the broad betaine absorption
in the 550 nm region.9 The usual porphyrin absorption is
superimposed upon this band. Thus, photoisomerization of the
dihydroindolizine still occurs when it is linked to the porphyrin-
fullerene dyad. This is vital to the proposed function of the
switch molecule. The 366 nm irradiation creates a photosta-
tionary state; in deuteriochloroform, the ratio of the DHI form
to the BT form in the model dihydroindolizine spiro[9H-
fluorene-9,5′(4′aH)-pyrrolo[1,2-b]pyridazine]-6′,7′-dicarboni-
trile is 15:85, as determined by 1H NMR experiments on ∼0.02
M solutions. No line broadening or other effects suggestive of
aggregation were noted in the NMR spectra. The molecule may
be returned virtually entirely to the DHI form by irradiation
with visible light g590 nm, or thermally.
The kinetics of the photoisomerization of 1 was investigated.
When DHI-P-C60 triad 1c is irradiated at 366 nm with a light
intensity of 1.5 mW/cm2, the photostationary state consisting
mainly of 1o is reached with a time constant of 17 s. For the
closing of BT-P-C60 triad 1o, isomerization under visible
irradiation (g590 nm, ∼33 mW/cm2) occurs with a lifetime of
2400 s at 29 °C. At this temperature, the time constant for
thermal closing of BT-P-C60 is 4000 s. Thus, at ambient
temperatures and with this light intensity, about 40% of the
closing of the BT is photochemical, and 60% is thermally driven.
The porphyrin fluorescence of DHI-P-C60 is comparable
in intensity to that from model porphyrin-fullerene dyad 9, but
when the molecule is switched to the BT-P-C60 form 1o, the
fluorescence is very strongly quenched. Time-resolved fluores-
cence measurements in 2-methyltetrahydrofuran yielded data
for 1c that were fitted (ø2 ) 1.08) with two exponential
components with time constants of 1.9 ns (86%) and 8.3 ns
(14%). The 1.9 ns component is ascribed to the triad and the
8.3 ns component to a porphyrin impurity or decomposition
product. After the molecule was opened to the BT-P-C60
form 1o, fitting the data (ø2 ) 1.17) required components of
60 ps (76%), 1.7 ns (18%), and 9.4 ns (8%). The 60 ps
component is due to BT-P-C60, the 1.7 ns component to
residual DHI-P-C60 in the photostationary state, and the 9.4
ns component to porphyrin impurity.
Subpicosecond transient absorption measurements were used
to further characterize the photochemistry of the triad in its
two forms. Representative results are shown in Figure 4. The
data for the closed DHI-P-C60 triad 1c in Figure 4b show
that at 1.0 ns after the excitation pulse, there is clear evidence
for an absorption maximum corresponding to the fullerene
radical anion at ∼1010 nm. This ion results from the
DHI-P•+-C60•- charge-separated state, as with model dyad 9.
The kinetic trace at 1000 nm (Figure 4a) shows the formation
of DHI-P•+-C60•- with a time constant of 2.0 ns and its decay
with a lifetime of ∼4 ns (the short time window available in
this experiment limits the accuracy of this lifetime). In contrast,
the results for BT-P-C60 triad 1o show no detectable fullerene
radical anion absorption in the 1000 nm region (Figure 4b).
The kinetic trace in Figure 4a shows only a decay of the
porphyrin first excited singlet state absorption at 1000 nm on
this time scale, with a time constant of 55 ps.
To obtain
a better measurement of the decay of
DHI-P•+-C60•- in 1c, transient absorbance experiments on the
nanosecond time scale were performed. Figure 5 shows the
kinetics at 1000 nm following excitation at 650 nm with 4.8 ns
•-
laser pulses. The formation of DHI-P•+-C60 is convoluted
into the excitation pulse on this time scale, but the decay is
seen to be biexponential, with a 4.7 ns decay constant and a
component that does not decay on the time scale shown. The
short component is ascribed to decay of DHI-P•+-C60 and
•-
the long component to the porphyrin and fullerene triplet excited
states. The triplet states form via normal intersystem crossing
in the porphyrin and possibly in part by recombination of
•-
DHI-P•+-C60
.
Figure 6 shows the effect of cycling the triad through a series
of UV-vis cycles. The set of high (∼0.015) absorbance values
signifies the detection of the fullerene radical anion of
DHI-P•+-C60 when the molecule is in the closed form 1c.
•-
The set of low absorbance values was measured after the
molecule was photoisomerized to the BT-P-C60 state 1o. To
9
2720 J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005

Photochromic Control of Photoinduced Electron Transfer
A R T I C L E S
Figure 5. Transient absorption data (O) at 1000 nm obtained following
excitation of a ∼1 × 10-4 M solution of DHI-P-C60 triad 1c in
2-methyltetrahydrofuran by a 4.8 ns laser pulse at 650 nm. The solid line
is a 2-exponential fit with a component that decays with a time constant of
4.7 ns and a nondecaying component. The dotted line is the spectrometer
response function, showing that the formation of the transient absorbance
is convoluted with the laser pulse.
Figure 7. Transient states and relevant interconversion pathways for triad
1 with the photochromic moiety in the open, 1o, form (left) and the closed,
1c, form (right).
Discussion
Energetics. To interpret the photochemical results, we need
estimates of the energies of the various excited states and charge-
separated states. The energy of the porphyrin first excited singlet
state, estimated from the wavenumber average of the longest
wavelength absorption maximum and the shortest-wavelength
emission band, is 1.91 eV above the ground state. Redox
potentials of model compounds measured by cyclic voltammetry
were used to estimate the energies of the charge-separated states.
The first oxidation potential of model porphyrin 3 is 1.03 V
versus SCE, measured in benzonitrile with 0.1 M tetra-n-
butylammonium hexafluorophosphate as supporting electrolyte
and ferrocene as an internal standard.16 The first reduction
potential of a model for the fullerene in 1, measured under the
same conditions, is -0.56 V.17 The first reduction potentials
for the DHI moiety are -1.18 V for the DHI form and -0.70
Figure 6. Photochemical cycling of the molecular double-throw switch 1.
The data are the amplitudes of the transient absorption at 1000 nm following
laser excitation at 650 nm. High absorbance indicates the presence of the
fullerene radical anion in DHI-P•+-C60•-. Low absorbance indicates the
absence of a fullerene radical anion in BT•--P•+-C60. At the beginning
of the experiment (left side of the figure), the molecule is in the betaine
form 1o. Red (g590 nm) light excitation converts the molecule to the DHI
form 1c. Excitation at 366 nm converts 1c back to 1o. The position of the
double-throw switch is shown schematically at the right.
V for the betaine form.9 Thus, the DHI-P•+-C60 state is
•-
estimated to be 1.59 eV above the molecular ground state, and
the BT•--P•+-C60 state 1.73 eV above the ground state. On
the basis of these data, a diagram showing the energetics of the
various relevant states of 1 and some of their potential
interconversion pathways has been constructed (Figure 7).
Electron-Transfer Kinetics. The spectroscopic results for
P-C60 dyad 9 and DHI-P dyad 2 will be discussed, and the
information obtained will then be used to decipher the data for
triad 1. The quenching of the porphyrin fluorescence in P-C60
dyad 9 is ascribed to photoinduced electron transfer to form
the P•+-C60•- charge-separated state. The formation of this state
is verified by the transient absorption results, showing the
fullerene radical anion absorption at ∼1010 nm and absorption
of the porphyrin radical cation in the visible region of the
spectrum. Although singlet energy transfer from 1P-C60 to the
fullerene to yield P-1C60 at 1.75 eV is also energetically
obtain these data, the transient absorbance of the molecule in
the BT-P-C60 form was first measured at 1000 nm (left
side of Figure 6). Laser excitation (650 nm) generates
BT•--P•+-C60 via photoinduced electron transfer from the
porphyrin to the betaine, and this state rapidly recombines,
leaving only residual absorption that comes mainly from the
fullerene radical anion generated in the portion of the sample
that is still in the DHI-P-C60 form. The sample was then
irradiated for 40 min with visible (g590 nm) light from a xenon
lamp, which also caused some local heating. The combined
effects of the light and heating caused virtually complete
isomerization of BT-P-C60 to the DHI-P-C60 form 1c. Laser
excitation of 1c led to photoinduced electron transfer from the
porphyrin to the fullerene, giving DHI-P•+-C60•-. Subse-
quently, the sample was irradiated for 1.5 min with 366 nm
light, driving isomerization of DHI-P-C60 back to the original
BT-P-C60 form 1o. Charge separation then switches back to
the other arm of the molecule. The cycles were repeated 7 times,
demonstrating reasonable photochemical stability for the system.
(16) Gould, S. L.; Kodis, G.; Palacios, R. E.; de la Garza, L.; Brune, A.; Gust,
D.; Moore, T. A.; Moore, A. L. J. Phys. Chem. B 2004, 108, 10566-
10580.
(17) Kodis, G.; Liddell, P. A.; de la Garza, L.; Moore, A. L.; Moore, T. A.;
Gust, D. J. Mater. Chem. 2002, 12, 2100-2108.
9
J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005 2721

A R T I C L E S
Straight et al.
possible and may occur to some extent, no evidence for this
process was observed in 2-methyltetrahydrofuran. In the absence
of energy transfer, the rate constant for photoinduced electron
transfer, kET, is given by eq 1
attempt was made to find the optimum wavelength for maxi-
mizing the population of the betaine form.
From these results, it is clear that triad 1 performs as a
molecular double-throw switch. This is easily seen from Figure
6. On the left side of Figure 6, the transient absorbance at 1000
nm 5 ns after 650 nm excitation was measured for the molecule
in the open, BT-P-C60 form 1o. In this form, 99% of the
photons absorbed by the porphyrin result in formation of
BT•--P•+-C60; photoinduced electron transfer occurs from the
porphyrin to the betaine moiety. The residual absorbance at 1000
nm is due to C60•- from the 1c still present and possibly some
decomposition due to irradiation. Irradiation of the sample with
light g590 nm for 40 min photoisomerizes the molecule to the
DHI-P-C60 form 1c. Excitation of the sample with 650 nm
kET ) (1/τs) - kD
(1)
where τs is the lifetime of the porphyrin excited singlet state of
9 (2.0 ns), and kD is the rate constant for decay of the excited
state in the absence of electron transfer. We estimate kD as 9.9
× 107 s-1 from the 10.1 ns lifetime of the first excited singlet
state of model porphyrin 3. Thus, kET ) 4.0 × 108 s-1. The
quantum yield of photoinduced electron transfer is kET × τs, or
•-
80%. The P•+-C60 state decays with a rate constant kCR of
laser pulses then generates the DHI-P•+-C60 charge-
•-
∼2.5 × 108 s-1, as determined from the 4 ns lifetime measured
separated state via photoinduced electron transfer from the
porphyrin to the fullerene. The switch has been thrown, and
electrons flow down the second pathway. Irradiation with 366
nm light for 1.5 min switches the molecule back to the
BT-P-C60 form, and the transient absorbance at 1000 nm again
drops. It is clear from Figure 6 that little or no photodecom-
position has occurred during the cycles shown. The variations
in absolute amplitude of the transient absorbance when the
molecule is in either the 1o or 1c form are due to variations in
light flux from the excitation laser.
by transient absorption spectroscopy.
The corresponding rate constants for model dyad 2 have been
reported.9 In the DHI-P form of the dyad, the porphyrin first
excited singlet state is unquenched, and electron transfer does
not occur. In the BT-P form, electron transfer occurs from
BT-1P to yield BT•--P•+ with a rate constant k′ of 2.0 ×
ET
1010 s-1 and a quantum yield of 1.0. Recombination of
BT•--P•+ to the ground state occurs rapidly, with k′CR ) 3.4 ×
1011 s-1
.
Turning now to triad 1, we find that the photochemical
behavior is basically a superposition of the processes in model
compounds 2 and 9. In the closed form of the triad,
DHI-P-C60, the porphyrin first excited singlet state decays
with a time constant of 1.9 ns (from the time-resolved
Conclusions
Molecule 1 has been shown to function as a double-throw
switch for photoinduced electron transfer, with switching
between production of two charge-separated states due to
photoisomerization of the dihydroindolizine moiety. The quan-
tum yield of one charge-separated state is 99% when the
molecule is in the open, BT-P-C60 form, while the yield of
the second charge-separated state is 82% when the triad is in
the DHI-P-C60 state. Thus, the excited porphyrin moiety sends
an electron down one arm of the molecule or the other, based
on the state of the photochromic switch. In either form of the
molecule, 1% or less of the porphyrin excited states decay by
electron transfer along the wrong pathway. The molecule is
relatively stable toward photodecomposition under the conditions
employed here. This result, coupled with other examples of
photochemically operated switches and logic gates, may help
point the way to photonic control of electron flow at the
molecular level, with possible applications to data processing,
storage, and transmission; biomimetic signaling; and solar
energy conversion.
•-
fluorescence results) to yield the DHI-P•+-C60 charge-
separated state. The rate constant for photoinduced electron
transfer (k2, for step 2 in Figure 7) may be calculated using an
equation similar to eq 1 and taking k1 (step 1 in Figure 7) as
9.9 × 107 s-1, based on the results for model porphyrin 3. The
value obtained is 4.3 × 108 s-1, and the quantum yield of
DHI-P•+-C60•- is 82%. From the 4.7 ns lifetime of this charge-
separated state measured using transient absorption spectroscopy
(Figure 5), k3 equals 2.1 × 108 s-1. Although Figure 7 depicts
recombination of DHI-P•+-C60•- directly to the ground state,
some recombination to give a triplet state cannot be excluded.
Excitation of the porphyrin moiety of the open, BT-P-C60
isomer 1o is also followed by photoinduced electron transfer,
but in this case the product is BT•--P•+-C60. The rate constant
for this process, k2′, may be calculated using eq 2
1/τs′ ) k1′ + k2′ + k2
(2)
Experimental Procedures
Instrumental Techniques. The 1H NMR spectra were recorded on
Varian Unity spectrometers at 300 or 500 MHz. Samples were dissolved
in deuteriochloroform with tetramethylsilane as an internal reference.
Mass spectra were obtained on a matrix-assisted laser desorption/
ionization time-of-flight spectrometer (MALDI-TOF). The solvent for
all spectroscopic measurements was freshly distilled 2-methyltetrahy-
drofuran, unless otherwise stated. All samples were deoxygenated by
bubbling with argon for 20 min. Ground-state absorption spectra were
measured on a Shimadzu UV-3101PC UV-vis-NIR spectrometer.
Steady-state fluorescence emission spectra were measured using a
Photon Technology International MP-1 fluorometer and corrected.
Excitation was produced by a 75 W xenon lamp and single grating
monochromator. Fluorescence was detected at 90° to the excitation
beam via a single grating monochromator and an R928 photomultiplier
where τs′ is the lifetime of the porphyrin first excited singlet
state of 1o (taken as 55 ps from the transient absorption data),
k1′ is the rate constant for decay of this state by the usual
photophysical processes in the absence of electron transfer (taken
as 9.9 × 107 s-1 from the results for model porphyrin 3), and
k2 is the rate constant for photoinduced electron transfer to the
fullerene moiety (taken as 4.3 × 108 s-1 from the results for
1c). A value for k2′ of 1.8 × 1010 s-1 is calculated. This
corresponds to a quantum yield of 99% for BT•--P•+-C60.
The yield of BT-P•+-C60 via electron transfer down the
•-
wrong path is only ∼1%. Of course, in an actual solution of
the molecules, a fraction is in the DHI-P-C60 form at the
photostationary state when 366 nm irradiation is employed. No
9
2722 J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005

Photochromic Control of Photoinduced Electron Transfer
A R T I C L E S
tube having S-20 spectral response operating in the single-photon-
counting mode.
Ar-H), 7.28 (4H, s, Ar-H), 8.00 (2H, d, J ) 8.7 Hz, Ar-H), 8.27
(2H, d, J ) 8.1 Hz, Ar-H), 8.36 (2H, d, J ) 8.1 Hz, Ar-H), 8.68-
8.90 (8H, m, â-pyrrole H); MALDI-TOF calcd for C55H51N5O2 813.40;
obsd 813.37; UV/vis (dichloromethane) 421, 517, 555, 595, 649 nm.
Fluorescence decay measurements were performed on ∼1 × 10-5
M solutions by the time-correlated single-photon-counting method. The
excitation source was a cavity-dumped Coherent 700 dye laser pumped
by a frequency-doubled Coherent Antares 76s Nd:YAG laser. Fluo-
rescence emission was detected at a magic angle using a single grating
monochromator and microchannel plate photomultiplier (Hamamatsu
R2809U-11). The instrument response time was ca. 35-50 ps, as
verified by scattering from Ludox AS-40 at the excitation wavelength.18
Nanosecond transient absorption measurements were made with
excitation from an OPOTEK optical parametric oscillator driven by
the third harmonic of a Continuum Surelight Nd:YAG laser (650 nm).
The pulse width was ∼5 ns, and the repetition rate was 10 Hz. The
detection portion of the spectrometer was manufactured by Ultrafast
Systems.
4-[15-{4-[(Pyridazine-4-carbonyl)-amino]-phenyl}-10,20-bis-(2,4,6-
trimethylphenyl)-21H,23H-porphine-5-yl]benzoic acid t-butyl ester
(6) was prepared by dissolving 170 mg (0.208 mmol) of 5 in 15 mL of
dichloromethane along with 56 mg (0.45 mmol) of pyridazine 4-car-
boxylate and 27 mg (0.22 mmol) of 4-(dimethylamino)pyridine. After
the mixture was cooled to 0 °C, 84 mg (0.44 mmol) of 1-[3-
(dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride was added;
after 10 min, the ice bath was removed. The reaction mixture was stirred
for 22 h, diluted with dichloromethane, and washed with water, and
the solvent was evaporated at reduced pressure. Chromatography (silica
gel, 30:70 acetone/dichloromethane) yielded 150 mg (0.163 mmol) of
6 (78%): 1H NMR (300 MHz, CDCl3) δ -2.62 (2H, s, N-H), 1.76
(9H, s, t-Bu-H), 1.84 (12H, s, -CH3), 2.62 (6H, s, -CH3), 7.28 (4H,
s, Ar-H), 8.08-8.12 (4H, m, Ar-H), 8.28 (2H, d, J ) 8.7, Ar-H),
8.70-8.84 (8H, m, â-pyrrole H), 9.52-9.54 (1H, m, pyridazine-H),
9.81 (1H, brs, pyridazine-H); MALDI-TOF m/z calcd for C60H53N7O3
919.42; obsd 919.41; UV/vis (dichloromethane) 419, 515, 550, 590,
646 nm.
The femtosecond transient absorption apparatus consisted of a kHz
pulsed laser source and a pump-probe optical setup. The laser pulse
train was provided by a Ti:Sapphire regenerative amplifier (Clark-MXR,
Model CPA-1000) pumped by a diode-pumped CW solid-state laser
(Spectra Physics, Model Millennia V). The typical laser pulse was 100
fs at 790 nm, with a pulse energy of 0.9 mJ at a repetition rate of 1
kHz. Most of the laser energy (80%) was used to pump an optical
parametric amplifier (IR-OPA, Clark-MXR). The excitation pulse was
sent through a computer-controlled optical delay line. The remaining
laser output (20%) was focused into a 1.2 cm rotating quartz plate to
generate a white light continuum. The continuum beam was further
split into two identical parts and used as the probe and reference beams,
respectively. The probe and reference signals were focused onto two
separated optical fiber bundles coupled to a spectrograph (Acton
Research, Model SP275). The spectra were acquired on a dual diode
array detector (Princeton Instruments, Model DPDA-1024).19 To
determine the number of significant components in the transient
absorption data, singular value decomposition analysis20,21 was carried
out using locally written software based on the MatLab 5.0 program
(MathWorks, Inc.).
Benzyl N-(4-Cyanophenyl)carbamate (10). A 1 g (8.47 mmol)
portion of p-cyanoaniline was suspended in 10 mL of 1 M NaOH, 3
mL THF was added, and the mixture was cooled to 0 °C. Benzylchlo-
roformate (1.8 mL, 12.6 mmol) was slowly added, followed by 5 mL
of 2 M aqueous NaOH. After 10 min, the yellow solution turned white;
10 mL of acetone and 5 mL of 2 M NaOH were added. After being
stirred overnight, the reaction mixture was diluted with ethyl acetate,
and the organic layer was washed with citric acid and water and dried
over MgSO4. The solvent was evaporated under reduced pressure. The
residue was chromatographed (silica gel, 10:90 ethyl acetate/toluene),
and 954 mg (3.78 mmol) of 10 was isolated (45%): 1H NMR (300
MHz, CDCl3) δ 5.22 (2H, s, -CH2-), 6.86 (1H, s, N-H), 7.37-7.41
(5H, m, Ar-H), 7.51 (2H, d, J ) 9 Hz, Ar-H), 7.59 (2H, d, J ) 9
Hz, Ar-H).
The photoinduced opening and closing kinetics of the triad were
studied using, respectively, a UVP UV lamp Model UVGL-25 and a
Xe/HgXe-lamp (ORIEL Corp. Model 66028). Before sample illumina-
tion with the Xe lamp, the IR portion of the light was reduced by the
beam being passed through two water-cooled IR filters (A ) 1.8 and
2.3 at 900 and 970 nm, respectively). Furthermore, a long-pass filter
was used to remove wavelengths shorter than 590 nm.
Synthesis. Using previously reported methods,15,16 4,4′-[10,20-bis-
(2,4,6-trimethylphenyl)-21H,23H-porphine-5,15-diyl]bis[benzoic acid]
dimethyl ester 3 was prepared and converted to 4-[15-(4-amino-
phenyl)-10,20-bis(2,4,6-trimethylphenyl)-21H-23H-porphine-5-yl]b en-
zoic acid 4.
Benzyl N-(4-Formylphenyl)carbamate (11). This previously re-
ported22 compound was prepared as follows. In 25 mL of tetrahydro-
furan was dissolved 0.750 g (2.98 mmol) of 10, and the mixture was
cooled to -40 °C. Diisobutylaluminum hydride in heptane (15 mL of
a 1.0 M solution, 15 mmol) was added slowly by addition funnel under
N2. The reaction mixture was allowed to warm gradually to room
temperature and was stirred overnight. Hydrochloric acid (10 mL of a
2 M solution) was added, and the mixture was stirred for 10 min before
being washed with water and aqueous NaHCO3. The organic layer was
diluted with toluene and dried with MgSO4, and the solvent was
evaporated at reduced pressure. The residue was chromatographed
(silica gel, 40:60 ethyl acetate/hexanes), and 165 mg (0.65 mmol) of
1
11 was isolated (22%): H NMR (300 MHz, CDCl3) δ 5.21 (2H, s,
-CH2-), 7.08 (1H, s, N-H), 7.35-7.41 (5H, m, Ar-H), 7.56 (2H, d, J
) 8 Hz, Ar-H), 7.82 (2H, d, J ) 8 Hz, Ar-H), 9.89 (1H, s, -CHO).
4-[15-(4-Aminophenyl)-10,20-bis(2,4,6-trimethylphenyl)-21H-
23H-porphine-5-yl]benzoic Acid t-Butyl Ester (5). Porphyrin 4 (220
mg, 0.290 mmol) was dissolved in 9 mL of anhydrous N,′N-
dimethylacetamide containing 128 mg (0.562 mmol) of benzyltrieth-
ylammonium chloride and 1.8 g (13 mmol) of dry K2CO3. The reaction
mixture was heated to 55 °C under N2 before 2.64 mL (23 mmol) of
2-bromo-2-methylpropane was added. After 17 h, the reaction was
cooled, diluted with dichloromethane, and washed repeatedly with
water. The organic layer was distilled at reduced pressure, and the
residue was chromatographed (silica gel, 8:92 MeOH:CH2Cl2) to give
173 mg (0.213 mmol) of 5 (73%): 1H NMR (300 MHz, CDCl3) δ
-2.60 (2H, s, N-H), 1.76 (9H, s, t-Bu-H), 1.83 (12H, s, Ar-CH3),
2.63 (6H, s, Ar-CH3), 4.04 (2H, brs, -NH2), 7.06 (2H, d, J ) 8.7 Hz,
Benzyl N-[4-(1′,5′-Dihydro-1′-methyl-2′H-[5,6]fullereno-C60-Ih-
[1,9-c]pyrrole-2′yl)phenyl ]carbamate (7). Aldehyde 11 (50 mg, 196
mmol) was added to 75 mL of toluene as were 282 mg (392 mmol) of
C60 and 174 mg (1.96 mmol) of sarcosine. The reaction mixture was
heated to reflux under N2 for 20 h, after which the solvent was
evaporated at reduced pressure. The residue was redissolved in
dichloromethane and filtered, and the solvent was again evaporated.
The residue was chromatographed (silica gel, 3:97 ethyl acetate/toluene)
to give 76 mg (0.076 mmol) of 7 (39%): 1H NMR (300 MHz,
CDCl3) δ 2.79 (3H, s, N-CH3), 4.24 (1H, d, J ) 9.3, -N-CH2-)
(18) Gust, D.; Moore, T. A.; Luttrull, D. K.; Seely, G. R.; Bittersmann, E.;
Bensasson, R. V.; Rouge´e, M.; Land, E. J.; de Schryver, F. C.; Van der
Auweraer, M. Photochem. Photobiol. 1990, 51, 419-426.
(19) Freiberg, A.; Timpmann, K.; Lin, S.; Woodbury, N. W. J. Phys. Chem. B
1998, 102, 10974-10982.
(20) Golub, G. H.; Reinsch, C. Numer. Math. 1970, 14, 403-420.
(21) Henry, E. R.; Hofrichter, J. In Methods in Enzymology; Ludwig, B., Ed.;
Academic Press: San Diego, 1992; p 219.
(22) Witek, S.; Bielawski, J.; Bielawska, A. J. Prakt. Chem./Chem.-Ztg. 1979,
321, 804-812.
9
J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005 2723

A R T I C L E S
Straight et al.
4.89 (1H, s, -N-CH-), 4.97 (1H, d, J ) 9.3, -N-CH2-), 5.19 (2H,
s, -CH2-), 6.70 (1H, s, -N-H), 7.31-7.40 (5H, m, Ar-H), 7.46
(2H, d, J ) 9 Hz, Ar-H), 7.73 (2H, brm, Ar-H); MALDI-TOF m/z
calcd for C77H18N2O2 1002.11; obsd 1002.14; UV/vis (dichloromethane)
702 nm.
CH2-), 4.96 (1H, s, pyrrolid-CH-), 4.99 (1H, d, J ) 9.3, pyrrolid-
CH2-), 7.25 (4H, s, Ar-H), 7.89 (4H, brs, Ar-H), 8.18-8.23 (3H, m,
Ar-H, amide-H), 8.27-8.33 (4H, m, Ar-H), 8.41 (2H, d, J ) 8.4
Hz, Ar-H), 8.68-8.73 (8H, m, â-pyrrole H); MALDI-TOF m/z calcd
for C122H54N6O3 1651.43; obsd 1651.41.
DHI-P dyad (8). In 18 mL of freshly distilled tetrahydrofuran in
a 50 mL round-bottom flask fitted with a septum was dissolved 58 mg
(0.063 mmol) of 6. The solution was degassed by three freeze-pump-
thaw cycles. After the flask was backfilled with argon, 118 mg (0.441
mmol) of spiro[fluorene-9,3′-[3H]pyrazole]-4′,5′-dicarbonitrile23,24 was
added as a solid, and the mixture was irradiated for 45 min with a 450
W medium-pressure mercury lamp through a Pyrex 7740 filter. After
this time, thin-layer chromatography indicated that the reaction was
not yet complete. Another 50 mg (0.19 mmol) of spiro[fluorene-9,3′-
[3H]pyrazole]-4′,5′-dicarbonitrile was added, and the solution was
irradiated for another 30 min. Following irradiation, the reaction was
kept in the dark overnight to allow thermal conversion of the product
to the closed form. Column chromatography (silica gel, dichlo-
romethane) followed by a second column (silica gel, 40:60 ethyl acetate/
hexanes) was performed to purify the product and separate it from its
isomer to yield 30 mg (0.026 mmol, 44%) of 8: 1H NMR (300 MHz,
CDCl3) δ -2.64 (2H, s, N-H), 1.75 (9H, s, t-Bu-H), 1.81 (12H, s,
Ar-CH3), 2.62 (6H, s, Ar-CH3), 5.52 (1H, d, J ) 2.4 Hz, pyridazine
CH), 5.77 (1H, dd, J ) 2.4 Hz, 2.1 Hz, pyridazine CdCHs), 7.27
(4H, s, Ar-H) 7.41-7.87 (9H, m, fluorene Ar-H, amide N-H), 7.69
(1H, d, J ) 2.1 Hz, pyridazine NdCHs), 7.73 (2H, d, J ) 8.7 Hz,
Ar-H), 8.12 (2H, d, J ) 8.7 Hz, Ar-H), 8.26 (2H, d, J ) 8.1 Hz,
Ar-H), 8.36 (2H, d, J ) 8.1 Hz, Ar-H), 8.65-8.75 (8H, m, â-pyrrole
H); MALDI-TOF m/z calcd for C77H61N9O3 1159.49; obsd 1159.47.
P-C60 Dyad (9). Fullerene precursor 7 (11.5 mg, 0.011 mmol) was
dissolved in 50 mL of dichloromethane and cooled to -78 °C under a
nitrogen atmosphere before 0.4 mL of 1.0 M BBr3 solution in
dichloromethane was added. After being stirred for 30 min, the reaction
was warmed to room temperature and stirred for 1 h, whereupon 25
mL of pyridine was added. The dichloromethane was evaporated at
reduced pressure, and 1.0 mg (0.0082 mmol) of 4-(dimethylamino)-
pyridine and 20 mg (0.025 mmol) of 4,4′-[10,20-bis(2,4,6-trimeth-
ylphenyl)-21H,23H-porphine-5,15-diyl]bis[benzoic acid] monomethyl
ester15 were added to the remaining pyridine solution. This solution
was cooled to 0 °C before 3 mg (0.016 mmol) of 1-[3-(dimethylamino)-
propyl]-3-ethylcarbodiimide hydrochloride was added. After 5 min,
the ice bath was removed, and the reaction mixture was stirred in the
dark for 20 h. The solvent was evaporated under reduced pressure,
and the residue was chromatographed (silica gel, 5:95 ethyl acetate/
toluene) to give 12 mg of 9 (63%): 1H NMR (300 MHz, CDCl3) δ
-2.63 (2H, s, N-H), 1.83 (12H, s, -CH3), 2.62 (6H, s, -CH3), 2.86
(3H, s, N-CH3), 4.10 (3H, s, COOCH3), 4.29 (1H, d, J ) 9.3, pyrrolid-
Triad 1. Dyad 8 (35 mg, 0.030 mmol) was dissolved in 4 mL of
dichloromethane; 10 µL of triethylsilane and 1.8 mL of trifluoroacetic
acid were added, and the reaction mixture was stirred in the dark. After
4.5 h, the reaction mixture was diluted with dichloromethane and
washed with water until neutral, the solvent was evaporated at reduced
pressure, and the residue was redissolved in 7 mL of dichloromethane.
Thin-layer chromatography indicated that the t-butyl ester had been
>95% cleaved to give the free acid. Meanwhile, 36 mg (0.036 mmol)
of 7 was dissolved in 40 mL of dichloromethane and cooled to -78
°C under N2, and 1 mL of 1 M BBr3 in heptane was added. This reaction
mixture was stirred under N2 for 30 min before being warmed to room
temperature. After being stirred for an additional 1.5 h, the reaction
mixture was washed with water, 5% aqueous NaHCO3, and then
repeatedly with water. Pyridine was added periodically throughout the
washings to keep 7 in its deprotected, amino form soluble. The resulting
solution was concentrated to a volume of 10 mL by evaporation of the
solvent at reduced pressure and then combined with the carboxylic acid
form of dyad 8, prepared as described previously, and dissolved in 7
mL of dichloromethane. To the resulting solution was added 10 mg
(0.048 mmol) of 1,3-dicyclohexylcarbodiimide, and the reaction mixture
was stirred under N2 for 40 h. The reaction mixture was then washed
twice with 0.25 M citric acid in water and repeatedly with water, the
organic layer was concentrated by evaporation of the solvent at reduced
pressure, and the residue was chromatographed (silica gel, 5:95 ethyl
acetate/toluene followed by silica gel 1:99 acteone/dichloromethane)
to give 4.5 mg (0.0023 mmol) of 1 (8.2%): 1H NMR (500 MHz CDCl3)
δ -2.66 (2H, s, NH), 1.81 (12H, s, Me-CH3), 2.62 (6H, s, Me-CH3),
2.86 (3H, s, N-CH3), 4.29 (1H, d, J ) 9 Hz, pyrrolid-CH2), 4.98 (1H,
s, pyrrolid-CH), 5.01 (1H, d, J ) 9 Hz, pyrrolid-CH2), 5.52 (1H, d, J
) 2.5 Hz, pyridazine CH), 5.76 (1H, dd, J ) 2.5 Hz, J ) 2.5 Hz,
pyridazine -CdCH-), 7.27 (4H, s, Ar-H), 7.43-7.68 (7H, m, fluorene
Ar-H, amide N-H), 7.70 (1H, d, J ) 2.5 Hz, pyridazine -NdCH-),
7.75 (2H, d, J ) 8.5 Hz, Ar-H), 7.82-7.87 (2H, m, fluorene Ar-H),
7.89 (4H, brm, pyrrolid-phenyl-H), 8.14 (2H, d, J ) 8.5 Hz, Ar-H),
8.19 (1H, s, C60-P2H amide N-H), 8.22 (2H, d, J ) 8 Hz, Ar-H),
8.32 (2H, d, J ) 8 Hz, Ar-H), 8.66-8.73 (8H, m, â-pyrrole H);
MALDI-TOF calcd for C142H63N11O2 1954.52; obsd 1954.43.
Acknowledgment. This work was supported by a grant from
the National Science Foundation (CHE-0352599). This is
publication No. 600 from the ASU Center for the Study of Early
Events in Photosynthesis. J.A. thanks the Carl Trygger Founda-
tion for Scientific Research for postdoctoral support.
(23) Du¨rr, H. Photochromism of dihydroindolizines and related systems; In
Organic Photochromic and Thermochromic Compounds; Crano, J. C.,
Guglielmetti, R. J., Eds.; Plenum Press: New York, 1999; pp 223-266.
(24) Gross, H.; Du¨rr, H. Angew. Chem., Int. Ed. Engl. 1982, 21, 216-217.
JA044128I
9
2724 J. AM. CHEM. SOC. VOL. 127, NO. 8, 2005