Through-bond excited energy transfer mediated by an amidinium-carboxylate salt bridge in Zn-porphyrin free-base porphyrin dyads

Article abstract of DOI:10.1002/chem.200701486

Excited energy-transfer processes were investigated for a supramolecular Zn-porphyrin free-base porphyrin dyad, ZnPA-2·FbPC-2, in which β-octaalkylated meso-diarylporphyrins are connected through an amidinium-carboxylate salt bridge. The rate of energy transfer in the dyad (1.3 × 10⁹ s^-1) is substantially slower than that in the previously reported dyad, ZnPA-1·FbPC-1 (4.0 × 10⁹ s^-1), in which meso-tetraarylporphyrins are connected through the same amidinium-carboxylate salt bridge. The Foerster-type mechanism can explain only minor parts of these rates (3.3 × 10⁸ and 5.1 × 10⁸ s^-1, respectively). Thus, Dexter-type through-bond energy transfer may be invoked. Indeed, bridge-mediated electronic processes would be favored in ZnPA-1·FbPC-1 over ZnPA-2·FbPC-2 on the basis of steric and electronic factors. Sterically, the phenyl groups in ZnPA-2 and FbPC-2 are more closely perpendicular to the porphyrin planes than those in ZnPA-1 and FbPC-1. Electronically, the energy and symmetry of the occupied frontier orbitais should favor ZnPA-1·FbPC-1 over ZnPA-2·FbPC-2 in terms of electronic interactions through the bridge. Therefore, the observed trend (ZnPA-1·FbPC-1 > ZnPA-2·FbPC-2), consistent with these considerations, lends further support to the through-bond mechanism. Thus, the amidinium-carboxylate salt bridge is effective in mediating through-bond energy transfer even though the bond is noncovalent.

Full text of DOI:10.1002/chem.200701486

DOI: 10.1002/chem.200701486
Through-Bond Excited Energy Transfer Mediated by an Amidinium–
Carboxylate Salt Bridge in Zn–Porphyrin Free-Base Porphyrin Dyads
Joe Otsuki,*[a] Yuki Kanazawa,[a] Atsuko Kaito,[a] D.-M. Shafiqul Islam,[b]
Yasuyuki Araki,[b] and Osamu Ito[b]
Abstract: Excited energy-transfer pro-
cesses were investigated for a supra-
molecular Zn–porphyrin free-base por-
phyrin dyad, ZnPA-2·FbPC-2, in which
b-octaalkylated meso-diarylporphyrins
are connected through an amidinium–
carboxylate salt bridge. The rate of
energy transfer in the dyad (1.3”
109 sꢀ1) is substantially slower than that
in the previously reported dyad, ZnPA-
1·FbPC-1 (4.0”109 sꢀ1), in which meso-
these rates (3.3”108 and 5.1”108 sꢀ1,
respectively). Thus, Dexter-type
perpendicular to the porphyrin planes
than those in ZnPA-1 and FbPC-1.
Electronically, the energy and symme-
try of the occupied frontier orbitals
should favor ZnPA-1·FbPC-1 over
ZnPA-2·FbPC-2 in terms of electronic
interactions through the bridge. There-
fore, the observed trend (ZnPA-
through-bond energy transfer may be
invoked. Indeed, bridge-mediated elec-
tronic processes would be favored in
ZnPA-1·FbPC-1 over ZnPA-2·FbPC-2
on the basis of steric and electronic
factors. Sterically, the phenyl groups in
ZnPA-2 and FbPC-2 are more closely
1·FbPC-1>ZnPA-2·FbPC-2),
consis-
tent with these considerations, lends
further support to the through-bond
mechanism. Thus, the amidinium–car-
boxylate salt bridge is effective in me-
diating through-bond energy transfer
even though the bond is noncovalent.
Keywords: energy transfer · por-
phyrinoids · salt bridges · supra-
molecular chemistry · through-bond
interactions
tetraarylporphyrins
are
connected
through the same amidinium–carboxyl-
ate salt bridge. The Fçrster-type mech-
anism can explain only minor parts of
Introduction
The supramolecular approach is a powerful methodology to
construct complex yet well-defined multichromophore as-
semblies. In such assemblies, chromophores are joined to-
gether through noncovalent bonds, such as hydrogen bonds
and coordination interactions. Therefore, it is important to
know whether or not these noncovalent bonds can mediate
excited energy migration from one chromophore to another
within noncovalent assemblies. Energy transfer may take
place by two mechanisms.[10] One is the Fçrster mechanism
which is due to Coulombic interactions of transition di-
poles.[11] This mechanism is also called the “through-space”
mechanism because Coulomb interactions operate through
space. The other is the Dexter mechanism, which is also
called the exchange mechanism because it can be considered
as an electron exchange process.[12] It is called the “through-
bond” mechanism as well, because the electron exchange
process operates via the overlap of molecular orbitals, that
is, through bonds.
Inspired by natural photosynthetic systems, attempts have
been made to mimic the antenna function by using artificial
chromophores, porphyrins in particular. These works aim
not only at a better understanding of the energy-transfer
processes but also at creating light-harvesting devices.[1–9]
[a]Prof. J. Otsuki, Y. Kanazawa, A. Kaito
College of Science and Technology
Nihon University, 1-8-14 Kanda Surugadai
Chiyoda-ku, Tokyo 101-8308 (Japan)
Fax : (+81)3-3259-0817
E-mail: otsuki@chem.cst.nihon-u.ac.jp
[b]Dr. D.-M. Shafiqul Islam, Dr. Y. Araki, Prof. O. Ito
Institute of Multidisciplinary Research for Advanced Materials
Tohoku University, Katahira, Aoba-ku
Sendai 980-8577 (Japan)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author and contains the
Experimental Section and the coordinates of the optimized structures
of the porphyrins.
We have previously reported the donor–acceptor dyad
ZnPA-1·FbPC-1[13] (Scheme 1; “ZnPA·FbPC” in the previ-
ous report), composed of a zinc-porphyrin and a free-base
3776
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3776 – 3784

FULL PAPER
substituents on the porphyrin
skeletons. The energy-transfer
rate in Lindseyꢁs dimer (4.2”
1010 sꢀ1)[34] is 17-fold larger than
that in Osukaꢁs dimer (2.4”
109 sꢀ1).[35] The difference was
ascribed to different HOMOs
for these porphyrins.[36,37] Lind-
seyꢁs dimer, which is of meso-
tetraphenyl type, has an a2u
HOMO with a significant elec-
tron density on the meso
carbon atoms, to one of which
the linker is attached. On the
other hand, the HOMO of
Osukaꢁs dimer is a1u, which has
nodes on the meso carbon
atoms. Therefore, the conduit
for an electron in the HOMO
in the process of Dexter energy
transfer is disconnected at the
meso carbon atom in Osukaꢁs
dimer. Faster energy-transfer
Scheme 1. The Zn–porphyrin free-base porphyrin dyads studied.
porphyrin connected through an amidinium–carboxylate salt
bridge.[13–23] The same intermolecular bond was also used for
the construction of an antenna-type (donor)4·(acceptor)
G
complex. We found that the rates of excited energy-transfer
processes in these supramolecules are much larger than ex-
pected from the Fçrster-type energy-transfer mechanism. In
the dyad ZnPA-1·FbPC-1, the observed rate was 4.0”
109 sꢀ1, which is an order of magnitude larger than that cal-
culated assuming the Fçrster mechanism (5.1”108 sꢀ1). We
suggested a Dexter-type energy-transfer process as a possi-
ble explanation for the fast excited energy-transfer rate.
This hypothesis has to invoke through-bond energy transfer
through hydrogen bonds. Our literature survey indicated
that energy-transfer rates were accounted for by the Fçrster
mechanism for all noncovalent porphyrin assemblies report-
ed so far.[24–34] Although the through-bond mechanism
through covalent bonds has been well documented, we did
not find any example in which this mechanism was implied
for noncovalent porphyrin assemblies, except our own case
described above. Therefore, we decided to seek further sup-
port for the hypothesis that through-bond energy transfer
can indeed be mediated by hydrogen bonds.
The rate of Dexter-type energy transfer is sensitive to the
nature of frontier orbitals, as it is the electrons in these orbi-
tals that are exchanged during the energy-transfer process.
The highest occupied molecular orbital (HOMO) and the
lowest unoccupied molecular orbital (LUMO) of chromo-
phores are the most critical orbitals, such that their overlap
is important. Scheme 2 shows a pair of covalently connected
Zn–prophyrin free-base porphyrin dyads with the same
linker moiety. It was shown that the rates of energy transfer
are quite different between these dyads due to the different
Scheme 2. Two covalently connected Zn–porphyrin free-base porphyrin
dyads with the same linker moiety.
processes for meso-tetraphenyl-type porphyrins than 5,15-
diaryl-b-octaalkyl-type porphyrins were confirmed for relat-
ed covalently linked porphyrin dyads as well.[38,39] It is also
likely that the mean angle made by the porphyrin plane and
the meso-phenyl ring are different between these dyads. The
b-methyl groups in Osukaꢁs dyad would force its phenyl
rings to orient in a more perpendicular manner to the por-
Chem. Eur. J. 2008, 14, 3776 – 3784
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3777

J. Otsuki et al.
phyrin plane than those in Lindseyꢁs dyad. This effect also
favors Lindseyꢁs dyad over Osukaꢁs in terms of the energy-
transfer process.
With this background, we set out to establish whether or
not the Dexter-type energy transfer operates in the amidini-
um–carboxylate dyad, ZnPA-1·FbPC-1. We prepared a new
dyad, ZnPA-2·FbPA-2, consisting of 5,15-diaryl-b-octaalkyl-
type porphyrins, and investigated the energy-transfer pro-
cess. Our working hypothesis was that if the Dexter-type
energy transfer plays a major role in these dyads, then the
observed rate for ZnPA-2·FbPA-2 would be significantly dif-
ferent from that observed for ZnPA-1·FbPC-1: 1) by a
margin larger than what would be accounted for by the Fçr-
ster-type energy-transfer process and 2) in a direction consis-
tent with steric and electronic factors.
Results
Self-assembling behavior: From previous studies on systems
containing both zinc-porphyrins and an amidine moiety, it
has been established that the amidine axially coordinates to
the zinc-porphyrin in noncoordinating solvents, such as tolu-
ene used in the present study.[21] Therefore, molecules of
ZnPA-2, which is a zinc-porphyrin with an amidine moiety,
self-coordinate to form aggregates. This is apparent in tolu-
ene from its red-shifted Soret and Q-band peaks of the ab-
sorption spectrum (see below). This is not so much a com-
plication, however, because FbPC-2 or any carboxylic acid,
such as benzoic acid, once added binds to the amidine group
much more strongly through amidinium–carboxylate salt
bridge formation, thus disrupting the amidine–zinc-porphy-
rin association.
Figure 1. 1H NMR spectroscopic chemical shifts of FbPC-2 (1 mm) with
addition of ZnPA-2 in CDCl3.
forming the amidinium–carboxylate salt bridge with ZnPA-
2. Thus, upon addition of benzoic acid to a solution of
ZnPA-2, the spectrum reverted to the one for isolated
ZnPA-2. Therefore, ZnPA-2·benzoic acid is more appropri-
ate as a model of the ZnPA-2 moiety in the complex ZnPA-
2·FbPC-2.
1
Figure 1 shows the H NMR spectroscopic chemical shift
changes of the aromatic protons in FbPC-2 on addition of
ZnPA-2. Addition of ZnPA-2 affected the chemical shifts of
the protons on the carboxyphenyl moiety but not of those
on the di-tert-butylphenyl moiety, which is placed across the
porphyrin core. This finding is consistent with the formation
of an amidinium–carboxylate salt bridge. The chemical shifts
changed until the added ZnPA-2 slightly exceeded one
equivalent. The chemical shift changes are a combined
effect of electronic changes associated with salt-bridge for-
mation and the ring-current effect exerted by the ZnPA-2
molecule placed nearby. The slight scatter in the chemical
shift values of the proton ortho to the carboxylic acid group
Figure 2a compares the absorption spectrum for the mix-
ture of ZnPA-2 and FbPC-2 (c) and the sum of the spec-
tra of individual species (”), that is, ZnPA-2·benzoic acid
and FbPC-2. These two spectra almost completely coincide
with each other. This observation for the mixture of ZnPA-2
and FbPC-2 indicates that 1) there is little ground-state elec-
tronic interaction between ZnPA-2 and FbPC-2, but
2) FbPC-2 does interact with ZnPA-2 in a way that prevents
the self-association of ZnPA-2, that is, via the formation of
the amidinium–carboxylate salt bridge.
Figure 2b compares the fluorescence spectra for ZnPA-
*
2·benzoic acid (-”-), FbPC-2 (- -), and the mixture of
&
( ) is due partly to some broadening of the resonance.
ZnPA-2 and FbPC-2 (c) with excitation at 555 nm, at
which the free-base porphyrin exhibits a minimum absorp-
tion. There are several features that indicate energy transfer
from ZnPA-2 to FbPC-2. The fluorescence intensity at
579 nm, at which only ZnPA-2 porphyrin emits fluorescence,
is greatly decreased on mixing with FbPC-2. On the other
hand, the intensity in the low-energy region is increased
when compared to FbPC-2 alone. In particular, the in-
creased fluorescence plateau around 700 nm indicates that
the free-base fluorescence is increased in the presence of
ZnPA-2. The excitation spectra, monitored at 697 nm,
Absorption and fluorescence spectra: It was found that the
absorption spectrum of ZnPA is concentration dependent.
As its concentration became higher, the Soret band peak at
413 nm shifted to 426 nm and the Q-band peaks at 544 and
574 nm merged into a single broad band peaking at 552 nm.
These spectral changes are accounted for by the self-associa-
tion of ZnPA-2 molecules through coordination of the ami-
dine group of one molecule to the zinc ion in another.[20,21]
Benzoic acid prevents this self-association of ZnPA-2 by
3778
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3776 – 3784

Energy Transfer in Porphyrin Dyads
FULL PAPER
Figure 2. Comparison of spectra for a mixed solution of ZnPA-2 and
FbPC-2 and those for the sum of individual species in toluene. a) Absorp-
tion spectrum for a mixture of ZnPA-2 and FbPC-2 (4 mm each; c) and
the sum of the spectra for ZnPA-2 (4 mm) in the presence of benzoic acid
(40 mm) and that for FbPC-2 (”). b) Fluorescence spectra (lex =555 nm)
for a mixture of ZnPA-2 and FbPC-2 (7 mm each; c), for ZnPA-2
(7 mm) in the presence of benzoic acid (70 mm; -”-), and for FbPC-2
Figure 3. Comparison of spectra for a mixed solution of ZnPA-2 and
FbPMe and those for the sum of individual species in toluene. a) Absorp-
tion spectrum for a mixture of ZnPA-2 and FbPMe (4 mm each; c) and
the sum of the individual spectra for ZnPA-2 and FbPMe (4 mm each; ”).
b) Fluorescence spectra (lex =555 nm) for a mixture of ZnPA-2 and
FbPMe (4 mm each; c), for ZnPA-2 (4 mm; -”-), and for FbPMe (4 mm;
*
-
-).
*
(7 mm; - -).
showed a sensitization by ZnPA-2 (data not shown). A
least-squares curve fitting to the fluorescence intensity at
697 nm indicated that the association constant is around 9”
104 mꢀ1.
the notion that the interaction of ZnPA-2 with FbPC-2 is
through the amidinium–carboxylate salt bridge. The fluores-
cence of these sets of porphyrins was also investigated to
see if there is any sign of energy-transfer processes. In this
case, no quenching of ZnPA-2 fluorescence was observed
even when FbPMe was present. Also, there was no sign of
the sensitization of FbPMe by the presence of ZnPA-2.
Compare the fluorescence intensities around 700 nm, a
region where FbPMe emits more strongly than ZnPA-2; the
observed fluorescence spectrum from the mixture of ZnPA-
2 and FbPMe (c) coincides with the calculated sum of
To provide evidence that the energy transfer occurs not
through a diffusional encounter of the porphyrins but within
a specific supramolecular complex, the methyl ester of
FbPC-2, FbPMe, was employed in the same set of absorp-
tion and fluorescence experiments for comparison (see
Figure 3). This compound cannot participate in the amidini-
um–carboxylate interaction. The observed absorption spec-
trum for the mixture of ZnPA-2 and FbPMe (c) nearly
completely coincides with the calculated sum of the spectra
of ZnPA-2 (without benzoic acid this time) and FbPMe (”).
The result indicates that the methyl ester, FbPMe, does not
interrupt the self-association of ZnPA-2, which corroborates
*
spectra of ZnPA-2 (-”-) and FbPMe (- -).
Time-resolved studies on excited energy transfer: Figure 4a
shows the results of picosecond time-resolved fluorescence
measurements. The decay of the fluorescence of ZnPA-2
Chem. Eur. J. 2008, 14, 3776 – 3784
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3779

J. Otsuki et al.
clearly seen in Figure 4b, which shows time-resolved spectra
for the mixture in time domains of 0–0.9 and 1.1–2.4 ns. It is
clearly seen that the fluorescence of ZnPA-2 in the shorter-
wavelength region is replaced by that of FbPC-2 in the
longer-wavelength region. No such time dependence of the
spectral profile was observed when FbPC-2 was replaced by
FbPCM, which is the methyl ester of FbPC-2, thus confirm-
ing that the amidinium–carboxylate salt bridge is indispens-
A
cess was calculated to be 1.26”109 sꢀ1 according to the equa-
tion kobs =tꢀ1ꢀt0
Rate of Fçrster-type energy transfer: The rate of Fçrster-
type energy transfer, kF (in sꢀ1), may be calculated by using
Equation (1), in which k2 is an orientation factor, F and t
are the fluorescence quantum yield and lifetime (in s) of
ZnPA-2, respectively, in the absence of FbPC-2,[40] JF is the
spectral overlap term (in mꢀ1 cm3), n is the refractive index
of the solvent, and r is the donor–acceptor center-to-center
distance (in cm). In Equation (2), F(n) is the fluorescence
spectrum of the donor and e(n) is the absorption spectrum
of the acceptor.
8:81 Â 10ꢀ25k2FJF
ð1Þ
ð2Þ
kF ¼
JF ¼
n4tr6
R
FðnÞeðnÞnꢀ4dn
R
FðnÞdn
The value of k2 =1.01 was used as the orientation factor,[41]
which is based on the assumption that the two porphyrin
planes are in a coplanar conformation. This assumption
would give the upper limit because there are some rotatable
single bonds along the linkage between the meso carbon
atoms connecting the two porphyrin macrocycles. The per-
pendicular porphyrins would give rise to k2 =0.81, as the
lower limit value. The center-to-center separation, r=2.28”
10ꢀ7 cm, was estimated from the crystal structures of a 5,15-
diphenyloctaalkylporphyrin[42] and an amidinium–carboxyl-
ate complex.[43] The quantum yield, F=0.011, was obtained
by comparing the fluorescence intensities integrated over
the entire reciprocal centimeter range with those of zinc-
meso-tetraphenylporphyrin (quantum yield: 0.030[34]). Other
parameter values used in the calculation were n=1.51 (tolu-
ene) and experimentally obtained values of JF =3.94”
10ꢀ14 mꢀ1 cm3 and t=1.6”10ꢀ9 s. The rate of Fçrster-type
energy transfer was obtained as kF =3.3”108 sꢀ1 by using
these parameter values. This accounts for about a fourth of
the observed rate, kobs =1.26”109 sꢀ1.
Figure 4. Time-resolved fluorescence measurements (toluene; each por-
phyrin 25 mm; lex =410 nm). a) Decays of ZnPA-2 fluorescence at 580–
&
*
620 nm in the absence ( ) and presence ( ) of FbPC-2. The time profile
of FbPC-2 fluorescence monitored at 690–720 nm in the mixture is also
~
shown ( ). b) Time-resolved spectra for the mixture in time domains of
0–0.9 and 1.1–2.4 ns.
&
alone ( ) was fitted by a single exponential function, which
gave a lifetime of t0 =1.6 ns. The lifetime was not altered by
the addition of excess (100 mm) benzoic acid to ZnPA-2. On
the other hand, the decay for an equimolar mixture of
*
ZnPA-2 and FbPC-2 ( ) was fitted by two exponential func-
tions with lifetimes of 1.58 (45%) and 0.53 ns (55%). As the
former value is identical with that of ZnPA-2 alone, it was
attributed to noncomplexed ZnPA-2 in the mixture. There-
fore, the latter value was ascribed to the complex, ZnPA-
2·FbPC-2. From the association constant estimated from
steady-state fluorescence titration (9”104 mꢀ1), the fraction
of ZnPA-2 incorporated in the dyad under the present con-
ditions is 52%, which is in good agreement with the above
results. We attributed the shortened lifetime to an energy-
transfer process from the singlet excited state of ZnPA-2 to
FbPC-2 within the complex on the basis of the observed
sensitized fluorescence in the steady-state measurements.
Further support for the energy transfer comes from the
delayed rise in the fluorescence intensity collected at 690–
DFT calculations: It is known that the HOMO and
HOMOꢀ1 of the porphine p system are nearly degener-
ate.[44] They respectively span the symmetry species A1u and
A2u in the D4h point group. This near degeneracy is lifted
upon substitution on the core. When electron-donating sub-
stituents, such as alkyl groups, are introduced to the b posi-
tions, the a1u orbital energy becomes higher, as this orbital
has a larger density on the b-carbon atoms. As a conse-
~
720 nm ( ), in which the fluorescence of FbPC-2 dominates
in the time-resolved measurements (Figure 4a). This is more
3780
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3776 – 3784

Energy Transfer in Porphyrin Dyads
FULL PAPER
quence, the a1u orbital becomes the HOMO. On the other
hand, the introduction of electron-donating substituents, say
phenyl groups, to the meso positions raises the a2u orbital
energy, as this orbital has a larger density on the meso
carbon atoms. In these circumstances, the a2u orbital be-
comes the HOMO.
Discussion
As reported previously, the observed energy-transfer rate in
dyad ZnPA-1·FbPC-1 was kobs =4.0”109 sꢀ1 and the estimat-
ed Fçrster rate, kF =5.1”108 sꢀ1, could only account for a
minor part of the energy-transfer rate.[13] We suggested a
DFT geometry optimization and molecular orbital calcu-
lations (B3LYP/6-31g*) were performed on the porphyrins
to obtain the energy and symmetry of frontier molecular or-
bitals on the porphyrin core. Calculations were separately
carried out for a ZnPA·benzoic acid pair and a FbPC·benz-
A
spectively, in the ZnPA·FbPC dyad. Table 1 shows the fron-
tier orbitals obtained, while Figure 5 shows their orbital en-
ergies. For all porphyrins the LUMO and the LUMO+1 are
nearly degenerate, as expected, within 0.002 eV. In contrast,
the ordering of the HOMO and HOMOꢀ1 and the energy
gap between these orbitals depends on the substituent pat-
terns on the porphyrin core more sensitively than the unoc-
cupied orbitals. For ZnPA-1 and FbPC-1, the HOMOs are
of an a2u-like symmetry, the energies of which are higher
than those of the HOMOꢀ1 (a1u) by 0.006 and 0.011 eV, re-
spectively. On the other hand, for ZnPA-2, the energy of the
a1u orbital is raised relative to that of the a2u orbital. As a
consequence, the ordering is reversed and the HOMO of
ZnPA-2 is now the a1u orbital. Likewise, the energy of the
a1u orbital is raised for FbPC-2 and the energy gap between
the HOMO and HOMOꢀ1 is appreciably narrowed to
0.002 eV, only a little short of the reversal of energy ordering.
Figure 5. Orbital energies, indicated by horizontal bars. Orbital localiza-
tion (i.e., on the porphyrin core or on the bridge) is also indicated by the
horizontal positions.
Table 1. Frontier molecular orbitals.[a]
ZnPA-1·benzoic acid
Benzamidine·FbPC-1
ZnPA-2·benzoic acid
Benzamidine·FbPC-2
LUMO+1
LUMO
HOMO
HOMOꢀ1
[a] tBu groups are omitted in the calculations.
Chem. Eur. J. 2008, 14, 3776 – 3784
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3781

J. Otsuki et al.
Dexter-type energy transfer to account for the remaining
major part of the energy-transfer process (kobsꢀkF =3.5”
109 sꢀ1). However, there was no precedent in which a
Dexter-type through-bond energy transfer was implicated
for supramolecular porphyrins connected through noncova-
lent interactions.[24–34] To strengthen our point, we prepared
a new dyad, ZnPA-2·FbPC-2, and measured the energy-
transfer rate in the present study. This dyad was designed to
have similar spectroscopic properties to the previous one to
minimize the effect on Fçrster-type energy-transfer process-
es. On the other hand, the substitution pattern on the por-
phyrin core was largely changed to have a significant impact
on the energetics of frontier orbitals, which would then have
a large effect on Dexter-type energy-transfer processes, if
they were involved. Thus, if the observed rate for the new
dyad is significantly different from that observed for ZnPA-
1·FbPC-1 in the right direction, it becomes more certain
that a Dexter-type energy-transfer process is indeed in oper-
ation.
The observed rate of energy transfer for dyad ZnPA-
2·FbPC-2 was kobs =1.26”109 sꢀ1, which was only about a
third of the rate for dyad ZnPA-1·FbPC-1. The Fçrster-type
energy-transfer rate for dyad ZnPA-2·FbPC-2 was estimated
to be kF =3.3”108 sꢀ1. This value can only account for a
fourth of the observed rate of energy transfer. The remain-
ing major part of the rate (kobsꢀkF =9.3”108 sꢀ1) for ZnPA-
2·FbPC-2 is only about a fourth of the corresponding rate
for ZnPA-1·FbPC-1 (3.5”109 sꢀ1). The appreciable differ-
ence in these rates between the dyads is in favor of the in-
volvement of Dexter-type energy-transfer processes, if the
difference is in the right direction. We discuss in the follow-
ing whether or not the difference in the rates is consistent
with the difference in the molecular structures when a
Dexter-type energy-transfer process is assumed.
1·FbPC-1 and ZnPA-2·FbPC-2, respectively, from relevant
spectroscopic data.
If we assume that the energy-transfer rates which are un-
accounted for by the Fçrster-type mechanism are due to
Dexter-type processes, that is, kD =3.5”109 and 9.3”108 sꢀ1
in ZnPA-1·FbPC-1 and ZnPA-2·FbPC-2, respectively, the
electronic coupling matrix element for ZnPA-1·FbPC-1 and
ZnPA-2·FbPC-2 would have to be V2 =41 and 7.2 cmꢀ1, re-
spectively. Thus, the value of V2 for ZnPA-2·FbPC-2 is only
about a sixth of that for ZnPA-1·FbPC-1. Now the problem
of whether or not the electronic interaction represented by
V2 in ZnPA-2·FbPC-2 is smaller than that in ZnPA-1·FbPC-
1 is reasonably explained by the difference in the molecular
structures.
There are steric and electronic factors to be considered.
With regard to steric factors, we examined the rotation of
the meso-phenyl rings. In the case of ZnPA-1·FbPC-1, the
angle made by the porphyrin plane and the meso-phenyl
ring was 658 in the optimized structure.[45] However, the
meso-phenyl group may nearly freely rotate with respect to
the porphyrin plane in solution.[46,47] For ZnPA-2·FbPC-2,
the angle of the phenyl ring with respect to the porphyrin
plane is more restricted to approximately 908 because the
phenyl group is flanked by the b-methyl groups, although
the rotation may not be completely hindered.[47,48] It is evi-
dent that these steric factors favor ZnPA-1·FbPC-1 over
ZnPA-2·FbPC-2 in a through-bond interaction. Thus, the
Dexter mechanism is consistent with the steric considera-
tions.
As for the electronic factors, we consider the energetics
and symmetry of the frontier orbitals, as it is the electrons
in these orbitals that are exchanged during the Dexter pro-
cess. The DFT calculations (Figure 5 and Table 1) show that
the LUMO and LUMO+1 of every porphyrin are nearly
degenerate, with their energy difference within 0.002 eV.
Thus, the symmetry difference in these two orbitals is insig-
nificant in determining electronic interactions between
ZnPA and FbPC. On the other hand, for occupied frontier
orbitals, the ordering of the HOMO and HOMOꢀ1 and the
energy gap between these orbitals would be important in de-
termining the electronic coupling through the bridge. The
lowest singlet excited state of porphyrins is a mixture of
HOMO!(LUMO, LUMO+1) and HOMOꢀ1!(LUMO,
LUMO+1) transitions due to configuration interactions, the
contribution of the former transition being larger. However,
the relative contribution of the latter increases as the energy
gap between the HOMO and HOMOꢀ1 is narrowed.[37] As
a guiding principle, a stronger electronic interaction is ex-
pected when the HOMO is a2u because the a2u orbital is
heavily populated at the meso carbon atoms, one of which is
just on the bridge connecting the porphyrins, whereas a
weaker interaction is expected when the HOMO is a1u be-
cause the a1u orbital has virtually no electron density at the
meso carbon atoms. It should be noted, however, that the
symmetry of HOMOꢀ1 plays an increasingly larger role as
the energy gap between the HOMO and HOMOꢀ1 be-
comes smaller.
The Dexter-type energy-transfer rate kD is given by Equa-
tion (3):
2p
ꢀh
ð3Þ
kD
¼
V 2JD
in which ꢀh is Planckꢁs constant, V is the electronic coupling
matrix element, and JD is the overlap integral. The electron-
ic coupling matrix element represents the magnitude of elec-
tronic interaction, while the overlap integral corresponds to
the Franck–Condon weighted density of states. The overlap
integral can be determined by using the following expression
[Eq. (4)]:[12]
R
FðnÞeðnÞdn
R
R
JD
¼
ð4Þ
FðnÞdn eðnÞdn
in which F(n) is the fluorescence spectrum of the donor and
e(n) is the absorption spectrum of the acceptor. For the ab-
sorption spectrum in this case, the Q band of the free-base
porphyrins may be used, as the relevant transition is the
S0!S1 transition. The values of the overlap integrals were
obtained as JD =7.1”10ꢀ5 and 1.1”10ꢀ4 cm for ZnPA-
3782
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3776 – 3784

Energy Transfer in Porphyrin Dyads
FULL PAPER
For ZnPA-1 and FbPC-1, the HOMOs are a2u-like while
the HOMOꢀ1s are a1u, the energy gaps between them being
substantial (0.006 and 0.011 eV, respectively). Accordingly,
the orbital ordering is favorable for a strong electronic inter-
action. For ZnPA-2, the orbital ordering is reversed; the
HOMO is a1u whereas the HOMOꢀ1 is a2u, the energy gap
being 0.005 eV, which is unfavorable for an electronic inter-
action through a bridge via the meso position. For FbPC-2,
the HOMO is a2u-like whereas the HOMOꢀ1 is a1u. This sit-
uation is favorable for an electronic interaction. However,
the energy gap between these two orbitals is only 0.002 eV.
Therefore, in the lowest excited state, there should be a sub-
stantial component of HOMOꢀ1!(LUMO, LUMO+1)
being mixed into the transition HOMO!(LUMO,
LUMO+1). This mixing works unfavorably for the elec-
tronic interaction through the meso carbon atoms. Taken to-
gether, the DFT calculations indicate that the frontier orbi-
tals in ZnPA-1·FbPC-1 are more favorable than those in
ZnPA-2·FbPC-2 for electronic interactions through the
meso carbon atoms. Thus, the through-bond mechanism is
also consistent with the electronic considerations. In conclu-
sion, steric as well as electronic features of the dyads ZnPA-
1·FbPC-1 and ZnPA-2·FbPC-2 are consistent with the exper-
imentally observed trend of energy-transfer rates (ZnPA-
1·FbPC-1>ZnPA-2·FbPC-2) when the Dexter mechanism is
assumed.
zoate!benzamidinium) low-energy excited states might be
involved to decrease the effective DEDB value, although cor-
responding optical transitions were not detected by absorp-
tion measurements. It is also noted that the benzamidinium
moiety contributes significantly to low-lying virtual orbitals
in ZnPA-1 and ZnPA-2 (Figure 5). Indeed, the LUMOs are
delocalized on both the porphyrin core and the benzamidini-
um moiety. In addition, the LUMO+2 is a pp* orbital lo-
calized on the benzamidinium moiety. For the free-base por-
phyrins, on the other hand, the HOMOꢀ3 (FbPC-1) and
HOMOꢀ4 (FbPC-2) are delocalized orbitals over the por-
phyrin core and the benzoate moiety. The delocalization
may contribute to the coupling terms VZnP-A and VC-FbP. Of
course, as the overall rate is likewise dependent on the cou-
pling through the term VA-C, the observation that the
through-bond energy transfer occurs at a significant rate
suggests that the through-hydrogen-bond electronic coupling
represented by the term VA-C is also significant.
Conclusion
The observed energy-transfer rates in dyads ZnPA-1·FbPC-1
and ZnPA-2·FbPC-2 were 4.0”109 and 1.3”109 sꢀ1, respec-
tively. The results corroborate the notion that the exchange
mechanism does operate in the energy-transfer processes for
the following reasons. First, these values are substantially
larger than those calculated assuming the Fçrster mecha-
nism (5.1”108 and 3.3”108 sꢀ1, respectively). Second, the
larger energy-transfer rate in ZnPA-1·FbPC-1 than in
ZnPA-2·FbPC-2 is consistent with the through-bond mecha-
nism, as the former dyad is more favorable for the through-
bond mechanism in terms of symmetries and energies of
frontier molecular orbitals as well as less encumbrance
around the phenyl groups along the linkage. These factors
would not exert any influence on the Fçrster energy transfer
except for a slight variation in the orientation factor. Thus,
this work has made it more certain that the exchange
energy transfer does operate in these dyads connected by
hydrogen bonds, or more specifically, the amidinium–car-
boxylate salt bridge. This conclusion means that hydrogen
bonds can provide an effective pathway for energy-transfer
processes via a through-bond mechanism. It has become
clear that the choice of noncovalent intermolecular linkage
is important not only in terms of structural aspects, but also
in terms of an electronic aspect for the design of supra-
molecular systems in which efficient energy transfer is de-
sired.
As the Zn- and free-base porphyrins are spatially separat-
ed in the dyad, direct electronic interactions between these
two chromophores is unlikely. Instead, a bridge-mediated
superexchange mechanism may be invoked.[49–51] In the su-
perexchange mechanism for the present system, the elec-
tronic coupling matrix element may be represented as
[Eq. (5)]:
VZnP-AVA-CVC-FbP
ð5Þ
V ¼
2
DEDB
in which V is the electronic coupling matrix element be-
tween the chromophores designated in the subscripts (A
and C represent benzamidinium and benzoate moieties, re-
spectively) and DEDB is the gap between the excited ener-
gies of the donor and the bridge. Thus, a larger electronic
coupling between consecutive components and a smaller
energy gap are more favorable for a superexchange-based
energy-transfer process.
The role of the energy gap was nicely demonstrated by
Albinsson and co-workers who examined singlet excited
energy-transfer rates in covalent Zn–porphyrin/bridge/free-
base porphyrin triads, in which the bridge is varied among
bicyclo[2.2.2]octane, benzene, naphthalene, and anthra-
G
1/2
cene.[52] They found a linear relation between kD and
ꢀ1
DEDB for the series. In our case, DEDB could be estimated
as approximately 14000 cmꢀ1 from the absorption edges for
ZnPA-2·benzoic acid (590 nm) and for benzamidinium/ben-
zoate (320 nm). However, as the HOMO/LUMO energies
are significantly offset between the benzamidinium and ben-
zoate moieties (see Figure 5), charge-transfer-type (ben-
Acknowledgements
This work was partially supported by the Tokyo Ohka Foundation for the
Promotion of Science and Technology and by the “High-Tech Research
Center” Project for Private Universities from the Ministry of Education,
Culture, Sports, Science, and Technology, Japan.
Chem. Eur. J. 2008, 14, 3776 – 3784
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3783

J. Otsuki et al.
[26]E. Iengo, E. Zangrando, E. Alessio, J.-C. Chambron, V. Heitz, L.
Flamigni, J.-P. Sauvage, Chem. Eur. J. 2003, 9, 5879–5887.
[27]A. Harriman, D. J. Magda, J. L. Sessler, J. Chem. Soc. Chem.
Commun. 1991, 345–348.
[28]A. Harriman, D. J. Magda, J. L. Sessler, J. Phys. Chem. 1991, 95,
1530–1532.
[29]J. L. Sessler, B. Wang, A. Harriman, J. Am. Chem. Soc. 1995, 117,
704–714.
[30]V. Krꢃl, S. L. Springs, J. L. Sessler, J. Am. Chem. Soc. 1995, 117,
8881–8882.
[31]S. L. Springs, D. Gosztola, M. R. Wasielewski, V. Krꢃl, A. Andriev-
sky, J. L. Sessler, J. Am. Chem. Soc. 1999, 121, 2281–2289.
[32]F. Hajjaj, Z. S. Yoon, M.-C. Yoon, J. Park, A. Satake, D. Kim, Y.
Kobuke, J. Am. Chem. Soc. 2006, 128, 4612–4623.
[33]H. Song, C. Kirmaier, J. K. Schwartz, E. Hindin, L. Yu, D. F. Bocian,
J. S. Lindsey, D. Holten, J. Phys. Chem. B 2006, 110, 19121–19130.
[34]J.-S. Hsiao, B. P. Krueger, R. W. Wagner, T. E. Johnson, J. K. Dela-
ney, D. C. Mauzerall, G. R. Fleming, J. S. Lindsey, D. F. Bocian, R. J.
Donohoe, J. Am. Chem. Soc. 1996, 118, 11181–11193.
[35]A. Osuka, N. Tanabe, S. Kawabata, I. Yamazaki, Y. Nishimura, J.
Org. Chem. 1995, 60, 7177–7185.
[36]J.-P. Strachan, S. Gentemann, J. Seth, W. A. Kalsbeck, J. S. Lindsey,
D. Holten, D. F. Bocian, J. Am. Chem. Soc. 1997, 119, 11191–11201.
[37]H. H. Tsai, M. C. Simpson, Chem. Phys. Lett. 2002, 353, 111–118.
[38]S. Kawabata, I. Yamazaki, Y. Nishimura, A. Osuka, J. Chem. Soc.
Perkin Trans. 2 1997, 479–484.
[39]A. Osuka, M. Ikeda, H. Shiratori, Y. Nishimura, I. Yamazaki, J.
Chem. Soc. Perkin Trans. 2 1999, 1019–1025.
[40]The fluorescence quantum yield and lifetime of ZnPA-2 were mea-
sured in the presence of benzoic acid.
[41]J. Mꢅrtensson, Chem. Phys. Lett. 1994, 229, 449–456.
[42]M. O. Senge, C. J. Medforth, T. P. Forsyth, D. A. Lee, M. M. Olm-
stead, W. Jentzen, R. K. Pandey, J. A. Shelnutt, K. M. Smith, Inorg.
Chem. 1997, 36, 1149–1163.
[43]D. Papoutsakis, J. P. Kirby, J. E. Jackson, D. G. Nocera, Chem. Eur.
J. 1999, 5, 1474–1480.
[44]M. Gouterman, J. Mol. Spectrosc. 1961, 6, 138–163.
[45]A value of ꢁ458 was reported: Y. Okuno, T. Kamikado, S. Yokoya-
ma, S. Mashiko, J. Mol. Struct. (Theochem) 2002, 594, 55–60.
[46]M. J. Crossley, L. D. Field, A. J. Forster, M. M. Harding, S. Sternhell,
J. Am. Chem. Soc. 1987, 109, 341–348.
[47]L. Noss, P. A. Liddell, A. L. Moore, T. A. Moore, D. Gust, J. Phys.
Chem. B 1997, 101, 458–465.
[48]A. Helms, D. Heiler, G. McLendon, J. Am. Chem. Soc. 1991, 113,
4325–4327.
[1]V. Balzani, S. Campagna, G. Denti, A. Juris, S. Serroni, M. Venturi,
Acc. Chem. Res. 1998, 31, 26–34.
[2]D. Gust, T. A. Moore, A. L. Moore, Acc. Chem. Res. 2001, 34, 40–
48.
[3]D. Holten, D. F. Bocian, J. S. Lindsey, Acc. Chem. Res. 2002, 35, 57–
69.
[4]M.-S. Choi, T. Yamazaki, I. Yamazaki, T. Aida, Angew. Chem. 2004,
116, 152–160; Angew. Chem. Int. Ed. 2004, 43, 150–158.
[5]Y. Kobuke, Eur. J. Inorg. Chem. 2006, 2333–2351.
[6]R. Berera, C. Herrero, I. H. M. van Stokkum, M. Vengris, G. Kodis,
R. E. Palacios, H. van Amerongen, R. van Grondelle, D. Gust, T. A.
Moore, A. L. Moore, J. T. M. Kennis, Proc. Natl. Acad. Sci. USA
2006, 103, 5343–5348.
[7]B. Zietz, V. I. Prokhorenko, A. R. Holzwarth, T. Gillbro, J. Phys.
Chem. B 2006, 110, 1388–1393.
[8]R. Berera, I. H. M. van Stokkum, G. Kodis, A. E. Keirstead, S.
Pillai, C. Herrero, R. E. Palacios, M. Vengris, R. van Grondelle, D.
Gust, T. A. Moore, A. L. Moore, J. T. M. Kennis, J. Phys. Chem. B
2007, 111, 6868–6877.
[9]T. Polivka, M. Pellnor, E. Melo, T. Pascher, V. Sundstroem, A.
Osuka, K. R. Naqvi, J. Phys. Chem. C 2007, 111, 467–476.
[10]F. Scandola, V. Balzani, J. Chem. Educ. 1983, 60, 814–823.
[11]T. Fçrster, Discuss. Faraday Soc. 1959, 27, 7–17.
[12]D. L. Dexter, J. Chem. Phys. 1953, 21, 836–850.
[13]J. Otsuki, K. Iwasaki, Y. Nakano, M. Itou, Y. Araki, O. Ito, Chem.
Eur. J. 2004, 10, 3461–3466.
[14]J. A. Roberts, J. P. Kirby, D. G. Nocera, J. Am. Chem. Soc. 1995, 117,
8051–8052.
[15]Y. Deng, J. A. Roberts, S.-M. Peng, C. K. Chang, D. G. Nocera,
Angew. Chem. 1997, 109, 2216–2219; Angew. Chem. Int. Ed. Engl.
1997, 36, 2124–2127.
[16]J. P. Kirby, J. A. Roberts, D. G. Nocera, J. Am. Chem. Soc. 1997, 119,
9230–9236.
[17]N. H. Damrauer, J. M. Hodgkiss, J. Rosenthal, D. G. Nocera, J. Phys.
Chem. B 2004, 108, 6315–6321.
[18]J. Rosenthal, J. M. Hodgkiss, E. R. Young, D. G. Nocera, J. Am.
Chem. Soc. 2006, 128, 10474–10483.
[19]J. M. Hodgkiss, N. H. Damrauer, S. Pressꢂ, J. Rosenthal, D. G.
Nocera, J. Phys. Chem. B 2006, 110, 18853–18858.
[20]J. Otsuki, M. Takatsuki, M. Kaneko, H. Miwa, T. Takido, M. Seno,
K. Okamoto, H. Imahori, M. Fujitsuka, Y. Araki, O. Ito, S. Fukuzu-
mi, J. Phys. Chem. A 2003, 107, 379–385.
[21]J. Otsuki, T. Narita, K. Tsutsumida, M. Takatsuki, M. Kaneko, J.
Phys. Chem. A 2005, 109, 6128–6134.
[22]J. Otsuki, Y. Kato, T. Sasahara, K. Iwasaki, A. Yasuda, Chem. Lett.
2005, 34, 1684–1685.
[23]L. Sꢃnchez, M. Sierra, N. Martꢄn, A. J. Myles, T. J. Dale, J. Re-
bek, Jr., W. Seitz, D. M. Guldi, Angew. Chem. 2006, 118, 4753–4757;
Angew. Chem. Int. Ed. 2006, 45, 4637–4641.
[49]H. H. McConnell, J. Chem. Phys. 1961, 35, 508–515.
[50]M. D. Newton, Chem. Rev. 1991, 91, 767–792.
[51]S. Speiser, Chem. Rev. 1996, 96, 1953–1976.
[52]K. Kilsꢅ, J. Kajanus, J. Mꢅrtensson, B. Albinsson, J. Phys. Chem. B
1999, 103, 7329–7339.
[24]I. V. Rubtsov, Y. Kobuke, H. Miyaji, K. Yoshihara, Chem. Phys.
Lett. 1999, 308, 323–328.
[25]A. Ambroise, J. Li, L. Yu, J. S. Lindsey, Org. Lett. 2000, 2, 2563–
2566.
Received: September 19, 2007
Published online: February 27, 2008
3784
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3776 – 3784