Long-lived charge separation in novel axial donor-porphyrin-acceptor triads based on tetrathiafulvalene, aluminum(III) porphyrin and naphthalenediimide

Article abstract of DOI:10.1002/chem.201202995

Two self-assembled supramolecular donor-acceptor triads consisting of Al^III porphyrin (AlPor) with axially bound naphthalenediimide (NDI) as an acceptor and tetrathiafulvalene (TTF) as a secondary donor are reported. In the triads, the NDI and TTF units are attached to Al^III on opposite faces of the porphyrin, through covalent and coordination bonds, respectively. Fluorescence studies show that the lowest excited singlet state of the porphyrin is quenched through electron transfer to NDI and hole transfer to TTF. In dichloromethane hole transfer to TTF dominates, whereas in benzonitrile (BN) electron transfer to NDI is the main quenching pathway. In the nematic phase of the liquid crystalline solvent 4-(n-pentyl)-4′-cyanobiphenyl (5CB), a spin-polarized transient EPR spectrum that is readily assigned to the weakly coupled radical pair TTF^.+NDI^.- is obtained. The initial polarization pattern indicates that the charge separation occurs through the singlet channel and that singlet-triplet mixing occurs in the primary radical pair. At later time the polarization pattern inverts as a result of depopulation of the states with singlet character by recombination to the ground state. The singlet lifetime of TTF^.+NDI^.- is estimated to be 200-300 ns, whereas the triplet lifetime in the approximately 350 mT magnetic field of the X-band EPR spectrometer is about 10 μs. In contrast, in dichloromethane and BN the lifetime of the charge separation is <10 ns. Long live the charge separation: New biomimetic donor-acceptor triads have been constructed by taking advantage of the ability of aluminum in Al^III porphyrin to form axial covalent and coordination bonds with carboxylic acids and Lewis bases, respectively (see figure; NDI=naphthalenediimide, TTF=tetrathiafulvalene). Upon light excitation the triads undergo sequential electron transfer and by using a liquid-crystalline solvent and high magnetic field, the lifetime of the charge separation can be extended into the microsecond range. Copyright

Full text of DOI:10.1002/chem.201202995

FULL PAPER
DOI: 10.1002/chem.201202995
Long-Lived Charge Separation in Novel Axial Donor–Porphyrin–Acceptor
Triads Based on Tetrathiafulvalene, Aluminum
Naphthalenediimide
ACHTUNGTREN(NUGN III) Porphyrin and
Prashanth K. Poddutoori,*[a] Niloofar Zarrabi,[a] Andrey G. Moiseev,[b]
Roger Gumbau-Brisa,[a] Serguei Vassiliev,[c] and Art van der Est*[a]
Abstract: Two self-assembled supramo-
dominates, whereas in benzonitrile
(BN) electron transfer to NDI is the
main quenching pathway. In the nemat-
ic phase of the liquid crystalline solvent
4-(n-pentyl)-4’-cyanobiphenyl (5CB), a
spin-polarized transient EPR spectrum
occurs through the singlet channel and
lecular donor–acceptor triads consisting
of AlIII porphyrin (AlPor) with axially
bound naphthalenediimide (NDI) as an
acceptor and tetrathiafulvalene (TTF)
as a secondary donor are reported. In
the triads, the NDI and TTF units are
attached to AlIII on opposite faces of
the porphyrin, through covalent and
coordination bonds, respectively. Fluo-
rescence studies show that the lowest
excited singlet state of the porphyrin is
quenched through electron transfer to
NDI and hole transfer to TTF. In di-
chloromethane hole transfer to TTF
that singlet–triplet mixing occurs in the
primary radical pair. At later time the
polarization pattern inverts as a result
of depopulation of the states with sin-
glet character by recombination to the
that is readily assigned to the weakly
ground state. The singlet lifetime of
+
+
C
Cꢀ
C
Cꢀ
coupled radical pair TTF NDI is ob-
tained. The initial polarization pattern
indicates that the charge separation
TTF NDI is estimated to be 200–
300 ns, whereas the triplet lifetime in
the approximately 350 mT magnetic
field of the X-band EPR spectrometer
is about 10 ms. In contrast, in dichloro-
methane and BN the lifetime of the
charge separation is <10 ns.
Keywords:
aluminum
·
electron transfer · photosynthesis ·
porphyrinoids · tetrathiafulvalene
Introduction
gether a photosensitizer and a series of redox-active compo-
nents in a supermolecular complex capable of generating
long-lived high-energy charge-separated states with high
quantum yield. Optimizing the light-induced charge separa-
tion in donor–acceptor (D–A) complexes requires simulta-
neous optimization of factors such as the electron-accepting
and -donating abilities and reorganization energies of the
components as well as the electronic coupling between
them. These factors can be varied by changing the number
of components, the distance between them, their redox and
optical properties, as well as their structures and orientation
with respect to each other.
In photosynthetic organisms, solar energy is converted into
energy-rich compounds by a highly efficient light-induced
multi-step electron-transfer reaction in the reaction center
protein complexes that is coupled to a series of exogenous
redox reactions.[1] There is widespread interest in trying to
mimic this process in artificial photosynthetic systems capa-
ble of converting solar energy into useful forms of energy.[2]
The main challenge in designing such systems is to bring to-
In recent years, many such donor–acceptor systems have
been synthesized[2m,3] and in most of them, one or more por-
phyrins adopt the role played by the chlorophylls in the nat-
ural systems. Porphyrins are an attractive choice for the pho-
tosensitizer because they strongly absorb light within the
visible region, are often highly fluorescent, and have rich
redox chemistry. More importantly, however, their optical
and redox properties are also easily tunable.[4] In combina-
tion with other redox-active species, especially electron ac-
ceptors, such as quinones and fullerenes, porphyrins have
been used to produce a wide array of covalent and non-co-
valent D–A conjugates and supramolecular assem-
[a] Dr. P. K. Poddutoori, N. Zarrabi, R. Gumbau-Brisa,
Prof. A. van der Est
Department Chemistry, Brock University
500 Glenridge Ave., St.Catharines, ON, L2S 3A1 (Canada)
Fax : (+1)905-682-9020
E-mail: ppodduto@brocku.ca
avde@brocku.ca
[b] Dr. A. G. Moiseev
Department of Chemistry, McGill University
801 Sherbrook Street West, Montreal QC, H3A 2K6 (Canada)
[c] Dr. S. Vassiliev
Department of Biological Sciences, Brock University
500 Glenridge Ave., St.Catharines, ON, L2S 3A1 (Canada)
blies.[2g,3c,f,h,5] Among the porphyrins, aluminum
ACHTNUTRGNE(NUG III) porphyr-
in (AlPor) is unique, because it can form two different axial
bonds that allow two different types of molecular compo-
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201202995.
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

nents to be attached. Carboxylic acids react with the axial
hydroxide group of AlPor-OH to form covalent ester linkag-
es, whereas Lewis bases, such as pyridine and imidazole,
form coordination bonds to the Al center, which is a Lewis
acid.[6] AlPor also has moderate oxidation and reduction po-
tentials, and therefore could be used as a primary electron
donor and/or acceptor.[6h,7] This combination of properties
makes AlPor an ideal candidate
factors governing the electron transfer in axial triads. As the
acceptor, we have chosen naphthalenediimide (NDI) be-
cause it is structurally very different from fullerene and has
been used in porphyrin-based D–A complexes before.[6d,8c,10]
Attaching NDI covalently to AlPor gives the new axial dyad
aluminumACHTUNGRTEN(NUNG III) porphyrin–naphthalenediimide (AlPor-Ph-
NDI; Ph=phenyl) shown in Figure 1. This dyad can then be
for constructing multi-compo-
nent D–A systems in which the
electron-transfer pathway is
perpendicular to the porphyrin
plane. One advantage of such
an arrangement is that the
donor and acceptor sides of the
complex are located on oppo-
site faces of the porphyrin,
which maximizes the spatial
separation of the charges pro-
duced by the electron transfer.
Moreover, the p* orbital of the
porphyrin resides in the same
region of space as the axial li-
gands, which should increase
the electronic coupling between
the excited state of the porphyr-
in and the acceptor and, hence,
promote electron transfer.
Figure 1. Molecular components (AlPor-Ph-NDI and TTF-py or TTF-Ph-py) for construction of self-assembled
axially connected vertical triads. The reference compound AlPor-Ph is also shown. TTF=tetrathiafulvalene.
There are a number of exam-
ples reported in the literature of D–A systems in which the
direction of the electron transfer is axial to the porphyrin
plane.[3c,6d,f–h,8] Most of these axial systems show efficient
electron transfer between the donor and acceptor compo-
nents. However, the factors governing electron transfer in
these types of complexes have not yet been fully explored.
Recently, we reported the first supramolecular triads with
an entirely axial electron-transfer pathway. The triads were
based on AlPor with fullerene (C60) covalently attached to
AlIII and phenothiazine (PTZ) (or 2-methylthiophenothia-
zine (TPTZ)) coordinated to AlIII through an appended pyr-
idine (py). We were able to show that excitation of the por-
phyrin led to an electron transfer to C60 followed by a hole
transfer to PTZ (or TPTZ).[6h] The electron-transfer rates
and DG values for the forward and reverse electron transfer
between AlPor and axially bound C60 were found to lie very
close to the Marcus curves determined for an equatorially
bound zinc porphyrin–fullerene dyad,[9] which suggested that
the placement of the C60 in the axial position did not lead to
the expected increase in the electronic coupling. However, it
is unclear whether this is a result of the unique properties of
fullerene and whether other factors, such as the nature of
the bridging group, may be critical in determining the cou-
pling. Understanding the role of these factors is crucial if
long-lived charge separation is to be achieved.
converted to a triad by coordinating a py-appended donor.
We have chosen tetrathiafulvalene as the secondary donor
because of its strong electron-donating ability, which makes
it an excellent candidate as a reductive electron quencher or
hole acceptor in D–A systems.[11] Despite its rich redox
chemistry, there are only a few reports of TTF in porphyrin-
containing D–A systems[12] and very little is known about its
efficiency as an electron donor when attached to a porphyr-
in through a coordination linkage.[13] The resulting triads,
TTF-py!AlPor-Ph-NDI and TTF-Ph-py!AlPor-Ph-NDI,
are shown in Figure 2. In these triads four important varia-
tions have been introduced compared to the PTZ-AlPor-C60
triads studied previously:[6h] 1) TTF has a lower oxidation
potential than PTZ or TPTZ and hence, the driving force
for electron donation to the porphyrin cation AlPor+ or ex-
cited singlet state 1AlPor*) is greater, 2) NDI is a much
smaller electron acceptor than C60 and has a very different
reorganization energy, 3) TTF is appended directly to the py
unit to reduce the flexibility of the bridge and ensure a
linear electron-transfer pathway, and 4) two different py-ap-
pended TTF donors have been prepared to allow the com-
parison of analogous complexes with different distances be-
tween the secondary donor and AlPor. We will show that ef-
ficient electron transfer occurs in the triads but that the life-
time of the charge-separated states is very strongly influ-
enced by the nature of the solvent and is only weakly
dependent on the length of the bridging group. We also
present DFT calculations, which, along with the optical
Here, we present the preparation and characterization of
two AlPor-based triads in which the nature of the donor, the
acceptor, and the bridging groups are varied to study the
&
2
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Charge Separation in Novel Axial Donor–Porphyrin–Acceptor Triads
FULL PAPER
upfield in AlPor-Ph-NDI com-
pared with unbound NDI-Ph-
COOH as indicated in Figure 3.
The signal of the naphthalic
protons c and d, which appears
as a singlet in unbound NDI-
Ph-COOH, is split into two
doublets with a small upfield
shift in the dyad. The observed
upfield shifts are due to the
shielding effect of the porphyr-
in ring and the magnitude of
the shift depends on the dis-
tance of the proton from the
porphyrin ring. As shown in
Figure 2, the protons in position
a are closer to the AlPor ring
than protons in position b and
hence, they experience a larger
shift. Similarly, protons c and d
exhibit smaller shifts because
they are further away from por-
phyrin ring.
Comparison of the 1H NMR
spectrum of a 1:1 mixture of
TTF-Ph-py and AlPor-Ph-NDI
Figure 2. Formation of self-assembled axially connected vertical supramolecular triads (R=phenyl).
spectra, indicate that the interaction of the lowest p* orbital
of the porphyrin with the bridge differs significantly, de-
pending on whether it is covalently bound or coordinated to
AlPor.
(Figure 3D) with the spectra of AlPor-Ph-NDI (Figure 3B)
and TTF-Ph-py (Figure 3C) confirms the expected coordina-
tion of the py group of TTF-Ph-py to the AlPor. In the triad
complex, shielding due to the porphyrin ring current causes
an upfield shift of the protons of the py moiety (protons e
and f), of the phenyl spacer (protons g and h), and of the
TTF unit (protons i and j). The pyridyl protons e and f dis-
play the greatest shift indicating that coordination occurs
through the py group. On the other hand, protons a and b
Results and Discussion
Synthesis: The assembly of the triads by coordination of py-
appended TTF derivatives to the dyad AlPor-Ph-NDI is
shown in Figure 2. Details of the synthetic procedure for
AlPor-Ph-NDI and the TTF derivatives (TTF-py and TTF-
Ph-py) are given in the Supporting Information. Briefly,
AlPor-Ph-NDI was prepared in quantitative yield by con-
densation
of
5,10,15,20-tetraACTHNGUTER(NNUG phenyl)porphyrinato-
aluminumACHTUNGTRENNUNG(III)hydroxide (AlPor-OH) with carboxylic acid
functionalized NDI (NDI-Ph-COOH) (Scheme S2 in the
Supporting Information) and the py-appended tetrathiaful-
valene derivatives (TTF-py and TTF-Ph-py) were prepared
by Stille and/or Suzuki coupling reactions.[14] The synthesis
and characterization of the reference compound, AlPor-Ph,
is reported in the literature.[6g]
The FAB mass spectrum of AlPor-Ph-NDI showed an in-
tense molecular ion [M]+ peak at m/z 1137 as well as a frag-
ment ion peak at m/z 639 due to [Mꢀ(NDI-Ph-COO)]+ for-
mation. The molecular ion peaks of TTF-py and TTF-Ph-py
were observed at m/z 281 and 357, respectively. Figure 3
Figure 3. 1H NMR spectra of A) NDI-Ph-COOH, B) AlPor-Ph-NDI,
C) TTF-Ph-py, and D) TTF-Ph-py!AlPor-Ph-NDI (all samples 5.6 mm
in CDCl3). In the case of NDI-Ph-COOH few drops of CD3OD were
added to dissolve the sample. NMR solvent or solvent impurity peaks are
indicated with an asterisk.
1
shows a comparison of the H NMR spectra of AlPor-Ph-
NDI (Figure 3B) and the unbound ligand NDI-Ph-COOH
(Figure 3A). The ring protons a and b of the benzoate
spacer (see Figure 2 for the proton assignments) are shifted
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
3
&
ÞÞ
These are not the final page numbers!

P. K. Poddutoori, A. van der Est et al.
of the bridging benzoate group between AlPor and NDI
show an increased upfield shift when TTF-Ph-py coordi-
nates. It is likely that the AlIII center lies out of the porphyr-
in plane when it is five-coordinate but that it moves closer
to the porphyrin plane when it becomes six-coordinate and
hence, the benzoate spacer also moves closer to the ring.
Similar spectra are obtained for a 1:1 mixture of TTF-py
and AlPor-Ph-NDI (data not shown) but with larger upfield
shifts for the TTF protons as expected because the shorter
spacer means they lie closer to the porphyrin ring.
UV/Vis absorption studies: The absorption spectra of
AlPor-Ph-NDI and the reference compounds NDI-Ph and
AlPor-Ph measured in dichloromethane shown in Figure 4.
Figure 5. Titration of AlPor-Ph-NDI with TTF-Ph-py in dichloromethane.
The inset shows the Benesi–Hildebrand plot of the absorbance change at
l=604 nm. TTF-Ph-py was added up to 2.6ꢁ10ꢀ3 m in 20 mL (2.4ꢁ10ꢀ4 m)
increments to 1 mL (6ꢁ10ꢀ5 m) of a solution of AlPor-Ph-NDI.
tic point is observed at l=555 nm, indicating complex for-
mation. The observed shifts in the porphyrin bands are typi-
cal of axial coordination of nitrogen ligands to AlPor.[6b,h,15]
Benesi–Hildebrand[16] analysis (Figure 5, inset) gives a linear
plot confirming that a 1:1 complex is formed and the slope
yields the binding constant K=1.1ꢁ103 mꢀ1. Similar spectral
changes and similar binding constants were observed in ti-
trations of TTF-py versus AlPor-Ph-NDI, py versus AlPor-
Ph-NDI, TTF-py versus AlPor-Ph, TTF-Ph-py versus AlPor-
Ph, and py versus AlPor-Ph (data not shown). The absence
of any additional bands in these titrations suggests that per-
turbation of the electronic structures of the photo- and
redox-active components due to formation of the complex is
relatively small. However, compared to covalent attachment
axial coordination causes a greater perturbation of the por-
phyrin. The same titrations were also performed in benzoni-
trile (BN), but because the solvent can also coordinate to
Figure 4. UV/Vis absorption spectra of NDI-Ph (dotted line), AlPor-Ph
(dashed line), and AlPor-Ph-NDI (solid line) in dichloromethane. The
inset shows the absorption spectra of TTF-py and TTF-Ph-py in dichloro-
methane.
The absorption maxima and extinction coefficients (lmax and
loge) are included with the synthetic procedures given in
the Supporting Information. The absorption spectrum of
AlPor-Ph-NDI shows the typical very strong B-band (Soret
band) at l=415 nm and a weak Q-band at l=547 nm from
AlPor and weak p–p* bands at l=342, 360, and 381 nm
from NDI. The spectrum is essentially a superposition of the
spectra of the individual chromophores indicating that in
the covalently linked dyad there is little or no perturbation
of the ground-state electronic structures of the chromo-
phores. The py-appended TTF derivatives, which will be
used to construct the self-assembled supramolecular triads,
have relatively weak and very broad absorption bands (see
inset in Figure 4) at l=304 (average of 285 and 323 nm
bands) and l=432 nm for TTF-py, and at l=299 and
427 nm for TTF-Ph-py.
ꢀ ꢁ
the porphyrin (Ph C N!AlPor) complexation by the py-
appended TTF does not lead to spectral shifts and the bind-
ing constants could not be determined. However, because
py is a better Lewis base than BN, complex formation is ex-
pected.
Density functional modeling calculations: The geometries
and electronic structures of the AlPor-Ph-NDI dyad and of
the self-assembled dyads and triads were investigated by
using B3LYP/3-21G*[17] level DFT calculations similar to
those carried out in previous studies of covalently linked
and self-assembled donor–acceptor systems.[3a,b,18] The opti-
mized structures of the dyads and triads are shown in
Figure 6 as well as Figures S7 and S8 in the Supporting In-
formation. In the calculations, the 2-ethylhexyl chain of the
NDI was truncated to simplify the optimization. This trunca-
tion is not expected to have any significant influence on the
calculated geometry of the dyad AlPor-Ph-NDI. The center-
to-center distances between the AlPor (Al center), the TTF
unit, and the NDI moiety are summarized in Table S1 in the
Supporting Information. The optimized structures also pro-
Figure 5 shows absorption titrations of AlPor-Ph-NDI
with TTF-Ph-py in dichloromethane. The Q-band region
was used to monitor the absorbance of the porphyrin in the
titrations because the TTF-py absorption band (lmax
=
432 nm) and the AlPor Soret band (lmax =415 nm) overlap.
Upon adding TTF-py, the Q-bands of the porphyrin shift
from l=546 and 585 nm to l=562 and 602 nm. An isosbes-
&
4
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Charge Separation in Novel Axial Donor–Porphyrin–Acceptor Triads
FULL PAPER
Figure 6. Frontier HOMO, HOMOꢀ1, and LUMO of a) dyad AlPor-Ph-NDI, b) dyad TTF-Ph-py!AlPor-Ph, and c) triad TTF-Ph-py!AlPor-Ph-NDI
calculated by ab initio B3LYP/3-21G* methods.
vide the radii of the AlPor, the NDI moiety, and the TTF
unit, which are found to be RAlPor =7.77, RNDI =6.85, and
to the migration of the hole to the TTF unit. In the case of
the triad TTF-Ph-py!AlPor-Ph-NDI (Figure 6c), the
HOMO, HOMO ꢀ1, and LUMO are located on the TTF,
AlPor, and NDI units as in the reference compounds. Thus,
excitation of the porphyrin can be expected to result in mi-
gration of the electron to the LUMO on the NDI moiety
and the hole to the HOMO on the TTF unit.
RTTF =5.47 ꢂ. As expected the HOMO and LUMO of
AlPor-Ph-NDI (Figure 6a) are localized on the porphyrin
macrocycle and the NDI unit, respectively, which implies
+
C
Cꢀ
that AlPor -NDI is the lowest excited state of the com-
plex. The computed gas-phase HOMO–LUMO gap is found
to be 2.05 eV but this value drops to 1.71 eV if higher level
calculations (e.g., B3LYP/6-31G
(d,p)/6-311G+(d,p)) are
Electrochemistry: Determination of the redox potentials of
the compounds is important for evaluating the energetics of
the electron-transfer reactions and possible electronic inter-
actions between the constituents. Figure 7 shows the cyclic
voltammograms of the dyad AlPor-Ph-NDI, its reference
compounds AlPor-Ph and NDI-Ph as well as the triad com-
ponents TTF-py and TTF-Ph-py in dichloromethane con-
taining 0.1m tetrabutylammonium perchlorate (TBAP). The
voltammogram for the dyad AlPor-Ph-NDI (Figure 7,
bottom) is essentially a sum of the traces for the AlPor-Ph
performed. For the dyad TTF-Ph-py!AlPor-Ph (Figure 6b)
the HOMO and LUMO are localized primarily on the TTF
and AlPor units, again as expected. However, in both cases
some delocalization onto the py bridge occurs. This delocali-
zation is more pronounced in TTF-py!AlPor-Ph (Figure S8
in the Supporting Information), which does not have the
phenyl group in the bridge. These HOMO and LUMO dis-
tributions suggest that excitation of the porphyrin, which re-
moves an electron from the HOMOꢀ1 orbital, would lead
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
5
&
ÞÞ
These are not the final page numbers!

P. K. Poddutoori, A. van der Est et al.
the solvent used for the photophysical studies (9.1 and 26
for dichloromethane and BN, respectively); eR is the dielec-
tric constant of the solvent used for measuring the redox po-
tentials, in this case, dichloromethane.
By using the radii from the DFT calculations, GS values
of ꢀ0.14 and ꢀ0.19 eV for AlPor·+-NDI·ꢀ in dichlorome-
thane and BN, respectively, are obtained. This leads to an
energy of 1.28 V for AlPor·+-NDI·ꢀ in dichloromethane,
which is lower than the gas-phase value of ꢀ1.71 eV calcu-
lated by DFT. This difference suggests that significant stabi-
lization of the charges by the surrounding solvent occurs. By
using, the energies obtained from the optical and electro-
chemical data, the energy-level diagram shown in Figure 8
Figure 7. Cyclic voltammograms of TTF-py, TTF-Ph-py, NDI-Ph, AlPor-
Ph, and AlPor-Ph-NDI in dichloromethane containing 0.1m TBAP. All
samples were measured with a scan rate of 100 mVsꢀ1, except TTF-py,
which was measured with a scan rate of 500 mVsꢀ1
.
and NDI-Ph references. Hence, the first two reduction proc-
esses at ꢀ0.55 and ꢀ1.00 V are assigned to the first and
second reduction of the NDI unit and the third and fourth
reduction processes at ꢀ1.16 and ꢀ1.54 V are the first and
second reduction of the AlPor unit. The oxidation scan is
virtually identical to that of AlPor-Ph and the observed two
processes are assigned to the first and second oxidations of
the porphyrin. As expected, the TTF derivatives show two
processes corresponding to the first and second oxidation.
The difference between the first oxidation potential of the
donor and the first reduction potential of the acceptor corre-
sponds to the HOMO–LUMO gap and is the energy of the
charge-separated state ECS. The free-energy changes for
charge separation (DGCS), charge recombination (DGCR),
and hole stabilization (DGHS) are given by Equations (1)
and (2):[19]
Figure 8. Energy-level diagram for AlPor-Ph-NDI, TTF-py!AlPor-Ph
(ECS =1.47 eV), TTF-Ph-py!AlPor-Ph (ECS =1.52 eV), TTF-py!AlPor-
Ph-NDI, and TTF-Ph-py!AlPor-Ph-NDI in dichloromethane.
was constructed for the dyads and triads in dichlorome-
thane. The free energy changes for charge separation ob-
tained from Equation (1) are summarized in Table 1 and
they suggest that the photoinduced electron transfer from
1AlPor* to NDI and hole stabilization onto TTF as well as
the electron shift to NDI are exothermic. The driving force
DGCS ¼ ECSꢀE0ꢀ0þGS
DGCR ¼ ꢀECS
ð1Þ
ð2Þ
of the hole shift (DGHS) and electron shift (DGES) from
+
C
Cꢀ
AlPor and AlPor to TTF and NDI, respectively, are esti-
mated from the difference of their corresponding ion-pair
energies.
where E0–0 is the singlet-state energy of AlPor (2.14 eV), es-
timated from the blue edge of the fluorescence spectrum at
l=579 nm. GS is the ion-pair stabilization and incorporates
both the solvent-dependent Coulomb energy change upon
ion-pair formation or recombination and the free energy of
solvation of the ions [Eq. (3)]:
Fluorescence studies: The steady-state fluorescence spectra
of the dyad AlPor-Ph-NDI and its reference compound
AlPor-Ph were measured with the excitation l=550 nm in
dichloromethane and BN. The only significant absorbance at
l=550 nm is by the porphyrin and the concentration of the
solution is the same for all of samples. Hence, the spectra
(Figure 9A) reflect the fluorescence yield following excita-
tion of the porphyrin. In each solvent, the band positions
are the same for the dyad (Figure 9A, dashed lines) and the
reference compound (Figure 9A, solid lines) but the fluores-
cence is quenched in the dyad. The positions and relative in-
tensities of the bands differ in the two solvents BN and di-
GS ¼ e2=ð4pe0Þ½ð1=ð2 RþÞþ1=ð2RꢀÞꢀ1=RDꢀAÞ1=eSꢀð1=ð2RþÞ
þ1=ð2 RꢀÞÞ1=eRꢂ
ð3Þ
where R+, Rꢀ, and RD–A are the donor radius, the acceptor
radius, and the center-to-center distance between the donor
and the acceptor, respectively; eS is the dielectric constant of
&
6
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Charge Separation in Novel Axial Donor–Porphyrin–Acceptor Triads
FULL PAPER
Table 1. Energies of the charge-separated states (ECS) and the free-energy changes of the electron transfer (ꢀDGCS), the hole shift (ꢀDGHS), and the
electron shift (ꢀDGES) for the dyads and triads in dichloromethane.[a]
Sample
ECS
(AlPor -NDI
ECS
(TTF -NDI
ECS
(TTF -AlPor
ꢀDGCS
ꢀDGCS
ꢀDGHS
ꢀDGES
+
+
+
+
+
(1AlPor*!
(1AlPor*!
(AlPor -NDI
!
(TTF -AlPor
!
C
Cꢀ
C
Cꢀ
C
Cꢀ
C
Cꢀ
)
C
Cꢀ
)
)
)
)
+
+
+
+
C
Cꢀ
C
Cꢀ
C
Cꢀ
C
Cꢀ
AlPor -NDI
)
TTF -AlPor
)
TTF -NDI
TTF -NDI
AlPor-Ph-NDI
TTF-py!
1.28
1.28
–
0.96
–
1.47
0.86
0.86
–
0.67
–
0.32
–
0.51
AlPor-Ph-NDI
TTF-Ph-py!
AlPor-Ph-NDI
1.28
0.96
1.52
0.86
0.62
0.32
0.56
C
Cꢀ
(A/A )+GS, where E
C
Cꢀ
(A/A ) is the first reduction potential of the ac-
[a] ECS =E
(D +/D)ꢀE
E
(D +/D) is the first oxidation potential of the donor, E
G
+
+
+
C
Cꢀ
C
Cꢀ
C
Cꢀ
ceptor, E0–0 is the singlet-state energy of AlPor (2.14 eV), ꢀDGHS =ECS(AlPor -NDI )ꢀECS(TTF -NDI ), and ꢀDGES
=
ECS(TTF -AlPor )
C
Cꢀ
ꢀECS(TTF -NDI ). All energies are given in [eV].
found to be ꢀ0.86 and
ꢀ0.91 eV in dichloromethane
and BN, respectively. Conse-
quently, electron transfer from
1AlPor* to NDI is the more
likely explanation for the
quenching of the fluorescence
of AlPor in AlPor-Ph-NDI.
Figures 9B–D show fluores-
cence titrations of the AlPor-
Ph-NDI dyad in the presence of
increasing amounts of py, TTF-
py, and TTF-Ph-py. The titra-
tions were carried out in di-
chloromethane and the samples
were excited at the isosbestic
point (l=555 nm, see Figure 4).
Upon addition of py to AlPor-
Ph-NDI, (Figure 9B) a red shift
of the fluorescence bands and a
change in their relative intensi-
ties occurs. However, no
quenching occurs and isosbestic
points are observed in this fluo-
rescence titration. In contrast,
addition of TTF-py to AlPor-
Ph-NDI (Figure 9C) leads to
quenching of the fluorescence
without any significant shift of
the emission bands. Titration of
Figure 9. Steady state fluorescence spectra. A) AlPor-Ph (solid lines) and AlPor-Ph-NDI (dashed lines) in di-
chloromethane and BN. The concentration is the same for all four spectra. B–D) Fluorescence titrations of
AlPor-Ph-NDI with B) py, C) TTF-py, and D) TTF-Ph-py in dichloromethane. Py, TTF-py, or TTF-Ph-py were
added up to 2.3ꢁ10ꢀ3 m in 20 mL (2.22ꢁ10ꢀ4 m) increments to 1 mL (6ꢁ10ꢀ5 m) of a solution of AlPor-Ph-NDI.
The excitation wavelength is l=555 nm.
chloromethane, because BN coordinates to the Al center
but dichloromethane does not. The observed fluorescence
quenching suggests that the excited state of the porphyrin
could be depopulated by processes such as energy and elec-
tron transfer between the components. Because there is no
spectral overlap between the fluorescence bands of AlPor in
the red part of the visible region (Figure 9) and the absorp-
tion bands of NDI in the ultraviolet (Figure 4), the yield of
the energy transfer from the singlet excited state of AlPor
AlPor-Ph-NDI with TTF-Ph-py (Figure 9D) shows both
quenching and a shift in the band positions. Corresponding
titrations with AlPor-Ph instead of AlPor-Ph-NDI show a
similar behavior (Figure S9 in the Supporting Information).
The quenching without any shift of the band positions ob-
served with TTF-py (Figure 9C) implies that the coordina-
tion complex TTF-py!AlPor-Ph-NDI has negligible fluo-
rescence. In contrast, TTF-Ph-py!AlPor-Ph-NDI (Fig-
ure 9D) is fluorescent but its fluorescence is weaker than
that of AlPor-Ph-NDI. Because coordination of AlPor-Ph
and AlPor-Ph-NDI with py does not cause significant
quenching, the strong quenching observed with TTF-py is
not simply a result of coordination by the py group. More-
(1AlPor*) to NDI is expected to be negligible. On the other
+
C
hand, the redox data (see Table 1) suggest that AlPor
-NDI lies energetically below 1AlPor* in both solvents.
Cꢀ
The free-energy change for charge separation (DGCS) is
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
7
&
ÞÞ
These are not the final page numbers!

P. K. Poddutoori, A. van der Est et al.
over, control experiments in which AlPor-Ph was titrated
against TTF without an appended py group (Figure S12 in
the Supporting Information), showed only a small amount
of fluorescence quenching. Stern–Volmer plots of the AlPor
fluorescence monitored at l=650 nm (Figure S13 in the
Supporting Information) give much larger slopes (KSV) for
TTF-py (1156mꢀ1 vs. AlPor-Ph) and (454mꢀ1 vs. AlPor-Ph-
NDI) than for TTF without an appended py group (48mꢀ1
vs. AlPor-Ph). Because TTF cannot coordinate to AlPor in
the absence of the py group, the latter KSV value represents
the contribution due to collisional quenching and we con-
clude that it plays only a minor role in the quenching seen
with TTF-py. The most likely explanation for the increased
quenching with py-appended TTF is an intramolecular hole
1
transfer from AlPor* to the TTF unit. This process is exo-
thermic (DG=ꢀ0.67 eV) and has been observed in other
systems in which TTF is attached axially to a porphyrin
macrocycle.[8e,13] Insertion of an additional phenyl spacer be-
tween the TTF and py units is expected to decrease the elec-
tronic coupling and to slow down the hole transfer rate.
Consistent with this expectation, the quenching is weaker in
the titrations with TTF-Ph-py (Figures 9D and S9C in the
Supporting Information).
Similar titrations were also carried out with BN as the sol-
vent. In this case, the solvent coordinates to AlPor and
therefore when AlPor-Ph is titrated with py no shift of the
fluorescence bands or change in the intensity are observed
(Figure S10A in the Supporting Information). In contrast, ti-
tration of AlPor-Ph with TTF-py or TTF-Ph-py, does lead to
fluorescence quenching (Figures S10B and C in the Support-
ing Information) as would be expected if hole transfer from
1AlPor* to TTF occurs. The amount of quenching in the BN
titrations is low compared to those in dichloromethane, pre-
sumably because the large excess of BN competes with
TTF-py (or TTF-Ph-py) for the Al coordination sites and
the concentration of the coordination complex TTF-py!
AlPor-Ph (or TTF-Ph-py!AlPor-Ph) is lower. In contrast to
the behavior in dichloromethane, titrations of AlPor-Ph-
NDI with py, TTF-py, or TTF-Ph-py in BN (Figures S11A–
C in the Supporting Information) all show a small increase
in the fluorescence intensity. This result suggests that coordi-
nation of py probably causes a slight slowing of the rate of
electron transfer from 1AlPor* to NDI. The absence of
quenching in the titrations with TTF-py or TTF-Ph-py im-
Figure 10. Fluorescence decay profiles 1) AlPor-Ph, 2) AlPor-Ph-NDI,
3) TTF-py!AlPor-Ph, and 4) TTF-py!AlPor-Ph-NDI in dichlorome-
thane (top) and BN (bottom). The excitation wavelength is l=406 nm
and the fluorescence was monitored at l=616 nm. The instrument re-
sponse function (IRF) is also shown.
rent from Figure 10 (top) the lifetimes of the profiles in di-
chloromethane follow the order AlPor-Ph (trace 1)>AlPor-
Ph-NDI (trace 2)>TTF-py!AlPor-Ph (trace 3)>TTF-py!
AlPor-Ph-NDI (trace 4). For AlPor-Ph, the decay is mono-
exponential with a lifetime of tf =5.70 ns. The decay traces
for the dyads are bi-exponential with a fast-decaying major
component and a minor component that decays with ap-
proximately the same lifetime as in AlPor-Ph.
The fast, major components have been used to calculate
the values of the rate constant of the quenching process (kq)
and the quantum yield (fq) as given in Table 2. If charge
separation from AlPor to NDI is assumed to be responsible
for the fluorescence quenching in AlPor-Ph-NDI then kq =
kET =1/tET and the data for the dyad and reference com-
pound with coordinated py give a value of kq =kET =1.13ꢁ
109 sꢀ1 (tET =880 ps) for the charge separation between
AlPor and NDI in dichloromethane. For TTF-py!AlPor-Ph
kq =kET =4.12ꢁ109 sꢀ1, which implies that the hole transfer
to TTF occurs with a lifetime of 240 ps. Based on these two
quenching rates, a lifetime of 180 ps is predicted for the flu-
orescence decay in the triad TTF-py!AlPor-Ph-NDI, which
is very close to the observed lifetime of 174 ps. Thus, the
data suggest that for TTF-py!AlPor-Ph-NDI in dichloro-
methane the reaction scheme is as follows [Eq. (4)]:
1
plies that in BN, the hole transfer from AlPor* to TTF is
too slow to compete effectively with the electron transfer
1
from AlPor* to NDI.
To estimate the hole and the electron transfer rates be-
tween TTF, 1AlPor*, and NDI, time-resolved fluorescence
studies were carried out by using a time-correlated single-
photon-counting apparatus with excitation at l=406 nm.
Figure 10 shows a comparison of the fluorescence decay
time profiles of the triad TTF-py!AlPor-Ph-NDI and its
components monitored at l=616 nm in dichloromethane
(top) and BN (bottom). The corresponding fluorescence
lifetimes (tf) obtained by curve fitting with mono- and bi-ex-
ponential functions are summarized in Table 2. As is appa-
hn
240 ps
880 ps
1
*
Cþ
Cꢀ
Cþ
Cꢀ
AlPor
AlPor
TTF -AlPor
TTF -AlPor-NDI
ƒ!
ƒƒ!
ƒƒ!
ð4Þ
&
8
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Charge Separation in Novel Axial Donor–Porphyrin–Acceptor Triads
FULL PAPER
3
Table 2. Fluorescence lifetimes (tf) and their relative amplitudes (A), quenching rates (kq), and quantum
band of the AlPor* moiety at
yields (fq) from time-resolved fluorescence data of the dyads and triads.[a]
l=470 nm (Figure S14 in the
Supporting Information), but
the intensity of the absorption
change is a factor of two small-
er and its lifetime is shorter
(17 ms) compared to AlPor-Ph.
Based on the energy level dia-
gram, (Figure 8), exergonic
electron transfer is possible
from both the singlet and the
triplet state of AlPor. However,
[c,d]
Sample
Solvent
tf [ns][b] (A [%])
kq [ꢁ109sꢀ1
]
fq ꢁA
dichloromethane
BN
dichloromethane
BN
dichloromethane
BN
dichloromethane
BN
dichloromethane
BN
dichloromethane
BN
dichloromethane
BN
dichloromethane
BN
5.70 (100)
6.61 (100)
5.66 (100)
6.62 (100)
–
–
–
–
–
–
–
AlPor-Ph
py!AlPor-Ph
–
5.45 (21), 0.852 (79)
6.64 (22), 0.283 (78)
5.53 (24), 0.763 (78)
6.55 (28), 0.268 (76)
4.44 (8), 0.174 (92)
6.25 (28), 0.231 (72)
3.99 (8), 0.216 (92)
6.11 (20), 0.256 (80)
4.75 (18), 0.233 (82)
6.22 (74), 0.623 (26)
4.08 (34), 0.293 (57)
6.33 (63), 0.706 (37)
1.00
3.38
1.13
3.58
5.57
4.18
4.45
3.75
4.12
1.45
3.24
1.27
85ꢁ79
96ꢁ78
–
AlPor-Ph-NDI
py!AlPor-Ph-NDI
TTF-py!AlPor-Ph-NDI
TTF-Ph-py!AlPor-Ph-NDI
TTF-py!AlPor-Ph
TTF-Ph-py!AlPor-Ph
–
96ꢁ92
96ꢁ72
96ꢁ92
95ꢁ80 additional absorption bands as-
96ꢁ82
90ꢁ26
94ꢁ57
89ꢁ37
sociated with the expected for-
+
C
Cꢀ
mation of AlPor -NDI are
not clearly seen. The most
likely explanation for the ab-
sence of these bands is that
charge recombination to the
ground state is faster than the
10 ns instrument response time.
The fact that the intensity of
[a] Excitation wavelength l=406 nm, emission collected at l=616 nm. [b] Where two lifetimes are given, the
shorter lifetime is a weighted average [t=(A1t1+A2t2)/(A1+A2)] of the two shortest lifetimes obtained from a
tri-exponential fit. [c] kq =(1/t)dyad/triad (1/t)reference. fq =[(1/t)dyad/triad (1/t)reference]/(1/t)dyad/triad [d] If intramolecu-
lar electron transfer is the dominant quenching mechanism kq =kET
AHCTUNGTRENNUNG
ꢀ
A
ꢀ
N
ACHTUNGTRENNUNG
.
Addition of a phenyl group to the bridge between TTF
and AlPor causes only a slight decrease in the rate of the in-
itial charge separation and the order of the steps remains
the same (Table 2). In contrast, changing the solvent to BN
(Figure 10, bottom), leads to a decrease of the electron-
transfer lifetime to 280 ps (kq =kET =3.58ꢁ109 sꢀ1) for
AlPor-Ph-NDI (trace 2) and an increase in the hole-transfer
lifetime to 690 ps (kq =kHT =1.45ꢁ109 sꢀ1) in TTF-py!
AlPor-Ph (trace 3). Thus, the data suggest that in BN the re-
action scheme is as follows [Eq. (5)]:
the absorption change at l=470 nm is a factor of two small-
er in the dyad is also consistent with the formation of a
short-lived charge-separated state. The observed shorter life-
time of the triplet state could be the result of slow electron
transfer from 3AlPor* to NDI, in addition to the singlet elec-
tron transfer, because the energy of the charge-separated
3
state is lower than that of AlPor* (Figure 8).
In the triads TTF-py!AlPor-Ph-NDI and TTF-Ph-py!
+
C
AlPor-Ph-NDI, absorbance changes due to TTF are antici-
pated at l=450 and 600 nm (Figure S16 in the Supporting
Information). Again, however, there is no clear evidence for
the formation of the charge-separated states and only a fur-
ther decrease in the triplet-state lifetime is seen. Thus, over-
all the transient absorbance data suggest that in dichlorome-
thane or BN at room temperature the charge-recombination
hn
280 ps
690 ps
1
*
Cþ
AlPor -NDI
Cꢀ
Cþ
Cꢀ
AlPor
AlPor
TTF -AlPor-NDI
ƒ!
ƒƒ!
ƒƒ!
ð5Þ
Again, only a small increase in the hole-transfer lifetime
is observed when the additional phenyl group in added to
the bridge between TTF and the porphyrin. Although it is
possible to deduce the lifetimes and the sequence of the ini-
tial charge separation events from the fluorescence data,
they do not provide direct evidence for the charge-separated
states or any information regarding their lifetimes. Hence,
nanosecond transient absorbance measurements were also
performed.
+
+
C
Cꢀ
C
Cꢀ
lifetimes of the states, AlPor -NDI , TTF -AlPor , and
+
C
Cꢀ
TTF -AlPor-NDI are shorter than a few ns. Such behavior
is consistent with that reported recently[6d] for a similar dyad
in which NDI is axially connected to the aluminum porphyr-
in through a flexible acetate spacer. By using ultrafast spec-
troscopy, the charge-separation and charge-recombination
lifetimes of the dyad were found to be 8.1 and 120 ps, re-
spectively, in dichloromethane solution at room tempera-
ture. The fluorescence data presented above, suggest that
the charge separation is slower in AlPor-Ph-NDI, which has
a benzoate spacer, than in the acetate-linked dyad reported
in the literature.[6d]
Transient absorption studies: Nanosecond transient absorp-
tion spectra of the dyads, triads, and their reference com-
pounds in Ar-saturated dichloromethane and BN solutions
were collected in the l=400–800 nm range by using l=
532 nm laser light to excite the AlPor moiety. For AlPor-Ph
the absorbance difference spectrum exhibits the expected
characteristic absorption increase at l=470 nm due to for-
mation of the porphyrin triplet state (data not shown). From
the decay of the absorption change at l=470 nm, the life-
time of 3AlPor* was evaluated to be 32 ms. The transient
spectra of the dyad AlPor-Ph-NDI also show the absorption
Transient electron paramagnetic resonance studies: Transi-
ent electron paramagnetic resonance (TREPR) measure-
ments of AlPor-Ph-NDI, TTF-py!AlPor-Ph, and TTF-Ph-
py!AlPor-Ph under the same conditions as used for the
transient absorbance experiments (room temperature, di-
chloromethane and BN as solvents) did not show any sig-
nals. This observation is consistent with a rapid charge re-
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
9
&
ÞÞ
These are not the final page numbers!

P. K. Poddutoori, A. van der Est et al.
combination in the dyads. However, it is also possible that
rapid motion and/or spin relaxation average the spectrum to
zero in these solvents. Previous studies[20] have shown that
the charge-recombination rates in some donor–acceptor
complexes are dramatically slowed down in partially or-
dered nematic liquid crystalline solvents. It has been pro-
posed that this change in the electron-transfer rate occurs as
a result of the anisotropic dielectric relaxation and rotation-
al motion of the liquid crystalline solvent.[20d,21] The high
degree of light scattering of nematic phases precludes transi-
ent absorbance studies but such samples are well suited for
TREPR experiments. Hence, we have measured the spin-
polarized transient EPR spectra of the dyads and triads in
the liquid crystal 4-(n-pentyl)-4’-cyanobiphenyl (5CB).
width of the radical-pair triplet spectrum is j2Dj and the ob-
served width of approximately 40 G is consistent with the
distance between AlPor and NDI. Thus, together the fluo-
rescence quenching and the transient EPR spectrum suggest
that upon photoexcitation electron transfer from the excited
singlet state of AlPor to the NDI unit occurs. The broad
spectrum from the porphyrin triplet state is probably
formed through intersystem crossing (ISC) because charge
recombination to the porphyrin triplet state is energetically
unfavorable. The two possible pathways to the porphyrin
triplet state can be distinguished by the resulting polariza-
tion patterns and the observed pattern is consistent with an
ISC. However, in the liquid crystalline solvent, the z compo-
nent of the spectrum, which would have A/E polarization if
the triplet state was formed by radical-pair recombination,
is very weak because orientations with the z axis of the por-
phyrin parallel to the field have a low probability. Thus, al-
though the triplet state is most likely formed by ISC, radi-
cal-pair recombination cannot be ruled out.
Figure 11 shows the TREPR spectrum for AlPor-Ph-NDI
in the nematic phase of 5CB at room temperature taken
For the TTF-py!AlPor-Ph and TTF-Ph-py!AlPor-Ph
dyads in 5CB at RT, only the spectrum of the porphyrin
triplet is observed in TREPR measurements (data not
shown). However, it is unclear whether this means that don-
1
ation from TTF to AlPor* is slower than the intersystem
crossing, the lifetime of the charge separation is short or if
only the singlet charge-separated state is populated.
Figure 12 shows the transient EPR spectrum of TTF-Ph-
py!AlPor-Ph-NDI in 5CB at room temperature. In a wide
scan taken 300 ns after the laser flash (Figure 12a), the spec-
trum is dominated by a narrow four-line E/A/E/A pattern
with a very weak, broad E/A pattern in the wings. The
broad pattern is due to the AlPor triplet state and the
narrow pattern is assigned to the weakly coupled radial pair
+
C
Cꢀ
TTF -NDI . The narrow pattern (Figure 12b) shows two
Figure 11. Room temperature (300 K) transient EPR spectrum of AlPor-
Ph-NDI in liquid crystal 5CB. The spectrum shows the signal intensity
300 ns after the laser flash.
anti-phase doublets centered at geff =2.010 and 2.0030. Qual-
itatively, these values are consistent with the expected g
+
C
Cꢀ
values for the TTF and NDI radicals oriented in the
liquid crystal. The sign of the polarization pattern is deter-
mined by the sign of the spin–spin coupling parameter a=
Jꢀd and whether the electron transfer occurs from the sin-
glet state or the triplet state of the chromophore. At a dis-
tance of approximately 25 ꢂ between the radicals the point-
dipole approximation predicts a dipolar coupling constant of
D=ꢀ0.167 mT. Thus, the sign of the dipolar splitting, d=
(D/3)h3cos2qꢀ1i, depends on the sign of the order parame-
ter h3cos2qꢀ1i. The long axis of the complex is expected to
align along the field and thus the order parameter should be
positive and therefore d should be negative. The value of J
is not known but it is expected to be small. Hence, we
expect that a=Jꢀd will be positive. Under these conditions,
singlet electron transfer gives an E/A/E/A pattern[22] as ob-
served. Thus, the polarization pattern at early time (Fig-
300 ns after the laser flash. A broad emission/absorption (E/
A) pattern that can be attributed to the triplet state of
AlPor is observed along with a narrow E/A pattern consis-
3
+
C
Cꢀ
tent with the radical-pair triplet state (AlPor -NDI ). The
sign of the polarization, E/A or A/E, for a radical-pair trip-
let state is determined by which of the triplet sublevels is
populated and by the sign of the splitting. In a strongly cou-
pled radical pair with nearly pure singlet and triplet eigen-
states, the splitting in the triplet state is approximately
(d+q2/4J) where q is the difference of the resonance fre-
quencies of the two radicals and d and J are the dipolar and
+
C
exchange couplings between them, respectively. For AlPor
-NDI in 5CB, we expect that jdj @ jq2/4Jj. Hence, the
Cꢀ
splitting is dominated by the dipolar coupling. Singlet elec-
tron transfer followed by singlet–triplet mixing in the radical
pair selectively populates the T0 level and at a distance of
11 ꢂ, the point dipole approximation predicts a dipolar cou-
pling constant of D=ꢀ20 G. The combination of T0 popula-
tion and D<0 gives an E/A pattern as observed. The full
ure 12b, solid line), leads to the conclusion that the radical
+
C
Cꢀ
pair TTF -NDI is formed by singlet electron transfer.
In the spectrum of a pure singlet-born radical pair, all
four lines of the two anti-phase doublets are of equal inten-
sity, that is, there is no net polarization of the radicals. How-
&
10
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Charge Separation in Novel Axial Donor–Porphyrin–Acceptor Triads
FULL PAPER
coupling. Theoretical[23] and experimental studies of bi-radi-
cals[24] and metal radical systems[25] indicate that the sign of
+
C
Cꢀ
J should be negative. Therefore, AlPor -NDI is the more
likely candidate as the primary radical pair and the polariza-
tion pattern in Figure 12b (solid line) is most easily ex-
1
+
C
plained by the reaction sequence AlPor*!AlPor
+
Cꢀ
C
Cꢀ
-NDI !TTF -NDI . At later time, the spin polarization
pattern decays and inverts (Figure 12b, dashed line). Such
behavior has been observed in other donor–acceptor sys-
tems[8d,26] and is usually the result of a spin-selective decay
of the radical pair. Figure 13 shows the transients taken at
Figure 12. Room temperature (300 K) transient EPR spectra of TTF-Ph-
py!AlPor-Ph-NDI in the nematic phase of the liquid crystal 5CB.
3
a) Wide scan showing broad, very weak contribution from AlPor and in-
+
C
Cꢀ
tense narrow peaks from TTF -AlPor-NDI 300 ns after the laser flash.
+
C
Cꢀ
b) Contribution from TTF -AlPor-NDI on an expanded scale; solid
line: 300 ns after the laser flash, dashed line: 1.1 ms after the laser flash.
The arrow under the spectra indicates the field position of the transients
shown in Figure 13.
ever, in the radical-pair spectrum in Figure 12b, emissive
+
C
Cꢀ
peak of TTF and the absorptive peak of NDI have great-
er intensity than the other two lines indicating net polariza-
tion of each radical. This net polarization is expected as a
result of S–T0 mixing during the sequential electron transfer
in the triad. The sign of the net polarization is determined
by the sign of the product b1q1q2, where b1 =2J1+d1 is the
non-secular part of the spin–spin coupling in the primary
radical pair and q1 and q2 are the differences in the reso-
nance frequencies in the primary and secondary radical
pairs, respectively.[22] The observed sign of the net polariza-
tion, E/A, implies that b1q1q2 is positive. Because the fluo-
rescence data indicate that the order of the two electron
transfer steps is solvent dependent, the primary radical pair
+
+
Cꢀ
C
Cꢀ
Figure 13. Signal decay and qualitative model of charge recombination in
the triads. The transients were measured at the field position indicated
C
in 5CB could either be TTF -AlPor or AlPor -NDI .
The sign of b1 can be predicted for each of these possibilities
by the arrow in Figure 12. The circles on the energy levels of the radical
based on the signs of q1 and q2, which are known from the g
+
C
Cꢀ
pair TTF -NDI represent their relative populations.
+
C
Cꢀ
factors of the radicals. If TTF -AlPor is the primary radi-
cal pair, then q1q2 is positive and, therefore, the spin polari-
zation predicts that b1 is positive. On the other hand, if
the field position indicated by the arrow in Figure 12 along
with a qualitative model that explains the signal decay. The
+
C
Cꢀ
AlPor -NDI is the primary radical pair, then q1q2 is nega-
tive and the sign of the polarization implies that b1 is nega-
tive. In both of the possible primary radical pairs, the ex-
change coupling J is expected to dominate the spin–spin
diagram on the right of Figure 13 shows the spin states Y1–
+
C
Cꢀ
Y4 of the weakly coupled radical pair TTF -NDI . The cir-
cles on the energy levels represent the population of the
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
11
&
ÞÞ
These are not the final page numbers!

P. K. Poddutoori, A. van der Est et al.
states. Initially, singlet electron transfer preferentially popu-
lates states Y2 and Y3, which have S and T0 character. Trip-
let electron transfer, spin relaxation, and/or microwave-in-
duced transitions also place a small amount of population in
states Y1 and Y4, which are the pure triplet states T+ and
Tꢀ. As time progresses, states Y2 and Y3 are depopulated
because the singlet recombination rate kS is much higher
than the triplet recombination rate kT. Thus, the signal in-
verts. Finally, the inverted signal decays due to a combina-
tion of spin relaxation and recombination. Kinetic fits of the
transients (Figure 13, left) give lifetimes of 350 ns for the
singlet recombination and 8.3 ms for the decay of the late
signal. Similar results were also found for the shorter triad
TTF-py!AlPor-Ph-NDI (see Figure S17 in the Supporting
Information). The singlet lifetime follows the expected
trend and decreases from 350 to 200 ns when the phenyl
spacer between the TTF and AlPor units is removed. The
decay of the late signal on the other hand increases from 8.3
to 12 ms probably as a result of differences in the T1 relaxa-
tion rates in the two complexes.
dichloromethane (Figure 10, top) they indicate that the hole
transfer to TTF is faster. Thus, there appears to be a subtle
balance of factors that changes the relative order of the
rates in different solvents. One possible reason for this can
be seen in the results of the DFT calculations. The orbitals
presented in Figure 6b show that when py-appended TTF
coordinates, the LUMO of AlPor, which is a p* orbital, is
delocalized onto the axially coordinated py bridge and away
from the covalently bound phenyl group on the opposite
face of the porphyrin. In the triads, the effect of this delocal-
ization should be to weaken the coupling between the
LUMO of AlPor and the LUMO of NDI and to stabilize
+
C
Cꢀ
the TTF -AlPor state. It is likely that these effects com-
pensate for the difference in driving force between the two
reactions and leave them with similar rates. Because both
the electronic coupling and the activation energy depend on
the nature of the solvent it is not surprising that the relative
order of the rates should change in different solvents. The
reorganization energy for charge separation/recombination
between TTF and NDI has been estimated to be between
0.8 and 1.0 eV in BN.[28] This range of values gives activation
Electronic coupling to axially bound ligands: For dichloro-
methane and BN, the electron transfer in the dyads and
triads can be discussed in terms of the classical Marcus
equation, which assumes non-adiabatic electron transfer
(ET). The longitudinal dielectric relaxation times of the sol-
vent (0.33 and 6.9 ps, for dichloromethane and BN, respec-
tively)[27] are much shorter than the electron-transfer life-
times estimated from the fluorescence quenching and hence,
the non-adiabatic assumption should be valid in these sol-
vents. Previous studies have shown that the reorganization
energy for electron transfer from ZnPor* to NDI is l=
(1.4ꢃ0.3) eV[10f] in polar solvents. In AlPor-Ph-NDI, the
free energy differences for charge separation and recombi-
nation are found to be ꢀ0.86 and ꢀ1.28 eV, respectively in
dichloromethane. By using these values, the classical Marcus
equation gives activation energies of 52 meV for the forward
reaction but only 2.5 meV for the back reaction. Consistent
with these estimates, competition between intersystem cross-
ing and charge separation occurs during the forward reac-
tion and long-lived charge separation is not observed for
AlPor-Ph-NDI in dichloromethane. For the hole transfer
from 1AlPor* to TTF, the reorganization energy is not
known but it is expected to be similar to or slightly larger
energies in the vicinity of 100 meV for recombination of the
+
C
Cꢀ
final charge-separated state TTF -AlPor-NDI
in the
triads. Because there is little or no spatial overlap between
the HOMO and LUMO in the triads (Figure 6c) the elec-
+
C
Cꢀ
tronic coupling between TTF and NDI should be small
and long-lived charge separation is expected. However, in
+
C
Cꢀ
dichloromethane and BN, TTF -AlPor-NDI is probably
not observed because the back reactions of the initial
+
+
C
Cꢀ
C
Cꢀ
charge-separated states AlPor -NDI and TTF -AlPor
lie near the top of the Marcus curve and recombination out
competes formation of the secondary radical pair.
Achieving long-lived charge separation: In the nematic
phase of 5CB the solvent dynamics change dramatically
compared to dichloromethane and BN and this clearly has a
pronounced effect on the lifetime of the charge separation,
as has been observed in a number of other systems.[20] In
particular, the higher viscosity and ordering of the nematic
phase result in much slower and highly anisotropic dielectric
relaxation compared to dichloromethane and BN. In liquid
crystals, which are structurally similar to 5CB, the compo-
nent of the molecular dipole along the director is known to
relax with a lifetime of approximately 100 ns, whereas the
relaxation lifetime of the perpendicular component is about
1 ns.[29] For such slowly relaxing solvents, the pre-exponential
term in the Arrhenius ET rate equation is predicted to be
proportional to the solvent relaxation rate[30] because the
slow reorganization of the solvent dipoles impedes the for-
mation of the transition state from which the electron trans-
fer occurs. Thus, slower ET is expected in 5CB because of
this effect. However, recent studies[20d,21] point to a more
complex situation in liquid crystalline solvents. An impor-
tant feature of such solvents is the correlation of the motion
of the solvent molecules over relatively long distances. It is
these correlated motion that leads to the partial ordering of
the solvent and restrict the rotational motion of solutes.[31]
1
than that of the electron transfer from AlPor* to NDI. The
driving force for the hole transfer (ꢀ0.64 eV) is therefore
considerably smaller than the reorganization energy, which
places the reaction in the normal region of the Marcus
+
C
Cꢀ
curve. The driving force for recombination of TTF -AlPor
is close to ꢀl and again long-lived charge separation is not
expected.
In the case of the triads TTF-py!AlPor-Ph-NDI and
TTF-Ph-py!AlPor-Ph-NDI, competition between hole
transfer to TTF and electron transfer to NDI exists. Based
on the activation energies, electron transfer to NDI is ex-
pected to be the faster process. The fluorescence data in BN
(Figure 10, bottom) agree with this expectation, however, in
&
12
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Charge Separation in Novel Axial Donor–Porphyrin–Acceptor Triads
FULL PAPER
Moore, Chem. Soc. Rev. 2009, 38, 25–35; i) K. Kalyanasundaram,
M. Graetzel, Curr. Opin. Biotechnol. 2010, 21, 298–310; j) G. F.
Moore, G. W. Brudvig, Annu. Rev. Condens. Matter Phys. 2011, 2,
303–327; k) D. G. Nocera, Acc. Chem. Res. 2012, 45, 767–776; l) A.
Magnuson, M. Anderlund, O. Johansson, P. Lindblad, R. Lomoth, T.
Polivka, S. Ott, K. Stensjo, S. Styring, V. Sundstrom, L. Hammar-
strom, Acc. Chem. Res. 2009, 42, 1899–1909; m) M. R. Wasielewski,
Acc. Chem. Res. 2009, 42, 1910–1921.
This restriction of the rotational motion leads to anisotropy
in the solvation energy, which in turn affects the ET rates.
Further, if the reorientation of the solvent occurs on a time-
scale similar to that of the electron transfer, the contribution
of the solvation energy to the activation barrier for electron
transfer is predicted to be much lower than its thermody-
namic value.[21] For the dyads and triads studied here, the
relative importance of these solvent effects is not known but
together they increase the lifetime of the charge-separated
states by at least an order of magnitude in 5CB compared to
the isotropic solvents dichloromethane and BN. Further ex-
tension of the lifetime of the charge-separated states is ach-
ieved by exploiting the spin selectivity of the charge recom-
bination in a magnetic field.
[3] a) F. DꢃSouza, S. Gadde, D. M. S. Islam, C. A. Wijesinghe, A. L.
Schumacher, M. E. Zandler, Y. Araki, O. Ito, J. Phys. Chem. A 2007,
111, 8552–8560; b) F. DꢃSouza, S. Gadde, M. E. Zandler, K. Arkady,
M. E. El-Khouly, M. Fujitsuka, O. Ito, J. Phys. Chem. A 2002, 106,
12393–12404; c) F. DꢃSouza, O. Ito, Chem. Commun. 2009, 4913–
4928; d) F. DꢃSouza, C. A. Wijesinghe, M. E. El-Khouly, J. Hudson,
M. Niemi, H. Lemmetyinen, N. V. Tkachenko, M. E. Zandler, S. Fu-
kuzumi, Phys. Chem. Chem. Phys. 2011, 13, 18168–18178; e) S. L.
Gould, G. Kodis, R. E. Palacios, L. de La Garza, A. Brune, D. Gust,
T. A. Moore, A. L. Moore, J. Phys. Chem. B 2004, 108, 10566–
10580; f) D. M. Guldi, Chem. Soc. Rev. 2002, 31, 22–36; g) H. Ima-
hori, D. M. Guldi, K. Tamaki, Y. Yoshida, C. P. Luo, Y. Sakata, S.
Fukuzumi, J. Am. Chem. Soc. 2001, 123, 6617–6628; h) H. Imahori,
Y. Mori, Y. Matano, J. Photochem. Photobiol. C 2003, 4, 51–83;
i) P. V. Kamat, J. Phys. Chem. C 2007, 111, 2834–2860; j) P. A. Lid-
dell, G. Kodis, J. Andreasson, L. de La Garza, S. Bandyopadhyay,
R. H. Mitchell, T. A. Moore, A. L. Moore, D. Gust, J. Am. Chem.
Soc. 2004, 126, 4803–4811; k) J. Y. Liu, M. E. El-Khouly, S. Fukuzu-
mi, D. K. P. Ng, Chem. Eur. J. 2011, 17, 1605–1613; l) C. Luo, D. M.
Guldi, H. Imahori, K. Tamaki, K. Sakata, J. Am. Chem. Soc. 2000,
122, 6535–6551; m) M. D. Ward, Chem. Soc. Rev. 1997, 26, 365–375.
[4] K. M. S. a. R. G. K. M. Kadish in Handbook of Porphyrin Science,
World Scientific, Singapore, 2010.
Conclusion
The results presented above show that self-assembled triads
capable of step-wise, light-induced charge separation can be
constructed by coordinating py-appended TTF derivatives
with the AlPor-Ph-NDI dyad. It is well known that hydrogen
bonds mediate electron transfer better than analogous cova-
lent bonds.[32] In AlPor-based triads, a similar comparison
can be made between axial coordination bonds versus axial
covalent bonds. The observation that donation by TTF and
electron transfer to NDI have similar rates suggests that the
two types of bonding have similar efficiencies in mediating
electron transfer. However, a better comparison could be
made if an analogous system was constructed with the ac-
ceptor bound to AlPor through the py linker. The DFT cal-
culations also suggest that more efficient electron transfer
might be obtained with this arrangement because the
LUMO of the porphyrin is delocalized onto the py bridge.
Currently, we are exploring these ideas and are preparing
similar triads with different arrangements and structures for
the donors and acceptors.
[5] a) D. Gust, T. A. Moore, Top. Curr. Chem. 1991, 159, 103–151; b) D.
Gust, T. A. Moore, A. L. Moore, Acc. Chem. Res. 1993, 26, 198–
205; c) M. R. Wasielewski, Chem. Rev. 1992, 92, 435–461.
[6] a) T. Aida, S. Inoue, Acc. Chem. Res. 1996, 29, 39–48; b) G. J. E.
Davidson, L. H. Tong, P. R. Raithby, J. K. M. Sanders, Chem.
Commun. 2006, 3087–3089; c) Y. Hirai, T. Aida, S. Inoue, J. Am.
Chem. Soc. 1989, 111, 3062–3063; d) E. Iengo, G. D. Pantos, J. K. M.
Sanders, M. Orlandi, C. Chiorboli, S. Fracasso, F. Scandola, Chem.
Sci. 2011, 2, 676–685; e) G. A. Metselaar, J. K. M. Sanders, J. de
Mendoza, Dalton Trans. 2008, 588–590; f) P. K. Poddutoori, P. Pod-
dutoori, B. G. Maiya, T. K. Prasad, Y. E. Kandrashkin, S. Vasil’ev, D.
Bruce, A. van der Est, Inorg. Chem. 2008, 47, 7512–7522; g) P. K.
Poddutoori, A. S. D. Sandanayaka, T. Hasobe, O. Ito, A. van der Est,
J. Phys. Chem. B 2010, 114, 14348–14357; h) P. K. Poddutoori,
A. S. D. Sandanayaka, N. Zarrabi, T. Hasobe, O. Ito, A. van der Est,
J. Phys. Chem. A 2011, 115, 709–717.
[7] P. P. Kumar, B. G. Maiya, New J. Chem. 2003, 27, 619–625.
[8] a) N. Armaroli, G. Marconi, L. Echegoyen, J. P. Bourgeois, F. Die-
derich, Chem. Eur. J. 2000, 6, 1629–1645; b) M. A. Fazio, A. Duran-
din, N. V. Tkachenko, M. Niemi, H. Lemmetyinen, D. I. Schuster,
Chem. Eur. J. 2009, 15, 7698–7705; c) S. Fukuzumi, Y. Kashiwagi, J.
Porphyrins Phthalocyanines 2007, 11, 368–374; d) A. G. Moiseev,
P. K. Poddutoori, A. van der Est, Appl. Magn. Reson. 2012, 42, 41–
55; e) P. K. Poddutoori, A. Dion, S. J. Yang, M. Pilkington, J. D.
Wallis, A. van der Est, J. Porphyrins Phthalocyanines 2010, 14, 178–
187; f) D. I. Schuster, P. Cheng, P. D. Jarowski, D. M. Guldi, C. P.
Luo, L. Echegoyen, S. Pyo, A. R. Holzwarth, S. E. Braslavsky, R. M.
Williams, G. Klihm, J. Am. Chem. Soc. 2004, 126, 7257–7270.
[9] H. Imahori, K. Tamaki, D. M. Guldi, C. P. Luo, M. Fujitsuka, O. Ito,
Y. Sakata, S. Fukuzumi, J. Am. Chem. Soc. 2001, 123, 2607–2617.
[10] a) S. J. Langford, M. J. Latter, C. P. Woodward, Photochem. Photo-
biol. 2006, 82, 1530–1540; b) N. Banerji, S. V. Bhosale, I. Petkova,
S. J. Langford, E. Vauthey, Phys. Chem. Chem. Phys. 2011, 13, 1019–
1029; c) Y. Mori, Y. Sakaguchi, H. Hayashi, J. Phys. Chem. A 2002,
106, 4453–4467; d) S. Wallin, C. Monnereau, E. Blart, J. R. Gankou,
F. Odobel, L. Hammarstrom, J. Phys. Chem. A 2010, 114, 1709–
1721; e) L. Flamigni, M. R. Johnston, L. Giribabu, Chem. Eur. J.
2002, 8, 3938–3947; f) H. Imahori, H. Yamada, D. M. Guldi, Y.
Acknowledgements
We thank Prof. Gonzalo Cosa, Department of Chemistry, McGill Univer-
sity for access to the nanosecond laser flash photolysis setup and Prof.
Doug Bruce at Brock University for use of the time-resolved fluores-
cence instrument. This work was supported by Discovery Grants from
Natural Sciences and Engineering Research Council Canada to A.v.d.E.
and D.B.
[1] R. E. Blankenship, Molecular Mechanisms of Photosynthesis, Black-
well Science, Oxford, 2002.
[2] a) B. Albinsson, J. Martensson, J. Photochem. Photobiol. C 2008, 9,
138–155; b) J. H. Alstrum-Acevedo, M. K. Brennaman, T. J. Meyer,
Inorg. Chem. 2005, 44, 6802–6827; c) J. Barber, Chem. Soc. Rev.
2009, 38, 185–196; d) S. V. Bhosale, C. H. Jani, S. J. Langford, Chem.
Soc. Rev. 2008, 37, 331–342; e) M. E. El-Khouly, O. Ito, P. M. Smith,
F. D’Souza, J. Photochem. Photobiol. C 2004, 5, 79–104; f) S. Fuku-
zumi, Eur. J. Inorg. Chem. 2008, 1351–1362; g) D. Gust, T. A.
Moore, A. L. Moore, Acc. Chem. Res. 2001, 34, 40–48; h) M. Ham-
bourger, G. F. Moore, D. M. Kramer, D. Gust, A. L. Moore, T. A.
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
13
&
ÞÞ
These are not the final page numbers!

P. K. Poddutoori, A. van der Est et al.
Endo, A. Shimomura, S. Kundu, K. Yamada, T. Okada, Y. Sakata, S.
Fukuzumi, Angew. Chem. 2002, 114, 2450–2453; Angew. Chem. Int.
Ed. 2002, 41, 2344–2347.
Gordon, E. S. Replogle, J. A. Pople, Gaussian, Inc., Pittsburgh, PA,
2002.
[18] a) F. D’Souza, G. R. Deviprasad, M. E. Zandler, M. E. El-Khouly,
M. Fujitsuka, O. Ito, J. Phys. Chem. A 2003, 107, 4801–4807; b) F.
D’Souza, P. M. Smith, S. Gadde, A. L. McCarty, M. J. Kullman,
M. E. Zandler, M. Itou, Y. Araki, O. Ito, J. Phys. Chem. B 2004, 108,
11333–11343.
[11] J. L. Segura, N. Martin, Angew. Chem. 2001, 113, 1416–1455;
Angew. Chem. Int. Ed. 2001, 40, 1372–1409.
[12] a) G. Kodis, P. A. Liddell, L. de La Garza, A. L. Moore, T. A.
Moore, D. Gust, J. Mater. Chem. 2002, 12, 2100–2108; b) P. A. Lid-
dell, G. Kodis, L. de La Garza, J. L. Bahr, A. L. Moore, T. A. Moore,
D. Gust, Helv. Chim. Acta 2001, 84, 2765–2783; c) J. Becher, T. Bri-
mert, J. O. Jeppesen, J. Z. Pedersen, R. Zubarev, T. Bjornholm, N.
Reitzel, T. R. Jensen, K. Kjaer, E. Levillain, Angew. Chem. 2001,
113, 2565–2568; Angew. Chem. Int. Ed. 2001, 40, 2497–2500;
d) H. C. Li, J. O. Jeppesen, E. Levillain, J. Becher, Chem. Commun.
2003, 846–847; e) K. A. Nielsen, E. Levillain, V. M. Lynch, J. L.
Sessler, J. O. Jeppesen, Chem. Eur. J. 2009, 15, 506–516; f) Y. B. Liu,
C. Y. Wang, M. J. Li, S. F. Lv, G. Q. Lai, Y. J. Shen, J. Porphyrins
Phthalocyanines 2007, 11, 729–735; g) M. Di Valentin, A. Bisol, G.
Agostini, P. A. Liddell, G. Kodis, A. L. Moore, T. A. Moore, D.
Gust, D. Carbonera, J. Phys. Chem. B 2005, 109, 14401–14409; h) S.
Sadaike, K. Takimiya, Y. Aso, T. Otsubo, Tetrahedron Lett. 2003, 44,
161–165; i) S. Saha, L. E. Johansson, A. H. Flood, H. R. Tseng, J. I.
Zink, J. F. Stoddart, Small 2004, 1, 87–90.
[19] a) D. Rehm, A. Weller, Ber. Bunsen Gesell. 1969, 73, 834–839; b) D.
Rehm, A. Weller, Isr. J. Chem. 1970, 10, 259–271.
[20] a) K. Hasharoni, H. Levanon, J. Phys. Chem. 1995, 99, 4875–4878;
b) L. E. Sinks, M. R. Wasielewski, J. Phys. Chem. A 2003, 107, 611–
620; c) G. P. Wiederrecht, W. A. Svec, M. R. Wasielewski, J. Phys.
Chem. B 1999, 103, 1386–1389; d) L. Sinks, M. J. Fuller, W. H. Liu,
M. J. Ahrens, M. R. Wasielewski, Chem. Phys. 2005, 319, 226–234;
e) M. Di Valentin, A. Bisol, G. Agostini, M. Fuhs, P. A. Liddell,
A. L. Moore, T. A. Moore, D. Gust, D. Carbonera, J. Am. Chem.
Soc. 2004, 126, 17074–17086.
[21] V. Kapko, D. V. Matyushov, J. Phys. Chem. B 2006, 110, 13184–
13194.
[22] Y. E. Kandrashkin, K. M. Salikhov, A. van der Est, D. Stehlik, Appl.
Magn. Reson. 1998, 15, 417–447.
[23] A. H. Goldberg, D. A. Dougherty, J. Am. Chem. Soc. 1983, 105,
284–290.
[24] M. D. E. Forbes, G. L. Closs, P. Calle, P. Gautam, J. Phys. Chem.
1993, 97, 3384–3389.
[25] P. K. Poddutoori, M. Pilkington, A. Alberola, V. Polo, J. E. Warren,
A. van der Est, Inorg. Chem. 2010, 49, 3516–3524.
[26] M. Jakob, A. Berg, E. Stavitski, E. T. Chernick, E. A. Weiss, M. R.
Wasielewski, H. Levanon, Chem. Phys. 2006, 324, 63–71.
[27] G. Grampp, S. Landgraf, K. Rasmussen, J. Chem. Soc. Perkin Trans.
2 1999, 1897–1899.
[13] X. W. Xiao, W. Xu, D. Q. Zhang, H. Xu, H. Y. Lu, D. B. Zhu, J.
Mater. Chem. 2005, 15, 2557–2561.
[14] a) Y. L. Wang, D. L. Frattarelli, A. Facchetti, E. Cariati, E. Tordin,
R. Ugo, C. Zuccaccia, A. Macchioni, S. L. Wegener, C. L. Stern,
M. A. Ratner, T. J. Marks, J. Phys. Chem. C 2008, 112, 8005–8015;
b) S. Bouguessa, A. K. Gouasmia, S. Golhen, L. Ouahab, J. M.
Fabre, Tetrahedron Lett. 2003, 44, 9275–9278.
[15] S. V. Zaitseva, S. A. Zdanovich, T. A. Ageeva, O. A. Golubchikov,
Russ. J. Gen. Chem. 2001, 71, 132–136.
[16] H. A. Benesi, J. H. Hildebrand, J. Am. Chem. Soc. 1949, 71, 2703–
2707.
[28] X. F. Guo, Z. H. Gan, H. X. Luo, Y. Araki, D. Q. Zhang, D. B. Zhu,
O. Ito, J. Phys. Chem. A 2003, 107, 9747–9753.
[17] Gaussian98 (RevisionA.11.3), M. J. Frisch, G. W. Trucks, H. B.
Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakr-
zewski, J. A. Montgomery, R. E. Stratmann, J. C. Burant, S. Dap-
prich, J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O.
Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C.
Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y.
Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K. Ra-
ghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, A. G. Baboul,
B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.
Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y.
Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P. M. W. Gill,
B. G. Johnson, W. Chen, M. W. Wong, J. L. Andres, M. Head-
[29] a) M. Davies, R. Moutran, A. H. Price, M. S. Beevers, G. Williams, J.
Chem. Soc. Faraday Trans. 2 1976, 72, 1456–1458; b) S. Urban, B.
Gestblom, W. Kuczynski, S. Pawlus, A. Wurflinger, Phys. Chem.
Chem. Phys. 2003, 5, 924–928.
[30] M. J. Weaver, G. E. Mcmanis, Acc. Chem. Res. 1990, 23, 294–300.
[31] a) A. J. van der Est, P. B. Barker, E. E. Burnell, C. A. Delange, J. G.
Snijders, Mol. Phys. 1985, 56, 161–169; b) A. J. van der Est, M. Y.
Kok, E. E. Burnell, Mol. Phys. 1987, 60, 397–413.
[32] P. J. F. De Rege, S. A. Williams, M. J. Therien, Science 1995, 269,
1409–1413.
Received: August 22, 2012
Published online: && &&, 0000
&
14
&
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Charge Separation in Novel Axial Donor–Porphyrin–Acceptor Triads
FULL PAPER
Long live the charge separation: New
biomimetic donor–acceptor triads have
been constructed by taking advantage
of the ability of aluminum in AlIII por-
phyrin to form axial covalent and coor-
dination bonds with carboxylic acids
and Lewis bases, respectively (see
figure; NDI=naphthalenediimide,
TTF=tetrathiafulvalene). Upon light
excitation the triads undergo sequen-
tial electron transfer and by using a
liquid crystalline solvent and high mag-
netic field, the lifetime of the charge
separation can be extended into the
microsecond range.
Artificial Photosynthesis
P. K. Poddutoori,* N. Zarrabi,
A. G. Moiseev, R. Gumbau-Brisa,
S. Vassiliev, A. van der Est* . . . &&& — &&&
Long-Lived Charge Separation in
Novel Axial Donor–Porphyrin–
Acceptor Triads Based on Tetrathiaful-
valene, AluminumACTHNUTRGENUGN(III) Porphyrin and
Naphthalenediimide
Chem. Eur. J. 2013, 00, 0 – 0
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
15
&
ÞÞ
These are not the final page numbers!