Solvent polarity effect on intramolecular electron transfer in a corrole-naphthalene bisimide dyad

Article abstract of DOI:10.1039/b916525h

A dyad (C3-NI) based on corrole and naphthalene bisimide has been synthesized and its photoreactivity compared to that of the model component corrole (C3) and naphthalene bisimide (NI) in solvents of different polarity: toluene (TL) and dichloromethane (DCM). The major emitting species in NI solutions, in TL, is identified as a dimeric species (λ = 470 nm, τ = 2.3 ns) but traces of monomer can also be detected (λ ca. 390 nm, τ = 40 ps). In DCM the major emitting component is the monomer (λ = 383 nm, τ = 20 ps) but traces of different aggregates (λ = 540 and 570 nm, τ = 4.5 and 11 ns) are present. C3 has a fluorescence nearly unaffected by solvent polarity, with a maximum around 655 and a lifetime of 3.5 or 3.8 ns in DCM and TL, respectively. The dyad C3-NI does not appear to be affected by aggregation problems in any of the solvents. Excitation of the imide component in C3-NI (C3-¹NI) results in an energy transfer to corrole ( ¹C3-NI) with rate k = 2.0 × 10¹¹ s^-1 in both solvents. The latter state reacts further via a LUMO-LUMO electron transfer to the naphthalene bisimide yielding the charge separated state C3 ⁺-NI^- (k = 1.8 × 10⁹ s^-1 in TL and k = 3.7 × 10⁹ s^-1 in DCM). The same type of reactivity is displayed by direct excitation of the corrole moiety in the dyad to ¹C3-NI. C3⁺-NI^- decays with a rate comparable to that of its formation in DCM (k = 4.0 × 10⁹ s^-1in DCM), precluding its accumulation, whereas it decays with a slower rate in TL (k = 7.1 × 10⁸ s^-1). The charge separated state recombines to the ground singlet state; recombination to the triplet state of corrole (excited state at the lowest energy) is in fact excluded on the basis of the experimentally determined triplet yields. The failure of the commonly used methods in the calculation of CS energy levels in apolar solvents is confirmed. the Owner Societies 2010.

Full text of DOI:10.1039/b916525h

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/pccp | Physical Chemistry Chemical Physics
Solvent polarity effect on intramolecular electron transfer
in a corrole–naphthalene bisimide dyad
Lucia Flamigni,*a Dagmara Wyrostek,b Roman Voloshchukb and
Daniel T. Gryko*b
Received 11th August 2009, Accepted 8th October 2009
First published as an Advance Article on the web 12th November 2009
DOI: 10.1039/b916525h
A dyad (C3–NI) based on corrole and naphthalene bisimide has been synthesized and its
photoreactivity compared to that of the model component corrole (C3) and naphthalene bisimide
(NI) in solvents of different polarity: toluene (TL) and dichloromethane (DCM). The major
emitting species in NI solutions, in TL, is identified as a dimeric species (l = 470 nm, t = 2.3 ns)
but traces of monomer can also be detected (l ca. 390 nm, t = 40 ps). In DCM the major
emitting component is the monomer (l = 383 nm, t = 20 ps) but traces of different aggregates
(l = 540 and 570 nm, t = 4.5 and 11 ns) are present. C3 has a fluorescence nearly unaffected by
solvent polarity, with a maximum around 655 and a lifetime of 3.5 or 3.8 ns in DCM and TL,
respectively. The dyad C3–NI does not appear to be affected by aggregation problems in any of
the solvents. Excitation of the imide component in C3–NI (C3–1NI) results in an energy transfer
to corrole (1C3–NI) with rate k = 2.0 ꢀ 1011 sꢁ1 in both solvents. The latter state reacts further
via a LUMO–LUMO electron transfer to the naphthalene bisimide yielding the charge separated
state C3+–NIꢁ (k = 1.8 ꢀ 109 sꢁ1 in TL and k = 3.7 ꢀ 109 sꢁ1 in DCM). The same type of
1
reactivity is displayed by direct excitation of the corrole moiety in the dyad to C3–NI. C3+–NIꢁ
decays with a rate comparable to that of its formation in DCM (k = 4.0 ꢀ 109 sꢁ1in DCM),
precluding its accumulation, whereas it decays with a slower rate in TL (k = 7.1 ꢀ 108 sꢁ1).
The charge separated state recombines to the ground singlet state; recombination to the triplet
state of corrole (excited state at the lowest energy) is in fact excluded on the basis of the
experimentally determined triplet yields. The failure of the commonly used methods in the
calculation of CS energy levels in apolar solvents is confirmed.
examination of new chromophores from the tetrapyrrole
Introduction
family, namely free-base corroles, which are one bond short
analogues of free-base porphyrins.4 Free-base corroles are
remarkably stable when substituted with electron withdrawing
groups at the meso positions. They have an intense absorption
extending up to 670 nm, a relatively strong fluorescence
peaking at 655 nm and, even in presence of strongly electron
withdrawing substituents, due to the high electron density of
the core, corroles oxidize at relatively low energy (ca. 0.85 V
vs. SCE). These properties make them valuable alternatives to
porphyrins as photosensitizers and electron donors in arrays
for charge separation. Almost all possible combinations of
porphyrins with electron or energy donor/acceptors have been
synthesized and probed, and a large variety of different motifs
connecting either covalently or by weak forces porphyrins to
photo- or electroactive partners has been explored,5 whereas
for corroles the study is at the beginning. So far corroles have
been connected via a limited number of spacers to partners as
porphyrins,6 aromatic imides7,8 aromatic amines9 and C60,10
and have proven to display good performances in terms of
charge separation yields and lifetimes. The existing literature
has just been reviewed and has shown that these arrays are as
promising as those of more popular porphyrins and deserve an
extended examination.11 In this respect we report on the
synthesis and photoinduced properties of a new dyad comprising
The interest in molecular assemblies able to collect and
convert light energy is increasingly growing in view of the
exploitation of solar energy. These arrays are composed of
molecules able to absorb light and convert the energy stored in
the excited state(s) into chemical potential via one or more
electron transfer steps. The chemical potential is in the form of
a charge separated state which is the reagent able to inject
electrons, starting an electron flow,1 or able to yield energy
demanding redox reactions, producing valuable products.2
There is demand for charge separated states which can store
a high energy and live long enough in order to be utilized also
by relatively endoergonic and slow reactions.
One of our groups has long been interested in the design and
study of structures for charge separation and has examined a
wide variety of covalently connected or self-assembled systems
with components of various nature.3 In order to increase the
variety of these systems we have recently undertaken the
a Istituto per la Sintesi Organica e Fotoreattivita’ (ISOF), CNR,
Via P. Gobetti 101, 40129 Bologna, Italy.
E-mail: flamigni@isof.cnr.it
b Institute of Organic Chemistry of the Polish Academy of Sciences,
Kasprzaka 44/52, 01-224 Warsaw, Poland.
E-mail: daniel@icho.edu.pl
ꢂc
This journal is the Owner Societies 2010
474 | Phys. Chem. Chem. Phys., 2010, 12, 474–483

View Article Online
corrole as electron donor and naphthalene bisimide as electron
acceptor (C3–NI) and explore the effect of solvent polarity on
photoinduced charge separation. Naphthalene bisimide has
frequently been used as electron acceptor in self assembled or
covalently linked arrays able to produce charge separation.12,13
NI can be reduced relatively easily (ca. ꢁ0.5 vs. SCE) and the
reduced radical anion has intense absorption signatures at 475
and 610 nm which allow a clear identification of the product
and of its time evolution by optical spectroscopic methods.12,13
Results and discussion
Synthesis
According to our previous strategy,8 in order to synthesize the
required C3–NI we decided to perform mix-condensation of
naphthalene bisanhydride (1) and two amines (Scheme 1).
Amine 2 bearing a formyl group protected as acetal was
chosen as one component and the second was decylamine as
solubilizing group. The condensation was performed according
to the general procedure in DMF and afforded three bisimides,
according to expectations. We found however that chromato-
graphy could not separate pure monoacetal from bisamide NI
and diacetal. Therefore the crude mixture was subjected to
1
TFA which resulted in cleavage of acetal as confirmed by H
NMR. Chromatographic separation afforded both 3 and NI in
yields of 20 and 25%, respectively. Finally the reaction of
aldehyde 3 with dipyrromethane 4 under typical conditions
resulted in the formation of corrole dyad C3–NI with 16%
yield. (Scheme 1).
Steady state spectroscopy
An investigation on dyad C3–NI was carried out and the
properties compared to those of naphthalene bisimide NI and
corrole C3 representing the component models of the array
(Fig. 1). The systems were examined in two different solvents,
apolar toluene (TL) and moderately polar dichloromethane
(DCM), in view of the effects that solvent polarity has on
photophysical properties and photoinduced processes. The
absorption spectrum of the components and of the dyad in
the two solvents is reported in Fig. 2. The absorption spectrum
of NI in the two solvents is different, broader and with a lower
molar absorption coefficient in TL, with bands at 379, 359
Scheme 1
This allows an approach based on a localized description of
the individual subunit. From inspection of the spectra of Fig. 2
one can see that selective excitation of the corrole component
in C3–NI is possible at wavelength 4400 nm, whereas
excitation of the bisimide unit is possible in a ca. 50% ratio
in the range 350–360 nm.
and 341 nm (e ca. 16 000, 15 000 and 10 000 Mꢁ1 cmꢁ1
respectively) whereas in DCM the bands peak at 378, 357
and 338 nm (e ca. 25 000, 20 000, 12 000 Mꢁ1 cmꢁ1
,
,
respectively). The behavior in TL is probably ascribable to
formation of aggregates (see the discussion below). C3 has the
well known spectrum presenting a broad Soret band around
420 nm and Q bands around 500–670 nm. The spectra,
discussed previously, can be interpreted in the frame of the
‘‘four-orbital model’’.4,14 In the dyad, the bands of the NI
component are detectable on the high energy side of the Soret
band of corrole, and appear at 342 (shoulder), 360 and 380
(shoulder) nm in TL and at 340 (shoulder), 358 and 379 nm in
DCM. In both solvents the spectrum of the dyad C3–NI is a
simple superposition of the spectra of the two components, as
already observed for similar corrole dyads,4,6–10 indicating
that the units are decoupled from an electronic viewpoint.
The luminescence spectra of NI in the two solvents are quite
different, both for the spectral distribution, see Fig. 3, and for
the quantum yield: ffl = 0.038 in TL and ffl = 0.0014 in
DCM (Table 1). In the more polar DCM the emission displays
a good mirror image, with the absorption showing maxima at
383, 404 and 429 nm ascribed to fluorescence from the singlet
excited state of the monomer emission. This is accompanied
by a much weaker, broad luminescence with a maximum at
540 nm and a shoulder at 570 nm. In TL no monomer
fluorescence can be detected, whereas a broad band with a
maximum at 470 nm appears. The shape and location of the
band is typical of excimer or excited dimer luminescence and is
ꢂc
This journal is the Owner Societies 2010
Phys. Chem. Chem. Phys., 2010, 12, 474–483 | 475

View Article Online
Fig. 1 Dyad and models examined.
Fig. 3 Normalized luminescence spectrum of NI in TL and DCM,
excitation at 358 nm.
Table 1 Fluorescence data of dyads C3–NI and models in air
equilibrated apolar toluene (TL) and moderately polar dichloro-
methane (DCM) at 295 K
a
Ffl
State
lmax/nm
Fb
t/nsc
TL
1NI
NI
—
—
—
0.14
0.040
2.3
3.8
0.005
0.480
1NI2
470
656
—
0.038
0.14
0.0002
0.0242
d
C3
C3–NI
1C3
C3-1NI
1C3-NI
DCM
1NI
656
0.0145
Fig. 2 Absorption spectra of NI–C3 (gray), NI (dash) and C3 (solid)
in TL (upper panel) and DCM (lower panel).
NI
382, 402, 429
0.0014
—
0.09
0.020
4.5; 11
3.5
0.005
0.250
1NI2
540, 570
654
—
d
1C3
0.17
in agreement with former reports on naphthalene bisimide
luminescence in apolar solvents.15 The broad band at 540 nm
in DCM can be assigned to the same excimer or excited dimer;
it is bathochromically shifted with respect to the band in TL as
expected by the stabilization of charge transfer interactions in
the excited assembly in the more polar solvent. However, the
complexity of the band in DCM seems to indicate the presence
of higher order aggregates in addition to simple excimers
(or excited dimers).
C3
C3–NI
1C3-1NI
1C3-NI
o0.00005
652
0.01
0.011
a
Luminescence yields in air equilibrated solutions, excitation at
b
c
358 nm. Luminescence yields upon excitation at 560 nm. Excitation
at 355 nm and 532 nm in the picosecond range and at 373 nm in the
nanosecond range. d 1NI2 is a generic excited dimer/excimer or higher
aggregate (see text).
The luminescence properties of the dyad C3–NI were
probed with reference to the model components. Excitation
was carried out at 358 nm, where the NI component absorbs
55% of the photons in DCM and 51% in TL, and at 560 nm,
where selective excitation of corrole can be performed. The
concentrations of the solutions of the references NI and C3
were adjusted to absorb the same number of photons absorbed
by the respective unit in the dyad. The results of the
ꢂc
This journal is the Owner Societies 2010
476 | Phys. Chem. Chem. Phys., 2010, 12, 474–483

View Article Online
luminescence experiments upon excitations at the two wave-
lengths for the different solvents are illustrated in Fig. 4 and
Fig. 5. In TL the excitation of the dyad at 358 nm (Fig. 4 upper
panel) leads to a quenching of the fluorescence from the NI
component to less than 1%, whereas the luminescence of the
C3 unit is reduced to less than 20%, compared to the
references. Conversely, upon selective excitation of C3 unit
in the dyad at 560 nm (Fig. 4 lower panel) a quenching to ca.
10% of the luminescence of the reference corrole is obtained.
In DCM solutions, upon excitation in the UV (Fig. 5 upper
panel) the structured luminescence of NI is nearly completely
quenched whereas the luminescence of corrole reduces to ca.
10% with respect to that of the model. Upon selective
excitation of corrole at 560 nm (Fig. 5, lower panel) the
luminescence from this component reduces to ca. 5% that of
C3 in the same conditions. Overall, a more efficient quenching
of the corrole unit luminescence is detected in more polar
DCM compared to TL solutions. It might also be noticed that
in both solvents a less efficient quenching of corrole luminescence
is noticed when this chromophore is excited in the UV together
with the other component rather than when selectively excited
in the visible range. Since the excited state energy of the NI
component is higher than that of C3, an energy transfer from
NI to C3 excited state could be feasible. In order to confirm
this hypothesis, an excitation spectrum in C3–NI with a
detection wavelength at 660 nm, the maximum of C3 emission,
was run. The results, displayed in Fig. 6 for the case of DCM
(the results are essentially identical in TL) clearly show the
presence of NI bands at 358 and 379 nm and consequently a
reasonable overlap with the absorption spectrum of the dyad.
This proves that excitation of the NI component leads to
luminescence from the corrole unit in the dyad, i.e. the
occurrence of an energy transfer C3–1NI - 1C3-NI. The
energy transfer is not quantitative as the non-perfect overlap
of excitation and absorption spectra in the 340–400 nm range
seems to suggest. In spite of energy transfer, no sensitization of
the corrole luminescence is detected in the dyad, whereas on
the contrary a quenching can be observed (Fig. 5 and 6). We
on the basis of the spectral distributions as the one due to the
monomer (40 ps) and to the excimer (2.3 ns), respectively. It
should be noticed that in a time resolved regime also the
fluorescence of the monomer, practically undetectable in the
steady state experiment in TL, can be revealed (Fig. 7). An
other important observation is that there is apparently no
generation of excimer/excited dimer upon decay of the monomer
but both are present immediately after the laser excitation
(Fig. 7); very likely the formation of the excited assembly
occurs only when pre-formed dimers in the ground state are
excited and can form immediately without requiring diffusion.
If this is the case the excited assembly should therefore be
more correctly identified as excited dimer rather than as
excimer. The decay of NI fluorescence in DCM is more
complex. On the picosecond scale a decay around 400 nm
with a lifetime of 20 ps, ascribable to the monomer fluorescence
is detected, whereas on the nanosecond time scale a weak
emission with a double exponential decay of 4.5 and 11 ns in
the wavelength range from 500 to 600 nm can be measured.
The monomer luminescence lifetime is consistent with a formerly
reported value of 10 ps in chloroform for a very similar
1
have to assume that the sensitized state C3-NI is depleted by
some other process whose nature will be addressed below.
In principle also dyads might aggregate through the NI
component but it is not possible to have direct information
on this from the luminescence signatures since the luminescence
of naphthalene bisimide is completely quenched in both
solvents. In the dyad steric hindrance by the connected corrole
inhibits the p–p stacking of NI component. In addition,
electronic factors within the dyad, i.e. the electron-poor
bisimide interacting with electron-rich corrole lead to unfavorable
conditions for electronic interactions between the NI moieties
of two C3–NI preventing association. Hence, the dyad C3–NI
can safely be considered free from aggregation phenomena in
both solutions and the reference model of the bisimide moiety
in the dyad will be monomeric NI.
Time resolved spectroscopy
Fig. 4 Luminescence spectra of NI–C3 (gray) A = 0.33, NI (dash)
A = 0.14 and C3 (solid) A = 0.14 in TL, excitation at 358 nm (upper
panel). Luminescence spectra of NI–C3 (gray) A = 0.24 and C3 (solid)
A = 0.24 in TL, excitation at 560 nm (lower panel).
The decay lifetime of model NI luminescence can be fitted over
the entire spectral range by a bi-exponential function with
lifetimes of 40 and 2.3 ns in TL. The lifetimes can be assigned
ꢂc
This journal is the Owner Societies 2010
Phys. Chem. Chem. Phys., 2010, 12, 474–483 | 477

View Article Online
Fig. 7 Normalized time profiles of the decay of NI luminescence in
TL measured at 400 nm (gray) with the bi-exponential fitting (black).
The time profile at 500 nm (black scattered) is also shown. Excitation
was at 355 nm, with a 35 ps pulse of 1.3 mJ. In the inset the time
resolved spectra collected with a gate of 40 ps at the end of pulse (gray)
and 200 ps after the end of the pulse (black).
in the TL case. In general, it might be observed that whereas
data on the triplet excited state of this class of compounds
widely agree,13,17 the photophysics of singlet excited state is
still far from being settled.13,16,18 Apparently both solvent and
substitution pattern at the imidic nitrogen have strong effects
on the early steps of light energy dissipation processes, very
likely due to aggregation phenomena and/or charge transfer
interactions.
The lifetime of C3 is 3.8 ns in TL and 3.5 ns in DCM,
respectively.4 In the dyad C3–NI excitation was performed
both at 355 nm where the corrole and naphthalene bisimide
components are excited, and at 532 nm where only corrole is
excited. Upon excitation at 355 nm, the luminescence lifetime
registered at 400 nm (NI fluorescence) in the C3–NI dyad both
in TL and DCM is reduced to 5 ps, as displayed for the DCM
case in Fig. 8. After excitation of C3–NI in the corrole band
at 532 nm, a lifetime of 480 ps in TL and 250 ps in DCM
is detected for the corrole fluorescence at 650 nm. This
corresponds to a reduction with respect to the model lifetimes
to ca. 13% in TL and 7% in DCM, in good agreement with the
reduction in steady state luminescence data (20% in TL and
5% in DCM).
Fig. 5 Luminescence spectra of NI–C3 (gray) A = 0.28, NI (dash)
A = 0.10 and C3 (solid) A = 0.13 in DCM, excitation at 358 nm
(upper panel). Luminescence spectra of NI–C3 (gray) A = 0.16 and C3
(solid) A = 0.16 in DCM, excitation at 560 nm (lower panel).
The reduction in corrole fluorescence yield and lifetime is
clearly indicative of the occurrence of a deactivation step
competitive to the intrinsic deactivation of the chromophores
which we can postulate to be electron transfer. In order to
confirm this, transient absorbance experiments with pico-
second resolution were performed. This type of experiment
allow to identify the presence of oxidized and reduced radicals,
signatures of the presence of charge separated (CS) states
formed upon intra-molecular electron transfer. The radical
NIꢁ is known to display characteristic bands at 475 nm
(e ca. 30 000) and 610 nm (e ca. 9000)13 whereas the C3+
cation is known to absorb around 650–700 nm with a rather
low absorption coefficient.19 The end of pulse spectrum after
Fig. 6 Normalized absorption (gray line) and excitation spectrum
(black line) of NI–C3 in DCM solutions. The excitation spectrum has
been registered at 660 nm.
derivative.16 We ascribe the nanosecond lifetimes to fluorescence
from aggregated forms of NI, including very likely a dimer as
ꢂc
This journal is the Owner Societies 2010
478 | Phys. Chem. Chem. Phys., 2010, 12, 474–483

View Article Online
Fig. 9 End of pulse absorption spectrum in a DCM solutions of
C3–NI, in the inset is the time evolution of the absorption at 475 nm
and the fitting according to a double exponential decay with a rise of
240 ps and a decay of 250 ps. Excitation at 532 nm with a 35 ps pulse of
3.5 mJ, A = 0.54.
Fig. 8 Normalized time profiles measured at 400 nm for the decay of
NI luminescence in DCM (line) and C3–NI (full circles). Excitation at
355 nm with a 35 ps pulse of 1.3 mJ.
excitation at 532 nm of C3–NI in DCM is displayed in Fig. 9
and exhibits the NIꢁ bands at 475 and 610 nm. The time
evolution registered at 475 nm can be satisfactorily fitted by a
bi-exponential with a rising part of 240 ps and a decaying
component of 250 ps, respectively. In TL the end of pulse
spectrum displays again the spectral features of the NIꢁ
radical and grows in after the end of pulse, Fig. 10. The time
evolution registered on the band maximum shows a rising and
a decaying component of 480 and 1.4 ns, respectively (inset of
Fig. 10). In both solvents the rising part of the radical anion
band is coincident with the fluorescence lifetime of the corrole
component in the dyad, see Table 1, which indicates that we
are observing the product of the reaction depleting the excited
1
state C3-NI. This confirms the nature of the reaction as an
electron transfer reaction. In order to check for further
absorbing products on longer time scales, flash photolysis
experiments were performed. These indicate the presence in
the microsecond time scale of an absorption with a maximum
at l o 450 nm, with bleaching bands at 560 nm and around
615–640 nm and no absorbance above 700 nm, in good
agreement with the triplet excited state of free-base corrole.4
Fig. 10 Absorption spectrum detected at 600 ps delay, corresponding
to the full evolution of the species, in a TL solutions of C3–NI.
The inset displays the time evolution of the absorption at 475 nm and
the fitting according to a double exponential decay with a rise of 480 ps
and a decay of 1.4 ns. Excitation at 532 nm with a 35 ps pulse of
3.5 mJ, A = 0.45.
3
By comparison with the C3 signal obtained from the model
solution in the same experimental conditions, a relative yield
3
for C3-NI of ca. 20% in TL and of ca. 10% in DCM can be
calculated. Fig. 11 displays the results for TL solutions.
Photoinduced processes
A qualitative energy level scheme for the dyad C3–NI in both
solvents is shown in Fig. 12. The singlet and triplet excited
state energy levels for C3 and NI (in the monomeric form)
derived from the luminescence maxima at 77 K, are available
from former reports.4,10,13 We can safely assume, given the
very modest electronic coupling of the moieties as evidenced
by the spectroscopic data, the same energy levels for the dyad
C3–NI, i.e. 1.9 eV for 1C3-NI and ca. 1.5 eV for 3C3-NI, 3.2 eV
for C3–1NI and 2.0 eV for C3-3NI. An approximate energy of
Fig. 11 Transient absorption spectrum detected with 5 ms delay after
a laser pulse (532 nm, 18 ns pulse, 3 mJ) in a TL solutions of C3–NI
(open circles) and of the reference C3 (full circles), A = 0.7 for both
solutions at 532 nm.
the charge separated state relative to the ground state, DGCS
,
can be derived from the literature values of the first oxidation
potential of the donor C3 (Eox = +0.86 V vs. SCE in
benzonitrile),19 and the first reduction potential of the acceptor
(Ered = ꢁ0.48 V vs. SCE in DMF),20 in polar solvents.
In order to derive the energy level corresponding to C3+-NIꢁ,
the electrochemical potentials need to be corrected by taking
ꢂc
This journal is the Owner Societies 2010
Phys. Chem. Chem. Phys., 2010, 12, 474–483 | 479

View Article Online
into account the Coulombic interaction term and the ion
solvation energy based on the Born dielectric continuum
model, when the solvent is different from the one used in the
electrochemical determinations, according to the following
equation:21
formation, allowing the state to accumulate. In principle,
recombination might occur to the triplet excited state localized
on corrole, 3C3-NI, which has an energy similar to the energy
of CS in DCM and lower than that in TL. This would require
a spin rephasing of the original CS state generated as a singlet
state. Otherwise, charge recombination can occur to the singlet
ground state C3–NI. Though recombination to the triplet state
has been reported for other corrole dyads,10 in this case
the evidence indicate a ground state recombination. In fact
the yield of triplet state fT is reduced, compared to C3, to
ca. 20% in TL and ca. 10% in DCM, consistent with the yield
from intersystem crossing from 1C3-NI which can be
calculated to be 13 and 7% in TL and DCM, respectively.
Quite clearly the contribution of CS recombination to triplet
formation is virtually zero.
DG0 = Eox ꢁ Ered ꢁ (14.32/RDAeS)
CS
+ 14.32 (1/2rD + 1/2rA) (1/eS ꢁ 1/eP)
(1)
In the present case, eS = 2.4 and eS = 8.9 are the dielectric
constant of TL and DCM, eP is taken as 31, an average of the
dielectric constant of benzonitrile (25.9) and of DMF (36.7).
The center to center donor–acceptor distance calculated after
energy minimization is 12.4 A whereas NI and C3 radii are
taken rA = 3.5 A and rD = 5 A, respectively.
The results of the calculation indicate for C3+-NIꢁ a value
of ca. 2.2 eV in TL and of ca. 1.5 eV in DCM. It is well
recognized that the correction overestimates the energy of the
charge separated state in low polarity solvents as TL.22 This is
quite evident also in this case, where the result of the calculations
place the CS state energy level at an energy well above that of
the excited 1C3-NI (1.9 eV), so that photoinduced charge
separation reaction starting from this state would be endergonic
by ca. 0.3 eV. It is quite obvious that this is not so, given the
fast rate measured (see below) for charge separation in TL. In
order to correct for this, a reduction of 0.35 eV has been
formerly proposed to yield a more realistic figure.22 By applying
a similar reduction, an energy of ca. 1.85 eV can be calculated
for C3+-NIꢁ. We are conscious that this is an approximation
and our discussion will accordingly be mostly qualitative.
Selective excitation of the singlet excited state of the corrole
unit of the dyad to the singlet excited state 1C3-NI is followed
by electron transfer from corrole to naphthalene bisimide
forming C3+-NIꢁ, as demonstrated by the detection of the
spectral features of naphthalene bisimide radical anion. The rate
constant of the charge separation reaction, kcs, can be
calculated by the equation kcs = 1/t ꢁ 1/t0, where t and t0
represent the lifetime of the fluorescence in the dyad and in the
model C3 and is, respectively 1.8 ꢀ 109 sꢁ1 in TL and 3.7 ꢀ 109 sꢁ1
in DCM. C3+-NIꢁ subsequently recombines with a rate
constant kcr (kcr = 1/t) which is 4.0 ꢀ 109 sꢁ1 in DCM,
similar to its formation rate. On the contrary, in TL, CS
decays with a lifetime of 7.1 ꢀ 108 sꢁ1 well lower than its
UV excitation of the singlet excited state of naphthalene
bisimide, C3–1NI, is followed by a very rapid decay with a
lifetime of 5 ps, corresponding to a rate (or combination of
rates) of 2.0 ꢀ 1011 sꢁ1 in both solvents. Responsible for the
quenching is energy transfer to the 1C3-NI state, as documented
by the excitation spectra and by some sensitization noticed in
this state after excitation in the UV compared to direct
excitation in 1C3-NI, but a minor contribution from a
HOMO–HOMO electron transfer from corrole to the excited
naphthalene imide unit is thermodynamically possible. By
inspecting the time evolution of the C3+-NIꢁ band at
475 nm, inset of Fig. 10, one can see a fast (instantaneous
with the laser excitation) and a slow formation of the DA.
However a quantization of the amount of C3+-NIꢁ present at
the end of pulse, and therefore derived by electron transfer
from C3–1NI, is made difficult by the presence of other
absorbing species in the same spectral region, e.g.
C3–1NI.13,16 Based on the excitation spectra and on the nearly
perfect overlap of the NI monomer emission with the C3 Soret
absorption band, which favors an energy transfer via a Forster
mechanism,23 we are prone to consider energy transfer, possibly
via upper excited states of corrole, as the main deactivation
1
mechanism of C3–1NI. The fate of the sensitized C3-NI state
will be identical to that directly excited, discussed above.
Formation of a CS state in C3–NI dyad has a yield of ca.
87% in TL and 93% in DCM, the difference reflecting the
higher driving force for the charge separation reaction in
DCM, see Fig. 12. Therefore DCM solutions are more efficient
in assisting charge separation, however the lifetime of the
charge separated state is lower in DCM, 250 ps as compared
to TL where the lifetime is of 1.4 ns. The reason for the
increase in lifetime has to be assigned to the higher driving
force for charge recombination DG0 in TL (ca. 1.85 eV)
compared to DCM (ca. 1.5 eV), in agreement with the well
known inverted behavior for highly exergonic reactions.24
According to this theory, when the reaction DG0 are higher
than the reorganization energy, the rate of the reaction
decreases for higher driving forces. Low polarity TL allows
a longer lifetime and the possibility to store an higher amount of
energy in the CS state and represents the optimal experimental
conditions, though the yield of CS is slightly lower than
in DCM.
The lifetime of C3+-NIꢁ is of the same order as that
Fig. 12 Energy level scheme and deactivation paths for excitation of
C3–NI in TL and DCM.
measured for the corresponding corrole–perylene bisimide
ꢂc
This journal is the Owner Societies 2010
480 | Phys. Chem. Chem. Phys., 2010, 12, 474–483

View Article Online
dyad, 2.5 ns, reflecting a similarity in the thermodynamic and
electronic conditions of the system. In fact, NI reduces at a
slightly more negative potential (o0.1 eV20,25) than the corres-
ponding perylene bisimide (PI), but other parameters, like
molecular dimensions, might counterbalance this. Several
porphyrin/naphthalene bisimide dyads have been reported so
far but apparently no identical structure containing free-base
porphyrin rather than corrole has been studied, making
impossible a direct comparison of the porphyrin and corrole
performances.
Mass spectra were obtained via EI or electrospray MS
(ESI-MS). The following molecules were prepared according to
literature procedures: 2 (ref. 7) and 4 (ref. 26).
Bisimides
3 and NI. Amine 2 (4 mmol, 880 mg) and decylamine (4 mmol,
800 mL) were dissolved in DMF (10 mL), at 130 1C. Subsequently,
bisanhydride 1 (4 mmol, 1072 mg) was added and the reaction
mixture was stirred at 170 1C for 16 h. Methanol (50 mL) was
added and the reaction mixture was cooled to RT. Filtration
afforded off-white solid (2.28 g), which was dried and dissolved
in CH2Cl2 (40 mL). TFA (10 mL) and H2O (1.5 mL) was
added to this solution and the reaction mixture was stirred at
RT for 24 h. Subsequently H2O and solid Na2CO3, were added
to neutralize TFA. Extraction with CH2Cl2 was performed
and organic layer was washed with water, dried (Na2SO4) and
evaporated. Chromatography (SiO2, CH2Cl2/heksan (4 : 1))
afforded bisimide NI (546 mg, 25%) and bisimide 3 (418 mg,
20%) as white solids. N,N0-Didecyl-1,4,5,8-naphthalenetetra-
carboxylic acid diimide (NI): (Found: C, 74.62; H, 8.68; N,
5.18; C34H46N2O4 requires C, 74.69; H, 8.48; N, 5.12);
dH(500 MHz; CDCl3, Me4Si) 0.87 (t, 6H, J 6.9, CH3), 1.26
(m, 20H, CH2), 1.40 (m, 8H, CH2), 1.74 (m, 4H, CH2),
Conclusions
A dyad comprising free-base corrole and the electron acceptor
naphthalene bisimide, C3–NI, was successfully synthesized
and the spectroscopy as well as the photo-reactivity of
the dyad was examined in solvents of different polarity. The
properties of the simple models were also examined in the
same solvents and association phenomena were evidenced in
NI solutions. In the dyad association is not possible due to
steric congestion and electronic factors. An efficient charge
separation reaction with yields of the order of 90% occurs in
apolar and moderately polar solvents upon excitation
throughout the UV and the visible wavelength range of
C3–NI. In the UV, the NI unit is excited and the resulting
4.19 (t, 4H,
J 7.6, CH2N), 8.75 (s, 4H, aromCH);
dC(125 MHz; CDCl3, Me4Si) 14.1, 22.7, 27.1, 28.1, 29.2,
29.3, 29.5, 29.6, 31.9, 41.0, 126.6, 126.7, 130.9, 162.8; m/z EI
546.3477 (M+, C34H46N2O4 requires 546.3458); found: lmax
(CH2Cl2)/nm 310, 341, 360, 380, 528 and 626.
1
C3–1NI state transfers energy to the lower lying C3–NI with
rates of the order of 1011 ꢁ1, irrespective of solvent polarity.
s
Direct excitation of the latter state in the visible wavelength
range, as well as its indirect formation by sensitization, is
followed by a LUMO–LUMO electron transfer producing
C3+–NIꢁ with rates of the order of 109 sꢁ1, about twice in
DCM than in TL. Charge recombination to the ground state
has a rate of the order of 108 sꢁ1 in TL and is an order of
magnitude faster in DCM. The corrections proposed by
Weller for the calculations of the CS energy levels from the
redox potentials appear to be overestimated in apolar solvents,
as formerly suggested.22a
N-Decyl-N0-[(4-formylphenyl)methylen]-1,4,5,8-
naphthalenetetracarboxylic acid diimide (3):
(Found: C,73.10; H, 5.96; N, 5.36; C32H32N2O5 requires C,
73.26; H, 6.15; N, 5.34); dH(200 MHz; CDCl3, Me4Si) 0.94
(t, 3H, J 6.9, CH3), 1.2–1.9 (m, 16H, CH2), 4.19 (t, 2H, J 7.6,
NCH2), 5.44 (s, 2H, NCH2Ar), 7.69 7.85 (AA0BB0, 4H, J 7.3,
arom), 8.78 (s, 4H, arom), 9.98 (s, 1H, CHO); dC(50 MHz;
CDCl3, Me4Si) 1.0, 14.1, 22.7, 27.1, 28.0, 31.9, 41.0, 43.8,
126.2, 126.7, 126.8, 127.0, 129.5, 130.0, 131.0, 131.4, 135.8,
143.1, 162.7, 162.9, 191.7; m/z EI 524.2330 (M+, C32H32N2O5
requires 524.2311); lmax (CH2Cl2)/nm (e ꢀ 10ꢁ3): 292 (15),
309 (15), 342 (21), 360 (30), 380 (36) nm.
With the synthesis and characterization of C3–NI dyad, a
further example of a stable corrole assembly able to convert and
store light energy into a charge separated state with good
efficiency and reasonable lifetime is presented. The possibility
to use these arrays in devices for the conversion of light energy
has still to be exploited, but they present performances at least as
good as those of more popular porphyrin arrays. The dyad will
serve as a building block of more complex, photo-functional
molecular architectures presently under examination.
Dyad C3–NI
Aldehyde 3 (0.2 mmol, 100 mg) was dissolved in TFA solution
in CH2Cl2 (3 mL of the solution containing TFA (200 mL)
w
CH2Cl2 (100 mL). Subsequently dipyrromethane 5
(0.4 mmol, 124 mg) was added. The reaction mixture was
stirred at room temperature for 40 min. Then triethylamine
(6 mL) was added and the reaction mixture was diluted with
CH2Cl2 (43 mL). Subsequently, DDQ (0.5 mmol, 117 mg) was
added. Chromatography (SiO2, CH2Cl2/heksan (3 : 1) afforded
corrole C3–NI (34 mg, 16%) as violet crystals. Rf 0.33
(SiO2, CH2Cl2/heksan 3 : 2); dH(200 MHz; CDCl3, Me4Si)
0.80 (t, 3H, J 6.9, CH3), 1.1–1.8 (m, 16H, CH2), 4.01 (t, 2H,
J 7.6, NCH2CH2), 5.58 (s, 2H, NCH2Ar), 7.81, 8.05 (AA0BB0,
4H, arom), 8.48 (d, J 4.2, 2H, b-H), 8.59 (s, 4H, b-H), 8.66
(d, J 7.6, 2H, arom), 8.73 (d, J 7.6, 2H, arom), 9.02 (d, J 4.4,
2H, b-H); m/z (ESI) 1125.3 (M + H+, C62H43F10N6O4
Experimental
Synthesis
All chemicals were used as received unless otherwise noted.
Reagent grade solvents (CH2Cl2, hexanes, cyclohexane) were
distilled prior to use. All reported 1H NMR and 13C NMR
spectra were recorded on Bruker AM 500 MHz or Varian 200
MHz spectrometers. Chemical shifts (d ppm) were determined
with TMS as the internal reference; J values are given in Hz. UV-
Vis spectra were recorded in CH2Cl2 (Cary). Chromatography
was performed on silica (Kieselgel 60, 200–400 mesh).
ꢂc
This journal is the Owner Societies 2010
Phys. Chem. Chem. Phys., 2010, 12, 474–483 | 481

View Article Online
requires 1124.31); lmax (CH2Cl2)/nm (e ꢀ 10ꢁ3): 360 (52),
Angew. Chem., 2008, 120, 8283–8387; (b) S. Gunes, H. Neugebauer
¨
and N. S. Sariciftci, Chem. Rev., 2007, 107, 1324–1338.
2 D. Gust, T. A. Moore and A. L. Moore, Acc. Chem. Res., 2001, 34,
40–48.
3 (a) E. Baranoff, J.-P. Collin, L. Flamigni and J.-P. Sauvage, Chem.
Soc. Rev., 2004, 33, 147–155; (b) L. Flamigni, J. Photochem.
Photobiol., C, 2007, 8, 191–210; (c) L. Flamigni, J.-P. Collin and
J.-P. Sauvage, Acc. Chem. Res., 2008, 41, 857–871; (d) A. I. Oliva,
382 (70), 410 (133), 559 (33) nm.
Photophysics and spectroscopy
Spectrophotometric grade toluene and dichloromethane
(C. Erba) was used as supplied. A Perkin-Elmer Lambda
950 UV/Vis spectrophotometer was used to measure absorption
spectra in 10 mm cells. Spex Fluorolog II spectrofluorimeter
was used to acquire fluorescence spectra in standard 10 mm
fluorescence cells. The reported luminescence spectra are
uncorrected, emission quantum yields were determined after
correction for the photomultiplier response, with reference to
an air-equilibrated toluene solution of C3 with a Ffl = 0.14.4
Luminescence lifetimes in the nanosecond range were obtained
with an IBH single photon counting equipment with excitation
at 373 nm from pulsed diode sources, resolutions was ca. 0.3 ns
and absorbance of the solutions at the exciting wavelength was
of the order of 0.3. For determination of emission lifetimes in
the picosecond range an apparatus based on a Nd:YAG laser
(35 ps pulse duration, 532 nm, 1.5 mJ) and a single shot Streak
Camera (Hamamatsu C1587 equipped with a High Speed
Streak Unit M1952) was used. Solutions with absorbance of
ca. 0.3 at the exciting wavelength were irradiated by the laser
pulse and light was collected at a right angle. The time
resolution of the apparatus after deconvolution is ca. 5 ps.27
Transient absorbance in the picosecond range made use
of a pump and probe system based on a Nd-YAG laser
(Continuum PY62/10, 35 ps pulse, 532 nm, 3.5 mJ) and an
optical multichannel analyzer based detection apparatus.
Solutions with absorbance of ca. 0.5 at the exciting wavelength
were used. In order to increase the sensitivity in the wavelength
range below 500 nm, where the intensity of the analyzing white
continuum light is extremely low, a short band pass filter at
500 nm was used. This, without preventing detection of signals
above this wavelength, where the white continuum intensity is
quite high, allowed registration of the NIꢁ radical band at
475 nm and its time evolution with a better signal to noise
ratio. More details on the apparatus can be found elsewhere.28
Laser flash photolysis in the nanosecond range was performed
with a Nd-YAG laser (18 ns pulse, 532 nm, 3 mJ) and an
apparatus previously described.29 For flash photolysis
determinations the samples were bubbled with argon for ca.
15 min and sealed in home made 10 mm optical cells, the
absorbance of the solutions was ca. 1. Molecular dimensions
were estimated after MM2 minimization by CS Chem 3D
Ultra 6.0 software.30 Estimated errors are 10% on lifetimes for
single exponentials 20% on more complex kinetics, 20% on
quantum yields, 20% on molar absorption coefficients and
3 nm on emission and absorption peaks.
B. Ventura, F. Wurthner, A. Camara-Campos, C. A. Hunter,
¨
P. Ballester and L. Flamigni, Dalton Trans., 2009, 4023–4037.
4 B. Ventura, A. Degli Esposti, B. Koszarna, D. T. Gryko and
L. Flamigni, New J. Chem., 2005, 29, 1559–1566.
5 For reviews see: (a) M. R. Wasielewski, Chem. Rev., 1992, 92,
435–461; (b) D. Gust, T. A. Moore and A. L. Moore, Acc. Chem.
Res., 1993, 26, 198–205; (c) H. Himahori, J. Phys. Chem. B, 2004,
108, 6130–6143; (d) M. D. Ward, Chem. Soc. Rev., 1997, 26,
365–375; (e) L. Flamigni, V. Heitz and J.-P. Sauvage, Struct.
Bonding, 2006, 121, 217–261; (f) M. R. Wasielewski, J. Org. Chem.,
2006, 71, 5051–5066.
6 (a) R. Paolesse, F. Sagone, A. Macagnano, T. Boschi, L. Prodi,
M. Montalti, N. Zaccheroni, F. Bolletta and K. M. Smith,
J. Porphyrins Phthalocyanines, 1999, 3, 364–370; (b) L. Flamigni,
B. Ventura, M. Tasior and D. T. Gryko, Inorg. Chim. Acta, 2007,
360, 803–813.
7 M. Tasior, D. T. Gryko, M. Cembor, J. S. Jaworski, B. Ventura
and L. Flamigni, New J. Chem., 2007, 31, 247–259.
8 L. Flamigni, B. Ventura, M. Tasior, T. Becherer, H. Langhals and
D. T. Gryko, Chem.–Eur. J., 2008, 14, 169–183.
9 M. Tasior, D. T. Gryko, J. Shen, K. M. Kadish, T. Becherer,
H. Langhals, B. Ventura and L. Flamigni, J. Phys. Chem. C, 2008,
112, 19699–19709.
10 F. D0Souza, R. Chitta, K. Ohkubo, M. Tasior, N. K. Subbaiyan,
M. E. Zandler, M. K. Rogacki, D. T. Gryko and S. Fukuzumi,
J. Am. Chem. Soc., 2008, 130, 14263–14272.
11 L. Flamigni and D. T. Gryko, Chem. Soc. Rev., 2009, 38,
1635–1646 and references therein.
12 (a) S. V. Bhosale, C. H. Jani and S. J. Langford, Chem. Soc. Rev.,
2008, 37, 331–342; (b) T. Yamazaki, I. Yamazaki and A. Osuka,
J. Phys. Chem. B, 1998, 102, 7858–7865; (c) O. Johansson,
M. Borgstrom, R. Lomoth, M. Palmblad, J. Bergquist,
¨
L. Hammarstrom, L. Sun and B. Akermark, Inorg. Chem., 2003,
¨
42, 2908–2918; (d) K. Okamoto, Y. Mori, H. Yamada,
H. Himahori and S. Fukuzumi, Chem.–Eur. J., 2004, 10,
474–483; (e) L. Flamigni, M. R. Johnston and L. Giribabu,
Chem.–Eur. J., 2002, 8, 3938–3947; (f) G. P. Wiederrecht,
M. P. Niemczyk, W. A. Svec and M. R. Wasielewski, J. Am.
Chem. Soc., 1996, 118, 81–88; (g) M. E. El-Khouly, J. H. Kim,
K.-Y. Kay, C. S. Choi, O. Ito and S. Fukuzumi, Chem.–Eur. J.,
2009, 15, 5301–5310; (h) R. F. Kelley, M. J. Tauber and
M. R. Wasielewski, J. Am. Chem. Soc., 2006, 128, 4779–4791.
13 L. Flamigni, E. Baranoff, J.-P. Collin and J.-P. Sauvage,
Chem.–Eur. J., 2006, 12, 6592–6606.
14 A. Ghosh, T. Wondimagegn and A. B. J. Parusel, J. Am. Chem.
Soc., 2000, 122, 5100–5104.
15 (a) V. Wintgens, P. Valat, J. Kossanyi, L. Biczok, A. Demeter and
T. Berces, J. Chem. Soc., Faraday Trans., 1994, 90, 411–421;
(b) T. C. Barros, S. Brochsztain, V. G. Toscano, P. B. Filho and
M. J. Politi, J. Photochem. Photobiol., A, 1997, 111, 97–104.
16 P. Ganesan, J. Baggerman, H. Zhang, E. J. R. Sudholter and
¨
Hong Zuilhof, J. Phys. Chem. A, 2007, 111, 6151–6156.
17 (a) B. M. Aveline, S. Matsugo and R. W. Redmond, J. Am. Chem.
Soc., 1997, 119, 11785–11795; (b) I. V. Sazanovich, M. A. H.
Alamiry, J. Best, R. D. Bennett, O. V. Bouganov, E. S. Davies,
V. P. Grivin, A. J. H. M. Meijer, V. F. Plyusnin, K. L. Ronayne,
A. H. Shelton, S. A. Tikhomirov, M. Towrie and J. A. Weinstein,
Inorg. Chem., 2008, 47, 10432–10445; (c) J. E. Rogers, S. J. Weiss
and L. A. Kelly, J. Am. Chem. Soc., 2000, 122, 427–436.
Acknowledgements
We thank CNR of Italy (PM.P04.010, Project MACOL) and
Polish Ministry of Research and Higher Education.
18 S. E. Miller, A. S. Lukas, E. Marsh, P. Bushard and M. R.
Wasielewski, J. Am. Chem. Soc., 2000, 122, 7802–7810.
19 J. Shao, J. Shen, Z. Ou, W. E. B. Koszarna, D. T. Gryko and
K. M. Kadish, Inorg. Chem., 2006, 45, 2251–2265.
20 D. Gosztola, M. P. Niemczyk, W. Svec, A. S. Lukas and
M. R. Wasielewski, J. Phys. Chem. A, 2000, 104, 6545–6551.
21 A. Weller, Z. Phys. Chem., 1982, 133(1), 93–98.
References
1 (a) H. Choi, S. Kim, S. O. Kang, J. Ko, M.-S. Kang, J. N. Clifford,
A. Forneli, E. Palomares, M. K. Nazeeruddin and M. Gratzel,
¨
ꢂc
This journal is the Owner Societies 2010
482 | Phys. Chem. Chem. Phys., 2010, 12, 474–483

View Article Online
22 (a) N. Mataga, H. Chosrowjan, S. Taniguchi, Y. Shibata,
N. Yoshida, A. Osuka, T. Kikuzawa and T. Okada, J. Phys. Chem.
A, 2002, 106, 12191–12201; (b) M. Fujitsuka, H. Shimakoshi,
S. Tojo, L. L. Cheng, D. Maeda, Y. Hisaeda Y and T. Majima,
J. Phys. Chem. A, 2009, 113, 3330–3335.
26 J. K. Laha, S. Dhanalekshmi, M. Taniguchi, A. Ambroise and
J. S. Lindsey, Org. Process Res. Dev., 2003, 7, 799–812.
27 L. Flamigni, B. Ventura, A. I. Oliva and P. Ballester, Chem.–Eur.
J., 2008, 14, 4214–4224 and references therein.
28 L. Flamigni, A. M. Talarico, S. Serroni, F. Puntoriero,
M. J. Gunter, M. R. Johnston and T. P. Jeynes, Chem.–Eur. J.,
2003, 9, 2649–2659 and references therein.
29 L. Flamigni, J. Phys. Chem., 1992, 96, 3331–3337.
30 CS Chem 3D Ultra CambridgeSoft. Com, Cambridge MA, USA,
2000.
23 Th. Forster, Discuss. Faraday Soc., 1959, 27, 7–17.
¨
24 R. A. Marcus, Angew. Chem., Int. Ed. Engl., 1993, 32, 1111–1121
and references therein.
25 A. Viehbeck, M. J. Goldberg and C. A. Kovac, J. Electrochem.
Soc., 1990, 137, 1460–1466.
ꢂc
This journal is the Owner Societies 2010
Phys. Chem. Chem. Phys., 2010, 12, 474–483 | 483