Excited-state dynamics of porphyrin-naphthalenediimide-porphyrin triads

Article abstract of DOI:10.1039/c2cp43595k

The excited-state dynamics of two triads consisting of a naphthalenediimide (cNDI) substituted at the core by two zinc (ZnP) or free-base tetraphenylporphyrins (FbP) was investigated by ultrafast fluorescence and transient absorption spectroscopy. The electronic absorption spectra of the triads are almost the composites of those of the constituents, pointing to a weak electronic coupling and to a localization of the excitation energy on one of the porphyrins. In cyclohexane, the excited-state dynamics of the triads are essentially the same as those of the individual porphyrins, with the exception of the Soret emission of the ZnP triad, whose lifetime exhibits a more than 10 fold shortening compared to ZnP. A similarly ultrafast fluorescence decay was measured in tetrahydrofuran and benzonitrile. In these two solvents, charge separation from the excited porphyrin to the cNDI was found to take place with ～1 ps and ～25 ps time constants in the ZnP and FbP triads, respectively. The build up of the charge-separated state population in the ZnP triad is independent of the excitation wavelength, indicating that charge separation takes place from the lowest singlet excited state. Charge recombination occurs with a time constant of around 8 ps in both triads, i.e. it is slower than charge separation in the ZnP triad but faster in the FbP triad. These differences are rationalized in terms of the driving forces for charge separation and recombination. the Owner Societies 2013.

Full text of DOI:10.1039/c2cp43595k

PCCP
View Article Online
View Journal | View Issue
PAPER
Excited-state dynamics of porphyrin–
naphthalenediimide–porphyrin triads†
Cite this: Phys. Chem. Chem. Phys., 2013,
15, 1177
Diego Villamaina,a Sheshanath V. Bhosale,z*b Steven J. Langfordb and
Eric Vauthey*a
The excited-state dynamics of two triads consisting of a naphthalenediimide (cNDI) substituted at the
core by two zinc (ZnP) or free-base tetraphenylporphyrins (FbP) was investigated by ultrafast
fluorescence and transient absorption spectroscopy. The electronic absorption spectra of the triads are
almost the composites of those of the constituents, pointing to a weak electronic coupling and to a
localization of the excitation energy on one of the porphyrins. In cyclohexane, the excited-state
dynamics of the triads are essentially the same as those of the individual porphyrins, with the exception
of the Soret emission of the ZnP triad, whose lifetime exhibits a more than 10 fold shortening
compared to ZnP. A similarly ultrafast fluorescence decay was measured in tetrahydrofuran and
benzonitrile. In these two solvents, charge separation from the excited porphyrin to the cNDI was
found to take place with B1 ps and B25 ps time constants in the ZnP and FbP triads, respectively. The
build up of the charge-separated state population in the ZnP triad is independent of the excitation
wavelength, indicating that charge separation takes place from the lowest singlet excited state. Charge
recombination occurs with a time constant of around 8 ps in both triads, i.e. it is slower than charge
separation in the ZnP triad but faster in the FbP triad. These differences are rationalized in terms of the
driving forces for charge separation and recombination.
Received 11th October 2012,
Accepted 20th November 2012
DOI: 10.1039/c2cp43595k
www.rsc.org/pccp
ZnP constitutes one of the few exceptions of Kasha’s rule, as
it exhibits fluorescence from the S2 state with a sufficient
Introduction
quantum yield (B8 ꢀ 10ꢁ4 11,12
)
to be detected using any
Over the past few years, considerable efforts have been invested
in the design and synthesis of complex molecular architectures
aimed to mimic the functions of the photosynthetic apparatus of
plants and bacteria.1–4 In nature, solar energy is harvested by
antenna complexes and funneled toward the reaction center,
where it is used to achieve long-lived charge separation (CS),
allowing its conversion into chemical energy. In these complex
natural systems, chlorophylls play a key role in both the light
collection and the CS steps and, therefore, many artificial
antenna systems and reaction centers are based on porphyrins,
in particular zinc and free-base tetraphenylporphyrins (ZnP and
FbP), which combine broad absorption spectra in the visible
region and suitable redox properties for photoinduced CS.1,2,5–10
standard fluorimeter.13 This is not only due to the very large
oscillator strength of the S2 ’ S0 transition, responsible for the
intense Soret or B band at around 420 nm, but also to the
relatively slow internal conversion from this state to the lower
S1 or Q state, namely between 1 and 3 ps, depending on the
solvent.12,14,15 The latter is explained by the large S2–S1 energy
gap and the small horizontal displacement of the S2 and S1
potential energy surfaces. This comparatively long lifetime
allows in principle other non-radiative processes to compete
with internal conversion. For example, S2 energy hopping
between ZnP units in porphyrin arrays has been observed to
take place on the subpicosecond timescale.8,16
Photoinduced CS from the S2 state offers the possibility to
store about 1 eV more energy than CS from the S1 state. Such CS
from the local S2 state of ZnP has been reported for several
porphyrin–acceptor dyads.17–20 For example, Mataga and
co-workers, using a series of directly linked zinc porphyrin–
imide dyads, have made the first observation of the Marcus
inverted region for intramolecular photoinduced CS reaction.18
Wasielewski and co-workers have shown that photoinduced
a Department of Physical Chemistry, University of Geneva, 30 quai Ernest-Ansermet,
`
CH-1211 Geneve 4, Switzerland. E-mail: eric.vauthey@unige.ch
b School of Chemistry, Monash University, Clayton, Victoria 3800, Australia
† Electronic supplementary information (ESI) available: Synthesis of the triads,
additional transient absorption spectra, and calculations of the excitation energy
hopping rate constants. See DOI: 10.1039/c2cp43595k
‡ Present address: School of Applied Sciences, RMIT University, GPO Box 2476V,
Melbourne, Victoria 3001, Australia. E-mail: sheshanath.bhosale@rmit.edu.au
c
This journal is the Owner Societies 2013
Phys. Chem. Chem. Phys., 2013, 15, 1177--1187 1177

View Article Online
Paper
PCCP
electron transfer from the S2 state of ZnP can occur even if the
Additionally, the covalent binding of electron donor units to
donor and acceptor units are not directly linked, but covalently the naphthalene core to build a D–A–D triad has been shown to
connected through a phenyl ring.19 They have also found that increase the CS rate constant by three orders of magnitude
the decay of the local S2 state through CS is faster when the compared to the conventional linkage via the imide N atoms.36
acceptor is linked to the ZnP ring at beta than at the meta We have recently reported on the charge transfer dynamics in
¨
position. Hammarstrom and co-workers have also observed CS two dyads consisting of a ZnP or a FbP unit linked through a
between the S2 state of ZnP and methylviologen in a donor– rigid diaza bridge to the core of a cNDI (Scheme S1, ESI†).37
acceptor complex.21 Our group has reported on the bimolecular Such binding resulted in a strong coupling between both
quenching of ZnP in the S2 state by both electron donors and porphyrin and cNDI units. As a consequence, the electronic
acceptors.22 In this case, the driving force for CS was such that absorption spectra of these dyads differed substantially from
the charge-separated state (CSS) was located between the S2 and the sum of those of the individual units and the electronic
S1 states of the ZnP unit and, as a consequence, charge excitation was delocalized over the whole dyads. Upon optical
recombination (CR) was found to populate the neutral ZnP in excitation, CS from the cNDI to the porphyrin unit was found to
the S1 state. Such ultrafast CR to the local S1 state of ZnP has take place within 1–3 ps in medium polarity solvents. However,
also been observed in a non-symmetric acceptor–ZnP–acceptor the large electronic coupling prevented long-lived CS, and CR
triad.20 In most of the above cases, the CS rate constants were time constants of 10–20 ps were measured.
determined by comparing the S2 lifetime of ZnP alone and in
the dyad or upon addition of electron donor/acceptor.
The triads investigated here differ, first by the number of
porphyrins, but more importantly by the linkage between the
The S1 state of ZnP consists of two degenerate levels, which porphyrin and the cNDI units. The conformational flexibility
are split in FbP because of its lower symmetry. This results in a introduced in these triads by the single bond should lead to
lower energy gap between the S2 state and the S1 state manifold, a substantially smaller electronic coupling. Consequently,
an acceleration of the S2 - S1 internal conversion and a electronic excitation should remain localized on either the
vanishingly small S2 fluorescence quantum yield. Therefore, porphyrin or the cNDI, allowing the occurrence of CS from the
all the relevant photoinduced processes in FbP occur from its S2 state of the ZnP moiety to be envisaged. Additionally, both CS
long-lived S1 state.
and CR can be expected to be substantially slower. We will show
We report here on an investigation of two triads consisting that although the coupling is small enough for the constituents
of a naphthalenediimide core substituted at 2 and 6 positions to keep their identity, it is still large enough to result in ultrafast
with two ZnP or FbP units (Scheme 1). Over the past few years, CS and CR. Additionally, we will show that the shortening of the
core-substituted naphthalenediimides (cNDIs) have regained S2 lifetime of the ZnP unit in the triad compared to ZnP alone is
interest because of their redox and photophysical proper- not due to the occurrence of CS but to a perturbation of the
ties.23,24 Indeed, in contrast to NDIs substituted at the imide electronic structure ZnP upon binding to cNDI.
nitrogen atoms, that have been often used in electron donor–
acceptor (D–A) systems,4,25–27 cNDIs substituted with electron-
Experimental details
Samples
donating groups absorb in the visible region and are generally
highly fluorescent.28 The S1 state of these cNDIs is character-
ized by a substantial charge transfer character and its energy The solvents cyclohexane (CHX), tetrahydrofuran (THF), and
relative to the ground state decreases with the electron donat- benzonitrile (BCN) were of the highest commercially available
ing ability of the substituents, allowing the whole visible region purity and were used as received. The synthesis of the triads 1
of the spectrum to be covered upon variation of a few atoms and 2 is described in detail in the ESI.† In brief, 2 was
only. The decrease in the S1–S0 energy gap with increasingly synthesized following standard Suzuki cross-coupling reaction
donating core substituents is accompanied by a parallel between dibromo-NDI and monoboronic ester porphyrin in
increase in the energies of the frontier molecular orbitals and 76% yield in the presence of Pd(PPh3)4 and Na2CO3 in
thus by a continuous variation of the redox potentials. This toluene.38 The triad 1 was prepared by reacting 2 with excess
makes cNDIs excellent building blocks for realizing multi- Zn(OAc)2ꢂ2H2O in chloroform/methanol (10 : 1, v/v).
chromophoric systems combining light harvesting, charge
separation, charge transport and redox cascades.29–35
Spectroscopic measurements
Absorption spectra were recorded on a Cary 50 spectro-
photometer, whereas fluorescence and excitation spectra were
measured on a Cary Eclipse fluorimeter. All fluorescence spectra
were corrected for the wavelength-dependent sensitivity of the
detection. For fluorescence measurements, the adsorbance at
the excitation wavelength was around 0.1 over 1 cm. The
fluorescence quantum yields were measured using ZnP and
FbP as standards.39,40
Fluorescence decay measurements on the nanosecond time-
Scheme 1
scale were performed using the time-correlated single photon
c
1178 Phys. Chem. Chem. Phys., 2013, 15, 1177--1187
This journal is the Owner Societies 2013

View Article Online
PCCP
Paper
counting (TCSPC) setup described in ref. 41 and 42. Excitation
was carried out at 395 nm using a laser diode (Picoquant Model
LDH-P-C-400B) generating B60 ps pulses at 10 MHz. The full
width at half maximum (FWHM) of the instrument response
function (IRF) was around 200 ps.
Fluorescence dynamics occurring on a shorter timescale
were measured using the fluorescence up-conversion setup
described before.22,43 Excitation was performed at 400 nm with
the frequency-doubled output of a Kerr lens mode-locked
Ti:sapphire laser (Mai Tai, Spectra Physics). The polarization
of the pump pulses was at magic angle relative to that of the
gate pulse at 800 nm. The pump intensity on the sample was of
the order of 5 mJ cmꢁ2, and the FWHM of the IRF was ca. 200 fs.
The sample solutions were located in a 0.4 mm rotating cell and
had an absorbance of about 0.1 at 400 nm.
The experimental setup for the transient absorption (TA)
measurements has been described in detail earlier in ref. 32
and 44. Excitation was performed at 400 nm with the frequency- Fig. 1 Absorption spectra of the triads 1 and 2 and of the individual constituents
in CHX (the insets show the B band region).
doubled output of a standard 1 kHz amplified Ti:sapphire
system (Spectra Physics) and at 555 nm with a home-built
two-stage non-collinear optical parametric amplifier. The pump
intensity on the sample was ca. 1–2 mJ cmꢁ2. The polarization to a small electronic coupling between the chromophores and
of the probe pulses was at magic angle relative to that of the indicates that, upon photoexcitation of the triads, the electronic
pump pulses. All spectra were corrected for the chirp of the excitation is essentially localized on either a porphyrin unit or
white-light probe pulses. The FWHM of the IRF was ca. 150 fs. on the cNDI. As a consequence, the excited states of the triads
The sample solutions were placed in a 1 mm thick quartz cell directly populated upon excitation in the Q and B band of a
and were continuously stirred by N2 bubbling. Their absor- porphyrin unit will be thereafter called local B and Q states
bance at the excitation wavelength was between 0.1 and 0.2.
(LBS, LQS), respectively, whereas the state populated upon cNDI
excitation will be designed as LNS, independently of whether
the 430 nm or the B375 nm transitions has been excited.
The probability to excite a porphyrin or the cNDI moiety of
Computational methods
Ground-state gas-phase geometry optimization was performed
at the density functional level of theory (DFT) using the B3LYP the triads at 400 nm can be estimated from the molar decadic
functional45 and a [3s2p1d] basis set.46 Electronic transitions absorption coefficient of the individual units. At 400 nm, the
were computed with time-dependent DFT (TD-DFT) using the absorption coefficient of cNDI amounts to 6200 Mꢁ1 cmꢁ1
same functional and basis set.47 The calculations were carried whereas those of ZnP and FbP are 40 000 and 77 400 Mꢁ1 cmꢁ1
,
,
out using Turbomole version 6.1.48
respectively.49 Thus, the relative contribution of the cNDI to the
absorbance of the triads at 400 nm is only 8% and 4% for 1 and 2,
respectively. As a consequence, direct population of LNS upon
400 nm excitation will be thereafter considered as negligible.
The emission spectra of the triads have been recorded upon
Results
Steady-state spectroscopy
The absorption spectra of the triads together with those of their LBS ’ S0 and LQS ’ S0 excitation in solvents of varying
constituting units in CHX are shown in Fig. 1. The triad spectra polarity, namely CHX, THF and BCN. Emission of 1 could only
consist of an intense band at 417 and 416 nm for 1 and 2, be observed in CHX and its spectrum consists of a broad band
respectively, and a series of weaker bands in the 500–650 nm centered at around 640 nm with a shoulder on both sides (Fig. 2).
region. These spectra are very similar to those of the porphyrin By comparison, the emission spectrum of ZnP exhibits three
units, with the B band at 416 and 415 nm for ZnP and FbP, bands, with the most intense centered at B640 nm as well, that
respectively, and the Q bands at 545 and 583 nm for ZnP and at of medium intensity at 595 nm and the weakest one at around
512, 546, 591 and 649 nm for FbP. The cNDI unit exhibits a 700 nm. Therefore, the shoulders in the spectrum of 1 seem to
broad absorption band centered at around 430 nm, which is coincide with the smaller emission bands of ZnP. Emission from
associated with the core substitution, and a structured band in 1 can thus be assigned to the LQS - S0 transition. Whereas the
the 350–390 nm region localized on the NDI center. The emission spectrum of ZnP contains a distinct S2 - S0 band at
contribution of the cNDI to the absorption spectra of the triads 430 nm, no LBS - S0 emission could be detected with 1.
can be seen as a broadening of the red edge of the B band and
as a small structured band on the blue side of the B band.
Emission of 2 could be observed in all three solvents,
although with different intensity. The spectrum is essentially
The fact that the absorption spectra of the triads are the same as that of FbP, which consists of two bands at 654 and
essentially the composites of those of the constituents points 719 nm (Fig. 2), and originates from the LQS.
c
This journal is the Owner Societies 2013
Phys. Chem. Chem. Phys., 2013, 15, 1177--1187 1179

View Article Online
Paper
PCCP
Because of the extremely short B fluorescence lifetime of
FbP, internal conversion to LQS can also be considered to be the
main deactivation pathway of LBS for 2.
Time-resolved fluorescence
The fluorescence dynamics of 1 and 2 upon LBS ’ S0 excitation
was measured using both TCSPC and fluorescence up-conversion.
Using the latter technique, it was possible to detect LBS
fluorescence of 1 at 430 nm. It was found to be much shorter-
lived than that measured with ZnP, with a decay time of around
200 fs in all three solvents vs. 2.2 ps for ZnP in CHX and THF
and 1.5 ps in BCN (Fig. 3, Table 1). These much shorter time
constants agree with the vanishingly small LBS fluorescence
quantum yield of 1.
The LQS fluorescence lifetime of 1 in CHX measured by
TCSPC amounts to 1.3 ns and is not much smaller than that of
1.8 ns reported for ZnP.5 In THF and BCN, the fluorescence
dynamics measured at 610, 640 and 660 nm was found to
be dominated by a B1 ps decay (Fig. 4). A small residual
intensity with a B2 ns component and B1% relative intensity,
Fig. 2 Fluorescence spectra of the triads and the porphyrin constituents in CHX.
The cNDI does not exhibit any stationary fluorescence. As a
consequence, the absence of any feature other than those of the
porphyrins in the fluorescence spectrum of the triads does not
imply that the LNS is never populated, for example via excita-
tion energy transfer (EET) from the porphyrin LBS.
The fluorescence quantum yield of 1 and 2 determined
relatively to ZnP and FbP, respectively,12,40 depends strongly
on the solvent polarity as illustrated in Table 1. The quantum
yield of both triads in CHX is more than 70% that of the ZnP
and FbP constituents, whereas in THF and BCN, it decreases
down to B3% for 2 and below the detection limit for 1. Such a
difference points to the opening of a new non-radiative deac-
tivation pathway in polar solvents, most probably CS.
The absence of LBS fluorescence with 1 can be due to an
acceleration of the internal conversion to LQS and/or to the
presence of a new non-radiative pathway, namely CS or EET, to
the cNDI center. In principle, the origin of this absence could
be determined by comparing the LQS fluorescence quantum
yield upon LBS and LQS excitation. This was only possible in
CHX, where 1 fluoresces. A decrease in LQS fluorescence
quantum yield of less than 10% was found on going from
LQS (544 nm) to LBS (400 nm) excitation (Table 1), indicating
that internal conversion to LQS is the main deactivation path-
way of the LBS, at least in CHX.
Fig. 3 LBS fluorescence decays at 430 nm measured with triad 1 and ZnP in THF.
Table 1 Fluorescence quantum yields of the triads, Ff, in different solvents
relative to ZnP and FbP, measured upon 400 nm (LBS) or 544 nm (LQS) excitation,
and fluorescence lifetimes measured by TCSPC (in CHX) or up-conversion
(in THF and BCN)
Triad
Solvent
Ff (LBS)
Ff (LQS)
tf (LBS) (ps)
tf (LQS) (ps)
1
1
1
2
2
2
CHX
THF
BCN
CHX
THF
BCN
0.85
—
—
0.73
0.03
0.02
0.93
—
—
0.13
0.23
0.22
1300
1.0
0.9
9200
28
24
Fig. 4 LQS fluorescence decays of the triads in CHX (A) and THF (B) upon LBS
excitation.
c
1180 Phys. Chem. Chem. Phys., 2013, 15, 1177--1187
This journal is the Owner Societies 2013

View Article Online
PCCP
Paper
independent of the wavelength, was also observed. The simi- of these TA spectra was performed using the sum of three
larity of this time constant to that found with ZnP suggests that exponential functions and yielded time constants of 6 ps, 1.3 ns
this residual fluorescence most probably arises from some ZnP and c2 ns (Fig. 5B). The same procedure, performed with the
monomers present as impurity. The rise of the LQS fluorescence TA spectra measured with ZnP, gave time constants of 2.3 ps,
from 1 upon LBS excitation was close to the response function 1.9 ns and c2 ns and the decay associated difference spectra
of the up-conversion setup, in agreement with the ultrafast (DADS) is displayed in Fig. S1 (ESI†). The residual TA spectra of
decay of the LBS fluorescence.
1 and ZnP are very similar and can thus be ascribed to the same
The LQS fluorescence decay of 2 is also strongly solvent intermediate, namely the lowest triplet state of ZnP. The 1.3
dependent. In CHX, it takes place on the nanosecond timescale and 1.9 ns time constants are identical to the LQS and S1
with a time constant of 9.2 ns, whereas in THF and BCN, the fluorescence lifetimes of 1 and ZnP, respectively, and can thus
fluorescence dynamics measured at several wavelengths be interpreted as the lifetime of these states decaying by
between 640 and 710 nm is dominated by more than 90% by fluorescence to the ground state and mainly by intersystem
a decay with 28 and 24 ps time constants, respectively (Fig. 4). crossing to the triplet state. The 2.3 ps time constant obtained
Although, cNDI does not exhibit any significant stationary with ZnP can be assigned to the S2 state lifetime, in agreement
emission, a very short-lived fluorescence with a decay time of with the presence of a negative band at B420 nm in the DADS
around 100 fs could be detected by up-conversion. Transient that is due to the S2 stimulated emission. The LBS lifetime of 1
absorption measurements that will be discussed in a forth- is too short for this state to contribute to the TA spectra.
coming paper point to an ultrafast decay of the emitting state to Consequently, the 6 ps time constant most probably arises
a long-lived dark state, most probably a triplet state. As will be from the vibrational relaxation of the hot LQS upon internal
shown below, the excited states of the cNDI unit do not play a conversion from LBS.
significant role in the photophysics of the triads.
Much faster changes can be observed in the TA spectra
measured with 1 in THF and BCN upon LQS ’ S0 excitation at
555 nm (Fig. 6). The early spectra are essentially the same as
those found in CHX and assigned to LQS absorption and
Transient absorption (TA)
The TA absorption spectra recorded with 1 in CHX (Fig. 5A) are
very similar to those measured with ZnP alone (Fig. S1, ESI†).
They consist of an intense negative band in the 400–430 nm
region, that can be ascribed to the bleach of the LBS ’ S0
transition, and of a broad positive band culminating at around
455 nm and extending to more than 700 nm with an over-
lapping negative feature around 545 nm that originates for the
bleach of the LQS ’ S0 transition. During the time window of
the experiment, i.e. 0–2 ns, the intensity of the band decreases
only partially, especially below 550 nm, and the 455 nm maxi-
mum shifts to B470 nm. Global analysis of the time evolution
Fig. 5 (A) TA spectra recorded with 1 in CHX at different time delays after 400 nm
excitation and (B) decay-associated difference spectra resulting from a global
multiexponential analysis.
Fig. 6 (A–C) TA spectra recorded at different time delays after 555 nm excitation
of triad 1 in THF and (D) species-associated difference spectra obtained from
target analysis.
c
This journal is the Owner Societies 2013
Phys. Chem. Chem. Phys., 2013, 15, 1177--1187 1181

View Article Online
Paper
PCCP
ground-state bleach. During the first 3 ps, the TA intensity in the with the data was obtained with the same three time constants
450–550 nm region decreases, whereas a parallel rise above and with the species associated difference spectra (SADS) shown
650 nm takes place (Fig. 6A). More importantly, a sharp band in Fig. 6D. These time constants and the SADS should not be
appears at 410 nm on the blue side of the LBS ’ S0 bleach. During considered too literally because they have been obtained assum-
the next B5 ps, this sharp feature as well as the TA intensity above ing that each of the three steps in the above scheme occurs from
450 nm decrease and a new narrow band grows around 435 nm on a thermally equilibrated state and follows an exponential
the red side of the LBS ’ S0 bleach (Fig. 6B). During the same time dynamics. This is of course not true as both CS and CR occur
interval, the LBS ’ S0 bleach decreases by more than 60%.
on a similar timescale to vibrational relaxation. Therefore,
From about 8 ps onwards, the shape of the TA spectra non-equilibrium and, thus, non-exponential dynamics can be
remains essentially unchanged, but the TA intensity decreases expected. Consequently, several processes could partially con-
continuously to zero within 50 ps (Fig. 6C). This strong accel- tribute to a given time constant and SADS. Despite this, the
eration of the excited-state dynamics of 1 on going from non- results support our assignments. Indeed, the first SADS is very
polar to polar solvents can be ascribed to the occurrence of CS similar to the TA spectra of 1 in CHX and ascribed to the LQS,
from the locally-excited ZnP unit to the cNDI center. In most whereas the second SADS exhibits the sharp 410 nm band due to
+
studies involving ZnP, occurrence of CS was observed as a rise ZnPꢃ and thus corresponds to the CSS. The radical anions of
in the TA intensity above 650 nm, as found here as well. various cNDIs have been shown to exhibit a main absorption
+
However, a stronger support for the presence of ZnPꢃ upon band centered at around 480 nm as well as some lower intensity
photoexcitation of 1 comes from the 410 nm band, which has features above 600 nm.30,32,34 Finally, the last SADS is dominated
been shown to be the most prominent feature in the electronic by the intense B430 nm band and is thus due to the hot ground
+ 50
absorption spectrum of ZnPꢃ
.
Upon photoexcitation at state population, which relaxes to the equilibrium ground state
400 nm instead of 555, this 410 feature is more difficult to with a 10 ps time constant. Such analysis was also performed on
discern, because of the scattering of the excitation light and the TA data measured upon 400 nm excitation and yielded very
also because the S2 state of ZnP absorbs in this region as well similar results. Comparison of the TA time profiles at 410 and
+
(Fig. S1, ESI†). The disappearance of the ZnPꢃ band observed 650 nm upon 400 and 555 nm excitation shows the same rise
in the 3 to B15 ps time interval is due to the decay of the CSS and decay of the CSS population (Fig. 7). The initial spike in the
population by recombination to the ground state. The band at 410 nm profile upon 400 nm excitation is most probably due to
435 nm that grows as CR takes place can be assigned to the the LBS. Global analysis of the TA data in BCN yielded time
absorption of the vibrationally hot ground state. Indeed, if CR constants very similar to those in THF (Table 2).
occurs on a timescale similar or shorter than that of vibrational
TA measurements with 2 were performed in CHX, THF and
cooling (VC), the instantaneous population of vibrational BCN upon 400 nm excitation only. In CHX, the TA spectra are
excited states of the electronic ground state is large enough to very similar to those measured with FbP alone, and consist of
be observed. The hot ground state is known to be characterized an intense negative band at 420 nm, due to the bleach of the
by a positive TA band located on the red edge of the bleach LBS ’ S0 transition, and a broad positive band going from
bands.51,52 The TA spectra recorded upon 400 nm excitation of
1 in THF are qualitatively similar to those obtained upon
555 nm excitation (Fig. S2, ESI†). The main differences are
the presence of a small feature around 400–410 nm at already
early time that is probably due to LBS absorption, and a more
intense hot ground state absorption band. This is in agreement
with the fact that about 0.9 eV more energy is deposited into the
vibrational modes of ZnP upon 400 nm than 555 nm excitation.
A global analysis of the time evolution of the TA spectra
measured upon 555 nm excitation was first performed using a
sum of exponential functions to determine the number of
relevant time constants. In THF, one time constant was kept
fixed at 1 ps to account for the decay of the LQS as measured by
fluorescence. The data could be well reproduced using three
Fig. 7 Time profiles of the TA intensity at 410 and 650 nm reflecting the CSS
exponential functions with 1, 4.5 and 10 ps time constants. As a
consequence, these data can be equally well analyzed assuming
a kinetic scheme with three consecutive first order processes,
with the same set of time constants.53 Target analysis of the TA
data was thus performed using the following scheme:
population measured upon LBS (400 nm) and LQS (555 nm) excitation.
Table 2 Time constants determined from the global analysis of the TA data
Triad
Solvent
tCS (ps)
tCR (ps)
tVC (ps)
CS
!
CR
ƒ!
VC
ƒ!
L S
Q
CSS
HGS
GS
1
1
2
2
THF
BCN
THF
BCN
1.0
0.9
28
5
8
10
5
10
11
where HGS and GS stand for the hot ground state and the
ground state, respectively. As expected, very good agreement
24
c
This journal is the Owner Societies 2013
1182 Phys. Chem. Chem. Phys., 2013, 15, 1177--1187

View Article Online
PCCP
Paper
440 to more than 700 nm, due to LQS absorption, with over- these two solvents. The fact that the observed decay time of the
lapping holes at 520, 555, 600 and 650 nm arising from the CSS band coincides with the LQS fluorescence lifetime is typical
bleach of the LQS ’ S0 transitions. Apart from minor variations of an inverted kinetics, i.e. a kinetics where the formation of the
during the first 10 ps after excitation, due most probably to intermediate is slower than its decay.57 In such case, the
vibrational relaxation of LQS, the TA spectra remain essentially measured rise time corresponds to the actual decay time of
unchanged within the time window of the experiment, in the intermediate, whereas the measured decay time corre-
agreement with the 9.2 ns fluorescence lifetime.
sponds to the actual time constant of formation of the inter-
Like for 1, the excited-state dynamics of 2 is substantially mediate. To verify whether the CSS population of 2 is indeed
accelerated when going to polar solvents as shown in Fig. 8A governed by such inverted kinetics, target analysis of the TA
and B. The early TA spectra are essentially the same as in CHX data was performed assuming the following scheme:
and can be ascribed to the LQS population and the ground-state
CS
CR
L S
Q
CSS
GS
!
ƒ!
depletion. During the first 7 ps after excitation, a small increase
in the TA intensity in the 450–500 nm region and above
B600 nm can be observed. Afterwards, the shape of the TA
spectrum remains the same but the intensity decreases con-
tinuously on a B100 ps timescale. A residual TA spectrum can
still be observed at the longest time delay. From its intensity
and shape, it most probably arises from free FbP present in
traces. The radical cation FbPꢃ+ has been shown to have a broad
absorption band between 430 nm and 700 nm with an absorp-
tion maximum centered at 436 nm.54–56 Thus, the observed
increase in TA intensity is most probably due to the population
of the CSS. The time evolution of the TA intensity at 436 nm,
and by keeping the CS time constant fixed at 28 and 24 ps in THF
and BCN, respectively. A good agreement with the TA data was
obtained with CR time constants of 10 and 5 ps in THF and BCN
and with the SADS shown in Fig. 8C. The first SADS is essentially
identical to the TA spectra measured in CHX and is assigned to the
LQS population. The second SADS exhibits a narrower LBS ’ S0
bleach, pointing to an overlapping absorption band in this region.
An intense positive band with maxima at B435 and 480 nm, and
extending to more than 700 nm is also present. The narrowing of
+
the bleach and the 435 nm maximum can be ascribed to FbPꢃ ,
whereas the 480 nm feature is most probably due to cNDIꢃꢁ. On
the other hand, the absorption above 550 nm could be due to
both ions. The fact that the contribution of the CSS is barely
visible in the TA data is in full agreement with CR faster than CS.
Because of this, the CSS population does not accumulate enough
to be clearly visible in the TA spectra. Due to the reduced CS time
constant compared to 1, the ground state recovery of 2 is
relatively slow compared to vibrational relaxation, and, conse-
quently, the hot ground state is not visible and does not need to
be accounted for in the target analysis.
+
where FbPꢃ absorbs most can be reproduced using a biexpo-
nential function with a rise time of 10 and 5 ps and a decay time
of 28 and 24 ps in THF and BCN, respectively. Both decay times
are identical to the fluorescence lifetimes of 1 measured in
Discussion
The above results point rather unambiguously to the occur-
rence of photoinduced CS in both triads 1 and 2 in polar
solvents. The existence of a low lying CS state is also supported
by TD-DFT calculations that predict the lowest singlet electro-
nic excited state to be associated with a one-electron transition
from the highest occupied molecular orbital (HOMO) to the
lowest unoccupied molecular orbital (LUMO) with essentially
zero oscillator strength. As illustrated in Fig. 9, the HOMO is
Fig. 8 (A and B) TA spectra recorded at different time delays after 400 nm
excitation of triad 2 in THF and (C) species-associated difference spectra obtained
from target analysis.
Fig. 9 Frontier molecular orbitals of triad 1 computed at the B3LYP/[3s2p1d]
level of theory.
c
This journal is the Owner Societies 2013
Phys. Chem. Chem. Phys., 2013, 15, 1177--1187 1183

View Article Online
Paper
PCCP
doubly degenerate with the electronic density localized almost
exclusively on one ZnP, whereas the LUMO is concentrated on
the central cNDI unit only. Depending on the reaction partner,
porphyrins as well as cNDI can act as both electron donor and
acceptor. According to both the quantum chemical calculations
and the TA data, CS takes place from the excited porphyrin to
the cNDI in both triads.
The CSS energy, ECSS, as well as the free energy of photo-
induced CS, DGCS, have been estimated from the Weller
equation:58
DGCS = ECSS ꢁ E* = e[Eox(D) ꢁ Ered(A)] + C + S ꢁ E*
(1)
where E* is the energy of the excited state from where CS
takes place, Eox(D) and Ered(A) are the oxidation and reduc-
tion potential of the donor and acceptor, respectively, C =
ꢁe2/(4pe0esdDA
) is a correction term that represents the Fig. 10 Energy levels of triad 1 in various solvents.
Coulombic interaction between the charged units at a distance
dDA and screened by a solvent of dielectric constant es. Finally,
S is a factor accounting for a dielectric constant of the solvent
different from that used for the determination of the redox
potentials, ers:
1 is energetically favorable in all three solvents. In principle, the
very short LBS fluorescence lifetime recorded with 1, i.e.
between 100 and 200 fs, compared with the 2.2 ps S2 lifetime
of ZnP could be ascribed to the occurrence of CS from the upper
excited state of 1. However, this hypothesis is not supported by
the other experimental data. For example, the LQS fluorescence
quantum yield of 1 is essentially the same upon 400 and
544 nm excitation (Table 1). As the CSS in CHX is above the
LQS, this could be explained by a CR to LQS and not to the
ground state, as reported previously.20,22 However, if such
sequential CS and CR were operative in CHX, the CSS should
be visible in the transient spectra, in contrast to the observa-
tion. The absence of TA features due to ZnPꢃ+ could in principle
be explained by a CR much faster than CS. Assuming that the
intrinsic LBS fluorescence lifetime of 1 in the absence of
CS amounts to 2.2 ps, the CS time constant would amount to
B150 fs. A CSS totally invisible by TA, even with an IRF of
200 fs, would require a CR time constant of at most 100 fs.
Given that, in polar solvents, CR to the ground state takes place
on the 5–10 ps timescale, such 100 fs CR to LQS is very unlikely.
Furthermore, the time evolution of the CSS population mea-
sured by TA in polar solvents is, within the limit of error, the
same upon LBS and LQS excitation, i.e. exhibits a B1 ps rise
followed by a B5 ps decay (Fig. 7). If the 200 fs LBS fluorescence
lifetime measured with 1 in polar solvents was due to CS, the
time evolution of the TA signal would strongly depend on the
excitation wavelength. We therefore conclude that, in this triad,
CS occurs essentially from the lowest ZnP-localized excited
state. The strong shortening of the LBS lifetime of 1 compared
with ZnP is most probably due to the perturbation of the ZnP
electronic structure introduced by the linkage to the cNDI
moiety. Such a decrease in the S2 lifetime of ZnP upon sub-
stitution has been reported earlier.8,15,59 Because of this, the
lifetime of the LBS is too short for CS to significantly compete
with internal conversion.
ꢀ
ꢁꢀ
ꢁ
e2
1
1
1
1
S ¼ ꢁ
þ
þ
ers es
ð2Þ
8pe0 rD rA
where rD and rA are the radii of the electron donating and
accepting units, respectively. The values of ECSS and DGCS
estimated for 1 and 2 using the excited-state energies and redox
potentials measured with the individual units are listed in
Table 3. Because the electron-donating and accepting units in
1 and 2 are far from spherical as assumed in the above
equations, and the solvent factor S is very sensitive to the rD
and rA values, especially in apolar solvents, the values listed in
Table 3 should only be considered as indicative. Despite this,
they predict rather unambiguously that, upon LQS excitation,
CS should be operative in both triads in THF and BCN, but not
in CHX. This is in perfect agreement with the experimental
data, i.e. the strong shortening of the fluorescence lifetime in
THF and BCN relative to CHX and the presence of TA features
in these two solvents, especially the 410 nm band due to ZnPꢃ
that can be ascribed to the CSS.
The energy level scheme of 1 shown in Fig. 10, together with
the CS driving forces (Table 3), indicates that CS from the LBS of
+
,
Table 3 Free energy of CS and energy of the charge-separated state calculated
from eqn (1)a
LEQ
LEB
Triad Solvent ECSS (eV) E* (eV) DGCS (eV) E* (eV) DGCS (eV)
1
1
1
2
2
2
CHX
THF
BCN
CHX
THF
BCN
2.3
1.6
1.5
2.6
1.9
1.7
2.1
2.1
2.1
1.9
1.9
1.9
0.2
ꢁ0.4
ꢁ0.6
0.6
2.95
2.9
2.9
ꢁ0.6
ꢁ1.3
ꢁ1.4
ꢁ0.05
ꢁ0.2
The relative magnitude of the CS and CR time constants
measured with 1 and 2 can be qualitatively well accounted for by
a
Using Eox(ZnP) = 0.82 V, Eox(FbP) = 1.08 V,37 Ered(NDI) = ꢁ0.65 V,38
rD = 5.5 Å, rA = 4.0 Å and dDA = 11.5 Å.
the free energy associated with these processes, with DGCR = ꢁECSS
.
c
This journal is the Owner Societies 2013
1184 Phys. Chem. Chem. Phys., 2013, 15, 1177--1187

View Article Online
PCCP
Paper
Table 4 Rate constants of Forster excitation energy hopping or transfer, kEEH(T), in
¨
Indeed, the faster CS in 1 compared to 2 can be explained by the
larger CS driving force and agrees with Marcus theory for
electron transfer reactions occurring in the normal regime,
i.e. a regime where the driving force is smaller than the
reorganization energy.60 Although the latter quantity for CS in
these triads cannot be calculated very reliably, it can be
expected to be larger than 0.6 eV in polar solvents.61 Similarly,
the slight acceleration of CS on going from THF to BCN with 2
coincides with a larger driving force in BCN, due to the better
solvent stabilization of the CSS. However, this effect might be
no longer operative for 1, for which CS occurs on a similar
timescale to that of solvent relaxation, the average solvation
time of THF and BCN amounting to 0.94 and 5.1 ps, respec-
tively.62 Both solvents have relaxation components that take
place on a subpicosecond timescale and, therefore, partial
solvation of the CSS might be enough to ensure ultrafast
electron transfer. In such a case however, the effective driving
force is time dependent, and both CS and solvation dynamics
are strongly entangled.63,64
Whereas CS is substantially faster in 1 than in 2, CR is
equally fast in both triads. This similarity is again in good
agreement with the driving force for CR, which does practically
not change when going from 1 to 2. The smaller electron
donating property of FbP relative to ZnP has only a weak effect
on the CR dynamics. However, the CS driving force depends
additionally on the energy of the excited state, E*, which is
lower for FbP than for ZnP, and is thus more favorable for ZnP
than for FbP.
the triads calculated from eqn (3)
Triad Donor
Acceptor |Vdd| (cmꢁ1
)
Y (cm)
1/kEEH(T) (ps)
1
1
1
2
ZnP (LEB) ZnP (LEB) 47
ZnP (LEB) NDI (LEN) 35
ZnP (LEQ) ZnP (LEQ) 3.8
FbP (LEQ) FbP (LEQ) 3.8
5.7 ꢀ 10ꢁ4 0.7
5.4 ꢀ 10ꢁ5 12
5.2 ꢀ 10ꢁ5 1100
3.7 ꢀ 10ꢁ5 1700
for EEH to the other ZnP or EET to cNDI to be very significant in
all solvents investigated. Exchange interaction is also known to
favor EEH(T), especially in bridged systems.65,69 However, as
the porphyrin units are relatively far apart (B23 Å), this
interaction energy should not exceed 10–20 cmꢁ1, and should
thus not lead to major LBS EEH. On the other hand, the
exchange interaction between ZnP and cNDI could be expected
to be larger, due to the smaller distance. Despite this, EET with
a B200 fs time constant would require an unrealistically large
coupling energy of about 300 cmꢁ1. In the same way, because of
the occurrence of CS, the LQS lifetime in polar solvents is too
short for EEH to take place even if exchange interaction is
accounted for. However, LQS EEH is most probably operative in
CHX. The presence of two porphyrin units in these triads leads
only to an increase in the absorption cross section relatively to a
dyad with only one porphyrin. This additional unit should have
a minor effect on the CS and CR dynamics. The situation would
be different if excitation was on the NDI centre. In this case, a
faster CS would be expected as the excited chromophore would
be surrounded by two quenching units.
Excitation energy hopping (EEH) and transfer (EET) are
major excited-state processes in porphyrin arrays.8,65,66 In the
triads, EEH between the porphyrin units as well as EET from
ZnP in the LBS and the cNDI unit could take place in parallel to
Conclusions
CS. In principle, EEH could be evidenced from the measure- The electron transfer dynamics of the two triads investigated
ments of the fluorescence anisotropy of the porphyrin units. here does not differ much from that measured in ZnP–cNDI
However, if the two porphyrins are coplanar, as predicted by the and FbP–cNDI dyads, where a rigid diaza bridge to the NDI core
quantum chemistry calculations, EEH does not lead to a introduced a large coupling between the constituents that was
depolarization of the fluorescence. In fact, the barrier for clearly visible in the absorption spectra.37 Charge separation
torsion of the porphyrins is low,67 and thus some depolariza- was found to be only operative in polar solvents, with time
tion of the fluorescence can be expected upon EEH. The constants of 1.2 and 3.6 ps for the ZnP and FbP dyads,
¨
probability of EEH(T) has been estimated using Forster theory, respectively, and with charge recombination time constants
where the EEH(T) rate constant, in psꢁ1, can be expressed as:68 of B7 and 16 ps. Going from ZnP to FbP leads to a 3 fold
slowing down of charge separation in the dyads compared to a
factor of B25 in the triads investigated here. This difference is
kEEH(T) = 1.18 (psꢁ1 cm)ꢂ|Vdd|2Y
(3)
where Vdd is the interaction energy in cmꢁ1 between the most probably due to the large coupling that results in a partial
transition dipole of the energy donor and that of the acceptor delocalization of the excited state over the whole dyad. This
and Y is the spectral overlap integral, obtained from the area- leads to a partial loss of identity of the individual chromophoric
normalized emission and absorption bands of the donor and units and, as a consequence, the differences between ZnP and
acceptor, respectively, represented on a wavenumber scale. This FbP in these dyads are much less marked than in triads 1 and 2.
expression was used to estimate the LBS and LQS EEH between
On the other hand, the timescale for charge separation in
the ZnP units in 1, the EET between the ZnP unit in the LBS and the dyads and triads is similar, despite the much larger
the cNDI center and finally the LQS EEH between the FbP coupling in the former being most probably due to the driving
moieties in 2. The resulting rate constants together with the force that is around zero. This comparison illustrates how a
Vdd values, calculated as described in the ESI,† are listed in more favorable driving force can balance a smaller electronic
Table 4.
Although these values should not be considered too literally,
coupling and vice versa.
No upper excited-state emission could be detected from the
they indicate that internal conversion from LBS of 1 is too fast ZnP–cNDI dyad, because of the large perturbation of the
c
This journal is the Owner Societies 2013
Phys. Chem. Chem. Phys., 2013, 15, 1177--1187 1185

View Article Online
Paper
PCCP
15 X. Liu, U. Tripathy, S. V. Bhosale, S. J. Langford and R. P. Steer,
J. Phys. Chem. A, 2008, 112, 8986–8998.
16 A. Nakano, Y. Yasuda, T. Yamasaki, S. Akimoto, I. Yamazaki,
H. Miyasaka, A. Itaya, M. Murakami and A. Osuka, J. Phys. Chem.
A, 2001, 105, 4822–4833.
electronic structure of the ZnP unit brought about by the
bridging. Ultrafast internal conversion of the upper excited
state was confirmed by the finding that the charge transfer
dynamics of this dyad was the same upon 400 and 625 nm
excitation. In contrast, emission from the localized B state, with
a lifetime B10 times smaller than that of free ZnP, is observed
with the ZnP triad. However, this shortening does not arise
from the occurrence of charge separation and is most probably
a consequence of the presence of the bridge as well. This result
shows that the observation of a reduced upper excited-state
lifetime is not sufficient for the occurrence of charge separation
from this state to be concluded.
´
17 D. LeGourrierec, M. Andersson, J. Davidsson, E. Mukhtar, L. Sun
¨
and L. Hammarstrom, J. Phys. Chem. A, 1999, 103, 557–559.
18 N. Mataga, H. Chosrowjan, S. Taniguchi and Y. Shibata, J. Phys.
Chem., 2002, 106, 12191–12201.
19 R. T. Hayes, C. J. Walsh and M. R. Wasielewski, J. Phys. Chem. A,
2004, 108, 2375–2381.
20 S. Wallin, C. Monnereau, E. Blart, J.-R. Gankou, F. Odobel and
¨
L. Hammarstrom, J. Phys. Chem. A, 2010, 114, 1709–1721.
¨
21 M. Andersson, J. Davidsson, L. Hammarstrom, J. Korppi-Tommola
and T. Peltola, J. Phys. Chem. B, 1999, 103, 3258–3262.
22 A. Morandeira, L. Engeli and E. Vauthey, J. Phys. Chem. A, 2002, 106,
4833–4837.
23 N. Sakai, J. Mareda, E. Vauthey and S. Matile, Chem. Commun., 2010,
46, 4225–4237.
24 S. V. Bhosale, S. V. Bhosale and S. K. Bhargava, Org. Biomol. Chem.,
2012, 10, 6455–6468.
25 S. R. Greenfield, W. A. Svec, D. Gosztola and M. R. Wasielewski,
J. Am. Chem. Soc., 1996, 118, 6767–6777.
26 K. P. Ghiggino, J. A. Hutchison, S. J. Langford, M. J. Latter, M. A. P.
Lee and M. Takezaki, Aust. J. Chem., 2006, 59, 179–185.
27 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.
28 F. Wu¨rthner, A. Shahadat, C. Thalacker and T. Debaerdemaeker,
Chem.–Eur. J., 2002, 8, 4742.
List of Abbreviations
CS, CSS
CR
charge separation, charge-separated state
charge recombination
D–A
electron donor–acceptor
D(S)ADS decay (species) associated difference spectrum
EEH(T)
LXS
excitation energy hopping (transfer)
locally-excited X state, X = B (porphyrin),
Q (porphyrin), N (NDI)
Acknowledgements
29 S. Bhosale, S. A. L. P. Talukdar, A. Fu¨rstenberg, N. Banerji,
¨
E. Vauthey, G. Bollot, J. Mareda, C. Roger, F. Wu¨rthner, N. Sakai
This work was supported by the Fonds National Suisse de la
Recherche Scientifique through Project Nr. 200020-124393 and
the University of Geneva. S. V. Bhosale gratefully acknowledges
the Australian Research Council for support under Future
Fellowship through their Discovery program and RMIT Uni-
versity for facilities.
and S. Matile, Science, 2006, 313, 84–86.
30 N. Banerji, A. Fu¨rstenberg, S. Bhosale, A. L. Sisson, N. Sakai,
S. Matile and E. Vauthey, J. Phys. Chem. B, 2008, 112, 8912–8922.
31 A. L. Sisson, N. Sakai, N. Banerji, A. Fu¨rstenberg, E. Vauthey and
S. Matile, Angew. Chem., Int. Ed., 2008, 47, 3727–3729.
32 N. Banerji, G. Duvanel, A. Perez-Velasco, S. Maity, N. Sakai, S. Matile
and E. Vauthey, J. Phys. Chem. A, 2009, 113, 8202–8212.
33 R. Bhosale, J. Misek, N. Sakai and S. Matile, Chem. Soc. Rev., 2010,
39, 138–149.
34 N. Sakai, M. Lista, O. Kel, S.-i. Sakurai, D. Emery, J. Mareda,
E. Vauthey and S. Matile, J. Am. Chem. Soc., 2011, 133, 15224–15227.
35 E. Vauthey, ChemPhysChem, 2012, 13, 2001–2011.
Notes and references
1 M. R. Wasielewski, J. Org. Chem., 2006, 71, 5051–5066.
2 D. Gust, T. A. Moore and A. L. Moore, Acc. Chem. Res., 2009, 42,
1890–1898.
3 G. Bottari, G. de la Torre, D. M. Guldi and T. s. Torres, Chem. Rev.,
2010, 110, 6768–6816.
36 F. Chaignon, M. Falkenstrom, S. Karlsson, E. Blart, F. Odobel and
¨
L. Hammarstrom, Chem. Commun., 2007, 64–66.
37 N. Banerji, S. V. Bhosale, I. Petkova, S. J. Langford and E. Vauthey,
Phys. Chem. Chem. Phys., 2011, 13, 1019–1029.
38 S. V. Bhosale, M. B. Kalyankar, S. V. Bhosale, S. J. Langford,
E. F. Reid and C. F. Hogan, New J. Chem., 2009, 33, 2409–2413.
39 H. Z. Liu, J. S. Baskin, B. Steiger, C. Z. Wan, F. C. Anson and
A. H. Zewail, Chem. Phys. Lett., 1998, 293, 1.
40 J. S. Baskin, H.-Z. Yu and A. H. Zewail, J. Phys. Chem., 2002,
106, 9837.
¨
4 M. Borgstrom, N. Shaikh, O. Johansson, M. F. Anderlund, S. Styring,
¨
B. Aakermark, A. Magnuson and L. Hammarstrom, J. Am. Chem.
Soc., 2005, 127, 17504–17515.
5 P. Brodard, S. Matzinger, E. Vauthey, O. Mongin, C. Papamicael and
A. Gossauer, J. Phys. Chem. A, 1999, 103, 5858.
6 E. K. L. Yeow, K. P. Ghiggino, J. N. H. Reek, M. J. Crossley,
A. W. Bosman, A. P. H. J. Schenning and E. W. Meijer, J. Phys. Chem.
B, 2000, 104, 2596–2606.
¨
¨
41 P.-A. Muller, C. Hogemann, X. Allonas, P. Jacques and E. Vauthey,
Chem. Phys. Lett., 2000, 326, 321–327.
42 A. Fu¨rstenberg, M. D. Julliard, T. G. Deligeorgiev, N. I. Gadjev,
A. A. Vassilev and E. Vauthey, J. Am. Chem. Soc., 2006, 128,
7661–7669.
43 G. Duvanel, J. Grilj, H. Chaumeil, P. Jacques and E. Vauthey,
Photochem. Photobiol. Sci., 2010, 9, 908–915.
44 G. Duvanel, J. Grilj, A. Schuwey, A. Gossauer and E. Vauthey,
Photochem. Photobiol. Sci., 2007, 6, 956–963.
45 J. P. Perdew, Phys. Rev. B: Condens. Matter Mater. Phys., 1986, 33,
8822–8824.
7 D. Holten, D. F. Bocian and J. S. Lindsey, Acc. Chem. Res., 2002,
35, 57.
8 A. Morandeira, E. Vauthey, A. Schuwey and A. Gossauer, J. Phys.
Chem. A, 2004, 108, 5741–5751.
9 N. Aratani, D. Kim and A. Osuka, Acc. Chem. Res., 2009, 42,
1922–1934.
10 M. C. O’Sullivan, J. K. Sprafke, D. V. Kondratuk, C. Rinfray, T. D. W.
Claridge, A. Saywell, M. O. Blunt, J. N. O’Shea, P. H. Beton,
M. Malfois and H. L. Anderson, Nature, 2011, 469, 72–75.
11 D. J. Quimby and F. R. Longo, J. Am. Chem. Soc., 1975, 97,
5111–5117.
¨
46 A. Schafer, H. Horn and R. Ahlrichs, J. Chem. Phys., 1992, 97,
2571–2577.
47 R. Bauernschmitt and R. Ahlrichs, Chem. Phys. Lett., 1996, 256,
12 H.-Z. Yu, J. S. Baskin and A. H. Zewail, J. Phys. Chem., 2002,
106, 9845.
13 O. Ohno, Y. Kaizu and H. Kobayashi, J. Chem. Phys., 1985, 82,
1779–1787.
454–464.
¨
¨
48 R. Ahlrichs, M. Bar and M. Haser, Chem. Phys. Lett., 1989, 162,
165–169.
14 G. G. Gurzadyan, T.-H. Tran-Thi and T. Gustavsson, J. Chem. Phys., 49 G. D. Dorough, J. R. Miller and F. M. Huennekens, J. Am. Chem. Soc.,
1998, 108, 385–388.
1951, 73, 4315–4320.
c
This journal is the Owner Societies 2013
1186 Phys. Chem. Chem. Phys., 2013, 15, 1177--1187

View Article Online
PCCP
Paper
50 A. N. Okhrimenko, A. V. Gusev and M. A. J. Rodgers, J. Phys. Chem. A, 60 R. A. Marcus and N. Sutin, Biochim. Biophys. Acta, 1985, 811,
2005, 109, 7653–7656. 265–322.
51 J. Grilj, C. Zonca, L. M. L. Daku and E. Vauthey, Phys. Chem. Chem. 61 N. Mataga, S. Taniguchi, H. Chosrowjan, A. Osuka and K. Kurotoby,
Phys., 2012, 14, 6345–6351. Chem. Phys. Lett., 2005, 403, 163.
52 J. Grilj, P. Buchgraber and E. Vauthey, J. Phys. Chem. A, 2012, 62 M. L. Horng, J. A. Gardecki, A. Papazyan and M. Maroncelli, J. Phys.
7516–7522. Chem., 1995, 99, 17311–17337.
53 I. H. M. van Stokkum, D. S. Larsen and R. van Grondelle, Biochim. 63 H. Sumi and R. A. Marcus, J. Chem. Phys., 1986, 84, 4814–4894.
Biophys. Acta, Bioenerg., 2004, 1657, 82–104. 64 O. Nicolet and E. Vauthey, J. Phys. Chem. A, 2002, 106, 5553–5562.
54 E. A. Aleman, J. Manrıquez Rocha, W. Wongwitwichote, 65 K. Kilså, J. Kajanus, J. Mårtensson and B. Albinsson, J. Phys. Chem.
´
´
´
L. A. Godınez Mora-Tovar and D. A. Modarelli, J. Phys. Chem. A,
B, 1999, 103, 7329–7339.
66 N. Aratani, A. Osuka, H. S. Cho and D. Kim, J. Photochem. Photobiol.,
C, 2002, 3, 25–52.
2011, 115, 6456–6471.
55 Z. Gasyna, W. R. Browett and M. J. Stillman, Inorg. Chem., 1985, 24,
2440–2447.
¨
67 K. Petterson, A. Kyrychenko, E. Ronnow, T. Ljungdahl,
56 C. Paliteiro and A. l. Sobral, Electrochim. Acta, 2005, 50, 2445–2451.
J. Mårtensson and B. Albinsson, J. Phys. Chem. A, 2006, 110, 310.
¨
¨
57 C. Hogemann and E. Vauthey, J. Phys. Chem. A, 1998, 102, 10051–10059. 68 T. Pullerits, S. Hess, J. L. Herek and V. Sundstrom, J. Phys. Chem. B,
58 A. Weller, Z. Phys.–Chem. Materialforsch., 1982, 133, 93. 1997, 101, 10560–10567.
59 H. S. Cho, N. W. Song, Y. H. Kim, S. C. Jeoung, S. Hahn, D. Kim, 69 B. Albinsson and J. Mårtensson, J. Photochem. Photobiol., C, 2008, 9,
S. K. Kim, N. Yoshida and A. Osuka, J. Phys. Chem. A, 2000, 104, 3287.
138–155.
c
This journal is the Owner Societies 2013
Phys. Chem. Chem. Phys., 2013, 15, 1177--1187 1187