Extensive reduction in back electron transfer in twisted intramolecular charge-transfer (TICT) coumarin-dye-sensitized TiO2 nanoparticles/film: A femtosecond transient absorption study

Article abstract of DOI:10.1002/chem.201303903

We report the synthesis, characterization, and optical and electrochemical properties of two structurally similar coumarin dyes (C1 and C2). These dyes have been deployed as sensitizers in TiO₂ nanoparticles and thin films, and the effect of molecular structure on interfacial electron-transfer dynamics has been studied. Steady-state optical absorption, emission, and time-resolved emission studies on both C1 and C2, varying the polarity of the solvent and the solution pH, suggest that both photoexcited dyes exist in a locally excited (LE) state in solvents of low polarity. In highly polar solvents, however, C1 exists in an intramolecular charge-transfer (ICT) state, whereas C2 exists in both ICT and twisted intramolecular charge-transfer (TICT) states, their populations depending on the degree of polarity of the solvent and the pH of the solution. We have employed femtosecond transient absorption spectroscopy to monitor the charge-transfer dynamics in C1- and C2-sensitized TiO₂ nanoparticles and thin films. Electron injection has been confirmed by direct detection of electrons in the conduction band of TiO ₂ nanoparticles and of radical cations of the dyes in the visible and near-IR regions of the transient absorption spectra. Electron injection in both the C1/TiO₂ and C2/TiO₂ systems has been found to be pulse-width limited (<100 fs); however, back-electron-transfer (BET) dynamics has been found to be slower in the C2/TiO₂ system than in the C1/TiO₂ system. The involvement of TICT states in C2 is solely responsible for the higher electron injection yield as well as the slower BET process compared to those in the C1/TiO₂ system. Further pH-dependent experiments on C1- and C2-sensitized TiO₂ thin films have corroborated the participation of the TICT state in the slower BET process in the C2/TiO₂ system.

Full text of DOI:10.1002/chem.201303903

DOI: 10.1002/chem.201303903
Full Paper
&
Electron Transfer
Extensive Reduction in Back Electron Transfer in Twisted
Intramolecular Charge-Transfer (TICT) Coumarin-Dye-Sensitized
TiO2 Nanoparticles/Film: A Femtosecond Transient Absorption
Study
Tushar Debnath,[a] Partha Maity,[a] Hyacintha Lobo,[b] Balvant Singh,[b]
Ganapati S. Shankarling,*[b] and Hirendra N. Ghosh*[a]
Abstract: We report the synthesis, characterization, and opti-
cal and electrochemical properties of two structurally similar
coumarin dyes (C1 and C2). These dyes have been deployed
as sensitizers in TiO2 nanoparticles and thin films, and the
effect of molecular structure on interfacial electron-transfer
dynamics has been studied. Steady-state optical absorption,
emission, and time-resolved emission studies on both C1
and C2, varying the polarity of the solvent and the solution
pH, suggest that both photoexcited dyes exist in a locally
excited (LE) state in solvents of low polarity. In highly polar
solvents, however, C1 exists in an intramolecular charge-
transfer (ICT) state, whereas C2 exists in both ICT and twisted
intramolecular charge-transfer (TICT) states, their populations
depending on the degree of polarity of the solvent and the
pH of the solution. We have employed femtosecond transi-
ent absorption spectroscopy to monitor the charge-transfer
dynamics in C1- and C2-sensitized TiO2 nanoparticles and
thin films. Electron injection has been confirmed by direct
detection of electrons in the conduction band of TiO2 nano-
particles and of radical cations of the dyes in the visible and
near-IR regions of the transient absorption spectra. Electron
injection in both the C1/TiO2 and C2/TiO2 systems has been
found to be pulse-width limited (<100 fs); however, back-
electron-transfer (BET) dynamics has been found to be
slower in the C2/TiO2 system than in the C1/TiO2 system.
The involvement of TICT states in C2 is solely responsible for
the higher electron injection yield as well as the slower BET
process compared to those in the C1/TiO2 system. Further
pH-dependent experiments on C1- and C2-sensitized TiO2
thin films have corroborated the participation of the TICT
state in the slower BET process in the C2/TiO2 system.
Introduction
rected towards gaining a better understanding of the factors
that actually control interfacial electron-transfer dynamics, so
that the efficiency of the devices may ultimately be im-
proved.[3] The most important processes, that are directly rele-
vant for achieving better efficiency of DSSCs, are electron injec-
tion from the photoexcited dye to the conduction band of the
semiconductor and slow back electron transfer from the semi-
conductor to the parent dye cation.[3] In this regard, RuII-based
complexes, such as N3, N719, and black dyes have been at the
forefront, showing photoconversion efficiencies above 11%.
Recently, Grꢀtzel and co-workers[2d] have reported porphyrin-
sensitized solar cells with a cobalt(II/III)-based redox electrolyte,
the efficiencies of which exceeded 12%. Although RuII-based
polypyridyl complexes continue to be investigated, metal-free
organic dyes have also been attracting much research effort
due to their comparatively low cost, high molar absorptivities,
and facile synthetic modification, whereby their spectral range
can be extended through judicious selection of attached func-
tional groups.[4] The efficiencies of DSSCs based on organic
dyes are generally lower than those of cells based on RuII dyes,
although the highest photoconversion efficiency of an organic
dye-sensitized solar cell hitherto reported is above 9%. Among
the organic dye sensitizers tested in DSSCs, coumarin dyes are
Dye-sensitized solar cells (DSSCs) based on organic dyes and
transition metal complexes have attracted great attention
since Grꢀtzel and co-workers reported their seminal work in
1991.[1] Globally, scientists have directed enormous research ef-
forts towards improving the power conversion efficiencies of
DSSCs to make them economically viable.[2] To date, the best
efficiencies obtained have been about 11–12% using the clas-
sical N3 dye and a porphyrin-based sensitizer system. In addi-
tion to testing the efficacies of different sensitizer molecules in
DSSCs, significant effort in this area of research has been di-
[a] T. Debnath, P. Maity, Dr. H. N. Ghosh
Radiation and Photochemistry Division, Bhabha Atomic Research Centre
Mumbai 400 085 (India)
Fax: (+91)22-25505151
E-mail: hnghosh@barc.gov.in
[b] H. Lobo, B. Singh, Dr. G. S. Shankarling
Department of Dyestuff Technology, Institute of Chemical Technology
Mumbai (India)
E-mail: ganu23@gmail.com
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201303903.
Chem. Eur. J. 2014, 20, 3510 – 3519
3510
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
strong candidates because of their good photoelectric conver-
sion properties.[4a–c] One of the main reasons for the high effi-
ciencies of N3, N719, and black dyes is reduced charge recom-
bination in these thiocyanate-substituted RuII dyes, with the
electron density of the HOMO extending over the RuII centre
and NCS ligands.[5] Holes can thereby be localized away from
the TiO2 surface. Ancillary ligands in RuII complexes and metal-
free organic dyes have likewise been used for hole localization
away from a TiO2 surface.[6] We have also demonstrated slow
charge recombination in TiO2 nanoparticles sensitized with
RuII-polypyridyl complexes by introducing an electron-donating
ligand at the RuII centre on which holes can be localized.[7]
Some organic dye molecules possessing donor and acceptor
groups can exhibit dual emission, the second emission being
associated with highly polar excited electronic structure.[8]
Such polar forms usually originate from the transfer of an elec-
tron from a donor group to an acceptor group within the mo-
lecular structure. The charge-separation process within the
molecule is facilitated by a polar environment, with the forma-
tion of a twisted intramolecular charge transfer (TICT) state.
Clearly, if the solvent polarity induces a change in the excited
electronic state leading to a highly polar form through charge
separation within the molecule, an ideal situation for electron
injection into the conduction band of the TiO2 nanoparticle
can be created. Previously, we demonstrated more efficient
electron injection from 7-diethylamino-coumarin 3-carboxylic
acid (D-1421) as compared to coumarin 343 (C-343).[9] These
dyes are structurally similar, except that D-1421 exists in a twist-
ed intramolecular charge-transfer state (TICT) whereas C-343
exists in an intramolecular charge-transfer (ICT) state after pho-
toexcitation in a polar solvent.[9a] Similarly, Sun and co-work-
ers[10] found evidence for the involvement of a TICT state in
a dye-sensitized solar cell and showed improved efficiency in
the case of TICT molecules. In both cases, a conformational
change in the excited state during electron injection seems to
provide an ideal environment for efficient electron injection
and slow charge-recombination kinetics between the injected
electron and the oxidized dye. To understand the effect of the
TICT state on the conversion efficiency in a DSSC, it is very im-
portant to monitor both the electron injection and charge re-
combination dynamics on the ul-
been studied. Both the ground- and excited-state optical prop-
erties suggest that photoexcited C1 and C2 both exist in a lo-
cally excited (LE) state in solvents of low polarity. However, in
highly polar solvents, C1 exists in an intramolecular charge-
transfer (ICT) state, whereas C2 exists in both ICT and twisted
intramolecular charge-transfer (TICT) states, depending on the
polarity of the solvent and the pH of the medium. To elucidate
the effect of the ICT and TICT states on the ET dynamics, we
have carried out femtosecond transient absorption spectrosco-
py on both C1- and C2-sensitized TiO2 nanoparticles and thin
films, and have monitored the transients in the visible and
near-IR regions. Electron injection and back-ET dynamics have
been monitored by direct detection of electrons in the con-
duction band of the nanoparticles and the radical cations of
the dyes. Electron injection and back-ET dynamics have been
compared in the C1/TiO2 and C2/TiO2 systems. To verify the
participation of the TICT state, pH-dependent experiments
have been carried out on C1- and C2-sensitized TiO2 thin films.
Results and Discussion
Solvatochromic behaviour of C1 and C2
The main aim of this investigation has been to ascertain the
effect of the structures of the newly synthesized coumarin
dyes C1 and C2 on the interfacial electron-transfer dynamics.
Hence, it was very important to study their solvatochromic be-
haviour, that is, the solvent-dependent optical properties of
these coumarins, in order to delineate their charge-transfer
properties. We measured the steady-state absorption and
emission properties of C1 and C2 in solvents of different polar-
ity as shown in the Supporting Information, and the related
photophysical data are summarized in Table 1. Interestingly, we
observed that both the absorption and emission bands of C1
and C2 were red-shifted with increasing polarity of the solvent.
Figure 1 (upper panel) shows the steady-state optical absorp-
tion and emission spectra of C1 in cyclohexane (low polarity)
and in water (high polarity). C1 shows optical absorption and
emission peaks at 430 nm (a) and 450 nm (b) in cyclohexane
and at 482 nm (c) and 510 nm (d) in water. The absorption
trafast time scale. However, to
the best of our knowledge, no
Table 1. Absorption (labs) and fluorescence (lem) maxima, Stokes shifts (DnSt), and lifetimes (tf) of C1 and C2 in
such reports have appeared in
the literature except our prelimi-
nary report on the D-1421/TiO2
system.[9b]
different solvents of different polarity.
Dye
C1
Solvent
Dielectric constant
labs [nm]
lem ICT/LE [nm]
DuSt [cmÀ1
]
t ICT/LE [ns]
water
80
482
443
442
436
430
510
498
494
483
1139
2493
2381
2232
1034
DuSt1,
DuSt2
4.19
3.46
3.37
2.94
acetonitrile
ethanol
ethyl acetate
cyclohexane
37.5
24.55
6.02
2.02
To understand the effect of
molecular structure on electron
injection and back electron
transfer in dye-sensitized TiO2
nanoparticles, in the present in-
vestigation we have synthesized
two new structurally similar cou-
marin dyes (C1 and C2). These
have been characterized and
their optical properties have
450
2.5
lem ICT/LE,
lem TICT
[nm]
498, 625
485, 523, 633
485, 555
474, 500
443
tav ICT/LE
tav TICT
[ns]
0.16, 0.09
1.08, 0.38
2.02
[cmÀ1
]
C2
water
80
442
421
420
414
406
2544, 6624
3134, 4632, 7955
3191, 5791
3058, 4155
2057
acetonitrile
ethanol
ethyl acetate
cyclohexane
37.5
24.55
6.02
2.02
1.311
2.4
Chem. Eur. J. 2014, 20, 3510 – 3519
www.chemeurj.org
3511
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
to LE states leads to ICT and TICT states, the populations of
which depend on the freedom of the 7-amino moiety to rotate
in the local environment.[2] A separate emission band from
a TICT state, red-shifted with respect to that of the ICT state,
was first detected by Rettig and co-workers[14] for for p-(N,N-di-
methylamino)benzonitrile (DMABN) due to charge localization
in this donor–acceptor molecule. Thus, in the present investi-
gation, the second red-shifted emission band can be attributed
to TICT emission. From the emission spectra of C1 and C2 re-
corded in different media, it is clear that the second emission
band appears for C2 in highly polar solvents, whereas no such
emission is observed for C1.
To confirm that the second emission band for C2 was due to
TICT emission, we examined the pH dependence of the emis-
sion spectra of C1 and C2 in water. Figure 2 (upper panel)
shows the emission spectra of C1 recorded at different pH of
Figure 1. Upper panel: Absorption (a, c) and emission spectra (b, d) of cou-
marin 1 (C1) in cyclohexane and in water, respectively. Lower panel: Absorp-
tion (e, g) and emission spectra (f, h) of coumarin 2 (C2) in cyclohexane and
in water, respectively.
spectra generally consist of one broad band in the visible
region and a less intense absorption band in the near-UV
region, which correspond to the locally excited (LE) p-p* transi-
tion in a nonpolar solvent (cyclohexane) and the ICT transition
in a highly polar solvent (water). The narrower emission band
in cyclohexane can be attributed to LE emission, whereas the
broader band in water can be attributed to ICT emission.
We also recorded the optical absorption and emission spec-
tra of C2 in cyclohexane and water (Figure 1, lower panel). The
optical absorption spectra of C2 are typical of a coumarin, the
absorption peak appearing at 406 nm in cyclohexane (e) and
being bathochromically shifted to 442 nm in water (g). We
have also monitored the emission behaviour of C2 in different
solvents with increasing polarity, and the results are reported
in the Supporting Information. Figure 1 f, h show the emission
spectra of C2 in cyclohexane and water, respectively. It can
clearly be seen that the emission spectra of C2 and C1 in cy-
clohexane are quite similar, being attributable to the LE state
emission. However, the spectrum of C2 in water features an
emission peak at 498 nm with a broad red-shifted shoulder at
620 nm. The molecular structures of C1 and C2 differ in that
C2 bears a diethylamino group, whereas the nitrogen atom is
incorporated into a ring in C1 (Scheme 1). In our previous in-
vestigation,[9a] we reported that the photophysical properties
of coumarin dyes are very sensitive to rotational hindrance of
7-amino substituents. The rotation of the amine functionality
results in excited-state conformers, that is to say, near-planar
and near-perpendicular twisted conformers.[3] Photoexcitation
Figure 2. Upper panel: Emission spectra of ICT state of C1 at different pH
values (decreasing order: 7–1) in water. Inset: Emission spectra of LE state of
C1 in the a) presence and b) absence of acid in cyclohexane. Lower panel:
Emission spectra of ICT and TICT states of C2 at different pH values (decreas-
ing order: 7–1) in water. Inset: Emission spectra of LE state of C2 in the
a) presence and b) absence of acid in cyclohexane.
the solution. It can clearly be seen that the ICT emission band
intensity of C1 gradually decreases with decreasing pH of the
solution. Interestingly, in the emission spectra of C2, which
consist of both ICT and TICT emission bands, the ICT band de-
creases with decreasing pH whereas the TICT band completely
vanishes at lower pH. It is clear from Scheme 1 that these cou-
marin molecules are donor–acceptor in nature, and upon pho-
toexcitation electrons are localized on the coumarin moiety
and holes are localized at the nitrogen atom. In C2, holes are
localized at the nitrogen atom of the diethylamino group. At
lower pH, due to protonation of the nitrogen atom, hole de-
localization will not take place, and as a result no TICT emission
can be observed. Interestingly, no change in emission intensity
due to the LE state (in cyclohexane) was observed for either
C1 or C2 in acidic solution (inset in Figure 2).
Scheme 1. Molecular structures of coumarin 1 and coumarin 2.
Chem. Eur. J. 2014, 20, 3510 – 3519
www.chemeurj.org
3512
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Time-resolved fluorescence (TCSPC) measurement
1421, which exists in a TICT state in a highly polar solvent such
as water. It is interesting to note that the emission lifetime and
quantum yield of C1 changed only marginally with the polarity
of the solvent, whereas the average emission lifetime and
emission quantum yield of C2 decreased drastically with in-
creasing polarity of the solvent (Table 1). The faster emission
decay for C2 in a highly polar solvent can be ascribed to the
presence of a fast non-radiative decay channel from the LE/ICT
state to the TICT state. However, there is still a minor contribu-
tion from a slower time component of 2–3 ns, which can be at-
tributed to the lifetime of the LE/ICT state. From the above ex-
perimental observations of steady-state absorption and emis-
sion, as well as the time-resolved studies, we can conclude
that both C1 and C2 exist in an LE state in solvents of low po-
larity, but that in a highly polar solvent C1 exists in an ICT state
whereas C2 exists in both ICT and TICT states, their populations
depending on the degree of polarity. We then proceeded to in-
vestigate the effect of molecular structure and the involvement
of the ICT and TICT states in interfacial electron-transfer dynam-
ics on the surface of TiO2 nanoparticles/thin films sensitized
with these coumarin dyes.
To understand the involvement of different charge-transfer
states for the coumarins C1 and C2, we measured their fluores-
cence lifetimes (tf) in solvents of varying polarity, and the re-
sults are summarized in Table 1. Figure 3 (upper panel) shows
Figure 3. Panel A: Time-resolved emission decay traces of coumarin 1 (C1)
a) in cyclohexane at 460 nm and b) in water at 500 nm after excitation at
405 nm. Panel B: Time-resolved emission decay traces of coumarin 2 (C2)
c) in cyclohexane at 460 nm, d) in water at 500 nm, and e) in water at
620 nm after excitation at 405 nm.
Optical absorption and emission studies on TiO2 nanoparti-
cle surfaces
The main aim of the present investigation has been to design
and synthesize new, tailor-made coumarin dyes capable of
serving as more efficient solar sensitizers. To delineate the
effect of molecular structure, which in turn affects solar conver-
sion efficiency, it was imperative to study interfacial electron
transfer (IET) dynamics on the surface of TiO2 nanoparticles.
For this, it was essential to study the optical absorption and
emission behaviour of C1 and C2 on TiO2 nanoparticles.
Figure 4 depicts the optical absorption and emission behaviour
the emission decay traces of C1 in cyclohexane at 460 nm (a)
and in water at 500 nm (b). Emission quantum yields for C1
were determined to be about 40% in cyclohexane and 20% in
water; however, those for C2 drastically reduced from about
50% in cyclohexane to 4.5% in water. The emission decay
traces of C1 could be fitted by single exponentials with time
constants of 2.5 ns in cyclohexane and 4.19 ns in water
(Table 1 and Table S1). The emission decay traces of C2 are de-
picted in Figure 3 (lower panel), as obtained in cyclohexane at
460 nm (c) and in water at 500 nm (d) and at 620 nm (e) after
excitation at 405 nm. It is interesting to note that the emission
decay trace in cyclohexane can be fitted by a single exponen-
tial with a time constant of t=2.4 ns, whereas that in water at
500 nm can be fitted multi-exponentially with time constants
of t1 =0.05 ns (80.8%), t2 =0.22 ns (16.5%), and t3 =3.15 ns
(2.7%), with tav =0.16 ns. That at 620 nm can be fitted multi-
exponentially with time constants of t1 =0.05 ns (87.6%), t2 =
0.18 ns (11.4%), and t3 =2.24 ns (1%), with tav =0.09 ns
(Table 1). It is clear that the majority of the luminescence in
water decays with pulse-width limited time. Rettig and co-
workers have reported that dimethylamino-benzoic acid ethyl
ester (DMABEE) decays non-exponentially with a very short life-
time in highly polar solvents. They attributed this to excitation
of the molecule to an LE (locally excited) or ICT state, followed
by very fast relaxation to the TICT state. Non-radiative decay
rates for this molecule were found to be very high. The transi-
tion from LE/ICT state to TICT was considered to be barrier-
less.[14] Similarly, in our previous investigation,[9a] we also ob-
served fast non-exponential emission decay for coumarin D-
Figure 4. Steady-state absorption spectra of C1 (upper panel) and C2 (lower
panel) in the absence and presence of different TiO2 nanoparticle concentra-
tions a) 0.0, b) 0.039, c) 0.078, d) 0.312, e) 0.625, f) 2.5, and g) 5 gLÀ1 in 35 mm
dye in water. Inset in upper panel: Emission spectra of C1 in the m) absence
and n) presence of 5 gLÀ1 TiO2 nanoparticles. Inset in lower panel: Emission
spectra of C2 in the p) absence and q) presence of 5 gLÀ1 TiO2
nanoparticles.
Chem. Eur. J. 2014, 20, 3510 – 3519
www.chemeurj.org
3513
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
of C1 and C2 free in solution and when adsorbed on a TiO2
nanoparticle surface. The effect of increasing amounts of TiO2
on the absorption spectra of C1 and C2 is shown. On addition
of TiO2 nanoparticles, the optical densities at the absorption
peaks of the respective dyes are seen to increase. The peaks
are also broadened and slightly red-shifted. These observations
strongly suggest that both the dyes interact quite strongly
with the surface of the TiO2 nanoparticles. The equilibrium con-
stants (Keq) for adsorption of C1 and C2 on TiO2 nanoparticles
were determined as 0.6ꢂ106 mÀ1 and 1.5ꢂ106 mÀ1, respectively,
according to the Benesi–Hildebrand (B-H) equation.[15]
The charge-transfer interaction between dye and TiO2 nano-
particles in the excited state was investigated by luminescence
spectroscopy. The emission spectra of C1 and C2 in the ab-
sence and presence of TiO2 nanoparticles are shown in the
inset in Figure 4. It can clearly be seen that in the presence of
TiO2 nanoparticles the emission intensities of both C1 and C2
are drastically reduced. This reduction of emission intensity
can be attributed to electron injection from the photoexcited
coumarin dyes to the conduction band of the TiO2 nanoparti-
cles. It is interesting to note that for C2 free in solution, two
emission bands are seen at 496 and 620 nm, assigned to ICT
and TICT states, respectively, whereas in the presence of TiO2
nanoparticles the ICT emission is greatly diminished and that
due to the TICT state completely vanishes. This suggests that
the TICT state is a better injecting state than the ICT/LE states
of coumarin dyes. Recently, Zietz and co-workers[16] have re-
ported that photoisomerization due to molecular rotation of
the dye D-149 is suppressed on a ZrO2 surface. One can envis-
age similar behaviour (suppression of molecular rotation) for
dye C2 on both ZrO2 and TiO2 surfaces (TiO2 and ZrO2 are
chemically very similar).[17] As a result, the TICT state might not
be observed on TiO2 or ZrO2 surfaces. In fact, however, we
clearly observed the TICT state on a ZrO2 surface (Supporting
Information). On this basis, we expected that the TICT state
might also be seen on a TiO2 surface. However, due to efficient
electron injection from the TICT state, no emission band due
to a TICT state on the TiO2 surface was observed. It is very diffi-
cult to elucidate the effect of TICT on IET dynamics from
steady-state absorption and emission spectra. To comprehend
the effect of TICT on IET dynamics, we carried out ultrafast
transient absorption measurements on the C1/TiO2 and C2/
TiO2 systems, exciting the samples with 400 nm laser light, and
compared the dynamics (see below).
Figure 5. Transient absorption spectra of the C1/TiO2 (upper panel) and C2/
TiO2 (lower panel) systems in water at different time delays after exciting the
samples with 400 nm laser light.
a peak at around 650 nm, and a broad positive feature in the
whole spectral region (750–900 nm). The bleach signal can be
attributed to the degradation of coumarin molecules adsorbed
on the TiO2 surface due to photoexcitation. The band in the
region 580–700 nm can be attributed to the radical cations of
the coumarin dyes (both C1 and C2). This band has been as-
signed on the basis of results obtained from a complementary
pulse radiolysis experiment, in which the radical cations of the
coumarin dyes were selectively generated by reactions of N3
radicals with C1 and C2 in N2O-saturated aqueous solution
(5% acetonitrile + 95% water), as shown in the Supporting In-
formation. The broad absorption band in the region 700–
900 nm has been attributed to conduction-band electrons in
the nanoparticles.[3b–f] It has previously been reported by our-
selves and many other research groups that conduction-band
electrons can be detected on the basis of characteristic visi-
ble[3b–f] and mid-IR bands.[3a] Injection times in the present in-
vestigation were determined by monitoring the appearance
times of the signals of the radical cations of 1 and 2 (1·+ and
2·+) at 650 nm and of the injected electron at 905 nm, and
were found to be pulse-width limited (<120 fs) in both the
C1/TiO2 and C2/TiO2 systems. To determine the back electron
transfer (BET) in both systems, we monitored the transient sig-
nals at 650 nm (due to the radical cations) and 905 nm (due to
injected electrons in TiO2), and the results are shown in
Figure 6. The kinetic decay trace at 650 nm was fitted multi-ex-
ponentially for the C1/TiO2 system, with time constants of
0.8 ps (71.2%), 5 ps (11%), 30 ps (7.6%), and >400 ps (10.2%)
(Table 2). The kinetic decay trace at 650 nm for the C2/TiO2
system was similarly fitted with time constants of 1 ps (43.8%),
5 ps (12.7%), 30 ps (10%), and >400 ps (33.5%) (Table 2). We
also fitted the kinetic decay traces multi-exponentially for the
injected electrons at 905 nm for both the C1/TiO2 and C2/TiO2
systems, as shown in Table 2. It is interesting to note that the
kinetic traces for both radical cations and injected electrons
decay faster for the C1/TiO2 system than for the C2/TiO2
Ultrafast transient absorption measurements on the C1/TiO2
and C2/TiO2 systems
To study the effects of molecular structure and the TICT states
of coumarin dyes on interfacial electron-transfer dynamics, we
carried out transient absorption measurements on both C1-
and C2-sensitized TiO2 nanoparticles excited with 400 nm fem-
tosecond laser pulses. Figure 5 shows the time-resolved transi-
ent absorption spectra of the C1/TiO2 (upper panel) and C2/
TiO2 (lower panel) systems in water. For both systems, the
spectrum at each time delay consists of bleach below 570 nm,
a positive absorption band in the region 580–700 nm with
Chem. Eur. J. 2014, 20, 3510 – 3519
www.chemeurj.org
3514
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 7. Panel A: Transient absorption decay kinetics at 650 nm for a) C1/
TiO2 thin film at pH 7 and b) C1/TiO2 thin film at pH 2. Panel B: Transient ab-
sorption decay kinetics at 650 nm for c) C2/TiO2 thin film at pH 7 and d) C2/
TiO2 thin film at pH 2. The samples were excited with 400 nm laser light.
Figure 6. Panel A: Transient absorption decay kinetics at 650 nm for a) C1/
TiO2 and b) C2/TiO2. Panel B: Transient absorption decay kinetics at 905 nm
for c) C1/TiO2 and d) C2/TiO2 after exciting the samples with 400 nm laser
light.
Figure 7 shows the normalized transient absorption kinetics at
650 nm at pH 7 and pH 2 for C1/TiO2 (Panel A) and C2/TiO2
films (Panel B). The decay traces were each fitted multi-expo-
nentially, as shown in Table 2. It was observed that the decay
dynamics, indicating BET dynamics, was slower in thin films
than on the surface of nanoparticles (Table 2). The BET process
in dye–nanoparticle systems depends on electronic coupling
between the injected electron and the parent cation. In the
case of a nanoparticle surface for the same dye–nanoparticle
pair, coupling will be much stronger due to the proximity of
the electron and the radical cation as compared to that in
a film. In the case of a film, the nanoparticles are interconnect-
ed, and hence immediately after injection the electrons can
diffuse to another particle within the film. As a result, the effec-
tive electronic coupling between the injected electron and the
parent cation (dye·+) becomes weak for the BET process on
the film surface. Consequently, a slower back-electron-transfer
process is observed on the film surface as compared to that
on the nanoparticles for both the C1/TiO2 and C2/TiO2 systems.
It is interesting to note that the effect of pH on the BET dy-
namics in a C1/TiO2 film is negligible; however, in a C2/TiO2
film, we have observed that the signal intensity of the transi-
ent absorption decreases, at the same transient signal decay,
faster at lower pH (pH 2). The kinetic decay trace at 650 nm for
the C2/TiO2 film at pH 7 could be fitted multi-exponentially
with time constants of 1 ps (49.4%), 5 ps (6.1%), 30 ps (6.2%),
and >400 ps (38.3%). However, at pH 2, the kinetic decay
trace could be fitted with time constants of 0.5 ps (77.2%),
3 ps (9.7%), 20 ps (6.95%), and >400 ps (6.1%). This observa-
tion clearly suggests the involvement of TICT states in the in-
terfacial ET dynamics in these dye–semiconductor nanoparticle
systems. We have already mentioned that at lower pH the C2
molecule exists only in an ICT state, whereas at pH 7 it exists in
a TICT state, which is primarily a charge-separated state.
Hence, the recombination (BET) dynamics is expected to be
slower in the TICT state as compared to that in the ICT state.
Table 2. Parameters for the multiexponential fits of the back-electron-
transfer (BET) kinetics for C1/TiO2 and C2/TiO2 nanoparticles and thin
films, after monitoring the coumarin radical cations at 650 nm and elec-
trons in the conduction band at 905 nm, under different conditions after
excitation of the samples with 400 nm laser light.
System/
t1 [ps]
t2 [ps]
t3 [ps]
t4 [ps]
probe wavelength [nm]
([%])
([%])
([%])
([%])
C1/TiO2/650
C1/TiO2/905
C2/TiO2/650
C2/TiO2/905
C1/TiO2 film/650
C1/TiO2 film (pH 2)/650
C2/TiO2 film/650
C2/TiO2 film (pH 2)/650
0.8 (71.2) 5 (11)
0.8 (70.7) 5 (5.7)
30 (7.6)
30 (10)
>400 (10.2)
>400 (13.6)
>400 (33.5)
>400 (36.9)
>400(17.8)
>400 (11.6)
>400 (38.3)
>400 (6.1)
1 (43.8)
1 (53.5)
1 (48.8)
1 (55.2)
1 (49.4)
5 (12.7) 30 (10)
5 (5.5) 30 (4.1)
5 (18.4) 30 (15.0)
5 (18)
5 (6.1)
30 (15.2)
30 (6.2)
20 (6.95)
0.5 (77.2) 3 (9.7)
system. This observation clearly suggests that the molecular
structure affects the BET dynamics, with photoexcited C2 exist-
ing in a TICT state and C1 existing in an ICT state in a highly
polar solvent.
To confirm the involvement of TICT states in the interfacial
ET dynamics of coumarin dye-sensitized TiO2 nanoparticles, we
carried out experiments on TiO2 thin films similarly sensitized
with dyes C1 and C2. So far, we have discussed the interfacial
electron-transfer dynamics of C1 and C2 on the surface of
nanoparticulate TiO2. However, a dye-sensitized solar cell is
constructed from a TiO2 thin film. Hence, it was considered
very important to study the IET dynamics on a TiO2 thin film in
order to assess the utility of the sensitizer molecules in real de-
vices. In the previous section, we noted that the existence of
TICT states is strongly affected by lowering the pH of the
medium. To assess the effect of pH on the IET dynamics in the
above systems, we carried out ultrafast transient studies on
TiO2 films sensitized with C1 and C2 at pH 7 and pH 2.
Chem. Eur. J. 2014, 20, 3510 – 3519
www.chemeurj.org
3515
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Charge-transfer dynamics between ICT and TICT states and
TiO2 nanoparticles
clearly be seen in Scheme 2 that the ICT state of C1 lies above
the conduction band edge, hence the majority of the photoex-
cited state of C1 can inject electrons (kinj), with a certain frac-
tion emitted as luminescence. In the case of the C2/TiO2
system, however, upon excitation with laser pulses, the dye
molecule attains the ICT state. From the ICT state, it can inject
an electron into the conduction band of TiO2 and relaxes very
rapidly to the TICT state (kBA). Since the TICT state is a charge-
separated state, in which an electron is localized on the cou-
marin moiety and a hole is localized on the nitrogen atom of
the diethylamino group, the electron injection (kinj’) is highly
facile. It is for this reason that we observed higher electron in-
jection yield in the C2/TiO2 system.
Finally, we consider the charge recombination reaction (BET)
for both systems. Although the coupling between dye and
TiO2 nanoparticle is the same for both C1 and C2, the free
energy of the BET process for the C2/TiO2 system is higher
(ÀDGC2/TiO2 =1.92 eV) than that for the C1/TiO2 system
(ÀDGC1/TiO2 =1.83 eV). As a result, the BET process for the C2/
TiO2 system is slower as it falls in the Marcus inverted region of
electron-transfer reaction. The difference in BET process can
also be envisaged in Scheme 3, which shows the orbital
schemes for BET for both the ICT state of C1 and the TICT state
of C2 with TiO2 nanoparticles. Prior to electron injection in the
C1/TiO2 system, holes are localized in the orbital of the ICT
state and electrons are localized in TiO2 and eventually BET can
take place (Scheme 3). On the other hand, in the C2/TiO2
system, holes are localized on the nitrogen atom of the diethy-
lamino group of the TICT state, in which the spatial charge sep-
aration is much larger than that in the ICT state of C1. As
a result, we observed much slower charge recombination in
the C2/TiO2 system than in the C1/TiO2 system. The participa-
From the above-mentioned steady-state optical absorption
and emission studies, as well as time-resolved emission studies,
changing the polarity of the solvents and the pH of the media,
we concluded that the photoexcited C1 molecule exists in an
ICT state, whereas the C2 molecule exists in both ICT and TICT
states in highly polar solvents. Ultrafast transient absorption
studies on C1/TiO2 and C2/TiO2 indicated that electron injec-
tion in both systems is pulse-width limited, and a slower back-
electron-transfer rate has been observed in the C2/TiO2 system
as compared to the C1/TiO2 system. The results clearly indicate
that the TICT state of C2 facilitates a higher electron injection
yield and a slower back-electron-transfer process. In a dye-sen-
sitized electron-transfer reaction, the excited molecular state of
the dye molecule overlaps with high density states of the con-
duction band of the TiO2 nanoparticles, which means that all
of the vibrational and rotational states of the excited electronic
state of the dye molecule overlap with the conduction band.
However, it has been reported in the literature[18] that for mod-
erate coupling (with carboxylic binding), electron injection and
vibrational relaxation are competing processes in photoexcited
dye/TiO2 nanoparticle systems. This model suggests that a por-
tion of the excited dye molecules will relax to the ground vi-
brational state of the first excited electronic state, from which
nonradiative and radiative (emission) processes are possible. In
the present investigation, we have also observed that a certain
percentage of photoexcited states do not inject an electron
(inset in Figure 4), with non-injecting states relaxing through
nonradiative and radiative transitions. To illustrate the mecha-
nism of electron transfer from different states (ICT and TICT) of
the coumarin dyes, we have
drawn Scheme 2, which shows
the energetics of the ground
and excited states, that is, the
energetic positions of the ICT
and TICT states of both C1 and
C2, and the conduction band
edge of the TiO2 nanoparticles.
The ground state (GS) oxidation
potentials for C1 and C2 were
measured by cyclic voltammetry
against an Ag/AgCl electrode.
The redox reactions were seen
to be reversible. The excited-
state oxidation potential of the
dyes was measured by adding
the E00 and GS oxidation poten-
tials. E00 was measured from the
crossing point of the excitation
and emission spectra of the dye
molecules. The approximate
Scheme 2. Schematic diagram of electron transfer from electronically excited dye to the conduction band of TiO2
nanoparticles for the coumarin dyes (C1 and C2). S0 is the ground electronic level of the coumarin dyes, S1 (ICT) is
the intramolecular charge-transfer state, and S1 (TICT) is the twisted intramolecular charge-transfer state. kBA indi-
cates the relaxation rate from the ICT state to the TICT state and kBET is the back-electron-transfer rate from TiO2 to
energy level of the TICT state of
C2 was determined from the
energy shift from the ICT state to
the TICT state in Figure 2. It can the S0 level of the coumarin dyes.
Chem. Eur. J. 2014, 20, 3510 – 3519
www.chemeurj.org 3516
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
hoped that our investigation
may expedite the design and
synthesis of more efficient or-
ganic solar-harvesting dyes that
might significantly improve the
efficiencies of DSSCs.
Experimental Section
Materials
Scheme 3. Orbital scheme showing back-ET processes in the ICT state of the C1/TiO2 system and the TICT state of
the C2/TiO2 system.
Titanium(IV) tetraisopropoxide (Ti-
[OCH(CH3)2]4; Aldrich, 97%) and
isopropyl alcohol (Aldrich) were
purified by distillation. Nanopure
tion of the TICT state was further confirmed by pH-dependent
water was used to prepare aqueous solutions. All solvents and
chemicals used for synthesis were procured from SD Fine Chemi-
cals (India) and were used without further purification. Chemical
reactions were monitored by TLC on 0.25 mm Merck silica gel 60
F254 precoated plates; spots were visualized under UV light.
1H NMR spectra were recorded on a 300 MHz Varian Mercury Plus
spectrometer. Chemical shifts are expressed in d (ppm) with refer-
ence to TMS as an internal standard. Mass spectral data were ob-
tained on Micromass-Q-TOF (YA105) spectrometer.
experiments on both systems on TiO2 thin films. It can clearly
be seen in Figure 7 (Panel A) that the BET dynamics was un-
changed in the C1/TiO2 system at pH 7 and pH 2. On the other
hand, in the C2/TiO2 system, the BET dynamics (Figure 7,
Panel B) was much faster at pH 2, at which the TICT states are
considered to be absent, on the basis of the results presented
herein. From the above experimental observations, we can
conclude that TICT states facilitate higher electron injection
yields and at the same time retard the back-electron-transfer
process at the dye–semiconductor nanoparticle surface.
Preparation of TiO2 nanoparticles
Nanometer-size TiO2 was prepared by controlled hydrolysis of tita-
nium(IV) tetraisopropoxide.[11] A solution of Ti[OCH(CH3)2]4 (5 mL) in
isopropyl alcohol (95 mL) was added dropwise (1 mLminÀ1) to
nanopure water (900 mL) at 28C and pH 1.5 (adjusted with HNO3).
The solution was continuously stirred for 10–12 h until a transpar-
ent colloid was formed. The colloidal solution was concentrated at
35–408C in a rotary evaporator, and the residue was dried with a ni-
trogen stream to yield a white powder. In the present work, all col-
loidal samples were prepared after dispersing dry TiO2 nanoparti-
cles in water (15gLÀ1). The coumarin dyes were added to the TiO2
colloid and sonicated for some time (2–3 min) to prepare the sen-
sitized nanoparticles.
Conclusion
To determine the effect of molecular structure on interfacial
electron dynamics in dye-sensitized TiO2, we have synthesized
and characterized two structurally similar coumarin dyes (C1
and C2). These have been used to sensitize TiO2 nanoparticles
and thin films. Steady-state optical absorption, emission, and
time-resolved emission studies on these dyes, changing the
polarity of the solvent, suggest that photoexcited C1 and C2
both exist in a locally excited (LE) state in solvents of low po-
larity. In highly polar solvents, however, C1 exists in an intra-
molecular charge-transfer (ICT) state, whereas C2 exists in both
ICT and twisted intramolecular charge-transfer (TICT) states,
their populations depending on the degree of polarity of the
solvent. The existence of the TICT state has been further sup-
ported by changing the pH of the medium. Electron-transfer
dynamics (both electron injection and back ET) has been deter-
mined by directly monitoring the radical cation and injected
electrons in the conduction band of TiO2 nanoparticles/films in
the visible and near-IR regions by femtosecond transient ab-
sorption spectroscopy. Pulse-width limited (<120 fs) electron
injection was observed in both the C1/TiO2 and C2/TiO2 sys-
tems; however, back ET dynamics was found to be slower in
the latter than in the former. Higher free energy of reaction
(ÀDG) in the C2/TiO2 system is one of the reasons for a slow
BET process. However, our results clearly indicate that involve-
ment of the TICT state in C2 is the main reason for the slower
BET process, which leads to higher electron yields. Ultrafast
transient absorption measurements performed at different pH
of the TiO2 thin film have clearly demonstrated the participa-
tion of the TICT state of C2 in the slower BET process. It is
Preparation and sensitization of the thin films
TiO2 nanocrystalline thin films were prepared by a method similar
to that used by Zaban and co-workers.[12] Briefly, a colloid of TiO2
nanoparticles was prepared by the controlled hydrolysis of titan-
ium(IV) isopropoxide in a mixture of glacial acetic acid and water
at 08C. It was heated at 808C for 8 h and then autoclaved at
2308C for 12 h. The resulting colloid was concentrated to 150 gLÀ1
and spread onto polished sapphire windows, which were baked at
4008C for 36 min. The TiO2 films were sensitized by first immersing
them in a solution of 200 mm coumarin dye and 20 mm cheno-
deoxycholic acid in ethanol at room temperature, whereby the dye
was adsorbed on the porous film surface. All experiments were
performed with thin films exposed to air under ambient condi-
tions. Film samples were moved rapidly during measurement to
avoid long-term photo-product build-up.
Synthesis of coumarin dyes
Synthesis of coumarin 1 sensitizer: A 100 mL three-necked
round-bottomed flask fitted with a mercury-sealed stirrer was
charged with a mixture of 3-chloro-3-(11-oxo-2,3,5,6,7,11-hexahy-
dro-1H-pyrano[2,3-f]pyrido[3,2,1-ij]quinolin-10-yl)acrylaldehyde [8,
Chem. Eur. J. 2014, 20, 3510 – 3519
www.chemeurj.org
3517
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Supporting Information] (1.5 g, 4.5 mmol) and cyanoacetic acid [9,
Supporting Information] (0.60 g, 6.8 mmol) in absolute ethanol
(15 mL) containing piperidine (0.5 mL). The stirred reaction mixture
was heated under reflux for 4 h. The progress of the reaction was
monitored by TLC, and after its completion the mixture was
poured into cold water and the product was extracted with ethyl
acetate. The ethyl acetate layer was washed with water and then
concentrated under vacuum in a rotary evaporator to obtain the
product, coumarin 1. The crude product was further purified by
column chromatography on silica gel, eluting with toluene/ethyl
acetate (6:4). Yield: 0.97 g (54%); 1H NMR (CDCl3, 300 MHz): d=
8.30 (s, 1H; aromatic CH), 8.18–8.09 (m, 2H; vinylic CH), 7.20 (s,
1H; aromatic CH), 2.75–2.60 (m, 4H; aliphatic CH2), 2.50–2.40 (m,
4H; aliphatic CH2), 1.90–1.75 ppm (m, 4H; aliphatic CH2); 13C NMR
(CDCl3, 75 MHz): d=163.2, 158.1, 151.1, 149.2, 148.7, 145.3, 142.5,
127.4, 119.9, 115.4, 109.8, 108.2, 104.5, 49.7, 49.2, 26.7, 20.4,
19.4 ppm; MS: m/z: 297, 269, 261 (corresponds to ion peak after
characteristic fragmentation).
18.4 W, 1 KHz). The amplified pulse at the power saturation point
of the regenerative cavity was extracted for second stage amplifi-
cation in a two-pass Ti:sapphire amplifier pumped by around 65%
of the output power of the Nd:YLF laser. Finally, after passage
through a compressor encompassing a two-grating assembly, the
amplified pulses of around 40 fs FWHM, 800 nm at 1 KHz repetition
rate were obtained. For the pump-probe set-up, 30% of the
output power of this amplifier system was split into two parts to
generate pump and probe pulses. Frequency-doubled 400 nm
pump pulses were generated using 200 mJ/pulses of 800 nm in
a barium b borate BBO crystal. Probe pulses in the region 480–
1000 nm were generated by focusing around 3 mJ of the 800 nm
beam onto a 1.5 mm thick sapphire window. The noise of the
white light continuum was controlled by adjusting the 800 nm
beam intensity through an iris and neutral density (ND) filter
placed before the sapphire window. The probe pulses were split
into signal and reference beams before the quartz cell (sample),
and were detected by two matched photodiodes with variable
gain. The pump and probe beams were focused on the same posi-
tion of the cell at the crossing point around 500 and 300 mm. The
excitation energy density (400 nm) was adjusted to around
2500 mJcmÀ2. The noise level of the white light was about 0.5%
with occasional spikes due to oscillator fluctuation. We have no-
ticed that most laser noise is low-frequency noise and can be elimi-
nated by comparing adjacent probe laser pulses (pump blocked vs
unblocked using a mechanical chopper). The typical noise in the
measured absorbance change was <0.3%. The instrument re-
sponse function (IRF) for 400 nm excitation was obtained by fitting
the rise time of the bleaching of the sodium salt of meso-tetra-
kis(4-sulfonatophenyl)porphyrin (TPPS) at 710 nm, and was found
to be 100 fs.[13] Kinetic decay traces were fitted using the Lab-view
program based on differential equations. The kinetic traces were
freely fitted, with both amplitudes and time constants being close-
ly coupled. The deviation or error bar in both lifetime and ampli-
tude measurements was no higher than Æ2%.
Synthesis of coumarin 2 sensitizer: A three-necked 100 mL
round-bottomed flask was charged with 3-chloro-3-[7-(diethylami-
no)-2-oxo-2H-chromen-3-yl]acrylaldehyde [12, Supporting Informa-
tion] (1 g, 3.2 mmol) in absolute ethanol (10 mL). Cyanoacetic acid
[9, Supporting Information] (0.55 g, 6.5 mmol) and piperidine (3–
4 drops) were then added and the reaction mixture was vigorously
stirred at reflux temperature for 4 h. The progress of the reaction
was monitored by TLC, and after its completion the mixture was
poured into cold water and the product was extracted with ethyl
acetate. The ethyl acetate layer was washed with water and then
concentrated under vacuum in a rotary evaporator to obtain the
product, coumarin 2. The crude product was further purified by
column chromatography on silica gel, eluting with toluene/ethyl
1
acetate (6:4). Yield: 1.0 g (89%); H NMR (CDCl3, 300 MHz): d=8.44
(s, 1H; aromatic CH), 8.18–8.15 (m, 1H; aromatic CH), 8.04–8.01 (d,
1H; vinylic CH), 7.64–7.62 (d, 1H; aromatic CH), 6.79–6.76 (d, 1H;
vinylic CH), 6.59 (s, 1H; aromatic CH2), 3.63–3.44 (m, 4H; aliphatic
CH2), 1.15–1.11 ppm (t, 6H; aliphatic CH3); 13C NMR (CDCl3,
75 MHz): d=172.0, 163.0, 158.2, 155.9, 152.2, 144.5, 144.4, 137.4,
131.3, 122.3, 117.5, 112.4, 110.1, 108.0, 95.7, 62.7, 44.3, 12.3 ppm;
MS: m/z: 373 [M+], 337 [MÀCl]+. The synthetic procedures for
other intermediates are described in detail in the Supporting Infor-
mation.
Cyclic voltammetry
Voltammetric experiments were performed with an Autolab
PGSTAT 20 (manufactured by Eco-Chemie, The Netherlands) cou-
pled to a Metrohm 663 VA standard electrode system comprising
glassy carbon (GC)/Pt/Ag/AgCl. The PG STAT was driven by Autolab
software. The temperature of the solution was maintained at 25Æ
0.18C. Measurements were made in solutions of tetrabutylammoni-
um hexafluorophosphate (TBAP) in acetonitrile as supporting elec-
trolyte under N2 atmosphere. The redox potentials of C1 and C2
were determined as 1.11 V and 1.19 V, respectively, against the Ag/
AgCl electrode.
Time-resolved emission spectrometer
Time-resolved fluorescence measurements were carried out using
a diode-laser-based spectrofluorimeter from IBH (UK). The instru-
ment works on the principle of time-correlated single-photon
counting (TCSPC). In the present work, 405 and 445 nm laser
pulses were used as excitation light sources, and a TBX4 detection
module (IBH) coupled with a Hamamatsu PMT was used for fluo-
rescence detection.
Acknowledgements
We thank Dr. Shilpa Tawade, Chemistry Division, BARC, and Dr.
A. Kumbhar of Pune University for their assistance with cyclic
voltammetric and spectroelectrochemistry measurements, re-
spectively. T.D. acknowledges the CSIR and P.M. acknowledges
the DAE for research fellowships. We sincerely thank Prof. D. K.
Palit and Dr. B. N. Jagatap for their encouragement.
Femtosecond visible spectrometer
A combined regenerative/multipass amplifier (1 kHz repetition rate
at around 800 nm, 50 fs, 1 mJ/pulse) from Thales (Alpha 1000) was
used for 400 nm pump/480–1000 nm probe measurements. The
output of a mode-locked Ti:sapphire laser (FEMTOSOURCE Synergy
20) pumped by a CW-Nd:YVO4 laser (Finesse, Laser Quantum,
532 nm, output power: 3.8 W) was used as a seed pulse train
(102 MHz). The 20 fs pulses from the oscillator were stretched
using an Offner stretcher (XS-type, supplied by Thales) and pream-
plified in a Ti:sapphire regenerative cavity pumped by around 35%
of the output power of an Nd:YLF laser (Jade–Thales, 532 nm,
Keywords: back electron transfer · coumarin dyes · electron
injection · solar cells · titania · ultrafast spectroscopy
Chem. Eur. J. 2014, 20, 3510 – 3519
www.chemeurj.org
3518
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[1] B. O’Regan, M. Grꢀtzel, Nature 1991, 353, 737–740.
J. Mater. Chem. 2008, 18, 4246–4253; c) K. Chen, Y.-H. Hong, Y. Chi, W.-
H. Liu, B.-S. Chen, P.-T. Chou, J. Mater. Chem. 2009, 19, 5329–5335; d) J.-
F. Yin, D. Bhattacharya, Y.-C. Hsu, C.-C. Tsai, K.-L. Lu, H.-C. Lin, J.-G. Chen,
K.-C. Ho, J. Mater. Chem. 2009, 19, 7036–7042.
[2] a) J. H. Delcamp, A. Yella, T. W. Holcombe, M. K. Nazeeruddin, M. Grꢀtzel,
Angew. Chem. Int. Ed. 2013, 52, 376–380; b) Q. Yu, Y. Wang, Z. Yi, N. Zu,
J. Zhang, M. Zhang, P. Wang, ACS Nano 2010, 4, 6032–6038; c) C.-Y.
Chen, M. Wang, J.-Y. Li, N. Pootrakulchote, L. Alibabaei, C. Ngoc-le, J.-D.
Decoppet, J. H. Tsai, C. Grꢀtzel, C.-G. Wu, S. M. Zakeeruddin, M. Grꢀtzel,
ACS Nano 2009, 3, 3103–3109; d) A. Yella, H.-W. Lee, H. N. Tsao, C. Yi,
A. K. Chandiran, M. K. Nazeeruddin, E. W.-G. Diau, C.-Y. Yeh, S. M. Zakeer-
uddin, M. Grꢀtzel, Science 2011, 334, 629–634; e) Y. Bai, J. Zhang, D.
Zhou, Y. Wang, M. Zhang, P. Wang, J. Am. Chem. Soc. 2011, 133, 11442–
11445; f) W. Zeng, Y. Cao, Y. Bai, Y. Wang, Y. Shi, M. Zhang, F. Wang, C.
Pan, P. Wang, Chem. Mater. 2010, 22, 1915–1925; g) A. Hagfeldt, G. Bos-
chloo, L. Sun, L. Kloo, H. Pettersson, Chem. Rev. 2010, 110, 6595–6663;
h) R. Argazzi, N. Y. M. Iha, H. Zabri, F. Odobel, C. A. Bignozzi, Coord.
Chem. Rev. 2004, 248, 1299–1316.
[3] a) J. B. Asbury, E. Hao, Y. Wang, H. N. Ghosh, T. Lian, J. Phys. Chem. B
2001, 105, 4545–4557; b) G. Ramakrishna, D. A. Jose, D. Krishna Kumar,
A. Das, D. K. Palit, H. N. Ghosh, J. Phys. Chem. B 2005, 109, 15445–
15453; c) T. Banerjee, S. Kaniyankandy, A. Das, H. N. Ghosh, Inorg. Chem.
2013, 52, 5366–5377; d) S. Verma, P. Kar, T. Banerjee, A. Das, H. N.
Ghosh, J. Phys. Chem. Lett. 2012, 3, 1543–1548; e) J. Teuscher, J.-D. De-
coppet, A. Punzi, S. M. Zakeeruddin, J.-E. Moser, M. Grꢀtzel, J. Phys.
Chem. Lett. 2012, 3, 3786–3790; f) S. Verma, P. Kar, A. Das, D. K. Palit,
H. N. Ghosh, Chem. Eur. J. 2010, 16, 611–619; g) S. Verma, H. N. Ghosh,
J. Phys. Chem. Lett. 2012, 3, 1877–1884.
[4] a) K. Hara, T. Sato, R. Katoh, A. Furube, Y. Ohga, A. Shinpo, S. Suga, K.
Sayama, H. Sugihara, H. Arakawa, J. Phys. Chem. B 2003, 107, 597; b) K.
Hara, Z.-S. Wang, T. Sato, A. Furube, R. Katoh, H. Sugihara, Y. Dan-oh, C.
Kasada, A. Shinpo, S. Suga, J. Phys. Chem. B 2005, 109, 476; c) Z.-S.
Wang, Y. Cui, K. Hara, Y. Dan-oh, C. Kasada, A. Shinpo, Adv. Mater. 2007,
19, 1138–1141; d) A. Mishra, M. K. R. Fischer, P. Buerle, Angew. Chem.
2009, 121, 2510–2536; Angew. Chem. Int. Ed. 2009, 48, 2474–2499;
e) K. Hara, Z.-S. Wang, T. Sato, A. Furube, R. Katoh, H. Sugihara, Y. Dan-
oh, C. Kasada, A. Shinpo, S. Suga, J. Phys. Chem. B 2005, 109, 15476–
15482; f) S. Kim, J. K. Lee, S. O. Kang, J. Ko, J.-H. Yum, S. Fantacci, F. D.
Angelis, D. Di Censo, M. K. Nazeeruddin, M. Grꢀtzel, J. Am. Chem. Soc.
2006, 128, 16701–16707; g) T. Horiuchi, H. Miura, K. Sumioka, S. Uchida,
J. Am. Chem. Soc. 2004, 126, 12218–12219; h) H. Qin, S. Wenger, M. Xu,
F. Gao, X. Jing, P. Wang, S. M. Zakeeruddin, M. Grꢀtzel, J. Am. Chem. Soc.
2008, 130, 9202–9203; i) G. Zhang, H. Bala, Y. Cheng, D.-G. Shi, X. Lv, Q.
Yu, P. Wang, Chem. Commun. 2009, 2198–2200.
[6] a) A. Abbotto, N. Manfredi, C. Marinzi, F. De Angelis, E. Mosconi, J.-H.
Yum, Z. Xianxi, M. K. Nazeeruddin, M. Grꢀtzel, Energy Environ. Sci. 2009,
2, 1094–1101; b) K. R. Justin Thomas, Y.-C. Hsu, J. T. Lin, K.-M. Lee, K.-C.
Ho, C.-H. Lai, Y.-M. Cheng, P.-T. Chou, Chem. Mater. 2008, 20, 1830–
1840; c) Y.-S. Yen, Y.-C. Hsu, J. T. Lin, C.-W. Chang, C.-P. Hsu, D.-J. Yin, J.
Phys. Chem. C 2008, 112, 12557–12567; d) J. T. Lin, P.-C. Chen, Y.-S. Yen,
Y.-C. Hsu, H.-H. Chou, M.-C. P. Yeh, Org. Lett. 2009, 11, 97–10.
[7] a) S. Verma, P. Kar, A. Das, H. N. Ghosh, Chem. Eur. J. 2011, 17, 1561–
1568; b) T. Banerjee, S. Rawalekar, A. Das, H. N. Ghosh, Eur. J. Inorg.
Chem. 2011, 4187–4197.
[8] a) M. Vogel, W. Rettig, Ber. Bunsen-Ges. 1985, 89, 962; b) W. Rettig,
Angew. Chem. 1986, 98, 969–986; Angew. Chem. Int. Ed. 1986, 25, 971–
988; c) E. Lippert, W. Rettig, V. Bonacic-Koutecky, F. Heisel, J. A. Miehe,
Adv. Chem. Phys. 1987, 68, 1; d) W. Rettig, Topics in Current Chemistry,
169, Electron-Transfer I (Ed.: J. Mattay), Springer, Berlin, 1994, p. 254.
[9] a) G. Ramakrishna, H. N. Ghosh, J. Phys. Chem. A 2002, 106, 2545–2553;
b) S. Verma, H. N. Ghosh, Ultrafast Phenomena XVIII, EPJ Web of Confer-
ences 2013, 41, Art. No. 08001.
[10] K.-F. Chen, C.-W. Chang, J.-L. Lin, Y.-C. Hsu, M.-C. P. Yeh, C.-P. Hsu, S.-S.
Sun, Chem. Eur. J. 2010, 16, 12873–12882.
[11] a) D. Bahnemann, A. Henglein, J. Lilie, L. Spanhel, J. Phys. Chem. 1984,
88, 709–711; b) H. N. Ghosh, J. Phys. Chem. B 1999, 103, 10382–10387.
[12] A. Zaban, S. Ferrere, J. Sprague, B. A. Gregg, J. Phys. Chem. B 1997, 101,
55–57.
[13] S. Kaniyankandy, S. Rawalekar, S. Verma, H. N. Ghosh, Phys. Chem. Chem.
Phys. 2010, 12, 4210–4216.
[14] W. Rettig, G. Wermuth, J. Photochem. 1985, 28, 351.
[15] J. Yarwood, Spectroscopy and Structure of Molecular Complexes, Plenum
Press, New York, 1970.
[16] A. El-Zohry, A. Orthaber, B. Zietz, J. Phys. Chem. C 2012, 116, 26144–
26153.
[17] A. Kay, R. Humphry-Baker, M. Grꢀtzel, J. Phys. Chem. 1994, 98, 952–959.
[18] J. B. Asbury, Y. Q. Wang, E. Hao, H. N. Ghosh, T. Lian, Res. Chem. Intermed.
2001, 27, 393.
Received: October 5, 2013
[5] a) A. Hagfeldt, M. Grꢀtzel, Acc. Chem. Res. 2000, 33, 269–277; b) F.
Matar, T. H. Ghaddar, K. Walley, T. Dos Santos, J. R. Durrant, B. O’Regan,
Revised: December 5, 2013
Published online on February 24, 2014
Chem. Eur. J. 2014, 20, 3510 – 3519
www.chemeurj.org
3519
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim