Effect of molecular structure on interfacial electron transfer dynamics of 7-N,N-dimethyl coumarin 4-acetic acid (DMACA) and 7-hydroxy coumarin 4-acetic acid (HCA) sensitized TiO2 and ZrO2 nanoparticles

Article abstract of DOI:10.1021/jp049701z

Ultrafast transient absorption spectroscopy has been employed to understand the effect of molecular structure on interfacial electron transfer (ET) dynamics of 7-N,N-dimethyl amino coumarin 4-acetic acid (DMACA) and 7-hydroxy coumarin 4-acetic acid (HCA) sensitized TiO₂ and ZrO₂ nanoparticles. Electron injection is confirmed by observing the cation radical of the dye molecules as well as the conduction band electron in the visible and near-IR regions. Electron injection efficiency has been found to be higher for the HCA/TiO₂ system as compared to the DMACA/TiO₂ system. Both dyes are structurally similar except that HCA has a hydroxyl group at the 7-position while DMACA has a dimethyl amino group at the 7-position. Steady-state and time-resolved fluorescence measurements confirmed that, in highly polar solvent, the excited state of DMACA dye exists both in twisted intramolecular charge transfer (TICT) and intramolecular charge transfer (ICT) states, whereas excited HCA exists only in the ICT state. Because the charge in the case of the TICT state of DMACA is localized away from the nanoparticle surface, electron injection from that state is not efficient. However, ICT states of both DMACA and HCA can inject electrons efficiently as the charge is delocalized. Hence, the quantum yield of electron injection is high in the case of HCA compared to DMACA. We have also observed that photoexcited DMACA whose energy level lies above the conduction band of ZrO₂ nanoparticle can inject electrons after exciting with 400 nm laser light. Back electron transfer (BET) rates have been determined by following the decay kinetics of the cation radical and conduction band electron in different systems. The BET rate is found to be slower for HCA/TiO₂ compared to DMACA/TiO₂, as they fall in the inverted region of Marcus electron transfer theory. The BET is faster in the DMACA/TiO₂ system as compared to that of the DMACA/ZrO₂ system. However, the residual absorption after 460 ps is less in the case of the DMACA/ZrO₂ system.

Full text of DOI:10.1021/jp049701z

J. Phys. Chem. B 2004, 108, 12489-12496
12489
Effect of Molecular Structure on Interfacial Electron Transfer Dynamics of 7-N,N-Dimethyl
Coumarin 4-Acetic Acid (DMACA) and 7-Hydroxy Coumarin 4-Acetic Acid (HCA)
Sensitized TiO2 and ZrO2 Nanoparticles
G. Ramakrishna, Ajay K. Singh, Dipak K. Palit, and Hirendra N. Ghosh*
Radiation Chemistry & Chemical Dynamics DiVision, Bhabha Atomic Research Center,
Trombay, Mumbais400 085, India
ReceiVed: January 21, 2004; In Final Form: June 8, 2004
Ultrafast transient absorption spectroscopy has been employed to understand the effect of molecular structure
on interfacial electron transfer (ET) dynamics of 7-N,N-dimethyl amino coumarin 4-acetic acid (DMACA)
and 7-hydroxy coumarin 4-acetic acid (HCA) sensitized TiO
confirmed by observing the cation radical of the dye molecules as well as the conduction band electron in the
visible and near-IR regions. Electron injection efficiency has been found to be higher for the HCA/TiO
system as compared to the DMACA/TiO system. Both dyes are structurally similar except that HCA has a
2 2
and ZrO nanoparticles. Electron injection is
2
2
hydroxyl group at the 7-position while DMACA has a dimethyl amino group at the 7-position. Steady-state
and time-resolved fluorescence measurements confirmed that, in highly polar solvent, the excited state of
DMACA dye exists both in twisted intramolecular charge transfer (TICT) and intramolecular charge transfer
(ICT) states, whereas excited HCA exists only in the ICT state. Because the charge in the case of the TICT
state of DMACA is localized away from the nanoparticle surface, electron injection from that state is not
efficient. However, ICT states of both DMACA and HCA can inject electrons efficiently as the charge is
delocalized. Hence, the quantum yield of electron injection is high in the case of HCA compared to DMACA.
We have also observed that photoexcited DMACA whose energy level lies above the conduction band of
ZrO
rates have been determined by following the decay kinetics of the cation radical and conduction band electron
in different systems. The BET rate is found to be slower for HCA/TiO compared to DMACA/TiO , as they
fall in the inverted region of Marcus electron transfer theory. The BET is faster in the DMACA/TiO system
as compared to that of the DMACA/ZrO system. However, the residual absorption after 460 ps is less in the
case of the DMACA/ZrO system.
2
nanoparticle can inject electrons after exciting with 400 nm laser light. Back electron transfer (BET)
2
2
2
2
2
1
. Introduction
important task to understand the relation between the solar cell
performance and the molecular and structural properties of the
sensitizer molecules. Many molecular properties, such as
excited-state properties and their lifetimes, injection efficiency,
and redox potential, have significant impact on the overall solar
cell performance.
Incessant efforts are going on throughout the world to
understand the interfacial electron transfer from molecular
adsorbates to wide band-gap semiconductors. The never-ending
thrust in this area is largely owed to their application in various
fields such as photochemical solar energy conversion and
storage,1 nonlinear optics and heterogeneous photocatalysis.
,2
4
5-7
The molecular structure of the sensitizing dye is an important
factor, which can affect the solar cell performance. Recently,
Arakawa and co-workers have reported the effect of the number
Among such studies, dye-sensitization of wide band-gap
semiconductor electrodes has gained sufficient attention in recent
years, mainly because of demonstration of the dye-sensitized
solar cells with conversion efficiency as high as 10%. The
mechanism of these devices is based upon the injection of an
electron from the excited state of the sensitized dye into the
conduction band of the semiconductor. Studies on dye-sensitiza-
24
of anchoring carboxyl groups and the effect of ligand structure
1
,2
on the efficiency of electron injection into TiO nanoparticles
2
from the excited states of Ru-phenanthroline complexes, which
in turn affect the solar cell performance. They have observed
that electron injection efficiency of a dye molecule with four
8
-15
anchoring groups is four times as large as that of a dye molecule
tion of TiO2 nanoparticles have been carried out by us
as
with one anchoring group. Lian and co-workers2
7,17
have
well as by many other groups.1
6-20
The efficiency of the dye-
reported the effect of molecular structure of Ru-polypyridyl
dyes on their ground and excited energy levels, which in turn
affect the quantum yield of electron injection. We have also
reported earlier the effect of molecular structure on the electron
injection efficiency of 7-diethyl amino coumarin 3-carboxylic
acid (D-1421) and coumarin 343 (C-343) sensitized TiO2
nanoparticles. We have shown that, with a little change in
molecular structure, excited-state properties of the dye can be
influenced drastically, which in turn affect the injection quantum
sensitized solar cells depends critically on the rates of the
forward (dye to semiconductor) and back (semiconductor to dye)
electron-transfer reactions. For efficient solar energy conversion,
it is necessary to establish the conditions of both higher electron
injection efficiency and slow recombination. For that purpose,
11
many new sensitizer molecules have been synthesized, and the
solar cell performance has been evaluated.2
1-26
It is a very
*
Corresponding author. E-mail: hnghosh@magnum.barc.ernet.in.
1
0.1021/jp049701z CCC: $27.50 © 2004 American Chemical Society
Published on Web 07/22/2004

12490 J. Phys. Chem. B, Vol. 108, No. 33, 2004
Ramakrishna et al.
SCHEME 1 : Molecular Structure of 7-N,N-Dimethyl
Coumarin 4-Acetic Acid (DMACA) and 7-Hydroxy
Coumarin 4-Acetic Acid (HCA)
only in the ICT state. Time-resolved absorption studies show
that electron injection efficiency is higher (∼1.6 times) in the
case of HCA sensitized TiO2 nanoparticles compared to
DMACA sensitized nanoparticles under similar conditions. It
is seen from Scheme 2 that, in the TICT state of DMACA, an
electron gets localized at the 3-position oxygen atom, and from
there the possibility of electron injection is remote. However,
ICT states of DMACA can inject electrons (Scheme 2). As the
population of ICT states is higher in HCA compared to
DMACA, injection quantum yield is high in the HCA/TiO2
system. The other important observation we made is the electron
injection into the conduction band of ZrO2 nanoparticles from
photoexcited DMACA. BET rates have been determined by
following the decay kinetics of the cation radical in different
systems. The BET rate is found to be slower for HCA/TiO2
compared to DMACA/TiO2, where both systems follow the
equation of the ET reaction in the Marcus inverted region. BET
dynamics is faster in DMACA/TiO2 as compared to the
DMACA/ZrO2 system. However, the residual absorption is less
in the case of ZrO2 as compared to TiO2.
yield. It is a very important task to optimize the molecular
structure of the sensitizer molecules which will have higher
injection quantum yield and at the same time will have a low
charge recombination rate.
Reports are available on interfacial electron transfer dynamics,
mainly based on dye-sensitized TiO2, SnO2, and ZnO nano-
particles, where energy levels of the photoexcited dyes are
higher than the conduction band edge of the semiconductor
nanoparticles. However, in most cases, the photoexcited dyes
cannot inject electrons into the ZrO2 nanoparticles because
energy level of the photoexcited dyes are lower than the
conduction band edge of ZrO2. Some reports are available on
2
. Experimental Section
a) Materials. 7-N,N-Dimethyl amino coumarin 4-acetic acid
(
(DMACA) and 7-hydroxy coumarin 4-acetic acid (HCA) were
obtained from Aldrich and used without further purification.
28
Titanium(IV) tetraisopropoxide {Ti[OCH(CH ) ] } (Aldrich,
electron injection into ZrO2 nanoparticles by Huber et.al., Hao
3 2 4
et al.29 and us,
14,15
97%) and isopropyl alcohol (Aldrich) were purified by dis-
where electron injection takes place in the
tillation. Zirconium(IV) isopropoxide 2-propanol complex
Zr[OCH(CH3)2]4‚(CH3)2CHOH (Aldrich, 99.9%) was used
without further purification. Nanopure water (Barnsted System)
was used for making aqueous solutions.
surface states of ZrO2. It will be very interesting to see the
electron injection dynamics from suitable photoexcited dyes into
the conduction band of ZrO2 nanoparticles and the corresponding
back electron-transfer reaction.
In the present investigation, we have studied electron injection
and back electron-transfer dynamics of of 7-N,N-dimethyl amino
coumarin 4-acetic acid (DMACA) and 7-hydroxy coumarin
(b) Sample Preparation. Nanometer-size TiO2 was prepared
by controlled hydrolysis of titanium(IV) tetraisopropoxide.3
A solution of 5 mL of Ti[OCH(CH3)2]4 (Aldrich, 97%) dissolved
in 95 mL of isopropyl alcohol (Aldrich) was added dropwise
(1 mL/min) to 900 mL of Nanopure water (2 °C) at pH 1.5
(adjusted with HNO3). The solution was continuously stirred
for 10-12 h until a transparent colloid was formed. The
colloidal solution was concentrated at 35-40 °C with a rotary
evaporator and then dried with a nitrogen stream to yield a white
powder. ZrO2 nanoparticles were also prepared by adopting the
above procedure. In the present work, all colloidal samples were
prepared after dispersing the dry TiO2 or ZrO2 nanoparticles in
water (15 g/L). A transparent clear colloidal solution was
formed. For all measurements, sample solutions were deoxy-
genated by continuously bubbling high purity nitrogen (99.95
Iolar grade from Indian Oxygen Co. Ltd., India) through the
0-32
4
-acetic acid (HCA) sensitized TiO2 and ZrO2 nanoparticles
using ultrafast transient absorption spectroscopy. Electron
injection has been confirmed by observing the cation radical of
the dye molecules as well as the conduction band electron in
the visible and near-IR region. To see the effect of molecular
structure on electron injection and BET dynamics, these two
dyes are chosen. Both the dyes have same anchoring group,
which binds to the nanoparticle with similar structure except
DMACA has a dimethyl amino at the 7-position and HCA has
a hydroxy group at the 7-position (Scheme 1). Steady-state and
time-resolved fluorescence measurements confirmed that, in
highly polar solvent, the excited state of DMACA dye exists
both in the TICT and the ICT state, whereas excited HCA exists
SCHEME 2 : Orbital Scheme Showing Electron Injection from ICT and TICT State of 7-N,N-Dimethyl Coumarin
-Acetic Acid (DMACA) into TiO2 Nanoparticles
4

Interfacial Electron Transfer Dynamics
J. Phys. Chem. B, Vol. 108, No. 33, 2004 12491
solutions. The solutions flowed through a 1 mm path quartz
cell during all the measurements.
(c) Femtosecond Spectrometer. The femtosecond tunable
visible spectrometer has been developed based on a multipass
amplified femtosecond Ti:sapphire laser system from CDP-
Avesta, Russia (1 kHz repetition rate at 800 nm, 50 fs, 300
µJ/pulse) and has been described earlier.12 The 800 nm output
pulse from the multipass amplifier is split into two parts to
generate pump and probe pulses. One part, with a 200 µJ/pulse,
is frequency doubled in BBO crystals to generate pump pulses
at 400 nm. To generate visible probe pulses, about 3 µJ of the
800 nm beam is focused onto a 1.5 mm thick sapphire window.
The intensity of the 800 nm beam is adjusted by iris size and
neutral density filters to obtain a stable white light continuum
in the 470 nm to over 1000 nm region. The probe pulses are
split into the signal and reference beams and are detected by
two matched photodiodes with variable gain. We have kept the
spot sizes of the pump beam and probe beam at the crossing
point around 500 and 300 µm, respectively. The excitation
Figure 1. Transient absorption spectra of (a) DMACA and (b) HCA
in water at the end of the laser pulse. The transient absorption peak at
690 nm has been attributed to the excited singlet state, and the negative
absorption band at 470 nm has been attributed to stimulated emission
for both the dyes. [Inset: transient absorption decay at 670 nm for (a)
DMACA and (b) HCA.]
2
energy density (at 400 nm) was adjusted to ∼2500 µJ/cm . The
noise level of the white light is about ∼0.5% with occasional
spikes due to oscillator fluctuation. We have noticed that most
laser noise is low-frequency noise and can be eliminated by
comparing the adjacent probe laser pulses (pump blocked vs
unblocked using a mechanical chopper). The typical noise in
the measured absorbance change is about <0.3%. The instru-
ment response function was obtained by fitting the rise time of
the bleach of sodium salt of meso-tetrakis(4 sulfonatophenyl)
porphyrin (TPPS) at 710 nm, which has an instantaneous
response.
light. Figure 1 shows the transient absorption spectra of
DMACA (Figure 1a) and HCA (Figure 1b) in water at the end
of the laser pulse. The HCA molecule does not absorb at 400
33
nm at pH 2.5. To perform transient absorption studies of HCA,
we have increased the pH of the free dye solution to 9. At this
pH, HCA exists as an anion, which absorbs at 400 nm. Transient
absorption spectra of the excited state of DMACA show an
absorption peak at 690 nm and negative absorption band from
4
70 to 580 nm (Figure 1a). On the other hand, the excited state
of HCA shows an absorption peak at 675 nm and a negative
absorption band from 470 to 580 nm (Figure 1b). The transient
absorption peak at 690 nm for the excited DMACA and the
peak at 670 nm for the excited HCA have been attributed to
the excited singlet state absorption (ESA), and corresponding
negative absorption bands have been attributed to the stimulated
emission of the respective molecules. The inset of Figure 1
shows the kinetic decay trace of ESA for both DMACA (Figure
(
d) Cyclic Voltammetry. Voltammetric experiments were
performed with Auto Lab PGSTAT 20 (Manufactured by Eco-
Chemie, Netharlands) coupled to a Metrohm 663 VA stand
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 25 ( 0.1 °C. Measurements
were made in acetonitrile solution with TEAP (triethylammo-
nium perchlorate) as a supporting electrolyte and in N2
atmosphere. Redox potentials of DMACA and HCA have been
determined to be 1.0 and 1.6 V against Ag/AgCl electrode,
respectively.
1c) and HCA (Figure 1d) molecules. Excited states of DMACA
and HCA show little decay up to 400 ps indicating longer
lifetimes. However, it is clear that the excited state of DMACA
decays faster than that of HCA. We have also carried out time-
resolved emission measurements of both the molecules using
time-correlated single photon counting. Emission lifetimes of
DMACA and HCA are found to be 1.4 ns and 5.65 ns,
respectively.
3
. Results and Discussion
(a) Steady-State Absorption and Emission Spectroscopy.
33
We have carried out steady-state absorption and emission
studies of both DMACA and HCA molecules in water and in
the presence of TiO2 and ZrO2 colloidal solutions. Our measure-
ments have shown that both the dye molecules are adsorbed on
both the nanoparticle surfaces. The fluorescence intensity of
both DMACA and HCA molecules has drastically reduced in
the presence of TiO2 nanoparticles as compared to that in
aqueous solution, which has been attributed to the electron
injection into the TiO2 nanoparticles. It is very interesting to
see that the fluorescence from DMACA is also quenched even
in the presence of ZrO2 nanoparticles indicating the electron
injection from DMACA into the ZrO2 nanoparticles. On the
other hand, fluorescence from the HCA molecule does not
change much on the ZrO2 nanoparticle surface as compared to
that in water showing negligible excited-state interaction of the
dye with the ZrO2 nanoparticle surface.
(c) Transient Absorption Measurements in DMACA/TiO2
and HCA/TiO Systems. To study the effect of molecular
2
structure on interfacial electron-transfer dynamics we have
carried out transient absorption measurements for both DMACA
and HCA sensitized TiO nanoparticles excited with a 400 nm
2
femtosecond laser pulse. Figure 2 shows the time-resolved
transient absorption spectra of the DMACA/TiO system in
2
water. The spectrum at each time delay consists of a positive
peak at ∼630 nm with a shoulder at 690 nm and a broad positive
feature in the whole spectral region (800-1000 nm). In the
DMACA/TiO2 system, we did not observe any ESA and
stimulated emission, which can arise due to a free excited
DMACA molecule in the solution that does not take part in the
interfacial process. The broad spectral absorption in the 800-
1000 nm region is attributed to the absorption of conduction
-
(b) Excited-State Dynamics of Free DMACA and HCA
band electrons in the nanoparticles (eCB ). It has already been
Molecules. We have carried out transient absorption measure-
ments to study the excited-state dynamics of both DMACA and
HCA molecules in water upon excitation with 400 nm laser
shown by many workers that the conduction band electrons can
34,35,8-14
16,17
be detected both by visible
and infrared absorption.
The positive peak at 630 nm has been attributed to the cation

12492 J. Phys. Chem. B, Vol. 108, No. 33, 2004
Ramakrishna et al.
Figure 2. Transient absorption spectra of DMACA sensitized TiO
2
Figure 4. Transient absorption spectra of DMACA sensitized ZrO
2
nanoparticle in water at (a) 200 fs, (b) 5 ps, and (c) 10 ps after excitation
at 400 nm (fwhm ) 50 fs). The spectrum at each time delay consists
of an absorption peak at 630 nm and a broad positive absorption feature
in the whole spectral region (800-1000 nm). These features are
assigned to the cation radical of DMACA dye and the injected electron
in nanoparticles, respectively. [Inset (d): kinetic decay trace of DMACA
cation radical at 630 nm.]
nanoparticle in water at (a) 500 fs, (b) 1 ps, (c) 5 ps, and and (d) 10 ps
after excitation at 400 nm. The spectrum at each time delay consists
of an absorption peak at 630 nm and a broad positive absorption feature
in the whole spectral region (800-1000 nm). These features are
assigned to the cation radical of DMACA dye and the injected electron
in nanoparticles, respectively. [Inset: kinetic decay trace of DMACA
cation radical at 630 nm.]
TABLE 1: Electron Injection and Back Electron Transfer
(
BET) Dynamics of Different Dye-Sensitized Nanoparticle
Systems Monitoring at Different Wavelength
DMACA/TiO
2
, 630 nm τinj ) 100 fs
〈τav〉 ) 5.05 ps (67.8%)
〈τav〉 ) 22.6 ps (88.1%)
〈τav〉 ) 68.2 ps (57.3%)
τ
1
τ
2
τ
3
) 1.1 ps (44%)
) 13 ps (23.8%)
> 400 ps (32.2%)
DMACA/ZrO , 630 nm τinj ) 100 fs
2
τ
1
τ
2
τ
3
) 2.7 ps (38.1%)
) 38 ps (50%)
> 400ps (11.9%)
HCA/TiO
2
, 610 nm
τ
τ
τ
τ
inj ) 100 fs
1
2
3
) 3.8 ps (22.2%)
) 110 ps (35.1%)
> 400 ps (42.7%)
Figure 3. Transient absorption spectra of HCA sensitized TiO
2
nanoparticle in water at (a) 200 fs, (b) 5 ps, and (c) 10 ps after excitation
at 400 nm. The spectrum at each time delay consists of an absorption
peak at 610 nm and a broad positive absorption feature in the whole
spectral region (800-1000 nm). These features are assigned to the
cation radical of HCA molecule and the injected electron in nanopar-
ticles, respectively. [Inset (d): kinetic decay trace of HCA cation radical
at 610 nm.]
Similarly, transient absorption experiments have been carried
out on the HCA/TiO2 system in water and are shown in Figure
3
. The spectrum at each time delay consists of a positive peak
at 610 nm with a shoulder at 710 nm and a broad positive feature
in the whole spectral region (750-1000 nm). The peak at 610
nm with the shoulder at 710 nm has been attributed to the HCA
•
+
cation radical (HCA ) as confirmed by pulse radiolysis
33
radical of DMACA molecule. Assignment of this band has been
made on the basis of the results obtained in separate pulse
radiolysis experiments,33 where DMACA was generated
experiments, and the band at 750-1000 nm has been attributed
to the electron in the conduction band of the nanoparticles. This
decay of the signals has been attributed to charge recombination
or BET with the nanoparticles (Figure 3, inset). The decay
kinetics at 610 nm has been found to follow a multiexponential
function with typical time constants of 3.8 ps (22.2%), 110 ps
(35.1%), and >400 ps (42.7%) (Table 1).
•+
•
selectively by the reaction of a N3 radical with a DMACA
molecule in N2O saturated aqueous solution. The transient
spectrum obtained from the above experiment shows an
absorption peak at 630 nm with a shoulder at 690 nm. We have
measured electron injection time by monitoring the appearance
(d) Transient Absorption Measurements in the DMACA/
ZrO System. As the surface nature of ZrO nanoparticles is
signal of DMACA• at 630 nm as well as eCB at 900 nm, and
it was found to be ∼ 100 fs (Figure 4). However, in the transient
spectra (Figure 2) we have observed that the transient signal in
the 630 nm region and also that in the near-IR region decay
rapidly with time. This decay of the transient signal has been
attributed to charge recombination or BET from the electron in
the nanoparticle to the parent dye cation radical (Figure 2, inset).
The decay kinetics at 630 nm has been fitted with a multi-
exponential function with the time constants of 1.1 ps (44%),
+
-
2
2
3
0
very similar to that of TiO nanoparticles, in our previous
investigations on dye-sensitized TiO nanoparticles reaction
2
8
,9
2
to understand the photophysics of the dye molecule on a
noninjecting nanoparticle surface, we have used ZrO nanopar-
2
ticles. As the conduction band edge of ZrO nanoparticles is
2
4
1
1.3 eV higher than that of TiO nanoparticles, electron injection
2
is not possible from the photoexcited dye molecules. So in the
8,9
earlier studies, we have observed that excited-state properties
of the dye molecules on ZrO2 nanoparticles surface are very
similar to those of bulk solvents. It is observed from steady-
state emission measurements that the fluorescence intensity of
DMACA is drastically reduced in the presence of ZrO2
1
3 ps (23.8%), and >400 ps (32.2%) (Table 1). We have also
observed that the kinetic decay trace of the injected electron at
00 nm can be fitted multiexponentially with similar time
constants to that of cation radical at 630 nm.
9

Interfacial Electron Transfer Dynamics
J. Phys. Chem. B, Vol. 108, No. 33, 2004 12493
SCHEME 3 : Mechanistic Scheme for Electron Transfer
from the Electronically Excited Dye to TiO2 and ZrO2
Nanoparticlesa
Figure 5. Relative magnitude of electron absorption signal at 900 nm
in (a) DMACA sensitized TiO
TiO nanoparticles.
2
nanoparticles and (b) HCA sensitized
2
and >400 ps (11.9%) (Table 1). However, for the longer time
>400 ps) residual charge separated (CS) species in DMACA/
(
ZrO2 is much less as compared to the situation in the DMACA/
TiO2 system (Table 1).
a
+
Here, S*/S is the excited sensitized dye/cation radical couple, E
is the conduction band edge, E is the valence band, SS is the surface
c
v
(e) Effect of Molecular Structure on Quantum Yield of
Electron Injection (Oinj). It will be interesting to find out the
effect of molecular structure on the quantum yield of electron
injection (φinj) in these studied systems. We have carried out
femtosecond transient absorption measurements of the above
systems in similar condition. The optical density of the
experimental solutions and the laser intensity was kept the same.
At similar laser intensity and the same optical density at
excitation wavelength for the samples, the measured change in
absorbance can be ascribed to the injection quantum yield
relative to one another. Figure 5 shows the kinetic trace of the
appearance signal of electrons in the conduction band at 900
nm of DMACA/TiO2 (Figure 5a), and HCA/TiO2 (Figure 5b).
It is clear from Figure 5 that φinj in HCA/TiO2 system is ∼1.6
times higher as compared to that of DMACA/TiO2 system.
In the present investigation, as the basic molecular structures
of both DMACA and HCA are very similar (Scheme 1), the
electronic couplings of the dye molecules with the TiO2
nanoparticle are similar (both the dyes have the same anchoring
group with the nanoparticle). Still, we have observed 60% more
electron injection in the HCA/TiO2 system as compared to that
in the DMACA/TiO2 system. Solvatochromic measurements
reveal that the locally excited state for both the dyes is an
intramolecular charge transfer (ICT) state. Emission quantum
yield (φem) measurements have been carried out for both the
dyes with changing polarity of the solvents. It has been observed
that φem of HCA monotonically decreases with solvent polarity;
however, φem of DMACA decreases drastically in highly polar
states, ICT is the intramolecular charge transfer state, and TICT is the
twisted intramolecular charge transfer state. kBET is the back electron-
transfer rate.
nanoparticles, which indicates that ZrO2 nanoparticle surface
is not a noninjecting surface for the photoexcited DMACA
molecule. However, the ZrO2 nanoparticle is an inert surface
for an excited HCA molecule. In the present investigation, we
have carried out femtosecond transient absorption experiments
on DMACA molecules adsorbed on the ZrO2 nanoparticle
excited with 400 nm, to study the excited-state dynamics of
the dye on ZrO2 surface. Figure 4 shows the time-resolved
transient absorption spectra of the DMACA/ZrO2 system in
water. The spectrum at each time delay consists of a positive
peak at ∼630 nm with a shoulder at 690 nm and a broad positive
feature in the whole spectral region (800-1000 nm) which can
be attributed to the cation radical and conduction band electron,
respectively. It will be interesting to find out the reason for
electron injection from the excited state of the DMACA
molecule into ZrO2 nanoparticles when a structurally similar
excited HCA molecule is unable to do so. We have measured
the energetics of the ground and excited redox potentials of both
+
+
DMACA (ED/D ) 1.0 V) and HCA (ED/D ) 1.6 V) molecules
Scheme 3). The GS redox potential of the dyes has been
(
measured by cyclic voltametry against a Ag/AgCl electrode.
The dye oxidations are reversible. The excited redox potentials
of the dyes were measured by adding E00 and GS redox
potential. E00 has been measured from the crossing point of the
excitation and emission spectra of the dye molecules. It is visible
3
3
solvent with a sudden break. This observation confirmed that
in highly polar solvent, the excited DMACA molecule relaxes
very fast to another nonemitting state. This nonemitting state
is often attributed to the twisted intramolecular charge transfer
from Scheme 3 that the excited redox potential of DMACA
+
(
ED*/D ) -1.9 V) is higher in energy than the conduction band
ZrO
30
edge of ZrO2 nanoparticles (ECB
2
) -1.8 V ). As a result,
3
6-39
40
photoexcited DMACA can easily inject electrons in the CB of
ZrO2 nanoparticles. However, the excited redox potential of
(TICT) state.
Nad et al. have shown the presence of a
nonemitting TICT state in the excited states of the coumarin-
152 and coumarin-1 molecules, which are structurally very
similar to DMACA. However, the transient absorption spectrum
of DMACA in both less and more polar solvents looks very
similar. From this observation, we can conclude that the TICT
state of DMACA does not show any transient absorption. The
transient absorption spectrum for DMACA in highly polar
solvent is attributed to that of the ICT state only.
+
HCA (ED*/D ) -1.6 V) is lower in energy than the conduction
band edge of ZrO2 nanoparticles. The transient absorption
spectrum DMACA/ZrO2 looks very similar to that of DMACA/
TiO2. The transient signal in the 630 nm region and also in the
near-IR region decay similarly as observed in the DMACA/
TiO2 system, which has been attributed to charge recombination
(CR) or BET with the nanoparticles (Figure 4, inset). The decay
kinetics at 630 nm has been fitted with a multiexponential
function with the time constants of 2.7 ps (38.1%), 38 ps (50%),
In our earlier studies, we have compared electron injection
(φinj) efficiency of two structurally similar dyes, 7-diethyl amino

12494 J. Phys. Chem. B, Vol. 108, No. 33, 2004
Ramakrishna et al.
coumarin 3-carboxylic acid (D-1421) and coumarin 343 (C-
11
3
43), adsorbed on TiO2 nanoparticles. Solvatochromic mea-
surements confirmed that the excited D-1421 molecule exists
as both ICT and TICT states and the excited C-343 molecule
exists as the ICT state in highly polar solvents. We have
observed that φinj in D-1421/TiO2 system is ∼1.5 times higher
compared to that in the C-343/TiO2 system. As the TICT state
of D-1421 is a charge-separated state, it can inject electrons
more efficiently into TiO2 nanoparticles. However, in the present
investigation we have observed that although excited DMACA
exists as a TICT state, still φinj in the DMACA/TiO2 system is
1.6 times lower than that in the HCA/TiO2 system. The observed
differences can be explained with the following explanation.
In the TICT states of both excited D-1421 and DMACA
molecules, the electron gets localized on the oxygen atom of
the keto group. D-1421 molecules couple with the TiO2
nanoparticles through phthalate type binding, where the keto
group also takes part in the coupling process. As a result,
electron injection is very efficient from the TICT state of the
D-1421 molecule to TiO2 nanoparticles. However, on the other
hand, in the case of the DMACA/TiO2 system, DMACA couples
with TiO2 nanoparticles through the carboxylic group, where
the keto group does not take part in the coupling process with
TiO2 nanoparticles (Scheme 2). As a result, the probability of
electron injection from the TICT state of DMACA will be less.
However, we have already discussed that excited DMACA
molecules exist both in ICT and TICT states. Electron injection
from the ICT state of DMACA can be highly possible (Scheme
Figure 6. Comparison of back electron transfer (BET) dynamics of
(a) DMACA sensitized TiO
ZrO nanoparticles monitoring the cation radical (DMACA ) at 630
nm. [Inset: kinetic decay trace at shorter time scale.]
2
nanoparticles and (b) DMACA sensitized
•+
2
TiO2 (2.1 eV), so BET dynamics is expected to be faster (in
the inverted regime) in the DMACA/TiO2 system which has
been reflected in our experimental results (Table 1). The inverted
behavior of BET dynamics has also been demonstrated by us
9
,42
43
and others in dye sensitized TiO2
surface.
and SnO2 nanoparticle
Similarly, we have compared the BET dynamics of DMACA
sensitized on both TiO2 and ZrO2 nanoparticles (see Figure 6).
As the injection is taking place from the DMACA molecule to
both the nanoparticles and it couples with the nanoparticles (both
TiO2 and ZrO2) in a similar fashion, we can assume that there
will be little difference in Λ and HAB for BET reaction. As
2
). The density of the ICT state will be less in the case of
DMACA because of the fact that there is another nonradiative
pathway leading to TICT, decreasing the density of ICT states.
As excited HCA molecules can exist only in ICT states, whose
density will be higher than that of ICT states of DMACA in
similar condition (at same excitation intensity), the electron
injection yield is higher in the HCA/TiO2 system as compared
to that of the DMACA/TiO2 system (Figure 5).
-∆G° in DMACA/TiO2 is lower (1.5 eV) compared to that in
DMACA/ZrO2 (2.8 eV), so BET dynamics is expected to be
faster in the DMACA/TiO2 system, which has been reflected
in our experimental results (Table 1). It may be reasonable to
observe Marcus inverted region back ET kinetics in the present
investigation, since the fast ET (<100 ps) is from electrons
located near the adsorbate where diffusion of electrons is not
the rate determining process. A similar argument was suggested
(f) Effect of Molecular Structure on ET Dynamics. In
addition to the injection quantum yield, it is important to
understand the effect of molecular structure on interfacial
electron transfer (ET) dynamics (both electron injection and
BET). In the present investigation, since both DMACA and
HCA molecules have the same electron donating orbital and
the same anchoring group to the surface, we can assume that
there is very little difference in electronic coupling for the
injection process. Shown in Figure 5 is comparison of the
electron injection time of TiO2 nanoparticles sensitized by
DMACA and HCA after 400 nm excitation. Electron injection
time for all the above systems has been found to be ∼100 fs
4
4
by Lian and co-workers in their catechols sensitized TiO2
nanoparticle systems. It is interesting to see that the rate of
charge recombination is slower for the case of the DMACA/
ZrO system as compared to that of the DMACA/TiO system,
2
2
but the residual absorption of charge separated species is less
for the case of ZrO as compared to that of TiO . About 88%
2
2
of the charge recombination takes place in DMACA/ZrO within
2
460 ps (our detecting limit) whereas only 68% of the CR
reaction takes place in the DMACA/TiO system. It is very
2
difficult to compare the back ET kinetics in TiO and ZrO2
2
(Table 1).
because they are highly nonexponential in nature, and we have
measured only a small cross-section (up to 460 ps) of the total
kinetic process. Still, we have tried to compare the BET process
by averaging the first two time constants adopting the following
equation.
Now it will be interesting to compare the back electron
transfer (BET) dynamics of the above systems and find out the
effect of molecular structure. We have already discussed that
BET dynamics is multiexponential in both DMACA/TiO2
(
Figure 2, inset) and HCA/TiO2 (Figure 3, inset) systems and
a τ + a τ
2 2
it is found to be faster in DMACA/TiO2 system (Table 1).
According to Marcus semiclassical theory,41 back electron-
transfer rate will be depending on ∆G°, the overall free energy
1
1
〈
τ 〉 )
(1)
av
a1 + a2
+
of reaction (∆G° ) EC - ES/S ), where Λ, the total reorganiza-
The observed average lifetime depends critically on the time
scale region that we are considering. In the present investigation,
we have considered the first two time constants (Table 1) to
calculate the average time constant where it can give information
up to first 100 ps of the total measured kinetics. At this point,
it is difficult for us to make any comment on the behavior of
the BET process of the above systems at longer time domains
tion energies, and HAB, the coupling elements of the system,
are the same. As the molecular structures of DMACA and HCA
are very similar and both the dyes couple with nanoparticles in
similar fashion, we can consider that there will be little
difference in Λ and HAB in the above systems. As -∆G° in
DMACA/TiO2 is lower (1.5 eV) as compared to that in HCA/

Interfacial Electron Transfer Dynamics
micro- and millisecond). However, within the time scale region
J. Phys. Chem. B, Vol. 108, No. 33, 2004 12495
(
the DMACA molecule exists in both ICT and TICT states. We
have observed that the TICT state of DMACA is a less efficient
injecting state, because in the TICT state charge is localized in
the keto group, which is away from the binding site of the dye
(-CH2COOH) with the nanoparticle. On the other hand, ICT
states of both HCA and DMACA can inject electrons into TiO2.
The higher density of ICT states in HCA compared to DMACA
resulted in a higher quantum yield of electron injection. Another
important observation that has been made is that the photo-
excited state of DMACA, which lies above the conduction band,
can inject electrons into the ZrO2 nanoparticles, which are
normally considered as a noninjecting surface for most photo-
excited dyes. However, the photoexcited state of HCA, which
lies below, cannot inject electrons in ZrO2. Electron injection
times have been found to be ∼100 fs in all the above systems
because both the dyes couple with both TiO2 and ZrO2 in a
similar fashion. However, BET dynamics is found to be slower
in the HCA/TiO2 system compared to the DMACA/TiO2 system,
following the equation of the ET reaction in the Marcus inverted
region. In the DMACA/ZrO2 system, BET dynamics is slower
compared to DMACA/TiO2 system; however, the charge
separated species left after 460 ps is less for the case of the
ZrO2 system.
of observation for about 100 ps, back ET is faster in the case
of TiO2 when compared to ZrO2. This reason behind the fast
back ET (<100 ps) is due to the recombination of the electrons
localized near the adsorbate, for which the diffusion into the
nanoparticle is so not important where the free energy of back
ET plays a significant role. For the >500 ps time scale, the
deep trap states and electron diffusion into the nanoparticles
can play a major role in back ET dynamics. As the two
semiconductors that we are considering are different, the
trapping/detrapping rate as well as the diffusion of electrons
between the trap states (hopping) may be different, and hence,
we can observe different dynamics for the >100 ps time scale
region in comparison to <100 ps time domain. Dynamics in
longer time domain can be explained by the fact that although
photoexcited DMACA can inject electron in the conduction band
in both TiO2 and ZrO2 nanoparticles, density of accepting states
in ZrO2 is less as compared to that in TiO2 at the injecting energy
level (-1.9 eV) (Scheme 3). Injected electrons in TiO2
nanoparticles will be hotter and will possess more energy
compared to those in ZrO2 nanoparticles; as a result diffusivity
of the injected electron will be greater within the TiO2
nanoparticles. So, the availability of injected electrons for the
recombination reaction will be greater in ZrO2 nanoparticles at
the injecting site. As a result, the amount of charge recombina-
tion is higher in ZrO2.
Acknowledgment. We are thankful to Dr. T. Mukherjee,
Associate Director, Chemistry Group, for his constant encour-
agement.
(g) Importance of Molecular Structure in Energy Conver-
sion Implications. In recent years, many researchers have been
trying hard to design and develop suitable low cost organic dyes,
which can be used in solar cells with higher efficiency.
Molecular structure of the organic dyes can be a very important
Supporting Information Available: Additional figures and
experimental details. This material is available free of charge
via the Internet at http://pubs.acs.org.
factor to get higher efficiency.4
5,24-26
In our previous studies
11
References and Notes
we have observed that a little structural change can increase
the injection efficiency in coumarin dye (D-1421 and C-343)
sensitized nanoparticles. Among them, excited D-1421 exists
in the TICT state, which eventually helps to inject electron
efficiently, because an electron localized at the keto group was
coupled with the nanoparticles. However, in the present
investigation we have observed that although the excited state
of DMACA exists as a TICT state, it cannot inject electrons
efficiently because its electron localized center is away from
the coupling site with the nanoparticles (Scheme 2). As we have
observed earlier,11 the TICT state of dye molecules can be of
help to increase the injection efficiency; on the other hand, it
can also diminish the same if the molecular structures of the
dye molecules are not designed properly.
(1) O’regan, B.; Gratzel, M. Nature 1991, 353, 737.
(2) Hagfeldt, A.; Gratzel, M. Chem. ReV. 1995, 95, 49.
(
3) Henglein, A. Chem. ReV. 1989, 89, 1861.
(4) Bawendi, M. G.; Steigerwald, M. L.; Brus, L. E. Annu. ReV. Phys.
Chem. 1990, 41, 477.
(5) Kamat, P. V. Chem. ReV. 1993, 93, 267.
(6) Alivisatos, A. P. J. Phys. Chem. 1996, 100, 13226.
(7) Fox, M. A. Top. Curr. Chem. 1987, 72, 142.
(8) Ghosh, H. N. J. Phys. Chem. B 1999, 103, 10382.
(9) Ramakrishna, G.; Ghosh, H. N. J. Phys. Chem. B 2001, 105, 7000.
(10) Ramakrishna, G.; Ghosh, H. N.; Singh, A. K.; Palit, D. K.; Mittal,
J. P. J. Phys. Chem. B 2001, 105, 12786.
(11) Ramakrishna, G.; Ghosh, H. N. J. Phys. Chem. A 2002, 106, 2545.
(12) Ramakrishna, G.; Singh, A. K.; Palit, D. K.; Ghosh, H. N. J. Phys.
Chem. B 2004, 108, 1701.
(
(
13) Ramakrishna, G.; Das, A.; Ghosh, H. N. Langmuir 2004, 20, 1430.
14) Ramakrishna, G.; Singh, A. K.; Palit, D. K.; Ghosh, H. N. J. Phys.
4
. Conclusion
Chem. B 2004, 108, 4775.
15) Ramakrishna, G.; Ghosh, H. N. Langmuir 2004, 20, ASAP.
(16) Ghosh, H. N.; Asbury, J. B.; Lian, T. J. Phys. Chem. B 1998, 102,
482.
17) Asbury, J. B.; Hao, E.; Wang, Y. Q.; Ghosh, H. N.; Lian, T. J.
Phys. Chem. B 2001, 105, 4545.
(18) Benko, G.; Hilgendroff, M.; Yartsev, A. P.; Sundstrom, V. J. Phys.
Chem. B 2001, 105, 967.
(
Sub-picosecond transient absorption spectroscopy has been
6
used to study the effect of molecular structure in photoinduced
electron-transfer dynamics in 7-N,N-dimethyl amino coumarin
4
(
-acetic acid (DMACA) and 7-hydroxy coumarin 4-acetic acid
(HCA) sensitized TiO2 and ZrO2 nanoparticles in aqueous
(
(
19) Dang, X.; Hupp, J. T. J. Am. Chem. Soc. 1999, 121, 8399.
20) Hannapel, T.; Burfeindt, B.; Storck, W.; Willig, F. J. Phys. Chem.
solution. The molecular structures of the studied dyes (DMACA
and HCA) are very similar with the same anchoring group which
binds with the nanoparticles except that HCA has a hydroxy
group at the 7-position while DMACA has a dimethyl amino
group at the 7-position. Electron injection has been confirmed
by direct detection of the conduction band electron in the 800-
B 1997, 101, 6799.
(21) Nazeeruddin, Md. K.; Pechy, P.; Gratzel, M. J. Chem. Soc., Chem.
Commun. 1997, 1705.
(
22) Argazzi, R.; Bignozzi, C. A.; Hasselmann, G. M.; Meyer, G. J.
Inorg. Chem. 1998, 37, 4533.
23) Nazeeruddin, Md. K.; Pechy, P.; Renourd, T.; Zakeeruddin, S. M.;
(
1
000 nm region and a cation radical in the visible region.
Humphry-Baker, R.; Comte, P.; Liska, P.; Cevey, L.; Costa, E.; Shklover,
V.; Spiccia, L.; Deacon, G. B.; Bignozzi, C. A.; Gratzel, M. J. Am. Chem.
Soc. 2001, 123, 1613.
Electron injection quantum yield is found to be 1.6 times higher
in HCA/TiO2 system as compared to that in DMACA/TiO2 in
similar condition. Solvatochromic measurements indicate that
the photoexcited HCA molecule exists in the ICT state; however,
(24) Hara, K.; Horiuchi, H.; Katoh, R.; Singh, L. P.; Sugihara, H.;
Sayama, K.; Murata, S.; Tachiya, M.; Arakawa, H. J. Phys. Chem. B 2002,
106, 374.

12496 J. Phys. Chem. B, Vol. 108, No. 33, 2004
Ramakrishna et al.
(
25) Hara, K.; Sugihara, H.; Singh, L. P.; Islam, A.; Katoh, R.; Yanagida,
M.; Sayama, K.; Murata, S.; Arakawa, H. J. Photochem. Photobiol., A 2001,
45, 117.
26) Hara, K.; Sugihara, H.; Tachibana, Y.; Islam, A.; Yanagida, M.;
(36) Rettig, W. Angew. Chem. Int. Ed. Engl. 1986, 25, 971.
37) Lippert, E.; Rettig, W.; Bonacic-Koutecky, V.; Heisel, F.; Miehe,
J. A. AdV. Chem. Phys. 1987, 68, 1.
38) Rettig, W. Electron-Transfer I; Mattay, J., Ed.; Topics in Current
(
1
(
(
Sayama, K.; Arakawa, H.; Fujihashi, G.; Horiguchi, T.; Kinoshita, T.
Langmuir 2001, 17, 5992.
Chemistry 169; Springer-Verlag: Berlin, 1994; p 254.
(
27) Asbury, J. B.; Wang, Y. Q.; Hao, E.; Ghosh, H. N.; Lian, T. Res.
Chem. Intermed. 2001, 27, 393.
28) Huber, R.; Sporlein, S.; Moser, J. E.; Gratzel, M.; Wachtveitl, J. J.
Phys. Chem B 2000, 104, 8995.
29) Hao, E.; Anderson, N. A.; Asbury, J. B.; Lian, T. J. Phys. Chem.
B 2002, 106, 10191.
30) Bahnemann, D.; Henglein, A.; Lilie, J.; Spanhel, L. J. Phys. Chem.
984, 88, 709.
(39) Grabowski, Z. R.; Rotkiewicz, K.; Rettig, W. Chem. ReV. 2003,
103, 3899.
(
(
40) Nad, S.; Kumbhakar, M.; Pal, H. J. Phys. Chem. A 2003, 107, 4808.
41) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265.
(
(
(42) Lu, H.; Prieskorn, J. N.; Hupp, J. T. J. Am. Chem. Soc. 1993, 115,
4927.
43) Dang, X.; Hupp, J. T. J. Am. Chem. Soc. 1999, 121, 8399.
44) Wang, Y.; Hang, K.; Anderson, N. A.; Lian, T. J. Phys. Chem. B
(
1
(
(
(
(
(
(
31) Ghosh, H. N.; Adhikari, S. Langmuir 2001, 17, 4129.
32) Ramakrishna, G.; Ghosh, H. N. Langmuir 2003, 19, 505.
33) Ramakrishna, G.; Ghosh, H. N. Supporting information.
34) Zhang, J. Z. Acc. Chem. Res. 1997, 30, 423.
(
2
003, 107, 9434.
(45) Hara, K.; Sato, T.; Katoh, R.; Furube, A.; Ohga, Y.; Shinpo, A.;
Suga, S.; Sayama, K.; Sugihara, H.; Arakawa, H. J. Phys. Chem. B 2003,
107, 597.
35) Colombo, D. P., Jr.; Bowman, R. M. J. Phys. Chem. 1996, 100,
1
8445.