Supramolecular Control of Charge-Transfer Dynamics on Dye-sensitized Nanocrystalline TiO2 Films

Article abstract of DOI:10.1002/chem.200305408

A [Ru(dcbpy)₂(NCS)₂ dye has been chemically modified by the addition of a secondary electron donor moiety, N,N-(di-p-anisylamino)phenoxymethyl. Optical excitation of the modified dye adsorbed to nanocrystalline TiO₂ films

Full text of DOI:10.1002/chem.200305408

J. R. Durrant, M. K. Nazeeruddin et al.
595
Chem. Eur. J. 2004, 10, 595 602
DOI: 10.1002/chem.200305408
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

FULL PAPER
Supramolecular Control of Charge-Transfer Dynamics on Dye-sensitized
Nanocrystalline TiO2 Films
[b]
Narukuni Hirata,[a] Jean-JacquesLagref, Emilio J. Palomares,[a] JamesR. Durrant,* [a]
M. Khaja Nazeeruddin,*[b] Michael Gratzel,[b] and Davide Di Censo[b]
Abstract: A [Ru(dcbpy)2(NCS)2] dye
has been chemically modified by the
addition of a secondary electron donor
cal calculations confirm that the
HOMO of the modified dye molecule
is localised on the electron donor
group. The retardation of the recombi-
nation dynamics relative to the un-
modified control dye is caused by the
increase in the spatial separation of the
HOMO orbital from the TiO2 surface.
The magnitude of the retardation is
shown to be in agreement with that
predicted from the non-adiabatic elec-
tron-tunnelling theory.
moiety,
N,N-(di-p-anisylamino)phen-
oxymethyl. Optical excitation of the
modified dye adsorbed to nanocrystal-
line TiO2 films shows a remarkably
long-lived charge-separated state, with
a decay half time of 0.7 s. Semiempiri-
Keywords: dyes/pigments ¥ electron
transfer ¥ ruthenium ¥ supramolec-
ular chemistry
Introduction
boxy-2,2’-bipyridine, TBA=tetrabutylammonium salt), ach-
ieving up to 9.18% solar to electric power conversion.[6] It is
necessary to achieve high rates of charge separation and col-
lection compared to interfacial charge-recombination pro-
cesses in order to achieve high device efficiencies. A
number of strategies are currently being developed to opti-
mise these dynamics, including the insertion of inorganic
barrier layers between TiO2 and the sensitiser dye,[7] the use
of alternative redox couples,[8] and the development of novel
sensitiser dyes.[9]On the other hand, the synthesis of supra-
molecular (multicomponent) molecules is attracting strong
interest due to their potential application in molecular elec-
tronic devices.[10,11] Such supramolecular structures provide a
potentially attractive approach to controlling the charge sep-
aration, and recombination dynamics at the molecule/nano-
crystal interfaces. In particular, by introducing a secondary
electron transfer function, or ™supersensitiser∫ into the
supramolecular complex, as illustrated in Scheme 1, it is pos-
sible to retard the interfacial charge-recombination dynam-
ics whilst retaining efficient light-induced charge separation.
This approach has previously been employed by using either
a RuII tris-bipyridine derivate to demonstrate the supersen-
sitiser function,[12] although the relatively weak red optical
absorbance of this dye relative to N719 precluded its practi-
cal device application, or by the use organic molecules such
as ™molecular spacers∫ linked to the redox chromophore.[13]
In this article, we will provide data showing that by appro-
priate design of a [RuL2(NCS)2] heteroleptic derivate it is
possible to achieve molecular control of the back electron-
transfer dynamics whilst maintaining good red-light absorp-
tion and efficient electron injection. This control is achieved
Light-driven electron-transfer processes between molecular
adsorbates, semiconductor nanomaterials and adsorbed
redox molecules have been a subject of intense research in
recent years.[1] Such electron-transfer processes are funda-
mental to many applications of semiconductor nanomateri-
als in photography,[2] photocatalytic degradation of pollu-
tants,[3] quantum dot devices[4] and solar energy conver-
sion.[5] In all of these applications, a key requirement is that
light absorption results in the efficient generation of a long-
lived charge-separated state. Of particular technological in-
terest are photoelectrochemical solar cells based on dye-sen-
sitised, nanocrystalline, mesoporous TiO2 films. These mole-
cules have emerged as potential low-cost, effective alterna-
tives to silicon-based devices. Efficient solar cells of this
type typically employ N719 dye (chemical name = bis(2,2’-
bipyridyl-4,4’-dicarboxylato)ruthenium(ii) bis-tetrabutylam-
monium sensitiser dye [(RuL2(NCS)2][TBA]2 L=4,4’-dicar-
[a] N. Hirata, Dr. E. J. Palomares, Dr. J. R. Durrant
Imperial College, Department of Chemistry
Centre for Electronic Materials and Devices
Exhibition Road, London SW7 2AY (UK)
E-mail: j.durrant@imperial.ac.uk
[b] Dr. J.-J. Lagref, Dr. M. K. Nazeeruddin, Prof. M. Gratzel,
D. Di Censo
Laboratory for Photonics and Interfaces
Institute of Molecular and Biological Chemistry
School of Basic Sciences, Swiss Federal Institute of Technology
1015 Lausanne (Switzerland)
E-mail: MdKhaja.Nazeeruddin@epfl.ch
596
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200305408
Chem. Eur. J. 2004, 10, 595 602

595 602
face and results in remarkably slow, interface limited recom-
bination dynamics.
Results and Discussion
Synthesis and characterisation of [Ru(L)(L’)(NCS)] hetero-
leptic dye: The details of the synthetic strategy adopted for
the preparation of heteroleptic complexes are shown in
Scheme 2. During the reaction the intermediate products
were drawn off and checked by UV-visible spectroscopy.
Reaction of dichloro(p-cymene)ruthenium(ii) dimer 1 with
4-[4-(N,N-di-p-anisylamino)phenoxymethyl]-4’-methyl-2,2’-
bipyridine (L’) in N,N’-dimethylformamide (DMF) solution
at 908C resulted in the mononuclear complex 2, [Ru
(L’)Cl(cymene)]Cl. In this step, the coordination of substi-
tuted bipyridine ligand to the ruthenium centre takes place
with cleavage of the doubly chloride-bridged structure of
the dimeric complex.
Heteroleptic dichloro complexes were prepared by react-
ing the mononuclear complex 2 with 4,4’-dicarboxy-2,2’-bi-
pyridine (L) under reduced light at 1508C. The displacement
of the cymene ligand from the coordination sphere of the
ruthenium metal by the 4,4’-dicarboxy-2,2’-bipyridine ligand
takes place efficiently in organic solvents such as DMF,
without scrambling. The absorption spectra of the resulting
complex 3 [Ru(L)(L’)(Cl)2] was measured in ethanol, which
Scheme 1. Schematic representation of electron-transfer (ET) processes
in dye-sensitised solar cells showing photogeneration of the dye excited
state: a) electron transfer from the secondary electron donor moeity,
b) electron injection into the semiconductor and c) the wasteful charge-
recombination pathway of the injected electron with the oxidised dye
molecules.
by ™supersensitising∫ the dye molecule by the addition of a
secondary electron-donor group, in this case N,N-(di-p-ani-
sylamino)phenoxymethyl(DAP) group. This increases the
physical separation of the dye cation from the electrode sur-
Scheme 2. Schematic of pathways for the synthesis of the N845 dye: a) 4(4-(N,N-di-p-anisolamino)phenoxymethyl)2,2-bpy, b) DMF, N2, 4,4’-dicarboxy-
2,2’-bpy, c) NH4NCS, N2, DMF.
597
Chem. Eur. J. 2004, 10, 595 602
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

J. R. Durrant, M. K. Nazeeruddin et al.
FULL PAPER
shows bands at 566, 414, 313 and 299 nm. The bands in the
visible region at 566 and 414 nm, and in the UV region at
313 and 299 nm, are due to metal-to-ligand charge-transfer
transitions (MLCT), and ligand-centered charge-transfer
transitions, respectively. The absorption spectra of the
[Ru(L)(L’)(Cl)2] complex is identical to that of the complex
prepared by means of a [RuCl2(DMSO)4] synthetic meth-
od.[14]The [Ru(L)(L’)(Cl)2] complex 3 was treated with a 30-
fold excess of ammonium thiocyanate ligand to obtain the
[Ru(L)(L’)(NCS)2] complex (N845). The UV-visible absorp-
tion spectra of N845 show bands at 300, 312, 400, and
535 nm (Figure 1). By comparing with the homoleptic com-
Figure 2. Emission spectra (lex =530 nm) of the complexes N719 (black
line), and N845 (grey line) at 298 K in DMF.
cyanate ligand. The ATR-FTIR spectra of N845 adsorbed
onto TiO2 films shows an absence of carboxylic acid bands
indicating that the two carboxylic acid groups are dissociat-
ed on the surface. The adsorbed dye shows bands at
1611 cmꢀ1 (-COOꢀas), and 1348 cmꢀ1 (-COOꢀ ) due to asym-
s
metric (as) and symmetric (s) vibrations of the carboxylate
groups, confirming the chelating mode of adsorption be-
tween the complex and the TiO2.
Semiempirical calculations: For the optimised geometry of
the RuII polypyridyl dyes, we assumed a linkage through the
two carboxyl groups (-COOH) to the Ti4+ surface, in agree-
ment with the ATR-FTIR data of the adsorbed dye. The
semiempirical calculations were performed on the cation
state of both dyes.
Figure 1. UV/Vis absorption spectra of the complexes N719 (black line),
and N845 (grey line) in DMF solution.
plex N719, the UV bands at 299, and 312 nm are assigned to
the ligand-centered charge-transfer transitions of the L’, and
L ligands, respectively. The absorption spectral data for both
complexes are shown in Table 1. When N845 is excited
within the MLCT absorption band at 298 K in an air-equili-
brated DMF solution, it exhibits a luminescence maximum
at 810 nm. The emission spectral profile is independent of
the excitation wavelength. For comparison the N719 emis-
sion spectra is included in Figure 2, which is blue-shifted by
40 nm and the emission intensity is significantly higher than
the N845 complex.
The ATR-FTIR spectra of N845 was measured as a solid,
and in the adsorbed form onto TiO2 films. The solid sample
of the complex N845 shows a strong broad band at
1705 cmꢀ1 due to carboxylic acid group. The intense peak at
1224 cmꢀ1 is assigned to the u(CꢀO) stretch. The other char-
acteristic band at 2095 cmꢀ1 is assigned to u(NC) of the thio-
We first considered the control dye (N719). Figure 3
shows a comparison between the optimised geometry of the
cation state determined by the semi-empirical ZINDO/1 cal-
culations (see Experimental Methods Section), against the
X-ray structure of the fully protonated analogue
[Ru(L)2(NCS)2] of this dye reported previously.[15] The cal-
culated RuꢀN(bpy) and RuꢀN(NCS) bond lengths, as sum-
marised in Table 2, are within 1% of those obtained from
crystallographic data, supporting the validity of our theoreti-
cal calculations. These data also indicate that the removal of
one electron from the complexes does not induce dramatic
geometric changes. Furthermore these results are in good
agreement with DFT calculations reported previously,[15,16]
providing further support for the validity of our semiempiri-
cal calculations.
Table 1. Absorption and electrochemical properties of the N845 and N719 complexes.
1
1
1
1
=
2
=
=
2
=
2
Abs. max. [nm][a]
Eox
Eox
Ered
Ered
2
Complex
(e[104 mꢀ1 cmꢀ1])
[V vs Fc][b]
L’+/L’
[V vs. Fc]
[V vs. Fc]
[V vs. Fc]
L’(p p*)
L(p p*)
4d p*
RuIII/II
L/Lꢀ
L’/L’ꢀ1
N845
N719
300(5.52)
312 (4.41)
312 (4.91)
372 (1.05), 535 (1.1)
396 (1.43), 535 (1.47)
0.200
0.34
0.35(irr)
ꢀ2.13
ꢀ2.21
ꢀ2.40
[a] l max and e values are ꢁ2 nm and ꢁ10%, respectively. [b] The ferrocene/ferrocinium complex used as a reference, which is +640 mV vs SHE.
598
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 595 602

Dye-Sensitised TiO2 Films
595 602
Figure 4. Graphical representation of the HOMOs of a) N719 and
b) N845 complexes. Atoms in turquoise, red, blue and pale yellow corre-
spond to carbon, oxygen, nitrogen and sulfur, respectively.
Cyclic voltammograms of the complex N845 in DMF re-
vealed two reversible, one-electron oxidation processes at
Figure 3. Comparison of bond (black) distances and angles (grey) calcu-
lated using ZINDO/1 for the cation state of N719 with the corresponding
experimental X-ray data.[27] The values in parentheses correspond to the
X-ray data.
E
1/2 =0.200 and 0.34 V versus ferrocene/ferrocinium couple
(Fc/Fc+). The 0.34 V couple, which is also observed for the
N719 control dye, is assigned to the RuII/III couple. The
0.20 V couple is therefore assigned to the oxidation of
donor N,N-(di-p-anisylamino)phenoxymethyl group, consis-
tent with the electron donating function of this group. In ad-
dition, two quasi-reversible reduction peaks are observed
for N845 at ꢀ2.13 and ꢀ2.4 V versus Fc/Fc+. Comparison
with the reduction peak observed for the control N719 dye
allows us to assign these reduction peaks for N845 to the re-
duction of L and L’ ligands, respectively. These data there-
fore indicate that the LUMO of the L’ ligand is higher than
that of the L ligand. Optical excitation of the N845 MLCT
state is therefore expected to result in electron localisation
on the 4,4’-dicarboxy-2,2’-bipyridine (L) ligand, thereby fa-
vouring efficient electron injection into the TiO2 electrode.
Table 2. Comparison of bond lengths [ä] and angles [8] of crystallo-
graphic data for cis-[Ru(L)2( NCS)2] with ZINDO/1-optimised geome-
tries for the cation states of N719 and N845.
cis-[Ru(L)2(NCS)2][c]
N719[d]
N845[d]
Ruꢀ-NCS
2.025
2.065
91.9
81.3
90.6
1.999
2.042
91.6
76.21
89.98
1.97
1.99
87.56
75.98
87.58
RuꢀN(bpy)
NCS-Ru-NCS[a]
N(bpy)-Ru-N(bpy)
(bpy)-(bpy)[b]
[a] NCS-Ru-NCS refers to the bpy ligands cis to the NCS ligands. [b] Di-
hedral angle between the mean planes of bipyridine ligands. [c] Crystallo-
graphic data from ref. [27]. [d] Data from semiempirical calculations.
The important distances and angles for the optimised
geometries of both dyes, N719 and N845, in their cation
states obtained from the ZINDO/1 calculations are com-
pared in Table 2. These data show that the Ru metal co-or-
dination centres of both dyes are expected to be very simi-
lar, as seen by the close match of their optical absorbance
spectra as detailed in Table 1. The orbital profiles for the
HOMO×s of both dyes are presented in Figure 4. The
HOMO of the N719 has a larger amplitude on NCS ligands,
(similar results were obtained by Rensmo et al.)[17] and within
the NCS ligands the amplitude is more on the sulfur atom.
The HOMO for the N845 is shifted to the secondary electron
donor moiety, displacing the positive charge (hole) in the
oxidised form of N845 from the NCS ligands closer to the
triphenyl amine group. This results in an increased separa-
tion of the HOMO orbital from the TiO2 surface, as desired
in order to achieve slower charge-recombination dynamics.
Transient absorption and luminescence measurements: The
primary function of the N,N-(di-p-anisylamino)phenoxy-
methyl (DAP) group is to accept the positive charge (hole)
after electron injection from the dye-excited state into the
conduction band of the nanocrystalline TiO2. This results in
an increase in the physical separation of the charged species.
To assay this function, we employed transient absorption
spectroscopy to measure the charge-recombination kinetics
(reaction c, Scheme 1). The decay of the induced absorption
of the N719 and N845 cations was monitored following
pulsed laser excitation of the dye-sensitised TiO2 films, in
the absence of any applied bias or redox-active electrolyte.
Typical data for the charge-recombination process is shown
in Figure 5. It is apparent that the decay dynamics of the
N845 dye cation are retarded 1000-fold relative to that of
the N719 cation, with decay half times, t50%, of 0.71 s and
0.85 ms, respectively. This retardation cannot be attributed
to differences in oxidation potential of the two dyes, as the
recombination reaction is thought to occur in the Marcus
™inverted region∫.[18] This is where the smaller reaction free
energy of the N845 dye would be excepted to result in accel-
Electrochemistry: Cyclic voltammetry was used to investi-
gate the electrochemistry of the N719 and N845 dyes. Mid-
point potentials of these data are summarised in Table 1.
599
Chem. Eur. J. 2004, 10, 595 602
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

J. R. Durrant, M. K. Nazeeruddin et al.
FULL PAPER
Figure 6. A) Transient absorption specta for TiO2 films sensitised with
N845 monitored at 8 ms after excitation at 516 nm and B) transient ab-
sorption of 4-(diphenylmaino) benzaldehyde (optical absorbance l410 nm
0.35) monitored at 8 ms after excitation at 400 nm.
=
small spatial separation of this cation from the electrode sur-
face.
Figure 5. Transient absorption data monitoring charge-recombination dy-
namics for TiO2 films sensitised with i) N719 and ii) N845. Both figures
show the same data with a) being a linear/log plot and b) being log/linear
plot. The grey lines are the fits to the decay kinetics corresponding to i) a
streched exponential (DO.D. /exp(ꢀ(t/t)a), a=0.44) and ii) a monoexpo-
nenetial decay. The signal size has been normalised for comparisson pur-
poses. Data collected at a probe wavelength of 800 nm for N719 and
850 nm for N845, following optical excitation at 550 nm.
The initial amplitudes of the dye cation transient absorp-
tion signals observed for both the N845 and N719 sensitised
TiO2 were of similar magnitudes (DODs of 1.4 at 700 nm
and 1 at 800 nm, respectively). This observation strongly
suggests that, as for N719, the optical excitation of the N845
adsorbed to the TiO2 films results in efficient electron injec-
tion into TiO2 conduction band. However, uncertainty over
the N845 cation extinction coefficient prevents quantitative
analysis from these data alone. Transient luminescence stud-
ies, as shown in Figure 7, further indicate efficient electron
erated recombination dynamics, contrary to the experimen-
tal observation. Instead the retardation is attributed to an
increased distance of the dye cation from the electrode sur-
face due to localisation of the N845 cation on the DAP
moiety, consistent with the semiempirical calculations and
electrochemical data detailed above.
Localisation of the N845 dye cation on the DAP moiety is
further supported by the transient absorption spectrum of
the dye cation. Figure 6 compares the transient absorption
spectrum of the N845 dye cation generated following photo-
excitation of the N845 dye adsorbed to the TiO2 electrode
(Figure 6a) with the transient spectrum observed for the 4-
(diphenylamino) benzaldehyde, closely analogous to the
DAP moiety alone (see inset to Figure 6b), adsorbed direct-
ly to the electrode (Figure 6b). In both cases, a photoin-
duced absorption maximum is observed at ~700 nm, as-
signed to an absorption maximum of the DAP cation, con-
sistent with previous observations,[19] and confirming the lo-
calisation of the N845 cation on the DAP moiety. We note
that in the case of Figure 6b, the photoinduced absorption
exhibits rapid kinetics (8 ms), consistent with the expected
Figure 7. Time-resolved single-photon-counting decays of emission from
N845 sensitised on a) ZrO2 and b) TiO2; c) corresponds to the instrument
response pulse.
600
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 595 602

Dye-Sensitised TiO2 Films
595 602
injection for the N845 sensitised TiO2 films. These compari-
sons show the transient luminescence decay of N845 sensi-
tised TiO2 and ZrO2 films, normalised to the same number
of absorbed photons. ZrO2 film was employed as a control
film, as the relatively high conduction band edge of this ma-
terial prevents electron injection, consistent with the long
excited state lifetime observed for dyes adsorbed to this
film.[20] It is apparent that relative to the control ZrO2 films,
the N845 luminescence is strongly quenched on the TiO2
film, consistent with efficient electron injection.
It can be concluded that this heterosupramolecular ap-
proach translates the oxidising equivalent (hole) away from
the nanostructured semiconducting surface. The increase in
the distance between the surface and the hole, results in an
extremely long-lived charge-separated pair. The N,N-(di-p-
anisylamino)phenoxymethyl
secondary
electron-donor
group employed in this study can be expected to interface
well with triarylamine-based hole conductors, making this
sensitiser dye particularly attractive for applications in solid
state dye-sensitised solar cells.
Electron transfer dynamics typically exhibit an exponen-
tial dependence upon spatial separation.[Eq. (1)]
Experimental Section
ket / eꢀbr
ð1Þ
Materials: The dichloro(p-cymene)ruthenium(ii) dimer and potassium
thiocyanate were obtained from Aldrich and used as received. The 4-(di-
phenylamino)benzaldehyde was purchased from Fluka. The HPLC grade
solvents from Fluka were used without purification. 4,4’-Dimethyl-2,2’-bi-
pyridine ligand (from Fluka) was dried under vacuum (9 mbar) at 408C
for 4 h before use. Water content was measured by Karl-Fisher automatic
titration using the coulometer Metrohm 684 KF. Silicagel 60 from Fluka
was used for chromatography as a stationary phase. Tris(p-anisyl)amine
and 4-(N,N-di-p-anisylamino)phenol were synthesised by using the proce-
dure developed by BonhÙte et al.[23]
In which r is the edge-to-edge separation of the donor
and acceptor wave-functions, and b is a measure of the bar-
rier height of the intervening medium. Typically b varies
from 2.8 äꢀ1 for a vacuum to 1 äꢀ1 for saturated-electron-
orbital densities. From consideration of the HOMO orbital
distributions shown in Figure 4, we estimate the dye-cation/
electrode-surface distance is increased by 4 ä for N845 rela-
tive to N719. Comparison with the ~1000 fold retardation
of the observed recombination dynamics yields a value for b
of 1.6 ꢁ 0.2 äꢀ1, typical of b values observed for through
solvent electron transfer.[21] It can thus be concluded that
the retardation of the recombination dynamics shown in
Figure 5 is in good quantitative agreement with that expect-
ed from the increased dye-cation/electrode separation deter-
mined from the HOMO orbital calculations.
The preparation of anatase nanocrystalline TiO2 films (average particle
diameter: 15 nm and film thickness 4 mm) and sensitisation of these films
were conducted as described previously.[24]
Electrochemistry experiments: Electrochemical data were obtained by
cyclic voltammetry using a three-electrode cell and an Auto lab System
(PGSTAT 30, GPES 4.8 software). The working electrode was a 0.03 cm2
gold disk, the auxiliary electrode was a glassy carbon rod and a silver
wire was used as quasi-reference electrode. Tetrabutyl ammonium per-
chlorate (TBAP) 0.1m was used as supporting electrolyte in DMF. Ferro-
cene was added to each sample solution at the end of the experiments
and the ferrocenium/ferrocene redox couple was used as an internal po-
tential reference.
In addition to the retardation of the recombination dy-
namics shown in Figure 5, it is apparent that the temporal
recombination shape differs significantly between these two
dyes. For dye N719, decay kinetics strongly suggest a
stretched exponential (DO.D. ~exp(ꢀ(t/t)a); a=0.44). Such
stretched exponential recombination dynamics are typical of
those we have observed previously for a range of different
dyes under similar experimental conditions. In contrast, for
dye N845 the recombination dynamics suggests an excellent
fit to a near monoexponential decay (DO.D. ~exp(ꢀ(t/t)a;
a ~0.9) as illustrated by the log/linear plot shown in Fig-
ure 5b.
We have previously demonstrated that the stretched expo-
nential recombination dynamics, as observed here for the
nanocrystalline TiO2 films sensitised by N719, are consistent
with the charge-recombination dynamics being rate limited
by electron transport within the metal oxide film.[22] The
stretched exponential nature of these kinetics have been at-
tributed to an inhomogeneous distribution of trap depths in
the film, resulting in a broad distribution of detrapping (and
therefore recombination) times. Our observation that N845
exhibits recombination dynamics three orders of magnitude
slower than N719 in itself suggests that for this sensitiser, in-
terfacial electron transfer is critical to determining the over-
all recombination dynamics. The near monoexponential be-
haviour is indicative of the recombination dynamics being
limited by the interfacial electron dynamics, as we have dis-
cussed previously in relation to experimental data obtained
with two porphyrin dyes.
Spectroscopy: UV/VIS and fluorescence spectra were recorded in a 1 cm
pathlength quartz cell on a Cary 5 spectrophotometer and Spex Fluoro-
log 112 Spectroflurimeter, respectively. The details of the transient ab-
sorption spectrometer system for measuring the charge-recombination
dynamics are given elsewhere.[25] Transient absorption data were collect-
ed with the films covered in ethylene carbonate/propylene carbonate
50:50 v/v. The excitation wavelength was 550 nm for both dyes (20 mJ).
The dye volume loaded was selected to achieve matched optical densities
of 0.8 absorption units at this wavelength.
Transient emission data were collected using a 635 nm excitation from a
IBH NanoLED-06 pulsed laser diode [repetition rate one-shot to 1 MHz,
pulse duration (1.3 ns), intensity (~1 mW)]. Emission was collected at
750 nm (bandwidth 50 nm) at a microchannel plate, with an instrument
response of 350 ps. Data were collected at matched data collection times
of 2500 s.
1H and 13C NMR spectra were measured with a Bruker ACP-200 spec-
trometer at 200 MHz and 50.3 MHz, respectively. The reported chemical
shifts were against TMS. FAB-MS spectra were obtained at the Universi-
ty of Lausanne from a nitrobenzyl alcohol matrix. The ATR-FTIR spec-
tra for all the samples were measured using a Digilab 7000 FTIR spec-
trometer. The FTIR data reported here were taken with the ™Golden
Gate∫ diamond anvil ATR accessory (Graseby-Specac) using typically
64 scans at a resolution of 2 cmꢀ1. The samples were all measured under
the same mechanical force pushing the samples in contact with the dia-
mond window. No ATR correction was applied to the data.
Synthesis and characterisation4-(Bromomethyl)-4’-methyl-2,2’-bipyridine
(L): This ligand was synthesised with a slight modification of Meyer×s
procedure.[26] In a 100 mL round-bottomed flask 4,4’-dimethyl-2,2’-bipyri-
dine (1.8 g, 9.78 mmol), N-Bromosuccinimide (1.8 g, 10.1 mmol) and azo-
bis(isobutyronitrile) (50 mg) were added to freshly distilled CCl4
(40 mL). The mixture was stirred magnetically and heated to reflux for
601
Chem. Eur. J. 2004, 10, 595 602
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

J. R. Durrant, M. K. Nazeeruddin et al.
FULL PAPER
3 h in the dark. The mixture was cooled to room temperature and the re-
sulting white precipitate filtered using a crucible. The filtrate was evapo-
rated under vacuum and dissolved in CH2Cl2 (10 mL) . Silica was added
to the solution and then the solvent was evaporated. The SiO2 with adsor-
bed product was charged onto a chromatography column (internal diam-
eter 15 mm, length 300 mm) containing silica, which was pre-treated with
CH2Cl2/NH4OH 99.5:0.5.4-(Bromomethyl)-4’-methyl-2,2’-bipyridine was
obtained by eluting with CH2Cl2/hexane 80:20 solvent mixture. The solu-
tion was evaporated under vacuum and a yellowish sticky oil was ob-
tained at a yield of 40%. 1H NMR (CDCl3, 298 K):d= 2.46 (s, 3H), 4.50
(s, 2H), 7.16 (dd, J=4.4 Hz, 1H), 7.34 (dd, J=4 Hz, 1H), 8.25 (d, J=
1.6 Hz, 1H), 8.43 (d, J=1 Hz, 1H), 8.56 (d, J=5 Hz, 1H), 8.68 ppm (d,
J=5.4 Hz, 1H).
Acknowledgement
We gratefully acknowledge the financial support from the EPSRC, the
European Union (Contract Number: ENK6-CT-2001 00560 Nanomax),
the Swiss Federal Office for Energy (OFEN) and US Air Force Research
Office under contract number F61775 00-C0003. E.P. is very grateful for
the support of a Marie Curie Fellowship and N.H. for the support of the
Bridgestone Corporation. Supply of the N719 dye from Johnson Matthey
is also gratefully acknowledged.
[1] C. A. Koval, J. N. Howard, Chem. Rev. 1992, 92, 411.
[2] I. R. Gould, J. R. Lenhard, A. A. Muenter, S. A. Godleski, S. Farid,
Pure Appl. Chem. 2001, 73, 455.
4-[4-(N,N-Di-p-anisylamino)phenoxymethyl]-4’-methyl-2,2’-bipyridine
(L’): In a dry 25 mL round bottom flask, under argon and at 293 K, 4-
(N,N-di-p-anisylamino)phenol (294 mg, 1.2 equiv, 0.91 mmol) and potassi-
um tert-butoxide (114 mg, 1.5 equiv, 1.18 mmol) were mixed with dry
THF (6 mL). The mixture was magnetically stirred for 20 min, which rap-
idly turned green. Then, 4-(bromomethyl)-4’-methyl-2,2’-bipyridine
(200 mg, 0.76 mmol) was dissolved in dry THF (8 mL) and added with a
syringe. A white precipitate of KBr formed progressively. After 35 h the
reaction was quenched with water (1 mL). The solvent was removed in a
rota-evaporator and tert-butyl methyl ether (TBME) (50 mL) and water
(40 mL) was added. The vial was shaken, and the organic phase was al-
lowed to separate. The aqueous phase was washed twice with TBME
(20 mL). The collected organic phase was evaporated under vacuum to
afford a yellow-green oil. The waxy solid was purified by chromatogra-
phy (SiO2, solid deposition, internal diameter 13 mm, high 200 mm) with
silica gel previously treated with hexane/triethylamine 99:1. A linear evo-
lution of the eluent composition (over a volume of 250 mL) from pure
hexane to hexane/tert-butyl methyl ether 50:50 was used. Fractions con-
taining the desired product were evaporated under vacuum to afford a
colourless oil, which was dissolved in DMF under argon and kept at 08C
to avoid oxidation by air. Air oxidation was characterised by a strong
[3] M. A. Fox, M. T. Dulay, Chem. Rev. 1993, 93, 341.
[4] A. Meier, D. C. Selmarten, K. Siemoneit, B. B. Smith, A. J. Nozik, J.
Phys. Chem. B 1999, 103, 2122.
[5] B. O×Reagan, M. Gratzel, Nature 1991, 353, 737.
[6] M. K. Nazeeruddin, R. Splivallo, P. Liska, P. Comte, M. Gratzel,
Chem. Commun. 2003, 1456.
[7] E. Palomares, J. N. Clifford, S. A. Haque, T. Lutz, J. R. Durrant, J.
Am. Chem. Soc. 2003, 125, 475.
[8] H. Nusbaumer, J.-E. Moser, S. M. Zakeeruddin, M. K. Nazeeruddin,
M. Graetzel, J. Phys. Chem. B 2001, 105, 10461.
[9] M. Yanagida, A. Islam, Y. Tachibana, G. Fujihashi, R. Katoh, H. Su-
gihara, H. Arakawa, New J. Chem. 2002, 26, 963.
[10] X. Marguerettaz, D. Fitzmaurice, J. Am. Chem. Soc. 1994, 116, 5017.
[11] X. Marguerettaz, R. O×Neill, D. Fitzmaurice, J. Am. Chem. Soc.
1994, 116, 2629.
[12] R. Argazzi, C. A. Bignozzi, T. A. Heimer, F. N. Castellano, G. J.
Meyer, J. Am. Chem. Soc. 1995, 117, 11815.
[13] P. G. Hoertz, R. A. Carlisle, G. J. Meyer, D. Wang, P. Piotrowiak, E.
Galoppini, Nano Lett. 2003, 3, 325.
[14] S. M. Zakeeruddin, M. K. Nazeeruddin, R. Humphry-Baker, P.
Pechy, P. Quagliotto, C. Barolo, G. Viscardi, M. Gratzel, Langmuir
2002, 18, 952.
[15] K. Srikanth, V. R. Marathe, M. K. Mishra, Int. J. Quantum. Chem.
2002, 89, 535.
[16] M. K. Nazeeruddin, S. M. Zakeeruddin, R. Humphry-Baker, S. I.
Gorelsky, A. B. P. Lever, M. Gratzel, Coord. Chem. Rev. 2000, 208,
213.
purple colour. Yield
=
200 mg, 52%. 1H NMR (CDCl3, 298 K,
200 MHz):d= 2.46 (s, 3H), 3.79 (s, 6H), 5.14 (s, 2H), 6.77 7.01 (m,
12H), 7.12 (d, J=7.6 Hz, 1H), 7.48 (d, J=3.8 Hz, 1H), 8.26 (s, 1H), 8.44
(s, 1H), 8.56 (d, J=5.03 Hz, 1H), 8.70 (d, J=5.05 Hz, 1H); 13C NMR
(CDCl3, 298 K): d=21.2, 55.4, 68.8, 114.53, 115.5, 119.1, 121.6, 122.2,
124.2, 124.9, 125.1, 134.0, 148.5, 149.4, 154.9, 155.0 ppm; Mass spectra
(FAB): calcd 503.59, measured 504.6 [MH+] (100%).
Synthesis of [Ru(L)(L’)(NCS)2] (3): [{RuCl(p-cymene)}2](0.115 g,
0.188 mmol) was dissolved in DMF (30 mL)and L’(0.19 g, 0.376 mmol)-
was added. The reaction mixture was heated at 808C under nitrogen for
4 h and then, L(0.0917 g, 0.376 mmol)was added . The reaction mixture
was refluxed at 1608C for another 4 h under reduced light to avoid light-
induced cis to trans isomerisation. Then an excess of NH4NCS (13 mmol)
was added to the reaction mixture and heated at 1308C for a further 5 h.
[17] H. Rensmo, S. Lunell, H. Siegbahn, J. Photochem. Photobiol. A
1998, 144, 117.
[18] S. Fukuzumi, Org. Biomol. Chem. 2003, 1, 609.
[19] A. C. Bhasikuttan, A. K. Singh, D. K. Palit, A. V. Sapre, J. P. Mittal,
J. Phys. Chem. A 1999, 103, 4703.
[20] Y. Tachibana, J.-E. Moser, M. Graetzel, D. R. Klug, J. R. Durrant, J.
Phys. Chem. 1996, 100, 20056.
[21] D. S. Bendall, Protein Electron Transfer, BIOS Scientific, Oxford,
1996.
[22] J. N. Clifford, G. Yahioglu, L. R. Milgrom, J. R. Durrant, Chem.
Commun. 2002, 126.
[23] P. Bonhote, J.-E. Moser, R. Humphry-Baker, N. Vlachopoulos, S. M.
Zakeeruddin, L. Walder, M. Gratzel, J. Am. Chem. Soc. 1999, 121,
1324.
[24] Y. Tachibana, S. A. Haque, I. P. Mercer, J. R. Durrant, D. R. Klug, J.
Phys. Chem. B 2000, 104, 538.
[25] S. A. Haque, Y. Tachibana, D. R. Klug, J. R. Durrant, J. Phys. Chem.
B 1998, 102, 1745.
[26] S. Gould, G. F. Strouse, T. J. Meyer, B. P. Sullivan, Inorg. Chem.
1991, 30, 2942.
[27] V. Shklover, Y. E. Ovchinnikov, L. S. Bragnisky, S. M. Zakeeruddin,
M. Graetzel, Chem. Mater. 1998, 10, 2533.
The solvent was removed using
a rotary-evaporator under vacuum.
Water was added to the resulting semisolid to remove excess NH4NCS.
The water-insoluble product was collected on a sintered glass crucible by
suction filtration and washed with distilled water, followed by diethyl
ether and dried. The crude complex was dissolved in a solution of tetra-
butyl ammonium hydroxide (0.4 g) in methanol (4 mL). The concentrated
solution was charged onto a Sephadex LH-20 (2î30 cm) column and
eluted with methanol. The main red band was collected and concentrated
to 3 mL. The required complex was isolated upon addition of a few drops
1
of 0.01m HNO3. Yield = 0.18 g, 50%. H NMR ([D4]CD3OD, 200 MHz):
d=2.78 (s, 3H), 3.79 (s, 6H), 5.42 (s, 2H), 6.79 7.03 (m), 7.38 (d, 1H),
7.65 (d, 1H), 7.82 (d, 1H), 7.85 (d, 1H), 7.96 (d, 1H), 8.3 (d, 1H). 8.38 (s,
1H), 8.5 (s, 1H), 8.89 (s, 1H), 9.02 (s, 1H), 9.51 (d, 1H), 9.54 ppm (d,
1H).
Semiempirical chemical calculations: Molecular orbitals and optimised
geometries for free molecules were calculated using Hyperchem 7 pro-
gram package running on a PC AMD Athlon XP-2000+ 1.25 GHz.
Geometrical optimisations were performed with the ZINDO/1 parameter
set. The overlap weighting factors s s and p p were set at 1.265 and
0.585, respectively. The number of singly excited configurations used was
1250 (25î25 occupiedîvirtual orbitals).
¾
Received: July 29, 2003
Revised: October 16, 2003 [F5408]
602
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 595 602