Reductive electron transfer quenching of MLCT excited states bound to nanostructured metal oxide thin films

Article abstract of DOI:10.1021/jp026823n

The ruthenium compounds Ru(deeb)(bpz)₂(PF₆)₂, Ru(deeb)₂(bpz)(PF₆)₂, and Ru(deeb)₂(dpp)(PF₆)₂, where deeb is 4,4a?2-(CO₂CH₂CH₃)₂-2,2a?2-b ipyridine, bpz is 2,2a?2-bipyrazine, and dpp is 2,3-bis(2-pyridyl)pyrazine, have been prepared, characterized, and anchored to mesoporous nanoparticle thin films comprised of the wide band gap semiconductor TiO₂ or the insulator ZrO₂. The metal-to-ligand charge-transfer (MLCT) excited states of these compounds are potent photooxidants (E^o(Ru^II*/+) > +1.0 V vs SCE) with long lifetimes (?? > 1 ??s) that efficiently oxidize iodide and phenothiazine with rate constants that approach the diffusion limit in acetonitrile. Photogalvanic cells based on the sensitized TiO₂ materials yield photocurrent action spectra that agree well with the Ru(II) absorptance spectra. The photocurrent efficiency was very low, ?? < 10^-4. Transient absorption data show that neither the excited nor the reduced state of the ruthenium compounds efficiently inject electrons into the TiO₂ particles. The cage escape yields following excited-state electron transfer are approximately 2/3 lower in the mesoporous thin films than in fluid solution. Intermolecular energy transfer across the nanoparticle surfaces is manifest in a second-order component to the excited-state relaxation kinetics.

Full text of DOI:10.1021/jp026823n

J. Phys. Chem. B 2003, 107, 245-254
245
Reductive Electron Transfer Quenching of MLCT Excited States Bound To Nanostructured
Metal Oxide Thin Films
Bryan V. Bergeron and Gerald J. Meyer*
Department of Chemistry and Department of Materials Science and Engineering, Johns Hopkins UniVersity,
3400 North Charles Street, Baltimore, Maryland 21218
ReceiVed: August 23, 2002; In Final Form: October 28, 2002
The ruthenium compounds Ru(deeb)(bpz)2(PF6)2, Ru(deeb)2(bpz)(PF6)2, and Ru(deeb)2(dpp)(PF6)2, where deeb
is 4,4′-(CO2CH2CH3)2-2,2′-bipyridine, bpz is 2,2′-bipyrazine, and dpp is 2,3-bis(2-pyridyl)pyrazine, have been
prepared, characterized, and anchored to mesoporous nanoparticle thin films comprised of the wide band gap
semiconductor TiO2 or the insulator ZrO2. The metal-to-ligand charge-transfer (MLCT) excited states of these
compounds are potent photooxidants (E°(RuII*/+) > +1.0 V vs SCE) with long lifetimes (τ > 1 µs) that
efficiently oxidize iodide and phenothiazine with rate constants that approach the diffusion limit in acetonitrile.
Photogalvanic cells based on the sensitized TiO2 materials yield photocurrent action spectra that agree well
with the Ru(II) absorptance spectra. The photocurrent efficiency was very low, φ < 10-4. Transient absorption
data show that neither the excited nor the reduced state of the ruthenium compounds efficiently inject electrons
into the TiO2 particles. The cage escape yields following excited-state electron transfer are approximately 2/3
lower in the mesoporous thin films than in fluid solution. Intermolecular energy transfer across the nanoparticle
surfaces is manifest in a second-order component to the excited-state relaxation kinetics.
Introduction
erative cells, D+ is reduced at a dark counter electrode and
electrical power is generated with no net chemistry.1-3
The vast majority of early dye sensitization studies were
performed on planar electrodes and relied strictly on photo-
electrochemical measurements from which it was not possible
to unambiguously distinguish between these two mechanisms.4
More recently, dye sensitization studies have focused on
mesoporous thin films of nanocrystalline (anatase) TiO2. These
materials have an enormous surface area, a long effective path
length, and a high photoconductivity that affords both spectro-
scopic and photoelectrochemical characterization that can firmly
establish sensitization mechanisms.5 The observation of ultrafast
electron injection6 coupled with the weak oxidizing power of
the excited sensitizers currently in use,7 strongly suggest an
excited-state injection mechanism is operative in regenerative
solar cells based on these materials.5
Strong spectroscopic evidence for the reduced sensitizer
injection mechanism with sensitized mesoporous thin films was
recently reported.8 Interfacial electron transfer was shown to
be rate-limited by reductive quenching of the sensitizer excited
state, eq 2. The sensitizer used in this work was Ru(deeb)(bpy)2-
(PF6)2, where deeb is 4,4′-(CO2CH2CH3)2-2,2′-bipyridine and
bpy is 2,2′-bipyridine, and the electron donor was phenothiazine,
PTZ. A drawback of the PTZ donors is that they produce
negligible photocurrents in the dye-sensitized nanocrystalline
solar cells.9 Iodide is the sole electron donor identified that yields
solar energy conversion efficiencies >10% under one sun of
AM 1.5 solar irradiation.10 Ru(deeb)(bpy)22+* and other Ru(II)
bipyridyl excited states are weak photooxidants that do not
efficiently oxidize iodide.8
Sensitization of semiconductor electrodes to visible light with
molecular compounds represents an attractive approach to solar
energy conversion.1-3 In photoelectrochemical cells, selective
light excitation of molecular chromophores, termed sensitizers,
S, proximate to n-type semiconductor surfaces often yields
anodic photocurrents. Two different interfacial electron-transfer
mechanisms have been identified by which this photocurrent
can be generated and sustained, Scheme 1. In one, termed
excited-state dye sensitization, a photoexcited sensitizer, S*,
injects an electron into the semiconductor, SC, and the sensitizer
ground state is then regenerated by an electron donor, D, eq 1.
SC/S* + D 9 9
1.8 SC(e-)/S+ + D 2.8 SC(e-)/S + D+ (1)
In the other mechanism, a donor present in the electrolyte
reductively quenches the excited sensitizer and the reduced state,
S-, transfers an electron into the semiconductor, eq 2. Solar
cells based on this mechanism are often referred to as photo-
galvanic cells.3
2.
SC/S* + D
9
1.8 SC/S- + D+ 98 SC(e-)/S + D+ (2)
In both mechanisms, [SC(e-), D+] charge-separated pairs are
formed and the thermodynamically downhill charge recombina-
tion process, eq 3, must be prevented to yield a high photocur-
rent.
SC(e-)/S + D+ 93.8 SC/S + D
(3)
In this work we have prepared new Ru(II) sensitizers, with
functional groups for surface attachment, that are potent excited-
state oxidants capable of efficient iodide oxidation. We present
photophysical and reductive electron-transfer studies of these
compounds in fluid acetonitrile and when bound to colloidal
In photoelectrosynthetic cells, SC(e-) and D+ undergo subse-
quent redox reactions that yield chemical products. For regen-
* Corresponding author.
10.1021/jp026823n CCC: $25.00 © 2003 American Chemical Society
Published on Web 12/03/2002

246 J. Phys. Chem. B, Vol. 107, No. 1, 2003
Bergeron and Meyer
SCHEME 1
TiO2 and ZrO2 mesoporous thin films. Unlike previous studies,
we find examples where neither the sensitizer’s reduced state
nor excited-state transfers electrons to the semiconductor
efficiently.8 The results demonstrate a general approach for
photogenerating oxidizing and reducing equivalents with well-
defined molecular structures that store >2 eV of free energy.
Hz), 7.96 (d, 2H, J ) 2.6 Hz), 8.03 (d, 2H, J ) 6.1 Hz), 8.67
(s, 2H, J ) 2.6 Hz), 9.34 (s, 4H), 10.14 (s, 2H).
Ru(Me2SO)4(Cl)2. Ru(Cl)3‚xH2O (660 mg, 2.74 mmol Ru)
was refluxed in 20 mL of DMSO under argon. The solution
first turned red within 1-2 h, then yellow. After the solution
cooled to room temperature, addition of 20 mL of acetone
yielded a yellow precipitate that was filtered and rinsed with
ether (302 mg, 23%).
Experimental Section
Ru(bpz)3(PF6)2. Ru(Me2SO)4(Cl)2 (400 mg, 1.1 mmol) and
bpz (800 mg, 5.06 mmol) were refluxed in 30 mL of H2O for
20 h. The reaction mixture was allowed to cool to room
temperature and filtered. The solution was combined with 10
mL of 2 mM KPF6(aq), resulting in the formation of an orange
precipitate. The filtered crude product was rinsed with 5 mL of
water and recrystallized in 70 mL of acetone/ethanol 6:1 (151
Materials. RuCl3.xH2O (∼ 42% Ru) was used as received
from Johnson Matthey. The following materials were obtained
from Aldrich and used as received: CuSO4‚5H2O (98%),
2-pyrazinecarboxylic acid (99%), hydroquinone (99%), 4,4′-
dimethyl-2,2′-dipyridyl (99%), LiCl (99%), ammonium hexaflu-
orophosphate (95%), tetrabutylammonium hexafluorophosphate,
TBAH (98%), tetrabutylammonium iodide (98%), tetraethyl-
ammonium chloride hydrate (98%), LiClO4 (95%), 2,3-bis(2-
pyridyl)pyrazine (98%), CH3CN, DMSO, CDCl3, and DMSO-
d6. Phenothiazine (Aldrich, 98%) was recrystallized from toluene
prior to use. Ru(deeb)2(Cl)2, Ru(deeb)(bpy)2(PF6)2, and 2,2′-
bipyrazine (bpz) were prepared as previously described.11-13
Preparations. Ru(deeb)2(dpp)(PF6)2. Ru(deeb)2Cl2 (535 mg,
0.693 mmol) and a 4-fold excess of 2,3-bis(2-pyridyl)pyrazine
(310 mg, 1.32 mmol) were combined in 115 mL of ethanol/
water (50:50) and the mixture was refluxed for 8-10 h. The
orange/red solution was allowed to cool to room temperature
and filtered to remove unreacted ligand. Approximately 40 mL
of a 0.2 M KPF6(aq) was combined with the solution and an
immediate orange precipitate formed that was filtered and dried
under vacuum (76%). Mass spectrum, (M + 1)+ ) 1082 (calcd.
1082), (M + 1)2+ ) 937 (calcd. 937). Anal. Calcd. for
C46N8O8H42P2F12Ru: C, 45.06; N, 9.14; H, 3.45. Found: C,
1
mg, 28%). H NMR ((CD3)2SO) δ: 8.00 (d, 6H, J ) 3.2 Hz),
8.71 (d, 6H, J ) 3.2 Hz), 10.14 (s, 6H).
Ru(bpz)2(Cl)2. A solution of Ru(bpz)3(PF6)2 (150 mg, 0.174
mmol) and tetraethylammonium chloride hydrate, TEACl‚H2O,
(2.5 g, 15 mM) were combined in 750 mL of argon-saturated
CH3CN and stirred. The mixture was irradiated using a 1000
W Xe lamp with a UV cutoff prefilter. Absorbance readings
were taken periodically to monitor the extent of the reaction
over approximately 12 h. The deep purple solution was filtered
and the solvent volume was reduced to approximately 60 mL
and this caused precipitation of the desired product, Ru(bpz)2-
(Cl)2. Filtration, followed by washing with acetonitrile, gave
purple crystals (30 mg, 35%).
Ru(deeb)(bpz)2(PF6)2. Ru(bpz)2(Cl)2 (40 mg, 0.0826 mmol)
and 4,4′-(CO2CH2CH3)2-2,2′-bipyridine (74 mg, 0.247 mmol)
were combined in 8 mL of ethanol/water (50:50) and the mixture
was refluxed for 8 h. The solution was allowed to cool to room
temperature and filtered. Approximately 10 mL of 0.1 M
KPF6(aq) was combined with the solution resulting in formation
of a precipitate that was isolated by filtration (43 mg, 52%).
Anal. Calcd. for C32N10O4H28P2F12Ru: C, 38.14; N, 13.90; H,
1
44.39; N, 9.19; H, 3.47. H NMR ((CD3)2SO) δ: 1.35 (t, 6H,
CH3), 1.37 (t, 6H, CH3), 4.42 (q, 4H, CH2), 4.46 (t, 4H, CH2),
7.16 (d, 1H, J ) 7.9 Hz), 7.32-7.45 (m, 1H), 7.69-8.27 (m,
14H), 8.66-8.70 (m, 2H), 9.36 (s, broad, 4H).
Ru(deeb)2(bpz)(PF6)2. Ru(deeb)2Cl2 (384 mg, 0.497 mmol)
and 2,2′-bipyrazine (315 mg, 1.99 mmol) were combined in
100 mL of ethanol/water (50:50) and refluxed for 3 days. The
orange/red solution was allowed to cool to room temperature
and filtered to remove unreacted ligand. Upon addition to 15
mL of 0.2 M KPF6(aq) an immediate orange/red precipitate
formed that was filtered and dried under vacuum (56%). Mass
spectrum, M+ ) 1005 (calcd. 1005), M2+ ) 860 ( calcd. 860).
Calcd. for C40N8O8H38P2F12Ru: C, 41.78; N, 9.75; H, 3.33.
1
2.80. Found: C, 36.94; N, 13.80; H, 2.88. H NMR ((CD3)2-
SO) δ: 1.36 (t, 6H, CH3), 4.45 (q, 4H, CH2), 7.86 (dd, 2H, J
) 5.9 Hz, J ) 2.2 Hz), 7.90 (d, 2H, J ) 2.2 Hz), 8.00 (d, 2H,
J ) 5.9 Hz), 8.05 (d, 2H, J ) 3.3 Hz), 8.67 (d, 2H, J ) 3.2
Hz), 8.72 (d, 2H, J ) 2.2 Hz), 9.34 (d, 2H, J ) 2.2 Hz), 10.13
(s, 2H), 10.14 (s, 2H).
Colloidal MO2 Films. Colloidal TiO2 and ZrO2 films were
prepared as previously described.13 The metal oxide films were
placed in a pH 11 aqueous solution (NaOH) for 1 h, rinsed in
neat acetonitrile, and were then exposed to mM concentrations
of the Ru(II) compounds for at least 12 h. The films were rinsed
1
Found: C, 41.24; N, 9.69; H, 3.41. H NMR ((CD3)2SO) δ:
1.35 (t, 6H, CH3), 1.36 (t, 6H, CH3), 4.44 (q, 4H, CH2), 4.45
(q, 4H, CH2), 7.85 (d, 4H, J ) 5.9 Hz), 7.92 (d, 2H, J ) 5.9

Reductive Electron Transfer Quenching of MLCT
J. Phys. Chem. B, Vol. 107, No. 1, 2003 247
in neat acetonitrile before spectroscopic, electrochemical, or
photoelectrochemical measurements.
Measurements. Absorption. All absorption measurements
were acquired by placing the TiO2 or ZrO2 on glass films
diagonally in an acetonitrile filled 10 mm × 10 mm quartz
cuvette, equipped with a 24/40 ground quartz joint. The cell
was closed with a PTFE stopper and purged with argon through
a glass capillary tube, when indicated. Ground-state absorption
spectra were acquired at ambient temperature in air using a
Hewlett-Packard 8453 diode array spectrophotometer.
Transient absorption spectra were acquired as previously
described. Briefly, an approximately 7 ns, 532 nm laser pulse
from a Surelite II Nd:YAG, Q-switched laser was used as the
excitation source. The approximately 1 cm beam was expanded
using a quartz concave lens as a means of both course
attenuation of the excitation energy and ensuring homogeneous
irradiation of the sample.13 The energy was often further
attenuated using a polarizing prism of local design. Each kinetic
trace was acquired by averaging 10-400 laser shots (typically
80) at a repetition rate of 1 Hz. Excitation was carried out such
that the entire exposed TiO2 surface of the sample, positioned
at an approximately 45° angle to the excitation beam, was
irradiated. The Xe probe (150 W, Applied Photophysics,
operating in pulsed mode) was positioned normal to the
excitation beam.
Figure 1. Normalized absorbance and steady-state photoluminescence
of Ru(deeb)2(bpz)(PF6)2 in acetonitrile (s), on TiO2 (- -), and on
ZrO2 (- - -).
Photoluminescence. Steady-state photoluminescence (PL)
spectra were obtained with a Spex Fluorolog that had been
calibrated with a standard NIST tungsten-halogen lamp. Time-
resolved photoluminescence (TRPL) decays were acquired with
an apparatus that has been previously described.14
solution. The measured reduction potentials, and the free energy
stored in the excited state, ∆Ges, were used to estimate the
excited-state reduction potentials with eqs 4 and 5, Table 2.
Steady-State Photoreductions. The ruthenium compounds
were excited with ∼120 mW/cm2 of 514.5 nm light (Coherent
Innova 70 Argon Ion Laser) in a 0.3 M triethylamine, argon-
purged acetonitrile solution. These photoreductions were per-
formed in the cuvette holder of an HP 8453 diode array
spectrophotometer and the changes in absorption as a function
of irradiance were monitored.
Electrochemistry. All cyclic voltammograms were recorded
on a BAS 50 electrochemical analyzer in 0.1 M TBAH
acetonitrile solution. A Ag/AgCl reference electrode with Pt
gauze auxiliary and Pt working electrode comprised the three
electrode cell. All samples were extensively purged with argon,
and the reference occasional monitored for a drift in potential
using ferrocene as a standard.
E°(RuII*/+) ) E°(RuII/+) + ∆Ges
E°(RuIII/II*) ) E°(RuIII/II) - ∆Ges
(4)
(5)
The energy stored in the thermally equilibrated excited state,
∆Ges, was estimated with a tangent line extrapolated to zero
intensity on the high-energy side of the corrected PL spectra.15
We were unable to measure sensitizer reduction potentials when
the sensitizer was anchored to the metal oxide surface. In
previous work, we have found that the solution and surface-
bound Ru(III/II) potentials are within (20 mV the same.13 There
is a measurable shift in the photoluminescence spectra of the
surface-bound compounds relative to solution values. Therefore,
in estimating the reducing and oxidizing power of the surface-
bound sensitizers, solution potentials and surface photolumi-
nescence data were used in conjunction with eqs 4 and 5. The
first oxidation of PTZ was found to be reversible with
E°(PTZ+/0) ) 0.58 V vs SCE.
Photoelectrochemistry. Photoelectrochemical measurements
were made as previously described.10a
1
NMR. All H NMR data was obtained on a Bruker AMX
300 Spectrometer.
Time-resolved PL decays of Ru(deeb)2(bpz)2+* and the other
Ru(II)* compounds in solution were fit to a first-order kinetic
model given in eq 6. The TRPL of Ru(II)* compounds bound
to TiO2 and ZrO2 were nonexponential and were well described
by a parallel first- and second-order kinetic model, eqs 7 and
8,
Results
Figure 1 shows the room-temperature absorption and pho-
toluminescence, PL, spectra of Ru(deeb)2(bpz)2+ in fluid solu-
tion, and when anchored to colloidal TiO2 and ZrO2 films and
immersed in argon-saturated acetonitrile. The metal-to-ligand
charge transfer (MLCT) bands broaden slightly and the PL
maximum red shifts upon surface attachment. The optical
properties of Ru(deeb)(bpz)22+ and Ru(deeb)2(dpp)2+ in solution
and when bound to the metal oxide surfaces are qualitatively
similar, Table 1. Throughout the Results Section, typical data
for Ru(deeb)2(bpz)2+ is shown; the corresponding data for Ru-
(deeb)(bpz)22+ and Ru(deeb)2(dpp)2+ are given in the Supporting
Information. The ligands are shown below for clarity.
I(t) ) A(exp(-k1t))
(6)
(7)
I(t) ) C(k1 exp(-k1t))/(k1 + p - p(exp(-k1t)))
p ) k2[Ru2+*]t)0
(8)
Reversible reductions of the Ru(II) compounds were observed
by cyclic voltammetry in 0.1 M TBAH acetonitrile electrolyte
where C is a constant, and k1 and k2 are the first- and second-
order rate constants, respectively.16

248 J. Phys. Chem. B, Vol. 107, No. 1, 2003
Bergeron and Meyer
processes.17 Stern-Volmer analysis was used to quantify the
TABLE 1: Photophysical Properties of Ru(II) Sensitizers on
MO2 and in Acetonitrilea
quenching dynamics, eq 9, Figure 2 inset.
sensitizerb
λ
abs.(nm)c
λ
PL (nm)d ꢀ (M-1 cm-1
)
τ (µs)e
τo/τ ) 1 + Ksv[Q]
(9)
Ru(deeb)(bpy)22+(sol)
Ru(deeb)(bpy)22+(TiO2)
Ru(deeb)(bpy)22+(ZrO2)
475
457
461
690
656
671
16000
0.990
1.04
1.30
In eq 9, the Stern-Volmer constant, Ksv, is equal to kqτï
where τï is the lifetime in the absence of PTZ.17 For the surface-
bound sensitizers, lifetimes were determined by fixing p in eqs
7 and 8 to the value determined in the absence of PTZ, and
allowing k1 to float until the best fit was obtained, with τ )
1/k1. Excellent fits via this method were obtained for both
solution and surface quenching. Table 3 summarizes these results
for Ru(deeb)2(bpz)2+ and the other compounds. Quenching
studies with iodide were also performed and reveal rapid and
efficient iodide oxidation, kq > 108 M-1 s-1. The transient
absorption details with iodide are more complex and will be
the focus of a separate paper.18
Figure 3 shows transient absorption difference spectra
recorded after pulsed 532 nm light excitation of Ru(deeb)2-
(bpz)2+ in fluid solution and when bound to TiO2 and ZrO2
films immersed in argon-saturated acetonitrile. Within reason-
able experimental error, the difference spectra are the same in
all three cases and are typical of MLCT excited states with
positive absorption bands at ∼ 380 nm, a bleach at ∼ 450 nm,
and an isosbestic point at 398 nm.16 In fluid acetonitrile, the
kinetics were first-order and when anchored to TiO2 and ZrO2
they were well-described by the parallel first- and second-order
kinetic model with rate constants that agreed well with those
obtained by TRPL measurements.
Transient absorption difference spectra recorded after pulsed
532 nm light excitation of Ru(deeb)2(bpz)2+ in argon-saturated
acetonitrile solution in the presence of PTZ are shown in Figure
4. The MLCT excited state was observed initially that evolved
into a new state with positive absorption bands at 380 and 520
nm, assigned to a [Ru(II)(L-)+/PTZ+] charge-separated state.
Within experimental error, the rate constants for appearance of
the charge-separated state agree with those measured by dynamic
PL quenching. At high PTZ concentrations, it was difficult to
transmit light below 400 nm due to competitive light absorption
by the PTZ. Therefore, the data shown was collected at 3-4
mM PTZ which corresponds to nearly complete quenching of
the Ru(II) excited states in solution and ∼75% within the
mesoporous thin films. Normalized absorption difference spectra
obtained at >5 µs after the laser pulse are within experimental
error the same on ZrO2 and TiO2 materials.
Ru(deeb)2(dpp)2+(sol)
Ru(deeb)2(dpp)2+(TiO2)
Ru(deeb)2(dpp)2+(ZrO2)
464
464
464
643
670
675
21200 ( 400 1.48
0.90
0.67
Ru(deeb)2(bpz)2+(sol)
458
658
683
683
18100 ( 600 1.31
Ru(deeb)2(bpz)2+(TiO2) 436-472f
0.56
0.63
Ru(deeb)2(bpz)2+(ZrO2)
466
Ru(deeb)(bpz)22+(sol)
Ru(deeb)(bpz)22+(TiO2)
Ru(deeb)(bpz)22+(ZrO2)
449
458
457
636
668
681
14460 ( 200 1.78
0.94
1.14
a All measurements made at room temperature. b The Ru(II) sensitiz-
ers with ligand abbreviations defined in the text. The measurements
were made in fluid acetonitrile (sol), and attached to (TiO2) and (ZrO2)
thin films immersed in an acetonitrile bath. c The visible MLCT
absorption maximum. d The photoluminescence maximum, (8 nm.
e Excited-state lifetime, ( 5%. Excited-state relaxation on TiO2 and
ZrO2 was found to be nonexponential. The first-order component
abstracted from a parallel first- and second-order kinetic model were
used to estimate the lifetime. See text for details. f A broad absorption
band with no clear maximum.
TABLE 2: Ground- and Excited-State Sensitizer Reduction
Potentialsa
sensitizer
E°(RuIII/II
)
E°(RuII/+ E°(RuIII/II*) E°(RuII*/+)
)
2+
Ru(deeb)(bpy)2
1.39
1.59
1.73
-1.0
-0.62
-0.57
-0.38
1.01
1.24
1.31
1.36
Ru(deeb)2(dpp)2+
-0.92
-0.82
-0.82
Ru(deeb)2(bpz)2+
2+ b
Ru(deeb)(bpz)2
a Potentials measured versus SCE in 0.1 M TBAH acetonitrile
electrolyte at room temperature. The ground-state potentials were
measured by cyclic voltammetry and the excited-state potentials were
estimated with eqs 4 and 5. b The reduction potentials of this compound
were not measured. The first reduction is expected to be bpz-based so
the E°(RuII/+) of Ru(deeb)2(bpz)2+ was assumed.
On a much longer time scale than is shown, the charge-
separated state decays to ground-state products. Figure 5 shows
transient absorption kinetics that correspond to electron transfer
from the reduced ruthenium compound to PTZ+ both in solution
and anchored to TiO2. The PTZ concentrations were kept high
to ensure that >99% of the excited states were quenched. Figure
5 shows that the yield of charge-separated pairs is about 3 times
larger in fluid solution. About 1/2 of the charge-separated states
formed in the mesoporous film recombine in the first micro-
second while there is negligible recombination in fluid solution
on this same time scale. UV-Vis absorption measurements
before and after the time-resolved studies revealed no evidence
for permanent photochemistry.
Chemical and photochemical redox reactions were performed
to obtain the visible absorption spectra of the oxidized phe-
nothiazine, PTZ+, and the reduced ruthenium compounds. The
stoichiometric addition of bromine to a PTZ acetonitrile solution
leads to the appearance of the PTZ+ absorption spectrum.
Photochemical reductions of the Ru(II) compounds were
obtained with 514.5 nm light excitation in the presence of the
Figure 2. Time-resolved photoluminescence quenching of Ru(deeb)2-
(bpz)(PF6)2 bound to TiO2 with PTZ following 460 nm excitation. Inset
shows a Stern-Volmer analysis for solution (9), bound to TiO2 (b),
and bound to ZrO2 (]).
The introduction of phenothiazine into the acetonitrile solution
resulted in excited-state quenching at homogeneous and het-
erogeneous environments. Figure 2 displays time-resolved PL
decays after pulsed light excitation of Ru(deeb)2(bpz)2+ in argon-
saturated acetonitrile with increasing concentrations of phe-
nothiazine, PTZ. Dynamic quenching was apparent for both
solution and surface experiments with no evidence for static

Reductive Electron Transfer Quenching of MLCT
J. Phys. Chem. B, Vol. 107, No. 1, 2003 249
TABLE 3: Excited-State Reductive Electron Transfer Rate Constantsa
-∆Gd
(eV)
b
b
Ksv
kq
kec
sensitizer
(M-1
)
(10-9 M-1 s-1
)
(10-9 M-1 s-1
)
Ru(deeb)(bpy)22+(sol)
Ru(deeb)(bpy)22+(TiO2)
Ru(deeb)(bpy)22+(ZrO2)
6030 ( 300
1190 ( 50
1200 ( 60
6.09 ( 0.30
1.15 ( 0.06
1.53 ( 0.08
8.58 ( 0.43
1.22 ( 0.06
1.65 ( 0.08
0.43
0.56
0.53
Ru(deeb)2(dpp)2+(sol)
Ru(deeb)2(dpp)2+(TiO2)
Ru(deeb)2(dpp)2+(ZrO2)
10200 ( 500
1130 ( 60
1080 ( 50
6.89 ( 0.34
1.25 ( 0.06
1.61 ( 0.08
10.3 ( 0.51
1.33 ( 0.07
1.74 ( 0.09
0.61
0.58
0.60
Ru(deeb)2(bpz)2+(sol)
Ru(deeb)2(bpz)2+(TiO2)
Ru(deeb)2(bpz)2+(ZrO2)
9900 ( 500
1010 ( 50
1101 ( 55
7.57 ( 0.38
1.79 ( 0.09
1.75 ( 0.09
11.8 ( 0.59
1.96 ( 0.10
1.91 ( 0.10
0.73
0.68
0.77
Ru(deeb)(bpz)22+(sol)
Ru(deeb)(bpz)22+(TiO2)
Ru(deeb)(bpz)22+(ZrO2)
13900 ( 700
2080 ( 100
2380 ( 120
7.79 ( 0.39
2.22 ( 0.11
2.09 ( 0.10
12.4 ( 0.62
2.48 ( 0.12
2.32 ( 0.12
0.78
0.71
0.79
a All measurements were performed in 0.1 M TBAH acetonitrile electrolyte with the electron donor phenothiazine. b Stern-Volmer, Ksv, and
observed quenching rate constants. c Electron-transfer rate constants calculated with eq 13. d Driving force for Ru(II)* reduction by phenothiazine.
Figure 3. Transient absorption spectra of Ru(deeb)2(bpz)(PF6)2 in 0.1 M TBAH acetonitrile electrolyte following 532 nm laser excitation (A) in
acetonitrile at 0 ns (9), 500 ns (]), 1200 ns (b), and 3000 ns (3); (B) bound to TiO2 at 0 ns (9), 80 ns (]), 300 ns (b), and 1500 ns (3); and (C)
bound to ZrO2 at 0 ns (9), 120 ns (]), 300 ns (b), and 1000 ns (3).
sacrificial donor triethylamine in argon saturated acetonitrile
solutions. Typical data for the excited-state reduction of Ru-
(deeb)2(bpz)2+ anchored to colloidal TiO2 and ZrO2 are shown
in Figure 6. The observed spectra were stable for short time
periods and returned to yield ground-state spectrum on a minutes
time scale. The absorption spectra of a 1:1 ratio of the reduced
ruthenium compound to PTZ+ were constructed and are
superimposed on the data in Figure 4. The agreement between
the simulated spectra and those observed transiently was good
over the visible range. Below 400 nm, where the materials
absorb strongly, the simulated spectra have a slightly lower
amplitude.
a Pt counter electrode and a 0.5 M tetrabutylammonium iodide/
0.05 M I2 acetonitrile electrolyte. The photocurrent action
spectrum agrees well with the sensitizer absorptance spectrum.
However, the photocurrent efficiency is very low, φ < 10-4
for all the sensitizers studied.
Discussion
A goal of this study was to prepare Ru(II) coordination
compounds with long-lived strongly oxidizing, excited states
that can be anchored to semiconductor surfaces for applications
in dye-sensitized photogalvanic cells. The Ru(II) compounds
synthesized with bipyrazine, bpz, or dipyridylpyrazine, dpp,
ligands are potent photooxidants, E°(RuII*/+) > +1.0 V vs SCE,
with microsecond excited-state lifetimes. The deeb ligands bind
The light-to-electrical energy conversion efficiency of these
materials was measured in a two-electrode configuration with

250 J. Phys. Chem. B, Vol. 107, No. 1, 2003
Bergeron and Meyer
Figure 4. Transient absorption difference spectra of Ru(deeb)2(bpz)(PF6)2 in 0.1 M TBAH acetonitrile electrolyte following 532 nm laser excitation
(A) in acetonitrile with 3 mM PTZ at 0 ns (9), 20 ns (]), 50 ns (b), and 300 ns (3); (B) bound to TiO2 with 3.5 mM PTZ at 0 ns (9), 15 ns (]),
60 ns (b), and 1000 ns (3); and (C) bound to ZrO2 with 1.6 mM PTZ at 0 ns (9), 30 ns (]), 120 ns (b), and 3000 ns (3). Superimposed on the
final difference spectra are simulated spectra (*).
(anatase) TiO2 surfaces. The results identify important loss
mechanisms relevant to solar energy conversion and also
demonstrate an approach for photogenerating oxidizing and
reducing equivalents with well-defined structures and potentials
on semiconductor surfaces.
Excited States. The steady-state absorption and photolumi-
nescence spectra of the Ru(II) compounds bound to the
nanoparticle oxide surfaces are only slightly shifted from that
observed in fluid solution, indicating that the compounds retain
their coordination sphere upon surface binding and the emissive
and absorptive states are largely unperturbed. The ethyl ester
groups on deeb ligands are known to hydrolyze upon reaction
with the TiO2 surface.3 For these sensitizers, the red-shifted
photoluminescence spectra upon surface binding reflect de-
creased π back-bonding to the Ru center by the deeb ligands.
Figure 5. Absorption change monitored at 510 nm following pulsed
Time-resolved absorption difference spectra are those ex-
532.5 nm light excitation of Ru(deeb)2(bpz)(PF6)2 in 0.1 M TBAH
pected for the MLCT excited state and indicate negligible
acetonitrile electrolyte in fluid electrolyte solution (upper trace) and
when anchored to TiO2 (lower trace).
excited-state electron injection into the semiconductor, φinj
<
0.05. Injection from the thermally equilibrated excited state,
vibrationally “hot” excited states, and the Franck-Condon
excited state are not expected on consideration of the interfacial
energetics. For example, 532.5 nm light excitation of Ru(deeb)2-
(bpz)2+ yields a Franck-Condon excited state with a reduction
potential of -0.60 V vs SCE and vibrational relaxation to the
thermally equilibrated state yields an even weaker photoreduc-
tant (Table 2). The energetic position of Ecb has been estimated
to be -2.24 V vs SCE in 0.2 M TBAClO4/CH3CN.30 Excited-
state injection is consequently not expected on energetic grounds
and is not observed. Therefore, these compounds provide an
opportunity to study molecular excited states anchored to
semiconductor surfaces without complications from interfacial
tenaciously to metal oxide surfaces resulting in high surface
coverages of the Ru(II) compounds on the semiconductor
surface.3 The light-to-electrical energy conversion efficiency of
the materials in photogalvanic cells was, however, disappoint-
ingly low. Time-resolved spectroscopic measurements provide
keen insights into the loss mechanisms and suggest methods
for future optimization.
Below, we discuss the photophysical and electron-transfer
studies of these molecular excited states. Since the metal-to-
ligand charge-transfer (MLCT) excited states of Ru(II) com-
pounds have been previously reported in the literature,19,20 we
restrict our discussion to the behavior on nanocrystalline

Reductive Electron Transfer Quenching of MLCT
J. Phys. Chem. B, Vol. 107, No. 1, 2003 251
Figure 6. Steady-state photolysis of Ru(deeb)2(bpz)(PF6)2 with 514 nm laser excitation in the presence of 0.3 M triethylamine with 0.1 M
tetrabutylammonium hexafluorophosphate in argon-purged acetonitrile in (A) solution, (B) bound to TiO2, and (C) bound to ZrO2.
electron transfer. We note that the prediction of interfacial
energetics for the operational solar cell (with I-/I3-) is
problematic as there exists no unambiguous method for deter-
mining Ecb under these conditions.
protocols for directly quantifying the fraction of excited states
that decay through a given pathway. Qualitatively, we have
observed that low sensitizer surface coverages and irradiances
favor the first-order relaxation while the second-order pathway
is predominant at high surface coverages and irradiances.16
However, low surface coverages result in less sunlight being
harvested and this is undesirable for solar conversion applica-
tions. At the lowest irradiances used, ∼ 1 mW/cm2, and the
optimal maximum surface coverages, a second-order component
was clearly present. A standard for solar cell testing is one sun
of air mass 1.5 solar irradiation that corresponds to a steady-
state irradiance of 100 mW/cm2. This suggests that excited-
state annihilation processes are a factor that would lower the
efficiency of photogalvanic cells based on this approach.
Photoinduced Electron Transfer. Electron transfer from
phenothiazine, PTZ, efficiently quenches the MLCT excited-
states donors, eq 10. Time-resolved and spectroelectrochemical
absorption
A notable difference between excited states in fluid solution
and attached to metal oxide nanoparticles is the relaxation
kinetics. In fluid solution, exponential kinetics are observed with
a rate constant equal to the sum of a radiative and a nonradiative
rate constant.19 This behavior is expected for MLCT excited
states.19 In contrast, the excited states anchored to ZrO2 or TiO2
are nonexponential and are adequately described by a parallel
first- and second-order kinetic model that has been previously
derived.16 The first-order rate constant corresponds to the
“normal’ solution relaxation pathways. The second-order com-
ponent is attributed to a competitive bimolecular reaction,
proposed to be a triplet-triplet annihilation reaction, rate-limited
by energy transfer migration between the ground and excited
states of adjacent, surface-attached molecules. In fluid solution,
rate limitations imposed by the solvent are sufficient to prevent
observation of such processes. However, these mesoporous
nanocrystalline films have a high local concentration of the Ru-
(II) compounds, thought by many to be monolayer surface
coverage.2 Constraining these compounds in close proximity
promotes intermolecular energy migration, and several examples
on metal oxide nanoparticles are now documented in the
literature.16,21 Energy migration across the nanocrystalline
semiconductor surface provides a mechanism for two excited
states to encounter and react with each other. The reaction is
reported to be a triplet-triplet annihilation process and proceeds
kq
98
Ru(deeb)2(bpz)2+* + PTZ
Ru(deeb)2(bpz-)+ + PTZ+ (10)
measurements clearly demonstrate that PTZ+ and the reduced
ruthenium compound are present following photoexcitation in
the expected 1:1 stoichiometry. Approximately 2.2 eV of free
energy are stored in the form of redox equivalent photoproducts.
Recombination requires tens of milliseconds in fluid solution
and tens of microseconds in the thin films. The recombination
follows the expected second-order equal concentration kinetics;
however, uncertainties in the appropriate path length make it
difficult to quantify the true second-order rate constant within
the mesoporous films.9
with a rate constant greater than 108 s-1 22
The products are
.
unknown in the present work, but have been the subject of much
discussion in the MLCT literature.22
Excited-state annihilation reactions are unwanted and can
lower the efficiency of solar energy conversion schemes based
on diffusional quenching. It is therefore of interest to identify
conditions that minimize the bimolecular excited-states decay
pathway. Remarkably, we have been unable to find literature
The transient spectroscopic data also demonstrate that the
reduced sensitizers do not inject electrons into TiO2 under these
conditions, step 2 in eq 2. The first reduction of the Ru(II)
compounds studied is localized on the bpz (or dpp) ligand and

252 J. Phys. Chem. B, Vol. 107, No. 1, 2003
Bergeron and Meyer
is not metal-based, i.e., RuII(deeb)2(bpz-)+. Therefore, electron
injection would be a remote process occurring from a ligand
that is not directly bonded to the semiconductor surface. Rapid
subnanosecond electron transfer from excited states localized
on ligands remote to the semiconductor surface have been
previously demonstrated.23 Since the reduced Ru compounds
have long microsecond lifetimes, the lack of injection is not
likely to be kinetic in origin and indicates that the conduction
band edge is not energetically accessible, i.e., Ecb < -1.0 V vs
SCE.
SCHEME 2
The quenching rate constants, kq, measured for eq 10 may
not correspond to the true electron-transfer rate constants and
must be corrected for possible contributions from diffusion.24
A kinetic mechanism that incorporates a diffusional preequi-
librium step is shown below, eqs 11 and 12.
least 2/3 lower than that in solution. Cage escape yields
following reductive electron transfer are known to be high in
fluid solution and this implies a cage escape yield <0.3 on the
nanocrystalline TiO2 surface.26 The low yield of charge-
separated products is unfortunate from an energy conversion
perspective and new strategies must be employed to increase
the quantum yields. Previous workers have found that ionic
strength and pH can be used to tune the cage escape yields at
colloidal SiO2 interfaces.25 A negatively charged particle surface
increases the cage escape yield when one of the charge-separated
products is also negatively charged. Electrostatics of this nature
appears to be relevant here as well. The TiO2 is negatively
charged and the oxidized phenothiazine product is positively
charged. Therefore, following electron transfer, electrostatic
interactions disfavor cage escape.
Ru(deeb)2(bpz)2+* + PTZ ykdiff
z
k-diff
[Ru(deeb)2(bpz)2+*, PTZ] (11)
ket
98
[Ru(deeb)2(bpz)2+*, PTZ]
[Ru(deeb)2(bpz-)+, PTZ+] (12)
A steady-state approximation yields eq 13 where Ka ) kdiff
/
k-diff
.
1
1
1
)
+
(13)
kq kdiff ketKa
The rapid recombination of the PTZ+ with the oxidized
sensitizer in the TiO2 film serves as direct evidence for
electrostatic PTZ+ surface adsorption. For diffusion-limited
charge recombination in acetonitrile at these concentrations,
there is no measurable loss of PTZ+ on a sub-microsecond time
scale. In contrast, within the mesoporous TiO2 films, almost
half of the pairs recombine in the first microsecond. This may
simply reflect the high local concentration within the meosporous
thin films, but is highly suggestive of recombination to PTZ+
absorbed on the semiconductor surface. Using pH, or specific
cation adsorption, to increase the cage escape yield would also
shift the energetic position of Ecb and could change the
sensitization mechanism. It would eliminate many of the possible
benefits of having the reduced sensitizer inject the electron rather
than the excited state. This point is elaborated upon further
below.
Implications for PhotogalVanic Cells. Photogalvanic cells
were the hallmark molecular solar cells of the 1940s to 1960s.27
In general, these cells operated by photoinduced charge separa-
tion in fluid solution followed by diffusion of the separated
charges to dark electrodes. The efficiency of such cells was
generally very low, < 1% in part due to competitive light
absorption and difficulties with collecting the separated charges.
Many of the obstacles have been overcome by the Gra¨tzel
approach.2 The mesoporous nanocrystalline thin films provide
a high surface area for sensitizer attachment, and hence efficient
light harvesting, that requires only one of the separated charges
to diffuse. To our knowledge, this report represents the first
attempt to prepare dye-sensitized photogalvanic regenerative
solar cells with these same materials. In terms of practical
applications, the attempt failed. The reduced sensitizers did not
inject electrons into the semiconductor, so the solar energy
conversion was disappointingly low. This occurs because the
conduction band is not energetically accessible, Ecb < E°(RuII/+),
Scheme 2A. Nevertheless, some insights into how efficient
When the time scales for diffusion and electron transfer are
comparable, the observed quenching rate constant has contribu-
tions from both. In the limit that kdiff , ketKa, the reaction is
diffusion controlled. In the other limit, electron transfer is slow
compared to diffusion and kq is equal to the product of Ka and
ket. The diffusion-controlled rate constant in acetonitrile at 25
°C can be calculated from the Debye-Smoluchowski-Stokes-
Einstein relations. For Ru(bpy)32+ with neutral organic donors
it has been calculated to be kdiff ) 2.1 × 1010 M-1 s-1 24
.
The
quenching constants measured experimentally approach this
limit and corrections for diffusion using eq 12 are therefore
warranted, Table 3.
The electron-transfer rate constants are found to increase with
driving force consistent with electron transfer in the Marcus
normal region. The trend is clear both in fluid solution and in
the mesoporous thin films. The rate constants measured in the
latter media are consistently a factor of 2 to 3 slower. Since
diffusion of the Ru(II) compounds is inhibited by carboxylate
binding to the metal oxide surface, quenching likely originates
mainly from PTZ diffusion. The quenching rate constant is
therefore lower, as has been previously found at aqueous
colloidal interfaces.25 If we assume that the electron-transfer
rate constants would be within experimental error the same, then
kdiff ∼ 3 × 109 M-1 s-1 within the mesoporous films. Sluggish
electron transfer may also reflect small changes in the thermo-
dynamics for electron transfer at the interface. In any case, it is
possible to quench > 99% of the excited states with high PTZ
concentrations, indicating that PTZ can access essentially all
the surface-bound excited states within the nanostructured thin
film.
The yield of electron transfer products in the mesoporous
films is clearly much lower than that observed in fluid solution.
It is experimentally difficult to quantify precise cage escape
yields in these materials. However a crude estimate based on
data like that shown in Figure 5 indicates that the yield is at

Reductive Electron Transfer Quenching of MLCT
J. Phys. Chem. B, Vol. 107, No. 1, 2003 253
SCHEME 3
a Ru* f Ru hopping rate of 1-4 ns, and suggest that an excited
state with a microsecond lifetime in a close-packed monolayer
could undergo a thousand hops.32 This would indicate that the
sensitizer excited states under study can migrate across the entire
nanoparticle surface within its lifetime.
Facile energy transfer allows translation of the excited state
to encounter other excited states, impurities, or other species
that could quench it, Scheme 3. For example, excited-state
reduction by a reduced ruthenium compound would be energeti-
cally favored (by over 2 eV for the compounds under study),
Scheme 3B. In cases where the reduced sensitizers inject
electrons, the TiO2 electron could also reduce the MLCT excited
state.33
Collectively, the electron-transfer quenching pathways in
Scheme 3B and C could lead to a type of “super quenching”
where multiple excited states are deactivated by a single
quencher.34 For these processes to be significant, the rate
constant for photoreduction would have to be competitive with
intermolecular energy transfer away from the quenching site.
No experimental evidence for these processes was obtained and
the decreased rate constants measured for the mesoporous films
relative to solution suggest that this is not a significant excited-
state decay pathway, presumably because of the low yield of
charge-separated products. Nevertheless, this reaction could be
a significant obstacle for the preparation of efficient photogal-
vanic dye-sensitized cells. One method for avoiding these
deleterious processes is to eliminate diffusion by covalently
binding an electron donor directly to the sensitizer.35
photogalvanic solar cells might be prepared in the future and
some possible advantages of this approach are discussed below.
A potential advantage of having the reduced Ru(II) state inject
an electron into the semiconductor rather than the excited state,
is that the former is a stronger reductant. For the sensitizers
under study, the energy difference is 300 to 400 mV and this is
typical of Ru(II) sensitizers.19 This means that Ru(II) sensitizers
that are weak excited-state reductants may not sensitize TiO2
under conditions where the reduced sensitizer could, Scheme
2B. Sensitizers with low-lying π* orbitals that efficiently absorb
light in the red and near-infrared region may therefore yield
higher photocurrent efficiencies when the reduced sensitizer
injects the electron.28
Reduced sensitizer injection may also improve the open
circuit photovoltages of regenerative dye-sensitized solar cells.
In the state-of-the-art cells, incident photons are converted to
electrons in an external circuit near-quantitatively.2 However,
less than 1/2 of the free energy stored in the sensitizer excited
state is converted to Gibbs free energy in the molecular solar
cell.2 Therefore, the sensitization kinetics are well optimized
while the thermodynamics are not. In their original Nature paper,
Gra¨tzel and O’Regan reported that the open circuit photovoltage
was increased, and the photocurrent decreased, when the cation
in the electrolyte was changed from Li+ to tetrabutylammo-
nium.5 Studies since have shown that Li+ adsorption to the TiO2
surface shifts the conduction band positive promoting excited-
state electron injection at the cost of lowering the quasi-Fermi
level of the TiO2.29,30 If the injection yield could be kept high
without introducing Li+, then the power conversion efficiency
would increase. In principle, the reduced sensitizer could
accomplish this.
A key difference between the two sensitization mechanisms
is the state that undergoes diffusional quenching, the excited
state versus the oxidized state. In the latter case, the excited
sensitizer can be very short-lived and indeed this has been found
to be the case.31 Long-lived excited states are needed for
diffusional quenching and when sensitizers are arranged in close
proximity, new energy-wasting pathways are introduced. One
such pathway was already mentioned above, excited-state-
excited-state annihilation reactions that are rate limited by
intermolecular energy transfer across the semiconductor surface,
Scheme 3A. The number of intermolecular Ru* f Ru “hops”
that an excited state can undergo is unknown in the present
system. However, recent Monte Carlo simulations for Ru(II)
complexes weakly coupled through a polymer backbone reveal
Conclusion
The products of reductive electron-transfer reactions at
sensitized semiconductor interfaces have been studied for the
first time. There is strong experimental evidence that the reduced
ruthenium compounds, like the excited states, do not inject
electrons into TiO2. This indicates that the acceptor states in
TiO2, presumably the conduction band, are more negative on
an electrochemical scale than -1.0 V vs SCE under these
conditions. In principle, it may be possible to shift the
conduction band edge with pH, cations, or by utilizing alterna-
tive semiconductor materials such that the reduced compounds
are energetically capable of interfacial electron transfer.
The possibility of making efficient photogalvanic cells based
on dye-sensitized nanocrystalline films appears to be very real,
even though it was not realized here. It may be possible to
significantly enhance both the spectral sensitivity and the open
circuit photovoltages over what has been previously reported.2
Rapid intermolecular energy transfer across the nanocrystalline
semiconductor surface leads to triplet-triplet annihilation
reactions that can lower the power conversion efficiency. In
addition, excited-state energy migration may lead to additional
excited-state deactivation pathways and a high sensitivity to
impurities. The use of molecular sensitizers that undergo
intramolecular electron transfer prior to interfacial electron
transfer will eliminate excited-state diffusion and should enhance
energy conversion efficiencies.
Acknowledgment. We thank the National Science Founda-
tion for research support.
Supporting Information Available: Typical data for for
Ru(deeb)(bpz)22+ and Ru(deeb)2(dpp)2+. This material is avail-
able free of charge via the Internet at http://pubs.acs.org.
References and Notes
(1) Gerischer, H. Photochem. Photobiol. 1972, 16, 243-260.
(2) Gra¨tzel, M. Nature 2001, 414, 338-344.

254 J. Phys. Chem. B, Vol. 107, No. 1, 2003
Bergeron and Meyer
(3) Qu, P.; Meyer, G. J. Dye-Sensitized Electrodes. In Electron Transfer
in Chemistry; Balzani, V., Ed.; 2000; Part 2, Vol. IV, Chapter 2, pp 355-
411.
Sykora, M.; Fleming, C. N.; Odobel, F.; Meyer, T. J. J. Phys. Chem. B
2001, 105, 8895-8904.
(22) (a) Milosavijevic, B. H.; Thomas, J. K. J. Phys. Chem. 1983, 87,
616-621. (b) Kennelly, T.; Gafney, H. D.; Braun, M. J. Am. Chem. Soc.
1985, 107, 4431-4440. (c) Fan, J.; Shi, W.; Strekas, T. C.; Gafney, H. D.
J. Phys. Chem. 1989, 93, 373-376.
(23) (a) Argazzi, R.; Bignozzi, C. A.; Heimer, T. A.; Meyer, G. J. Inorg.
Chem. 1997, 36, 2-3. (b) Qu, P.; Thompson, D. W.; Meyer, G. J. Langmuir
2000, 16, 4662-4671.
(4) (a) Kirsch and co-workers have correlated Stern-Volmer constants
calculated from luminescence and photocurrent measurements at planar
electrodes: Ortmans, I.; Moucheron, C.; Kirsch-De Mesmaker, A. Coord.
Chem. ReV. 1998, 168, 233-271. (b) In many other cases, it was not possible
to identify the sensitization mechanism, see for example: Fox, M. A.; Nobs,
F. J.; Voynick, T. A. J. Am. Chem. Soc. 1980, 102, 4036-4039.
(5) O’Regan, B.; Gra¨tzel, M. Nature 1991, 353, 737-739.
(6) (a) Tachibana, Y.; Moser, J. E.; Gra¨tzel, M.; Klug, D. R.; Durrant,
J. J. Phys. Chem. 1996, 100, 20056-20062. (b) Hannappel, T.; Burfeindt,
B.; Storck, W.; Willig, F. J. Phys. Chem. B 1997, 101, 6799-6807. (c)
Heimer, T. A.; Heilweil, E. J. J. Phys. Chem. B 1997, 101, 10990-10996.
(d) Benko, G.; Kallioinen, J. E.; Korppi-Tommola, J. E. I.; Yartsev, A. P.;
Sundstrom, V. J. Am. Chem. Soc. 2002, 124, 489-494.
(7) Nasr, C.; Hotchandani, S.; Kamat, P. V. J. Phys. Chem. B 1998,
102, 4944-4951.
(24) Meyer, T. J. Prog. Inorg. Chem. 1982, 30, 389-439.
(25) (a) Willner, I.; Yang, J.-M.; Laane, C.; Otvos, J. W.; Calvin, M. J.
Phys. Chem. 1981, 85, 3277-3282. (b) Degani, Y.; Willner, I. J. Am. Chem.
Soc. 1983, 105, 6228-6233.
(26) Mallouk, T. E.; Krueger, J. S.; Mayer, J. E.; Dymond, C. M. G.
Inorg. Chem. 1989, 28, 3507-3510.
(27) Albery, W. J. Acc. Chem. Res. 1982, 15, 142-149.
(28) (a) Argazzi, R.; Bignozzi, C. A.; Heimer, T. A.; Castellano, F. N.;
Meyer, G. J. Inorg. Chem. 1994, 33, 5741-5749. (b) Nazeeruddin, Md.
K.; Pechy, P.; Renouard, T.; Zakeeruddin, S. M.; 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-
1620.
(29) (a) Kelly, C. A.; Thompson, D. W.; Farzad, F.; Stipkala, J. M.;
Meyer, G. J. Langmuir 1999, 15, 7047-7054. (b) Qu, P.; Meyer, G. J.
Langmuir 2001, 17, 6720-6728.
(30) (a) Redmond, G.; Fitzmaurice, D. J. Phys. Chem. 1993, 97, 1426-
1430. (b) Enright, B.; Redmond, G.; Fitzmaurice, D. J. Phys. Chem. 1994,
98 (8), 6195-6200.
(31) (a) Ferrere, S.; Gregg, B. A. J. Am. Chem. Soc. 1998, 120, 843-
844. (b) Islam, A.; Hara, K.; Singh, L. P.; Katoh, R.; Yanagida, M.; Murata,
S.; Takahashi, Y.; Sugihara, H.; Arakawa, H. Chem. Lett. 2000, 490-494.
(8) Thompson, D. W.; Kelly, C. A.; Farzad, F.; Meyer, G. J. Langmuir
1999, 15, 650-653.
(9) Argazzi, R.; Bignozzi, C. A.; Heimer, T. A.; Castellano, F. N.;
Meyer, G. J. J. Phys. Chem. B 1997, 101, 2591-2597.
(10) Alternative redox mediators for dye-sensitized solar cells represent
an active area of research. See for example: (a) Oskam, G.; Bergeron, B.
V.; Meyer, G. J.; Searson, P. C. J. Phys. Chem. B 2001, 105, 6867-6873.
(b) Busbammer, H.; Moser, J. E.; Zakeeruddin, S. M.; Nazeeruddin, M.
K.; Gra¨tzel, M. J. Phys. Chem. B 2001, 105, 10461-10464.
(11) Crutchley, R. J.; Lever, A. B. P. Inorg. Chem. 1982, 21, 2276-
2282.
(12) Evans, I. P.; Spencer, A.; Wilkinson, G. J. C. S. Dalton 1973, 204-
209.
(13) (a) Heimer, T. A.; D’Arcangelis, S. T.; Farzad, F.; Stipkala, J. M.;
Meyer, G. J. Inorg. Chem. 1996, 35, 5319-5324. (b) Galoppini, E.; Guo,
W.; Zhang, W.; Hoertz, P. G.; Qu, P.; Meyer, G. J. J. Am. Chem. Soc.
2002, 124, 7801-7811.
(14) Castellano, F. N.; Heimer, T. A.; Thandasetti, M.; Meyer, G. J.
Chem. Mater. 1994, 6, 1041-1048.
(15) (a) Arnold, D. R.; Baird, N. C.; Bolton, J. R.; Brand, J. C. D.;
Jacobs, P. W. M.; DeMayo, P.; Ware, W. R. Photochemistry. An Introduc-
tion; Academic Press: New York and London, 1974; p 13. (b) Rehm, D.;
Weller, A. Isr. J. Chem. 1970, 8, 259-267.
(32) Fleming, C. N.; Maxwell, K. A.; DeSimone, J. M.; Meyer, T. J.;
Papanikolas, J. M. J. Am. Chem. Soc. 2001, 123, 10336-10347.
(33) Long-lived excited states have been observed on reduced TiO2 under
conditions where reductive quenching like that shown in Scheme 3C would
be expected. (a) O’Regan, B.; Moser, J. E.; Anderson, M.; Gra¨tzel, M. J.
Phys. Chem. 1990, 94, 8720-8726. (b) Kamat, P. V.; Bedja, I.; Hotchandi,
S.; Patterson, L. K. J. Phys. Chem. 1996, 100, 4900-4905. (c) Heimer, T.
A.; Meyer, G. J. J. Lumin. 1996, 70, 468-478. (d) Lemon, B. I.; Hupp, J.
T. J. Phys. Chem. B 1999, 103, 3797-3799.
(34) (a) Swager, T. M. Acc. Chem. Res. 1998, 31, 201-207. (b) Chen,
L.; McBranch, D. W.; Wang, H.-L.; Helgeson, R.; Wudl, F.; Whitten, D.
G. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 12287-12292. (c) Harrison, B.
S.; Ramey, M. B.; Reynolds, J. R.; Schanze, K. S. J. Am. Chem. Soc. 2000,
122, 8561-8562.
(35) (a) Argazzi, R.; Bignozzi, C. A.; Heimer, T. A.; Castellano, F. N.;
Meyer, G. J. J. Am. Chem. Soc. 1995, 117, 11815-11816. (b) Bonhote, P.;
Moser, J.-E.; Humphry-Baker, R.; Vlachopoulos, N.; Zakeeruddin, S. M.;
Walder, L.; Gratzel, M. J. Am. Chem. Soc. 1999, 121, 1324-1331. (c)
Kleverlaan, C. J.; Alebbi, M.; Argazzi, R.; Bignozzi, C. A.; Hasselmann,
G. M.; Meyer, G. J. Inorg. Chem. 2000, 39, 1342-1343.
(16) Kelly, C. A.; Thompson, D. W.; Farzad, F.; Meyer, G. J. Langmuir
1999, 15, 731-737.
(17) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 2nd ed.;
Plenum Press: New York, 1999.
(18) Bergeron, B. V. Unpublished results.
(19) (a) Crutchley, R. J.; Lever, A. B. P. J. Am. Chem. Soc. 1980, 102,
7128-7130. (b) Prasad, D. R.; Hessler, D.; Hoffman, M. Z.; Serpone, N.
Chem. Phys. Lett. 1985, 121, 61-64. (c) Rillema, D. P.; Allen, G.; Meyer,
T. J.; Conrad, D. Inorg. Chem. 1983, 22, 1617-1622.
(20) Kalyanasundaram, K. Photochemistry of Polypyridine and Por-
phyrin Complexes; Academic Press: London, 1992.
(21) (a) Farzad, F.; Thompson, D. W.; Kelly, C. A.; Meyer, G. J. J.
Am. Chem. Soc. 1999, 121, 5577-5578. (b) Trammell, S. A.; Yang, J.;