Characterization of photoinduced self-exchange reactions at molecule-semiconductor interfaces by transient polarization spectroscopy: Lateral intermolecular energy and hole transfer across sensitized TiO 2 thin films

Article abstract of DOI:10.1021/ja200652r

Transient anisotropy measurements are reported as a new spectroscopic tool for mechanistic characterization of photoinduced charge-transfer and energy-transfer self-exchange reactions at molecule-semiconductor interfaces. An anisotropic molecular subpopul

Full text of DOI:10.1021/ja200652r

ARTICLE
pubs.acs.org/JACS
Characterization of Photoinduced Self-Exchange Reactions at
MoleculeꢀSemiconductor Interfaces by Transient Polarization
Spectroscopy: Lateral Intermolecular Energy and Hole Transfer
across Sensitized TiO Thin Films
2
Shane Ardo and Gerald J. Meyer*
Departments of Chemistry and Materials Science and Engineering, Johns Hopkins University, 3400 North Charles Street, Baltimore,
Maryland 21218, United States
S Supporting Information
b
ABSTRACT: Transient anisotropy measurements are reported as a new spectro-
scopic tool for mechanistic characterization of photoinduced charge-transfer and
energy-transfer self-exchange reactions at moleculeꢀsemiconductor interfaces. An
anisotropic molecular subpopulation was generated by photoselection of randomly
oriented Ru(II)ꢀpolypyridyl compounds, anchored to mesoscopic nanocrystalline
TiO or ZrO thin films, with linearly polarized light. Subsequent characterization
2
2
of the photoinduced dichromism change by visible absorption and photolumines-
cence spectroscopies on the nano- to millisecond time scale enabled the direct
comparison of competitive processes: excited-state decay vs self-exchange energy
transfer, or interfacial charge recombination vs self-exchange hole transfer. Self-
exchange energy transfer was found to be many orders-of-magnitude faster than
hole transfer at the sensitized TiO interfaces; for [Ru(dtb) (dcb)](PF ) , where
2
2
6 2
0
0
0
0
dtb is 4,4 -(C(CH ) ) -2,2 -bipyridine and dcb is 4,4 -(COOH) -2,2 -bipyridine,
3
3 2
2
anchored to TiO , the energy-transfer correlation time was θ = 3.3 μs while the average hole-transfer correlation time was
2
ent
Æθ æ = 110 ms, under identical experimental conditions. Monte Carlo simulations successfully modeled the anisotropy decays
h+
associated with lateral energy transfer. These mesoscopic, nanocrystalline TiO thin films developed for regenerative solar cells
2
thus function as active “antennae”, harvesting sunlight and transferring energy across their surface. For the dicationic sensitizer,
[
Ru(dtb) (dcb)](PF ) , anisotropy changes indicative of self-exchange hole transfer were observed only when ions were present in
2 6 2
the acetonitrile solution. In contrast, evidence for lateral hole transfer was observed in neat acetonitrile for a neutral sensitizer, cis-
0
0
Ru(dnb)(dcb)(NCS) , where dnb is 4,4 -(CH (CH ) ) -2,2 -bipyridine, anchored to TiO . The results indicate that mechanistic
2
3
2 8 2
2
models of interfacial charge recombination between electrons in TiO and oxidized sensitizers must take into account diffusion of
2
the injected electron and the oxidized sensitizer. The phenomena presented herein represent two means of concentrating potential
energy derived from visible light that could be used to funnel energy to molecular catalysts for multiple-charge-transfer reactions,
such as the generation of solar fuels.
1
,2
’
INTRODUCTION
have been garnered. Self-exchange reactions are also of interest
in natural and artificial photosynthesis, as they provide a general
mechanism by which energy or charge can be transferred be-
Outer-sphere electron and energy self-exchange represent a
simple class of reactions in which no bonds are broken and no
bonds are formed, eqs 1 and 2, respectively, where bpy is 2,2 -
bipyridine.
1,2
0
tween molecules without the loss of free energy. For example,
in photosynthetic light harvesting antennae systems, chloro-
phyll*-to-chlorophyll energy transfer can rapidly and quantita-
tively deliver energy to a reaction center, behavior that arises
from a recently observed quantum coherent energy-transfer me-
chanism wherein nearly perfect excited-state alignment and envir-
onmental screening inhibit collisional decoherence. Self-exchange
energy transfer has also been observed in artificial photosynthetic
assemblies like those based on the dye-sensitized, mesoscopic nano-
crystalline TiO thin films developed by Gr €a tzel and co-workers for
3
RuðbpyÞ3 þ þ RuðbpyÞ33þ f RuðbpyÞ33þ þ RuðbpyÞ32þ
2
ð1Þ
4
ꢁ
ꢁ
RuðbpyÞ3 þ þ RuðbpyÞ32þ f RuðbpyÞ32þ þ RuðbpyÞ32 þ
2
ð2Þ
2
The equivalence of the reactants and products simplifies theore-
tical expressions and has enabled critical analysis of experimental
data from which deep insights into the intrinsic reaction barriers
Received: February 3, 2011
Published: August 23, 2011
r 2011 American Chemical Society
15384
dx.doi.org/10.1021/ja200652r
|J. Am. Chem. Soc. 2011, 133, 15384–15396

Journal of the American Chemical Society
ARTICLE
5
ꢀ8
application in regenerative photoelectrochemical cells.
Inter-
Scheme 1. Cross Section of a Hypothetical MoleculeꢀTiO2
estingly, lateral intermolecular self-exchange electron transfer,
like that in eq 1, can also be induced in these same materials with
Arrangement for Ru-Trisbipyridyl Coordination Compounds
a
Anchored to a Spherical TiO Nanocrystallite
2
9
ꢀ13
an electrochemical bias.
Historically, such self-exchange reactions have been difficult to
quantify experimentally in heterogeneous materials due to the
1
equivalence of reactants and products. Conventional line-broad-
ening measurements are nontrivial and complicated by interfacial
heterogeneity that can in itself result in significant spectral
1
,2
broadening. Such heterogeneity may induce a distribution of
environments such that the reactants and products are not truly
identical. It is therefore a simplification to consider eqs 1 and 2 as
self-exchange processes in photosynthetic assemblies where ag-
gregation, partial solvation, and other environmental factors may
alter the reaction chemistry. Nevertheless, there still exists a real
need to quantify interfacial “self-exchange reactions” that may
be utilized for solar energy conversion and to provide funda-
mental insights into interfacial chemistry. For example, in dye-
sensitized solar cells, one might want to know whether an
oxidized dye molecule formed after excited-state injection stays
localized on the injecting dye or hops to its neighbor or to a
catalytic site. The extent to which lateral photoinduced self-
exchange occurs at any semiconductor surface remains largely
unknown. Given this impetus, we report herein transient
1
4,15
anisotropy measurements
as a new spectroscopic assay
for mechanistic characterization of photoinduced charge-trans-
fer and energy-transfer self-exchange reactions at moleculeꢀ
semiconductor interfaces.
The anisotropy assay can be understood with the aid of the
idealized TiO nanocrystallite with surface-anchored Ru(II)ꢀ
2
polypyridyl compounds, shown in Scheme 1, where each com-
pound’s lowest-energy charge-transfer transition dipole mo-
ment is depicted by a singly degenerate, vectorial transition (blue
arrow). An anisotropic molecular subpopulation can be gen-
erated by photoselection of these randomly oriented com-
pounds with linearly polarized light excitation. The magnitude
of the overlap between the molecular transition dipole moments
and the polarization vector of the excitation light (polarized
vertically and propagating in the plane of the page) is depicted as
thick brown arrows, where ε is the absorption coefficient
measured in isotropic fluid solution. This illustrates that com-
pounds positioned closer to the vertical poles of a nanocrystallite
will be preferentially photoexcited by the incoming vertically
polarized light relative to those near the equator. For anchored
molecules on immobilized nanoparticles, like those sintered
together in the mesoporous thin films employed here, the
direction of the photoselected transition dipole moments will move
little due to molecular or nanoparticle diffusion; however, when
molecules participate in lateral intermolecular energy- or hole-
transfer reactions, the detected transition dipole moments will
change. With pulsed-laser excitation, anisotropy measurements
thus provide information on energy- and hole-transfer self-
exchange dynamics that are central to this report.
a
For the sake of clarity, only five molecules are depicted while ∼400 are
expected for each anatase nanocrystallite (∼15 nm in diameter).
Polarized light excitation was assumed to be vertical, propagating in
the plane of the page, with detection in the direction of the reader.
due to lateral self-exchange reactions that translate the energy or
charge around the spherical nanoparticle.
This anisotropy approach enabled the determination of self-
exchange kinetic data under conditions where unwanted recom-
bination processes were operative. Self-exchange energy transfer
occurred in competition with excited-state decay, while self-
exchange hole transfer often occurred on the time scale of
interfacial charge recombination. The results presented below
indicate that interfacial self-exchange energy-transfer reactions
are many orders-of-magnitude faster than hole-transfer reactions
at these sensitized TiO interfaces. Furthermore, hole transfer
2
was found to be sensitive to the identity of the rutheniumꢀ
polypyridyl compound and the chemical environment. The
implications of these findings for solar energy conversion are
discussed.
’
EXPERIMENTAL SECTION
We note that fluorescence anisotropy has widely been utilized
14,15
as a rotational probe in biophysical assays.
In such assays, θ is
Materials. The following reagents and substrates were reagent grade
calculated in the same manner as described herein and is often
referred to as the rotational correlation time. In the present work,
θ is now understood to be a self-exchange correlation time, and
hence, its significance differs from that of physical rotation of
molecules. The surface-anchored molecules are not undergoing
rotational diffusion; instead, the loss in anisotropy over time is
or better and were used as received from the indicated commercial
suppliers: acetonitrile (Burdick & Jackson, spectrophotometric grade);
absolute, anhydrous ethanol (PharmcoꢀAaper, ACS/USP grade, > 99.5%);
methanol (SigmaꢀAldrich, spectrophotometric grade, > 99.9%); acetone
(bulk solvent); lithium perchlorate (Aldrich, 99.99%); tetra-n-butylammo-
nium perchlorate (TBAClO ; Fluka, > 99.9%); tetra-n-butylammonium
4
1
5385
dx.doi.org/10.1021/ja200652r |J. Am. Chem. Soc. 2011, 133, 15384–15396

Journal of the American Chemical Society
ARTICLE
hydroxide (TBAOH; Fluka, 1 M aqueous); cobalt(III) meso-5,10,15,20-
tetrakis(4-carboxyphenyl)porphyrin chloride (Frontier Scientific); ar-
gon gas (Airgas, > 99.998%); oxygen gas (Airgas, industrial grade);
titanium(IV) isopropoxide (SigmaꢀAldrich, 97%); zirconium(IV)
propoxide (Aldrich, 70 wt % solution in 1-propanol); fluorine-doped,
Stokes-shifted 683 nm excitation. A polarizer was employed even though
the vertical polarization was nearly maintained due to the extremely
20ꢀ22
small depolarization ratio of H
2
.
The excitation fluence was measured
by a thermopile power meter (Molectron) and was ∼200 μJ/pulse so
that the absorbed fluence was typically e50 μJ/pulse, unless noted
otherwise. A 150 W xenon arc lamp served as the probe (Applied Photo-
physics) and was aligned orthogonal to the laser excitation light. Before
arriving at the sample, the probe was directed through a 1/4 m mono-
chromator(SpectralEnergy Corp., GM 252). For detectionon sub-100μs
time scales, the lamp was pulsed with 100 V. Detection was achieved in
a T format with a monochromator (Spex 1702/04) optically coupled
to an R928 photomultiplier tube (Hamamatsu). For some measure-
ments a GlanꢀTaylor polarizer positioned at the magic angle (∼54.7ꢀ)
relative to the polarization of the excitation light was placed before the
detection monochromator to remove polarization-dependent signals.
Transient data was acquired on a computer-interfaced digital oscillo-
scope (LeCroy 9450, Dual 350 MHz). The instrument response time
was ∼10 ns. Typically, 100 laser pulses were averaged at each observa-
tion wavelength and two to six identical measurements were taken and
averaged, to help increase the signal-to-noise ratio of the magic angle
and anisotropy data.
SnO -coated glass (FTO; Hartford Glass Co., Inc., 2.3 mm thick,
2
15 Ω/0); and glass microscope slides (Fisher Scientific, 1 mm thick).
The sensitizers employed were available from previous studies:
1
6
0
0
[
Ru(dtb) (dcb)](PF ) , where dtb is 4,4 -(C(CH ) ) -2,2 -bipyri-
2
6 2
0
3 3 2
0
dine and dcb is 4,4 -(COOH)
2
-2,2 -bipyridine; [Ru(bpy)
2
(deebq)]-
1
7
0
16
0
(
PF
6
)
2
,
where bpy is 2,2 -bipyridine and deebq is 4,4 -diethyl
0
ester-2,2 -biquinoline; Z907, cis-Ru(dnb)(dcb)(NCS)
2
, where dnb is
0 0
,4 -(CH (CH ) ) -2,2 -bipyridine; and N3, cis-Ru(dcb) (NCS) .
3 2 8 2 2 2
18
4
Sensitized Metal Oxide Thin Film. Transparent TiO
anatase, ∼15 nm in diameter) and ZrO nanoparticles were prepared
by hydrolysis of the appropriate precursors [Ti(i-OPr) or Zr(OPr) ]
2
nanocrystallites
(
2
4
4
9
using a solꢀgel technique previously described in the literature. The
sols were cast as mesoporous thin films using Scotch transparent
film tape as a spacer (∼10 μm thick) by doctor blading onto glass
microscope slides for spectroscopic measurements and transparent
FTO conductive substrates for electrochemical measurements. In
all cases, the thin films were annealed at 420 ꢀC for 30 min under
All measurements employed an excitation polarizer (Pex) before the
sample and a detection polarizer (Pdet) after the sample. For anisotropy
2
O flow. Imaging of platinum-sputtered thin film samples on FTO
substrates was performed with a JEOL JSM-6700F cold cathode
field emission scanning electron microscope (SEM) to determine
the average nanoparticle size.
measurements, P was set to vertical, the same polarization of the laser,
ex
and Pdet was set to either vertical (V) or horizontal (H) such that the
magic angle and anisotropy values could be calculated via eqs 4 and
1
4,15
Thin films were first pretreated with aqueous base (TBAOH, pH 11)
for 15 min followed by an acetone wash and heated to 75 ꢀC. The films
were then briefly placed in the neat solvent that was to be used for surface
binding to wet the surface with the desired solvent. Sensitization for all
coordination compounds was achieved by immersing the supported thin
films in sensitizer solutions (nanomolar to millimolar concentrations)
for minutes to days. Low surface coverage films were prepared by over-
night reactions in low concentration dying solutions. Films were then
immersed in the neat solvent that was used for sensitizer binding,
acetonitrile, absolute ethanol, or DMSO, for 5ꢀ10 min, followed by a
thorough washing with the experimental solvent. Unless noted other-
wise, the thin films were sensitized to roughly maximum surface cover-
5,
respectively
ðIVV þ 2GIVHÞ
ΔAbsmagic angle ¼
ð4Þ
ð5Þ
3
ðIVV ꢀ GIVHÞ
ΔAbsanisotropy ¼
3ΔAbsmagic angle
where IVY is the intensity of the detected light with excitation polariza-
tion V and detection polarization Y = V or H, and G is a wavelength-
dependent correction factor for the polarization-dependent response of
the detection system. Corrections for the wavelength-dependent polar-
ization response were not necessary in the transient absorption data due
to their cancellation during calculation of ΔAbs, and thus G = 1. Unless
specifically mentioned, all time-resolved pumpꢀprobe experiments were
performed under conditions where less than one excited-state was formed
per nanoparticle.
ꢀ
8
2
age Γ ≈ (5ꢀ7) ꢂ 10 mol/cm , which was calculated by a previously
11
published method. Briefly, the molar decadic extinction (absorption)
ꢀ
1
ꢀ1
coefficient (ε in M cm ) at the maximum of the metal-to-ligand
charge transfer transition was assumed to be the same in solution and on
the surface. This value was used along with the modified BeerꢀLambert
2
law formula to calculate a macroscopic surface coverage (Γ in mol/cm )
Photoluminescence. Steady-state photoluminescence (PL) measure-
ments were obtained with a fluorimeter (SLM Instruments SLM-
by eq 3,
4
8000S) employing a Rhodamine B/ethylene glycol “quantum coun-
Abs ¼ εcl ¼ 1000 ꢂ εΓ
ð3Þ
ter”. The PL spectra were collected near room temperature and were
corrected for the wavelength-dependent response of the excitation and
emission detection systems by calibration with a tungstenꢀ(halogen)
irradiance-standard lamp. The SLM Instruments SLM-48000S consisted
of a single excitation monochromator (MC200) optically coupled to a
450 W xenon arc lamp and a single detection monochromator (MC200)
with a GaAs photomultiplier tube (Hamamatsu, R928P). The excitation
beam was directed 45ꢀ to the film surface, and the emitted light was
monitored at a right angle. The SSPL spectra were collected at 21.0(1) ꢀC
with calcite polarizers and a 585 nm long-pass filter.
Time-resolved PL data were acquired on the same apparatus used for
the transient-absorption measurements in an L format. Corrections for
the polarization-dependent response of the detection system were
achieved by calculating G factors. A two-mirror beam-steering device
was used to convert the polarization of the excitation laser to a ∼45ꢀ
linear polarization, so that the correction factor, G, could be measured.
This allowed for in situ correction of the wavelength-dependent
polarization response of the detection system for each sample by setting
The samples were then positioned diagonally in a 1 cm cuvette
containing the experimental solution. For transient absorption and
electrochemical studies, the cuvettes containing the sample and electro-
lyte solution were purged with Ar(g) for at least 30 min prior to
experimentation. Unless noted otherwise, all thin films were sensitized
to near maximum surface coverage and immersed in Ar-purged acet-
onitrile at 21 ꢀC.
Spectroscopy. UVꢀVisible Absorption. Steady-state UVꢀvisible
absorbance spectra were obtained on a Varian Cary 50 spectrophoto-
meter at room temperature.
Nanosecond transient absorption measurements were obtained with
1
9
an apparatus similar to that which has been previously described.
Briefly, samples were excited by a Q-switched, pulsed Nd:YAG laser
Quantel USA (BigSky) Brilliant B; 5ꢀ6 ns full width at half-maximum
[
(
fwhm), 1 Hz, ∼10 mm in diameter] directed 45ꢀ to the film surface and
tuned to vertically polarized 532 nm light with the appropriate nonlinear
optics. An H -filled Raman shifter (∼400 psi) was employed to obtain
2
1
5386
dx.doi.org/10.1021/ja200652r |J. Am. Chem. Soc. 2011, 133, 15384–15396

Journal of the American Chemical Society
ARTICLE
Scheme 2. Chemical Structures of the Transition-Metal Coordination Compound Sensitizers, and Their Abbreviations, Used in
the Studies
P
ex to horizontal so that the correction factor, G = IHV/IHH, could be
Figure 1 depicts normalized magic-angle photoluminescence
(PL) excitation and PL spectra for Ru(dtb) (dcb)/ZrO im-
mersed in neat acetonitrile at room temperature in the presence
calculated. See the Supporting Information for further details of the
correction method. For each wavelength, the calculated G factor was
then used to scale the IVH values as explicitly written in eqs 4 and 5.
Variable-temperature PL spectra were acquired with a Neslab ULT95
bath circulator equipped with a four-window Dewar filled with methanol.
The temperature was monitored with a thermocouple probe (Omega
DPi32) and was allowed to equilibrate at each temperature for at least
2
2
and absence of LiClO . The excitation spectra were found to be
4
in good agreement with the absorptance spectra, which is related
ꢀ
Abs
to the absorbance by 1 ꢀ 10
, measured under the same
conditions. Broad overlapping Ru f dcb and Ru f dtb metal-to-
ligand charge transfer (MLCT) absorptions bands were observed
in the visible region (Figure S1a, Supporting Information). The
absorption and PL spectra displayed a noticeable red shift in the
presence of this electrolyte; however the PL excitation spectra
varied very little (Figure S2a, Supporting Information). The PL
1
5 min prior to a measurement. For measurements at ∼77 K, a 2ꢀ3 mm
wide sample was submerged in MeOH:EtOH (∼1:4, v/v) in a standard
NMR tube and immersed into an N (l)-filled glass coldfinger Dewar
2
such that only about half of the MeOH:EtOH solution was sub-
merged in N
and those that did not crack were used.
Electrochemistry. A potentiostat (Epsilon electrochemical ana-
lyzer) was employed for measurements in a standard three-electrode
arrangement with a sensitized TiO thin film deposited on an FTO sub-
2
strate working electrode, a Pt gauze counter electrode, and an aqueous
Ag/AgCl (NaCl saturated) reference electrode (Bioanalytical Scientific
Instruments, Inc.). The ferrocenium/ferrocene [Fe(Cp)2 ] half-wave
potential measured in a 200 mM LiClO acetonitrile electrolyte was used
as an internal standard. Conversion to normal hydrogen electrode (NHE)
used the published values for the reference electrode, i.e. +197 mV vs
2
(l). The resultant solvent glass cracked ∼80% of the time,
intensity decreased in the presence of LiClO (data not shown).
4
17
The absorption spectrum of Ru(bpy) (deebq)/ZrO2 displayed
2
well-resolved Ru f deebq (550 nm) and Ru f bpy (450 nm)
MLCT absorption bands (Figure S1b, Supporting Information).
The absorption spectra of all compounds were the same within
experimental error when anchored to ZrO or TiO thin films.
2
2
+
/0
At low laser fluences, the magic angle and anisotropy time-
resolved PL kinetics were well-modeled as first-order decay
processes, eq 6
4
ꢂ ꢀ ꢁꢃ
23
+/0
t
NHE, and corrected for the expected E1/2(Fe(Cp)
the KCl-saturated aqueous calomel electrode (SCE), where SCE is
2
) of +310 mV vs
I ¼ Ioexp ꢀ
ð6Þ
h
23
+
241.2 mV vs NHE.
Computation. Data Analysis. Kinetic and spectral data were
where for PL kinetics I = PL and h is the characteristic lifetime, τ,
and for anisotropy kinetics I = r and h is the anisotropy
correlation time constant, θent. Mean values from multiple
measurements/samples were reported with standard deviation
of the final digit in parentheses. The kinetics was independent of the
monitoring wavelength beyond 625 nm with lifetime [τ = 1.04(3)
vs 0.68(1) μs] and anisotropy correlation time [θent = 2.8(3) vs
modeled in Origin 7.0. Least-squares error minimization was accom-
plished with the LevenbergꢀMarquardt iteration method. The photo-
luminescence kinetic data were smoothed via adjacent averaging ((5
points) before calculating the magic-angle and anisotropy data. The
transient-absorption kinetic data were first converted to log-time space
and then interpolated to 0.01 logarithmic deviations, followed by smoothing
via adjacent averaging ((5 points). For the spectral modeling, a method
for the standard addition of known spectra, written in the C program-
ming language, was implemented in Origin’s error minimization routine.
Simulations of Anisotropy Data. Monte Carlo simulations to model
the energy-transfer anisotropy decay kinetics and calculate the average
number of self-exchange hole-transfer reactions required to reach a
metalloporphyrin catalyst were performed with Wolfram Mathematica
2
.0(3) μs] in the absence and presence of LiClO , respectively. The
4
most significant effect introduced by addition of LiClO was a
4
decrease in the excited-state lifetime with lesser change in the PL
anisotropy (Figure S2b, Supporting Information).
Monte Carlo simulations under vertically polarized simulated
low light conditions, i.e., only one excited state per nanoparticle,
could accurately reproduce the first-order anisotropy decays,
with the nearest-neighbor hopping time constant as the only
adjustable parameter. The energy-transfer model employed
spherical nanoparticles (15 nm in diameter) with a maximum,
near evenly distributed packing of sensitizers longitudinally and
7.0 on a PC running an Intel Core2 Duo CPU P9700 at 2.79 GHz.
’
RESULTS
2
4
The heteroleptic sensitizers utilized for the self-exchange
latitudinally at a minimum of van der Waals contact (7 Å)
(Scheme 3a). The initial excited-state distribution was randomly
studies along with their abbreviations are shown in Scheme 2.
1
5387
dx.doi.org/10.1021/ja200652r |J. Am. Chem. Soc. 2011, 133, 15384–15396

Journal of the American Chemical Society
ARTICLE
Figure 1. (A) Normalized magic angle photoluminescence (PL) spectra for a Ru(dtb)
2
(dcb)/ZrO
2
thin film before (filled black circles) and after (open
4
red triangles) introduction of 100 mM LiClO . Photoluminescence spectra (right) were acquired with λex = 532 nm, while excitation spectra (left) were
monitored at the PL maxima (λmon-neat = 650 nm, λmon-Li+ = 680 nm). (B) Normalized magic angle (open magenta circles) and anisotropy (open blue
circles) PL changes (λmon = 700 nm) after pulsed 532 nm laser excitation (λ = 532 nm) for the thin film from panel A immersed in neat acetonitrile, on a
ex
logarithmic intensity scale. Overlaid on the anisotropy changes, as a solid gray line, is a representative Monte Carlo simulation from which a Dexter
ꢀ
1
energy-transfer hopping rate constant of (710 ns) was abstracted.
selected in spherical polar coordinates with equal circular latitu-
dinal weights and longitudinal weights of (cos ϕ) with appro-
Scheme 3. Monte Carlo Simulations for a Maximum Packing
2
Coverage of RuꢀTrisbipyridyl Coordination Compounds on
a,b
priate (sin ϕ) degeneracy, where ϕ is the inclination angle. Energy
transfer at any distance beyond van der Waals contact was
modeled by a Dexter-type mechanism with an exponential decay
a Spherical Nanocrystallite
ꢀ
1
for the probability of transfer, βtunnel. Given βtunnel = 0.35 Å
,
which was determined previously for energy-transfer processes
25
between metalꢀpolypyridyl compounds on ZrO and by far
2
26
represents a lower limit for most solvents, a transfer to a
non-nearest-neighbor occurred, on average, about once per 60
simulated hops.
The probability that an excited state hopped to any other
molecule was calculated using the rate constant at van der Waals
contact, the degeneracy of each hop (i.e., the number of locations
per circular latitude from the excited sensitizer), and an expo-
nentially decaying distance-dependent term employing the elec-
tronic-coupling matrix-element βtunnel factor. Random numbers
were generated with an extended cellular automaton pseudoran-
dom number generator. Mathematical movement of the excited
state was controlled using Quaternion matrices and conversion
of the excited state’s location to Cartesian coordinates, which
allowed for ease of its movement and regeneration of the dis-
a
Even distribution of molecules (red spheres) anchored to a 15 nm
diameter nanocrystallite, assuming a 7 Å van der Waals radius, which is
depicted for the seven molecules at the top pole. Monte Carlo
b
simulation (ꢀ) for energy transfer across a single nanocrystallite, based
on a Dexter mechanism, for the initial distribution shown in panel A
from the initially photoexcited molecule (green sphere) to the final
molecule (red sphere) in the simulation. Also shown as blue arrows are
the approximate transition dipole moments for the excited states
involved in the simulation.
27
tribution of neighboring sensitizers. Conversion back to sphe-
rical polar coordinates allowed the inclination angle, and thus the
relative angular location of the excited molecule versus the initial
vertical polarization of the excitation light, to be retained. This
ꢀ
1
28
rate constant per energy transfer “hop” of (120 ns) , which
1
also made calculation of the anisotropy straightforward via eq 7,
ꢀ
included nearest-neighbor degeneracy, and thus (710 ns) for a
single self-exchange reaction (Figure 1b).
2
3
hcos ϕi ꢀ 1
ð7Þ
ꢀ8
r ¼
The absorption spectra of one of the highest (5.0 ꢂ 10 mol/
2
2
ꢀ9
2
cm ) and one of the lowest (7 ꢂ 10 mol/cm ) Ru(dtb) -
2
where (cos ϕ) for each sensitizer was simply the cosine of its in-
clination angle stored in the spherical polar coordinate matrix. This
process was repeated for a given number of iterations (Scheme 3b),
corresponding to an amount of time, and then the resulting
anisotropy decay was generated. Care was taken such that the
probability of a transfer event per time period, i.e. frequency
factor, was <1%, so as not to mistakenly miss two hopping events
per random number generated. Under conditions of maximum
sensitizer coverage, the experimental data was well-modeled using a
(dcb)/ZrO surface-coverage thin films are depicted in Figure 2a.
2
PL decays measured at the magic angle and anisotropy kinetics
for the same materials are shown in Figure 2b. At low surface
coverage, the excited-state lifetime was within error the same,
whereas the rate of anisotropy decay decreased slightly. In general,
there was only a minor increase in the anisotropy decay rate with
surface coverage (Figure S3, Supporting Information). Note that
the initial fundamental anisotropy at 700 nm was significantly
smaller than the theoretical maximum of 0.4 for both surface
1
5388
dx.doi.org/10.1021/ja200652r |J. Am. Chem. Soc. 2011, 133, 15384–15396

Journal of the American Chemical Society
ARTICLE
Figure 2. (A) UVꢀVis absorption spectra of two Ru(dtb)
2
(dcb)/ZrO
2
thin films immersed in Ar-purged neat acetonitrile at 21 ꢀC at the indicated
surface coverages. (B) PL magic-angle (bottom data; left axis) and anisotropy (top data; right axis) changes (λ = 532 nm, λ
= 700 nm) for the thin
ex
mon
films from panel A, where the dashed kinetics correspond to the sample with the higher surface coverage, on logarithmic intensity scales. Similar behavior
was also observed for Ru(dtb) (dcb)/TiO thin films (Figure S3, Supporting Information).
2
2
Figure 3. (A) Normalized magic angle and (B) anisotropy PL changes (λ = 532 nm, λ = 700 nm) of a Ru(dtb) (dcb)/ZrO thin film immersed in
ex
mon
2
2
Ar-purged, neat acetonitrile at the indicated temperatures (ꢀC), on logarithmic intensity scales. Similar behavior was also observed for a Ru(dtb)
thin film in a solvent glass (Figure S4, Supporting Information).
2 2
(dcb)/TiO
coverages, r ≈ 0.21(3). The initial anisotropy increased to
was the same within error (Figure S5, Supporting Information).
o
∼
0.35 when a Ru(bpy) (deebq)/ZrO -sensitized thin film was
The initial fundamental anisotropies were, within error, the same
but decayed with an increasing fraction of second-order relaxa-
tion as the pulsed 532 nm laser excitation fluence was increased.
2
2
used under otherwise identical conditions, which is reflected in
its steady-state PL spectrum at long wavelengths (Figure S1b,
Supporting Information).
The decays exhibited long-time limiting anisotropies, r ≈ 0, as
∞
Figure 3 depicts the temperature-dependence of the magic
explained in the Supporting Information.
angle and anisotropy PL signals of a Ru(dtb) (dcb)/ZrO thin
film. Both the magic angle and anisotropy PL signals were well-
described by a first-order kinetic model when monitored from
To promote excited-state injection, a Ru(dtb) (dcb)/TiO
thin film was immersed in a 10 mM LiClO /CH CN electrolyte.
4 3
2
2
2
2
This increased the yield for excited-state injection, but not to
unity. The magic angle and anisotropy signals for PL at 700 nm
and transient absorption at 486 nm after pulsed 532 nm laser
6
50 to 750 nm. Temperature-dependent magic angle kinetics
were apparent, τ = 0.97 vs 2.8 μs for 283 and 77 K, respectively;
however, the anisotropy signal was nearly temperature-indepen-
dent (Figure 3b and Figure S4b, Supporting Information). From
the magic angle PL kinetics at four temperatures, an activation
energy of 230 cm and frequency factor of 3.4 ꢂ 10 s were
abstracted via an Arrhenius analysis (Figure S4c, Supporting
Information).
2
excitation at 2.5 mJ/cm are shown in Figure 5. The entire PL
decays were nonexponential and well-described by a parallel first-
and second-order kinetic model; a best fit resulted in the lifetime
of the first-order component of τ = 1.2 μs. The anisotropy kinetics
were well-modeled as a first-order kinetic decay process at 300 ns
and beyond, resulting in an energy-transfer correlation time,
θent = 3.3 μs; this value was larger than that obtained under
low-fluence conditions, consistent with Figures 4b and S5b
(Supporting Information).
ꢀ
1
6
ꢀ1
Figure 4 depicts the fluence dependence to the magic angle
2+
and anisotropy PL decays of [Ru(dtb) (dcb)] anchored to
2
TiO thin films immersed in neat acetonitrile; data using ZrO2
2
1
5389
dx.doi.org/10.1021/ja200652r |J. Am. Chem. Soc. 2011, 133, 15384–15396

Journal of the American Chemical Society
ARTICLE
Figure 4. (A) Normalized magic angle and (B) anisotropy PL changes (λ = 532 nm, λ = 700 nm) for a Ru(dtb) (dcb)/TiO thin film immersed in
ex
mon
2
2
2
Ar-purged, neat acetonitrile at 21 ꢀC at the indicated absorbed photon fluxes (in nmol/cm ), using the same key and on logarithmic intensity scales. The
initial time points in panel B (open magenta circles) highlight the nearly coincident fundamental anisotropies. Similar behavior was also observed for
2 2
Ru(dtb) (dcb)/ZrO thin films (Figure S5, Supporting Information).
while for the self-exchange hole-transfer process, θh+ = 16.7 μs,
using the same βKWW value. Note that a higher laser fluence was
required under these low injection conditions to generate
enough charge-separated states for anisotropy analysis.
Figure 6a shows the absorption spectra of Z907/TiO and
2
N3/TiO2 thin films immersed in neat acetonitrile. Pulsed
683 nm laser excitation of either sensitized thin film resulted in
the prompt appearance of a transient-absorption spectrum con-
sistent with an interfacial charge-separated state, comprised of an
8
injected electron in TiO and an oxidized sensitizer, k > 10
s
2
inj
ꢀ
1
. Shown in Figure 6b are magic angle and anisotropy absorp-
tion transients monitored at 465 nm. For cis-Ru(dnb)(dcb)-
NCS) it was determined that τ = 14.5 μs and βKWW
(
=
2
KWW
0.183, from the magic angle data, and using the same βKWW value to
fit the anisotropy data, the hole-transfer correlation time was
determined to be θh+ = 9.6 μs. The anisotropy signals for the cis-
Ru(dcb) (NCS) /TiO sample displayed little change over the
lifetime of the charge-separated state and made quantification of
useful kinetic parameters difficult.
Figure 5. PL (λmon = 700 nm) and transient-absorption (λmon
=
2
2
2
4
86 nm), magic angle (bottom data, left axis), and anisotropy (top data,
2
right axis) after pulsed laser excitation (λ = 532 nm, 2.5 mJ/cm ) for a
ex
Ru(dtb)
2 2 4
(dcb)/TiO thin film immersed in Ar-purged 10 mM LiClO /
Transient absorption anisotropy changes for self-exchange hole
transfer were also quantified for Ru(dtb) (dcb)/TiO (Figure 5).
CH CN, on a logarithmic time scale.
3
2
2
Interestingly, the anisotropy measured after pulsed 532 nm laser
excitation of Ru(dtb) (dcb)/TiO at 465 nm was essentially
time-independent when neat acetonitrile was used as the external
solvent bath, but a surface-coverage-dependent rate of decay was
The measured transient-absorption data were fit to kinetic models
on 1 μs and longer time scales to ensure that there were no excited-
2
2
III
ꢀ
state contributions, such that Ru (dtb) (dcb)/TiO (e ) f
2
2
II
Ru (dtb) (dcb)/TiO charge recombination was solely ob-
served, where TiO (e ) stands for electrons in TiO . The
2
2
ꢀ
observed when g10 mM LiClO
time-resolved anisotropies measured for cis-Ru(dnb)(dcb)-
NCS) /TiO were, within experimental error, the same in the
4
was introduced. In contrast,
2
2
kinetics for charge recombination and the corresponding aniso-
tropy changes were nonexponential but were well-modeled by
the KohlrauschꢀWilliamsꢀWatts (KWW) model as a stretched-
exponential function, eq 8
(
2
2
presence and absence of 100 mM LiClO . A comparison of the
4
results for self-exchange energy- and hole-transfer reactions across
the surface of sensitized mesoporous thin films is shown in Table 1.
"
#
ꢀ
ꢁ
βKWW
Transient absorption studies of TiO thin films functionalized
2
t
I ¼ I exp ꢀ
ð8Þ
with CoTCPP [TCPP is meso-5,10,15,20-tetrakis(4-carboxyphe-
nyl)porphyrin] and Z907 in a 1:100 molar ratio were also
o
b
investigated. The strategy was that after excited-state injection
+
where βKWW is inversely related to the width of an underlying L ꢀe vy
distribution of rate constants, 0 < βKWW < 1, b = τKWW or θh+,
I = ΔAbs for transient absorption signals, and the other variables
were previously defined. The values obtained for the charge-
recombination process were τKWW = 0.14 μs and βKWW = 0.136,
into TiO , the Z907 formed was thermodynamically capable of
2
oxidizing the cobalt porphyrin from the formal oxidation state of
II to III by lateral hole transfer with a driving force of ΔGꢀ =
ꢀ1.1 eV. An isotropic distribution of molecules would result in
each porphyrin being separated by an average of four Z907
1
5390
dx.doi.org/10.1021/ja200652r |J. Am. Chem. Soc. 2011, 133, 15384–15396

Journal of the American Chemical Society
ARTICLE
Figure 6. (A) UVꢀvis absorption spectrum of a cis-Ru(dcb)
2
(NCS)
2
/TiO
2
(black, solid) and a cis-Ru(dnb)(dcb)(NCS)
2
/TiO
2
(purple, dashed) thin
film immersed in neat acetonitrile. (B) Transient absorption difference magic-angle (left axis) and anisotropy (right axis) changes for the samples in
2
panel A monitored at 465 nm after pulsed laser excitation (λ = 683 nm, ∼700 μJ/cm ), using the same color scheme, and on a logarithmic time scale.
ex
Overlaid in green, on the anisotropy kinetics, are fits to a stretched exponential function.
Table 1. Photophysical Properties of RuꢀPolypyridyl-Sensitized Thin Films in Acetonitrile Solutions at 21 ꢀC Obtained from
Time-Resolved Spectroscopic Measurements (λ = 532 nm), Monitored at the Indicated Wavelength
ex
a
b
c
d
sensitizer/film
Ru(dtb) (dcb)*/ZrO
λ
mon / nm
[LiClO
4
]/mM
τ/μs
β
KWW
θ/μs
r
o
r
ss-calculated
Φ
2
2
700
0
1.0
1
1
2.8
2.0
0.20
0.22
0.24
0.20
0.17
0.15
0.16
0.263
0.254
7
00
100
10
0.68
e
e
e
e
e
Ru(dtb)
2
(dcb)*/TiO
2
700
486
465
1.2
1
3.3
0.18
0.267
0.008
0.602
III
-
Ru (dtb)
2
2
(dcb)/TiO (e )
10
0.14
14.5
0.14
0.18
17
0.20
0.07
III
-
Ru (dnb)(dcb)(NCS) /TiO
2
2
(e )
100
9.6
a
b
The excited-state lifetime or the interfacial charge-separated-state lifetime. The anisotropy correlation time for energy transfer from first-order excited-
state anisotropy decays, θent, or the hole-transfer anisotropy correlation time abstracted from analsyis of the charge-recombination kinetics with the
KWW function, θh+. The steady-state anisotropy calculated via the Perrin equation. The calculated quantum yield for randomization of an initial
c
28 d
e
anisotropic population per eq 10. A parallel first- and second-order kinetic model was required to adequately model the data that was acquired at higher
laser fluence; solely the first-order components were used to determine rSS and Φ.
sensitizers. Pulsed-light excitation of Z907 resulted in subnano-
assignment of the data as being due to self-exchange reactions is
described below, followed by a discussion of the relevance of each
self-exchange reaction to solar energy conversion.
+
ꢀ
second excited-state injection to yield Z907 /TiO (e ) and the
2
appearance of a very small concentration of oxidized porphyrins,
III
+
Co TCPP . Over hundreds of microseconds, a sharp absorp-
Assignment as Self-Exchange Reactions. Unlike the situa-
tion normally encountered in fluid solution, the anisotropy decay
measured for the rutheniumꢀpolypyridyl compounds anchored
to TiO or ZrO mesoporous thin films was not attributed to
tion growth at ∼440 nm and bleach at ∼420 nm appeared that
II
III
were expected for oxidation of Co to Co . Spectral modeling
indicated that by 85 μs, ∼10% of the porphyrins had been
oxidized with a hole-transfer quantum yield of ∼3% based on the
2
2
physical motion of the compounds but to self-exchange energy-
or hole-transfer reactions across the surface of the nearly spherical
nanoparticles. This assignment was supported by the following
observations. Under conditions where the excited-state injection
yield was less than unity, the anisotropy decays associated with
energy and hole transfer measured on the same sensitized thin
film occurred with drastically different rates (Figure 5). If the
compounds were indeed mobile, one would anticipate that both
the excited and oxidized forms would translate and/or rotate
with nearly identical rates, contrary to what was observed. An
approximately order-of-magnitude increase in the correlation
time between self-exchange energy- and hole-transfer reactions,
+
number of Z907 initially created. Control experiments done in
the absence of Z907 showed no evidence for oxidation of the
porphyrin catalyst.
’
DISCUSSION
A new photoselection approach to monitor self-exchange
reactions between molecules anchored to semiconductor mate-
rials enabled the relative rates for isoenergetic, lateral energy- and
hole-transfer reactions, and their competing undesirable exergo-
nic recombination reactions, to be quantified. Specifically, it was
determined that self-exchange energy-transfer reactions, to ran-
domize initially photoselected anisotropic conditions, occur over
the nano- to microsecond time scales. It was also shown that hole-
transfer reactions are greatly affected by the choice of sensitizer and
electrolyte and occur over longer time scales. The basis for the
θ
ent = 3.3 μs vs θh+ = 17 μs, respectively, was measured after light
excitation of Ru(dtb) (dcb)/TiO . Furthermore, if an average
2
2
hole-transfer correlation time, Æθh+æ, is calculated as the first
moment of the underlying L ꢀe vy distribution of rate constants by
1
5391
dx.doi.org/10.1021/ja200652r |J. Am. Chem. Soc. 2011, 133, 15384–15396

Journal of the American Chemical Society
ARTICLE
2
9
the following relation (eq 9),
energy transfer measurably. In addition, the previously reported
ꢀ1
(
30 ns) hopping rate constant was obtained when hops to
ꢀ
ꢁ ꢀ
ꢁ
θ
1
beyond nearest neighbors were forbidden; when modeled with
hþ
βKWW
hθhþi ¼
Γ
ð9Þ
this assumption, the hopping rate constant from this work was
βKWW
ꢀ1
(
110 ns)
.
as suggested by excellent agreement between the experimen-
What follows is a discussion of the anisotropy’s dependence
on the sensitizer employed, its surface coverage, the solution
temperature, and the excitation laser fluence. Interestingly, an
tal data and the KohlrauschꢀWilliamsꢀWatts kinetic model,
another 4 orders of magnitude discrepancy is realized, Æθ æ =
h+
1
10 ms. Therefore, the extremely disparate rates of anisotro-
initial fundamental anisotropy, r
o
> 0.3, was not attained for
(dcb)] * anchored to TiO or ZrO . As the ther-
mally equilibrated excited (thexi) state is proposed to be well-
2
+
py decay for energy and hole transfer support an underlying
mechanism that does not involve physical motion of the
compounds.
[Ru(dtb)
2
2
2
ꢀ
2+
formulated as [Ru(dtb)
dcb transition should yield the theoretical maximum anisotro-
py value for randomly distributed chromophores, i.e., r = 0.4,
(dcb )] *, excitation into the Ru f
2
It was also shown that there was little-to-no measurable
temperature dependence to the anisotropy decay associated with
energy transfer over a ∼77ꢀ283 K range. If the observed
anisotropy was indeed physical rotation of the molecules, an
inverse temperature dependence to the correlation time would
have been expected based on the StokesꢀEinsteinꢀSmoluchowski
relation. Energy-transfer reactions do not follow Brownian motion
and are instead governed by Dexter electron-exchange interac-
tion, F €o rster dipoleꢀdipole interaction, quantum coherence, trivial
emission followed by reabsorption, and/or hybrid or alternative
mechanisms. Each has its own temperature dependence to the
energy-transfer rate and many display small temperature effects,
like that observed here.
o
while excitation into either of the doubly degenerate opposing
dtb ligands should yield a negative initial fundamental PL
45,46
anisotropy.
Notwithstanding, anisotropies less than the
theoretical maximum of 0.4 are typical for Ru(II)ꢀpolypyridyl
compounds, where multiple and overlapping charge-transfer
4
5ꢀ47
transitions are present.
PL excitation anisotropy spectrum for a sample containing
[Ru(bpy) (deebq)]*/ZrO was investigated. This compound
To test this notion, the steady-state
2
2
was chosen as it has a single well-resolved Ru f deebq absorp-
tion band in the red portion of the visible spectrum, which is
significantly separated from the blue Ru f bpy transitions, and a
short excited-state lifetime, i.e., τ ≈ 245 ns, which attenuates
significant energy transfer, and thus anisotropy loss, during
excited-state decay. Steady-state PL anisotropy values as high as
0.35 and as low as ꢀ0.03 resulted, which agreed reasonably well
withthefundamental anisotropies obtainedfrom time-resolved PL
measurements. Therefore, the low initial fundamental anisotropies
likely resulted from overlapping charge-transfer electronic transi-
tions. Thin films of ZrO , instead of TiO , were chosen initially so
Energy-Transfer Reactions. Energy transfer assemblies
have been reported previously in dye-sensitized solar cells. An
early strategy was to covalently link energy-transfer donors to a
central unit that accepted the charge and subsequently trans-
30ꢀ36
ferred an electron to TiO2.
If the energy-transfer donors did
not increase the molecular footprint of the central unit on the
semiconductor surface, the light harvesting efficiency could be
II
enhanced. Indeed, the trinuclear Ru ꢀcompound utilized in the
2
2
5
celebrated 1991 Nature paper was designed to function as such
as to exclude side effects due to other phenomena that often occur
33
an antenna. An issue with the cis-Ru(dcb) (CN) acceptor group
under conditions of favorable excited-state electron injection into
2
2
18,48ꢀ50
used was the cis geometry of the ambidentate cyano ligands,
which resulted in a footprint that increased with the number of
TiO2.
The surface-coverage dependence for energy transfer was small.
As the energy-transfer mechanism involves interactions between
II
Ru chromophores. In this regard, a trans geometry was more
3
7
51ꢀ54
preferred. A related approach that did not require synthesis of
elaborate molecules involved introduction of an energy-transfer
donor in the electrolyte that harvested and transferred energy to
a surface-anchored compound. This was recently realized by
McGehee, Gr €a tzel, and colleagues using an efficient organic
a singlet ground state and a predominantly triplet excited state,
a Dexter energy-transfer mechanism previously has been pro-
posed for MLCT excited states. This implies an exponential
dependence of the energy-transfer, nearest-neighbor hopping
rate constant. Because the experimentally observable parameters
were the self-exchange correlation times and the sensitizer sur-
face coverage or number of sensitizers per nanoparticle, deduced
from absorption measurements, Monte Carlo simulations were
performed to determine how these parameters were related. It
was confirmed that, under the conditions given in the Experi-
mental Section, the nearest-neighbor distance was proportional
to the square root of the inverse of the number of sensitizers
per particle (Figure S6a, Supporting Information). It was also
determined that the logarithm of the nearest-neighbor energy
transfer rate constants and logarithm of the self-exchange
correlation times were linearly related to one another (Figure
S6b, Supporting Information). The experimental data did not
show a clear exponential dependence for energy transfer, which
implied that intermolecular energy transfer may not be the only
mechanism resulting in a loss of anisotropy (Figure S3, Support-
ing Information).
energy-transfer donor in the electrolyte surrounding a phthalo-
6
cyanine-sensitized TiO thin film. Since then, other research
2
groups have investigated the same phenomenon and have
7,38ꢀ41
realized similar successes.
Most relevant to the studies described herein, evidence for
lateral energy transfer between rutheniumꢀpolypyridyl com-
pounds anchored to the same surface has been reported. Kinetic
data at high laser fluences and high surface coverages were found
to obey a second-order, equal-concentration mechanism at early
times attributed to tripletꢀtriplet annihilation reactions that
42ꢀ44
resulted from lateral self-exchange energy transfer.
Studies
of films cosensitized with Ru(II) and Os(II) chromophores later
25
provided more direct evidence for the lateral energy transfer.
Monte Carlo simulations of the fluence-dependent data revealed
an energy-transfer, nearest-neighbor hopping rate constant of
ꢀ
1
44
(
30 ns) for Ru(bpy) (dcb)/TiO . This value is in reasonable
2 2
1
ꢀ
agreement with (120 ns) abstracted from the maximum sur-
face-coverage anisotropy data herein, where the minor discre-
pancies could be due to the bulky tert-butyl groups, which may slow
A small temperature dependence to the energy-transfer self-
exchange was measured at low laser fluences. The quantum yield
for randomization, Φ, of an initial anisotropic population was
1
5392
dx.doi.org/10.1021/ja200652r |J. Am. Chem. Soc. 2011, 133, 15384–15396

Journal of the American Chemical Society
ARTICLE
calculated per the following equation
last milliseconds, thereby providing a large temporal range for
characterization of hole-hopping reactions. The three sensiti-
12
τ
Φ ¼
ð10Þ
zers cis-Ru(dcb) (NCS) (N3), cis-Ru(dnb)(dcb)(NCS) (Z907),
2
2
2
τ þ θent
2+
and Ru(dtb) (dcb) were compared for their ability to perform
2
where τ is the excited-state lifetime and θ is the energy-transfer
self-exchange hole-transfer reactions in acetonitrile in the ab-
ent
2
8
correlation time. The Φ parameters are given in Table 1 and
assist in quantifying the effectiveness by which a molecule can
deliver its free energy to sparsely located “reaction centers”. In
addition, the temperature required to achieve a certain efficiency
can be calculated. For example, for the lifetime measured at 283
K, a Φ < 0.20 quantum yield was obtained. As θent was nearly
temperature independent over a 200 K range while τ was not,
extrapolation of the Arrhenius relation predicted that at ∼130 K
the excited-state lifetime and energy-transfer correlation time
sence and presence of 100 mM LiClO . Under most conditions,
4
hole transfer was observed on a micro- to millisecond time scale.
+
Interestingly, little-to-no hole transfer was evident for N3 /
ꢀ
TiO
(e ) over the course of its charge-separated-state lifetime.
2
On the other hand, another bis-isothiocyanate-based sensitizer,
+
ꢀ
Z907 /TiO (e ), was found to hole transfer regardless of the
2
presence of a supporting electrolyte. The significant transient-
absorption anisotropy kinetics indicative of lateral self-exchange
+
ꢀ
+
hole transfer observed for Z907 /TiO
(e ) but not N3 /
2
ꢀ
should be equal, τ = θ . Thus, Φ = 0.5, implying that half of the
TiO
2
(e ) is consistent with previous electrochemical measure-
ent
sensitizers in the initially photoselected subpopulation, excited
by low fluence excitation, would circumnavigate the nanosphere
before relaxing back to the ground state.
ments made on much longer time scales and was attributed to
13
the molecularꢀsurface orientation. For N3/TiO
boxylic acid groups on opposite dcb ligands, which are trans to
the isothiocyanato ligands, were proposed to anchor to TiO
whereas for Z907/TiO both carboxylate groups of the single dcb
, two car-
2
The initial anisotropy decay rate increased with laser fluence.
A likely explanation for this observation is based on the increased
yield of tripletꢀtriplet annihilation reactions previously reported
,
2
2
13
III
ligand bind to TiO . As partial hole transfer from Ru to the
2
44
under such conditions. Since the excitation laser light was vertically
polarized, the majority of the excited states were generated at the
north and south poles along the vertical axes of the nanocrys-
tallites. The close proximity of these excited states suggests that at
early times they were removed most rapidly, via annihilation re-
actions. The excited states that were located near the equator of
the nanoparticles were much less likely to have been excited. As
such, the resulting low local concentration of chromophores
meant little-to-no rapid annihilation reactions near the equator.
Since radiative and nonradiative decay are much slower pro-
cesses, these excited states would decay little over this initial time
period. The disparate rates for excited-state deactivation near the
poles vs the equator of the nanoparticles were therefore expected
to result in a large and rapid initial drop in anisotropy, consistent
with the experimental data.
isothiocyanto ligands is known, the more radial orientation of the
ꢀ
NCS ligands relative to the surface in the former compound
12
attenuates rapid intermolecular hole transfer.
Although hole transfer was not observed for Ru (dtb)
III
(dcb)/
2
ꢀ
TiO
(e ) in neat acetonitrile, it did occur after introduction of
2
just 10 mM LiClO
tions for Ru (dtb)
. The lack of significant hole-transfer reac-
4
III
ꢀ
(dcb)/TiO
(e ) in neat acetonitrile sug-
2
2
gests that ions were required to decrease the work terms associated
1
with self-exchange hole transfer in this more highly charged tris-
diimine compound. Heteroleptic ruthenium compounds, like
II
2+
[Ru (dtb)
Ru
(dcb)] , anchored to TiO are known to have effective
2
2 13
2
III/II
ꢀ10
diffusion coefficients of ∼3 ꢂ 10
cm /s, roughly an
+
ꢀ
order-of-magnitude smaller than that of Z907 /TiO (e ), due
2
1
to electron tunneling through the bipyridine ligands. A difficulty
in performing a more thorough comparison of the self-exchange
hole-transfer rates for these sensitizers is that the functions re-
quired to fit their anisotropy kinetics yielded very different dis-
tributions of self-exchange hole-transfer rate constants. However,
the fact that the distributions and the average hole-transfer cor-
Hole-Transfer Reactions. The change in anisotropy mea-
sured after excited-state injection was reasonably attributed to
lateral self-exchange hole transfer across the nanocrystalline
TiO surface. It is well-known that Ru-polypyridyl sensitizers
2
+
ꢀ
anchored to these thin films can be fully oxidized in a standard
electrochemical cell provided that the surface coverage exceeds a
percolation threshold of about 50% of the saturation value.
relations times were larger for Z907 /TiO (e ) indicates that the
2
rates for hole transfer were on average faster for these compounds.
A discrete quantum yield for randomization of the initial
anisotropy by hole transfer self-exchange was quantified with
eq 10, as was done for energy transfer, where τ was the charge-
separated-state lifetime with hole-transfer correlation time, θh+.
The KWW function used to model the data is based on a L ꢀe vy
distribution of rate constants, where β is inversely related to the
width of the distribution of rate constants. Since the β values
abstracted from transient data that corresponded to charge re-
combination and hole hopping were, within experimental error,
the same for a given compound, a comparison of any single value
within the distribution was of interest. For the data in Table 1, the
9
ꢀ11,13
Potential-step experiments have allowed effective diffusion coef-
ficients for self-exchange hole transfer to be quantified on the
minutes time scale. It was of specific interest to see whether this
available electrochemical data was correlated in any way to that
from these anisotropy measurements. Since the entire mesopor-
ous film was oxidized in the electrochemical studies, the data
were expected to be influenced by interparticle hole transfer across
necking regions between anatase nanocrystallites. Significant hole
transfer between molecules on different TiO nanocrystallites
2
was less likely on the short time scale probed spectroscopically, so the
combined techniques represent a powerful approach for character-
ization of lateral charge transfer on semiconductor materials.
One advantage of anisotropy over electrochemical measure-
ments is that self-exchange hole transfer can be quantified in
many environments without the need for high ionic strength
electrolytes. Furthermore, hole hopping can be quantified on
short time scales and is limited only by excited-state injection, a
process that is known to occur on ultrafast time scales under
III
efficiency of hole hopping for Ru (dtb) (dcb)/TiO (e-) was
2
2
+
ꢀ
only 0.008 but increased to 0.6 for Z907 /TiO (e ). Hole
2
transfer self-exchange, while much slower than energy transfer
self-exchange, was relatively efficient, as it competed kinetically
with the much slower charge recombination process. This data
shows that lateral hole transfer can dramatically randomize the
initial anisotropy with efficiencies that are markedly dependent
on the rutheniumꢀpolypyridyl compound and the environ-
mental conditions.
5
5
many conditions. The interfacial charge-separated states often
1
5393
dx.doi.org/10.1021/ja200652r |J. Am. Chem. Soc. 2011, 133, 15384–15396

Journal of the American Chemical Society
ARTICLE
Scheme 4. Proposed Mechanisms for Multiple-Charge-Transfer Oxidation at a Dye-Sensitized TiO Interface: (A) (1) Lateral
2
Energy Transfer, (2) Excited-State Injection at a Charge-Separation Unit (red box), and (3) Catalysis (Cat) or (B) (1) Excited-
State Injection, (2) Lateral Hole Transfer, and (3) Catalysis
Efficient lateral hole-transfer self-exchange reactions after
indicate that the artificial assembly proposed in Scheme 4 could soon
excited-state injection could be exploited for delivery of multiple
be realized.
56
redox equivalents to a catalysts capable of water oxidation. In a
small step toward this goal, and to test whether lateral hole
transfer to any catalyst would occur prior to charge recombina-
tion, a molecular cobalt compound, CoTCPP, was coanchored to
’
CONCLUSIONS
Photoinduced lateral energy- and hole-transfer self-exchange
a Z907/TiO -sensitized thin film in a 1:100 ratio. The strategy
across semiconductor surfaces was observed for the first time by
time-resolved spectroscopic anisotropy measurements. The data
has clear implications for fundamental studies and for applications in
solar energy conversion. First, at a minimum, both photolumi-
nescence and transient-absorption studies of dye-sensitized films
that utilize polarized laser sources need employ a depolarizer,
without which nonuniform illumination will result. Second, me-
2
was that after excited-state injection into TiO , the oxidized
2
rutheniumꢀpolypyridyl compounds would undergo lateral hole
transfer to the catalyst. This was indeed observed to occur on a
hundreds of microseconds time scale with a yield of 0.03. The
slow time scale for the reaction and low surface coverage of the
II
metalloporphyrin support the hypothesis that Co TCPP/TiO
2
III/II
oxidation was preceded by lateral Ru
self-exchange hole
chanistic models of interfacial charge recombination between
II
ꢀ
transfer. The poor Co oxidation yields can be rationalized by
the large number of hops, kinetic competition with charge
recombination, and/or nonuniform loading of CoTCPP. Never-
theless, the experiments were successful and supported our
TiO
2
(e )s and oxidized sensitizers must take into account diffu-
8
2
13
sion of the injected electron8 and the oxidized sensitizer, not
3,84
simply the injected electron.
Third, in the absence of excited-
state injection, the mesoscopic TiO thin films developed for
2
III/II
regenerative solar cells can function as active “antennae”, harvest-
ing sunlight and transferring energy across the surface of nano-
crystalline materials. Fourth, lateral hole transfer can be utilized
to translate holes present as oxidized sensitizers to catalysts.
Such hole transfer is not necessary for iodide oxidation in dye-
sensitized solar cells; however, it is required for water oxidation,
where four oxidizing equivalents must be accumulated on a single
catalytic site. The phenomena presented herein represent two
means of concentrating potential energy derived from visible
light. If one were able to funnel such energy to molecular catalysts,
multiple-charge-transfer reactions that generate solar fuels could
be realized. Finally, we anticipate that time-resolved photolumi-
nescence and transient-absorption anisotropy will be useful tools for
the study of a wide variety of lateral self-exchange reactions across
many classes of semiconductor materials.
previous results that catalyst oxidation preceded by lateral Ru
57
self-exchange hole transfer did occur (Scheme 4b). To the best
of our knowledge, this represents the first direct observation of
lateral hole transfer across a semiconductor surface to oxidize a
85
5
7,58
molecular catalyst.
55
While this report focused on characterization of fundamental
energy- and hole-transfer self-exchange reactions, it should im-
minently be possible to utilize them for the photocatalytic oxidation
of water (Scheme 4). Energy transfer to a catalyst capable of elec-
56
tron injection into TiO would leave an oxidizing equivalent (hole)
2
on the catalyst. Alternatively, excited-state injection followed by
hole transfer could be utilized to deliver an oxidizing equivalent
to a catalyst. Accumulation of four holes at a single catalyst is a
prerequisite for direct water splitting. In principle, the multiple
energy- and charge-transfer reactions required for water splitting
could occur after only a few photons were absorbed by chromo-
phores on a single nanoparticle. Photoelectrosynthetic cells that
operate by lateral hole transfer to catalysts that accumulate
charge and mediate OꢀO bond formation have indeed been
’
ASSOCIATED CONTENT
S
Supporting Information. Anisotropy data, Arrhenius anal-
b
5
9
reported. The discovery of a wide variety of water oxidation
ysis, and Monte Carlo simulations, and the rationale for the method
chosen to correct for the polarization-dependent response of the
6
0ꢀ64 56,59,65ꢀ70
71ꢀ75
76ꢀ80
81
catalysts based on Ru,
Ir,
Co,
Mn,
and Ni
1
5394
dx.doi.org/10.1021/ja200652r |J. Am. Chem. Soc. 2011, 133, 15384–15396

Journal of the American Chemical Society
ARTICLE
detection system are given. This material is available free of charge
via the Internet at http://pubs.acs.org.
(24) Meyer, T. J.; Meyer, G. J.; Pfennig, B. W.; Schoonover, J. R.;
Timpson, C. J.; Wall, J. F.; Kobusch, C.; Chen, X.; Peek, B. M.; Wall,
C. G.; Ou, W.; Erickson, B. W.; Bignozzi, C. A. Inorg. Chem. 1994,
3
3, 3952–3964.
25) Trammell, S. A.; Yang, J.; Sykora, M.; Fleming, C. N.; Odobel,
F.; Meyer, T. J. J. Phys. Chem. B 2001, 105, 8895–8904.
26) Gray, H. B.; Winkler, J. R. Proc. Natl. Acad. Sci. U. S. A. 2005,
02, 3534–3539.
27) Ericson, T.; Zinoviev, V. Codes on Euclidean Spheres; Elsevier:
Asterdam, 2001; Vol. 63.
28) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 3 ed.;
Springer Science+Business Media: New York, 2006.
’
AUTHOR INFORMATION
(
Corresponding Author
meyer@jhu.edu
(
1
(
’
ACKNOWLEDGMENT
(
We acknowledge support by a grant from the Division of
(29) Lindsey, C. P.;Patterson, G. D.J. Chem. Phys. 1980, 73, 3348–3357.
30) Balzani, V.; Moggi, L.; Scandola, F. In Supramolecular Photo-
Chemical Sciences, Office of Basic Energy Sciences, Office of
Energy Research, U.S. Department of Energy (DE-FG02-
(
chemistry; Balzani, V., Ed.; D. Reidel Publishing Co.: Dordrecht, Holland,
987; p 1ꢀ28.
31) Bignozzi, C. A.; Argazzi, R.; Indelli, M. T.; Scandola, F. Sol.
Energy Mater. 1994, 32, 229–244.
32) Bignozzi, C. A.; Argazzi, R.; Scandola, F.; Schoonover, J. R.;
Meyer, G. J. Sol. Energy Mater. 1995, 38, 187–198.
33) Amadelli, R.; Argazzi, R.; Bignozzi, C. A.; Scandola, F. J. Am.
Chem. Soc. 1990, 112, 7099–7103.
34) Nazeeruddin, M. K.; Liska, P.; Moser, J.; Vlachopoulos, N.;
9
6ER14662). S.A. acknowledges a Johns Hopkins University
1
Greer graduate student fellowship. We thank Prof. Ludwig Brand
and Dr. Dmitri Toptygin for stimulating discussions on polariza-
tion spectroscopy and assistance using their steady-state emis-
sion detection system.
(
(
(
’
REFERENCES
(
(1) Sutin, N. Acc. Chem. Res. 1982, 15, 275–282.
Gr €a tzel, M. Helv. Chim. Acta 1990, 73, 1788–1803.
(
(
(
2) Sutin, N. Prog. Inorg. Chem. 1983, 30, 441–498.
3) Fleming, G. R.; Scholes, G. D. Nature 2004, 431, 256–257.
4) Beljonne, D.; Curutchet, C.; Scholes, G. D.; Silbey, R. J. J. Phys.
(35) Holten, D.; Bocian, D. F.; Lindsey, J. S. Acc. Chem. Res. 2002,
35, 57–69.
(36) Forster, R. J.; Keyes, T. E.; Vos, J. G. Interfacial Supramolecular
Assemblies; John Wiley & Sons Ltd.: Chichester, 2003.
(37) Gajardo, F.; Leiva, A. M.; Loeb, B.; Delgadillo, A.; Stromberg,
J. R.; Meyer, G. J. Inorg. Chim. Acta 2008, 361, 613–619.
(38) Shankar, K.; Feng, X.; Grimes, C. A. ACS Nano 2009, 3, 788–794.
(39) Basham, J. I.; Mor, G. K.; Grimes, C. A. ACS Nano 2010,
4, 1253–1258.
Chem. B 2009, 113, 6583–6599.
(
(
P.; Torres, T.; Frechet, J. M. J.; Nazeeruddin, M. K.; Gr €a tzel, M.;
McGehee, M. D. Nat. Photon 2009, 3, 406–411.
(
5) O’Regan, B.; Gr €a tzel, M. Nature 1991, 353, 737–740.
6) Hardin, B. E.; Hoke, E. T.; Armstrong, P. B.; Yum, J.-H.; Comte,
7) Hardin, B. E.; Yum, J.-H.; Hoke, E. T.; Jun, Y. C.; Pechy, P.;
Torres, T.; Brongersma, M. L.; Nazeeruddin, M. K.; Gr €a tzel, M.;
McGehee, M. D. Nano Lett. 2010, 10, 3077–3083.
(40) Buhbut, S.; Itzhakov, S.; Tauber, E.; Shalom, M.; Hod, I.; Geiger,
T.; Garini, Y.; Oron, D.; Zaban, A. ACS Nano 2010, 4, 1293–1298.
(41) Mor, G. K.; Basham, J.; Paulose, M.; Kim, S.; Varghese, O. K.;
Vaish, A.; Yoriya, S.; Grimes, C. A. Nano Lett. 2010, 10, 2387–2394.
(42) Kelly, C. A.; Farzad, F.; Thompson, D. W.; Meyer, G. J.
Langmuir 1999, 15, 731–737.
(43) Kelly, C. A.; Meyer, G. J. Coord. Chem. Rev. 2001, 211, 295–315.
(44) Higgins, G. T.; Bergeron, B. V.; Hasselmann, G. M.; Farzad, F.;
Meyer, G. J. J. Phys. Chem. B 2006, 110, 2598–2605.
(45) Wallin, S.; Davidsson, J.; Modin, J.; Hammarstr €o m, L. J. Phys.
Chem. A 2005, 109, 4697–4704.
(46) Wallin, S.; Davidsson, J.; Modin, J.; Hammarstr €o m, L. J. Phys.
Chem. A 2005, 109, 9378–9378.
(47) Terpetschnig, E.; Szmacinski, H.; Malak, H.; Lakowicz, J. R.
Biophys. J. 1995, 68, 342–350.
(48) Kelly, C. A.; Farzad, F.; Thompson, D. W.; Stipkala, J. M.;
Meyer, G. J. Langmuir 1999, 15, 7047–7054.
(49) Ardo, S.; Sun, Y.; Castellano, F. N.; Meyer, G. J. J. Phys. Chem. B
2010, 114, 14596–14604.
(
8) Yum, J.-H.; Baranoff, E.; Hardin, B. E.; Hoke, E. T.; McGehee,
M. D.; Nuesch, F.; Gr €a tzel, M.; Nazeeruddin, M. K. Energy Environ. Sci.
010, 3, 434–437.
9) Heimer, T. A.; D’Arcangelis, S. T.; Farzad, F.; Stipkala, J. M.;
Meyer, G. J. Inorg. Chem. 1996, 35, 5319–5324.
10) Bonhote, P.; Gogniat, E.; Tingry, S.; Barbe, C.; Vlachopoulos,
N.; Lenzmann, F.; Comte, P.; Gr €a tzel, M. J. Phys. Chem. B 1998, 102,
498–1507.
2
(
(
1
(11) Trammell, S. A.; Meyer, T. J. J. Phys. Chem. B 1999, 103, 104–107.
(12) Wang, Q.; Zakeeruddin, S. M.; Cremer, J.; Bauerle, P.; Humphry-
Baker, R.; Gr €a tzel, M. J. Am. Chem. Soc. 2005, 127, 5706–5713.
13) Wang, Q.; Zakeeruddin, S. M.; Nakeeruddin, M. K.; Humphry-
Baker, R.; Gr €a tzel, M. J. Am. Chem. Soc. 2006, 128, 4446–4452.
14) Kliger, D. S.; Lewis, J. W.; Randall, C. E. Polarized Light in Optics
and Spectroscopy; Academic Press, Inc.: Boston, 1990.
15) Michl, J.; Thulstrup, E. W. Spectroscopy with Polarized Light;
VCH Publishers, Inc.: Weinheim, Germany, 1986.
16) Staniszewski, A.; Ardo, S.; Sun, Y.; Castellano, F. N.; Meyer,
G. J. J. Am. Chem. Soc. 2008, 130, 11586–11587.
17) Hoertz, P. G.; Thompson, D. W.; Friedman, L. A.; Meyer, G. J.
J. Am. Chem. Soc. 2002, 124, 9690–9691.
18) Ardo, S.; Sun, Y.; Staniszewski, A.; Castellano, F. N.; Meyer,
G. J. J. Am. Chem. Soc. 2010, 132, 6696–6709.
19) Argazzi, R.; Bignozzi, C. A.; Heimer, T. A.; Castellano, F. N.;
Meyer, G. J. Inorg. Chem. 1994, 33, 5741–5749.
20) Holzer, W.; Le Duff, Y.; Altmann, K. J. Chem. Phys. 1973,
8, 642–643.
(
(
(
(
(50) Cappel, U. B.; Feldt, S. M.; Schoneboom, J.; Hagfeldt, A.;
Boschloo, G. J. Am. Chem. Soc. 2010, 132, 9096–9101.
(51) Crosby, G. A.; Hipps, K. W.; Elfring, W. H. J. Am. Chem. Soc.
1974, 96, 629–630.
(
(
(52) Kober, E. M.; Meyer, T. J. Inorg. Chem. 1982, 21, 3967–3977.
(53) Meyer, T. J. Pure Appl. Chem. 1986, 58, 1193–1206.
(54) Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.;
von Zelewsky, A. Coord. Chem. Rev. 1988, 84, 85–277.
(55) Ardo, S.; Meyer, G. J. Chem. Soc. Rev. 2009, 38, 115–164.
(56) Youngblood, W. J.; Lee, S.-H. A.; Kobayashi, Y.; Hernandez-
Pagan, E. A.; Hoertz, P. G.; Moore, T. A.; Moore, A. L.; Gust, D.;
Mallouk, T. E. J. Am. Chem. Soc. 2009, 131, 926–927.
(57) Ardo, S.; Meyer, G. J. J. Am. Chem. Soc. 2010, 132, 9283–9285.
(58) Kellett, R. M.; Spiro, T. G. Inorg. Chem. 1985, 24, 2378–2382.
(59) Youngblood, W. J.; Lee, S.-H. A.; Maeda, K.; Mallouk, T. E. Acc.
Chem. Res. 2009, 42, 1966–1973.
(
(
5
(21) Le Duff, Y. J. Phys. B: At. Mol. Phys. 1981, 14, 55.
(22) Bischel, W. K.; Bamford, D. J.; Jusinski, L. E. Appl. Opt. 1986,
2
5, 1215–1221.
23) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Funda-
mentals and Applications, 2nd ed.; John Wiley & Sons, Inc.: New York,
001.
(
2
1
5395
dx.doi.org/10.1021/ja200652r |J. Am. Chem. Soc. 2011, 133, 15384–15396

Journal of the American Chemical Society
ARTICLE
(
60) Geletii, Y. V.; Botar, B.; K €o gerler, P.; Hillesheim, D. A.; Musaev,
D. G.; Hill, C. L. Angew. Chem., Int. Ed. 2008, 47, 3896–3899.
61) Concepcion, J. J.; Jurss, J. W.; Brennaman, M. K.; Hoertz, P. G.;
(
Patrocinio, A. O. T.; Murakami Iha, N. Y.; Templeton, J. L.; Meyer, T. J.
Acc. Chem. Res. 2009, 42, 1954–1965.
(
62) Romain, S.; Bozoglian, F.; Sala, X.; Llobet, A. J. Am. Chem. Soc.
009, 131, 2768–2769.
63) Romain, S.; Vigara, L.; Llobet, A. Acc. Chem. Res. 2009, 42,
944–1953.
64) Xu, Y.; Fischer, A.; Duan, L.; Tong, L.; Gabrielsson, E.; Åkermark,
B.; Sun, L. Angew. Chem., Int. Ed. 2010, 49, 8934–8937.
65) McDaniel, N. D.; Coughlin, F. J.; Tinker, L. L.; Bernhard, S.
J. Am. Chem. Soc. 2008, 130, 210–217.
66) Hull, J. F.; Balcells, D.; Blakemore, J. D.; Incarvito, C. D.;
Eisenstein, O.; Brudvig, G. W.; Crabtree, R. H. J. Am. Chem. Soc. 2009,
2
(
1
(
(
(
1
2
2
31, 8730–8731.
67) Nakagawa, T.; Beasley, C. A.; Murray, R. W. J. Phys. Chem. C
009, 113, 12958–12961.
68) Nakagawa, T.; Bjorge, N. S.; Murray, R. W. J. Am. Chem. Soc.
009, 131, 15578–15579.
69) Blakemore, J. D.; Schley, N. D.; Balcells, D.; Hull, J. F.; Olack,
(
(
(
G. W.; Incarvito, C. D.; Eisenstein, O.; Brudvig, G. W.; Crabtree, R. H.
J. Am. Chem. Soc. 2010, 132, 16017–16029.
(70) Tilley, S. D.; Maurin, C.; Sivula, K.; Gr €a tzel, M. Angew. Chem.,
Int. Ed. 2010, 49, 6405–6408.
(71) Kanan, M. W.; Nocera, D. G. Science 2008, 321, 1072–1075.
(72) Jiao, F.; Frei, H. Angew. Chem., Int. Ed. 2009, 48, 1841–1844.
(73) Kanan, M. W.; Surendranath, Y.; Nocera, D. G. Chem. Soc. Rev.
2
009, 38, 109–114.
(
74) Nocera, D. G. Inorg. Chem. 2009, 48, 10001–10017.
(75) Yin, Q.; Tan, J. M.; Besson, C.; Geletii, Y. V.; Musaev, D. G.;
Kuznetsov, A. E.; Luo, Z.; Hardcastle, K. I.; Hill, C. L. Science 2010,
28, 342–345.
76) Brimblecombe, R.; Swiegers, G. F.; Dismukes, G. C.; Spiccia, L.
Angew. Chem., Int. Ed. 2008, 47, 7335–7338.
77) Dismukes, G. C.; Brimblecombe, R.; Felton, G. A. N.; Pryadun,
R. S.; Sheats, J. E.; Spiccia, L.; Swiegers, G. F. Acc. Chem. Res. 2009,
2, 1935–1943.
78) Li, G.; Sproviero, E. M.; Iii, R. C. S.; Iguchi, N.; Blakemore, J. D.;
Crabtree, R. H.; Brudvig, G. W.; Batista, V. S. Energy Environ. Sci. 2009,
, 230–238.
79) Magnuson, A.; Anderlund, M.; Johansson, O.; Lindblad, P.;
3
(
(
4
(
2
(
Lomoth, R.; Polivka, T.; Ott, S.; Stensj €o , K.; Styring, S.; Sundstr €o m, V.;
Hammarstr €o m, L. Acc. Chem. Res. 2009, 42, 1899–1909.
(
80) Brimblecombe, R.; Koo, A.; Dismukes, G. C.; Swiegers, G. F.;
Spiccia, L. J. Am. Chem. Soc. 2010, 132, 2892–2894.
81) Dinc ꢁa , M.; Surendranath, Y.; Nocera, D. G. Proc. Natl. Acad. Sci.
U. S. A. 2010, 107, 10337–10341.
82) Kopidakis, N.; Schiff, E. A.; Park, N. G.; van de Lagemaat, J.;
Frank, A. J. J. Phys. Chem. B 2000, 104, 3930–3936.
83) Nelson, J.; Chandler, R. E. Coord. Chem. Rev. 2004, 248,
181–1194.
84) Nelson, J.; Haque, S. A.; Klug, D. R.; Durrant, J. R. Phys. Rev. B
001, 63, 205321.
85) Bonhote, P.; Moser, J. E.; Humphry-Baker, R.; Vlachopoulos,
N.; Zakeeruddin, S. M.; Walder, L.; Gr €a tzel, M. J. Am. Chem. Soc. 1999,
21, 1324–1336.
(
(
(
1
(
2
(
1
1
5396
dx.doi.org/10.1021/ja200652r |J. Am. Chem. Soc. 2011, 133, 15384–15396