Distance dependent electron transfer at TiO2 interfaces sensitized with phenylene ethynylene bridged RuII-isothiocyanate compounds

Article abstract of DOI:10.1021/ja402193f

Excess electrons present in semiconductor nanocrystallites generate a significant electric field, yet the role this field plays in molecular charge transfer processes remains poorly understood. Three ruthenium bipyridyl cis-Ru(bpy)(LL)(NCS)₂ compounds, where LL is a 4-substituted bpy, with zero, one, or two phenylene ethynylene bridge units, were anchored to mesoporous nanocrystalline TiO₂ thin films to specifically quantify interfacial charge transfer with chromophores designed to be set at variable distances from the surface. Injection of electrons into TiO₂ resulted in a blue shift of the metal-to-ligand charge transfer absorption consistent with an underlying Stark effect. The electroabsorption data were used to quantify the electric field experienced by the compounds that decreased from 0.85 to 0.22 MV/cm as the number of OPE spacers increased from 0 to 2. Charge recombination on the 10^-8-10^-5 s time scale correlated with the magnitude of the electric field with an apparent attenuation factor β = 0.12 A^-1. Slow components to charge recombination observed on the 10^-4-10^-1 s time scale that were unaffected by temperature, irradiance, or the bridge units present on the molecular sensitizer were attributed to electron tunneling between TiO ₂ acceptor states. The photocurrent efficiencies of solar cells based on these compounds decreased markedly when the bridge units were present on the sensitizer. Iodine was found to form adducts with all three compounds, K = 1.8 ± 0.2 × 10⁴ M^-1, but only significantly lowered the excited state injection yield for those that possessed the bridge units.

Full text of DOI:10.1021/ja402193f

Article
pubs.acs.org/JACS
Distance Dependent Electron Transfer at TiO2 Interfaces Sensitized
with Phenylene Ethynylene Bridged RuII−Isothiocyanate Compounds
Patrik G. Johansson,† Andrew Kopecky,‡ Elena Galoppini,*,‡ and Gerald J. Meyer*,†
†Departments of Chemistry and Materials Science & Engineering, Johns Hopkins University, 3400 North Charles Street, Baltimore,
Maryland 21218, United States
‡Department of Chemistry, Rutgers University, 73 Warren St, Newark, New Jersey 07102, United States
S
* Supporting Information
ABSTRACT: Excess electrons present in semiconductor
nanocrystallites generate a significant electric field, yet the
role this field plays in molecular charge transfer processes
remains poorly understood. Three ruthenium bipyridyl cis-
Ru(bpy)(LL)(NCS)2 compounds, where LL is a 4-substituted
bpy, with zero, one, or two phenylene ethynylene bridge units,
were anchored to mesoporous nanocrystalline TiO2 thin films
to specifically quantify interfacial charge transfer with
chromophores designed to be set at variable distances from
the surface. Injection of electrons into TiO2 resulted in a blue
shift of the metal-to-ligand charge transfer absorption
consistent with an underlying Stark effect. The electro-
absorption data were used to quantify the electric field experienced by the compounds that decreased from 0.85 to 0.22
MV/cm as the number of OPE spacers increased from 0 to 2. Charge recombination on the 10−8−10−5 s time scale correlated
with the magnitude of the electric field with an apparent attenuation factor β = 0.12 Å−1. Slow components to charge
recombination observed on the 10−4−10−1 s time scale that were unaffected by temperature, irradiance, or the bridge units
present on the molecular sensitizer were attributed to electron tunneling between TiO2 acceptor states. The photocurrent
efficiencies of solar cells based on these compounds decreased markedly when the bridge units were present on the sensitizer.
Iodine was found to form adducts with all three compounds, K = 1.8 0.2 × 104 M−1, but only significantly lowered the excited
state injection yield for those that possessed the bridge units.
hole transfer from the oxidized sensitizer to a redox mediator,
behavior that is particularly important for sensitization to
infrared light, where S+ is a weak oxidant and the driving force
for hole transfer is small.14 Here we describe studies where
charge recombination was inhibited by the presence of
phenylene ethynylene, or what are sometimes called oligo-
(phenylene ethynylene) (OPE), bridge units.
The introduction of a rigid bridge that controls the distance
between the donor and acceptor is one approach by which
electron transfer has been controlled in molecular com-
pounds.15−17 Quite often the influence of the bridge on the
electron transfer rate constants is quantified through eq 1,
INTRODUCTION
■
The fundamental behavior of metal-to-ligand charge-transfer
(MLCT) excited states at conductive interfaces is relevant to
applications in dye-sensitized solar cells (DSSCs) and organic
light emitting diodes (OLEDs).1−3 In DSSCs, ruthenium
polypyridyl compounds, generically referred to as “sensitizers”,
anchored to nanocrystalline (anatase) TiO2 have emerged as
the most successful.4,5 Under most conditions, the MLCT
excited states are short-lived due to rapid electron transfer to
the TiO2 acceptor states, S*/TiO2 → S+/TiO2(e−).6−10 Indeed,
electron injection on sub-picosecond time scales has been
reported for many sensitized TiO2 interfaces.9,11 Under
conditions where the energetics for excited state electron
transfer are unfavorable, lateral intermolecular energy transfer
across the semiconductor surface has been observed.12,13 Since
MLCT excited states are often long-lived, ultrafast injection is
not necessarily required for high efficiency DSSCs. For many
sensitizers, injection on nanosecond time scales would still be
expected to occur with quantum yields near unity. A potential
advantage of slower injection is that the rate constant for the
unwanted charge recombination reaction between the oxidized
sensitizer and the injected electron, S+/TiO2(e−) → S/TiO2,
may also be retarded. Slow recombination allows more time for
ket = k0exp(−βRDA
)
(1)
where k0 is the rate constant at van der Waals separation, RDA is
the donor−acceptor distance, and β is the attenuation factor.
The magnitude of β has been correlated with the molecular
details of the bridge in many literature reports.15−17 At
sensitized TiO2 interfaces, distance dependencies are much
less well understood in spite of many studies designed to
Received: March 1, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja402193f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
quantify them.18−28 Sensitizers with methylene bridges or rigid
OPE and phenyl bridges positioned between the sensitizer and
carboxylic acid surface binding groups have failed to provide
consistent β values. In a materials chemistry approach, core−
shell structures comprising spherical TiO2 cores coated with
thin insulating metal oxide layers that electrons must tunnel
through have been sensitized to visible light, but have also failed
to show a systematic distance dependence for electron
transfer.22−24 Distance arguments have in fact been invoked
to rationalize interfacial charge recombination trends where the
“hole” was translated from the oxidized sensitizer to spatially
isolated molecular orbitals on a pendant electron donor, but
such behavior does not appear to be general.25−28 In all these
molecular−semiconductor interfaces, the true charge transfer
distance remains speculative. Here we provide evidence that
electroabsorption spectroscopy can be used as a direct in situ
tool for the characterization of donor−acceptor distances at
molecular−semiconductor interfaces.
we take advantage of a recently discovered electro-absorption
signature, similar to that observed in Stark spectroscopy,30 and
correlate electric field strength to charge recombination rate
constants with the expanded series of sensitizers shown in
Chart 1 where the distance from the Ru metal center to the
anchoring oxygen atoms on the carboxylic acid groups are
estimated to be 7, 12, and 18.6 Å. The data suggest that Stark
spectroscopy can provide information on the magnitude of
surface electric fields of molecular sensitizers positioned at
variable distances from a semiconductor surface. The
implications of these findings for solar energy conversion are
discussed.
EXPERIMENTAL SECTION
■
Materials. The following reagents and substrates were used as
received, unless otherwise specified, from the indicated commercial
suppliers: acetonitrile (Burdick& Jackson, spectrophotometric grade);
toluene (OmniSolv, 99.99%); lithium perchlorate (Aldrich, 99.99%);
n-tetrabutylammonium perchlorate (TBAP; Fluka, >99.9%); n-
tetrabutylammonium chloride (TBACl; Sigma Aldrich, 98%); bis-
(triphenylphosphine)palladium(II) chloride (PdCl2(PPh3)2, Strem);
triphenylphosphine palladium (Pd(PPh3)4, Strem); di-μ-chlorobis[(p-
cymene)chlororuthenium(II)] (Strem); n-tetrabutylammonium iodide
(TBAI; Aldrich, >99% or Fluka, >98%); platinum(IV) chloride (Sigma
Aldrich); silver nitrate (Bioanalytical Scientific Instruments, Inc.);
hydrochloric acid (Fisher Scientific, 37.2% aqueous solution); argon
gas (Airgas, >99.999%); oxygen gas (Airgas, industrial grade);
dimethyl isophthalate (Sigma-Aldrich); N-bromosuccinimide (Sigma-
Aldrich); tetra-N-butylammonium fluoride (Sigma-Aldrich); [(4-
bromophenyl)ethynyl] (trimethyl)silane (Sigma-Aldrich); sodium
nitrite (Sigma-Aldrich), potassium iodide (Sigma-Aldrich); ammonium
thiocyanate (Sigma-Aldrich); copper(I)bromide (Sigma-Aldrich);
titanium(IV) isopropoxide (Sigma-Aldrich, 97%); zirconium(IV)
propoxide (Aldrich, 70 wt % solution in 1-propanol); SnO2 colloidal
solution (Alfa Aesar, 15% in water); anhydrous sodium sulfate
(Na2SO4 VWR); N,N-dimethylformamide (VWR); 2,2′-dipyridyl
(VWR); sodium hydroxide (NaOH, VWR); 10% Pd/C (Fisher-
Acros); sodium bisulfate (Fisher-Acros); potassium dichromate
(Fisher-Acros); tetra-N-butylammonium hydroxide (1.0 M in
methanol, Alfa Aesar); fluorine-doped SnO2-coated glass (FTO;
Hartford Glass Co., Inc., 2.3 mm thick, 15 Ω/□); and microscope
slides (Fisher Scientific, 1 mm thick); silica gel (230−400 mesh,
Sorbent Technologies); hexanes (Pharmco) were distilled prior to use
in column chromatography; diisopropyl amine (Pharmco) and
dichloromethane (Pharmco) were distilled over calcium hydride;
THF (Pharmco) was distilled with sodium and benzophenone;
deuterated NMR solvents were purchased from Cambridge Isotopes.
Sensitizer Synthesis. The ruthenium isothiocyanate compounds
AK0−2 were prepared according to Scheme 1, adapting a tandem
synthesis reported by Wang and co-workers for heteroleptic Ru
complexes.31 Compound AK0 was prepared as a carboxylic acid by
using the dcb ligand (2,2′-bipyridine-4,4′-dicarboxylic acid) in the first
step. Ligands 5a and 5b were synthesized as previously reported.32−36
Rigid-rod complexes AK1 and AK2 were prepared as esters, 6a and
6b, respectively to improve their solubility in inorganic solvents for the
purification and characterization process, and were then hydrolyzed to
acids. Compounds AK1 and AK2 were present as stereoisomeric
mixtures since the isophthalic unit can be trans to the pyridyl unit or to
the isothiocyanate ligand.
Electron transfer in molecular donor−bridge−acceptor
compounds is usually characterized when the concentration
of oxidized donors and reduced acceptors are equal. At
sensitizer−semiconductor interfaces, however, one is often
most interested in situations where this is not the case. For
example, in DSSCs under 1 sun illumination, the number of
electrons in each TiO2 nanocrystallite at the power point
condition is about 20, while the steady state value for oxidized
dyes is much less than 1.5 The excess electrons have been
quantified spectroscopically and by electrochemical means such
as the charge extraction method.1,5 These disparate concen-
trations of TiO2(e−) and oxidized sensitizers may change the
mechanism(s) of charge recombination from the second-order
kinetics expected when the numbers are equal. Furthermore,
the presence of excess electrons in TiO2 generates an electric
field that can influence redox potentials and electron transfer
dynamics in a manner that is not well understood. This study
reports the first example of interfacial charge recombination
where the bridge and the number of TiO2(e−)s were
systematically varied.
A previous letter reported S+/TiO2(e−) → S/TiO2 charge
recombination data for sensitizers AK0 and AK1 shown in
Chart 1, under conditions where the number of injected
electrons and oxidized sensitizers were equivalent.29 The kinetic
data were complex and modeled by a distribution of rate
constants whose average value was larger for AK0/TiO2. Here
Chart 1. The Ru(II)-Isothiocyanate Compounds Studied in
This Work and Their Abbreviations
General Procedure for the Synthesis of Complexes AK0,
AK1, and AK2.31,37,38 Di-μ-chlorobis[(p-cymene)chlororuthenium
(4)] (1 equiv, 0.25−0.75 mmol) and the substituted dipyridyl ligand
5a, 5b, or dcb (2 equiv) were dissolved in DMF (160 mL/mmol of 4)
under N2 and heated to 85 °C for 4 h, forming an orange-red solution.
2,2′-Dipyridyl (2 equiv) was added, and the reaction mixture was
heated to 150 °C for an additional 4 h, upon which time the solution
darkened. An excess of ammonium thiocyanate (15 equiv) was then
added and the reaction mixture was maintained at 150 °C with stirring
B
dx.doi.org/10.1021/ja402193f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 1. Synthesis of Sensitizers AK0, AK1, and AK2
for an additional 4 h. The reaction mixture was allowed to cool to
room temperature overnight and then the solvent was removed in
vacuum. Water was added to the purple crude product, and the
suspension was sonicated for 10 min. At this point, esters 6a and 6b
were collected via vacuum filtration and washed with water. For
carboxylic acid 1, the pH of the suspension was adjusted to
approximately pH 2 with 0.2 M HNO3(aq.). The suspension was
refrigerated (5 °C) overnight, then filtered, the purple solid washed
with water, and the precipitate was collected to afford 1.
possible. ESI+: calculated, 769.0244 [M+Na]+, 784.9983 [M+K]+;
found, 769.0328 [M+Na]+, 785.0070 [M+K]+.
Methyl Ester 6b. The complex was prepared from 7 (357 mg,
0.758 mmol), 4 (231 mg, 0.379 mmol), DMF (55 mL), 2,2′-dipyridyl
(116 mg, 0.757 mmol), and ammonium thiocyanate (432 mg, 5.69
mmol). Yield: 0.225 g, 33.6%. 1H NMR (DMSO): δH 9.31 (d, J = 5.9),
9.28 (dd, J = 7.9, 5.7), 8.95 (s), 8.85 (d, J = 8.4), 8.80−8.78 (m),
8.78−8.68 (m), 8.61 (dd, J = 13.2, 6.0), 8.46 (d, J = 9.8), 8.34 (d, J =
1.2), 8.31 (d, J = 1.2), 8.29−8.20 (m), 8.04 (dd, J = 6.1, 1.0), 7.99−
7.85 (m), 7.79 (s), 7.72 (d, J = 8.3), 7.69−7.61 (m), 7.60−7.49 (m),
7.32 (dd, J = 6.0, 1.1), 7.31−7.23 (m), 3.93 (s), 3.91 (s). The 1H NMR
data shows an integration ratio of 22:6 for protons in the aromatic
region to protons on the methyl ester, with the protons from the
methyl ester appearing at 3.93 and 3.91 ppm for each stereoisomer.
Due to the splitting of peaks, isomerization, and the low solubility,
more accurate proton assignment was not possible. The 13C NMR
spectrum was not recorded due to the low solubility. ESI+: calculated,
869.0559 [M+Na]+; found, 869.0550 [M+Na]+.
Ru(NCS)2[2,2′-Dipyridyl][2,2′-bipyridine-4,4′-dicarboxylic
acid], AK0 (1). The complex was prepared from 4 (303 mg, 0.496
mmol), 4,4′-dicarboxylic acid-2,2′-bipyridine (248 mg, 1.00 mmol),
DMF (87 mL), 2,2′-dipyridyl (158 mg, 1.00 mmol), and ammonium
thiocyanate (578 mg, 7.605 mmol) were used. Yield: 0.440 g, 71%. 1H
NMR (DMSO): δH 9.43 (d, J = 5.7, 1H), 9.27 (d, J = 5.5, 1H), 9.07 (s,
1H), 8.91 (s, 1H), 8.77 (d, J = 8.1, 1H), 8.62 (d, J = 8.1, 1H), 8.28 (t, J
= 7.3, 2H), 7.97 (t, J = 6.7, 1H), 7.91 (t, J = 8.2, 1H), 7.76 (d, J = 5.6,
1H), 7.61 (d, J = 5.5, 1H), 7.52 (d, J = 5.5, 1H), 7.29−7.18 (m, 1H).
ESI+: calculated, 640.9615 [M+Na]+; found: 640.9793 [M+Na]+.
Methyl Ester 6a. The complex was prepared from 4 (150 mg,
0.245 mmol), 5a (182 mg, 0.489 mmol), DMF (43 mL), 2,2′-dipyridyl
(78 mg, 0.489 mmol), and ammonium thiocyanate (250 mg, 3.2
General Procedure for Hydrolysis of Esters to Acids.39
Methyl esters 6a and 6b were dissolved in freshly distilled
dichloromethane (2500 mL/mmol complex), and 2 N solution of
NaOH in methanol (250 mL/mmol complex) was added; and the
mixture was stirred at room temperature for 4 days. Water was added,
and the mixture was extracted with water. Approximately 0.5 M HCl
was used to neutralize the aqueous solution to approximately pH 3 and
precipitate the product. The product was then filtered, and DMF was
used to recover the purple solid.
AK1 (2): Methyl Ester 6a. 6a (23 mg, 0.031 mmol), freshly
distilled dichloromethane (55 mL), and 2 N NaOH solution in
methanol (7 mL) were used. Yield: 0.013 g, 58.5%. ESI+: calculated,
740.9930 [M+Na+], 756.9669 [M+K+]; found, 741.0118 [M+Na+],
756.9857 [M+K+].
AK2 (3): Methyl Ester 6b. 6b (18 mg, 0.021 mmol), freshly
distilled dichloromethane (50 mL), 2 N NaOH solution in methanol
(6 mL) were used. Yield: 0.005 g, 29.3%. ESI+: calculated, 841.0245
[M+Na]+; found, 841.0250 [M+Na]+.
Sensitized Metal Oxide Thin Films. Anatase TiO2 nano-
crystallites were prepared by hydrolysis of a Ti(i-OPr)4 precursor by
a previously described sol−gel technique.18 The sols were cast as
mesoporous thin films (∼10 μm thick) by doctor blading onto glass
1
mmol). Yield: 0.300 g, 82.12%. H NMR (DMSO): δH 9.33 (d, J =
5.8), 9.29 (d, J = 4.7), 9.02 (s), 8.87 (s), 8.85 (d, J = 8.0), 8.79−8.73
(m), 8.71 (d, J = 8.1), 8.65−8.58 (m), 8.54 (s), 8.48 (s), 8.46 (s), 8.33
(s), 8.30−8.19 (m), 8.10 (d, J = 5.8), 7.99−7.91 (m), 7.91−7.84 (m),
7.67−7.62 (m), 7.59 (d, J = 5.6), 7.57−7.51 (m), 7.37 (d, J = 5.8),
7.32−7.22 (m), 3.95 (s), 3.91 (s). 1H NMR data shows an integration
ratio of 17.5:6 for protons in the aromatic region to protons on the
methyl ester, with the proton peak for the two stereoisomers occurring
at 3.95 and 3.91 ppm. 13C NMR (DMSO): δC 164.54, 164.47, 162.27,
159.12, 158.35, 158.19, 158.13, 157.85, 157.63, 157.04, 156.96, 156.77,
156.56, 152.28, 151.53, 151.48, 136.77, 136.65, 136.49, 136.12, 136.02,
135.92, 135.77, 133.58, 133.39, 131.22, 131.11, 128.84, 128.02, 127.96,
127.29, 127.04, 126.92, 126.34, 126.33, 126.25, 125.36, 125.12, 123.79,
123.69, 123.52, 123.50, 123.43, 123.39, 123.35, 123.31, 122.50, 122.37,
93.49, 88.43, 87.98, 52.82 (2C, −methyl ester), 52.77 (2C′, −methyl
ester). The presence of stereoisomers results in splitting of the 52.82
and 52.77 methyl ester. Due to the low solubility and the presence of
stereoisomers, an accurate assignment of the NMR spectra is not
C
dx.doi.org/10.1021/ja402193f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Table 1. Redox and Photoelectrochemical Properties of the Sensitized Thin Films
a
E1/2(RuIII/II) (mV)
c
d
e
f
g
sensitized film
TiO2
FTO
E1/2(RuIII/II*) (mV)
i
IPCE
τ
ϕinj
b
AK0/TiO2
AK1/TiO2
AK2/TiO2
xx
940
930
910
−660
−700
−730
1.85 0.1
3.20 0.3
3.40 0.4
0.78
0.48
0.39
27
31
26
5
5
5
0.98 0.03
0.93 0.03
0.70 0.03
900
880
a
b
The RuIII/II reduction potential measured for the indicated sensitizer anchored to fluorine-doped tin oxide (FTO) and TiO2. Irreversible redox
chemistry was observed. The excited state reduction potential. Ideality factor (i). Incident photon-to-current efficiency measured at 525 nm in 0.5
M LiI/0.05 M I2. Excited state lifetime at room temperature in CH3CN. Excited state injection yield quantified by comparative actinometry.
c
d
e
f
g
Figure 1. (A) Visible absorption spectra of AK0/TiO2 (red), AK1/TiO2 (green), and AK2/TiO2 (blue) in 0.3 M LiClO4/CH3CN with a bias of
+300 mV () and −300 mV vs NHE (---). The inset shows the magnitude of the long wavelength absorption change measured from the data. (B)
The visible absorption spectra of the sensitized thin films measured without () and with 1.03 mM I2 (---). Inset show the fraction of I2-adduct, θ,
as a function of the I2 concentration with an overlaid fit to the Langmuir adsorption isotherm model, K = 1.8 0.2 × 104 M−1.
microscope slides for spectroscopic measurements and transparent
conductive substrates (FTO) for electrochemical measurements. The
thin films were annealed at 450 °C for 30 min under an atmosphere of
O2.
a frequency doubled Nd:YAG Brilliant B Blue Sky. A pulsed 150 W Xe
lamp coupled to a Spectral Energy GM-252 monochromator served as
the probe light. The excitation fluency was measured with a thermopile
power meter (Molectron). A Spex 1702/04 monochromator optically
coupled to a 928 Hamamatsu photomultiplier tube with a computer
interfaced LeCroy 9450 oscilloscope was employed for detection.
Typically the absorption change measured after 60−100 laser pulses
were averaged at each observation wavelength. For short time scales
(<90 μs), the Xe lamp was pulsed and the sample was protected with a
fast shutter, appropriate UV and heat absorbing glass, and solution
filter combinations that also prevented PMT fatigue. The instrument
response time was ∼10 ns.
In some experiments, an external bias was applied to the sensitized
thin film in a spectroelectrochemical cell. In other experiments, the
514.5 nm line of an argon ion laser (Coherent) was used to illuminate
the sensitized film. Comparative actinometry with RuII(bpy)32+ and cis-
RuII(dcb)2(NCS)2/TiO2 was used to quantify the injection yields on
nanosecond and longer time scales as previously described.21,40 Low-
temperature data were obtained with a liquid nitrogen CoolSpeK UV
USP-203 four window cryostat (Unisoku).
The sensitizers were anchored to the metal oxide film by reaction in
a 0.5−2 mM solution of acetonitrile/tert-butanol (50/50 mixture, v:v).
The sensitized metal oxide thin films were immersed in the solutions
and then kept at 25 °C until the desired surface coverage was achieved.
The cuvettes containing the thin film and electrolyte solution were
purged with Ar(g) for at least 30 min prior to experimentation. The
macroscopic surface coverage, Γ in mol/cm2, was determined from the
measured absorption with a modified Beer−Lambert law: Abs = 1000
× Γ × ε, where ε was the molar decadic extinction (absorption)
coefficient, M−1 cm−1, that was assumed to be the same in solution and
on the surface. The saturation surface coverages were typically Γ ∼ 4.4
× 10−8 mol/cm2 which corresponds to roughly 500 sensitizers per
TiO2 nanocrystallite.5 Sensitized films were stored in acetonitrile for at
least an hour before use.19,20
Counter electrodes were prepared by deposition of PtCl4 dissolved
in isopropanol onto FTO that was heated at 450 °C for 30 min under
an atmosphere of O2. The solar cell was completed by sandwiching the
sensitized TiO2 thin film and the Pt electrodes together with a 100 μm
thick Surlyn film heated to ∼70 °C. An electrolyte solution of 0.5 M
LiI and 0.05 M I2 in acetonitrile was introduced between the
electrodes through a hole drilled in the Pt-counter electrode. The hole
was later sealed with Surlyn.
Photoluminescence. Steady-state photoluminescence measure-
ments were obtained with 514.5 nm excitation from an argon ion laser
(Coherent Innova) and a double monochromator (SPEX 1682)
optically coupled to a red sensitive Hamamatsu photomultiplier tube
(Hamamatsu 636-10) for detection.
Electrochemistry. A potentiostat (BAS model CV-50W or Epsilon
electrochemical analyzer) was employed for measurements in a
standard three-electrode arrangement with a sensitized TiO2 thin film
deposited on an FTO substrate working electrode or a sensitized FTO
slide working electrode,41 a Pt gauze counter electrode, and a
nonaqueous silver reference electrode (Bioanalytical Scientific Instru-
ments, Inc.). The ferrocenium/ferrocene (Fc+/Fc) half-wave potential
measured in a 0.3 M LiClO4 acetonitrile electrolyte was used as a
standard to calibrate the reference electrode. Conversion to the normal
Absorption Spectroscopy. UV−vis absorbance spectra were
collected using a Varian Cary 50 spectrometer or an Aviv
spectrophotometer model 14NT-UV−vis at room temperature in a
standard 1 cm path length quartz cuvette. Sensitized metal oxide slides
were positioned at a 45° angle in a 0.3 M LiClO4/CH3CN filled
cuvette. The solutions were purged with argon gas for at least 30 min
prior to absorption studies.
Nanosecond transient absorbance measurements were performed
with a 532 nm laser excitation (ca. 8 ns fwhm at the magic angle) from
D
dx.doi.org/10.1021/ja402193f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
band centered at 1630 cm−1 assigned to the asymmetric νco
stretch.
hydrogen electrode (NHE) was achieved with a conversion constant
of −630 mV from NHE to Fc+/Fc in acetonitrile at 25 °C.42
The absorption attributed to TiO2(e−) increased exponen-
tially with an applied negative bias, behavior consistent with
many previous studies.46 The absorption data were converted
to a chemical capacitance using Faraday’s Law, the electrode
area, and the measured extinction coefficient of 1000 M−1 cm−1
at 800 nm, as was previously described.46 We note that while
this extinction coefficient is in good agreement with previous
measurements,1 it differs from a value recently reported for
related materials.47 The use of an alternative extinction
coefficient would quantitatively change the magnitude of the
capacitance, but would otherwise have no impact on the results
of this paper that is largely comparative in nature. The
capacitance as a function of applied potential is shown in Figure
S1, with an overlaid fit to eq 2,
Photoelectrochemistry. Photocurrent action spectra were
measured in a two electrode configuration with a previously described
apparatus that consisted of a PhotoMax 100 W Xe-lamp optically
coupled to a 1/4 m Oriel Corner Stone monochromator.43 Incident
irradiances were measured with a calibrated silicon diode. Photo-
currents were quantified with a Keithley 617 electrometer. The
incident photon-to-current conversion efficiency (IPCE) was calcu-
lated as the number of injected electrons measured in the external
circuit divided by the number of photons incident on the solar cell.
The open circuit photovoltages were measured as a function of
incident 514.5 nm irradiance with an Ar ion laser. A two electrode
configuration was employed with a sensitized TiO2 thin film as the
illuminated electrode and a Pt coated FTO electrode as the reference.
Experiments with a Ag/AgCl reference electrode (Bioanalytical
Scientific Instruments, Inc.) in 0.3 M LiClO4 acetonitrile solution
were conducted to ensure that the measured values were independent
of the reference electrode. The voltage difference was measured with
the Keithley electrometer before and after illumination. The incident
irradiance was attenuated with neutral density filters.
⎛
⎞
e
C = C exp
V
⎜
⎟
⎠
0
ak T
⎝
(2)
B
where C is the measured chemical capacitance, C0 is the
capacitance intrinsic to the materials, that is, the total
distribution of states that can store charge,48 T is the absolute
temperature, and kB is the Boltzmann constant. The best fit
gave α ≈ 6, a value that was within experimental error the same
for sensitized and unsensitized thin films. An α = 6 value was
extracted from data measured on an unsensitized TiO2 thin film
that enabled characterization over a wider potential range.
The influence of 0 to 1.33 mM I2 on the absorption spectra
of the sensitized thin films was quantified in 0.3 M LiClO4/
CH3CN, Figure 1B. The addition of I2 to the external solution
surrounding the films resulted in a significant blue shift in the
MLCT absorption with the maintenance of an isosbestic point
near 500 nm. Concentrations of I2 greater than 1.33 mM had
no further influence on the MLCT absorption. Shown in Figure
1B are the MLCT absorption spectra measured at 0 and 1.33
mM I2 concentrations after contributions from I2 were
subtracted. The inset shows a Langmuir adsorption isotherm
for the spectral changes observed at 450 nm with a best fit to eq
3,
RESULTS
■
The three cis-Ru(bpy)(LL)(NCS)2 coordination compounds
shown in Chart 1 were synthesized and characterized
spectroscopically. The visible absorption spectra of the
compounds measured in fluid CH3CN were assigned to
metal-to-ligand charge transfer (MLCT) transitions. Intense
π-to-π* transitions were observed at 300 nm.44 Both AK1 and
AK2 displayed an additional absorption band in the 340−360
nm range that was assigned to the OPE bridge.45 Light
excitation into the MLCT absorption resulted in room
temperature photoluminescence, PL. The PL maxima were
770 nm for AK0 and AK1, and the maximum was 750 nm for
AK2. Excited state relaxation was first-order with lifetimes of τ
= 30 4 ns, Table 1.
The compounds were anchored to nanocrystalline (anatase)
TiO2 thin films by immersing the films overnight in
acetonitrile/tert-butanol solutions of the sensitziers. Typical
surface coverages attained by this method were 4 × 10−8 mol/
cm2. Shown in Figure 1A are representative absorption spectra
of the sensitized thin films measured in a three electrode
configuration with a +300 and a −300 mV applied bias. The
spectra measured at +300 mV were within experimental error
the same as that measured in the absence of an applied
potential. With an applied bias of −300 mV, the characteristic
absorption spectrum of reduced TiO2, abbreviated TiO2(e−),
was observed and was subtracted from the data shown. The
MLCT absorption was found to blue shift in the presence of
TiO2(e−).
To help visualize the MLCT absorption changes that
accompanied TiO2 reduction for all three sensitizers, the
absorption shift in wavenumbers measured between 0.30 and
0.35 absorbance units are shown as the Figure 1 inset. A 190
cm−1 change was observed for AK0/TiO2(e−), and 100 and 40
cm−1 change for AK1/TiO2(e−) and AK2/TiO2(e−), respec-
tively. These spectral shifts were reversible when the TiO2(e−)
were removed from TiO2 with a positive bias of +300 mV.
Cathodic excursions negative of −350 mV resulted in
absorption changes that were irreversible and were therefore
avoided. Spectra that include contributions from TiO2(e−) as
well as unsensitized TiO2 thin films are given in Figure S1 in
the Supporting Information. The ATR-FTIR spectra for the
sensitized thin films measured at open circuit revealed a broad
K[I2]
θ =
1 + K[I2]
(3)
where θ is the fraction of sensitizer present in the I2-adduct
form.49 The adduct formation constant, K, was estimated to be
1.8
0.2 × 104 M−1 for all three sensitizers. The spectral
changes were found to be >90% reversible when CH3CN or
iodide were added to the external solution.
Cyclic voltammograms of the sensitizers anchored to TiO2 or
to fluorine doped tin oxide (FTO) are shown in Supporting
Information Figures S2 and S3. The electrochemistry measured
on FTO was reversible as the cathodic and anodic currents
were equal, ipa/ipc = 1, and the peak-to-peak separation ΔEpp
<
59 mV at a scan rate of 100 mV/s. The formal reduction
potential for the three sensitized coordination compounds were
as follows: AK0III/II = 950 mV, AK1III/II = 940 mV, and AK2III/II
= 900 mV vs NHE. On TiO2, the electrochemistry was
irreversible for AK0/TiO2 and quasi-reversible for the other
two sensitizers with ipa/ipc = 1, and ΔEpp > 100 mV. The excited
state reduction potentials E0(RuIII/II*) were calculated with eq
4,
E0(RuIII/II*) = E0(RuIII/II) − ΔGES
(4)
E
dx.doi.org/10.1021/ja402193f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 2. Transient absorption data monitored at 500 nm after pulsed 532 nm laser excitation (0.2 mJ/pulse) of AK0/TiO2 (black), AK1/TiO2
(red), and AK2/TiO2 (green) in 0.3 M LiClO4 CH3CN with an applied bias of (A) +150 mV, (B) −150 mV, and (C) −350 mV versus NHE.
Overlaid are fits to a KWW kinetic model (white).
Figure 3. Transient absorption data monitored at 500 nm after pulsed 532 nm laser excitation (0.2 mJ/cm2) of AK0/TiO2 (black), AK1/TiO2 (red),
and AK2/TiO2 (green) in 0.3 M LiClO4/CH3CN under (A) steady state illumination (180 mW/cm2 of 514.5 nm light); the inset shows without
steady state illumination; (B) pulsed 1.4 mJ/cm2 excitation irradiance while the inset shows 0.2 mJ/cm2. The data in panel (C) was measured after
pulsed laser excitation (1.0 mJ/cm2) of AK0/TiO2 at +40 °C (black), AK0/TiO2 at −20 °C (red), AK1/TiO2 at +40 °C (green), and AK1/TiO2 at
−20 °C (blue). The inset shows the first 80 μs.
Figure 4. (A) Transient absorbance monitored at 375 nm after pulsed 532 nm excitation of AK0/TiO2 (black), AK1/TiO2 (red), and AK2/TiO2
(green) in 0.5 M LiI/CH3CN. Overlaid are solid line fits to the Kohlrausch−Williams−Watts model. (B) Transient absorbance monitored at 500 nm
after pulsed 532 nm excitation of AK0/TiO2 (black), AK1/TiO2 (red), and AK2/TiO2 (green) in 0.3 M LiClO4/CH3CN in the absence () and
presence (---) of 0.11 mM I2.
where ΔGES is the free energy stored in the MLCT excited
state. A tangent line on the high energy side of the corrected PL
spectra, measured in solution, was used to obtain ΔGES. The
electrochemical data is summarized in Table 1.
Nanosecond transient absorption data were measured after
pulsed laser excitation of the samples as the applied potential
was varied from +350 to −350 mV vs. NHE. Transient data
monitored at 500 nm with +150, −150, and −350 mV applied
bias are shown in Figure 2 on a log time scale. The insets of
Figure 2 show the first 5 μs of the same data plotted on a linear
time scale. The ground state absorption and incident irradiance
at 532 nm were the same for this study. Data recorded at other
applied potentials are given in Figure S4 and Videos S1 and S2
in the Supporting Information. Overlaid on the data are best fits
to the Kohlrausch−Williams−Watts (KWW) kinetic model, eq
5.
Abs(t) = A0 exp[−(t/τ0)β′]
(5)
F
dx.doi.org/10.1021/ja402193f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Here A0 is the initial absorption change, β′ is inversely related
TiO2 surface. The sensitized interfaces were thus conceived
with the hope that the interfacial energetics would be
unchanged as the number of phenylene ethynylene bridge
units was increased. This was found to approximately be the
case. The RuIII/II reduction potentials varied by only 30 mV and
the TiO2 acceptor states were independent of the identity of
the molecular sensitizer. Similar to the prototypical “N3”
sensitizer,52 that is, cis-Ru(dcb)2(NCS)2, the excited states
characterized herein are potent reductants with favorable
overlap with the TiO2 acceptor states for efficient excited
state injection, Figure 5.
́
to the width of the underlying Levy distribution of rate
constants, 0 < β′ < 1, and τ0 is a characteristic lifetime.50 Note
that the data measured with a −350 mV applied bias is shown
on a shorter time scale as recombination was much faster under
these conditions.
The influence of steady state illumination on charge
recombination was investigated with an argon ion laser (λ =
514.5 nm, irradiance =180 mW/cm2). This steady state
irradiance increased the open circuit photovoltage by ∼200
mV relative to that measured with only the white light probe
beam of the transient absorption apparatus. The spectral
features observed after pulsed laser excitation were the same as
that measured without steady state illumination, but the
kinetics differed. This is shown by the comparison of Figure
3A and the inset. Significant increases in the charge
recombination rate with the steady state irradiance were
observed for AK0/TiO2, while AK1/TiO2 showed a very small
increase with no measurable change for AK2/TiO2.
The influence of pulsed 532 nm excitation fluency on charge
recombination was also quantified at the open circuit condition.
Figure 3B shows data measured with 1.4 mJ/pulse (for a 1 cm2
area) and the inset show the same samples with an excitation
fluency of 0.2 mJ/pulse. The higher fluency enhanced
recombination at short time scales, but had no measurable
influence on the long time scale data, see Supporting
Information Video S3. Transient absorption data for the
sensitized thin films was quantified from +40 to −20 °C, Figure
3C. The short time scale data was temperature dependent with
no or very little dependence on the long time scale.
Figure 5. Energetics for excited state injection and charge
recombination at the sensitized TiO2 interfaces under study.
Transient absorption changes monitored at 375 nm after
pulsed 532 nm excitation of the sensitized thin films immersed
in 0.5 M LiI/CH3CN are shown in Figure 4A. Note that the
data acquired on the sub 100 μs time were smoothed with a 20
data point average. The kinetic data was well described by the
KWW model with β = 0.61 and k = 5.26 s−1 for all three
sensitized materials. The experiment was also performed with
0.63 mM I2 in place of the 0.5 M LiI, Figure 4B. The initial
amplitude of the absorption change was attenuated by about a
factor of 6 for AK1/TiO2 and AK2/TiO2 yet had a much
smaller ∼20% effect on AK0/TiO2. In addition, kinetic data
Furthermore, Stark spectroscopic data suggested that the cis-
Ru(NCS)2(bpy) core was on average further from the TiO2
surface when phenylene ethynylene spacer were present
between the bipyridyl ligand and the surface anchoring
carboxylic acid groups. The magnitude of the electric field
abstracted from such data was correlated with the S+/TiO2(e−)
→ S/TiO2 charge recombination rate constants with a bridge
length dependence indicating that the oxidized metal center
was also further from the surface. The photophysical data
measured in inert electrolytes indicated that sensitizers with
phenylene ethynylene bridges would yield high efficiencies
when utilized in a dye sensitized solar cell, yet this expectation
was not realized and an unwanted quenching pathway with
molecular iodine was identified. These behaviors are detailed
below in the broader context of the relevant literature.
•‑
•‑
were also collected at 400 and 700 nm, where I2 /I3− and I2 /
TiO2(e−) absorb light, respectively.51 The presence of I2 was
found to lower the excited state injection yield without the
formation of new chemical products.
The incident photon-to-current efficiency (IPCE) for the
sensitized thin films were measured in 0.5 M LiI and 0.05 M I2/
CH3CN, Supporting Information Figure S6A. With 525 nm
excitation, the IPCE values were as follows: AK0/TiO2 = 0.78,
AK1/TiO2 = 0.48, and AK2/TiO2 = 0.39, Table 1. The open
circuit voltage (Voc) was measured in the absence of redox
mediator in 0.3 M LiClO4/CH3CN versus a reference
electrode. Plots of Voc versus the log of the irradiance were
linear with slopes of i = 2, 3.2, and 3.4 for AK0/TiO2, AK1/
TiO2, and AK2/TiO2, respectively. This data is given in
Supporting Information Figure S6B and summarized in Table
1.
I. Stark Effects. The visible absorption spectra of the Ru(II)
compounds are characteristic of metal-to-ligand charge-transfer
(MLCT) transitions in ruthenium polypyridyl compounds. The
spectra were specifically in agreement with expectations for cis-
Ru(NCS)2 polypyridyl compounds that display significant solar
light harvesting through the visible region to about 700 nm.1
Reflectance infrared studies show that the sensitizers bind to
TiO2 in their carboxylate form.53
The photoinduced or thermal transfer of electrons to TiO2 is
known to induce a blue shift of the MLCT absorption
consistent with an underlying electric field, referred to as a
Stark effect.54 Here we found that the magnitude of the shift
was largest for the sensitizer without a phenylene ethynylene
bridge and smallest for that with two bridge units. When
measured as an absorption difference with and without the
presence of TiO2(e−), the spectral data were well modeled by a
first-derivative of the ground state absorption spectrum,
consistent with an antiparallel alignment of the surface electric
DISCUSSION
■
The goal of this work was to elucidate whether the kinetic rate
constant for interfacial charge recombination, kcr, was
influenced by the surface electric field when the oxidized
ruthenium compound was set at a variable distance from the
G
dx.doi.org/10.1021/ja402193f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
field with the change in the sensitizer dipole moment vector, θ
= 180°. The magnitude of the spectral shift was directly related
to the field strength through the change in the dipole moment,
eq 6.55
The slow charge recombination kinetics processes were
insensitive to the identity of the molecular sensitizer, the
temperature, and the incident irradiance. The data suggest a
temperature independent rate limiting electron tunneling step
as was proposed by Furube and Katoh.58 It was curious that this
slower tunneling process was insensitive to the number of
bridge units, given that the faster recombination processes were
(see below). A mechanism where electron tunneling between
acceptor states rate limits a more rapid electron transfer to the
oxidized sensitizer was therefore most consistent with this data.
The fast charge recombination dynamics were found to be
sensitizer dependent, behavior that was most pronounced when
the steady state illumination, TiO2(e−) concentration, and/or
the pulsed laser fluence were increased. Potentiostatic control
of the sensitized thin film was employed to increase the
concentration of TiO2(e−)s, a mechanistic approach first
reported by O′Regan and co-workers,59 that was later
effectively utilized by the Durrant group60 who found that an
increased TiO2(e−) concentration resulted in enhanced charge
recombination rates. Here we find similar behavior with the
extent of the enhancement being directly related to the number
of bridge units present in the sensitizer.
|Δμ||E| cos θ
Δv = −
̅
(6)
100hc
The change in dipole moment for these compounds is
unknown, however a value for cis-Ru(dcb)2(NCS)2 of |Δμ| =
13.3 D was reported.56 Since the compounds under study all
have a cis arrangement of the isothiocyanate ligands and the
MLCT absorption is predominantly a Ru → bpy′ localized
transition, we assert that this value represents a reasonable
approximation for all three compounds. With this assumption,
electric fields of magnitude E = 0.85, 0.45, and 0.20 MV/cm
were calculated for these sensitized materials as the number of
OPE units was increased from zero to two. While some surface
heterogeneity and the presence of necking regions between the
TiO2 nanocrystallites within the mesoporous thin film are
expected, the data indicates that the “average” field strength
experienced by the chromophoric portion of the sensitizers is
that expected when the sensitizers are oriented normal to the
TiO2 surface. Abstraction of a true distance between the cis-
Ru(NCS)2(bpy)-core and the TiO2 interface was complicated
by the aforementioned heterogeneity, the unknown location of
the trapped electrons, and the unknown interfacial dielectric
constant. Nevertheless, the Stark data support the notion that
on average AK2 was further from the TiO2 surface then was
AK1 than AK0.
The significant electroabsorption measured for AK0/
TiO2(e−) that decreased in magnitude by about 1/2 and 1/4
as the phenylene ethynylene bridge units were incorporated are
reminiscent of previous redox titrations reported by Gregg,
Zaban, and Ferrere.57 These investigators found that the
reduction potentials of molecules located within the electric
double layer were profoundly influenced by surface electric
fields, while molecules outside were far less sensitive.57 This
raises the question of what the actual E0(RuIII/II) reduction
potentials are when electrons are injected into TiO2 and the
electric field is large. The persistence of such large electric fields
indicates that the anatase TiO2 and the acetonitrile electrolyte
do not quantitatively screen the field. A small distance between
the TiO2(e−) and the sensitizer would inhibit substantial
screening by the electrolyte while the bridge units enable more
substantial screening. Indeed the notion that AK0 is within the
inner Helmholtz layer while AK1 and AK2 are outside the
double layer represents a simplified model in which much of
the electron transfer dynamics described below can be
understood.
When the TiO2(e−) concentration was high, the slow
components for charge recombination were not observed,
presumably because the redox active states responsible for them
were filled and tunneling between them was therefore absent.
Under such conditions, the use of the KWW function, which is
a paradigm for modeling transport in disordered materials like
these TiO2 thin films,61 was of questionable value even though
it did accurately model the experimental data. We therefore
utilized a simple t1/2 analysis or considered the fraction of
charge recombination that occurred under specific conditions.
With low irradiances and at open circuit or with a forward bias
of +200 mV, change recombination over the first 85 μs was
negligible for all the sensitized materials under study. As the
quasi-Fermi level of the TiO2 nanocrystallites was raised to +50
mV, about 20% of the electrons injected from AK0/TiO2 had
recombined while there was no measurable change for AK1 or
AK2. A forward bias of −50 and −150 mV were required to
promote the same 20% increase in charge recombination for
AK1 and AK2, Figure 6. Stark spectroscopy showed that the
field generated by TiO2(e−)s was strongly felt by AK0, but to a
much lesser extent by AK1 and AK2. Taken together, these
II. Interfacial Electron Transfer. Recombination of an
injected electron with an oxidized sensitizer represents an
unwanted reaction that wastes energy and yields ground state
products. This reaction has previously been studied by several
research groups, mainly by time-resolved absorption spectros-
copy.1,2 The results described herein provide new insights into
this important reaction. Comparative experiments were
performed where the number of absorbed photons, the applied
potential (and hence the TiO2(e−) concentration), and the
temperature were controlled. In discussions of these results, it is
convenient to consider two separate time scales of charge
recombination: slow (i.e., 10−4−10−1 s) reactions and fast (i.e.,
10−7−10−4 s) reactions.
Figure 6. Fraction of charge recombination that occurred over the first
85 μs after pulsed 532 nm laser excitation as a function of the applied
bias (black square, AK0/TiO2), (red circle, AK1/TiO2), and (green
up-triangle, AK2/TiO2).
H
dx.doi.org/10.1021/ja402193f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
data indicate that charge recombination can be inhibited with
distance particularly when the number of injected electrons is
large. These conditions are relevant to the power point of
operational solar cells where over 20 electrons are expected to
reside in each TiO2 nanocrystallite as well as the open circuit
condition where the number is even larger.5 Such behavior is
particularly evident with videos provided in the Supporting
Information, and this almost certainly contributes to the
unusually high irradiance dependence of the open circuit
photovoltage, Voc, measured with ideality factors greater than
3.29
Previous time-of-flight and intensity-modulated photocurrent
measurements have shown that the apparent diffusion constant
for charge transport through the TiO2 thin films increased with
the number of TiO2(e−)s, behavior that has been rationalized
with trap filling models.62−64 Such transport also involves the
movement of ions, such as Li+, and is often termed ambipolar
diffusion. It is thus postulated that the observed rate constants
abstracted from data measured under forward bias represent
and hence high TiO2(e−) concentrations most accurately reflect
the true interfacial electron transfer rate constants for charge
recombination. The transient data measured at −350 mV were
thus fit to a distributional model based on a Levy distribution of
rate constants, or simply to the time required for half of the
charge separated states to recombine, t1/2. The latter analysis
revealed a significant distance dependence with an attenuation
factor of β ∼ 0.12 Å−1. Note that the sensitizer AK0 was not
included in this analysis as it did not contain a bridge unit and
hence was not homologous with the other two sensitizers.
While this value was abstracted from only the two bridged
sensitizers and is subject to a rather large uncertainty,
previously published data indicate that the mesoporous nature
of the TiO2 thin films likely preclude analysis over larger
distances.65
The notion that the short time scale data measured at open
circuit reports predominantly on charge recombination to the
oxidized sensitizer while the slower time scale data reports on
the electron tunneling between acceptor states is consistent
with most previous literature reports at sensitized TiO2
interfaces,1,5 and helps rationalize why some investigators
have found that charge recombination was highly sensitive to
the thermodynamic driving force66 while others found
recombination to be essentially independent of driving
force.24,67 High injection fluxes and a low density of acceptor
states with short observation times reflect conditions relevant to
interfacial electron transfer while low injection fluxes, high
acceptor densities, and long observation times reflect electron
tunneling between acceptor states.
redox active iodide−iodine mixtures, it was of interest to
quantify reactivity under these conditions.
The oxidized sensitizer was found to react with iodide to
•‑
yield di-iodide, I2 , as a product. The detailed mechanism by
which I2•‑ was formed remains unknown at the TiO2 interfaces,
but iodine atom and I2•‑ intermediates have been proposed.70,71
Regardless of the I−I bond forming mechanism, I2•‑ is unstable
with respect to disproportionation chemistry that yields tri-
−
iodide and iodide. Reaction of the injected electrons with I3
was quantified and found to be independent of the structure of
the sensitizer. Collectively, the data indicated that regeneration
and charge recombination to oxidized iodide acceptors were
within experimental error the same for all three sensitizers and
did not explain the poor energy conversion efficiency measured
with the OPE bridged sensitizers. The data do not support the
recent suggestion of a possible kinetic advantage to initiating
regeneration reactions further from the TiO2 interface, at least
in polar CH3CN solution.72
The standard redox mediator solutions of 0.5 M LiI and 0.05
M I2 were utilized in this study. The equilibrium constant for eq
7 has been reported to be 5.6 × 106 M−1 in acetonitrile.
Therefore, the equilibrium concentrations of the iodine species
present were 0.05 M I3 , 0.45 M I− and ∼10−7 M I2.68,69
−
I− + I2 ⇌ I−3
(7)
Titration of 1 M iodide or tri-iodide into the acetonitrile
electrolyte surrounding a sensitized thin film had no measurable
influence on the MLCT absorption spectra of the sensitizers.
However, the addition of I2 to the external acetonitrile
electrolyte surrounding a sensitized thin film resulted in
significant spectral changes could be reversed by dilution or
by addition of TBAI that decreased the I2 concentration
through the equilibrium in eq 7. Taken together, these
observation are consistent with the formation of a sensitizer−
I2 adduct as was recently reported by O’Regan and co-
workers.73 Adsorption isotherm data revealed an adduct
formation constant of about K = 1.8
0.2 × 104 M−1 that
was in reasonable agreement with their data for similar cis-
Ru(NCS)2 sensitizers that possess an isothiocyante ligand for I2
coordination.73
Pulsed laser excitation of these adducts did not appreciably
change the injection yield for AK0*/TiO2, but decreased the
yield by about 80% for the phenylene ethynylene bridged
sensitized films. Molecular I2 is an electron acceptor that could
reductively quench the excited state to yield the oxidized
74,75
•‑
sensitizer and I2 , thereby lowering the injection yield.
While the concentration of I2 is expected to be quite low based
on the equilibrium in eq 7, the value might be larger within the
sensitized semiconductor mesopores and could account, at least
in part, for the low photocurrent efficiencies measured in
operational solar cells. Indeed, recent solid state studies indicate
that cis-Ru(dcb)2(NCS)2 displays enhanced interactions with I2
III. Light-to-Electrical Energy Conversion. Rapid excited
state injection with slow microsecond recombination represents
kinetic behavior ideal for electrical power generation in dye
sensitized solar cells.1 However, when the sensitizers with OPE
76
−
relative to I3 . The data here reveal for the first time an
unwanted consequence of remote excited state injection in
redox active electrolytes: excited state quenching by redox
mediators that lowers the efficiency for electron injection.
The Voc values measured for phenylene ethynylene bridged
sensitized TiO2 in the absence of redox mediators displayed the
largest dependence on irradiance that has been measured. The
slopes abstracted from plots of Voc versus the log of the incident
irradiance were greater than 180 mV/decade where “ideal”
behavior, based on a Schottky junction model of the
semiconductor electrolyte interface, is 59 mV/decade. As this
−
bridge units were employed with standard I−/I3 redox
mediators, the incident photon-to-current efficiencies were
poor, less than 0.50. Under the same conditions, AK0/TiO2
gave nearly quantitative conversion efficiencies, behavior
consistent with this class of cis-Ru(NCS)2 type sensitizers.52
The ∼30% lower excited state injection yield measured for the
longest sensitizer certainly contributes to the lower efficiency,
but cannot explain the magnitude of the decreased photo-
current. As the electron transfer kinetics were measured under
inert electrolyte conditions while the solar cells also contained
I
dx.doi.org/10.1021/ja402193f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
model is inapplicable to these nanocrystalline thin films, the
origin of the open circuit photovoltage has been the subject of
many theoretical and experimental studies.77 One wishes to
understand how the quasi-Fermi level of the fluorine doped tin
oxide substrate is influenced by the number of electrons present
within the TiO2 nanocrystallites. There exists significant
experimental and theoretical data that implicate the presence
of an exponential distribution of acceptor states in TiO2.77 The
data measured herein support this picture. The chemical
capacitance measured experimentally is proportional to the
number density of acceptor states per unit energy and reveals
that a factor of 10 increase in the injected charge corresponds
to a 360 mV increase in voltage. Therefore, ideal behavior at
these interfaces is not the 59 mV of an ideal Schottky junction,
but 360 mV, and deviations from this must be attributable to
charge recombination.
ASSOCIATED CONTENT
■
S
* Supporting Information
Physical methods, spectroscopic, and electrochemical data, as
well as videos of the charge recombination data as the TiO2(e−)
concentration was increased with an external bias. This material
is available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
meyer@jhu.edu; galoppin@andromeda.rutgers.edu
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was funded by the Division of Chemical Sciences,
Geosciences, and Biosciences, Office of Basic Energy Sciences
of the U.S. Department of Energy through Grant DE-FG02-
01ER15256 (E.G.) and Grant DE-FG02-96ER14662 (G.J.M.).
The authors would also like to thank Professor Doug Barrick
and Jake Marold for use of the Model 14NT-UV−vis Aviv
spectrophotometer.
The more rapid charge recombination for AK0+/TiO2(e−)
results in the largest deviation from the expected 360 mV/
decade and hence the smallest ideality factor. To fully correlate
the Voc dependence on irradiance, one must quantify how
charge recombination rates depend on the quasi-Fermi level of
the TiO2. Qualitatively, the phenylene ethynylene bridges slow
down recombination particularly when the TiO2(e−) concen-
tration is large and the quasi-Fermi level is high which would
certainly give rise to the more ideal ∼180 mV/decade change in
Voc measured with these sensitizers. This discussion under-
scores the importance of measuring ideality factors and charge
recombination rate constants over a wide range of irradiances
rather than a single Voc value measured under one sun, that is,
100 mW/cm2, as is commonly done.2
REFERENCES
■
(1) Ardo, S.; Meyer, G. J. Chem. Soc. Rev. 2009, 38, 115−164.
(2) Hagfeldt, A.; Gratzel, M. Acc. Chem. Res. 2000, 33, 269−277.
(3) Yella, A.; Lee, H.-W.; Tsao, H. N.; Yi, C.; Chandiran, A. K.;
Nazeeruddin, M. K.; Diau, E. W.-G.; Yeh, C.-Y.; Zakeeruddin, S. M.;
Gratzel, M. Science 2011, 334, 629−634.
(4) OʼRegan, B.; Gratzel, M. Nature 1991, 353, 737−740.
(5) O′Regan, B. C.; Durrant, J. R. Acc. Chem. Res. 2009, 42, 1799−
1808.
(6) Benko, G.; Kallioinen, J.; Korppi-Tommola, J. E. I.; Yartsev, A. P.;
Sundstrom, V. J. Am. Chem. Soc. 2001, 124, 489−493.
(7) Asbury, J. B.; Ellingson, R. J.; Ghosh, H. N.; Ferrere, S.; Nozik, A.
J.; Lian, T. J. Phys. Chem. B 1999, 103, 3110−3119.
(8) Monat, J. E.; McCusker, J. K. J. Am. Chem. Soc. 2000, 122, 4092−
4097.
CONCLUSION
■
The photophysical properties of three polypyridyl compounds
varying in length, from ∼6 to 18.6 Å, that maintain a
Ru(NCS)2(bpy) core with variable number of OPE bridge
units were characterized by spectroscopic and electrochemical
methods in fluid solution and when anchored to nanocrystalline
TiO2 thin films immersed in CH3CN electrolyte. For the first
time, Stark spectroscopy was utilized as an in situ probe of the
average electric field experienced by sensitizers with variable
bridge lengths, and this data was correlated with a distance
dependence inferred from electron transfer kinetic measure-
ments. The bridge length dependence for charge recombination
was consistent with the attenuation factor of β = 0.12 Å−1.
Inhibited S+/TiO2(e−) → S/TiO2 charge recombination is
beneficial for solar energy conversion as larger steady state
concentrations of injected electrons can be realized. Slow
recombination is important in regenerative dye sensitized solar
cells when S+ is a weak oxidant as well as for photo-
electrosynthetic water splitting cells where the rates of water
oxidation are generally small. Furthermore, the strong
irradiance dependence of the recombination rate constants
and Voc values reveal the dangers of comparative studies
measured only at one irradiance and surface coverage. Finally,
an unwanted consequence of excited state injection from
sensitizers with long bridge units was identified, excited state
quenching by redox mediators present in the electrolyte, that
can possibly be avoided in the future through judicious choice
of the redox mediator.
(9) Hannappel, T.; Burfeindt, B.; Storck, W.; Willig, F. J. Phys. Chem.
B 1997, 101, 6799−6802.
(10) Tachibana, Y.; Moser, J. E.; Gratzel, M.; Klug, D. R.; Durrant, J.
R. J. Phys. Chem. 1996, 100, 20056−20062.
(11) Benko, G.; Kallioinen, J.; Korppi-Tommola, J. E. I.; Yartsev, A.
P.; Sundstrom, V. J. Am. Chem. Soc. 2002, 124, 489−493.
(12) Farzad, F.; Thompson, D. W.; Kelly, C. A.; Meyer, G. J. J. Am.
Chem. Soc. 1999, 121, 5577−5578.
(13) Higgins, G. T.; Bergeron, B. V.; Hasselmann, G. M.; Farzad, F.;
Meyer, G. J. J. Phys. Chem. B 2006, 110, 2598−2605.
(14) Altobello, S.; Argazzi, R.; Caramori, S.; Contado, C.; Da Fre, S.;
Rubino, P.; Chone, C.; Larramona, G.; Bignozzi, C. A. J. Am. Chem.
Soc. 2005, 127, 15342−15343.
(15) Smalley, J. F.; Sachs, S. B.; Chidsey, C. E. D.; Dudek, S. P.; Sikes,
H. D.; Creager, S. E.; Yu, C. J.; Feldberg, S. W.; Newton, M. D. J. Am.
Chem. Soc. 2004, 126, 14620−14630.
(16) Pettersson, K.; Kyrychenko, A.; Ronnow, E.; Ljungdahl, T.;
Martensson, J.; Albinsson, B. J. Phys. Chem. A 2005, 110, 310−318.
(17) Scott, A. M.; Wasielewski, M. R. J. Am. Chem. Soc. 2011, 133,
3005−3013.
(18) Heimer, T. A.; D′Arcangelis, S. T.; Farzad, F.; Stipkala, J. M.;
Meyer, G. J. Inorg. Chem. 1996, 35, 5319−5324.
(19) Neale, N. R.; Kopidakis, N.; van de Lagemaat, J.; Gratzel, M.;
Frank, A. J. J. Phys. Chem. B 2005, 109, 23183−23189.
(20) Trammell, S. A.; Meyer, T. J. J. Phys. Chem. B 1998, 103, 104−
107.
(21) Bergeron, B. V.; Kelly, C. A.; Meyer, G. J. Langmuir 2003, 19,
8389−8394.
J
dx.doi.org/10.1021/ja402193f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(22) Zaban, A.; Chen, S. G.; Chappel, S.; Gregg, B. A. Chem.
Commun. 2000, 2231−2232.
(56) Waterland, M. R.; Kelley, D. F. J. Phys. Chem. A. 2001, 105,
4019−4028.
(57) Zaban, A.; Ferrere, S.; Gregg, B. A. J. Phys. Chem. B 1998, 102,
452−460.
(23) Palomares, E.; Clifford, J. N.; Haque, S. A.; Lutz, T.; Durrant, J.
R. J. Am. Chem. Soc. 2003, 125, 475−482.
(58) Katoh, R.; Furube, A. J. Phys. Chem. Lett. 2011, 2, 1888−1891.
(59) O′Regan, B.; Moser, J.; Anderson, M.; Graetzel, M. J. Phys.
Chem. 1990, 94, 8720−8726.
(24) Clifford, J. N.; Palomares, E.; Nazeeruddin, M. K.; Gratzel, M.;
Nelson, J.; Li, X.; Long, N. J.; Durrant, J. R. J. Am. Chem. Soc. 2004,
126, 5225−5233.
(25) Argazzi, R.; Bignozzi, C. A.; Heimer, T. A.; Castellano, F. N.;
Meyer, G. J. J. Am. Chem. Soc. 1995, 117, 11815−11816.
(26) Bonhote, P.; Moser, J.; Humphry-Baker, R.; Vlachopoulos, N.;
Zakeeruddin, S. M.; Walder, L.; Gratzel, M. J. Am. Chem. Soc. 1999,
121, 1324−1336.
(27) Haque, S. A.; Handa, S.; Peter, K.; Palomares, E.; Thelakkat, M.;
Durrant, J. R. Angew. Chem., Int. Ed. 2005, 44, 5740−5744.
(28) Hu, K.; Robson, K. C. D.; Johansson, P. G.; Berlinguette, C. P.;
Meyer, G. J. J. Am. Chem. Soc. 2012, 134, 8352−8354.
(29) Abrahamsson, M.; Johansson, P. G.; Ardo, S.; Kopecky, A.;
Galoppini, E.; Meyer, G. J. J. Phys. Chem. Lett. 2011, 1, 1725−1728.
(30) Bublitz, G. U.; Boxer, S. G. Annu. Rev. Phys. Chem. 1997, 48,
213−242.
(31) Wang, P.; Zakeeruddin, S. M.; Moser, J. E.; Nazeeruddin, M. K.;
Sekiguchi, T.; Gratzel, M. Nat. Mater. 2003, 2, 402−407.
(32) Grosshenny, V.; Romero, F. M.; Ziessel, R. J. Org. Chem. 1997,
62, 1491−1500.
(33) Hofmeier, H.; El-ghayoury, A.; Schenning, A. P. H. J.; Schubert,
U. S. Tetrahedron 2004, 60, 6121−6128.
(34) Rajesh, K.; Somasundaram, M.; Saiganesh, R.; Balasubramanian,
K. K. J. Org. Chem. 2007, 72, 5867−5869.
(60) Haque, S. A.; Tachibana, Y.; Klug, D. R.; Durrant, J. R. J. Phys.
Chem. B 1998, 102, 1745−1749.
(61) Nelson, J. Phys. Rev. B 1999, 59, 15374−15380.
(62) Cao, F.; Oskam, G.; Meyer, G. J.; Searson, P. C. J. Phys. Chem.
1996, 100, 17021−17027.
(63) Huang, S. Y.; Schlichthorl, G.; Nozik, A. J.; Gratzel, M.; Frank,
A. J. J. Phys. Chem. B 1997, 101, 2576−2582.
(64) Peter, L. M.; Duffy, N. W.; Wang, R. L.; Wijayantha, K. G. U. J.
Electroanal. Chem. 2002, 524−525, 127−136.
(65) Vargas-Florencia, D.; Edvinsson, T.; Hagfeldt, A.; Furo, I. J.
Phys. Chem. C 2007, 111, 7605−7611.
(66) Kuciauskas, D.; Freund, M. S.; Gray, H. B.; Winkler, J. R.; Lewis,
N. S. J. Phys. Chem. B 2000, 105, 392−403.
(67) Hasselmann, G. M.; Meyer, G. J. J. Phys. Chem. B 1999, 103,
7671−7675.
(68) Boschloo, G.; Hagfeldt, A. Acc. Chem. Res. 2009, 42, 1819−1826.
(69) Rowley, J. G.; Farnum, B. H.; Ardo, S.; Meyer, G. J. J. Phys.
Chem. Lett. 2011, 1, 3132−3140.
(70) Gardner, J. M.; Abrahamsson, M.; Farnum, B. H.; Meyer, G. J. J.
Am. Chem. Soc. 2009, 131, 16206−16214.
(71) Clark, C. C.; Marton, A.; Srinivasan, R.; Narducci Sarjeant, A.
A.; Meyer, G. J. Inorg. Chem. 2006, 45, 4728−4734.
(72) Clark, C. C.; Meyer, G. J.; Wei, Q.; Galoppini, E. J. Phys. Chem.
B 2006, 110, 11044−11046.
(35) Lee, C.-H.; Galoppini, E. J. Org. Chem. 2010, 75, 3692−3704.
(36) Sprecher, M.; Breslow, R.; Uziel, O.; Link, T. M. Org. Prep.
Proced. Int. 1994, 26, 696−701.
(73) Li, X.; Reynal, A.; Barnes, P. R. F.; Humphry-Baker, R.;
Zakeeruddin, S. M.; De Angelis, F.; O′Regan, B. C. Phys. Chem. Chem.
Phys. 2012, 14, 15421−15428.
(74) Stanbury, D. M.; Sykes, A. G. In Advances in Inorganic Chemistry;
Academic Press: San Diego, 1989; Vol. 33, pp 69−138.
(75) Wilmarth, W. K.; Stanbury, D. M.; Byrd, J. E.; Po, H. N.; Chua,
C.-P. Coord. Chem. Rev. 1983, 51, 155−179.
(76) Tuikka, M.; Hirva, P.; Rissanen, K.; Korppi-Tommola, J.;
Haukka, M. Chem. Commun. 2011, 47, 4499−4501.
(77) Bisquert, J.; Fabregat-Santiago, F.; Mora-Sera, I. n.; Garcia-
Belmonte, G.; Barea, E. M.; Palomares, E. Inorg. Chim. Acta 2008, 361,
684−698.
(37) Jang, S.-R.; Yum, J.-H.; Klein, C.; Kim, K.-J.; Wagner, P.; Officer,
D.; Gratzel, M.; Nazeeruddin, M. K. J. Phys. Chem. C 2009, 113,
1998−2003.
(38) Klein, C.; Nazeeruddin, M. K.; Liska, P.; Di Censo, D.; Hirata,
N.; Palomares, E.; Durrant, J. R.; Gratzel, M. Inorg. Chem. 2004, 44,
178−180.
(39) Theodorou, V.; Skobridis, K.; Tzakos, A. G.; Ragoussis, V.
Tetrahedron Lett. 2007, 48, 8230−8233.
(40) Nazeeruddin, M. K.; Zakeeruddin, S. M.; Humphry-Baker, R.;
Jirousek, M.; Liska, P.; Vlachopoulos, N.; Shklover, V.; Fischer, C.-H.;
Gratzel, M. Inorg. Chem. 1999, 38, 6298−6305.
(41) Chen, Z.; Concepcion, J. J.; Luo, H.; Hull, J. F.; Paul, A.; Meyer,
T. J. J. Am. Chem. Soc. 2011, 132, 17670−17673.
(42) Pavlishchuk, V. V.; Addison, A. W. Inorg. Chim. Acta 2000, 298,
97−102.
(43) Morris, A. J.; Marton, A.; Meyer, G. J. Inorg. Chem. 2008, 47,
7681−7685.
(44) De Angelis, F.; Fantacci, S.; Selloni, A.; Nazeeruddin, M. K.
Chem. Phys. Lett. 2005, 415, 115−120.
(45) Wang, D.; Mendelsohn, R.; Galoppini, E.; Hoertz, P. G.;
Carlisle, R. A.; Meyer, G. J. J. Phys. Chem. B 2004, 108, 16642−16653.
(46) Morris, A. J.; Meyer, G. J. J. Phys. Chem. C 2008, 112, 18224−
18231.
(47) Ondersma, J., W.; Hamann, T. W. Energy Environ. Sci. 2012, 5,
9476−9480.
(48) Fabregat-Santiago, F.; Mora-Sero, I.; Garcia-Belmonte, G.;
Bisquert, J. J. Phys. Chem. B 2002, 107, 758−768.
(49) Langmuir, I. J. Am. Chem. Soc. 1918, 40, 1361−1403.
(50) Williams, G.; Watts, D. C. Trans. Faraday. Soc 1970, 66, 80−85.
(51) Rowley, J. G.; Meyer, G. J. J. Phys. Chem. C 2011, 115, 6156−
6161.
(52) Nazeeruddin, M. K.; Kay, A.; Rodicio, R.; Humphry-Baker, R.;
Muller, P.; Liska, P.; Vlachopoulos, N.; Gratzel, M. J. Am. Chem. Soc.
1993, 115, 6382−6390.
(53) Qu, P.; Meyer, G. J. Langmuir 2001, 17, 6720−6728.
(54) Ardo, S.; Sun, Y.; Staniszewski, A.; Castellano, F. N.; Meyer, G.
J. J. Am. Chem. Soc. 2011, 132, 6696−6709.
(55) Boxer, S. G. J. Phys. Chem. B 2009, 113, 2972−2983.
K
dx.doi.org/10.1021/ja402193f | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX