Increase in the coordination number of a cobalt porphyrin after photo-induced interfacial electron transfer into nanocrystalline TiO2

Article abstract of DOI:10.1021/ic301300h

Spectroscopic, electrochemical, and kinetic data provide compelling evidence for a coordination number increase initiated by interfacial electron transfer. Light excitation of Co^I(meso-5,10,15,20-tetrakis(4- carboxyphenyl)porphyrin) anchored to a nanocrystalline TiO₂ thin film, abbreviated Co^IP/TiO₂, immersed in an acetonitrile:pyridine electrolyte resulted in rapid excited state injection, k_inj > 10⁸ s^-1, to yield Co ^IIP/TiO₂(e^-), followed by axial coordination of pyridine to the Co^IIP and hence an increase in coordination number from four to five. The formal oxidation state and coordination environment of the Co metalloporphyrin on TiO₂ were assigned through comparative studies in fluid solution as well as by comparisons to previously reported data. The kinetics for pyridine coordination were successfully modeled with a pseudo-first order kinetic model that yielded a second-order rate constant of k_+py = 2 × 10⁸ M^-1 s^-1. Spectro-electrochemical measurements showed that pyridine coordination resulted in a ～200 mV negative shift in the Co^II/I reduction potential, E°(Co^II/I/TiO₂) = -0.72 V and E°(Co ^II/I(py)/TiO₂) = -0.85 V vs NHE. With some assumptions, this indicated an equilibrium formation constant K_f = 400 M ^-1 for the Co^IIP(py)/TiO₂ compound. The kinetics for charge recombination were non-exponential under all conditions studied, but were successfully modeled by the Kohlrausch-Williams-Watts (KWW) function with observed rate constants that decreased by about a factor of 100 when pyridine was present. The possible mechanisms for charge recombination are discussed.

Full text of DOI:10.1021/ic301300h

Article
pubs.acs.org/IC
Increase in the Coordination Number of a Cobalt Porphyrin after
Photo-Induced Interfacial Electron Transfer into Nanocrystalline TiO2
Darren Achey,† Shane Ardo,† and Gerald J. Meyer*,†,‡
‡
†Department of Chemistry and Department of Materials Science & Engineering, Johns Hopkins University, 3400 North Charles
Street, Baltimore, Maryland 21218, United States
S
* Supporting Information
ABSTRACT: Spectroscopic, electrochemical, and kinetic data
provide compelling evidence for a coordination number increase
initiated by interfacial electron transfer. Light excitation of CoI(meso-
5,10,15,20-tetrakis(4-carboxyphenyl)porphyrin) anchored to a nano-
crystalline TiO2 thin film, abbreviated CoIP/TiO2, immersed in an
acetonitrile:pyridine electrolyte resulted in rapid excited state
injection, kinj > 108 s−1, to yield CoIIP/TiO2(e−), followed by axial
coordination of pyridine to the CoIIP and hence an increase in
coordination number from four to five. The formal oxidation state
and coordination environment of the Co metalloporphyrin on TiO2
were assigned through comparative studies in fluid solution as well
as by comparisons to previously reported data. The kinetics for
pyridine coordination were successfully modeled with a pseudo-first
order kinetic model that yielded a second-order rate constant of k+py = 2 × 108 M−1 s−1. Spectro-electrochemical measurements
showed that pyridine coordination resulted in a ∼200 mV negative shift in the CoII/I reduction potential, E°(CoII/I/TiO2) =
−0.72 V and E°(CoII/I(py)/TiO2) = −0.85 V vs NHE. With some assumptions, this indicated an equilibrium formation constant
Kf = 400 M−1 for the CoIIP(py)/TiO2 compound. The kinetics for charge recombination were non-exponential under all
conditions studied, but were successfully modeled by the Kohlrausch−Williams−Watts (KWW) function with observed rate
constants that decreased by about a factor of 100 when pyridine was present. The possible mechanisms for charge recombination
are discussed.
most stable and widely utilized Ru polypyridyl sensitizers are
substitution-inert coordinatively saturated octahedral com-
pounds.2
INTRODUCTION
■
Electron transfer from a molecular excited state to a semi-
conductor material provides a means for conversion of the free
energy stored in the excited state into a charge separated state
composed of an oxidized molecule and a reduced semi-
conductor.1,2 An advantage of separating charges at conductive
interfaces is that one of the redox equivalents formed can
transport from the interface to an external location where
subsequent redox reactions are initiated. For example, in
regenerative dye-sensitized solar cells, an electron injected into
TiO2 transports to a Pt electrode where tri-iodide is reduced.3 In
what have been termed photoelectrosynthetic cells for water
oxidation, injected electrons reduce protons and generate
hydrogen gas at an external electrode.4,5
We recently reported excited-state electron injection into
mesoporous, nanocrystalline TiO2 films by cobalt meso-
5,10,15,20-tetrakis(4-carboxyphenyl) porphyrin (CoP) when in
the formal oxidation state of CoI.11 Well documented in the
Vitamin B12 (cobalamin) and cobalt porphyrin literature is the
tendency of Co to ligate a specific number of axial ligands in a
given oxidation state: CoIII has two axial ligands, CoII has one
axial ligand, and CoI has no axial ligands.12,13 An example of this
for pyridine ligation to a cobalt macrocycle is shown in Scheme 1.
With this background, it seemed likely that after CoIP*/TiO2 →
CoIIP/TiO2(e−) excited-state injection, the CoIIP generated
In older embodiments of photoelectrosynthetic cells, the
oxidized dye molecule itself was proposed to drive the
multielectron transfer reactions necessary for water oxidation.6,7
The mechanistic details that have recently emerged for molecular
water oxidation, and other multielectron transfer reactions
relevant to solar fuel generation, consistently invoke inner-sphere
activation with transition metal catalysts.8−10 Therefore, if one
attempts to use a single coordination compound to sensitize a
wide bandgap semiconductor to visible light and drive catalysis,
open coordination sites may be necessary. Unfortunately, the
Scheme 1. Typical Axial Ligation of Pyridine to a Co
Macrocycle in the Indicated Formal Oxidation States
Received: June 19, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ic301300h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
150 W xenon arc lamp served as the probe beam (Applied
Photophysics) that was aligned orthogonally to the laser excitation
and directed to a monochromator (Spex 1702/04) optically coupled to
an R928 photomultiplier tube (Hamamatsu). Transient data were
acquired on a computer-interfaced 350 MHz digital oscilloscope
(LeCroy 9450). In typical experiments, 30 laser pulses were averaged
at each monitoring wavelength, and full spectra were generated by
averaging multiple data points.
Electrochemistry. Electrochemical measurements utilized a
potentiostat (BAS model CV-50W or Epsilon electrochemical analyzer)
with a standard three electrode arrangement with a glassy carbon disk
(solution) or sensitized TiO2 thin film deposited on FTO glass (surface)
working electrode, a Pt gauze or Pt disk (Bioanalytical Scientific
Instruments, Inc.) counter electrode, and an aqueous Ag/AgCl (KCl
saturated) reference electrode. All potentials are reported vs NHE unless
otherwise noted. The ferrocenium/ferrocene (Fe(Cp)2+/0) half-wave
potential was measured both before and after experiment in a 100 mM
TBAClO4/acetonitrile electrolyte and was used as a standard to calibrate
the reference electrode. Conversion to vs NHE was achieved by
correcting for the expected ferrocenium/ferrocene E1/2 of +310 mV vs
the KCl-saturated aqueous calomel electrode (SCE), where SCE is +241
mV vs NHE.22
would ligate a fifth ligand. Here we present compelling evidence
that this does indeed occur. While light-induced coordination
number changes have been observed for excited and charge-
separated states in fluid solution, to our knowledge this report
represents the first example initiated by photoinduced interfacial
electron transfer.14−17 Whether this finding can one day be
exploited for photocatalysis remains to be demonstrated. More
fundamentally, inner-sphere ligation provides a general means by
which the thermodynamics and kinetics for the unwanted CoIIP/
TiO2(e−) → CoIP/TiO2 charge recombination reaction can be
controlled. In one extreme, ligand dissociation and charge
recombination occur in one concerted step such that the
enthalpic contribution of breaking the Co-ligand bond is
included within the total reorganization energy for electron
transfer. Indeed, it is shown herein that charge recombination is
significantly slowed by inner-sphere coordination of pyridine.
EXPERIMENTAL SECTION
■
Materials. The following reagents were used as received from the
indicated commercial suppliers: acetonitrile (Burdick and Jackson,
spectrophotometric grade); dimethylsulfoxide (DMSO; Fisher Scien-
tific, 99.9%); pyridine (py; Fisher Scientific, 99.9%); deionized water;
acetone (bulk solvent); tetra-n-butylammonium perchlorate (TBA-
ClO4; Fluka, >99.9%); tetra-n-butylammonium hydroxide (TBAOH;
Fluka 1 M aqueous); cobalt(III) meso-5,10,15,20-tetrakis(4-
carboxyphenyl)porphyrin chloride (CoP; Frontier Scientific >97%);
argon gas (Airgas, >99.998%); oxygen gas (Airgas, industrial grade);
titanium(IV)isopropoxide (Sigma-Aldrich, 97%); fluorine-doped SnO2-
coated glass (FTO; Hartford Glass Co., Inc., 2.3 mm thick, 15 Ω/□);
glass microscope slides (Fisher Scientific, 1 mm thick); and triethyl-
amine (Fisher Scientific, 99%).
The reduction potentials for CoP in fluid solution were determined
by differential pulse voltammetry as an average of the anodic and
cathodic peak potentials.
Spectroelectrochemistry was executed through simultaneous appli-
cation of an applied potential bias while monitoring the UV−vis
absorption spectra of TiO2 thin film electrodes in the standard
electrolytes. At each potential step, a spectrum that was invariant with
time was recorded. Single-wavelength absorption features plotted as a
function of potential bias were proportional to the cumulative
formation/loss of states; for the TiO2(e−) absorption features this was
directly related to the cumulative TiO2 density of states.
Reductive Photochemistry. Photochemical reduction of ∼5 μM
CoP was performed in argon-purged acetonitrile/DMSO (∼ 1% v/v of
DMSO) or acetonitrile/pyridine (∼ 15% v/v of pyridine) solutions
containing ∼50 μM [fac-Re(deeb)(CO)3(4-ethylpyridine)](OTf) and
50 mM triethylamine. A lens focused 150 W Xe arc lamp was utilized as
the photolysis source and absorption spectra were recorded
perpendicular to the light excitation.
[fac-Re(deeb)(CO)3(4-ethylpyridine)](OTf), where OTf− is triflate
anion and deeb is 4,4′ diethylester-2,2′ bipyridine, was prepared as
previously described.18,19
Preparations. Sensitized Metal-Oxide Thin Film Electrode.
Transparent TiO2 nanocrystallites (anatase, ∼15 nm diameter) were
prepared by hydrolysis of Ti(i-OPr)4 using a sol−gel technique
previously described in the literature.20 The sols were cast as
mesoporous thin films (∼ 5 μm thick), using Scotch tape as a spacer,
by doctor blading onto transparent FTO conductive substrates. The thin
films were annealed at 420 °C for 30 min under an atmosphere of O2
flow.
The films were pretreated with aqueous base (TBAOH, pH ∼ 11) for
10 min, followed by an acetonitrile wash, and were then immersed in a
μM CoP/DMSO solution. Within an hour, the films became brightly
colored and were washed thoroughly with 100 mM TBAClO4/
acetonitrile and placed diagonally in a standard 1 cm2 quartz cuvette
containing the electrolyte solution. The electrolyte solution were purged
with Ar(g) for at least 30 min prior to experimentation. The surface
coverage, Γ, in mol/cm2, was quantified from the measured absorption
data with the modified Beer−Lambert formula given in eq 1,
RESULTS
■
Steady-state illumination of [Re(CO)3(deeb)(4-ethyl
pyridine)](OTf), where OTf− is triflate anion and deeb is 4,4′-
diethylester-2,2′-bipyridine, in the presence of a sacrificial
electron donor, triethylamine (TEA), and CoIIIP, where P is
meso-5,10,15,20-tetrakis(4-carboxyphenyl)porphyrin, in an ace-
tonitrile:DMSO solution, 1:100 (v/v), resulted in spectral
changes associated with reduction of the CoIII metal center to
CoII, followed by further reduction to CoI, and finally observation
of the reduced Re complex, [Re(CO)3(deeb−)(4-ethyl pyr-
idine)]. These photochemical experiments were repeated in the
added presence of ∼2 M pyridine where the same sequence of
redox reactions was observed. Representative extinction
coefficient spectra for the metalloporphyrin in the indicated
oxidation states in the absence, CoP, and presence of 2 M
pyridine, CoP(py), are shown in Figure 1. Note that the CoP(py)
abbreviation refers only to the presence of pyridine and does not
imply a specific number of coordinated pyridines. The minor
visible absorption from [Re(CO)3(deeb)(4-ethyl pyridine)]-
(OTf) or [Re(CO)3(deeb−)(4-ethyl pyridine)] were subtracted
from the spectral data shown in Figure 1.
(1)
Abs = 1000 × Γ × ε
where ε is the molar decadic extinction (absorption) coefficient, M−1
cm−1, that was assumed to be unchanged whether in solution or on the
surface. The cobalt porphyrins were typically anchored to the films in
coverages of Γ = 1−4 × 10−9 mol/cm2. These values are equivalent to
approximately 10% of the saturation surface coverage.
Spectroscopy. Steady-state UV−visible absorption spectra were
obtained on either a Varian Cary 50 spectrophotometer or a Hewlett-
Packard 8453 Photodiode Array UV−vis Spectrophotometer at room
temperature. Nanosecond transient absorption were obtained on an
apparatus similar to that which has been previously described.21 Samples
were irradiated with 532 nm light from a frequency doubled Q-switched,
pulsed Nd:YAG laser (Quantel U.S.A. (BigSky) Brilliant B; 5−6 ns full
width at half-maximum, 1 Hz, ∼ 10 mm in diameter) directed 45° to the
FTO side of a TiO2 film. The excitation fluence was measured by a
thermopile power meter (Molectron) and was typically <2 mJ/cm2. A
The CoIIIP compound was anchored to a mesoporous,
nanocrystalline TiO2 (anatase) thin film supported on a fluorine
doped tin oxide (FTO) glass electrode, herein abbreviated CoP/
TiO2, and immersed in an argon-saturated 100 mM TBAClO4
acetonitrile solution. Typical surface coverages were 1−4 × 10−9
B
dx.doi.org/10.1021/ic301300h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
without (432 nm) pyridine as well as the sharper, more intense
absorption band at 410 nm observed when pyridine was present.
Figure 3 presents the absorbance monitored at 373 and 800
nm as a function of applied potential for a CoP/TiO2 thin film in
Figure 1. Extinction coefficient spectra of CoP in the indicated formal
oxidation states in the absence and presence of 2 M pyridine (py)
measured in argon-purged acetonitrile:DMSO 1:100 (v/v) solution
with 50 mM triethylamine.
Figure 3. Absorbance measured at 373 and 800 nm as a function of
applied potential for a CoP/TiO2 thin-film electrode immersed in 0.1 M
TBAClO4 in acetonitrile in the absence and presence of 360 mM
pyridine. The 373 nm data reported on the formation of CoIP/TiO2
while the absorbance at 800 nm was due to TiO2(e−)s.
mol/cm2. The visible absorption spectra of CoIIIP/TiO2
immersed in DMSO were, within experimental error, the same
as that for the compound dissolved in fluid solution. The CoP/
TiO2 thin film was used as the working electrode in a standard 3-
electrode electrochemical cell positioned in a 1 cm cuvette in a
UV−vis spectrophotometer. Forward bias raised the Fermi level
of the functionalized film, with spectral changes associated with
the reduction of the CoIII center to the formal oxidation state of
CoII, followed by the appearance of absorption features
characteristic of reduced TiO2, abbreviated as TiO2(e−) and
shown in Supporting Information, Figure S1. As the Fermi level
was raised further an increase in the TiO2(e−) concentration was
observed concurrent with the reduction of CoII/TiO2(e−) to
CoI/TiO2(e−). The same sequence of redox processes were
observed when 360 mM pyridine was present in the electrolyte
solution, herein abbreviated as CoP(py)/TiO2. Representative
absorption spectra of CoP/TiO2 and CoP(py)/TiO2 after
spectral features from the TiO2(e−)s were subtracted in the
indicated oxidation states are shown in Figure 2A. Figure 2B
shows absorption difference spectra measured after complete
conversion of CoIP/TiO2 to CoIIP/TiO2 in the presence (red)
and absence (black) of pyridine. Of note in Figure 2B were the
unique CoII−CoI isosbestic points observed with (426 nm) and
0.1 M TBAClO4/acetonitrile, both with (red and blue) and
without (black and green) 360 mM pyridine. An isosbestic point
for the appearance of TiO2(e−) was measured at 373 nm;
therefore, measurements at this wavelength reported only on the
reduction of CoIIP/TiO2 to CoIP/TiO2. The 800 nm absorption
was due solely to TiO2(e−)s.
The 800 nm absorption data were fit to an exponential
function while the 373 nm data were fit with eq 2,23
1
x =
°
app−E )/(a×59 mV)
1 + 10(E
(2)
where x is the fraction of the CoP in the CoI oxidation state, Eapp
is the applied potential, E0 is the formal reduction potential, and a
is the ideality factor that accounts for the non-Nernstian behavior
of the reduction. The ideality factor was 2.2 when the
measurements were made in the absence of pyridine, compared
to 1.6 in the presence of pyridine.
Figure 2. (A) Absorption spectra of a CoP/TiO2 thin-film in the indicated formal oxidation states measured in 100 mM TBAClO4/CH3CN electrolyte
with and without 360 mM pyridine (py). (B) The CoIIP/TiO2 absorption spectrum minus the CoIP/TiO2 absorption spectrum with noted isosbestic
points measured with (red) and without (black) 360 mM pyridine.
C
dx.doi.org/10.1021/ic301300h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
The derivatives of the CoII/I concentrations extracted from the
spectroelectrochemical data as well as the exponential distribu-
tion of the TiO2(e−)s both versus applied potential were plotted
as the chemical capacitance as shown in Figure 4.
CoIIP/TiO2 with respect to CH3CN was calculated based on the
II
spectroelectrochemical data in Table 1 and yielded log(KCo ‑py
)
of 2.6 0.3.
Shown in Figure 5 are transient absorption spectra measured
at the indicated delay times after pulsed 532 nm laser excitation
of a CoIP/TiO2 thin film immersed in 0.1 M TBAClO4/CH3CN
with 0, 0.03, and 2.5 M pyridine. When no pyridine was present
in the external solution, the spectra recorded at each delay time
were within experimental error the same when normalized. The
contribution due to TiO2(e−)s over this wavelength range was
very small and thus afforded realistic CoII/IP/TiO2 isosbestic
points. At the extreme of ∼2.5 M pyridine in Figure 5C, the
measured spectra were again unchanged when normalized, but
with an isosbestic point at 426 nm. At intermediate pyridine
concentrations (0.03 M), Figure 5B, an isosbestic point at 432
nm was initially observed that shifted to 426 nm at observation
times longer than 3 μs. Under these conditions, the spectra
observed at delay times less than 3 μs were time dependent.
The transient absorption decay monitored at 415 nm, the
approximate CoIIP/CoIIP(py) isosbestic point, was measured at
different pyridine concentrations with a constant applied bias of
−920 mV vs NHE, Figure 6A. At this observation wavelength,
spectral features due to charge recombination between the
injected electron and the oxidized cobalt compound, that is,
TiO2(e−) + CoIIP → TiO2 + CoIP, were observed without
contribution from pyridine ligation. The data were satisfactorily
fit to the Kohlrausch−Williams−Watts (KWW) function, eq
4,24,25
Figure 4. Chemical capacitance of a CoP/TiO2 thin-film electrode
immersed in 0.1 M TBAClO4/CH3CN in the absence and presence of
360 mM pyridine as a function of applied potential.
The CoP formal reduction potentials, ideality factors, and
absorbance data in the presence and absence of pyridine, both
anchored to TiO2 and in solution are summarized in Table 1. The
I(t) = I0 exp[−(kobst)β]
(4)
Table 1. CoII/IP Reduction Potentials and Ideality Factors, a
̀
where β is inversely related to the width of a Levy distribution of
sample
E°(CoII/I) V vs NHE
a
rate constants, and was fixed at 0.365. Increased pyridine
concentrations resulted in slower loss of the absorption change
associated with CoII. Shown in Figure 6B are absorption decays
monitored at 426 nm where two competitive processes were
expected: charge recombination and pyridine association to CoII.
The kinetic data was therefore fit to a kinetic model based on a
sum of a KWW function (with kobs and β fixed from the 415 nm
data) and a first-order reaction. Therefore, data monitored at 426
nm on these time scales reported on the coordination of pyridine
to CoII and was adequately described by a pseudo-first order
kinetic model at high pyridine concentrations. Plots of the
abstracted rate pseudo-first order rate constants measured as a
function of pyridine concentration from 4−20 mM were linear
and yield a second order rate constant of 2 1 × 108 M−1 s−1,
Figure 6B inset.
Shown in Figure 7 are plots of the observed rate constants
versus the 800 nm absorbance at the indicated pyridine
concentrations. The observed rate constants, kobs, were measured
as a function of applied bias, which was used to control the
concentration of TiO2(e−) as was quantified through their
absorption at 800 nm. The 800 nm absorbance was measured
directly for the extreme pyridine concentrations of 0 and 0.36 M,
intermediate values were approximated based on a logarithmic
dependence of the [TiO2(e−)] with pyridine concentration. The
absorbance at 800 nm was converted to the number of TiO2(e−)s
in a 1 cm2 geometric area by eq 1, using a molar decadic
extinction coefficient of 1000 M−1 cm−1.2
CoP/TiO2
−0.72
−0.85
−0.61
−0.83
2.2 0.1
1.6 0.1
i
CoP(py)/TiO2
CoP/DMSO
ii
CoP(py)
i
ii
Measured in acetonitrile with 0.36 M pyridine. Measured in neat
pyridine, 12.4 M.
solution E1/2 values were obtained via differential pulse
voltammetry in 0.1 M TBAClO4 solutions of both DMSO
(CoP is insoluble in acetonitrile) and pyridine. The CoII/IP/TiO2
reduction potentials in acetonitrile and acetonitrile/pyridine
solutions were taken to be the equilibrium potential at which
50% of the surface-bound molecules were reduced to the CoI
state in the spectroelectrochemical experiment shown in Figure
3.
The equilibrium constant for axial ligation of pyridine to CoIIP
measured relative to DMSO was determined by electrochemical
measurements as previously described.13 The use of DMSO was
required because of poor solubility of CoIIIP in CH3CN. The
half-wave potentials in DMSO and pyridine electrolyte,
E
1/2CoII/I(py) and E1/2CoII/I(DMSO), were measured by differential
pulse voltammetry. These potentials were used in conjunction
with eq 3
E
1/2CoII/I
= E1/2CoII/I
− 0.059 × log(KCo ‐py
)
II
(py)
(DMSO)
− 0.059 × log[py]
(3)
DISCUSSION
to calculate the equilibrium constant for pyridine ligation to
■
CoIIP, KCo ‑py. A value of ∼400 M−1 (log(KCo ‑py) = 2.6 0.3)
The results presented herein illustrate the first instance of an
excited state interfacial electron transfer event that initiates
II
II
was calculated. The equilibrium constant for pyridine ligation to
D
dx.doi.org/10.1021/ic301300h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 5. Transient absorption data collected at the indicated delay times following pulsed 532 nm laser excitation of a CoIP/TiO2 thin film immersed in
0.1 M TBAClO4/CH3CN with the indicated pyridine concentrations. For reference, the vertical dashed line indicates the 432 nm isosbestic point for
CoII to CoI reduction in the absence of pyridine.
Figure 6. (A) Absorption change monitored at 415 nm (indicative of charge recombination) under the same experimental conditions from Figure 5 with
0 to 0.3 M pyridine. A bias of −920 mV vs NHE was applied throughout the experiment, and the data were normalized to the maximum signal. Overlaid
on the data are kinetic fits to the KWW model. (B) Absorption change monitored at 426 nm (indicative of pyridine ligation) for a CoP/TiO2 thin film
with increasing pyridine concentrations at an applied bias of −830 mV vs NHE. Overlaid on the black data (0 M pyridine) in red is a fit to the KWW
model, the other three data have overlaid fits to a sum of a KWW and a first-order kinetic model. In both A and B the direction of the arrow indicates
increasing pyridine concentrations. The inset illustrates the observed first-order rate constants as a function of titrated pyridine concentration.
formation of a metal−ligand bond and hence an increase in the
coordination number. Light excitation of CoIP/TiO2 resulted in
rapid interfacial electron transfer to yield CoIIP/TiO2(e−),
followed by ligation of pyridine to the CoII metal center. Pyridine
ligation was found to alter the thermodynamics and kinetics for
electron transfer as is described further below.
Interfacial Thermodynamics. Spectroelectrochemistry was
utilized as an in situ tool to characterize the interfacial redox
thermodynamics of CoP/TiO2 with and without pyridine in the
external acetonitrile electrolyte that surrounded the sensitized
thin film. The cobalt sensitizer was anchored to TiO2 in the
formal oxidation state of CoIII. By raising the Fermi-level of the
sensitized TiO2 film toward the vacuum level with an external
E
dx.doi.org/10.1021/ic301300h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
absence and presence of pyridine, respectively. The direction and
magnitude of the pyridine-induced shift in the E°(CoII/I
)
reduction potential was consistent with literature reports.13
Spectroscopic titration data revealed significant spectral changes
indicative of pyridine ligation to CoIIP/TiO2, but no evidence for
pyridine coordination when the Co center was in the formal
oxidation state of CoI. Therefore, the shift in reduction potentials
can be attributed to the equilibrium shown in eq 5.
Keq
CoIIP/TiO2 + py HooI CoIIP(py)/TiO2
(5)
In the absence of pyridine, the acetonitrile solvent or a TiO2
surface site likely serves as the fifth ligand for CoIIP, and the true
equilibrium is therefore expected to be more complex than that
shown in eq 5. If one assumes that inductive effects due to this
nonpyridine ligation are small, the measured 130 mV shift in
formal reduction potentials corresponds to Keq > 102 M−1 for
pyridine ligation to CoII, a value in good agreement with previous
studies in CH2Cl2 electrolytes.13,27
Anodic or cathodic potential steps greater than 59 mV were
required to induce a factor of 10 change in the [CoII]:[CoI] ratio
when anchored to TiO2, behavior that necessitated the inclusion
of ideality factors to model this interfacial redox chemistry.23 An
ideality factor of 2.2, that is, about 130 mV/decade, was required
to accurately model the data in the absence of pyridine and 1.6,
95 mV/decade, in the presence of pyridine. The molecular
origins of nonideality at semiconductor interfaces remain
unknown and speculative.23,28 Nevertheless, this data demon-
strates that ideality factors can be controlled by coordination of
nonredox active molecules.
As the focus of this work was on electron transfer driven
changes in the cobalt coordination number, it would have been
convenient if pyridine had no influence on the TiO2 redox
chemistry; however, this was not found to be the case. Indeed
pyridine, particularly tert-butyl pyridine, is a common additive in
dye-sensitized solar cells that is known to increase open circuit
photovoltages.29−33 In LiI/I2 acetonitrile electrolytes, the
addition of tert-butyl pyridine to the electrolyte has been
Figure 7. Observed rate constants for charge recombination measured
at 415 nm after pulsed 532 nm laser excitation of a potentiostatically
controlled CoIP/TiO2 thin film in acetonitrile electrolyte with the
indicated pyridine concentration. The rate constant data are plotted
versus the absorbance at 800 nm, which is directly related to the
TiO2(e−) concentration.
bias, the reduction of CoIII to CoII was first observed followed by
the reduction of TiO2 itself, abbreviated TiO2(e−), and finally the
reduction of CoII to CoI. In all cases, the spectral changes that
accompanied this redox chemistry were reversible with
maintenance of isosbestic points when the applied bias was
reversed. The formal oxidation states of CoP were unambigu-
ously assigned by independent characterization of the metal-
loporphyrin in fluid solution and by comparisons with literature
data.26 The presence of pyridine in the external electrolyte,
CoP(py)/TiO2, did not change the order of this redox chemistry,
but did influence the potentials at which they occurred.
The reduction of CoIIP/TiO2 to CoIP/TiO2 was non-
Nernstian under all conditions studied. The equilibrium
potential at which the adjacent oxidation states were present in
equal concentration was taken as the formal reduction potential.
This was measured to be −720 mV and −850 mV vs NHE in the
Scheme 2. Summary of Interfacial Electron Transfer and Pyridine Coordination Kinetics
F
dx.doi.org/10.1021/ic301300h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
shown to induce a negative potential shift in the TiO2 density of
states.34,35 Under the conditions reported here with TBAClO4/
CH3CN, the shift was in the same direction but smaller. Lewis
base coordination to other semiconductor materials has also
been shown to have a similar effect and, in general, is expected to
lower the work function of solids.36 Sigma donation from the
pyridine lone pair would inductively be expected to increase the
TiIV/III reduction potential of acceptor states relative to a weaker
Lewis base like CH3CN. As the Ti centers in TiO2 have a d0
electronic configuration, π-backbonding is not expected to be
significant in the fully reduced form although it may become
more relevant after excited state injection and/or a forward bias
generates Ti(III) states. An alternative explanation for the
pyridine-induced shift is that pyridine acts as a Brønsted base that
deprotonates surface titanol groups that gives rise to the
Nernstian band edge shifts commonly reported for metal-oxide
semiconductors with pH in aqueous solution.37 Indeed, the TiO2
thin films used herein were pretreated with aqueous base and
studied in acetonitrile electrolytes without intentionally added
“potential determining cations”, such as Li+ and H+, so that CoIP
could be generated with only a small concentration of
TiO2(e−)s.2
Establishment of the pyridine equilibrium and rate constant
affords calculation of the rate constant for pyridine dissociation
from CoIIP(py)/TiO2(e−) to be ∼5 × 105 s−1. Therefore, on
average pyridine deligates from CoIIP(py)/TiO2 every 1.4 μs and
remains dissociated for 1 μs and 10 ns at 0.003 and 0.300 M
pyridine concentrations, respectively. As a result, charge
recombination that was observed to occur on a much slower
millisecond time scale could indeed occur by a mechanism where
pyridine dissociation occurred prior to back electron transfer.
The derived rate law for such a mechanism is inversely correlated
with the concentration of pyridine as was observed exper-
imentally. Consequently, a stepwise mechanism where ligand
dissociation occurs prior to electron transfer is most consistent
with the data. We note that concerted electron transfer
mechanisms have previously been invoked by Saveant and co-
workers in cobalamin chemistry for dissociation of cobalt−
carbon bonds, behavior that cannot be fully ruled for the Co-py
bond breaking that accompanies the interfacial electron transfer
data reported herein.40
CONCLUSIONS
■
The first instance of photoinduced interfacial electron transfer
incurring a change in coordination number was observed and
quantified. The coordination of pyridine to the CoIIP/TiO2(e−)
charge-separated state influenced both the kinetics and the
thermodynamics for interfacial charge recombination. A factor of
100 decrease in the observed rate constant for charge
recombination was quantified when pyridine was present. The
charge recombination mechanism was not fully established, but
the data were consistent with ligand dissociation followed by
electron transfer.
Kinetics and Mechanisms. Pulsed-light excitation of CoIP/
TiO2 resulted in spectral changes consistent with rapid excited
state injection, kinj > 108 s−1, to yield CoIIP/TiO2(e−). In the
absence of pyridine, this was the only product observed
spectroscopically. At high pyridine concentrations, a single
product was again observed with spectroscopic features assigned
to CoIIP(py)/TiO2(e−) indicating that both excited state
injection and pyridine coordination occurred within 10 ns, the
response time of our apparatus. When pyridine was present at
intermediate concentrations, a time dependent spectral change
assigned to pyridine coordination to CoII was observed that
allowed the kinetics for CoII coordination chemistry to be
quantified. A second-order rate constant of 2 1 × 108 M−1 s−1
for pyridine coordination CoIIP/TiO2(e−) was abstracted from
kinetic data measured as a function of pyridine concentration and
was consistent with that reported in solution, Scheme 2.27,38
The ∼130 mV negative shift of E°(CoII/I) that accompanied
pyridine coordination to Co(II) decreased the driving force for
TiO2(e−) + CoIIP(py) charge recombination by a corresponding
value. We note that pyridine altered the reduction of TiO2 itself,
however this influence was small relative to the change in the
metalloporphyrin reduction potential. Pyridine ligation de-
creased significantly the observed charge recombination rate
constant. In making this comparison, care was taken to ensure
that the TiO2(e−) concentration was the same as it is now well
established that recombination rates at sensitized TiO2 interfaces
are highly sensitive to the concentration of trapped electrons.2,11
Indeed, charge recombination was on average 100 times faster
when pyridine was absent over a factor of 3 change in TiO2(e−)
concentration, presumably because of the less favorable free
energy change with electron transfer in the normal kinetic region.
Three mechanisms for CoIIP(py)/TiO2(e−) charge recombi-
nation can be envisioned: (1) a concerted process wherein
electron transfer and CoII-py bond cleavage occur in one step;
(2) electron transfer followed by CoI-py bond breaking; and (3)
CoII-py bond cleavage followed by interfacial electron transfer.
The second mechanism could have been unambiguously
identified by the appearance of a CoI-py intermediate. Despite
many attempts, no evidence for such intermediates were
garnered. Density functional calculations of cobalamin indicate
that a 5-coordinate CoI species is energetically unfavored.39
ASSOCIATED CONTENT
■
S
* Supporting Information
Absorption spectra of TiO2 thin films with and without
TiO2(e−)s. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: meyer@jhu.edu.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The National Science Foundation is gratefully acknowledged for
research support.
REFERENCES
■
(1) Hagfeldt, A.; Gratzel, M. Acc. Chem. Res. 2000, 33, 269−277.
̈
(2) Ardo, S.; Meyer, G. J. Chem. Soc. Rev. 2009, 38, 115−164.
(3) O’Regan, B.; Gratzel, M. Nature 1991, 353, 737−740.
̈
(4) 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.
(5) Brimblecombe, R.; Koo, A.; Dismukes, G. C.; Swiegers, G. F.;
Spiccia, L. J. Am. Chem. Soc. 2010, 132, 2892−2894.
(6) Gleria, M.; Memming, R. Z. Phys. Chem. 1975, 98, 303−316.
(7) Anderson, S.; Constable, E. C.; Dare-Edwards, M. P.; Goodenough,
J. B.; Hamnett, A.; Seddon, K. R.; Wright, R. D. Nature 1979, 280, 571−
573.
G
dx.doi.org/10.1021/ic301300h | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(8) Gilbert, J. A.; Eggleston, D. S.; Murphy, W. R., Jr.; Geselowitz, D.
A.; Gersten, S. W.; Hodgson, D. J.; Meyer, T. J. J. Am. Chem. Soc. 1985,
107, 3855−3864.
(9) Hurst, J. K.; Cape, J. L.; Clark, A. E.; Das, S.; Qin, C. Inorg. Chem.
2008, 47, 1753−1764.
(10) Wasylenko, D. J.; Ganesamoorthy, C.; Henderson, M. A.;
Koivisto, B. D.; Osthoff, H. D.; Berlinguette, C. P. J. Am. Chem. Soc.
2010, 132, 16094−16106.
(11) Achey, D.; Ardo, S.; Xia, H.-L.; Siegler, M. A.; Meyer, G. J. J. Phys.
Chem Lett. 2011, 2, 305−308.
(12) Lexa, D.; Saveant, J. M. Acc. Chem. Res. 1983, 16, 235−243.
(13) Kadish, K.; Bottomley, L. A.; Beroiz, D. Inorg. Chem. 1978, 17,
1124−1129.
(14) Blaskie, M. W.; McMillin, D. R. Inorg. Chem. 1980, 19, 3519−
3522.
(15) Chen, L. X.; Jennings, G.; Liu, T.; Gosztola, D. J.; Hessler, J. P.;
Scaltrito, D. V.; Meyer, G. J. J. Am. Chem. Soc. 2002, 124, 10861−10867.
(16) Ruthkosky, M.; Kelly, C. A.; Zaros, M. C.; Meyer, G. J. J. Am.
Chem. Soc. 1997, 119, 12004−12005.
(17) Scaltrito, D. V.; Kelly, C. A.; Ruthkosky, M.; Zaros, M. C.; Meyer,
G. J. Inorg. Chem. 2000, 39, 3765−3770.
(18) Hino, J. K.; Ciana, L. D.; Dressick, W. J.; Sullivan, B. P. Inorg.
Chem. 1992, 31, 1072−1080.
(19) Hasselmann, G. M.; Meyer, G. J. J. Phys. Chem B 1999, 103,
7671−7675.
(20) Heimer, T. A.; D’Arcangelis, S. T.; Farzad, F.; Stipkala, J. M.;
Meyer, G. J. Inorg. Chem. 1996, 35, 5319−5324.
(21) Argazzi, R.; Bignozzi, C. A.; Heimer, T. A.; Castellano, F. N.;
Meyer, G. J. Inorg. Chem. 1994, 33, 5741−5749.
(22) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals
and Applications, 2nd ed.; Wiley: New York, 2001.
(23) Ardo, S.; Achey, D.; Morris, A. J.; Abrahamsson, M.; Meyer, G. J. J.
Am. Chem. Soc. 2011, 133, 16572−16580.
(24) Williams, G.; Watts, D. C. Trans. Faraday Soc. 1970, 66, 80−85.
(25) Lindsey, C. P.; Patterson, G. D. J. Chem. Phys. 1980, 73, 3348−
3357.
(26) D’Souza, F.; Villard, A.; Van Caemelbecke, E.; Franzen, M.;
Boschi, T.; Tagliatesta, P.; Kadish, K. M. Inorg. Chem. 1993, 32, 4042−
4048.
(27) Tetreau, C.; Lavalette, D.; Momenteau, M. J. Am. Chem. Soc. 1983,
105, 1506−1509.
(28) Zaban, A.; Ferrere, S.; Gregg, B. A. J. Phys. Chem. B 1998, 102,
452−460.
(29) Nazeeruddin, M. K.; Kay, A.; Rodicio, I.; Humphry-Baker, R.;
Mueller, E.; Liska, P.; Vlachopoulos, N.; Gratzel, M. J. Am. Chem. Soc.
̈
1993, 115, 6382−6390.
(30) Gothelid, M.; Yu, S.; Ahmadi, S.; Sun, C.; Zuleta, M. Int. J.
Photoenergy 2011, 2011.
(31) Yu, S.; Ahmadi, S.; Sun, C.; Palmgren, P.; Hennies, F.; Zuleta, M.;
Gothelid, M. J. Phys. Chem. C 2010, 114, 2315−2320.
(32) Shi, C.; Dai, S.; Wang, K.; Pan, X.; Kong, F.; Hu, L. Vib. Spectrosc.
2005, 39, 99−105.
(33) Xiong, Y.; Weiwei, T.; Jingbo, Z.; Yuan, L.; Xurui, X.; Xiaowen, Z.;
Xueping, L.; Sankapal, B. R. J. App. Electrochem. 2009, 39, 147−154.
(34) Hara, K.; Dan-oh, Y.; Kasada, C.; Ohga, Y.; Shinpo, A.; Suga, S.;
Sayama, K.; Arakawa, H. Langmuir 2004, 20, 4205−4210.
(35) Boschloo, G.; Haggman, L.; Hagfeldt, A. J. Phys. Chem. B 2006,
110, 13144−13150.
(36) Stair, P. C. J. Am. Chem. Soc. 1982, 104, 4044−4052.
(37) Finklea, H. O. Semiconductor Electrodes; Elsevier: New York, 1988.
(38) Gouterman, M.; Tait, C. D.; Holten, D. J. Am. Chem. Soc. 1984,
106, 6653−6659.
(39) Brunold, T. C.; Liptak, M. D. J. Am. Chem. Soc. 2008, 128, 9144−
9156.
(40) Costentin, C.; Robert, M.; Saveant, J.-M. J. Am. Chem. Soc. 2005,
127, 12154−12155.
H
dx.doi.org/10.1021/ic301300h | Inorg. Chem. XXXX, XXX, XXX−XXX