Visible light-induced hole injection into rectifying molecular wires anchored on Co3O4 and SiO2 nanoparticles

Article abstract of DOI:10.1021/ja306162g

Tight control of charge transport from a visible light sensitizer to a metal oxide nanoparticle catalyst for water oxidation is a critical requirement for developing efficient artificial photosynthetic systems. By utilizing covalently anchored molecular w

Full text of DOI:10.1021/ja306162g

Article
pubs.acs.org/JACS
Visible Light-Induced Hole Injection into Rectifying Molecular Wires
Anchored on Co3O4 and SiO2 Nanoparticles
Han Sen Soo, Anil Agiral, Andreas Bachmeier,† and Heinz Frei*
Physical Biosciences Division, Lawrence Berkeley National Laboratory, University of California, Berkeley, California 94720, United
States
S
* Supporting Information
ABSTRACT: Tight control of charge transport from a visible
light sensitizer to a metal oxide nanoparticle catalyst for water
oxidation is a critical requirement for developing efficient
artificial photosynthetic systems. By utilizing covalently
anchored molecular wires for hole transport from sensitizer
to the oxide surface, the challenge of high rate and
unidirectionality of the charge flow can be addressed.
Functionalized hole conducting molecular wires of type p-
oligo(phenylenevinylene) (3 aryl units, abbreviated PV3) with
various anchoring groups for the covalent attachment to
Co3O4 catalyst nanoparticles were synthesized and two
alternative methods for attachment to the oxide nanoparticle surface introduced. Covalent anchoring of intact PV3 molecules
on Co3O4 nanoparticles (and on SiO2 nanoparticles for control purposes) was established by FT-Raman, FT-IR, and optical
spectroscopy including observation, in some cases, of the vibrational signature of the anchored functionality. Direct monitoring of
the kinetics of hole transfer from a visible light sensitizer in aqueous solution ([Ru(bpy)3]2+ (and derivatives) light absorber,
[Co(NH3)5Cl]2+ acceptor) to wire molecules on inert SiO2(12 nm) particles by nanosecond laser absorption spectroscopy
revealed efficient, encounter controlled rates. For wire molecules anchored on Co3O4 nanoparticles, the recovery of the reduced
sensitizer at 470 nm indicated similarly efficient hole transfer to the attached PV3, yet no transient hole signal was detected at
600 nm. This implies hole injection from the anchored wire molecule into the Co3O4 particle within 1 μs or shorter, indicating
efficient charge transport from the visible light sensitizer to the oxide catalyst particle.
important for the case of multistep CO2 reduction, which may
involve 4, 6, or even 8 electron transfer events. Moreover,
coupling of a nanoparticulate water oxidation catalyst with the
visible light chromophore by molecular wires might overcome
the inefficient hole transfer between catalyst and light absorber
commonly observed thus far for metal oxide catalysts.4,5
1. INTRODUCTION
For developing an efficient artificial photosynthetic system, a
critical requirement is the delivery of charge from the light
absorber to multielectron catalysts for water oxidation or fuel
production with high quantum yield at low overpotential. Tight
control of charge transport is favored for linkages that are
molecularly defined and allow flow of electrons (or holes) only
in the desired direction, while blocking transfer in the reverse
sense. Such stringent control may be provided by molecular
wires, especially the ones that impose directionality on the
charge flow. We are exploring a method for coupling Co3O4
nanoparticles with a visible light chromophore through a hole
conducting molecular wire embedded in a dense phase, few
nanometers thin, silica shell that separates the catalyst from the
chromophore. We have recently shown that the Co3O4
nanoparticles act as efficient multielectron catalysts for water
oxidation at close to neutral pH.1,2 Amorphous silica layers of
nanometer depth are known to provide facile transport for
protons while blocking passage of molecular products.3 Using
proper geometries, this approach could open up constructs for
integrated systems that afford closure of the photosynthetic
cycle on the nanoscale under separation of the water oxidation
catalysis from the reduction chemistry. Spatial separation of the
two chemistries by a product-impermeable layer is particularly
A class of well-established molecular wires for hole
conduction are p-oligo(phenylenevinylene) molecules, which
exhibit a very small rate dependence on distance for properly
matched potentials of the attached components.6,7 For the
three membered representative 1,3-di((E)-styryl)benzene (ab-
breviated PV3, referring to 3 aryl units), the highest occupied
molecular orbital (HOMO) has sufficient potential (E0 = 1.32
V vs SHE)7,8 to drive a Co3O4 nanoparticle catalyst for water
oxidation. At the same time, hole injection into this wire
molecule is energetically accessible by mild donors such as Ru
bipyridyl-type visible light sensitizers.9 With the PV3 LUMO
situated at −1.7 V (vs SHE), electrons from the visible light-
excited sensitizer cannot be accepted by the wire, thus,
imposing unidirectional charge flow. Efficient transfer of charge
from wire molecules to the nanoparticle catalyst may require
Received: July 2, 2012
Published: August 30, 2012
© 2012 American Chemical Society
17104
dx.doi.org/10.1021/ja306162g | J. Am. Chem. Soc. 2012, 134, 17104−17116

Journal of the American Chemical Society
Article
covalent anchoring on the metal oxide surface. While covalent
charge transport linkers to Co oxide surfaces have not been
explored so far to our knowledge, functionalities such as
carboxylate, phosphonate, ether, or acetyl acetonate commonly
used for anchoring of visible light sensitizers on TiO2 in dye
sensitized solar cells offer a starting point.4,10−14 Moreover,
photochemical grafting of terminal vinyl groups results in stable
attachment to TiO2 surfaces as well.15 For achieving orientation
of the wire molecules vertical to the nanoparticle surface,
polydentate and tripodal anchors with terminal OH groups are
of particular interest.11,16−20
In this paper, we report on the synthesis, characterization,
and attachment of molecular wires on Co3O4 catalyst surfaces,
and on the demonstration of visible light-sensitized hole
transfer to the anchored wire molecules. The syntheses and
characterization of 1,3-di((E)-styryl)benzene molecules featur-
ing several different functionalities for covalent anchoring on
oxide nanoparticle surfaces are presented. Optical, FT-infrared,
and FT-Raman spectroscopy are employed for monitoring
covalent attachment to the oxide particle surface. Transient
optical measurements on the nano- and microsecond time scale
provide direct evidence for efficient hole injection from a Ru
bipyridyl-type visible light sensitizer into molecular wires
anchored on Co3O4 and SiO2 nanoparticles. Embedding of
the anchored wire molecules into an amorphous, nanoscale
silica layer, and core−shell constructs with proper geometry
(e.g., nanotubes) for the development of an integrated system
that affords water oxidation catalysis separated from chromo-
phore and reduction chemistry will be reported in forthcoming
papers.
Figure 1. Core−shell construct with tripodally anchored, hole-
conducting PV3 molecular diodes embedded in a proton-conducting
silica shell.
2. RESULTS
2.1. Design, Synthesis, and Characterization of p-
Oligo(phenylenevinylene) Molecular Wires. The PV3
derivatives were assembled by the trans-selective Wittig−
Horner reaction21−23 after incorporating the desired hydro-
philic24 and anchoring functional groups in the precursors. A
perpendicular (radial) orientation of the molecular wires with
respect to the oxide particle surface is required in order to
achieve spatial separation of the water oxidation from
chromophore/carbon dioxide reduction catalytic sites by a
few nm and minimize through-space back charge transfer.
Furthermore, radial orientation prevents aggregation and
intermolecular π−π interactions, thus will enable subsequent
casting of a gas separating but proton-conducting silica layer
around the wire molecules (schematically shown in Figure 1).
To minimize perturbation of the electronic structure of the
conjugated π-electron system, the anchoring groups for the
molecular wires were introduced only at the para position of
one terminal benzene ring of the PV3 molecules. For the PV3
molecules to be water-soluble, sulfonate groups were
incorporated into the benzene ring at the other end of the
PV3 molecular wires.22 In addition to facilitating water
solubility, sulfonate groups are poor conjugate bases and
nucleophiles, and therefore unlikely to covalently attach to
metal oxide surfaces, which might otherwise result in cross-
linking and aggregation of the functionalized nanoparticles.
For the determination of hole injection efficiency from a
visible light sensitizer into wire molecules in homogeneous
aqueous solution, the water-soluble sodium 2,2′-((1E,1′E)-1,4-
phenylenebis(ethene-2,1-diyl))dibenzenesulfonate
(Na2PV3_SO3_SO3) was prepared in good yields by a
straightforward one-step Wittig−Horner reaction between
sodium 2-formylbenzenesulfonate (1) and tetraethyl (1,4-
phenylenebis(methylene))bis(phosphonate) (2) (Scheme 1).
Asymmetrically functionalized oligo(phenylenevinylene)
wires were synthesized in a convergent and modular fashion.
The routes to the mono para-sulfonated PV3 oligomer ethyl 4-
((E)-4-((E)-4-cyanostyryl)styryl)benzenesulfonate
(PV3_CN_SO2OEt) and its more hydrophilic derivatives are
illustrated in Schemes 2 and 3. The nitrile precursor (E)-4-(4-
(diethoxymethyl)styryl)benzonitrile (PV2_CN_CH(OEt)2) is
isolated with high trans regioselectivity and high yields from the
Wittig−Horner reaction between 4-(diethoxymethyl)-
benzaldehyde (3) and diethyl 4-cyanobenzylphosphonate (4).
The nitrile group is versatile and can be readily transformed
into anchors such as carboxylates, aldehydes, alcohols, amines,
or further converted into chelating groups. The acetal can be
hydrolyzed under mild acidic conditions with aqueous acetic
acid in a biphasic reaction to provide (E)-4-(4-formylstyryl)-
benzonitrile (PV2_CN_C(O)H), with another formyl group
for a subsequent Wittig−Horner reaction with an appropriately
derivatized hydrophilic aryl phosphonate. Methyl 4-
(bromomethyl)benzenesulfonate (5) is converted via the
Arbuzov reaction with triethyl phosphite (6) into ethyl 4-
((diethoxyphosphoryl)methyl)benzenesulfonate (7) in quanti-
tative yields.22 Subsequent reaction between PV2_CN_C(O)H
and 7 in the presence of a strong base gave PV3_CN_SO2-
(OEt) as a yellow solid in good yields without the need for
column chromatography in this entire reaction sequence.
Attempts at basic hydrolysis of PV3_CN_SO2(OEt) with
aqueous NaOH or KOH yielded a yellow solid material that is
poorly soluble in water. Hydrolysis at elevated temperatures
and pressures in an autoclave with CsOH produced cesium 4-
17105
dx.doi.org/10.1021/ja306162g | J. Am. Chem. Soc. 2012, 134, 17104−17116

Journal of the American Chemical Society
Article
Scheme 1. Preparation of the Water-Soluble Na2PV3_SO3_SO3 by the Wittig−Horner Reaction
Scheme 2. Convergent Synthesis of PV3_CN_SO2(OEt)
Scheme 3. Hydrophilic Functionalization of PV3_CN_SO2(OEt) by Hydrolysis and Conversion into an Amide
( (E ) - 4 - ( ( E ) - 4 - s u l f o n a t o s t y r y l ) s t y r y l ) b e n z o a t e
(Cs2PV3_CO2_SO3, Scheme 3), which is still only sparingly
soluble in water and dimethyl sulfoxide (DMSO), but sufficient
(hydroxymethyl)propan-2-yl)carbamoyl)-styryl)styryl)-
benzenesulfonate (CsPV3_tripod_SO3) has only slightly
enhanced aqueous solubility, prompting further synthetic
modifications to the hydrophilic component for coupling with
PV2_CN_C(O)H.
1
amounts dissolved for characterization by H NMR spectros-
copy. In accordance to observations in previous studies,22 para-
sulfonated PV3 molecules have poor solubility in water,
presumably due to the hydrophobic π-stacking interactions.
In an attempt to disrupt the strong intermolecular forces of the
PV3 molecules and install three anchor points in a tripodal
perpendicular arrangement for the molecular wires on metal
oxide surfaces, Cs2PV3_CO2_SO3 is condensed with tris-
(hydroxymethyl)aminomethane (8) using O-(6-chlorobenzo-
triazol-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophos-
phate (HCTU) as the peptide coupling agent.25,26 However,
the isolated yellow cesium 4-((E)-4-((E)-4-((1,3-dihydroxy-2-
On the basis of the reasonable water solubility of
Na2PV3_SO3_SO3, we surmised that ortho substitution on
PV3 molecules may minimize the π-stacking interactions. The
derivative Cs3PV3_CO2_2SO3 was thus designed with this
insight in mind and the assembly is presented in Scheme 4.
Hydrogenation of sodium 4-formylbenzene-1,3-disulfonate (9)
with NP Pd(0)EnCat 3027 as the catalyst provides sodium 4-
(hydroxymethyl)benzene-1,3-disulfonate (10) in good yields
without need for further purification. The use of 10% Pd/C
with various additives such as ethylenediamine28 in different
17106
dx.doi.org/10.1021/ja306162g | J. Am. Chem. Soc. 2012, 134, 17104−17116

Journal of the American Chemical Society
Article
Scheme 4. Convergent Preparative Route to the Water-Soluble Cs3PV3_CO2_2SO3
Scheme 5. Transformation of Cs3PV3_CO2_2SO3 into Chelating, Poly-ol Anchors by Amide Coupling
solvents invariably leads to a mixture of unreacted starting
material, product, and the over-reduced byproduct 4-methyl-
benzene-1,3-disulfonate (11), which cannot be readily sepa-
rated without reverse-phase column chromatography. Chlori-
nation of finely crushed 10 with SOCl2 and catalytic amounts of
N,N-dimethylformamide (DMF) to produce 4-(chloromethyl)-
benzene-1,3-disulfonyl chloride (12),29 followed by methyl-
ation using sodium methoxide (13), yields dimethyl 4-
(chloromethyl)benzene-1,3-disulfonate (14) as a viscous,
orange oil.30,31 Subsequent Arbuzov reaction22 between 6 and
14 led to an unexpectedly complex mixture, from which diethyl
4-((diethoxyphosphoryl)methyl)benzene-1,3-disulfonate (15)
is isolated as a colorless oil after purification by column
chromatography on silica. A Wittig−Horner reaction between
PV2_CN_C(O)H and 15 also resulted in a mixture of
products, and diethyl 4-((E)-4-((E)-4-cyanostyryl)styryl)-
benzene-1,3-disulfonate (PV3_CN_2SO2(OEt)) was purified
by silica column chromatography and collected in moderate
yields. Hydrolysis of PV3_CN_2SO2(OEt) at elevated
temperatures and pressures in an autoclave with a concentrated
aqueous solution of CsOH produced cesium 4-((E)-4-((E)-2,4-
disulfonatostyryl)styryl)benzoate (Cs3PV3_CO2_2SO3). The
compound is readily soluble in water and somewhat soluble in
methanol.
dihydroxy-2-(hydroxymethyl)propan-2-yl)carbamoyl)-styryl)-
styryl)benzene-1,3-disulfonate (Cs3PV3_tripod_2SO3) were
synthesized. These congeners have two and three alcohol
groups for condensation with surface M−OH (M = Si, Co)
groups, and were prepared in similar fashion to the
monosulfonated version above with HCTU as the dehydrating
agent (Scheme 5).25,26 All precursors were characterized by 1H
and 13C NMR spectroscopy and high resolution mass
spectrometry. In addition, the PV2 and PV3 oligomers were
further studied using FT-IR, FT-Raman, and UV−visible
spectroscopy because these techniques offer unique features
of the molecules that are essential for verifying the covalent
anchoring of the PV3 molecular wires on the metal oxide
surfaces.
2.2. Transient Absorption Spectroscopy of Hole
Injection into Free Molecular Wires in Aqueous
Solution. The ground state optical absorption spectra of
representative examples of the PV2 and PV3 molecular wires
measured in chloroform or in water are displayed in Figure 2,
while the absorption maxima (λmax) and extinction coefficients
(ε) are summarized in Table 1. All spectra exhibit at least one
intense absorption band between 320 and 380 nm, with
extinction coefficients equal or exceeding 29 000 M−1 cm−1,
attributable to an aromatic π−π* transition.32 The observed red
shift of the π−π* band from PV2_CN_C(O)H to
PV3_CN_2SO2(OEt) is consistent with the extended
conjugation of the PV3 congeners compared to the PV2
precursors. The absence of a substantial solvent shift of the PV3
To achieve covalent attachment to metal oxide surfaces (vide
infra), the two derivatives cesium 4-((E)-4-((E)-4-(bis(2-
hydroxyethyl)carbamoyl)styryl)styryl)benzene-1,3-disulfonate
(Cs2PV3_O3N_2SO3) and cesium 4-((E)-4-((E)-4-((1,3-
17107
dx.doi.org/10.1021/ja306162g | J. Am. Chem. Soc. 2012, 134, 17104−17116

Journal of the American Chemical Society
Article
molecules in aqueous solution. The transient optical experi-
ments were carried out with [Ru(bpy)3]2+ or [Ru(bpy-
(CO2Me)2)3]2+ (bpy = 2,2′-bipyridine; bpy(CO2Me)2 = 2,2′-
bipyridine-4,4′-dimethylcarboxylate) as the visible light sensi-
tizers and Na2S2O8 or Co(NH3)5Cl3 as the sacrificial electron
acceptor. These are established visible light sensitization
systems35 in which a RuIII complex is generated at a potential
suitable for hole injection into the wire molecules. The
energetics and kinetics of the electron transfer processes
leading to the oxidized Ru complex have been reported.7,36 The
standard redox potentials of these reagents and PV3 (1,3-
di((E)-styryl)benzene) are listed in Table 2 (relative to NHE in
Table 2. Reduction Potentials of Reagents Used in Time-
Resolved Spectroscopic Studies
a
reagent
E1/2 (V)
b
b
PV3+/PV3
PV3/PV3−
1.327,41
−1.697,41
1.4539−41
−1.1739−41
0.8439−41
−0.8439−41
1.8041,42
−0.6641,42
1.9644,45
2.4344,45
0.3343
Ru(bpy)33+/ Ru(bpy)3
Ru(bpy)32+/ Ru(bpy)3
Ru(bpy)32+*/ Ru(bpy)3
Ru(bpy)33+/ Ru(bpy)3
Ru(bpy[CO2Me])33+/ Ru(bpy[CO2Me])3
Ru(bpy[CO2Me])32+/ Ru(bpy[CO2Me])3
2+
+
+
2+
Figure 2. (a) UV−visible spectra of 10 μM PV2_CN_C(O)H (black)
and 25 μM PV3_CN_2SO2(OEt) (red) in CHCl3. (b) UV−visible
spectra of 12.5 μM Cs2PV3_tripod_2SO3 (blue) and 25 μM
Na2PV3_SO3_SO3 (green) in water. All spectra were collected in
10 × 5 mm cuvettes.
*
c
2+
c
+
−
S2O82−/SO4
*
2−
SO4−*/SO4
Co(NH3)5Cl2+/Co2+
O2/H2O
Table 1. Photophysical Properties of New PV2 and PV3
1.2344,45
a
Oligomers in Solution
a
Potentials relative to standard hydrogen electrode in water unless
b
oligomer
λmax (nm)
ε (104 M−1cm−1
)
otherwise noted. The potentials were measured in butyronitrile with
the sodium calomel reference electrode7 and the value reported here
PV2_CN_CH(OEt)2
PV2_CN_C(O)H
327
241
325
354
244
371
243
356
239
349
242
306
361
250
323
6.09
2.29
9.29
7.52
2.31
6.09
1.73
3.68
1.49
4.14
1.80
1.90
2.92
2.51
3.72
0.99
1.53
5.05
3.53
7.95
2.09
5.38
c
has been corrected to SHE in aqueous solutions.41 The potentials
were measured in acetonitrile with ferrocene as the internal standard42
and the value reported here has been corrected to SHE in aqueous
solutions.41
PV3_CN_SO2(OEt)
PV3_CN_2SO2(OEt)
Na2PV3_SO3_SO3
Cs2PV3_CO2_SO3
aqueous solution).7,21,38−45 The energetics of the Ru complexes
and the parent PV3 molecule confirm that the metal-to-ligand
charge transfer (MLCT) excited state [Ru(bpy)3]2+* does not
have sufficient potential to oxidize or reduce PV3 molecules.38
The only process that may occur in the absence of an electron
acceptor for quenching of the excited Ru complex is triplet−
triplet energy transfer from [Ru(bpy(CO2Me)2)3]2+* to
Na2PV3_SO3_SO3. The lifetime of the excited 3PV3
CsPV3_tripod_SO3
molecules so formed was found to be 1.39
0.02 μs in
b
405
aqueous solutions (Supporting Information Figure S1). Upon
2‑
Cs3PV3_CO2_2SO3
Cs2PV3_O3N_2SO3
Cs2PV3_tripod_2SO3
242
363
241
361
243
356
quenching with a sacrificial electron acceptor S2O8 or
Co(NH3)5Cl2+, oxidized sensitizers [Ru(bpy)3]3+ or [Ru(bpy-
(CO2Me)2)3]3+so formed have sufficient potential to oxidize
PV3 molecules (unfunctionalized). On the other hand,
reduction of PV3 by the excited sensitizer is thermodynamically
not possible (Table 2 and Figure 3). The parent [Ru(bpy)3]3+
sensitizer only has a modest overpotential available to oxidize
unfunctionalized PV3 molecules. Since all PV3 derivatives
investigated in this work have electron-withdrawing groups
such as sulfonates, carboxylates, and amides, the required
potential for oxidation is expected to be more positive than that
for unfunctionalized PV3. For this reason, [Ru(bpy-
(CO2Me)2)3]2+ was used as the standard chromophore in the
time-resolved experiments.
a
Ionic compounds in aqueous solution, and neutral organic molecules
b
in CHCl3. Shoulder.
π−π* peak between chloroform and water indicates that, in
dilute solution, species in both solvents are well dispersed and
do not undergo aggregation by π-stacking. Therefore, we do
not expect excited state−excited state annihilation processes to
play a role in dilute solutions.33,34
Transient absorption measurements were conducted to
probe for spectral signatures of hole injection into PV3
For typical nanosecond pump−probe measurements, sam-
ples purged with N2 in quartz cuvettes beforehand were
17108
dx.doi.org/10.1021/ja306162g | J. Am. Chem. Soc. 2012, 134, 17104−17116

Journal of the American Chemical Society
Article
Figure 3. (a) Redox potential (vs SHE) diagram of the reagents used in time-resolved optical spectroscopy; the detailed potentials are summarized in
Table 2. (b) Plausible reaction paths (solid arrows) and thermodynamically unfavorable reactions (dashed arrow with cross) in pump−probe
measurements with PV3 wires.
2−
Figure 4. (a) Transient absorption signal at 600 nm upon 450 nm excitation of sensitizer in the presence of Na2PV3_SO3_SO3 (0.26 mM)(S2O8
,
0.50 M; [Ru(bpy(CO2Me)2)3]2+, 5 μM). Inset: phosphorescence recovery at 600 nm on expanded time scale. (b) Decay of PV3+ 600 nm signal. (c)
Bleach recovery of [Ru(bpy(CO2Me)2)3]2+ and growth of PV3+ at 470 nm. Inset: bleach recovery at 470 nm on expanded scale. (d) Transient
absorption signal at 600 nm upon 450 nm excitation of sensitizer in the presence of Cs2PV3_O3N_2SO3 (50 μM)(Co(NH3)5Cl2+, 10 mM;
[Ru(bpy(CO2Me))3]2+, 10 μM). Inset: Phosphorescence recovery at 600 nm on expanded scale. (e) Decay of PV3+ 600 nm signal. (f) Bleach
recovery of [Ru(bpy(CO2Me)2)3]2+ at 470 nm. Inset: Bleach recovery at 470 nm on expanded scale.
illuminated at 450 nm (3−8 mJ/pulse, 8 ns pulse duration) to
excite the MLCT transition of [Ru(bpy(CO2Me)2)3]2+, a
wavelength that avoids direct excitation of the PV3 molecules.
recovery of the ground state from the sensitizer photobleach,
shown on an expanded scale in the inset, is followed by growth
of the PV3 radical absorption. By contrast, only partial recovery
of the photobleach of [Ru(bpy(CO2Me)2)3]2+ is observed in
the absence of Na2PV3_SO3_SO3 (Supporting Information
Figure S2a). When probed at 600 nm (Figure 4a), the rise time
of the PV3 radical absorption is 4.5 0.2 μs (preceded on the
nanosecond time-scale by [Ru(bpy(CO2Me)2)3]2+ phosphor-
escence decay shown in the inset), in agreement with the
bleach recovery at 470 nm (4.9 0.1 μs). The 600 nm PV3+
2−
A large excess (0.5 M, 2 × 104 to 105 equivalents) of S2O8
relative to [Ru(bpy(CO2Me)2)3]2+ was used to ensure
formation of substantial concentrations of [Ru(bpy-
(CO2Me)2)3]3+. In time-resolved measurements with excess
Na2PV3_SO3_SO3 relative to [Ru(bpy(CO2Me)2)3]2+, the
reaction was probed at 470 nm near the MLCT λmax of the
sensitizer, and at 600 nm where the absorption λmax of PV3
radical cations is anticipated (Figure 4).7 The radical cation
absorption at 600 nm overlaps with the phosphorescence of
[Ru(bpy(CO2Me)2)3]2+. As shown in Figure 4c, at 470 nm, the
absorption signal decays slowly with a time constant of 300
5
μs (Figure 4b). An exponential fit of the lifetime for the 650 nm
phosphorescence emission gives a 1/e time of 56 4 ns, which
17109
dx.doi.org/10.1021/ja306162g | J. Am. Chem. Soc. 2012, 134, 17104−17116

Journal of the American Chemical Society
Article
Figure 5. Proposed elementary steps involved in the generation and decay of PV3 radicals after photosensitization by [Ru(bpy(CO2Me))3]2+.
is consistent with a bimolecular quenching process between
spontaneous secondary redox reactions following the electron
transfer step do not occur. To probe if the alkyl groups of the
multidentate anchors will adversely affect the hole injection
kinetics due to oxidative degradation, Cs2PV3_O3N_2SO3
was examined in time-resolved optical spectroscopic studies.
2‑
[Ru(bpy(CO2Me)2)3]2+* and S2O8 with a lower limit of k =
3.6 × 107 L mol−1 s−1 (lower limit because of consumption of
2‑
S2O8 during the photolysis).
These results are satisfactorily explained by the elementary
steps summarized in Figure 5. After quenching by the electron-
acceptor S2O82‑ at near diffusion-controlled rates, the [Ru(bpy-
(CO2Me)2)3]3+ thus formed injects a hole into
Na2PV3_SO3_SO3 resulting in growth of the radical cation
absorption with a 4.5 μs time constant. Subsequent degradation
of the PV3 radical by excimer formation or other processes
leads to the gradual decay of the transient absorption signal.
The transient absorption spectrum, shown in Figure 6a,
[Ru(bpy(CO2Me)2)3]2+* is quenched with a lifetime of 81
2
ns, and the radical cation absorption at 600 nm grows with a
time constant of 25.8 0.4 μs (Figure 4d). The latter agrees
well with the bleach recovery of the sensitizer of 27.7 0.4 μs
monitored at 470 nm (Figure 4f). The corresponding average
bimolecular hole transfer constant is 7.5 × 108 L mol−1 s−1. The
cation subsequently decays with a time constant of 343 3 μs;
the characteristic transient absorption spectrum is shown in
Figure 6b. The chemical stability of the PV3
(Cs2PV3_O3N_2SO3) wire molecule is much higher with
[Co(NH3)5Cl]2+ as the sacrificial acceptor compared to the
2‑
experiments with S2O8 , as can be seen from the negligible loss
of UV band intensity below 400 nm after irradiation (Figure S3,
experiment with similar concentration and number of laser
shots as Na2PV3_SO3_SO3 measurements). The electron
acceptor [Co(NH3)5Cl]2+ has a lower aqueous solubility than
S2O82‑, slightly slower quenching rates possibly due to
electrostatic repulsion with [Ru(bpy(CO2Me))3]2+, and sub-
stantial UV−visible absorption intensity that limits the usable
concentration due to the optical inner filter effect. Despite these
limitations, the better stability of the PV3 derivatives in the
presence of [Co(NH3)5Cl]2+ made the Co complex the
preferred sacrificial quencher in subsequent transient optical
measurements of hole injection into molecular wires anchored
on metal oxide surfaces.
2 . 3 . C o v a l e n t A t t a c h m e n t o f p - O l i g o -
(phenylenevinylene) on SiO2 and Co3O4 Nanoparticles.
Inspired by postsynthetic bioconjugation protocols on silica
surfaces,46−52 we sought a procedure that would be suitable for
covalent attachment of PV3 wire molecules on SiO2 and Co3O4
nanoparticle surfaces. In initial attempts, peptide coupling
condensation reagents25,26 to form covalent linkages between
surface Si−OH or Co−OH groups and the carboxylates from
the PV3 wires were employed. For example,
Cs3PV3_CO2_2SO3 was activated with HCTU as the
dehydrating agent prior to adding 12 nm SiO2 nanoparticles
to the aqueous solution, which resulted in a maximum of 8 mol
% PV3 surface coverage. After repeated centrifugation and
rinsing cycles to remove physisorbed PV3 molecules, the
functionalized SiO2_PV3_CO2_2SO3 was isolated as a pale
yellow, fluorescent solid. FT-IR spectra of a powder of SiO2
particles with attached PV3 molecules mixed with KBr are
shown in Figure 7. After subtracting the spectrum of plain SiO2
particles (Figure 7b) from the SiO2_PV3 spectrum (Figure
7a), the resulting difference spectrum (Figure 7c) shows
characteristic bands of Cs3PV3_CO2_2SO3. Figure 7d
presents the infrared spectrum of solid Cs3PV3_CO2_2SO3
in KBr.49,53 There is good agreement between the spectra of
Figure 6. (a) Time evolution of transient absorption spectra of radical
cation of Na2PV3_SO3_SO3 with vibronic structure. Excitation of
[Ru(bpy(CO2Me)2)3]2+ at 450 nm. (b) Transient absorption spectra
of radical cation of Cs2PV3_O3N_2SO3.
confirms the assignment of the 600 nm signal as the λmax of a
PV3 radical cation. The absorption profile is in good agreement
with PV3 radical cation reported in tetrahydrofuran solvent.7 It
is interesting to note that the transient absorption spectrum
shows the characteristic vibronic fine structure with a frequency
of about 690 cm−1. However, due to formation of the strongly
−
oxidizing SO4 radical when persulfate is used as the sacrificial
quencher, Na2PV3_SO3_SO3 is rapidly degraded in solution.
Therefore, time-resolved measurements had to be conducted
with optimized concentrations of the molecular wire and just a
few laser shots before replacing the sample. The static UV−
visible spectra before and after the laser photolysis experiments
clearly demonstrate significant loss of the π−π* band below
400 nm (Figure S3). To avoid this problem and the
complications that arose for the kinetic treatment, an alternative
electron acceptor [Co(NH3)5Cl]2+ was used for the majority of
the remaining experiments.
The reduced aquated CoII ions generated by quenching of
[Ru(bpy(CO2Me)2)3]2+* by [Co(NH3)5Cl]2+ cannot be
reoxidized to CoIII by the radical cation (Table 2). Therefore,
17110
dx.doi.org/10.1021/ja306162g | J. Am. Chem. Soc. 2012, 134, 17104−17116

Journal of the American Chemical Society
Article
the water byproduct. Excessively high temperatures were
avoided to prevent decomposition of the PV3 side-chains and
aggregation of the SiO2 particles by condensation reactions
between surface silanol groups of different silica particles.
Addition of about 20 mol % of lithium dodecylsulfate as a co-
surfactant improved the loading, probably by dispersing the
SiO2 particles. After centrifugation and an aqueous workup
procedure similar to the method mentioned before, the
functionalized silica material was characterized with vibrational
and optical spectroscopy and found to contain the expected
spectral features of PV3 wires (Supporting Information Figure
S5). The FT-Raman spectrum of SiO2_PV3_tripod_2SO3
shown in Figure 8a further corroborates the attachment of
Figure 7. FT-IR spectra of PV3_CO2_2SO3 molecules anchored on
12 nm SiO2 nanoparticles. (a) Spectrum of SiO2_PV3_CO2_2SO3 in
a KBr pellet. (b) Spectrum of SiO2 nanoparticles in a KBr. (c)
Difference spectrum of SiO2_PV3_CO2_2SO3 and SiO2. The black
arrow indicates the CO stretch of anchored silyl ester. (d) Spectrum
a KBr. Strong IR bands of
Figure 8. FT-Raman spectra of PV3_tripod_2SO3 molecules on 12
nm SiO2 nanoparticles and PV3_tripod_SO3 molecules on 5 nm
Co3 O4 nanoparticles. (a) Spectrum of powder of
SiO2_PV3_tripod_2SO3. The two red arrows indicate Raman
bands attributed to C−H stretch modes of methylene groups of the
anchored tripod. (b) Spectrum of Cs2PV3_tripod_2SO3. (c)
Spectrum of Co3O4_PV3_tripod_SO3. (d) Spectrum of CsPV3_tri-
pod_SO3. Strong Raman bands of the wire molecules are indicated by
dashed lines for convenience. Bands with red asterisks originate from
Co3O4. The sharp spike at 600 cm−1 in traces (a) and (c) is due to an
instrumental artifact.
of Cs3PV3_CO2_2SO3 in
Cs3PV3_CO2_2SO3 are indicated by dashed lines for convenience.
the neat wire and the anchored wire molecule except for small
frequency shifts and broadening of some bands in the case of
the anchored molecule. The most significant change is the
replacement of the CO stretch of the carboxylic acid group
around 1663 cm−1 for pristine Cs3PV3_CO2_2SO3 (Figure
7d) by an ester CO stretch at 1710 cm−1 upon covalent
anchoring (arrow, Figure 7c).54 Transmission electron
microscopy confirms that the silica nanoparticles remain well
dispersed with the original size distribution, although some
aggregates exist (Figure S4).
In pursuit of an even simpler protocol to anchor the PV3
wires to metal oxide surfaces, an alternative method that
exploits covalent attachment via dehydration was developed.
Since multidentate photosensitizers have been successfully
adsorbed on TiO2 surfaces,4,11−20,55−58 the poly alcohol
derivatives CsPV3_tripod_SO3, Cs2PV3_O3N_2SO3 and
Cs2PV3_tripod_2SO3 were used to ensure a perpendicular
orientation of the wires while enhancing chemical stability
through the chelating effect of three anchor points. Briefly, after
dispersing the finely crushed PV3 salts with the 12 nm SiO2
particles in a polar aprotic solvent with a moderate boiling
point such as 1,2-dimethoxyethane (DME) and 1,2-difluor-
obenzene, the reaction mixture was heated to reflux, capped
with a Soxhlet extractor containing CaH2 to dehydrate the
solvent. The dehydration is driven to completion by removal of
intact PV3 wires since all bands of the wire molecule (trace b)
appear in the spectrum of the SiO2 anchored molecule (trace
a). Moreover, the C−H stretch absorptions of the tripodal
anchor at 2911 and 2868 cm−1 are observed with enhanced
intensity compared to the free molecule, as was previously
noted in the case of tripodal anchoring on another oxide surface
(TiO2). This observation further confirms that PV3_tripod_2-
SO3 is attached to the solid oxide surface via the three alcohol
groups.20,49,53,59
The method using HCTU proved equally effective under the
same experimental conditions for the covalent anchoring of
PV3 molecules on Co3O4 catalyst nanoparticles surfaces. Figure
9 shows FT-IR spectra of PV3_tripod_SO3 wire molecules
anchored on 5 nm Co3O4 particles. When subtracting the
spectrum of plain 5 nm Co3O4 particles (Figure 9b) from the
Co3O4_PV3 spectrum (Figure 9a), the resulting difference
spectrum (Figure 9c) shows characteristic bands of CsPV3_tri-
pod_SO3. Figure 9d presents the infrared spectrum of solid
CsPV3_tripod_SO3 in KBr. There is good agreement between
the spectra of the pristine wire and the anchored wire molecule
17111
dx.doi.org/10.1021/ja306162g | J. Am. Chem. Soc. 2012, 134, 17104−17116

Journal of the American Chemical Society
Article
p a r t i c l e s . T h e U V − v i s i b l e s p e c t r u m o f
SiO2_PV3_CO2_2SO3 dispersed in water in the presence of
200 mM Li2SO4 is illustrated in the Supporting Information
Figure S6. The silica constructs aggregate upon drying and
scatter light significantly when redispersed in aqueous solutions.
The high concentration of Li+ aids in the dispersion of the silica
constructs, presumably by electrostatically adhering to the
negatively charged silica surface under neutral aqueous
conditions. Moreover, the Li+ ions also minimizes adsorption
of [Ru(bpy(CO2Me)2)3]2+ on the silica surface, thus
minimizing triplet energy transfer from MLCT excited
[Ru(bpy(CO2Me)2)3]2+ to PV3 prior to electron transfer to
the acceptor molecules. The Li+ thus serves to enhance the
formation of [Ru(bpy(CO2Me)2)3]3+ and subsequent hole
transfer to PV3 wire molecules. Assuming that the extinction
coefficient of the PV3 moiety remains unchanged after
attachment on SiO2 and taking into account the 200 m2
g−1surface area of the commercial 12 nm particles, the surface
concentration is estimated to be 0.051 PV3 molecules per nm2
of silica, corresponding to about 22 PV3 molecules per
nanoparticle. The optical and photophysical data of free and
surface attached PV3 molecules are summarized in Table 3.
Nanosecond optical spectroscopy of SiO2_PV3_O3N_2SO3
sample in 250 mM Li+ aqueous solutions probed at 600 nm
gave a rise time of the radical cation of 18 2 μs (Figure 10a)
and a decay time of 208 4 μs (Figure 10b). On the basis of a
SiO2 particle concentration of 1.5 × 10−6 M, a diffusion
controlled bimolecular hole transfer constant of 3.6 × 1010 M−1
s−1 is calculated. Contribution to the signal from any free PV3
molecules detached from the silica particle surface is very small
as illustrated in Figure 10a by the trace (gray color) measured
for the supernatant solution after sonication and removing the
SiO2_PV3_O3N_2SO3 particles by centrifugation. Further-
more, control sensitization experiments of free PV3 molecules
in the presence and absence of Li+ (250 mM) confirmed that
Li+ does not affect the hole transfer kinetics within uncertainties
(Figure S7). The recovery of the [Ru(bpy(CO2Me)2)3]2+
bleach at 470 nm fits an exponential function with a time
Figure 9. FT-IR spectra of PV3_tripod_SO3 molecules anchored on
5
nm Co3 O4 nanoparticles. (a) FT-IR spectrum of
Co3O4_PV3_tripod_SO3 in a KBr. (b) FT-IR spectrum of 5 nm
Co3O4 nanoparticles in a KBr. (c) Difference spectrum of
Co3O4_PV3_tripod_SO3 and 5 nm Co3O4 (normalized to the
Co−O band at 665 cm−1). Strong IR bands of CsPV3_tripod_SO3,
are indicated by vertical lines for convenience. (d) FT-IR spectrum of
CsPV3_tripod_SO3 in a KBr pellet.
constant of 16
1 μs (Figure 10c), in agreement with the
except for small frequency shifts and broadening of some bands
in the case of the anchored molecule. Comparison of the FT-
Raman spectra of anchored Co3O4_PV3_tripod_SO3 and neat
PV3_tripod_SO3 powder (Figure 8c,d) confirms that the wire
molecules remain intact after anchoring. We conclude that this
synthetic method provides a general, straightforward protocol
for attachment of molecular wires on metal oxide surfaces.
2.4. Kinetics of Hole Injection into p-Oligo-
(phenylenevinylene) Wires Attached to SiO2 Nano-
growth kinetics of the PV3 radical cation. The identity of the
PV3 radical cation is confirmed by the transient absorption
spectrum shown in Figure 10d. We conclude that hole transfer
from the oxidized Ru sensitizer to the anchored PV3 wire
molecule is very efficient at a driving force (overpotential) of
480 mV. The inset of Figure 10a shows on an expanded scale
the quenching of the [Ru(bpy(CO2Me)2)3]2+* emission by
electron transfer to the [Co(NH3)5Cl]2+ acceptor with a time
Table 3. Photophysical Properties in Aqueous Solutions of Selected PV3 Derivatives and 12 nm Silica Nanoparticles with PV3
Wires Anchored on the Surface
sample
λmax (nm)
ε (104 M−1 cm−1
Na2PV3_SO3_SO3
349
4.1
Cs2PV3_O3N_2SO3
361
SiO2_PV3_CO2_2SO3 SiO2_PV3_O3N_2SO3 Co3O4_PV3_tripod_SO3
324
5.1
350
8.0
365
3.7
8.3
)
8.0
Surface Coverage
(molecules nm−2
0.051
0.097
)
Surface Coverage (per
nanoparticle)
22
42
417
a
PV3 Formation τ (μs) 4.5 0.2 (600 nm)
25.8 0.4 (600 nm)
27.7 0.4 (470 nm)
25 8 (600 nm)
30 2 (470 nm)
18 2 (600 nm)
16 1 (470 nm)
10.4 0.2 (470 nm)
4.9 0.1 (470 nm)
a
PV3 Formation k (109 0.82
0.75
>0.51
36
−
M−1s−1
)
a
No PV3 radical cation observed on the 10 ns and longer time scale, only recovery of RuII at 470 nm is detected.
17112
dx.doi.org/10.1021/ja306162g | J. Am. Chem. Soc. 2012, 134, 17104−17116

Journal of the American Chemical Society
Article
Figure 11. (a) Transient absorption trace (black) at 600 nm upon 450
n m e x c it a t i o n o f s e n s it i z e r i n t h e p r e s e nc e o f
Co3O4_PV3_tripod_SO3 (Co(NH3)5Cl3, 15 mM; [Ru(bpy-
(CO2Me)2)3]2+, 20 μM; Li+, 0.25 M; CH3CN:H2O = 1/5). Gray
trace shows transient observed for supernatant. (b) Black trace: Bleach
recovery of [Ru(bpy(CO2Me)2)3]2+ at 470 nm. The gray trace shows
transient observed for supernatant.
Figure 10. (a) Transient absorption signal at 600 nm upon 450 nm
excitation of sensitizer in the presence of SiO2_PV3_O3N_2SO3 (4
mg) in a 2 mL suspension (Co(NH3)5Cl3, 10 mM; [Ru(bpy-
(CO2Me)2)3]2+, 10 μM). The gray trace is the time-resolved signal of
the supernatant. Inset: Sensitizer phosphorescence recovery at 600 nm
with a lifetime of 160 2 ns. (b) Decay of SiO2_PV3_O3N_2SO3
radical cation at 600 nm. (c) Bleach recovery of [Ru(bpy-
(CO2 Me)2 )3]2+ at 470 nm after hole injection into
SiO2_PV3_O3N_2SO3. The gray trace is the time-resolved signal
of the supernatant. Inset: Bleach recovery at 470 nm with a lifetime of
substantially higher concentration of attached PV3 wire
molecules for Co3O4_PV3_tripod_SO3: 417 PV3_tripod_-
SO3 molecules per Co3O4 particles versus 22 PV3−O3N
molecules per SiO2 particle. Yet, no significant PV3+ radical
cation signal can be discerned from the 600 nm transient
absorption experiment; only the emission of [Ru(bpy-
(CO2Me)2)3]2+* and its decay by conversion to [Ru(bpy-
(CO2Me)2)3]3+ is observed. By contrast, bleach recovery of
[Ru(bpy(CO2Me)2)3]2+ with a time constant of 10.4 0.2 μs
can clearly be seen from Figure 11b, indicating hole injection
into anchored PV3 molecules. The same observations were
made for Co3O4−PV3 samples with lower PV3 loading. No
significant recovery of the bleach at 470 nm is detected on the
tens of μs time scale when conducting the sensitization
experiments with bare (wire-free) Co3O4 particles (the hole
injection time from [Ru(bpy(CO2Me)2)3]3+ into Co3O4
particles has been measured as 25 ms).60 Contributions to
the signal from any free PV3 molecules detached from the
Co3O4 particle surface are absent, as can be seen in Figure 11
by the faint traces of photolysis experiments with supernatant
solutions recorded under the same conditions used for the
Co3O4_PV3_tripod_SO3 suspensions. We conclude that hole
injection from [Ru(bpy(CO2Me)2)3]3+ to PV3 wire molecules
attached to Co3O4 molecules occurs, yet no concurrent radical
cation signal on the time scale of tens of nanoseconds or
microseconds is observed.
151
6 ns. (d) Transient absorption spectral decay of
SiO2_PV3_O3N_2SO3 radical cation after excitation at 450 nm.
constant of 160 2 ns (k = 6.3 × 108 M−1 s−1). These results
indicate that rise and decay of the radical cation generated by
hole injection into SiO2-anchored wire molecules is similar to
that observed for the corresponding free PV3_O3N_2SO3
molecules.
A similar study of time-resolved hole injection was conducted
for PV3 molecules anchored on the SiO2 particle surface by a
carboxylate group (SiO2_PV3_CO2_2SO3). The results,
presented in Table 3 and Figure S8 are again the same as for
the free wire in solution, thus, confirming that anchoring of
PV3 wire molecules on an inert oxide surface, whether through
a carboxylate group or via a combination of two alcohols plus
an amide group, does not substantially affect the fate of the
injected hole.
2.5. Hole Injection into p-Oligo(phenylenevinylene)
Wires Attached to Co3O4 Nanocatalyst Particles. When
exploring hole injection into PV3 wire molecules covalently
attached to Co3O4 nanocatalyst particles using [Ru(bpy-
(CO2Me)2)3]2+ as the sensitizer and [Co(NH3)5Cl]2+ as the
acceptor (pH 7), no significant absorption of the radical cation
was detected at 600 nm on the time scale of nanoseconds and
microseconds. Twenty percent acetonitrile was added to the
aqueous solution in order to prevent precipitation of the
Co3O4_PV3 particles at the desired concentrations. For
nanosecond pump−probe measurements with Co3O4_PV3
particles, samples were purged with N2 in 10 × 2 mm cuvettes
and excited at 450 nm (20−30 mJ/pulse). The higher power
and shorter path length for the probe light compared to the
SiO2_PV3 study were necessary to provide sufficient signal-to-
noise due to light absorption by the black-colored Co3O4
nanoparticles. Figure 11a shows the transient response at 600
nm for a Co3O4_PV3_tripod_SO3 sample in 250 mM Li+
monitored under concentration conditions similar to those of
SiO2_PV3 (Co3O4 particle concentration 0.46 μM), but with
3. DISCUSSION
p-Oligo(phenylenevinylene) molecules with 3 units (PV3) offer
suitable prototypes as hole conducting molecular wires with
reasonable photochemical stability for charge transport from a
visible light sensitizer to a Co3O4 nanoparticle catalyst for water
oxidation. We have synthesized a small library of PV3
derivatives for the covalent anchoring of the wire molecules
on the oxide particle surface that includes anchoring groups
such as tripodal −C(O)NHC(CH2OH)3 or bidentate
−C(O)N(CH2CH2OH)2. Sulfonate groups in ortho and/
or para position on the opposite end of the wire molecule
17113
dx.doi.org/10.1021/ja306162g | J. Am. Chem. Soc. 2012, 134, 17104−17116

Journal of the American Chemical Society
Article
facilitate radial arrangement of the wire molecules and at the
same time afford adequate water solubility for spectroscopic
characterization and hole injection experiments. Anchoring of
PV3 wire molecules on Co3O4 nanoparticle catalyst surfaces,
and on insulating SiO2 for control experiments using either
HCTU coupling or the Soxhlet extractor dehydration method,
was found to provide stable surface attachment for spectro-
scopic studies and photosensitized charge injection experiments
in aqueous solution. These anchors are known to impose radial
(vertical) orientation of the wire molecule according to
previous work by Meyer and Gallopini in the case of TiO2
surfaces,11,16−20 thus enforcing proper geometry of the wire
molecules in addition to electrostatic repulsion between the
negatively charge Co3O4 surface and the sulfonate groups.
With covalent attachment of intact wire molecules to
insulating SiO2 and catalytic Co3O4 nanoparticle surfaces
established by FT-Raman, FT-IR and UV−visible spectroscopy,
time-resolved optical absorption measurements reveal efficient
hole injection from the oxidized sensitizer into the anchored
wire molecules. While hole transfer from oxidized sensitizer
into PV3 wires attached to SiO2 nanoparticles is observed both
by the growth of the 600 nm transient absorption of the PV3+
radical cation and the recovery of the reduced Ru sensitizer at
470 nm, the absence of a signal at 600 nm for wires attached to
Co3O4 indicates a much shorter residence time of the hole on
PV3 attached to these catalyst particles. Comparison of the
transient signal amplitude at 600 nm for PV3 on SiO2 particles
(Figure 10a) and the noise of the kinetic trace for PV3 on
Co3O4 at the same wavelength (Figure 11a) allows us to
estimate an upper limit for the hole residence time in the case
of Co3O4−PV3. For SiO2−PV3, the buildup of PV3+
absorbance is 0.007 with negligible decay on the time scale of
its 18 μs rise. In the case of Co3O4−PV3, the 1/e time for PV3+
growth is 10.4 μs (1/k1) as inferred from the rate of 470 nm
recovery of the reduced sensitizer. Buildup of intensity at 600
nm (noise 0.0008 absorbance units) is prevented by a fast
decay process of PV3+. From these observations, we calculate
that PV3+ is depleted at a rate of k2 = 7 × 105 s−1 or larger (k2
calculated according to 0.0008/0.007 = 1 − k2/(k2 + k1), with
k1 = 9.6 × 104 s−1).
4. CONCLUSIONS
Hole conducting wire molecules of type 1,3-di((E)-styryl)-
benzene were functionalized with carboxylate, or with bi- or
tridentate polyalcohol amide anchors for covalent attachment
to Co3O4 and SiO2 nanoparticle surfaces with a preferential
radial arrangement. No drastic change of HOMO and LUMO
energy levels by the functionalization was noted as indicated by
the UV−vis spectra and the fact that hole transfer from
[Ru(bpy(CO2Me)2)3]3+ to anchored wire molecules was
observed for all derivatives. Two methods were developed for
anchoring of the wire molecules. The first uses a peptide
coupling agent (HCTU) for activating the OH groups while
the other is a more general method for driving condensation of
the OH anchor groups of PV3 with surface hydroxyl groups to
completion using dehydration by CaH2. FT-infrared and FT-
Raman signatures present spectroscopic evidence for covalent
surface attachment.
Observation of the PV3 radical cation by transient optical
absorption spectroscopy, upon excitation of Ru(bpy-
(CO2Me)2)3 in the presence of an electron acceptor in
homogeneous aqueous solution, provides direct evidence for
hole injection into wire molecules. The same radical cation
signal with similar kinetics (10−20 μs growth, hundreds of μs
decay) is observed for wire molecules attached to SiO2
nanoparticles. The growth kinetics of PV3+, which is mirrored
by the recovery of reduced Ru(bpy(CO2Me)2)3 sensitizer,
indicates very efficient charge transfer into free wire molecules
as well as those anchored on the silica surface. For PV3
anchored on Co3O4 catalyst particles, no radical cation signal is
observed despite the fact that hole injection into wire molecules
is indicated by fast (10.4 μs) recovery of reduced Ru(bpy-
(CO2Me)2)3 sensitizer. This observation implies that the hole is
transferred from the wire to the Co3O4 particle within a
microsecond or faster. This is consistent with the favorable
energy alignment of the HOMO of wire and catalyst. Hence,
the approach of using anchored wire molecules may open up an
efficient hole transfer pathway from oxidized donor [Ru(bpy-
(CO2Me)2)3]3+ to Co3O4 catalyst particles. In the next step,
casting of a dense phase, few nanometer thin silica layer around
the wire molecules will introduce a physical barrier for
separating the water oxidation catalysis from the light absorber
and the reductive chemistry.66 At the same time, the
surrounding silica protects the wire molecules from oxidative
damage. When implemented in the form of asymmetrically
functionalized core−shell nanotubes instead of the spherical
core−shell particles, separate physical spaces will be available
for O2 evolution and light absorber/reduction chemistry.
Separation of the two half reactions by a physical barrier is a
critical ingredient of an artificial photosynthetic system.
The most probable pathway of PV3+ loss is hole transfer to
the attached Co3O4 particle. Alternative hole transfer pathways,
in particular reaction with other wire molecules or unknown
species, are very unlikely because of the observed long hole
lifetime, under otherwise similar experimental conditions, for
wire molecules anchored on (insulating) SiO2 particles. The
fast hole injection from anchored PV3 to Co3O4 is plausible in
particular because the HOMO of the wire molecule (1.4 V)6 is
energetically well aligned with the HOMO of the Co3O4
acceptor particle (1.2 V). While we are not aware of precedents
for hole transport across conducting wire molecules covalently
attached to a metal oxide surface, hole transfer by an incoherent
hopping mechanism on submicrosecond time scales was
reported for organic donor−acceptor systems featuring oligo-
(phenylenevinylene) bridges.6,7,18,55,61−65 The fastest reported
hole transfer so far from a visible light sensitizer, a Ru(bpy)3
congener attached to a metal oxide nanoparticle catalyst, IrO2,
via an acetylacetonate anchor was determined by Mallouk to be
2.2 ms.4,5 We conclude that coupling of a sensitizer with Co3O4
particles via oligo-para(phenylenevinylene) molecular wire may
offer a new approach for efficient light-induced hole transfer to
a catalyst nanoparticle.
5. EXPERIMENTAL SECTION
5.1. Synthetic Materials and Methods. Silica gel (SiliaFlash
P60, SiliCycle) was used for column chromatography, while analytical
thin layer chromatography was performed using EMD Chemicals, Inc.
60 F254 silica gel (precoated plastic sheets, 0.25 mm thick). All
moisture-sensitive reactions were performed using Schlenk techniques
under a nitrogen atmosphere. Deuterated solvents and isotopically
labeled reagents were procured from Cambridge Isotope Laboratories
(Cambridge, MA). Solvents such as acetonitrile, chloroform, dichloro-
methane, ethyl acetate, hexanes, and toluene were obtained from EMD
Chemicals, Inc. or Honeywell and used without further purification.
All other chemicals, including anhydrous DMF and THF, were
purchased from Sigma-Aldrich (St. Louis, MO) and were used as
17114
dx.doi.org/10.1021/ja306162g | J. Am. Chem. Soc. 2012, 134, 17104−17116

Journal of the American Chemical Society
Article
received. Details for the synthesis of functionalized PV3 molecular
wires are presented in the Supporting Information.
ASSOCIATED CONTENT
* Supporting Information
■
S
1
5.2. Instrumentation. H and 13C NMR spectra were recorded
Detailed synthesis methods of functionalized PV3 molecular
wires. Spectroscopic measurements, time-resolved data, and
TEM images on some of the other PV3 derivatives and silica
constructs. This material is available free of charge via the
Internet at http://pubs.acs.org.
using a Bruker Biospin Avance II 500 MHz High Performance NMR
Spectrometer in the Chemistry NMR Facility at the Molecular
Foundry Organic and Macromolecular Synthesis Facility, LBNL. The
NMR chemical shifts are reported in the standard δ notation of parts
per million (ppm) with respect to residual protonated solvent;
coupling constants are reported in Hz. High-resolution mass spectral
analyses were performed by the QB3/Mass Spectrometry facility at the
College of Chemistry, University of California, Berkeley. UV−visible
absorption spectra and kinetic runs were recorded using a Shimadzu
UV-2100 spectrophotometer, with solution samples in 10 mm × 5 mm
or 10 mm × 2 mm cuvettes. FT-IR spectra were recorded at 0.5 cm−1
resolution using a Bruker IFS88 or IFS66V spectrophotometer
equipped with liquid nitrogen cooled MCT detectors. FT-Raman
spectra were recorded with the IFS66V spectrophotometer with a FT-
Raman module FRA-106 at 1 cm−1 resolution.
AUTHOR INFORMATION
Corresponding Author
HMFrei@lbl.gov
■
Present Address
†Dept. of Chemistry, University of Oxford, South Parks Road,
Oxford
Notes
The authors declare no competing financial interest.
5.3. Time-Resolved Optical Absorption Spectroscopy. For
the nanosecond optical kinetics measurements, an Edinburgh
Instruments model LP920 transient absorption spectrometer equipped
with a pulsed Xe probe lamp or a CW halogen lamp (for
measurements >1 ms) was used in conjunction with a Nd:YAG
laser-pumped tunable Optical Parametric Oscillator (Continuum
model Surelite II and Surelite OPO Plus) as the excitation source.
The laser pulse width was 8 ns and the repetition rate was 10 Hz. For
measurements of the silica nanoparticle constructs, transient
absorption spectroscopy for lifetimes shorter than 20 ns is complicated
by visible light scattering of the incident pulse and emission from
defect sites of the silica material.
In typical nanosecond pump−probe measurements, samples of the
dissolved PV3 and silica constructs in sealed 10 × 5 mm cuvettes are
purged with N2 and illuminated with 450 nm laser pulses at 3−8 mJ
for kinetics measurements, and up to 22 mJ pulse energy when
recording transient absorption spectral maps. The laser excitation
beam was at right angle to the Xe (or halogen lamp) probe beam. For
samples of the Co3O4 constructs, sealed 10 × 2 mm cuvettes
containing the nanoparticulate suspension are purged with N2 and
illuminated with 450 nm laser pulses at 20−30 mJ for kinetics
measurements. A colinear geometry of the laser excitation beam and
the Xe probe beam was adopted to maximize the overlap of the two
beams. The wavelength of 450 nm was used to excite the MLCT
transition of [Ru(bpy(CO2Me))3]2+ while avoiding direct excitation of
the PV3 molecules. A large excess of the sacrificial electron acceptors
S2O82‑ or Co(NH3)5Cl3 was used to ensure a large quantum efficiency
for generating [Ru(bpy(CO2Me))3]3+. To obtain the time constants
for the elementary steps, the data were fit to single exponential or
biexponential functions:
ACKNOWLEDGMENTS
■
This work was funded by the Helios Solar Energy Research
Center, which is supported by the Director, Office of Science,
Office of Basic Energy Sciences of the U.S. Department of
Energy under Contract No. DE-AC02-05CH11231. Anil Agiral,
Rubicon Postdoctoral Fellow, acknowledges support by The
Netherlands Organization for Scientific Research (NWO).
Portions of this work (NMR measurements) were performed as
a User Project at the Molecular Foundry, Lawrence Berkeley
National Laboratory, which is supported by the Office of
Science, Office of Basic Energy Sciences, of the U.S.
Department of Energy under Contract No. DE-AC02-
05CH11231. The authors thank Dr. Beth Anne McClure for
assistance with the transient absorption measurements.
REFERENCES
■
(1) Jiao, F.; Frei, H. Angew. Chem., Int. Ed. 2009, 48, 1841.
(2) Jiao, F.; Frei, H. Energy Environ. Sci. 2010, 3, 1018.
(3) Fogarty, J. C.; Aktulga, H. M.; Grama, A. Y.; van Duin, A. C. T.;
Pandit, S. A. J. Chem. Phys. 2010, 132, 174704.
(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.
(5) Youngblood, W. J.; Lee, S. H.; Maeda, K.; Mallouk, T. E. Acc.
Chem. Res. 2009, 42, 1966.
(6) Davis, W. B.; Svec, W. A.; Ratner, M. A.; Wasielewski, M. R.
Nature 1998, 396, 60.
(7) Davis, W. B.; Ratner, M. A.; Wasielewski, M. R. Chem. Phys. 2002,
281, 333.
(8) Pavlishchuk, V. V.; Addison, A. W. Inorg. Chim. Acta 2000, 298,
97.
(t−τ0)/τ1
y = y + A1 e
0
(9) Kalyanasundaram, K. Coord. Chem. Rev. 1982, 46, 159.
(t−τ0)/τ1
(t−τ0)/τ2
(10) Gratzel, M. Acc. Chem. Res. 2009, 42, 1788.
̈
y = y + A1 e
+ A2 e
0
(11) Hoertz, P. G.; Carlisle, R. A.; Meyer, G. J.; Wang, D.;
Piotrowiak, P.; Galoppini, E. Nano Lett. 2003, 3, 325.
(12) Hardin, B. E.; Sellinger, A.; Moehl, T.; Humphry-Baker, R.;
Moser, J. E.; Wang, P.; Zakeeruddin, S. M.; Gratzel, M.; McGehee, M.
D. J. Am. Chem. Soc. 2011, 133, 10662.
The parameters y0, τ0, An, and τn were determined by a least-squares
fitting procedure. The term y0 corresponds to the vertical intercept at
infinite time and depends on whether the signal recovers to the
baseline or retains a long-term bleach or absorption in the
measurement window. The variable τ0 is the delay time of the
excitation pulse after start of the probe measurement. An is the change
in optical density after irradiation for the nth exponential term, while τn
is the time constant of the corresponding nth fit. Biexponential fits are
used only when two consecutive elementary reactions are anticipated.
This is the case for probing at 600 nm where the decay of
[Ru(bpy(CO2Me))3]2+* phosphorescence is immediately followed by
growth of the absorption of PV3 radical cation.
(13) McNamara, W. R.; Snoeberger, R. C.; Li, G.; Schleicher, J. M.;
Cady, C. W.; Poyatos, M.; Schmuttenmaer, C. A.; Crabtree, R. H.;
Brudvig, G. W.; Batista, V. S. J. Am. Chem. Soc. 2008, 130, 14329.
(14) Xiao, D. Q.; Martini, L. A.; Snoeberger, R. C.; Crabtree, R. H.;
Batista, V. S. J. Am. Chem. Soc. 2011, 133, 9014.
(15) Franking, R. A.; Landis, E. C.; Hamers, R. J. Langmuir 2009, 25,
10676.
(16) Galoppini, E.; Guo, W.; Zhang, W.; Hoertz, P. G.; Qu, P.;
Meyer, G. J. J. Am. Chem. Soc. 2002, 124, 7801.
17115
dx.doi.org/10.1021/ja306162g | J. Am. Chem. Soc. 2012, 134, 17104−17116

Journal of the American Chemical Society
Article
(17) Galoppini, E.; Guo, W. Z.; Qu, P.; Meyer, G. J. J. Am. Chem. Soc.
2001, 123, 4342.
(18) Meyer, G. J. Inorg. Chem. 2005, 44, 6852.
(19) Piotrowiak, P.; Galoppini, E.; Wei, Q.; Meyer, G. J.; Wiewior, R.
J. Am. Chem. Soc. 2003, 125, 5278.
(20) Wang, D.; Mendelsohn, R.; Galoppini, E.; Hoertz, P. G.;
Carlisle, R. A.; Meyer, G. J. J. Phys. Chem. B 2004, 108, 16642.
(21) Garner, L. E.; Park, J.; Dyar, S. M.; Chworos, A.; Sumner, J. J.;
Bazan, G. C. J. Am. Chem. Soc. 2010, 132, 10042.
(22) Nielsen, C. B.; Johnsen, M.; Arnbjerg, J.; Pittelkow, M.; McIlroy,
S. P.; Ogilby, P. R.; Jorgensen, M. J. Org. Chem. 2005, 70, 7065.
(23) Abbel, R.; Grenier, C.; Pouderoijen, M. J.; Stouwdam, J. W.;
Leclere, P. E. L. G.; Sijbesma, R. P.; Meijer, E. W.; Schenning, A. P. H.
J. J. Am. Chem. Soc. 2009, 131, 833.
(54) Colthup, N. B.; Daly, L. H.; Wiberley, S. E. Introduction to
Infrared and Raman Spectroscopy, 3rd ed.; Academic Press: New York,
1990; p 301.
(55) Galoppini, E. Coord. Chem. Rev. 2004, 248, 1283.
(56) Myahkostupov, M.; Piotrowiak, P.; Wang, D.; Galoppini, E. J.
Phys. Chem. C 2007, 111, 2827.
(57) Wei, Q.; Galoppini, E. Tetrahedron 2004, 60, 8497.
(58) Yi, K. J.; Wang, H.; Lu, Y. F.; Yang, Z. Y. J. Appl. Phys. 2007,
101.
(59) Moskovits, M.; Suh, J. S. J. Phys. Chem. 1984, 88, 5526.
(60) De Respinis, M. Master Thesis, Technical University of
Denmark, Copenhagen, 2011.
(61) Giacalone, F.; Segura, J. L.; Martin, N.; Guldi, D. M. J. Am.
Chem. Soc. 2004, 126, 5340.
(62) Gray, H. B.; Halpern, J. Proc. Natl. Acad. Sci. U.S.A. 2005, 102,
3533.
(63) Hanss, D.; Walther, M. E.; Wenger, O. S. Coord. Chem. Rev.
2010, 254, 2584.
(64) Gust, D.; Moore, T. A.; Moore, A. L. Acc. Chem. Res. 2009, 42,
(24) Gaylord, B. S.; Wang, S. J.; Heeger, A. J.; Bazan, G. C. J. Am.
Chem. Soc. 2001, 123, 6417.
(25) Albericio, F. Curr. Opin. Chem. Biol. 2004, 8, 211.
(26) Albericio, F.; Bofill, J. M.; El-Faham, A.; Kates, S. A. J. Org.
Chem. 1998, 63, 9678.
(27) Ley, S. V.; Stewart-Liddon, A. J. P.; Pears, D.; Perni, R. H.;
Treacher, K. Beilstein J. Org. Chem. 2006, 2, 15.
(28) Sajiki, H.; Hattori, K.; Hirota, K. J. Org. Chem. 1998, 63, 7990.
(29) Cui, Z.; Zhang, S.; Yang, J.; Tang, L. Front. Chem. China 2008, 3,
425.
1890.
(65) Zhao, Y.; Swierk, J. R.; Megiatto, J. D.; Sherman, B.;
Youngblood, W. J.; Qin, D.; Lentz, D. M.; Moore, A. L.; Moore, T.
A.; Gust, D.; Mallouk, T. E. Proc. Natl. Acad. Sci. U.S.A. 2012, 109,
15612.
(66) Agiral, A.; Soo, H. S.; Frei, H. J. Am. Chem. Soc., submitted for
publication, 2012.
(30) Geng, M.; Zhang, D.; Wu, X.; He, L.; Gong, B. Org. Lett. 2009,
11, 923.
(31) Reuter, C.; Wienand, W.; Hubner, G. M.; Seel, C.; Vogtle, F.
Chem.Eur. J. 1999, 5, 2692.
(32) Colaneri, N. F.; Bradley, D. D. C.; Friend, R. H.; Burn, P. L.;
Holmes, A. B.; Spangler, C. W. Phys. Rev. B 1990, 42, 11670.
(33) Meyer, G. J. Inorg. Chem. 2005, 44, 6852.
(34) Liu, D.; Hug, G. L.; Kamat, P. V. J. Phys. Chem. 1995, 99, 16768.
(35) Morris, N. D.; Suzuki, M.; Mallouk, T. E. J. Phys. Chem. A 2004,
108, 9115.
(36) White, H. S.; Becker, W. G.; Bard, A. J. J. Phys. Chem. 1984, 88,
1840.
(37) Sivasankar, N.; Weare, W. W.; Frei, H. J. Am. Chem. Soc. 2011,
133, 12976.
(38) Bachmeier, A. Diploma Thesis, Friedrich-Alexander-Universitat,
̈
Erlangen-Nurnberg, 2010.
̈
(39) Eskelinen, E.; Luukkanen, S.; Haukka, M.; Ahlgren, M.;
Pakkanen, T. A. J. Chem. Soc., Dalton 2000, 2745.
(40) Kavarnos, G. J.; Turro, N. J. Chem. Rev. 1986, 86, 401.
(41) Pavlishchuk, V. V.; Addison, A. W. Inorg. Chim. Acta 2000, 298,
97.
(42) Schwalbe, M.; Schafer, B.; Gorls, H.; Rau, S.; Tschierlei, S.;
Schmitt, M.; Popp, J.; Vaughan, G.; Henry, W.; Vos, J. G. Eur. J. Inorg.
Chem. 2008, 3310.
(43) Curtis, N. J.; Lawrance, G. A.; Sargeson, A. M. Aust. J. Chem.
1983, 36, 1327.
(44) Mills, A.; Valenzuela, M. A. J. Photochem. Photobiol., A 2004,
165, 25.
(45) Huie, R. E.; Clifton, C. L.; Neta, P. Radiat. Phys. Chem. 1991, 38,
477.
(46) Gerion, D.; Herberg, J.; Bok, R.; Gjersing, E.; Ramon, E.;
Maxwell, R.; Kurhanewicz, J.; Budinger, T. F.; Gray, J. W.; Shuman, M.
A.; Chen, F. F. J. Phys. Chem. C 2007, 111, 12542.
(47) Gerion, D.; Pinaud, F.; Williams, S. C.; Parak, W. J.; Zanchet,
D.; Weiss, S.; Alivisatos, A. P. J. Phys. Chem. B 2001, 105, 8861.
(48) Kumar, R.; Roy, I.; Hulchanskyy, T. Y.; Goswami, L. N.; Bonoiu,
A. C.; Bergey, E. J.; Tramposch, K. M.; Maitra, A.; Prasad, P. N. ACS
Nano 2008, 2, 449.
(49) Li, H. Y.; Tripp, C. P. J. Phys. Chem. B 2004, 108, 18318.
(50) Rossi, L. M.; Shi, L. F.; Quina, F. H.; Rosenzweig, Z. Langmuir
2005, 21, 4277.
(51) Wang, L.; Zhao, W. J.; Tan, W. H. Nano Res. 2008, 1, 99.
(52) Xiao, H. N.; Cezar, N. J. Colloid Interface Sci. 2003, 267, 343.
(53) Saavedra, H. M.; Thompson, C. M.; Hohman, J. N.; Crespi, V.
H.; Weiss, P. S. J. Am. Chem. Soc. 2009, 131, 2252.
17116
dx.doi.org/10.1021/ja306162g | J. Am. Chem. Soc. 2012, 134, 17104−17116