Efficient lateral electron transport inside a monolayer of aromatic amines anchored on nanocrystalline metal oxide films

Article abstract of DOI:10.1021/jp972890j

A monolayer of a phosphonated triarylamine adsorbed on nanocrystalline TiO₂, ZrO₂, or Al₂O₃ film deposited on conducting glass displays reversible electrochemical and electrochromic behavior although the redox potential of the electroactive molecules (0.80 V vs NHE) lies in the forbidden band of the semiconducting or insulating oxides. The mechanism of charge transport was found to involve hole injection from the conducting support followed by lateral electron hopping within the monolayer. The apparent diffusion coefficient ranged from 2 8 ?? 10^-12 m² s^-1 in the neat 1-ethyl-2-methylimidazolium bis(trifluoromethylsulfonyl)imide (EtMeIm⁺Tf₂N^-) to 1.1 ?? 10^-11 m² s^-1 in acetonitrile + 2 M EtMeIm⁺Tf₂N^-. A percolation threshold for electronic conductivity was found at a surface coverage corresponding to 50% of a full monolayer.

Full text of DOI:10.1021/jp972890j

1498
J. Phys. Chem. B 1998, 102, 1498-1507
Efficient Lateral Electron Transport inside a Monolayer of Aromatic Amines Anchored on
Nanocrystalline Metal Oxide Films
Pierre Bonhoˆte,* Eric Gogniat, Sophie Tingry, Christophe Barbe´, Nicolas Vlachopoulos,
Frank Lenzmann, Pascal Comte, and Michael Gra1tzel
Laboratoire de photonique et interfaces, De´partement de chimie, Ecole Polytechnique Fe´de´rale de Lausanne,
CH-1015 Lausanne, Switzerland
ReceiVed: September 4, 1997; In Final Form: December 17, 1997
A monolayer of a phosphonated triarylamine adsorbed on nanocrystalline TiO2, ZrO2, or Al2O3 film deposited
on conducting glass displays reversible electrochemical and electrochromic behavior although the redox potential
of the electroactive molecules (0.80 V vs NHE) lies in the forbidden band of the semiconducting or insulating
oxides. The mechanism of charge transport was found to involve hole injection from the conducting support
followed by lateral electron hopping within the monolayer. The apparent diffusion coefficient ranged from
2.8 × 10-12 m2 s-1 in the neat 1-ethyl-2-methylimidazolium bis(trifluoromethylsulfonyl)imide (EtMeIm+Tf2N-)
to 1.1 × 10-11 m2 s-1 in acetonitrile + 2 M EtMeIm+Tf2N-. A percolation threshold for electronic conductivity
was found at a surface coverage corresponding to 50% of a full monolayer.
Introduction
oxide layers. It appears thus that the electrons are transported
by the triarylamine 1. In such a case, the second and the third
mechanism can be implicated. In the second hypothesis, the
triarylamine molecules should be sufficiently mobile inside the
monolayer or between the monolayer and the contacting
solution, to account for a diffusive electron transport to the
underlying conducting glass. In the third hypothesis, the
molecules should be firmly enough bound to the metal oxide
surface to prevent significant diffusion, and charge motion could
only take place by electron hopping between molecules inside
the monolayer. These two types of electron transport, which
often coexist in a given material, were extensively studied for
various types of electroactive materials deposited on electrodes8
and were the subject of extensive theoretical descriptions and
simulations by Blauch and Save´ant.9 Diffusive and hopping
charge transport exhibit very different behavior. Ideally, in a
purely diffusive process, the apparent diffusion coefficient Dapp
of the electroactive species should be independent of the
concentration of that species in the considered material. Practi-
cally, structural changes often lead to a decrease in Dapp with
increasing concentration.10 Moreover, the conductivity remains
even at low concentration. On the contrary, in an electron
hopping process, Dapp increases with the concentration C of the
electroactive species, as expressed by the Dahms-Ruff (DR)11
and the Laviron-Andrieux-Save´ant (LAS)12 equations (eq 1),
in which kex is the electron self-exchange rate constant and δ
the intermolecular distance at which the electron transfer takes
place.
Nanocrystalline semiconductor films, in particular of TiO2,
are a proven and very versatile component1 in the fabrication
of dye-sensitized solar cells,2 electrochromic devices,3 and
sensors,4 as well as photoelectrosynthetic5 or photoelectrochro-
mic6 cells, owing to their unique properties of high surface area,
transparency, and electronic conductivity. However, based on
the existence of an electronic band gap in the material, it was
assumed until now that interfaces to the nanocrystalline
semiconductor layers are conductive only in the accumulation
regime and are electronically blocking under reverse bias. Thus,
charge-transfer reactions should be restricted to adsorbed species
whose redox potential lies above the conduction band edge. For
TiO2, the flat band potential (Vfb) in aprotic solvents such as
CH3CN is very negative (as far as - 2 V)7 but can be raised to
0 V by addition of Li+.3 Thus, surface-anchored viologens such
as 3 (E° ) -0.4 V) display a reversible behavior only in the
presence of Li+. In the absence of lithium ions, the reduction
is irreversible.
In view of this fact, it was striking to observe that compound
1 did exhibit reversible electrochemistry when grafted onto
nanocrystalline TiO2 layers supported by conducting glass, even
though its redox potential (0.80 V) is located in a domain where
the TiO2 interface is expected to be blocking. The phosphonated
tetraalkyl-p-phenylenediamine 2 (E° ) 0.28 V) also displays
electroactivity but in a much weaker and less reversible fashion.
Three electron transport mechanisms can in principle be
proposed to explain the observed electroactivity. First, the
electrons could be carried by the supporting semiconductor.
Second, they could be transported by the triarylamine molecules
in a diffusive process. Third, they could be transported by the
triarylamine molecules in an electron hopping process. The first
hypothesis implies that an electron flow can exist on the
semiconductor surface, in the middle of the band gap, by
hopping from surface state to surface state. Such a mechanism,
even though it cannot be ruled out on a theoretical basis, is
very unlikely in the present case, as the same electroactivity of
1 as a monolayer was observed on nanocrystalline aluminum
Dapp ) kexδ2C/6
(1)
The meaning of δ is slightly different in the DR and in the
LAS equation which apply to redox centers that are respectively
freely diffusing and linked to a backbone with a restricted
motion freedom around an equilibrium position. None of these
two models describe an assembly where the redox centers are
completely immobilized and where the electron transfer does
not require molecular displacement. In such an extreme case,
S1089-5647(97)02890-3 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/10/1998

Aromatic Amines Anchored on Metal Oxide Films
J. Phys. Chem. B, Vol. 102, No. 9, 1998 1499
classical percolation behavior is expected, characterized by a
sharp threshold below which the conductivity drops to zero as
the contact between the electroactive centers breaks down. If,
however, the electron hopping is not strictly restricted to adjacent
redox sites, the predicted transition at the percolation limit is
softer, and Dapp, even if minute, is not nil below that limit.9b
To our knowledge, the only study on electron transport inside
a monolayer of electroactive molecules on a mesoporous
material was carried out by Majda and co-workers on an
aluminum oxide film in which 80 nm wide pores were organized
perpendicularly to a gold support. The alumina was derivatized
with various polymers13 or bilayers bearing redox active
centers.14 In all those cases, the observed conductivity was
essentially due to the diffusion of the loosely bound charge
carriers in the polymer or in the bilayer. For the latter structure,
the apparent diffusion coefficient was reported to be independent
of the concentration of electroactive species. Even though the
supporting material was not nanocrystalline as in the present
study, the structures of Majda’s electrodes and of ours are
similar, from the fact that in both systems a monolayer of
electroactive molecules is organized on average perpendicularly
to a conductive support, on the inner side of mesopores.
Comparison between the systems is hence relevant.
An example of modified electrodes where conductivity occurs
only by electron hopping was described by Anson et al.15 The
investigated system consisted of a poly(vinylpyridine) film on
a conducting support, loaded with coordinately linked Fe(CN)5
redox units, whose diffusion is negligible. When the concentra-
tion of Fe(CN)5 was varied, a percolation threshold appeared,
below which the electroactive fraction of the redox centers
sharply dropped to zero. Two-dimensional percolation behavior
was observed by Majda et al.16 in a Langmuir monolayer of an
osmium complex diluted by 1-octadecanol at the water surface.
The apparent diffusion coefficient decreased rapidly from 2.4
× 105 m2 s-1 to 0 as the mole fraction of the electroactive
species was reduced from 1 to 0.6.
In the present work, electro- and spectroelectrochemical
experiments were carried out on nanocrystalline electrodes
derivatized by 1 in order to allow discrimination between a
diffusive and an electron hopping mechanism of charge
transport.
array spectrophotometer from Hewlett-Packard, in a 1 cm by 1
cm spectroelectrochemical cell.
Experimental Section
Dynamic viscosity measurements were carried out on a Haake
VT500 microviscosimeter, thermally controlled by a Haake D8
thermostat-cryostat.
Reagents were purchased from Fluka, unless otherwise stated.
Acetonitrile, ethanol, acetic acid, and cyclohexane were puriss,
and the supporting electrolytes were electrochemical grade
quality. Propylene carbonate was obtained from Burdick &
Jackson. tert-Butyl methyl ether (hereafter TBME) was ob-
tained from Siegfried. These compounds were not submitted
to further purification before use. 3-Methoxypropionitrile was
distilled over a Fischer HMS500 column.
1H NMR spectra were recorded on a Bruker Spectrospin 200
MHz spectrometer.
Data processing and curve fitting were done with the help of
the Igor Pro software form Wave Metrics Inc.
Synthesis. Synthesis of 1, 4, and 5 will be reported
1
elsewhere. Synthesis of 3 is reported in ref 3c. All gave H
NMR spectra consistent with their expected structure and free
of foreign signals. Unless otherwise stated, the 1H NMR spectra
given below were recorded in CDCl3 with TMS as an internal
standard, and the chemical shifts are given in ppm versus this
reference. Diazabicylcooctane (δ ) 2.83) was added to protect
the p-phenylenediamines from oxidation by air.
N-Ethyl-N′,N′-dimethyl-p-phenylenediamine. In a dry,
three-necked, round-bottomed flask equipped with a reflux
condenser, a thermometer, and a magnetical stirrer, under Ar,
N,N-dimethyl-p-phenylenediamine (4.08 g, 30 mmol) was
dissolved in 300 mL of dry THF. The solution was cooled to
-65 °C, and n-butyllithium (10 M in hexane, 3.2 mL, 32 mmol)
was slowly added under stirring, keeping the temperature below
Electrochemistry. Potentials are given vs NHE. Cyclic
voltammetric and chronocoulometric experiments were carried
out with a PC-controlled Ecochemie model Autolab P20
potentiostat. In all the experiments, the reference electrode was
Ag/Ag+-saturated NaCl (+0.197 V vs NHE), and the counter
electrode was glassy carbon. Both were separated by a salt
bridge from the solution in contact with the working electrode.
For cyclic voltammetry, the working electrodes were a 4.8 µm
film of TiO2 on a 1 cm2 conducting oxide substrate (SnO2)
derivatizated in 0.5 mM solutions in EtOH for 12-15 h at room
temperature (r.t.). The electrolytic solution was 0.2 M tetrabu-
tylammonium triflate in propylene carbonate. The spectropho-
tometric measurements were carried out with a HP8453 diode

1500 J. Phys. Chem. B, Vol. 102, No. 9, 1998
Bonhoˆte et al.
-60 °C. After 15 min at -65 °C, bromoethane (3.24 g, 30
mmol) was added in one portion. The mixture was allowed to
warm to r.t. and then heated to 50 °C for 1 h. The solution
was concentrated to 10 mL, diluted to 100 mL with heptane,
and washed twice with water. This water was extracted with
dichloromethane (2 × 50 mL). The combined organic phases
were dried over MgSO4 and concentrated to 10 mL. The
concentrate was purified by chromatography over silica gel,
eluting by heptane followed by heptane/TBME (4:1). The
fraction eluted first gave a Rf ) 0.58 by TLC on silica gel with
TBME as eluent. The second one, with Rf ) 0.47, was
concentrated to 50 mL and cooled to -30 °C. The red crystals
formed were filtered below 0 °C and dried under 12 mbar. The
product is liquid at 20 °C and was kept in the solid state at
-20 °C. Yield: 2.85 g (58%). Caution: N-ethyl-N′,N′-
dimethyl-p-phenylenediamine is strongly irritating to the skin
and must be handled with gloves. 1H NMR: 6.74 (dt, 2H, J )
9 and 2.7 Hz), 6.60 (dt, 2H, J ) 9 and 2.7 Hz), 3.10 (q, 2H, J
) 7 Hz), 2.81 (s, 6H), 1.22 (t, 3H, J ) 7 Hz).
Diethyl 3-(Ethyl(p-(N,N-dimethylamino)phenyl)amino)-
propyl-1-phosphonate. N-Ethyl-N′,N′-dimethyl-p-phenylene-
diamine (0.656 g, 4 mmol), diethyl 3-bromopropyl-1-phospho-
nate (1.04 g, 4 mmol), and 0.1 mL of sym-collidine were
dissolved in acetonitrile (20 mL). The solution was refluxed
for 3 h. Silica gel (1.5 g) was added, and the solvent was
removed by a Rotavapor. The remaining solid was suspended
in TBME and placed on the top of a silica gel chromatography
column. Elution was done with TBME. The starting amine
was eluted first (Rf ) 0.63 by TLC on silica gel with acetone
as eluent) followed by the desired product (Rf ) 0.46). Yield:
0.42 g (31%). 1H NMR: 6.74 (s, 4H), 4.07 (quintet, 4H, J ) 7
Hz), 3.22 (m, 4H), 2.80 (s, 6H), 1.86-1.70 (m, 6H), 1.30 (t,
6H, J ) 6.5 Hz), 1.07 (t, 3H, J ) 7 Hz).
Sodium 3-(Ethyl(p-(N,N-dimethylamino)phenyl)amino)-
propyl-1-phosphonate (2). Conversion of the diethylphos-
phonate to the free acid was carried out following Katz et al.:17
diethyl 3-(ethyl(p-(N,N-dimethylamino)phenyl)amino)-propyl-
1-phosphonate (0.42 g, 1.23 mmol) and bromotrimethylsilane
(1.90 mL, 15 mmol) were dissolved in dry dichloromethane
(10 mL) under Ar, and the resulting violet solution was stirred
at r.t. for 15 h. The solvents and excess bromotrimethylsilane
were removed by a Rotavapor under Ar. To the remaining
brown gum was added a mixture of diethyl ether (10 mL),
ethanol (10 mL), and collidine (1 mL), whereby a white solid
was formed, which was filtered and washed with ether to afford
52 mg of 2. The filtrate was concentrated to leave a paste that
was washed with THF and acetonitrile to give another 120 mg
white solid. The combined fractions were purified by dissolu-
tion in ethanol (5 mL), precipitation by addition of acetonitrile
(50 mL), concentration to 20 mL, addition of more acetonitrile
(30 mL), and so on, until no more precipitation occurred.
Yield: 153 mg (44%) of 3-(ethyl(p-(N,N-dimethylamino)-
phenyl)amino)propyl-1-phosphonic acid. 1H NMR (CD3OD):
7.40 (d, 2H, J ) 9 Hz), 6.83 (d, 2H, J ) 9 Hz), 3.54-3.42 (m,
4H), 3.00 (s, 6H), 1.90-1.67 (m, 4H), 1.08 (t, 3H, J ) 7 Hz).
Solutions of 2 were prepared by addition of 2 equiv of NaOH
to the phosphonic acid in ethanol.
and condensation reactions. Nitric acid (68%, 6 mL) was then
added to bring the pH of the suspension to 1. The suspension
was sonicated using a titanium ultrasonic horn from Bioblock
Scientific Inc. and stirred for 2 h at 80 °C under reflux to achieve
peptization (i.e., destruction of the agglomerates and redispersion
into primary particles). The growth of these particles up to 20
nm was achieved under hydrothermal conditions in a titanium
autoclave with a Teflon jacket heated for 12 h at 230 °C.
Sedimentation took place during the autoclaving, and the
particles were redispersed using an ultrasonic horn (400 W, 15
× 2 s pulses) and readjusting the pH to 1. After two sonications
the colloidal suspension was concentrated in a Rotavapor (35
°C, 30 mbar) to a final TiO2 concentration of 11 wt %. To
prevent cracking during film drying, poly(ethylene glycol) (MW
) 20 000, Merck) was added in a proportion of 50 wt % TiO2.
The resulting paste is stored in a screw-thread glass bottle until
deposition.
The synthetic procedure for the preparation of the ZrO2
colloids and films was similar. Zirconium n-propoxide (6 mL,
13.5 mmol, 70% in propanol) was mixed with acetic acid (0.775
mL, 13.5 mmol) under Ar. Under vigorous stirring, this mixture
was added in one shot to H2O (25 mL) at r.t., resulting in the
immediate formation of a colorless, jelly precipitate. After
stirring for another 60 min, the propanol of the heterogeneous
reaction mixture was distilled off in a Rotavapor. Then, HNO3
(1 mL, 68%) was added, and the mixture was stirred at r.t.
overnight for peptization. The peptized sol had the appearance
of a homogeneous, transparent yellowish liquid. It was diluted
with water to a total volume of 35 mL, transferred to a screw-
capped Teflon container, and autoclaved at 230 °C for 12 h,
which resulted in the formation of a stable, white colloid. The
colloid was further processed as described for TiO2 in order to
obtain viscous, tape-castable pastes. Higher sol concentrations
(15-17 wt %) were necessary though in order to obtain the
appropriate viscosities.
Al2O3 collo¨ıds were prepared as previously described.18
The colloidal pastes were deposited using a simple doctor
blade technique on glass coated with conductive fluorine-doped
SnO2 (Nippon sheet glass, 8-10 ohms/square). The resulting
layer was dried in air at r.t. for 10 min followed by 15 min at
50 °C. The film was then heated to 450 °C at 20-50°/min
and left at 450 °C for 30 min before cooling to room
temperature.
Treatment of the electrodes was effected as follows. To
remove water and organic material adsorbed in the pores during
storage, the electrodes were heated to 350 °C for 30 min in an
air stream. Derivatization was done by immersing the still hot
(100 °C) electrode in a submillimolar solution of the derivatizing
agent in absolute ethanol. After 12 h at r.t., the electrodes were
rinsed with absolute ethanol, dried briefly (1 min) at 100 °C,
and used directly or stored in cyclohexane.
Results
Structure of the Nanocrystalline Layers. The specific
surface area of the TiO2 nanocrystalline electrodes was found
to be 91 m2/g, which gives a crystallite size around 17 nm,
assuming the particles are quasi-spherical. This was confirmed
by scanning electron microscopy (Figure 1). The pore size
distribution (Figure 1) is centered around 20 nm. X-ray
diffraction analysis showed that the particles are composed of
anatase (no rutile detected), and transmission electron micro-
graphy revealed that the main orientation of the lattice planes
at the particle surface is (101). For ZrO2, XRD analysis showed
that the powder consists of the thermodynamically stable
Preparation and Characterization of the Nanocrystalline
Layers. In an argon glovebox, acetic acid (24.2 mL) was added
to titanium isopropoxide (125 mL, Aldrich, 97%). This
modified TiO2 precursor was then added in one shot and at r.t.,
to deionized water (728 mL) under vigorous stirring. A white
precipitate formed instantaneously, and the solution was left
under stirring for 1 h to ensure completion of the hydrolysis

Aromatic Amines Anchored on Metal Oxide Films
J. Phys. Chem. B, Vol. 102, No. 9, 1998 1501
Figure 2. Visible absorption spectra of 1 (---) and 1+ (s) in
acetonitrile/THF (4:1). Oxidation of 1 to 1+ was done in a photoelec-
trochemical cell, on carbon felt, increasing the potential stepwise from
0.7 to 1.2 V, allowing for spectrum stabilization between steps.
Absorption maximum was reached at 1.0 V.
acterized by an extinction coefficient ꢀ and a molecular surface
Smol is given by eq 2.
A ) ηꢀ/(NASmol
)
(2)
The absence of aggregation of the dye on the surface is of
course a prerequisite for proper roughness determination. The
complex 4 fulfills that condition. Its ꢀ is 15 300 M-1 cm-1 at
482 nm, and Smol was calculated to be 200 Å2. The absorbance
at that wavelength of the TiO2 electrodes derivatized with 4
was A ) 0.80 ( 0.05, yielding a roughness factor equal to 630
( 40. For the Al2O3 electrodes, we obtained a value of 520 (
40.
Adsorption Behavior. The adsorption isotherms of 1 were
established by derivatizing electrodes from solutions of different
concentrations, oxidizing the adsorbed molecules chemically by
NOBF4, and measuring the absorbance of the electrodes at 750
nm, where ꢀ of 1+ is 15 200 M-1 cm-1 (see Figure 2). That
wavelength was preferred to the absorption maximum (λmax
)
730 nm, ꢀ ) 24 000 M-1 cm-1) in order avoid too high
absorbance values. The surface concentrations Γ were obtained
by eq 3.
Figure 1. Top: scanning electron micrograph of a nanocrystalline TiO2
film. Bottom: pore size distribution in a nanocrystalline TiO2 film as
determined by nitrogen adsorption, using the BET formalism.
Γ ) A/(ηꢀ)
(3)
monoclinic crystal modification (baddeleyite) with no detectable
amounts of any of the other crystal phases. The specific surface
area of the powders is 121 m2/g as determined by nitrogen
adsorption on a Gemini 2375 apparatus (Micromeretics) using
the BET formalism. This corresponds to an average particle
diameter of about 8.5 nm, assuming spherical, monodispersed
particles with the density of baddeleyite (5.89 g/cm3).
The roughness factor η of the electrodes, defined as the ratio
of the effective surface area of the nanocrystalline layer to its
projected area, was determined by derivatization with a phos-
phonated dye. The ruthenium complex (4,4′,4′′-Me3terpy)Ru-
(terpy-PO3H) (4) was used, where 4,4′,4′′-Me3terpy is 4,4′,4′′-
trimethyl-2,2′:6′,2′′-terpyridine and terpy-PO3H is 2,2′:6′,2′′-
terpyridine-4′-hydrogenophosphonate. The diameter of the
molecule (16 Å), determined by molecular modeling, is close
to the diameter of 1, as measured between the two methoxy
groups. On can thus assume that both molecules will be able
to penetrate more than 90% of the pores, based on the size
distribution shown in Figure 1. The absorbance A of a
nanocrystalline layer bearing a monolayer of molecules char-
The experimental values were fitted to a Langmuir adsorption
isotherm (Figure 3), as expressed by eq 4,
Γ ) Γ°C/(C + 1/K)
(4)
where C is the concentration in solution, Γ° is the surface
concentration in the saturated monolayer, and K is the adsorption
constant. The best fit for TiO2 was obtained with Γ° ) 3.3 (
0.1 µmol/m2 and K ) 26 000 ( 2000 M-1. For Al2O3, the
respective values were 2.9 ( 0.1 µmol/m2 and 33 000 ( 4000
M-1
.
The adsorption constant of 1 on TiO2 is close to that obtained
for the phosphonated viologen 3 (K ) 24 000 M-1).19
Smol ) 1/(Γ°NA)
(5)
Electrochemistry. The radical cations of trianisylamine and
N,N,N′,N′-tetraalkyl-p-phenylenediamines as well as the radical
anions of viologens are known to be fairly stable in solution.20
As expected, the corresponding phosphonated molecules 1, 2,

1502 J. Phys. Chem. B, Vol. 102, No. 9, 1998
Bonhoˆte et al.
Figure 3. Adsorption isotherms of 1 on nanocrystalline TiO2 and Al2O3
layers from ethanolic solutions: measured points (squares) and best
fit of the Langmuir equation (line).
and 3 displayed reversible electrochemical behavior in aceto-
nitrile or propylene carbonate, with the redox potentials given
in the Introduction. The situation was different in the adsorbed
state. Nanocrystalline TiO2 electrodes derivatized with 1, 2,
and 3 were submitted to cyclic voltammetry in propylene
carbonate. The recorded traces, presented in Figure 4a, clearly
show the particular behavior of the triarylamine 1. Integra-
tion of the oxidation wave of 1 gave 13.2 mC, corresponding
to Γ° ) 2.3 µmol/m2, but only 7.53 mC for the reduction
wave, denoting probably a partial desorption of 1+ in the
polar solvent used. In the ambient temperature molten salt
1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide
(EtMeIm+Tf2N-, specific conductivity 8.8 mS cm-1)21 oxidation
and reduction charges were equal. This medium was chosen
for further electrochemical investigations as it grants a satisfac-
tory long-term stability to the monolayer, as shown by the
absence of noticeable degradation established during 100
voltammetric cycles. In acetonitrile containing lithium salt, the
system exhibited also a completely reversible behavior (Figure
4b). Observations were similar on nanocrystalline ZrO2 films
(Figure 4c).
Figure 4. Cyclic voltammograms of nanocrystalline electrodes de-
rivatized by 1, 2, and 3: (a) in propylene carbonate + 0.2 M
NBu4+TfO- at (s) 20 mV/s and (---) 2 mV/s with vertical expansion
7×; (b) and (c) in acetonitrile + 0.5 M Li+Tf2N- at 2 mV/s. (d) Optical
Anson plot for nanocrystalline TiO2 electrodes derivatized with 1, in
EtMeIm+Tf2N-. The absorbance A was recorded as a function of the
square root of time, following a potential step from 0.2 to 1.0 V.
Linearity is observed over 60% of the oxidation process, following
what A tends asymptotically toward Af ) 2.7.
in the presence of lithium ions, the latter lowering the conduction
band edge below the redox potential of the viologen.3
The adsorbed p-phenylenediamine 2 underwent a slow
oxidation at 0.5 V, but the reduction only occurred below -0.3
V. The viologen 3 behaved completely irreversibly. Its
reduction took place only when electrons were injected from
the conduction band of the TiO2, but no oxidation wave was
observed, these electrons being trapped in the reduced viologen,
below the conduction band edge. Reversibility was obtained
The electrodes derivatized with 1 displayed striking electro-
chromism. Typically, their aspect turned from colorless to deep
blue (∆A > 2 between 730 and 770 nm) when the applied
potential changed from 0 to 1 V or more. In polar, highly
conductive solvents such as acetonitrile, propylene carbonate,

Aromatic Amines Anchored on Metal Oxide Films
J. Phys. Chem. B, Vol. 102, No. 9, 1998 1503
or 3-methoxypropionitrile, the process was very fast: coloration
or discoloration occurred in less than 2 s following a potential
step up from 0.2 to 1.0 V or down from 1.0 to -0.15 V,
respectively. For practical electrochromic applications that
require long-term stability, EtMeIm+Tf2N- appears better suited
despite the longer switching time (∆A ) 1.5 in 7 s, ∆A ) 2 in
30 s).
The apparent diffusion coefficients were measured at different
surface concentrations of 1, using a chronospectrophotometric
method derived from the classical chronocoulometric method.22
Following a potential step, the charge Q exchanged with a
modified electrode evolves with time following eq 6, where S
is the electrode surface area and C is the concentration of the
electroactive species in the film. C can be obtained via eq 7,
from the film thickness d and the total charge Qf necessary to
obtain full redox change of the electroactive centers. Substitu-
tion affords eq 8. As this model describes a semiinfinite
diffusion phenomenon, it holds only as long as the oxidative or
reductive front has not reached the limit of the electroactive
layer. Hence, in an Anson plot (Q versus t0.5), the first part is
linear, according to eq 6, followed by a curved region in which
Q asymptotically tends toward Qf.
as in the derivatizing solution. Clearly, Dapp varies with x1, with
a percolation threshold around x1 ) 0.5.
For the experiment of surface dilution without coadsorbent,
TiO2 electrodes were derivatized with 0.005-1 mM solutions
of 1. The surface concentrations were determined as for the
Langmuir isotherm and ranged from 0.35 to 3.3 µmol/m2.
Figure 5b shows the variation of Dapp with the surface
concentration. Here again, it is clear that Dapp varies with Γ,
above a percolation limit situated around 1.7 µmol/m2, corre-
sponding, to 50% of a full monolayer. The fraction of 1 that
was electrochemically oxidizable is plotted in Figure 5d as a
function of the surface concentration. The measure was taken
340 s after the potential step, a delay during which an absorbance
plateau was reached in all the cases. The percolation threshold
is particularly striking in this type of representation. Below
1.7 µmol/m2, almost no molecules of 1 can be oxidized
electrochemically. Above that limit, all the molecules are
accessible.
On Al2O3, similar behavior was observed (Figure 5c). The
wider points distribution and the lower apparent diffusion
coefficients can be attributed to a worse adherence of the
nanocrystalline layer on the conducting glass.
To investigate the correlation between the apparent diffusion
coefficient and the properties of the contacting medium, Dapp
was recorded in a series of blends of EtMeIm+Tf2N- with
various solvents. The values obtained for a full monolayer of
1 with different compositions of a EtMeIm+Tf2N--methoxy-
propionitrile (MPN) mixture are reported in Figure 6a. Dapp
increases rapidely between 1% and 25% of EtMeIm+Tf2N- and
then decreases for higher salt concentrations. Comparison of
the apparent diffusion coefficient measured in 1:1 mixtures of
EtMeIm+Tf2N- with solvents of different dielectric constants
is presented in Table 1. In Figure 6b, the apparent diffusion
coefficients measured in EtMeIm+Tf2N--MPN (1:1) at various
surface concentrations are compared with the values obtained
in the neat liquid salt. Clearly, Dapp is higher in the blend than
in the pure salt, at any concentration above the percolation
threshold.
Q ) 2nFSCDapp0.5t0.5/π0.5
C ) Qf/nFSd
(6)
(7)
Q ) 2QfDapp0.5t0.5/(dπ0.5)
(8)
In the case of an electrochromic film, where the absorbance
change ∆A is proportional to the charge passed, eqs 6 and 8
can be transformed into the corresponding eqs 9 and 10. Optical
equivalents of Anson plots are obtained by measuring ∆A versus
t0.5. Compared to chronocoulometry, chronospectrophotometry
ensures that only the redox phenomena involving the studied
species are recorded. Figure 4d shows a typical optical Anson
plot. The linearity observed for the first half of the absorption
change supports the diffusion control of the process. The
highest apparent diffusion coefficients recorded on TiO2 in
Discussion
The average surface occupied by each molecule of 1 in the
saturated monolayer was calculated by eq 5 to be Smol ) 50 (
2 Å2. That area is particularly small. In fact, if the molecule
would lie horizontally on the TiO2 or if, while perpendicular to
the surface, the triarylamine moiety would be freely rotating,
the molecule would occupy at least 200 Å2. The much smaller
value obtained for Smol indicates a high degree of self-assembly
in the monolayer. If we consider the planar triarylamine
moieties as perpendicular to the surface, with a projected
methyl-methyl distance of 13-16 Å, depending on the position
of these groups, then the interplanar distance should be 3-4 Å
(see Figure 8a), which indeed corresponds to a full stacking of
the molecules.
Both the dependence of the apparent diffusion coefficient on
the surface concentration and the existence of a percolation limit
for the conductivity of the monolayers of 1 indicate that the
charge motion takes place by hopping and not by molecular
diffusion, as depicted in Figure 8. The percolation threshold
(pc) for the conductivity inside a composite material corresponds
to a critical concentration above which an infinite cluster of
conductive sites spans the network. pc is dependent on both
the dimensionality and the coordination number Z of the lattice.24
The observed percolation threshold, at x1 ) 0.5 or Γ ) 0.5Γ°,
is consistent with a two-dimensional triangular (Z ) 6) or square
(Z ) 4) lattice in the site percolation model. The calculated
EtMeIm+Tf2N- were 2.8 × 10-12 m2 s-1
.
∆A ) 2CꢀDapp0.5t0.5/π0.5
(9)
∆A ) 2∆AfDapp0.5t0.5/(dπ0.5)
(10)
To discriminate between a charge transport by molecular
diffusion to the conducting glass and a charge transport by
electron hopping, the apparent diffusion coefficient was mea-
sured as a function of the surface concentration of the electro-
active molecule 1. This surface concentration was controlled
in two ways: in a first series of experiments by coadsorbing a
nonelectroactive molecule and in a second one by varying the
concentration of 1 in solution, in the submonolayer domain of
the Langmuir isotherm.
For the experiment involving a redox-inert coadsorbent,
compound 5 was used as a surface diluant. As it is composed,
like 1, of an aromatic head and of a propylphosphonate tail, it
was supposed to form with 1 homogeneous mixed monolayers
of a similar structure as the pure monolayer. The apparent
diffusion coefficient is plotted in Figure 5a as a function of the
molar fraction x1 of 1 in a solution of 1 and 5. The total
concentration of the two species was always 0.50 mM. The
molar fraction in the monolayer was measured to be the same23

1504 J. Phys. Chem. B, Vol. 102, No. 9, 1998
Bonhoˆte et al.
Figure 5. Dapp for the oxidation of a monolayer of 1 on nanocrystalline films, in EtMeIm+Tf2N-, following a potential step from 0.2 to 1.0 V:
(a) as a function of the molar fraction of 1 (x1) in a mixed monolayer of 1 and 5 on TiO2; (b) as a function of the surface concentrations of 1, in
the absence of coadsorbent, on TiO2; (c) idem on Al2O3. (d) Electrochemically oxidizable fraction of a monolayer of 1 on a TiO2 nanocrystalline
film, measured at different surface concentrations of 1, after 340 s at a potential of 1.0 V.
values for these types of lattices are pc ) 0.5 and pc ) 0.59,
respectively. Those lattices correspond to an electron hopping
possible between each molecule and its four to six closest
neighbors, as illustrated in Figure 8b.
for electron transfers between aromatic molecules in rigid
glasses (γ ) 0.58-0.83 Å).25 This fast damping means that
the electron hopping does not extend significantly beyond the
next neighbor.
In the model of Blauch and Save´ant,9b the percolation
behavior was analyzed as a function of the spatial extension of
the electronic coupling characterizing the hopping process. The
classical exponential decrease of the electron-transfer rate
constant kex(r) with the intermolecular distance r was considered,
according to eq 11, in which k°ex is the rate at the closest
distance δ, and γ is the damping parameter of the electron
transfer.
The structure and electrochemical properties of the adsorbed
molecules seem essential for the appearance of an efficient
lateral charge transport. The reason electron hopping is efficient
in monolayers of 1 is very likely related to the high self-
exchange rate of these amines. Viologens or electrochemically
reversible ruthenium complexes grafted on nanocrystalline layers
do not display lateral charge percolation, at least not to a
comparable extent.26 Self-exchange rate constants in acetonitrile
were reported to be 8.4 × 106 M-1 s-1 for dimethyl
kex(r) ) k°exe-(r-δ)/γ
(11)
viologen2+/+ 27
,
3.5 × 107 M-1 s-1 for Ru(Me2bipy)(hfac)2 (Me2-
bipy ) 4,4′-dimethyl-2,2′-bipyridine, hfac ) hexafluoroacety-
In a classical percolation case, the electron hopping occurs
only between adjacent sites (γ ) 0) and Dapp ) 0 below the
critical fractional loading pc. For γ > 0, electrons can hop
across distances larger than δ and can hence escape from an
isolated cluster, thereby allowing some conductivity of the
material below the percolation threshold. We observed in fact
a nonzero value of Dapp below x1 ) 0.5 (or below Γ ) 0.5Γ°).
In Figure 7, the observed Dapp versus x1 variation is compared
with the predictions of the model for different values of γ/δ in
a square lattice. On the basis of that comparison, we attribute
to our system a value γ/δ ) 0.13 ( 0.05. Assuming δ ) 4 Å,
we get γ ) 0.52 ( 0.20 Å, which is not far from values reported
lacetonate),28 and (5.3-7.7) × 109 M-1 s-1 for N,N,N′,N′-
tetramethyl-p-phenylenediamine+/0 29
The latter value is close
.
to the diffusion-limited rate constant kD ) 2.1 × 1010 M-1 s-1
.
We assume that for triarylamines the self-exchange rate constant
is of the same order of magnitude as for p-phenylenediamines.
Hence, as the conductivity in a monolayer of 2 is much lower
than in a monolayer of 1, it seems that a fast electron self-
exchange in solution is a necessary but not sufficient condition
for efficient lateral charge motion. The fast charge percolation
inside monolayers of 1 must thus be attributed to the almost
planar geometry of the molecules, which allows compact

Aromatic Amines Anchored on Metal Oxide Films
J. Phys. Chem. B, Vol. 102, No. 9, 1998 1505
Figure 7. Variation of log(Dapp) (9) with the molar fraction of 1 (x1)
in a mixed monolayer of 1 and 5 on TiO2 compared with the simulations
of Blauch and Save´ant9b (lines with indicated value of γ/δ).
Figure 6. Dapp for the oxidation of a monolayer of 1 on nanocrystalline
TiO2 films (a) in different methoxypropionitrile-EtMeIm+Tf2N- blends
between 99:1 and 0:100 ratios by volume and (b) as a function of the
surface concentration of 1, in 1:1 methoxypropionitrile-EtMeIm+Tf2N-
(+) and in neat EtMeIm+Tf2N- (b).
TABLE 1: Apparent Diffusion Coefficients in a Full
Monolayer of 1 on TiO2, in Different
EtMeIm+Tf2N--Solvent (1:1) Blends
nanocrystalline
material
Dappc/10-12
solvent
diglymea
ꢀb
m2 s-1
TiO2
TiO2
TiO2
TiO2
TiO2
Al2O3
ZrO2
7
2.6
2.8
5.0
9.1
11
0.035
4.6
EtMeIm+Tf2N-
4-methyl-2-pentanone
3-methoxypropionitrile
acetonitrile
E10
13
20
37
E10
37
EtMeIm+Tf2N-
acetonitrile
a Diethylene glycol dimethyl ether. EtMeIm+Tf2N- is not miscible
with solvents of dielectric constant lower than 7. b Dielectric constant
of the solvent. c Highest measured values are given, assumed to
correspond to the best nanocrystalline films. Estimated error 10%.
Figure 8. Proposed mechanism for a lateral charge transport inside a
monolayer of amines 1 on porous metal oxide on conductive sub-
strate: (a) cross section of a half pore; (b) top view of a monolayer
with possibilities of electron hopping in a square (left) and in a triangular
(right) lattice; (c) scheme of a metal oxide surface bearing various
submonolayer surface concentrations of 1 in a square lattice, with charge
paths opening above the percolation threshold (Γ G 0.5Γ°).
stacking perpendicular to the metal oxide surface and thus
probably a high degree of self-organization in the mono-
layer.
The physical meaning of the apparent diffusion coefficient
in our system can be equivocal and deserves to be discussed.
During the oxidation or reduction of the adsorbed monolayer,
electrons hop from molecule to molecule while ions move at
the monolayer boundary to compensate the charge variation and
further flow through the pores of the metal oxide. Thus, the
apparent diffusion coefficient can reflect the motion of either
the electrons or the ions or a combination of both. From the
observation that the apparent diffusion coefficient strongly
depends on the surface concentration of the electroactive
molecules, one can infer that the resistance of the electrolyte in

1506 J. Phys. Chem. B, Vol. 102, No. 9, 1998
Bonhoˆte et al.
the pores was never limiting in our experiments. In fact, if the
ion motion in the pores would be the slowest process, one would
expect that the apparent diffusion coefficient would remain
independent of the surface concentration over a range comprised
between the full coverage and a value at which the electron
hopping process would become the rate limiting process. This
is obviously not the case, as Dapp continuously varies with Γ
above the percolation threshold. Hence, we conclude that the
conductivity of the solutions was high enough to avoid current
limitation by bulk resistance of that medium. However, it
appears clearly from the entire Figure 6 that the apparent
diffusion coefficient is not only controlled by the surface
concentration of the amine 1 but also depends strongly on the
nature of the contacting medium. Diluting the liquid salt by a
polar solvent leads to an increase of Dapp. What is the relevant
parameter? Both the viscosity31 and the dielectric constant32
of the medium are known to influence the rate of an intermo-
lecular electron transfer in solution. As the proportion of liquid
salt diminishes, the viscosity decreases while, with polar
solvents, the dielectric constant increases. It was in fact shown
that at the molecular level EtMeIm+Tf2N- behaves like a solvent
of low dielectric constant (ꢀ e 10).21 In the present case, solvent
molecules are not supposed to be present between the closely
packed electroactive molecules of the monolayer, and thus, the
viscosity is not expected to affect the rate of electron transfer.
Consistently, similar values of Dapp were obtained in EtMeIm+-
Tf2N--diglyme blends and in the pure liquid salt despite very
different viscosities (5.3 and 34 cP, respectively). The dielectric
constant, on the contrary, does affect the charge-transfer rate
via the solvent reorganization energy which makes an important
contribution to the total activation energy of the process, even
when there are no solvent molecules between an electron donor
and an electron acceptor, like in intramolecular electron
transfers. However, the observed effect is opposite to the
expectation: the solvent reorganization energy increases with
ꢀ, and thus, the electron-transfer rate should decrease, as
formalized in the Marcus theory.32 We observed, on the
contrary, that Dapp increases with ꢀ. The only way to deal with
these apparent contradictions is to postulate that the overall
charge percolation process, measured by Dapp, is comprised of
more than one elementary process. At the molecular level, each
molecule is oxidized or reduced by electron transfer to its
neighbor. That first process is probably the fastest and, hence,
not rate limiting. At the same time, charge compensation must
take place by ion motion at the monolayer boundary. In the
pure EtMeIm+Tf2N- or blends with low-ꢀ solvents, ion pairing
is strong, and hence, the process is slow because pairs need to
be dissociated to afford free ions. With a solvent of high
dielectric constant added, ion pairs are separated, which allows
faster ion motion for charge compensation, leading to higher
values of Dapp. If however the ion concentration decreases too
much (below ca. 1 M in MPN), the advantage of dissociating
the ion pairs is overcompensated by the longer average diffusion
distance to cross for the ion closest to the monolayer. As a
result, Dapp drops, the curved region of the Anson plot
progressively vanishes, and the absorbance of the electrode
ultimately grows linearly with the square root of the time, up
to Af. The whole process of space charge compensation is
obviously rate limiting. The third factor controlling Dapp is
charge percolation at the mesoscopic oxide surface. The
individual electron-transfer step being limited by charge com-
pensation, we can assume that it takes place at the same rate
for every molecule pair at any surface concentration of 1 above
the percolation limit. As a consequence, the origin of the effect
of surface concentration on Dapp is not molecular but very likely
reflects only the density of conducting paths on the surface. In
fact, there cannot be two rate-limiting processes at the molecular
level. If we postulate that charge compensation limits the
intermolecular electron hopping rate, we must admit that the
surface concentration does not restrict it under the same
conditions. However, as both the dielectric constant and the
surface concentration affect Dapp simultaneously, the control by
the latter parameter must take place at a superior scale: the
mesoscopic architecture of the entire monolayer where the
complexity and thus the length of the way a charge has to follow
to reach the electrode depends on the density of open paths.
One can formalize this model by eq 12. It is in that way possible
to account for the existence of two limiting processes affecting
the apparent diffusion coefficient simultaneously.
Dapp ∼ rate of electron transfer ×
density of conducting paths (12)
The apparent rate constant kex for the bimolecular electron-
transfer process can be obtained via eq 13,
kex ) 6Dappdꢀ/(δ2∆Af)
(13)
derived from eq 1, which is assumed to be valid at full fractional
loading (Γ ) Γ°), by substituting the concentration by ∆Af /ꢀd.
With Dapp ) 2.8 × 10-12 m2 s-1 in EtMeIm+Tf2N- and 2 Å <
δ < 4 Å, we get 106 M-1 s-1> kex > 3 × 105 M-1 s-1. The
theoretical diffusion-limited rate constant (kD) in this medium
is 107 M-1 s-1, according to the Smolukovsky-Debye relation-
ship.30 kex is 1 order of magnitude lower than kD, which is
consistent with a rate limitation imposed by charge compensa-
tion. The rate of the pure electron hopping step is probably
much higher as it does not imply any diffusion process. This
value could be obtained in a steady-state conductivity measure-
ment, with a nanocrystalline TiO2 film derivatized with 1
deposited on a microelectrodes array.33 This work is in progress.
Conclusions. The self-organized monolayer of a phospho-
nated triarylamine anchored on a nanocrystalline metal oxide
behaves like an electronically conducting polymer, as illustrated
by the dependence of the apparent diffusion coefficient on the
concentration of the electroactive species and by the existence
of a percolation threshold for the conductivity of the assembly.
The possibility of lateral charge transport at the surface of
nanocrystalline semiconductor layers further broadens the field
of their applications. Not only species having redox levels in
the vicinity of the conduction band of the material can be
electroactive but also species prone to fast electron hopping and
able to form compact self-organized monolayers, in which case
the nanocrystalline semiconductor only acts as an inert sup-
porting material exploited for its high surface area. This opens
the way to new electrochromic, electrocatalytic, and sensor
applications.
Acknowledgment. This work was supported by the Swiss
National Science Foundation (FNRS) and (for S.T.) by the
European TMR program. The authors are grateful to Dr. A. J.
McEvoy for useful discussions.
References and Notes
(1) Gerfin, T.; Gra¨tzel, M.; Walder, L. Prog. Inorg. Chem. 1997, 44,
345-393.
(2) (a) O’Reagan, B.; Gra¨tzel, M. Nature 1991, 353, 737-739. (b)
Nazeeruddin, M. K.; Kay, A.; Rodicio, J.; Humphrey-Baker, R.; Mu¨ller,
E.; Liska, P.; Vlachopoulos, N.; Gra¨tzel, M. J. Am. Chem. Soc. 1993, 115,

Aromatic Amines Anchored on Metal Oxide Films
J. Phys. Chem. B, Vol. 102, No. 9, 1998 1507
6382-6390. (c) Nazeeruddin, M. K.; Liska, P.; Vlachopoulos, N.; Gra¨tzel,
M. HelV. Chim. Acta 1990, 73, 1788-1803.
(3) (a) Hagfeldt, A.; Vlachopoulos, N.; Gilbert, S.; Gra¨tzel, M. Proc.
SPIEsInt. Soc. Opt. Eng. XIII 1994, 2255, 297-304. (b) Hagfeldt, A.;
Walder, L.; Gra¨tzel, M. Proc. SPIEsInt. Soc. Opt. Eng. XIV 1995, 2531,
60-69. (c) Campus, F.; Bonhoˆte, P.; Gra¨tzel, M.; Heinen, S.; Walder, L.
Sol. Energy Mater., in press.
(18) Nu¨esch, F.; Moser, J.-E.; Shklover, V.; Gra¨tzel, M. J. Am. Chem.
Soc. 1996, 118, 5420-5431.
(19) Campus, F. Unpublished data, Ph.D. Thesis, EPFL.
(20) (a) Bard, A. J.; Ledwith, A.; Shine, H. J. AdV. Phys. Chem. 1976,
13, 155-278. (b) Dapperheld, S.; Steckhan, E.; Grosse Brinkhaus, K. H.;
Esch, T. Chem. Ber. 1991, 124, 2557-2567. (c) Steckhan, E. Top. Curr.
Chem. 1987, 142, 3-69.
(4) Shinohara, H.; Gra¨tzel, M.; Vlachopoulos, N.; Aizawa, M.
Bioelectrochem. Bioenerg. 1991, 26, 307-320.
(21) Bonhoˆte, P.; Dias, A. P.; Papageorgiou, N.; Kalyanasundaram, K.;
Gra¨tzel, M. Inorg. Chem. 1996, 35, 1168-1178.
(5) Mayor, M.; Hagfeldt, A.; Gra¨tzel, M.; Walder, L. Chimia 1996,
50, 47-49.
(22) Bard, A. J.; Faulkner, L. R. Electrochemical Methods, Fundamental
and Applications; J. Wiley: New York, 1980; pp 535-536.
(23) The absorbance of the electrode measured after chemical oxidation
by NOBF4 was proportional to the molar fraction x1 in the mixed monolayer,
with a correlation coefficient of 0.992 over seven points.
(24) (a) Sahimi, M. Application of Percolation Theory; Taylor &
Francis: London, 1994; pp 1-22. (b) Bunde, A., Havlin, S., Eds. Fractal
and Disordered Systems; Springer: Heidelberg, 1991; pp 59-171.
(25) (a) Miller, J. R.; Beitz, J. V. J. Chem. Phys. 1981, 74, 6746-6756.
(b) Miller, J. R.; Beitz, J. V.; Huddleston, R. K. J. Am. Chem. Soc. 1984,
106, 5057-5068.
(6) Bonhoˆte, P.; Moser, J. E.; Vlachopoulos, N.; Walder, L.; Zakeer-
uddin, S. M.; Humphry-Baker, R.; Pe´chy, P.; Gra¨tzel, M. J. Chem. Soc.,
Chem. Commun. 1996, 1163-1164.
(7) Finklea, H. O., Ed. Semiconductor Electrodes; Elsevier: Amster-
dam, 1988; pp 108-114.
(8) Majda, M. In Molecular Design of Electrode Surfaces; Murray, J.,
Ed.; Wiley-Interscience: New York, 1992; Chapter IV, pp 159-184.
(9) (a) Blauch, D. N.; Save´ant, J.-M. J. Am. Chem. Soc. 1992, 114,
3323-3332. (b) Blauch, D. N.; Save´ant, J.-M. J. Phys. Chem. 1993, 97,
6444-6448.
(10) See note 10 in ref 9b.
(26) Heimer, T. A.; D’Archangelis, S. T.; Farzad, F.; Stipkala, J. M.;
Meyer, G. J. Inorg. Chem. 1996, 35, 5319-5324.
(11) (a) Dahms, H. J. Phys. Chem. 1968, 72, 362-365. (b) Ruff, I.;
Friedrich, V. J. J. Phys. Chem. 1971, 75, 3297-3302. (c) Ruff, I.; Friedrich,
V.; Demeter, K.; Csaillag, K. J. Phys. Chem. 1971, 75, 3303-3309.
(12) (a) Laviron, E. J. Electroanal. Chem. 1980, 112, 1-9. (b) Andrieux,
C. P.; Save´ant, J.-M. J. Electroanal. Chem. 1980, 111, 377-381. (c) Anson,
F. C.; Blauch, D. N.; Save´ant, J.-M.; Shu, C. F. J. Am. Chem. Soc. 1991,
113, 1922-1932. (d) Fritsch-Faules, I.; Faulkner, L. R. J. Electroanal.
Chem. 1989, 263, 237-255.
(13) Miller, C. J.; Majda, M. J. Electroanal. Chem. 1986, 207, 49-72.
(14) (a) Miller, C. J.; Majda, M. J. Am. Chem. Soc. 1986, 108, 3118-
3120. (b) Miller, C. J.; Widrig, C. A.; Charych, D. H.; Majda, M. J. Phys.
Chem. 1988, 92, 1928-1936. (c) Miller, C. J.; Majda, M. J. Phys. Chem.
1988, 92, 2, 1937-1942.
(15) Shigehara, K.; Oyama, N.; Anson, F. C. J. Am. Chem. Soc. 1981,
103, 2552-2558.
(27) Bock, C. R.; Connor, J. A.; Gutierrez, A. R.; Meyer, T. J.; Whitten,
D. G.; Sullivan, B. P.; Nagle, J. K. Chem. Phys. Lett. 1979, 61, 522-525.
(28) Chan, M. S.; Wahl, A. C. J. Phys. Chem. 1982, 86, 126.
(29) (a) Bock, C. R.; Connor, J. A.; Gutierrez, A. R.; Meyer, T. J.;
Whitten, D. G.; Sullivan, B. P.; Nagle, J. K. Chem. Phys. Lett. 1979, 61,
522-525. (b) Grampp, G.; Jaenicke, W. Ber. Bunsen-Ges. Phys. Chem.
1984, 88, 335-340.
(30) Debye, P. Trans. Electrochem. Soc. 1942, 82, 265-272.
(31) Noyes, R. M. Prog. React. Kinet. 1961, 1, 129-160.
(32) (a) de Schryver, F. C.; Boens, N.; Put, J. AdV. Photochem. 1977,
10, 359-465. (b) de Schryver, F. C.; Collart, P.; Vandenriessche, J.;
Goedeweeck, R.; Swinnen, A. M.; Van der Auweraer, M. Acc. Chem. Res.
1987, 20, 159-166. (c) Borkent, J. H. Dissertation, Universiteit van
Amsterdam, 1976.
(16) (a) Majda, M. Thin Films 1995, 20, 331-347. (b) Charych, D. H.;
Anvar, D. J.; Majda, M. Thin Solid Films 1994, 242, 1.
(17) Katz, H. E.; Bent, S. F.; Wilson, W. L.; Schilling, M. L.; Ungashe,
S. B. J. Am. Chem. Soc. 1994, 116, 6631-6635.
(33) (a) Chidsey, C. E.; Feldman, B. J.; Lundgren, C.; Murray, R. W.
Anal. Chem. 1986, 58, 601-607. (b) Natan, M. J.; Wrighton, M. S. Prog.
Inorg. Chem. 1989, 37, 391-494.