Ultrafast electrochromic windows based on redox-chromophore modified nanostructured semiconducting and conducting films

Article abstract of DOI:10.1021/jp001763b

Described is the construction of an ultrafast electrochromic window. One electrode of this window is based on a transparent nanostructured TiO₂ (anatase) film (4.0 μm thick) supported on conducting glass (F-doped tin oxide, 10 Ω cm^-2, 0.5 μm thick) and modified by chemisorption of a monolayer of the redox chromophore bis(2-phosphonoethyl)-4,4′-bipyridinium dichloride. The other electrode is based on a transparent nanostructured SnO₂ film (3.0 μm thick) supported on conducting glass (F-doped tin oxide, 10 Ω cm^-2, 0.5 μm thick) and modified by chemisorption of a monolayer of the redox chromophore [β-(10-phenothiazyl)propoxy]phosphonic acid. The electrolyte used is LiClO₄ (0.2 mol dm^-3) in γ-butyrolactone. The excellent performance of a 2.5 cm × 2.5 cm window over 10000 electrochromic test cycles-switching times (coloring and bleaching) of less than 250 ms, coloration efficiency of 270 cm² C^-1, steady-state currents (colored and bleached) of less than 6 μA cm^-2, and memory of greater than 600 s (time required for low end transmittance to increase by 5%) - suggest a practical technology.

Full text of DOI:10.1021/jp001763b

J. Phys. Chem. B 2000, 104, 11449-11459
11449
ARTICLES
Ultrafast Electrochromic Windows Based on Redox-Chromophore Modified Nanostructured
Semiconducting and Conducting Films
David Cummins, Gerrit Boschloo, Michael Ryan, David Corr, S. Nagaraja Rao, and
Donald Fitzmaurice*
Department of Chemistry, UniVersity College Dublin, Belfield, Dublin 4, Ireland
ReceiVed: May 11, 2000; In Final Form: September 15, 2000
Described is the construction of an ultrafast electrochromic window. One electrode of this window is based
on a transparent nanostructured TiO2 (anatase) film (4.0 µm thick) supported on conducting glass (F-doped
tin oxide, 10 Ω cm-2, 0.5 µm thick) and modified by chemisorption of a monolayer of the redox chromophore
bis(2-phosphonoethyl)-4,4′-bipyridinium dichloride. The other electrode is based on a transparent nanostructured
SnO2 film (3.0 µm thick) supported on conducting glass (F-doped tin oxide, 10 Ω cm-2, 0.5 µm thick) and
modified by chemisorption of a monolayer of the redox chromophore [â-(10-phenothiazyl)propoxy]phosphonic
acid. The electrolyte used is LiClO4 (0.2 mol dm-3) in γ-butyrolactone. The excellent performance of a 2.5
cm × 2.5 cm window over 10 000 electrochromic test cyclessswitching times (coloring and bleaching) of
less than 250 ms, coloration efficiency of 270 cm2 C-1, steady-state currents (colored and bleached) of less
than 6 µA cm-2, and memory of greater than 600 s (time required for low end transmittance to increase by
5%)ssuggest a practical technology.
Introduction
Second, there are devices that change color upon reduction
or oxidation of a redox chromophore. The redox chromophore
is either dissolved in a solution or incorporated in a polymeric
film deposited on a transparent conducting substrate. In the case
of the viologens, a well-known class of redox chromophore,
their reduction leads to a deep blue or green color. The rates of
coloring and bleaching are, therefore, limited by the rate of
diffusion of the chromophore to the conducting substrates, or
the rate of migration of electrons within the polymeric film. As
a consequence, the switching times of such devices are typically
of the order of tens of seconds even for relatively small area
devices. While a rapidly switching device may be constructed
by adsorbing a monolayer of a redox chromophore at the surface
of a conducting substrate, the extent to which the device colors
is not sufficient for practical applications.
Transparent nanoporous-nanocrystalline TiO2 films on con-
ducting glass referred to as nanostructured TiO2 films, have been
used as photoanodes in efficient dye-sensitized photoelectro-
chemical solar cells.2 This has led to a great deal of activity in
the area of nanostructured semiconductor materials in general
and nanostructured TiO2 electrodes in particular. A characteristic
feature of nanostructured TiO2 electrodes is that they typically
have a porosity of 50% and an internal surface area that greatly
exceeds their geometric area. When used in an electrochemical
system, application of a negative potential leads to a rapid
accumulation of electrons in the nanostructured TiO2 electrode
as charge may be readily compensated by ions adsorbed or
intercalated at or near the electrode-electrolyte interface.
Electrochromic (EC) materials, i.e., materials that change in
color on application of a potential, are the subjects of an
increasing number of reports in the literature and a growing
number of patent applications.1 This increased activity is due,
generally, to a growing awareness of the commercial potential
of EC materials and, specifically, to the fact that a number of
leading manufacturers have now demonstrated the commercial
viability of products incorporating EC materials.
It is still the case, however, that only a few of the commercial
applications foreseen for EC materials, namely rear-view EC
mirrors and large segment static displays, have been brought
to the market. Other potential applications, including large-area
privacy/architectural glazing and high contrast angle-independent
dynamic displays have yet to reach the market. Failure to
commercialize these applications reflects the limitations of
existing EC materials, which may be divided in two classes as
outlined below.
First, there are devices that change color upon intercalation
of small ions into a thin metal oxide film. Examples of metal
oxides used include WO3, V2O5 and TiO2; examples of ions
used include Li+ and Na+. These materials are usually deposited
as amorphous or polycrystalline thin films onto conducting glass
substrates and the ions present in the electrolyte intercalated
under an applied potential. To obtain sufficient optical absorp-
tion, ion-intercalation must extend into the bulk of the metal
oxide film. The rates of coloring and bleaching are, therefore,
limited by the rate of ion transport in to and out of the metal
oxide film. As a consequence, the switching times of such
devices are typically of the order of tens of seconds even for
relatively small area devices.
The rapid accumulation of charge in nanostructured TiO2
electrodes, compensated in an aprotic electrolyte by intercalation
of small ions, leads to formation of blue color centers. The extent
of coloration of a nanostructured TiO2 electrode is, however,
10.1021/jp001763b CCC: $19.00 © 2000 American Chemical Society
Published on Web 11/10/2000

11450 J. Phys. Chem. B, Vol. 104, No. 48, 2000
Cummins et al.
not sufficient for many practical applications.3 To overcome
this limitation, nanostructured TiO2 electrodes have been
modified by chemisorption of a monolayer of a viologen, which
colors blue when reduced by an accumulated electron. This
approach combines the advantages of EC devices based on metal
oxide films with those based on redox chromophores in solution
or within a polymer film.4-6
SCHEME 1: Reagents and Conditions for Synthesis of
Bis(2-phosphonylethyl)-4,4′-bipyridinium Dichloride (I)a
Here we describe the construction of a novel electrochromic
window based on modified nanostructured metal oxide films,
which exhibits unprecedentedly fast switching times.7,8 One
electrode of this window is based on a transparent nanostructured
TiO2 film (anatase, 4.0 µm thick) supported on conducting glass
(F-doped tin oxide, 15 Ω cm-2, 0.5 µm thick) and modified by
chemisorption of a monolayer of the redox chromophore bis-
(2-phosphonoethyl)-4,4′-bipyridinium dichloride. The other
electrode is based on a transparent nanostructured SnO2:Sb film
(3.0 µm thick) supported on conducting glass (F-doped tin oxide,
15 Ω cm-2, 0.5 µm thick) and modified by chemisorption of a
monolayer of the redox chromophore [â-(10- phenothiazyl)-
propoxy]phosphonic acid. The electrolyte used is 0.2 M LiClO4
in γ-butyrolactone. The excellent performance of a 2.5 cm ×
2.5 cm window over 10 000 electrochromic test cycless
switching times (coloring and bleaching) of less than 250 ms,
coloration efficiency of 270 cm2 C-1, steady-state currents
(colored and bleached) of less than 6 µA cm-2, and memory of
greater than 600 s (time required for low end transmittance to
increase by 5%)ssuggest a practical technology.
a (a) Reflux, water, 72 h. (b) 50% HCl, reflux 24 h.
to link this redox chromophore to the surface of the nanostruc-
tured TiO2 electrode prepared as described above10,11 4,4′-
bipyridine (4.4 g) and diethyl-2-bromoethyl phosphonate (15.0
g) were added to water (75 mL). The reaction mixture was
refluxed for 72 h and allowed to cool. Following addition of
concentrated hydrochloric acid (50%, 75 mL) the reaction
mixture was refluxed for a further 24 h. To recover the product,
the reaction mixture was concentrated (to 50 mL), 2-propanol
(200 mL) added dropwise, stirred on ice for an hour and filtered.
The white crystalline product was washed with cold 2-propanol
and air-dried to give the pure redox chromophore bis(2-
phosphonoethyl)-4,4′-bipyridinium dichloride (I) (12.72 g,
84.24% yield).
Experimental Section
Preparation of Transparent Nanostructured TiO2 Elec-
trodes. Transparent nanostructured TiO2 films were deposited
on F-doped tin oxide glass substrates (15 Ω cm-2, 0.5 µm thick,
TEC 15 supplied by Libby-Owen-Ford) as described in detail
elsewhere.2,5 Briefly, a colloidal TiO2 dispersion was prepared
by hydrolysis of titanium tetraisopropoxide and autoclaved at
200 °C for 12 h to yield a dispersion of 10 nm diameter
nanocrystals. Concentrating this dispersion (160 g L-1) and
addition of Carbowax 20000 (40 wt % equivalent of TiO2) yields
a white viscous paste. This paste was spread using a glass rod
on the conducting glass substrate masked by Scotch tape.
Following drying in air for 1 h the film was fired, also in air,
at 450 °C for 2 h. The resulting transparent nanostructured
electrodes are 4 µm thick and possess a surface roughness of
about 1000.
Calculated for I (C14H20N2Cl2O6P2): C, 37.77; H, 4.53; N,
1
6.29. Found: C, 35.09; H, 4.49; N, 6.09. H NMR (water-d2):
δ 2.31-2.43 (m, 4H); δ 4.68-4.80 (m, 4H); δ 8.33 (d,
unresolved meta-coupling, 4H); δ 8.94 (d, unresolved meta-
coupling, 4H).
The modified phenothiazine VIII was prepared as shown in
Scheme 2. As phosphonic acid moieties are chemisorbed at a
metal oxide surface, it is possible to link this redox chromophore
to the surface of the nanostructured SnO2:Sb film described
above.10,11
III: â-(10-Phenothiazyl)propionitrile. Triton B (0.6 mL of a
40% aq soln) was added dropwise to a solution of phenothiazine
(II, 50 g) in acrylonitrile (45 mL) on ice resulting in a vigorous
reaction. The reaction mixture was refluxed for 1 h and allowed
to cool. The resulting crude product was recrystallized from a
30:70 mixture of hot ethanol and acetone to yield orange crystals
of III (31.27 g, and 49.6%).
Preparation of Transparent Nanostructured SnO2 Elec-
trodes. Transparent nanostructured SnO2:Sb films were depos-
ited on F-doped tin oxide glass substrates (10 Ω cm-2, 0.5 µm
thick, TEC 15 supplied by Libby-Owen-Ford) as described in
detail elsewhere.9 Briefly, 20 drops of acetic acid (2.0 mol dm-3
)
were added with stirring to an aqueous dispersion (50 g) of 5
nm diameter Sb-doped SnO2 nanocrystals (15% by wt. SnO2:
Sb, supplied by Alfa). The gel, which formed immediately, was
diluted by addition of water (15 mL) and autoclaved at 200 °C
for 12 h. Addition of Carbowax 20000 (3.75 g) with stirring
for 8 h yields an amber viscous paste, which was diluted with
water (15 mL) to make it suitable for spreading. This paste was
spread using a glass rod on the conducting glass substrate
masked by Scotch tape. Following drying in air for 1 h the film
was fired, also in air, at 450 °C for 1 h. The resulting transparent
nanostructured SnO2:Sb electrodes are 3.0 µm thick and possess
a surface roughness of about 1000.
1H NMR (acetone-d6): δ 2.22-2.62 (t, 2H, J ) 6.0 Hz); δ
3.82-4.04 (t, 2H, J ) 6.0 Hz); δ 6.4-7.00 (m, 8H).
IV: â-(10-Phenothiazyl)propionic Acid. Compound III (31.27
g) was added to a mixed solvent (350 mL methanol, 105 mL
water) NaOH (35 g) solution, refluxed for 15 h and allowed to
cool. The crude product was poured on ice water and acidified
by the addition of dil. sulfuric acid until a white precipitate
formed. The crude product was recrystallized to yield IV (17.0
g, 52.26%).
1H NMR (acetone-d6): δ 2.56-2.80 (t, 2H, J ) 7.2 Hz); δ
4.04-4.28 (t, 2H, J ) 7.2 Hz); δ 6.88-7.20 (m, 8H).
V: â-(10-Phenothiazyl)propionate Ester. Compound IV (17
g) was dissolved in 1:2 by volume mixture of ethanol and
toluene (700 mL) acidified by addition of concentrated sulfuric
Synthesis of Redox Chromophores. The modified viologen
I was prepared as shown in Scheme 1. As phosphonic acid
moieties are chemisorbed at a metal oxide surface it is possible

Ultrafast Electrochromic Windows
J. Phys. Chem. B, Vol. 104, No. 48, 2000 11451
SCHEME 2: Reagents and Conditions for Synthesis of [â-(10-Phenothiazyl)propoxy]phosphonic Acid (VIII)a
a (a) Acrylonitrile, Triton B (40% aq solution), 0 °C. (b) Reflux 1 h. (c) Methanolic sodium hydroxide, reflux 15 h. (d) Ethanol-toluene, concentrated
H2SO4, reflux 12 h. (e) Diethyl ether (dry), LiAlH4. (f) Pyridine-chloroform (dry), phosphorus oxychloride, stir, -15 °C, 2 h. (g) Stir, RT, 1.5 h.
(h) Water, stir, RT, 12 h.
acid (4 mL) and refluxed overnight. The solution was concen-
trated (to approximately 50 mL) and diluted by addition of water
(500 mL). The crude product was extracted into ethyl acetate
(4 × 200 mL), washed with water, dried over MgSO4, filtered
and the solvent removed under reduced pressure. White crystals
of V precipitated from the solution on cooling (11.85 g, 63.9%).
1H NMR (chloroform-d): δ 1.34-1.61 (t, 3H, J ) 7.1 Hz);
δ 2.97-3.21 (t, 2H, J ) 7.4 Hz); δ 4.28-4.65 (m, 4H); δ 7.08-
7.62 (m, 8H).
1H NMR (chloroform-d): δ 2.11-2.27 (m, 2H); δ 3.56-
3.59 (m, 2H); δ 3.98-4.01 (m, 2H); δ 6.86-7.13 (m, 8H).
VIII: [â-(10- Phenothiazyl)propoxy]phosphonic Acid. A
solution of VII (0.9 g) in deionized water (60 mL) was stirred
overnight. The crude product was extracted in ethyl acetate (4
× 50 mL), washed with water, and dried over sodium sulfate.
The white crystals that formed were removed by filtration and
the filtrate recrystallized a further three times to yield the impure
product VIII (0.30 g, 40%). The impurity, also a white
crystalline solid, was determined by 1H NMR, 31P NMR, mass
spectrometry and elemental analysis to be â-(10-phenothiazyl)-
propyl chloride, IX. It is noted that following recrystallization,
the product still contained significant quantities of IX as an
impurity (approximately 80% by mass). It is also noted, since
this impurity does not incorporate phosphonic acid groups it is
not adsorbed at the surface of a nanostructured SnO2:Sb
electrode.
VI: â-(10-Phenothiazyl)propanol. A solution of the compound
V (11.85 g) in dry diethyl ether (33 mL) was added dropwise
to a suspension of LiAlH4 (4.74 g) in dry diethyl ether (70 mL)
and stirred overnight at room temperature. Excess LiAlH4 was
decomposed by the dropwise addition of water and filtered.
Removal of the solvent under reduced pressure gave the green
solid VI (5.57 g, 54.7%).
1H NMR (chloroform-d): δ 2.00-2.05 (m, 2H); δ 3.72-
3.76 (t, 2H, J ) 6.4 Hz); δ 4.00-4.05 (t, 2H, J ) 5.9 Hz); δ
6.92-7.21 (m, 8H).
Calculated for VIII (C15H16O4NSP): C, 53.43; H, 4.76; N,
1
4.15; P, 9.19. Found: C, 63.58; H, 5.42; N, 4.77; P, 1.86. H
VII: [â-(10-Phenothiazyl)propoxy]phosphonic Acid Dichlo-
ride. A solution of VI (1.0 g) and pyridine (1.0 mL) in dry
chloroform (60 mL) was cooled to -15 °C. A solution of
phosphorus oxychloride (4.73 mL) and pyridine (1.0 mL) and
dry chloroform (40 mL) was added dropwise over 0.5 h. The
reaction mixture was stirred at -15 °C for 2 h and the resulting
homogeneous solution allowed to reach ambient temperature
over 1.5 h. The chloroform was removed under reduced pressure
and the crude product washed with toluene (3 × 50 mL) to
remove any unreacted phosphorus oxychloride affording a green
oil VII (0.9 g, 65.2%).
NMR (CDCl3): δ 2.24-2.28 (t, 2H, J ) 6.3 Hz); δ 3.67-3.70
(t, 2H, J ) 6.2 Hz); δ 4.09-4.12 (t, 2H, J ) 6.5 Hz); δ 6.91-
7.19 (m, 8H). 31P NMR (CDCl3): δ 1.69-1.89 (H3PO4); δ
-11.96
Calculated for IX (C15H14NSCl): C, 65.00; H, 5.10; N,
5.10;S, 11.63; Cl, 12.70. Found: C, 64.78; H, 5.07; N, 4.99; S,
1
12.31; Cl, 12.12. H NMR (CDCl3): δ 2.25-2.28 (t, 2H, J )
6.3 Hz), δ 3.66-3.70 (t, 2H, J ) 6.0 Hz); δ 4.07-4.12 (t, 2H,
J ) 6.0 Hz); δ 6.898-7.21 (m, 8H).
All elemental analysis was performed at the Chemical
Services Unit, Department of Chemistry, University College

11452 J. Phys. Chem. B, Vol. 104, No. 48, 2000
Cummins et al.
SCHEME 3: Construction of Electrochromic Window
1
Dublin. H NMR spectra were recorded using a JEOL GPX
Sb electrode was modified by chemisorption of the phenothi-
azine VIII from a chloroformic solution (0.02 mol dm-3) during
6 h, washed with distilled chloroform, dried in air and stored
in a darkened vacuum desiccator for 24 h prior to use.
270 MHz spectrometer at the Chemical Services Unit, Depart-
ment of Chemistry, University College Dublin. 31P NMR were
measured using a Brucker 400 MHz spectrometer at the NMR
Center, Department of Chemistry, Trinity College Dublin.
Preparation and Characterization of Modified Nanostruc-
tured Electrodes. A 2.5 cm × 2.5 cm transparent nanostruc-
tured film of TiO2 or SnO2:Sb was deposited on a 3.9 cm ×
3.9 cm conducting glass substrate as described above. The TiO2
film was modified by chemisorption of the viologen I from an
aqueous solution (0.02 mol dm-3) during 24 h, washed with
distilled 2-propanol, dried in air and stored in a darkened vacuum
desiccator for 24 h prior to use. The SnO2:Sb film was modified
by chemisorption of the phenothiazine VIII from a chloroformic
solution (0.02 mol dm-3) during 6 h, washed with distilled
chloroform, dried in air and stored in a darkened vacuum
desiccator prior to use.
Electrochemical and spectroelectrochemical characterization
of the modified electrode employed the following: A glass
single-compartment 3-electrode cell with a modified nanostruc-
tured TiO2 or SnO2:Sb electrode (geometric surface area 1 cm2)
as the working electrode, a platinum sheet counter electrode,
and a Ag/AgCl/1 M KCl (aq) reference electrode. The electrolyte
was thoroughly deaerated by bubbling with Ar for 30 min prior
to experiments. The cell was incorporated in the sample
compartment of a Hewlett-Packard 8452A diode array spec-
trophotometer and connected to a Solartron SI 1287 potentiostat.
Construction of Electrochromic Window. A 2.5 cm × 2.5
cm transparent nanostructured film of TiO2 or SnO2:Sb was
deposited on a 3.9 cm × 3.9 cm conducting glass substrate.
The TiO2 electrode was modified by chemisorption of the
viologen I from an aqueous solution (0.02 mol dm-3) during
24 h, washed with distilled 2-propanol, dried in air and stored
in a darkened vacuum desiccator for 24 h prior to use. The SnO2:
A cell, with an internal spacing of about 400 µm, was
constructed from a modified nanostructured TiO2 electrode
mounted on a modified nanostructured SnO2:Sb electrode, as
shown in Scheme 3, and sealed using a thermoplastic gasket
(IPBOND 2025, supplied by Industria Plastica Monregalese).
The above gasket had an opening at one corner. The sandwich
structure was evacuated in a modified vacuum desiccator, the
opening dipped in electrolyte solution, and filled by admitting
Ar into the vacuum desiccator. The electrolyte solution consisted
of LiClO4 (0.2 mol dm-3) in γ-butyrolactone. It should be noted
that both the LiClO4 and γ-butyrolactone were carefully purified,
rigorously dried and degassed by bubbling with Ar for 30 min
prior to use. Finally, the cell was sealed using a UV-curable
epoxy resin.
Findings are also reported for EC windows with the following
nominal areas: 0.5 cm × 0.5 cm, 1.0 cm × 1.0 cm, 3.0 cm ×
3.0 cm, and 10 cm × 10 cm. These windows were all assembled
as described above and as shown in Scheme 3.
Testing of Electrochromic Window. A National Instruments
DAQ board in a Macintosh Power PC was used for electro-
chromic cycling. One cycle consisted of applying a voltage of
1.2 V which biases the viologen modified nanostructured TiO2
electrode negative of the phenothiazine modified SnO2:Sb
electrode for 15 s, and a voltage of 0.0 V for 15 s. Optical
absorption spectra were recorded using a Hewlett-Packard
8452A diode array spectrophotometer. A Solartron SI 1287
potentiostat was used to record potential-current characteristics.
All reported testing was done at room temperature.

Ultrafast Electrochromic Windows
J. Phys. Chem. B, Vol. 104, No. 48, 2000 11453
the further observation that, as in Figure 1, the region where
these electrodes overlap (Position C) the EC window colors
blue-red.
To permit a complete assignment of the optical absorption
spectrum of the EC window in the colored state, cyclic
voltammograms (CVs) were measured for both the viologen
modified nanostructured TiO2 electrode and the phenothiazine
modified nanostructured SnO2:Sb electrode, see Figure 3. The
optical absorption spectra of the same electrodes were measured
simultaneously at the indicated applied potentials.
As may be seen from the CV in Figure 3a, all the viologen
molecules chemisorbed at the surface of the nanostructured TiO2
electrode are all singly reduced at applied potentials more
negative than -0.9 V. As may be seen from the corresponding
optical absorption spectra in Figure 3b, due to formation of the
radical cation of the viologen, this electrode colors blue.12 At
applied potentials more negative than -1.0 V all the viologen
molecules chemisorbed at the surface of the nanostructured TiO2
electrode are doubly reduced and the electrode colors a faint
yellow-green.12,13
Figure 1. Optical absorption spectra of the electrochromic window in
Scheme 3 prior to (bleached) and following application of a voltage of
1.20 V (colored) which biases the viologen modified nanostructured
TiO2 electrode negative of the phenothiazine modified nanostructured
SnO2:Sb electrode.
The nanostructured TiO2 electrode used to prepare the EC
window shown in Scheme 3 is 4 µm thick and has a surface
roughness of about 1000 [2]. The concentration of surface Ti4+
sites is approximately 2 × 1013 cm-2 implying; assuming each
viologen is chemisorbed through one phosphonate group, a
surface concentration of 2 × 1013 cm-2 viologens.14 Taking into
account the surface roughness (1000) and the extinction
coefficient of the reduced form of this viologen at 608 nm
(14 000 mol-1 dm3 cm-1),12 this implies an upper limit for
absorbance at this wavelength in the colored state of about 0.5.
Having singly reduced all the viologen molecules chemisorbed
at the surface of the nanostructured TiO2 film in Figure 3b, the
absorbance at 608 nm is increased by 0.5.
As may be seen from the CV in Figure 3c, all the phenothi-
azine molecules chemisorbed at the surface of the nanostructured
SnO2:Sb electrode are singly oxidized at applied potentials more
positive than +0.9 V. As may be seen from the corresponding
optical absorption spectra in Figure 3d, due to formation of the
radical cation form of the phenothiazine, this electrode colors
red. At an applied potential of +1.1 V all the phenothiazine
molecules chemisorbed at the surface of the nanostructured
SnO2:Sb electrode are doubly oxidized and the electrode is a
faint yellow-green in appearance.
Figure 2. Optical absorption spectra measured at the indicated point
of the electrochromic window in Scheme 4 following application of a
voltage of 1.20 V (colored) which biases the viologen modified
nanostructured TiO2 electrode negative of the phenothiazine modified
nanostructured SnO2:Sb electrode.
Results
Optical Absorption Spectra of EC Window. Shown in
Figure 1 are the optical absorption spectra of a 2.5 cm × 2.5
cm EC window assembled as shown in Scheme 3. In the
bleached state the EC window has a slight yellow color due to
the phenothiazine derivative adsorbed at the surface of the
nanostructured SnO2:Sb film. The relatively high absorbance
in the bleached state is due to reflection losses in the multilayer
assembly, which have not been corrected for. Application of a
voltage of 1.2 V, which biases the viologen modified nano-
structured TiO2 electrode negative of the phenothiazine modified
SnO2:Sb electrode, causes the EC window to color blue-red.
The nanostructured SnO2:Sb electrode used to prepare the
EC window in Figure 1 is 3 µm thick and has a surface
roughness of about 1000.9 The concentration of surface Sn4+
sites is assumed to be 2 × 1013 cm-2 implying a surface
concentration of 2 × 1013 cm-2 phenothiazine. Taking into
account the surface roughness (1000) and the extinction
coefficient of the oxidized form of this phenothiazine at 515
nm (8000 mol-1 dm3 cm-1),15 this implies an upper limit for
absorbance at this wavelength in the colored state of about 0.3.
Having reduced all the phenothiazine molecules adsorbed at the
surface of the nanostructured SnO2:Sb film in Figure 3d, the
absorbance at 515 nm is increased by 0.3.
Shown in Figure 2 are the optical absorption spectra of the
EC window assembled as shown in Scheme 4. These spectra
were measured at the indicated positions at an applied voltage
of 1.2 V which biased the viologen modified nanostructured
TiO2 electrode negative of the phenothiazine modified SnO2:
Sb electrode.
On the basis of the CVs and spectra, shown in Figure 3, it is
possible to conclude the following: First, that the viologen
molecules chemisorbed at the nanostructured TiO2 electrode
undergo a one electron reduction to form the corresponding
radical cation and color blue; and second, that the phenothiazine
molecules chemisorbed at the nanostructured SnO2:Sb electrode
undergo a one electron oxidation to form the corresponding
Clearly under the reported conditions the viologen modified
nanostructured TiO2 electrode colors blue (Position A), while
the phenothiazine-modified nanostructured SnO2:Sb electrode
colors red (Position B). Consistent with these observations is

11454 J. Phys. Chem. B, Vol. 104, No. 48, 2000
Cummins et al.
SCHEME 4: Construction of Test Electrochromic Window
radical cation and color red. Together, these observations
account for the blue-red, or essentially neutral, color of the EC
window.
a voltage of 0.0 V to a previously colored device. As may also
be seen from Figure 4a the bleaching time, defined as the
average time taken for the transmittance to increase by two-
thirds of the difference between the steady-state transmittances
in the colored and bleached states, is about 105 ms.
Also shown in Figure 4a are the transmittance transients
measured on coloring and bleaching a 0.8 cm × 0.8 cm and a
8.7 cm × 8.7 cm device. The former show coloring and
bleaching times of 80 and 35 ms, respectively. The latter shows
coloring and bleaching times of 1450 and 530 ms, respectively.
Interestingly, the measured switching times are proportional to
the area of the EC window.
Switching Times of EC Window. The rate of coloration of
a 2.5 cm × 2.5 cm EC window assembled as shown in Scheme
3 was measured following application of a voltage of 1.2 V
which biased the viologen modified nanostructured TiO2 film
negative of the phenothiazine modified SnO2:Sb film. As may
be seen from Figure 4a the coloration time, defined as the time
taken for the transmittance to decrease by two-thirds of the
difference between the steady-state transmittances in the
bleached and colored states, is about 250 ms. The rate of
bleaching of the same EC window was measured by applying
Because of the above windows being assembled by hand,
Figure 3. (a) Cyclic voltammogram of viologen modified nanostructured TiO2 electrode versus Ag/AgCl (1.0 mol dm-3 KCl) in an electrolyte
consisting LiClO4 (0.2 mol dm-3) in γ-butyrolactone. (b) Potential dependent optical absorption spectra of viologen modified nanostructured TiO2
electrode versus Ag/AgCl (1.0 mol dm-3 KCl) in an electrolyte consisting LiClO4 (0.2 mol dm-3) in γ-butyrolactone. (c) As in (a) for phenothiazine
modified nanostructured SnO2:Sb electrode. (d) As in (b) for a phenothiazine modified nanostructured SnO2:Sb electrode.

Ultrafast Electrochromic Windows
J. Phys. Chem. B, Vol. 104, No. 48, 2000 11455
Figure 5. (a) Transient optical absorbance at 608 nm and current
consumption by the electrochromic window in Scheme 3 following
application of a voltage of 1.20 V which biases the viologen modified
nanostructured TiO2 electrode negative of the phenothiazine modified
nanostructured SnO2:Sb electrode or application of a voltage of 0.00
V. (b) Absorbance versus accumulated charge calculated from data in
(a). (c) The persistence of the optical absorbance in (a) under open-
circuit conditions.
Figure 4. (a) Transient optical transmission at 608 nm of electrochro-
mic windows similar to that shown in Scheme 3 versus area following
application of a voltage of 1.20 V which biases the viologen modified
nanostructured TiO2 electrode negative of the phenothiazine modified
nanostructured SnO2:Sb electrode. (b) Coloring times obtained from
transients in (a) and similar data. (c) Bleaching times as in (b).
there is significant variability in these values. It is expected that
much of this variability will be eliminated by automation of
key assembly steps. Nevertheless, the coloring and bleaching
times of a number of different EC windows, whose nominal
active areas were 0.5 cm × 0.5 cm, 1.0 cm × 1.0 cm, 2.5 cm
× 2.5 cm, 3.0 cm × 3.0 and 9.0 cm × 9.0 cm, have been
measured. These times, with the exception of those measured
for the 9.0 cm × 9.0 cm devices, are plotted versus the active
area of the EC window in Figures 4b and 4c. A linear regression
of the above data yields parameters (slope and intercept) that
may be used to predict the average coloring and bleaching times
of any EC window whose active area is known. The average
coloring and bleaching times measured for 9.0 cm × 9.0 cm
EC windows are 3.0 and 0.7 s, respectively. The measured
coloring and bleaching times are, as far as the authors are aware,
the fastest switching times reported for EC windows of their
respective areas.
Coloration Efficiency of EC Window. The peak and steady-
state currents of a 2.5 cm × 2.5 cm EC window, assembled as
shown in Scheme 3, have been measured during coloring and
bleaching and are shown in Figure 5a.
The peak and steady-state currents measured on coloring are
6.5 mA cm-2 and 5 µA cm-2, respectively. The peak and steady-

11456 J. Phys. Chem. B, Vol. 104, No. 48, 2000
Cummins et al.
TABLE 1: Stability of Window under Electrochromic
Cyclinga
number of electrochromic cyclesb
transmittance in bleached state (%)
transmittance in colored state (%)
coloring time (ms)
1
10 100 1000 10000
64 61 67 57
13 12 17 14
460 443 605 448 422
245 270 215 265 212
10 12
16 17 17 12
33 28 17 13
64
23
bleaching time (ms)
peak coloring current (mA cm-2
)
7
12
9
peak bleaching current (mA cm-2
steady-state colored current (µA cm-2
steady-state bleached current (µA cm-2
)
11
15
1
)
)
1
2
1
2
a This test was performed under ambient conditions on a 2.5 cm ×
2.5 cm device assembled as shown in Scheme 3. b Each electrochromic
cycle involved applying a voltage of 1.2 V which biased the viologen
modified nanostructured electrode negative of the phenothiazine
modified electrode for 15 s and then applying a voltage of 0.0 V for
15 s.
Figure 6. Transient optical absorption at 608 nm by the electrochromic
window in Scheme 3 during electrochromic cycling (application of a
voltage of1.20 V which biases the viologen modified nanostructured
TiO2 electrode negative of the phenothiazine modified nanostructured
SnO2:Sb electrode for 15 s, followed by application of a voltage of 0.0
V for 15 s after 1 and 10, 000 cycles.
technology for which results have been reported above is
discussed in detail.
Existing Nanomaterials Based EC Window Technologies.
The electrochromic properties of nanostructured TiO2 electrodes
were first reported by Marguerettaz et al.4 These authors
described the spectroelectrochemical properties of a nanostruc-
tured TiO2 electrode modified by chemisorption of a monolayer
of viologen molecules in a protic electrolyte. At negative applied
potentials electrons were accumulated in the nanostructured TiO2
electrode, the chemisorbed viologen molecules were reduced,
and the electrode colored a deep blue.
state currents measured on bleaching the same EC window are
also 6.5 mA and 5 µA cm-2, respectively.
The coloration efficiency CE(λ) at 608 nm, defined by eq 1,
was determined from the slope of the plot of the increase in
absorbance ∆A(λ) versus the charge accumulated in the device
∆Q, see Figure 5b. The measured CE (608 nm) is 270 C-1 cm2.
∆A(λ)
∆Q
EC window devices based on a nanostructured TiO2 film were
first reported by Cinnsealach et al.5,6 These authors described
an EC window assembled, as shown in Scheme 5, from one
electrode consisting of a viologen modified nanostructured TiO2
film supported on conducting glass, a second electrode consist-
ing of conducting glass, and a liquid electrolyte consisting of
LiClO4 and ferrocene in γ-butyrolactone. On application of a
voltage which biased the nanostructured TiO2 electrode negative
of the conducting glass electrode, the viologen molecules (redox
chromophores) chemisorbed at the surface of the nanostructured
TiO2 electrode were reduced. At the same time a fraction of
the ferrocene molecules (redox promoters) present in the
electrolyte were oxidized at the conducting glass electrode.
CE(λ) )
(1)
Both the above peak and steady-state currents are very low and
suggest that the power consumption of the EC window will be
low and that it should have a long-term memory.
Concerning power consumption, a 2.5 cm × 2.5 cm EC
window will have an associated steady-state current of 31 µA
in the colored state. This implies that the rate of charge
consumption is 3.1 × 10-5 C s-1 or 5.0 × 1014 electrons s-1
.
Concerning the long-term memory, if a voltage of 1.2 V is
applied to the EC window for 60 s and the circuit opened the
EC window first colors and then bleaches on the time scale of
hours. More quantitatively, and as may be seen from Figure
5c, the absorbance of the EC window measured at 608 nm takes
about 3 h to return to the initially measured value, while the
time required for the minimum transmittance in the colored state
to increase by 5% is 600 s.
In the EC window described above, the redox chromophore
is chemisorbed at the surface of a nanostructured TiO2 film.5,6
As a consequence, the rate of reduction of the redox chro-
mophore is not limited by the rate of diffusion of this molecule
to the electrode. This is generally the rate-limiting step in
commercial EC windows based on the reduction of solution
phase redox chromophores. Furthermore, the charge accumu-
lated in the nanostructured electrode is readily compensated by
ions adsorbed at the surface of the constituent nanocrystals of
the nanostructured film.16 As a consequence, the rate of
reduction of the redox chromophore is not limited by the rate
of ion intercalation in the bulk electrode material. This is
generally the rate-limiting step in commercial EC windows
based on ion intercalation in a compact metal oxide film. It is
assumed that the rate-limiting step in the EC window in Scheme
5 is diffusion of the redox promoter to the anode. The fact that
the oxidized redox promoter can diffuse to the modified
nanostructured electrode and accept electrons from the reduced
redox chromophore accounts for the fact that the above EC
window exhibits a significant leakage current and is self-
erasing.5,6 The fact that the redox chromophores are in effect
stacked, the surface roughness of the nanostructured TiO2 film
being about 1000,2 accounts for the fact that the electrode colors
a deep blue upon application of a suitable voltage. In short,
Stability of EC Window. The stability of a 2.5 cm × 2.5
cm EC window, assembled as shown in Scheme 3, was tested
under ambient conditions by subjecting it to 10 000 electro-
chromic cycles. Each electrochromic cycle consisted of applying
a potential of 1.2 V, which biases the viologen modified
nanostructured TiO2 electrode negative of the phenothiazine
modified nanostructured SnO2:Sb electrode, for 15 s and
applying a voltage of 0.0 V for 15 s, see Figure 6a. The
parameters used to characterize cell performance were measured
after 1, 10, 100, 1000 and 10 000 electrochromic cycles and
are summarized in Table 1.
Generally, the findings summarized in Table 1 establish that
a 2.5 cm × 2.5 cm EC window, assembled as shown in Scheme
3, is relatively stable under ambient laboratory conditions is
relatively stable over 10 000 cycles.
Discussion
Existing nanomaterials based EC window technologies are
discussed in outline. The novel nanomaterials based EC window

Ultrafast Electrochromic Windows
J. Phys. Chem. B, Vol. 104, No. 48, 2000 11457
SCHEME 5: Electrochromic Window Designed by Cinnsealach et al.5,6 Based on Viologen Modified Nanostructured
TiO2 and Conducting Glass Electrodes
SCHEME 6: Electrochromic Window Designed by Campus et al.18 Based on Viologen Modified Nanostructured TiO2
and Prussian Blue (Electrodeposited on Conducting Glass or on Nanostructured TiO2 on Conducting Glass) Electrodes
this self-erasing EC window is characterized by relatively fast
switching times and low transmissivity in the colored state.
In a subsequent development of this technology, Cinnsealach
et al. described differently colored (blue and green) EC windows
based on a series of modified viologens which could be
chemisorbed at the surface of a nanostructured TiO2 film.5,6
Boehlen et al. also described a modified viologen, which could
be chemisorbed at the surface of a nanostructured TiO2 film
and which colored (pink) on being reduced.17
In a further development of this emerging technology Campus
et al. described an EC window assembled, as shown in Scheme
6, from an electrode consisting of a viologen modified nano-
structured TiO2 film supported on a conducting glass, a second
electrode consisting of either a thin-film of Prussian Blue
electrodeposited on conducting glass or a very thin-film of
Prussian Blue electrodeposited on a nanostructured TiO2 film,
and a liquid electrolyte consisting of LiN(SO2CF3)2 in glu-
taronitrile.18,19 On application of a voltage which biased the
viologen modified nanostructured TiO2 electrode negative of
the Prussian Blue modified conducting glass or nanostructured
TiO2 electrode, the viologen molecules (redox chromophores)
were reduced. At the same time the Prussian Blue molecules
(redox chromophores) electrodeposited at the surface of the
second nanostructured TiO2 electrode were oxidized.
ions in the bulk electrode material. The second redox chro-
mophore, namely the Prussian Blue, is electrodeposited as a
film on a conducting glass substrate. As a consequence the rate
of oxidation of the Prussian Blue molecules is limited by the
rate of hole diffusion in the bulk film. Alternatively, the Prussian
Blue is electrodeposited as a thin film on the surface of a
nanostructured TiO2 electrode. While this might be expected
to increase the rate of oxidation of this molecule, the absence
of a population of acceptor states isoenergetic with the donor
states of the redox chromophore inhibit electron transfer and,
in fact, results in a device with slower switching speeds.18 The
fact that neither of the redox chromophores, either viologen or
Prussian Blue, are present in solution ensures that the colored
states persist and that this EC window has a long-lived memory.
In short, this nonself-erasing EC window is characterized by
relatively slow switching times and very low transmissivity in
the colored state.
New Nanomaterials Based EC Window Technology. In an
attempt to combine the best features of the EC windows
described above the 2.5 cm × 2.5 cm EC window shown in
Scheme 3 has been assembled.7,8 This EC window consists of
a nanostructured TiO2 electrode modified by chemisorption of
a viologen. The above electrode consists of a network of
interconnected anatase nanocrystals supported on a conducting
glass substrate.4 As there are Ti atoms, which require additional
oxygen atoms to complete their coordination shell at the surface
of these nanocrystals, oxygen-containing ligands that act as
Lewis bases are strongly chemisorbed.10,11 As a consequence,
the viologen I is strongly chemisorbed at the surface of the
nanostructured TiO2 electrode.
In the above EC window one redox chromophore, namely
the viologen, is chemisorbed at the surface of a nanostructured
TiO2 film.18 As a consequence, the rate of reduction of the redox
chromophore is not limited by the rate of diffusion of this
molecule to the electrode. Furthermore, the charge accumulated
in the nanostructured electrode is compensated by ions adsorbed
at the surface of the constituent nanocrystals of the nanostruc-
tured film.16 As a consequence, the rate of reduction of this
redox chromophore is not limited by the rate of diffusion of
The EC window in Scheme 3 also consists of a nanostructured
SnO2:Sb electrode modified by chemisorption of a phenothi-
azine. This electrode consists of a network of interconnected

11458 J. Phys. Chem. B, Vol. 104, No. 48, 2000
Cummins et al.
SCHEME 7: Energy Level Diagram of Electrochromic Window in Scheme 7
nanocrystals supported on a conducting glass substrate.9 As there
are undoubtedly Sn atoms which require additional oxygen
atoms to complete their coordination shell at the surface of these
nanocrystals, oxygen containing ligands that act as Lewis bases
are strongly chemisorbed. As a consequence, the phenothiazine
VIII is strongly chemisorbed at the surface of the nanostructured
SnO2:Sb electrode.
by the conduction band states of the nanostructured TiO2
electrode at which these molecules are adsorbed.
The highly doped nature of the SnO2:Sb electrode means that
there is a sufficiently high density of states present in the
intraband region to ensure this material is highly conducting.9
The oxidation potential of the phenothiazine VIII adsorbed at
the surface of the nanostructured SnO2:Sb film has been
measured in solution (+0.81 V).22 As may also be seen from
Figure 3, phenothiazine oxidation is observed at potentials close
to that of the oxidation potential of the molecule in solution.
Clearly, phenothiazine oxidation is mediated by the intraband
band states of the nanostructured SnO2:Sb electrode at which
these molecules are adsorbed.
The electrolyte used in the present device is γ-butyrolatone
containing added LiClO4 (0.2 mol dm-3). This electrolyte has
the advantages of a low freezing-high boiling solvent that is
stable to photodecomposition at the surface of nanostructured
metal oxide electrodes. This is not the case for solvents more
typically employed in EC windows such as ethylene carbonate
and propylene carbonate.20 The requirement for a redox
promoter in the electrolyte is eliminated by adsorbing redox
chromophores at the surface of the nanostructured TiO2 and
SnO2:Sb electrodes, which are reduced and oxidized, respec-
tively. This ensures, in turn, that this EC window is characterized
by low power consumption and an extended memory.
As stated, on applying a voltage which biases the viologen
modified nanostructured TiO2 electrode 1.2 V negative of the
phenothiazine modified nanostructured SnO2:Sb electrode the
viologen is reduced to the corresponding radical cation, while
the phenothiazine is oxidized to the corresponding radical cation.
Since reduction of the viologen is efficiently mediated by the
conduction band states of the nanostructured TiO2 film, and
since oxidation of the phenothiazine is efficiently mediated by
the intraband states of the Sb doped SnO2 nanostructured film,
the coloring (and bleaching) times of the EC window in Scheme
3 are very rapid as may be seen from Figure 4. It is clear,
therefore, that doped metal oxide nanostructured films will be
important in constructing ultrafast EC window and display
technologies.
As may be seen from Figure 1, application of a voltage of
1.2 V which biases the viologen modified nanostructured TiO2
electrode negative of the phenothiazine modified nanostructured
SnO2:Sb electrode causes the EC window in Scheme 3 to
develop a blue-red color. As may be seen from Figure 2, this
effect has its origin in the fact that under these conditions the
viologen modified nanostructured TiO2 electrode colors blue,
while the phenothiazine modified nanostructured SnO2:Sb
electrode colors red. The mechanism of operation of the EC
window outlined above is described in more detail below. These
findings are summarized in Scheme 7.
As the viologen and phenothiazine molecules are chemisorbed
at the surface of nanostructured TiO2 and SnO2:Sb films, the
principal source of the leakage current in the device shown in
Scheme 4, namely diffusion of the oxidized redox promoter to
the reduced redox chromophore, is eliminated. Consequently,
as may be seen in Figure 5, the steady-state current measured
in the colored and bleached states are very low and endows the
device with a memory.
The redox potential of the conduction band edge of the
nanostructured TiO2 film at the semiconductor-liquid electrolyte
interface is -0.9 V in an electrolyte consisting of γ-butyrolac-
tone containing added LiClO4 (0.2 mol dm-3). This value is
determined from measurements of the onset of the absorption
assigned to electrons accumulated in the conduction band of
the nanostructured TiO2 film (not shown), and is consistent with
values obtained for similar electrodes in electrolytes consisting
of an aprotic solvent containing added LiClO4.21 The reduction
potential of viologen I adsorbed at the surface of a nanostruc-
tured TiO2 film has been measured in solution (-0.61 V).22 As
may be seen from Figure 3, viologen reduction is observed at
potentials close to that of the conduction band edge and,
therefore, at potentials more negative than that of the reduction
potential of the viologen. Clearly, viologen reduction is mediated
Finally, as may be seen from Figure 6, the EC window in
Scheme 3 is relatively stable, both under electrochromic cycling
and under prolonged coloration.
Conclusions
Using a viologen modified semiconducting nanostructured
metal oxide film as one electrode, and a phenothiazine modified
conducting nanostructured metal oxide electrode it has proved
possible to construct an ultrafast electrochromic window.

Ultrafast Electrochromic Windows
J. Phys. Chem. B, Vol. 104, No. 48, 2000 11459
(8) Fitzmaurice, D.; Rao, S. N.; Cummins, D.; Corr, D.; Boschloo, G.
Irish Pat. Appl. 1999.
The ultrafast nature of this window is a direct consequence
of combining a semiconducting nanostructured metal oxide
electrode modified by a redox chromophore which colors upon
being reduced with a conducting nanostructured metal oxide
electrode modified by a redox chromophore which colors upon
being oxidized.
(9) Boschloo, G.; Fitzmaurice, D. J. Phys. Chem. B 1999, 103, 3093.
(10) (a) Ko¨lle, U.; Moser, J.; Gra¨tzel, M. Inorg. Chem. 1985, 24, 2253.
(b) Frei, H.; Fitzmaurice, D.; Gra¨tzel, M. Langmuir 1990, 6, 198. (c) Moser,
J.; Punchihewa, S.; Infelta, P.; Gra¨tzel, M. Langmuir 1991, 7, 3012. (d)
Redmond, G.; Fitzmaurice, D.; Gra¨tzel, M. J. Phys. Chem. 1993, 97, 6951.
(e) Moser, J.; Gra¨tzel, M. Chem. Phys. 1993, 176, 493.
(11) (a) Pe´chy, P.; Rotzinger, F.; Nazeeruddin, M.; Kohle, O.; Zakeer-
uddin, M.; Humphry-Baker, R.; Gra¨tzel, M. J. Chem. Soc., Chem. Commun.
1995, 65. (b) Yan, S.; Hupp, J. J. Phys. Chem. 1996, 100, 6867.
(12) (a) Hunig, S.; Berneth, H. Top. Curr. Chem. 1980, 92, 1. (b) Bird,
C.; Kuhn, A.; Chem. Soc. ReV. 1981, 10, 49.
Acknowledgment. The relevant patent application was
funded jointly by Enterprise Ireland and the University Industry
Program at UCD.
(13) (a) Kok, B.; Rurainski, H.; Owens, O. Biochim. Biophys. Acta 1965,
109, 347. (b) Trudinger, P. Anal. Biochem. 1970, 36, 222. (c) Wantanabe,
T.; Honda, K. J. Phys. Chem. 1982, 86, 2617.
(14) Siripala, W.; Tomkiewics, M. J. Electrochem. Soc. 1982, 129, 1240.
(15) The extinction coefficient of this compound was determined using
standard spectroelectrochemical techniques.
(16) Boshcloo, G.; Fitzmaurice, D. J. Phys. Chem. B 1999, 103, 7860.
(17) Boehlehn, R.; Felderhoff, M.; Michalek, R.; Walder, L. Chem. Lett.
1998, 8, 815.
References and Notes
(1) (a) Lampert, C. Sol. Energy Mater. Solar Cells 1984, 11, 1. (b)
Large-Area Chromogenics: Materials and DeVices for Transmittance
Control; Lampert, C., Granqvist, C., Eds.; SPIE Optical Engineering
Press: Washington, DC, 1990. (c) Greenberg, C. Thin Solid Films 1994,
251, 81.
(2) O’Regan, B.; Gra¨tzel, M. Nature 1991, 353, 737.
(3) Hagfeldt, A.; Vlachopoulos, N.; Gra¨tzel, M. J. Electrochem. Soc.
1994, 14, L82.
(4) Marguerettaz, X.; O’Neill, R.; Fitzmaurice, D. J. Am. Chem. Soc.
1994, 116, 2629.
(18) Campus, F.; Bonhote, P.; Gratzel, M.; Heinen, S.; Walder. L. Sol.
Energy Mater. Sol. Cells 1999, 56, 281.
(5) (a) Cinnsealach, R.; Boschloo, G.; Rao, S. N.; Fitzmaurice, D. Sol.
Energy Mater. Sol. Cells 1998, 55, 215. (b) Cinnsealach, R.; Boschloo, G.;
Rao, S. N.; Fitzmaurice, D. Sol. Energy Mater. Sol. Cells 1999, 57, 107.
(6) Fitzmaurice, D.; Rao, S. N.; Cinnseleach, R.; Enright, B. Eur. Pat.
Appl. 98/9032735.
(7) Cummins, D.; Boschloo, G.; Ryan, M.; Corr, D.; Rao, S. N.;
Fitzmaurice, D. Science, paper submitted.
(19) Bonhote, P.; Walder, L.; Gratzel, M. Eur. Pat. EP0886804.
(20) Gratzel, M. Private communication.
(21) Enright, B.; Redmond, G.; Fitzmaurice, D. J. Phys. Chem. 1994,
98, 6195.
(22) These reduction and oxidation potentials were measured in the
electrolyte used to fill the assembled EC window under test.