DOI: 10.1002/cphc.

201200073

The Electrochemistry of Nanostructured Titanium Dioxide

Electrodes
Thomas Berger,[a, b] Damin Monllor-Satoca,[a, c] Milena Jankulovska,[a] Teresa Lana-Villarreal,[a]
and Roberto Gmez*[a]
Dedicated to the memory of Dieter M. Kolb, who made important contributions to the field of electrochemistry.

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim &1&
These are not the final page numbers! ÞÞ
Several of the multiple applications of titanium dioxide nano- carrier transfer to solution determines the sign and value of
materials are directly related to the introduction or generation the photocurrent. Furthermore, methods for extracting kinetic
of charge carriers in the oxide. Thus, electrochemistry plays information from open-circuit potential and photocurrent
a central role in the understanding of the factors that must be measurements are briefly presented. Some aspects of the com-
controlled for the optimization of the material for each applica- bination of electrochemical and spectroscopic measurements
tion. Herein, the main conceptual tools needed to address the are also dealt with. Finally, some of the applications of TiO2
study of the electrochemical properties of TiO2 nanostructured nanostructured samples derived from their electrochemical
electrodes are reviewed, as well as the electrochemical meth- properties are concisely reviewed. Particular attention is paid
ods to prepare and modify them. Particular attention is paid to to photocatalytic processes and, to a lesser extent, to photo-
the dark electrochemical response of these nanomaterials and synthetic reactions as well as to applications related to energy
its direct connection with the TiO2 electronic structure, interfa- from the aspects of both saving (electrochromic layers) and ac-
cial area and grain boundary density. The physical bases for cumulation (batteries). The use of TiO2 nanomaterials in solar
the generation of currents under illumination are also present- cells is not covered, as a number of reviews have been pub-
ed. Emphasis is placed on the fact that the kinetics of charge- lished addressing this issue.

1. Introduction Significantly, most of these applications (see Figure 1) are

based on the generation, separation, transport and recombina-
Recent years have witnessed an exponential growth in the tion of charge carriers in the oxide, which points out the im-
number of publications on diverse aspects of the preparation portance of the electrochemical methods in studying the oxide
and characterization of nanoscopic TiO2 materials.[1–3] This has properties relevant in such applications. In addition, these
been caused by the increasing number of applications in methods provide guides for the rational design of the TiO2-
which TiO2 nanomaterials play a central role. In addition, the based nanostructures appropriate in each case. The purpose of
nature of some of these applications also favors the wide- this review is to give an account of the current understanding
spread interest in TiO2 nanomaterials. In fact, promising tech- of the electrochemistry of TiO2 nanostructures, focused mainly,
nologies in the crucial fields of energy generation, accumula-
tion and saving, as well as environmental decontamination,[4–6]
are based on TiO2 nanomaterials. As for energy generation, ti-
tanium dioxide nanomaterials are being intensely investigated
as the electron-conducting phase in third-generation solar
cells. Different configurations are being assayed, including
solid-state and liquid electrolyte systems employing either
dyes or quantum dots as sensitizers.[7–10] Together with photo-
voltaic devices, much effort is being devoted to the generation
of solar fuels by means of photoelectrochemical systems based
on titania.[11–13] The photogeneration of either hydrogen or
simple C1–C2 organic fuels (such as formic acid, methanol, eth-
anol…) is being pursued. As for energy accumulation, TiO2
nanomaterials have found application in the development of
negative electrodes with maximized interfacial areas (superca-
pacitors) for lithium-ion batteries.[14] Finally, some effort is Figure 1. Fields of application of TiO2 nanomaterials that are directly or indi-
being devoted to exploring the use of TiO2 nanomaterials in rectly related to their electrochemical properties.
the development of electrochromic (and to a lesser extent
photochromic) layers with fast response and high coloration
efficiency.[15] These layers could be employed, among other [a] Dr. T. Berger, Dr. D. Monllor-Satoca, M. Jankulovska, Dr. T. Lana-Villarreal,
uses, in the design of smart windows, thus allowing for the Dr. R. Gmez
regulation of natural light and heat in buildings, with the con- Institut Universitari d’Electroqumica
i Departament de Qumica Fsica
sequent energy saving.
Universitat d’Alacant, Apartat 99, 03080 Alacant (Spain)
The use of nanoparticles (NPs) and nanostructures of TiO2 in Fax: (+ 34) 965903537
environmental applications is linked to the high redox activity E-mail: Roberto.Gomez@ua.es
of illuminated TiO2 NPs, which can be employed as heteroge- [b] Dr. T. Berger
neous photocatalysts for the degradation of organic contami- Departamento de Sistemas Fsicos, Qumicos y Naturales
Universidad Pablo de Olavide, 41013 Sevilla (Spain)
nants present in different environments such as water or air. In
[c] Dr. D. Monllor-Satoca
the latter case, the application of TiO2 films on different surfa-
School of Environmental Science and Engineering
ces may result not only in decontamination, but also in self- Pohang University of Science and Technology (POSTECH)
cleaning and antifogging properties.[16] Pohang 790-784 (Korea)

&2& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

but not exclusively, on aqueous media, employed in some of nium electrodes in contact with alkaline aqueous solutions in
the applications mentioned above. 1955, the results being interpreted in the framework of the
Let us put the subject into historical perspective.[17] The elec- solid-state physics theories of the p–n junction.[20] In the late
trochemistry of semiconductors, particularly under illumination, 1960s, the first articles on the electrochemistry of oxides,
has rather ancient precedents. In the first half of the 19th cen- namely ZnO, NiO, SnO2 and KTaO3, appeared.[17] It is worth
tury, Becquerel discovered that illuminating silver halides de-
posited on Pt or Au electrodes caused the appearance of tran-
Milena Jankulovska obtained her Dipl.-
sient currents attributed to the decomposition of the halide,
Ing. degree in chemistry at “St Kiril and
which in the case of AgCl lasted for more than 2 h.[18] In close
Metodius” University in Skopje, Mace-
connection with the development of photography, the first
donia. She received her diploma of ad-
photoelectrochemical experiments on sensitization to the visi-
vanced studies from the University of
ble of AgCl crystals were reported in 1931. In these experi-
Alicante (Spain) in 2010. She is current-
ments a dye adsorbed on the chloride particles was able to
ly completing her Ph.D. at the Univer-
inject electrons upon photoexcitation with visible light.[19]
sity of Alicante, under the direction of
Apart from these precedents, modern photoelectrochemical in-
Dr. Roberto Gmez Torregrosa and Dr.
vestigations only began in the 1950s, fueled by the fast devel-
Teresa Lana-Villarreal. Her thesis re-
opment that the semiconductor industry was experiencing at
search involves development of thin
that time. In fact, the first experiments were done with germa-
films consisting of TiO2 nanostructures
with different morphologies (nanoparticles, nanowires, nanocol-
Thomas Berger received his Ph.D. in umns and nanotubes) and crystalline structures (anatase and rutile)
physical chemistry (2005) from the and investigation of their (photo)electrochemical properties.
Vienna University of Technology (Aus-
tria) where he worked under the direc- Teresa Lana-Villarreal obtained her di-
tion of Prof. E. Knçzinger and Dr. O. ploma in chemistry in 1999 from the
Diwald focusing on the synthesis of in- University of Navarra (Spain). During
sulating and semiconducting nanocrys- her Ph.D., she worked on the photoox-
tals via chemical vapor reaction and idation mechanism of organics on TiO2
on the study of their photoelectronic electrodes at the University of Poitiers
properties. During his Ph.D. studies he (France) and at the University of Ali-
performed research work at the Sur- cante (Spain). After her postdoctoral
face Science Center in Pittsburgh (PA) research work with Prof. A. Hagfeldt at
with Prof. J. T. Yates, Jr. After a postdoctoral stay with Dr. R. Gmez KTH (Sweden) on DSCs, she moved
at the University of Alicante (Spain) he spent a period in the solar back to the University of Alicante
industry. In 2009 he moved to the University Pablo de Olavide in where she became associate professor
Sevilla (Spain) where he presently holds a tenure track position. His of physical chemistry in 2011. Her research activities include funda-
main scientific interests are concerned with the preparation of mental aspects of metal oxide semiconductors and their use in
nanoscopic metal oxides and the application of spectroscopic and/ technological applications.
or electrochemical techniques to the study of the semiconductor/
gas and the semiconductor/solution interface. Roberto Gmez obtained his diploma
in chemistry in 1990 and received his
Damin Monllor-Satoca obtained his doctoral degree from the University of
B.Sc. in Chemistry in 2003 (University Alicante (Spain) in 1994 where he
of Alicante, Spain) and received his worked on the electrochemistry of
Ph.D. in Materials Science in 2010 from single crystals of platinum group
the same university under the guid- metals. After postdoctoral research
ance of Dr. Roberto Gmez. He worked work with Prof. M. J. Weaver at Purdue
on the characterization, modeling and University (USA), he moved back to
(photo)activity optimization of nano- the University of Alicante where he
crystalline semiconductor thin films by has been associate professor of physi-
means of electrochemical and spectro- cal chemistry since 1999 and leads the
scopic methods. He is presently work- research group of Photochemistry and Electrochemistry of Semi-
ing at the Pohang University of Sci- conductors. In 1999 he spent a period at Bath University (UK) with
ence and Technology (POSTECH), South Korea, as a postdoctoral Prof. L. M. Peter and a year later at the National Renewable Energy
researcher with Prof. Wonyong Choi. His research work focuses on Laboratory (USA) with Dr. A. J. Nozik. His research interests include
photoinduced water splitting and pollutant remediation with the electrochemistry of semiconductors and its applications in
modified semiconductor thin films and suspensions. solar fuel generation and photoelectrochemical solar cells.

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &3&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

noting that at that time, the electrodes employed were always first photoelectrochemical study with electrodes composed of
monocrystalline in nature. The oxide single crystals were inten- NPs doped with different transition-metal ions.[41] The Uppsala
tionally doped to confer them the conductivity needed in the group made at that time some seminal contributions on the
electrochemical experiments. However, in 1960, an article ap- theoretical treatment of both I–V curves and photocurrent
peared in Russian on the electrochemical behavior (mainly transients with nanoporous TiO2 electrodes.[42, 43] In this context
open-circuit potential measurements) of oxide thin films the enormous impact of Gerischer’s work in this field should
grown on the corresponding metal. Titanium dioxide was be mentioned. Undoubtedly the way we understand semicon-
among the oxides studied in that early work.[21] ductor electrochemistry is impregnated with his ideas.[44]
Apart from the precedent cited above, the electrochemistry It is worth noting that, in addition to the typical nanoparti-
of TiO2 was born in 1969. After devising a way of preparing n- culate electrodes constituted by immobilized NPs, Bard and
type rutile single crystals, Fujishima et al. performed the first co-workers presented at the beginning of the 1980s a novel
electrochemical experiments with TiO2 electrodes in aqueous electrode configuration adequate for NPs dispersed in a con-
solution (pH 4.7). In the dark, hydrogen was evolved under ductive liquid medium, called “slurry electrode”.[45] In this
negative polarization, whereas no currents were observed system, an inert, large-area collector electrode is introduced
under positive polarization. Under illumination, anodic photo- into a suspension of oxide NPs, which in their collision with
currents associated with the oxidation of water were observed the electrode are able to exchange photogenerated carriers
at potentials lower than that of the O2/H2O couple.[22] Fujishima with it.
et al. also determined for the first time the flatband potential This review is focused on fundamental aspects of the elec-
of a TiO2 electrode.[23] The first article in English appeared in trochemistry of nanostructured TiO2, both in the dark and
1971,[24] shortly before the seminal paper in Nature in 1972.[25] under illumination. Some applications are also included in
The latter work had a high impact at that time as the photo- a summarized way. Morphologically, we can distinguish two
electrolysis of water was demonstrated for the first time. Inter- types of nanocrystalline electrodes: 1) random nanoparticulate
estingly, the first paper on dye photosensitization of TiO2 elec- electrodes, where crystallites are deposited without any partic-
trodes was published by this Japanese group in 1971.[26] An- ular control; and 2) ordered nanostructured electrodes, where
other breakthrough in the electrochemistry of TiO2 occurred in crystallites are deposited or grown to form organized assem-
1975, when an oxide thin film, produced by flame-annealing blies[46] (i.e. nanorods, nanowires, nanotubes, etc.). Both of
a Ti foil, was demonstrated to be almost equally photoactive them are covered in this review. Regarding the TiO2 crystalline
as a single-crystal electrode.[27] structure, most of the results shown here correspond to ana-
More than 20 years elapsed between the appearance of the tase and rutile. In any case, many of the ideas presented in this
first reports on the electrochemistry of monocrystalline TiO2 review are general, and equally applicable to nanostructured
electrodes and those on nanocrystalline TiO2 electrodes. This electrodes made of other oxides and semiconductors.
type of electrode is formed by porous tridimensional networks This article intends to provide the reader with the theoretical
of NPs with an average grain size in the tens-of-nanometer tools needed to fully exploit electrochemical measurements
range, deposited on a conducting substrate (metal or conduct- and technologies. It is addressed not only to electrochemists
ing glass). In 1990, O’Regan et al. published an article on the but also to scientists in other related fields (materials science,
sensitization of a semiconductor transparent membrane, re- photocatalysis,…) who could benefit from an electrochemical
porting the first photoelectrochemical experiments with these background in their research. The structure of the review is as
electrodes.[28] The following year witnessed the publication of follows. First, a brief account of the main methods for prepar-
the seminal paper by O’Regan and Grtzel introducing the ing nanostructured TiO2 electrodes is presented. Next, the fun-
concept of a dye-sensitized solar cell based on a nanocrystalline damental basis for the interpretation of their dark electrochem-
photoanode.[29] In the following years, the first UV/Vis spectroe- istry and photoelectrochemistry is developed. Finally, some of
lectrochemical experiments with these samples were pub- the applications of TiO2 nanostructured samples directly relat-
lished.[30–32] In 1993–1994, nanoporous electrodes were em- ed to their electrochemistry are briefly reviewed.
ployed in the context of electrochemically assisted photocatal-
ysis by the groups of Kamat and Anderson.[33–35] Also, Augus-
tynski and co-workers studied several photooxidation reactions
2. Methods for the Preparation of Nanostruc-
in aqueous media and compared the behavior of compact and
tured TiO2 Electrodes
nanoporous anatase electrodes.[36] At that time, studies fo-
cused on the potential distribution in these electrodes, em- The deposition of TiO2 as thin films has been a subject of in-
ploying different electrochemical techniques, also appeared. tense research over the past few decades. Besides the synthe-
The first articles analyzing the dark voltammetric and impe- sis of NPs, different preparation techniques of nanostructured
dance spectroscopy behavior of TiO2 nanoparticulate electro- TiO2 electrodes have been developed.[1, 47, 48] The nature of the
des date from 1995,[37, 38] while electrolyte electroreflection was synthesized particles (morphology, size, crystalline structure,
used soon after.[39] In 1998, a systematic study was published etc.) and, consequently, their photoelectrochemical properties,
on the effects of different electron and hole acceptors in solu- show a strong dependence on the synthesis route adopted.[49]
tion on the photoelectrochemical properties of nanoparticulate Nowadays, it is still a challenge to prepare electrodes with an
TiO2 electrodes.[40] On the other hand, Wang et al. authored the optimized structure for each particular application.

&4& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

The methods used for electrode preparation can be classi- the suspension,[56, 57] and also there is the possibility of prepar-
fied initially into two groups: those based on presynthesized ing multilayered TiO2 samples when thicker films are required.
TiO2 NPs and those where the TiO2 nanoporous layer is grown A deposition method similar to doctorblading is screen
directly on a conducting substrate (Figure 2). The latter can be printing. This technique allows one to prepare porous films
from TiO2 pastes, with the advantage that it can easily be
scaled up.[58] The composition of the precursor paste is crucial
for the homogeneity, adherence and roughness of the final
TiO2 film.[59, 60] It can be prepared from commercially available
TiO2 nanopowders or by hydrolysis of Ti alkoxides in water. The
technique is depicted in Figure 3. The TiO2 paste is placed on

Figure 2. Outline of the main preparation methods of TiO2 nanostructures.

a metal, but the most frequently used substrate is conducting

glass. This is a glass sheet covered with a thin film of either F-
doped SnO2 (F:SnO2, FTO, fluorine tin oxide) or In2O3 :SnO2 (ITO,
indium tin oxide). The heavy F-doping of SnO2 gives rise to the
metal-like properties of the conducting glass.[50] Its use is par-
ticularly convenient for some applications and experiments
due to its transparency to visible light. The preparative meth- Figure 3. Schematic representation of the screen-printing process. Reprinted
ods of TiO2 can also be classified according to their nature into with permission from [58]. Copyright 2005 Elsevier.
nonelectrochemical and electrochemical (Figure 2). The latter
also include a method based on presynthesized NPs (electro- top of the screen. Afterwards, it is extended using a squeegee
phoresis). This section is ordered according to these ideas, the that is drawn across the screen, by applying pressure and
electrochemical methods being gathered at the end. thereby forcing the paste to pass through the open areas (or
pores) of the screen. The printing quality depends on the com-
position of the paste, pressure and speed of the squeegee,
2.1. Preparation of Thin Films from Presynthesized NPs
among other parameters.[58]
One of the most employed and simple methods for preparing Other routes for preparing TiO2 thin films composed of pre-
TiO2 thin films is the so-called doctor blade technique.[51, 52] This synthesized NPs are spin coating and dip coating.[61, 62] Spin
method is based on the preparation of a concentrated slurry coating allows one to prepare uniform thin films on flat sub-
of TiO2 NPs, which is subsequently spread on a conducting strates.[62] In this case, a relatively large amount of a NP disper-
substrate. This slurry (usually called paste) can be produced sion is deposited on the substrate, which is then rotated at
using commercial or homemade nanopowders. Usually it con- high speed to spread the fluid by centrifugal force. Finally, the
tains some additives such as Triton X-100 (which acts as a sur- substrates are annealed in air.
factant),[53] acetylacetone (which favors the disaggregation of The dip-coating technique[63–65] allows preparation of uni-
the NPs),[54] polyethylene oxide, or polyethylene glycol,[55] form thin films by dipping the substrate in a dispersion con-
which allows better control over the porosity of the electrode. taining TiO2 NPs and organic binders and pulling it out at
The films are formed by drop casting the suspension onto the a slow and uniform rate. This procedure can be repeated to in-
substrate, over an area previously defined with an adhesive crease the film thickness. The organic binders are most of the
tape, and then spreading it with a glass rod. Subsequently, the time ionic species, which allow for a layer-by-layer deposi-
deposited layers are allowed to dry in air, prior to an annealing tion.[66] The thin films are finally thermally annealed.
treatment, typically at 450 8C for 1 h in air. Annealing is, in gen- The procedures described so far can also be used in combi-
eral, a requirement for the preparation methods based on pre- nation with sol–gel methods. The sol–gel process is a wet
formed NPs as it promotes sintering of the NPs among them chemical technique, based on the generation of a colloidal
and with the substrate. During this treatment, the organics TiO2 solution composed of discrete NPs, via the hydrolysis of
coming from the slurry additives are burnt up. The thickness of metal reactive precursors, usually alkoxides in alcoholic/aque-
the films can be controlled by changing the solvent content in ous mixtures.[67] The rates at which hydrolysis and condensa-

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &5&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

tion take place determine the

properties of the final TiO2 nano-
material.[68] A slower and more
controlled hydrolysis typically
leads to smaller particle sizes
and more homogeneous proper-
ties.[69] Depending on the nature
of the precursor solution, the
sol–gel synthesis can yield
single-crystalline NPs at low tem-
perature.[70] The NPs can be de-
posited as described previously
(by spin coating, dip coating,
etc.) and fired to obtain thin
films.
The sol–gel method has some
advantages, such as its reprodu-
cibility, controllability and sim- Figure 4. TEM micrographs (a, b, d and e), and AFM images (c and f) of anatase (a–c) and rutile (d–f) NW electro-
des. Insets: electron diffraction patterns. Reprinted with permission from [83]. Copyright 2012 Elsevier.
plicity. Employing this method
combined with a template-
based synthesis allows one to obtain different TiO2 morpholo- sis of the precursor and the growth of the TiO2 film take place
gies such as nanotubes, nanofibers or nanocolumns.[71–73] at higher temperature and pressure than in the corresponding
chemical bath. The composition of the precursor solution and
its temperature (and pressure) determine the final film struc-
2.2. Directly Grown TiO2 Nanoporous Layers
ture and morphology.[84–86]
Among the wet chemical processes, we would like to highlight Spray pyrolysis deposition is a film processing technique in
chemical bath deposition (CBD) due to its simplicity and low which a source solution containing the Ti precursor is sprayed
cost, besides the capability to achieve the coating of large on the heated substrate.[87] When the source solution is atom-
areas. This method is also known as solution growth, con- ized, small droplets splash and vaporize on the substrate and
trolled precipitation, or simply chemical deposition. The as-pre- leave a dry precipitate that thermally decomposes.[88] This
pared films are of comparable quality to those obtained by method is often used for deposition of compact layers, which
more expensive physical deposition processes requiring so- can be used as blocking layers to hinder the contact between
phisticated instrumentation, such as vacuum systems and the electrolyte and the conducting substrate.[89]
[74–79]
other expensive equipment. Chemical vapor deposition (CVD) has also been used to pro-
CBD methods employ different titanium precursors (TiCl4, duce highly pure and uniform TiO2 films.[90–94] In a typical CVD
TiCl3, Ti(SO4)2, TiF4, Ti(OC3H7)4, Ti(OC4H9)4). The hydrolysis of the process, the substrate is exposed to a Ti volatile precursor
precursor results in TiO2 precipitation in solution and/or depo- (most commonly titanium(IV) isopropoxide and TiCl4), which
sition on a substrate immersed in the bath (generating a film). reacts and/or decomposes on the substrate surface to produce
Depending on the deposition conditions, highly crystalline TiO2 the TiO2. Oxygen or another strong oxidant is often added to
NPs with different sizes and shapes can be obtained.[80] The increase the rate of deposition and the quality of the film.
shape control is attributed to the tuning of the growth rate of Figure 5 shows a sketch of a CVD process. The thickness of the
the different crystal planes of TiO2 NPs by the specific adsorp- film is controlled by the coating time.
tion of species present in the bath on these facets.
In the context of this review, we would like to mention the
CBD method developed by Yamabi and Imai,[81] which allows
for the formation of one-dimensional TiO2 structures directly
grown on a substrate. The crystal phases of the resulting TiO2
thin films can be controlled by pH and/or the nature of the Ti
precursor. Berger et al. have employed this methodology to
prepare rutile and anatase electrodes with the same interfacial
area, composed of 2 nm nanowires (NWs) arranged in bundles
(Figure 4). Concretely, TiOSO4 was used for preparing rutile and
TiF4 for anatase NWs.[82, 83]
A variation of the CBD process used to synthesize TiO2 nano-
particulate films is hydrothermal (solvothermal) synthesis. The
synthesis is carried out in a stainless steel autoclave with Figure 5. Schematic diagram of the CVD process for fabricating TiO2 thin
a Teflon liner to be used in a furnace. In this case, the hydroly- films. Reprinted with permission from [95]. Copyright 2009 Elsevier.

&6& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

2.3. Electrochemical Methods O2 ions from H2O. The system is under constant applied po-
tential, and thus the field within the oxide is progressively re-
2.3.1. Electrodeposition
duced as the oxide layer thickness increases, that is, the pro-
Electrodeposition is a useful tool for the preparation of TiO2 cess is self-limiting. The formation of NTs is governed by the
thin films on different conducting substrates. The electrodepo- balance between the anodic growth of a compact oxide layer
sition process is cost effective, environmentally friendly and on the surface of Ti [Eq. (1)] and its chemical dissolution as
allows facile control of the film thickness.[56, 96–98] a consequence of the generation of soluble species, most
Zhitomirsky[99] first reported the cathodic electrodeposition often fluoride complexes [Eq. (2)]:[111]
of TiO2 using a titanium peroxocomplex. Nevertheless, nowa-
days the deposition of TiO2 films has been achieved by both Ti þ 2 H2 O Ð TiO2 þ 4 Hþ þ 4 e ð1Þ
anodic[100] and cathodic[101] electrolysis from solutions contain-
TiO2 þ 6 F þ 4 Hþ Ð ½TiF6 2 þ 2 H2 O ð2Þ
ing precursors such as TiCl3, TiCl4 or TiOSO4.[102]
The as-prepared electrodeposited films are insulating amor-
phous TiIV hydroxides which can be converted into crystalline Figure 7 shows a sketch of the NT growth during anodiza-
TiO2 by heat treatment.[103] Due to the electrically insulating tion under potentiostatic conditions. The NT arrays prepared
nature of the as-deposited layers, only films of limited thick- by anodization are amorphous, and there is a need of thermal
ness and porous structures can be obtained by this method. annealing for triggering crystallization.[112]
However, in applications that require very thin films with high
uniformity, electrodeposition has some clear advantages com-
pared to other deposition techniques.[104] It should be men-
tioned that the addition of surfactants to the working solution
results in the direct deposition of crystalline porous films with-
out further heat treatment.[105]

2.3.2. Electrochemical Anodization

Anodization of Ti is one of the simplest, cheapest and most
straightforward approaches yielding ordered nanostructures.
The process is generally conducted in a two-electrode electro-
chemical cell at a constant applied voltage at room tempera-
ture (Figure 6). The preparation of self-organized porous TiO2

Figure 7. a) Current–time (j–t) characteristics in the absence (a) and pres-

ence (c) of fluoride ions in the electrolyte. I) formation of a compact
oxide layer (CO), II) formation of an irregular porous structure (PO) III) forma-
tion of TiO2 NTs. The inset shows typical linear sweep voltamograms result-
ing in electropolished metal (EP), compact oxide or tube formation.
b,c) Scheme showing the field-aided transport of mobile ions through the
oxide layers in the absence and presence of fluoride ions. Reprinted with
permission from [110]. Copyright 2011 John Wiley and Sons.
Figure 6. a) The electrochemical anodization process of Ti metal (M), b) pos-
sible resulting morphologies and c) highly ordered TiO2 NTs in top view.
Adapted with permission from [110]. Copyright 2011 John Wiley and Sons. Chronologically, the first generation of TiO2 NT arrays was
grown in aqueous HF electrolytes. These NT layers had a thick-
ness not exceeding 500–600 nm.[108] By using buffered neutral
films by anodizing a Ti-based alloy in an acidic fluoride-based electrolytes containing NaF or NH4F instead of HF, NT layers
electrolyte was first reported in 1999 by Zwilling and co-work- with thicknesses higher than 2 mm could be obtained (second
ers.[106, 107] In the last few years Ti anodization has become quite generation).[118] Third-generation NTs were grown in water-free
popular because it can induce the growth of closely packed, electrolytes (ethylene glycol and glycerol). NTs prepared in
vertically aligned nanotubes (NTs; Figure 6). In particular, the these electrolytes show extremely smooth walls and their
groups of Grimes[108, 109] and Schmuki[110] have greatly contribut- length exceeds 7 mm. By a further optimization of electrolyte
ed to the optimization of these one-dimensional TiO2 struc- parameters, NTs as long as 260 mm have been prepared.[113, 114]
tures. Allam et al.[115, 116] have reported on a fourth generation of NTs,
During anodization, the formation of a compact oxide layer grown by Ti anodization using HCl, and HCl in combination
occurs as a consequence of the reaction of metal species with with H2O2 in the electrolyte. The latest studies focused on the

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &7&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

precise control and extension of the NT length, pore size and

Table 1. Summary of the most common preparation routes of TiO2 elec-
wall thickness.[117–119] trodes and some examples of their applications.

Method Morphology of Applications References

2.3.3. Electrophoretic Deposition the nanostruc-
ture
The electrophoretic deposition method is used to deposit pre- Doctor blade According to Photocatalysis [51, 52]

formed TiO2 NPs uniformly on a conducting substrate that preformed Li-ion batteries [123]

[15]
serves as working electrode in a two- or three-electrode cell nanoparticles Electrochromic
whose working solution contains the TiO2 NPs in suspended devices
[124, 125]
Screen printing According to Photocatalysis
form. The method is based on the motion of charged particles preformed Solar cells [59]

under an applied electric field. The deposition rate depends on nanoparticles Gas sensing [126]

[127, 128]
the applied voltage and the properties of the gel or suspen- Dip coating According to Electrochromic
sion.[120–122] The main advantages of the electrophoretic deposi- preformed devices
nanoparticles
tion method are its low cost and the fact that it is relatively Spin coating According to Photocatalysis [61]

fast and reproducible. The film thickness can be readily con- preformed Electrochromic [129]

trolled by changing certain parameters such as the electric nanoparticles devices

[67, 68, 71, 130]
field (applied bias), concentration of electrolyte, working–coun- Sol–gel synthesis Nanoparticles Photocatalysis
[131]
Li-ion batteries
ter electrode distance and deposition time. Electrochromic [132]

A summary of the most common preparation routes for TiO2 devices

[133]
nanoporous electrodes together with some examples of their Chemical bath Nanowires Photocatalysis
[134–136]
applications is shown in Table 1. deposition Li-ion batteries
[85]
Water splitting
[137]
Nanocolumns Sensing appli-
cations
3. Dark Electrochemistry of Nanostructured Solar cells [79, 138]

TiO2 Electrodes Li-ion batteries [139]

[140]
Nanotubes Water splitting
In this section, the fundamental aspects of the dark electro- Electrochromic [141]

chemistry of nanostructured TiO2 electrodes will be discussed. devices

[142]
Li-ion batteries
First, those electrode characteristics facilitating the homogene- [87, 143]
Spray pyrolysis Nanoparticles Photocatalysis
ous charging of mesoporous semiconductor thin films will be Solar cells [89]

highlighted and a quantitative description of the chemical film Chemical vapor Nanoparticles Photocatalysis [90, 95, 144]

[145]
capacitance will be given. Experimental results associated with deposition Electrochromic
devices
charge accumulation in these films will be discussed. Then, the [101, 103–105]
Electrodeposition Nanoparticles Photocatalysis
possibility of addressing the density of states (DOS) by electro- Solar cells [56,96–98, 146]

chemical and spectroelectrochemical approaches will be Li-ion batteries [100]

[110, 114, 116, 119, 147–150]

stressed. Finally, electron-transfer reactions in the dark will be Electrochemical Nanotubes Photocatalysis
[151]
anodization Gas sensing
briefly reviewed. [152, 153]
Solar cells
[110, 133, 154]
Li-ion batteries
[155–157]
Water splitting
3.1. Fermi-Level Control in Nanostructured Electrodes Electrochromic [119, 158]

devices
Electrochemical interfaces usually consist of a solid electron Electrophoretic According to Solar cells [121, 122]

[120]
conductor (electrode) in contact with an ionic solution (electro- deposition preformed Photocatalysis
lyte). When an electron conductor is used as working electrode nanoparticles

in an electrochemical cell, the electrochemical potential, or

equivalently the Fermi level of electrons in the electrode (eF)  1 1
can be modified by changing the applied potential (E) with re- C ¼ Csolid þ CH1 ð4Þ
spect to a reference electrode [Eq. (3)]:

eF ¼ eE þ const: ð3Þ Capacitance is an essential electrical quantity that relates

a change in charge-carrier concentration to a change in poten-
tial (C = dQ/dE). The response of the interface to a change of
where e is the elementary charge. In the case of a concentrated
the electrode potential will thus depend on Csolid/CH. The po-
electrolyte solution (in which the diffuse layer contribution is
tential drop over the interfacial part of the solid (Dsolid ) is
negligible), the interface is defined by a series connection of
given by [Eq. (5)]:[160, 161]
two capacitances, one associated with the solid (Csolid) and the
other with the double layer at the electrode/electrolyte inter-
face (or Helmholtz layer, CH).[159] The total capacitance is then dDsolid CH
¼ ð5Þ
given by [Eq. (4)]: dE CH þ Csolid

&8& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

For Csolid @ CH a change of electrode potential leads mainly to semiconductor. Under depletion, CSC can be obtained by solv-
a change of the electrostatic potential drop across the Helm- ing the Poisson–Boltzmann equation (additional details can be
holtz layer. As dDsolid /dEffiCH/Csolid ! 1, the Fermi level in the found in ref. [159]). Under such conditions, the potential de-
solid remains unchanged with respect to the energy levels of pendence of CSC (per electrode unit area) is given by the Mott–
the solid. This situation is typical of metal electrodes or highly Schottky equation, which for an n-type semiconductor such as
doped semiconductors, where the DOS at the Fermi level is TiO2 becomes [Eq. (6)]:
very high (Figure 8 a). In this case, there is no significant  
change of the charge-carrier concentration in the solid upon 1 2kT eDSC
2 ¼ 22 N e2 kT
1 ð6Þ
a potential change, as the Fermi level is pinned at a fixed CSC 0 D

energy level (Fermi level pinning).

where ND is the donor density,
220 is the permittivity of the
semiconductor and DSC is the
potential difference associated
with band bending. Since the
potential drop across the Helm-
holtz layer is unknown, it is im-
possible to theoretically predict
DSC from energy considerations
exclusively. In this context, ex-
perimental capacity measure-
ments combined with Equa-
tion (6) furnish valuable informa-
tion. Extrapolation of the Mott–

Schottky plot (1 C 2 vs. E) to
 SC
1 2 ¼ 0 yields the electrode
CSC
potential at which the potential
across the space charge region
approaches zero, DSC ! 0. This
potential value is the so-called
flatband potential (EFB) which is
an essential parameter for bulk
semiconductor electrodes.
Nanostructured semiconduc-
tor electrodes comprise a special
Figure 8. Different types of electrodes at open circuit (left) and upon negative polarization (right) in a concentrat- case. These electrodes typically
ed electrolyte. a) Metal electrode, b) compact semiconductor electrode and c) nanocrystalline semiconductor elec-
trode. DH : electrostatic potential drop across the double layer; EC and EV : potentials corresponding to the
consist of a mesoporous net-
bottom CB edge and top VB edge, respectively; EF : potential corresponding to the Fermi level of the electrode, work of low-doped crystallites in
ERed/Ox : redox potential of the electrolyte, Cm : chemical capacitance; CH : double-layer capacitance; e: elementary the nanometer range, permeat-
charge; dH : Helmholtz layer thickness. ed by the electrolyte. The crystal
size in nanocrystalline systems is
commonly so small that band
A different situation is observed for low-doped semiconduc- bending is negligible. In fact, these particles have a diameter
tors where the DOS at the Fermi level is low (zero in an ideal normally smaller than the width of the space charge region
case). The transfer of mobile charges from the semiconductor (W),[162] provided that the particles are not intentionally doped
to the electrolyte induces in this case the formation of an in- (as a lower density of donors, ND, implies a larger value of
terfacial layer where the conduction and valence bands are W[159]). This is the case for randomly organized nanoparticulate
bent (space charge layer). For Csolid ! CH, dDsolid /dEffi1 and electrodes. Nanostructured semiconductor electrodes show in-
a change of electrode potential has the effect of displacing the sulating behavior if the Fermi level lies far below the conduc-
Fermi level in the space charge layer with respect to the tion band (CB) edge. However, when the Fermi level ap-
energy levels of the solid, whereas the energy levels of the proaches the CB edge upon negative polarization (or upon ex-
solid are pinned at the surface (band edge level pinning). Csolid posure to photons with supra-band-gap energy, see Section 4),
may thus be associated with the space charge capacitance, CSC. the nanostructured film can adopt a conducting behavior.[163]
As illustrated in Figure 8 b for a bulk n-type semiconductor in Before approaching a quasi-metallic state there exists a wide
depletion, electrode polarization may induce the transition potential range where the semiconductor film exhibits a certain
from an insulating to a conducting behavior by changing the degree of conductivity, which allows for a homogeneous
electron concentration, at least in the interfacial region of the charging of the electrode and thus for a homogeneous dis-

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &9&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

placement of the semiconductor Fermi level with respect to ceptual. In any case, for ideally polarizable nanostructured
the CB (where dDEsolid/dEffi1, DEsolid = EFEC ; Figure 8 c) as long films in quasi-equilibrium both  and m can be considered ap-
as the band edges are pinned. This range is delimited by the proximately constant throughout the whole film. When the
onset potential for efficient electronic communication between inner potential is not modified upon variation of the electron
the conducting substrate and the semiconductor network, and density, a displacement of the Fermi level induces a variation
by the CB edge, which defines the onset potential of quasi- of the chemical potential of electrons deF = dm. The chemical
metallization (where dDEsolid/dEffiCH/Csolid ! 1). Electron trans- capacitance per unit volume is defined as [Eq. (9)]:
port within the nanostructured film is mainly diffusive and
occurs as a consequence of temporary thermal excitation of dn
Cm;v ¼ e2 ð9Þ
electrons from band-gap states to the CB and/or direct hop- dm
ping between electronic states. Thus, a homogeneous equilib-
rium occupancy of band-gap states corresponding to the ex- Assuming a DOS function D(e) and neglecting many-particle
ternally controlled substrate potential can be established. Elec- effects, that is, electrons are treated as noninteracting entities
tron accumulation in the electrode is compensated by electro- and the energy of a state, therefore, does not depend on the
sorption of ions at the huge internal surface of the film or by electrochemical potential of electrons (one-particle DOS), the
ion absorption/intercalation (mainly for cations such as Li + and carrier density can be calculated according to [Eq. (10)]:
H + ). Charge injection and compensation is thus truly three-di-
mensional in nanostructured electrodes. Zþ1
As mentioned above, Fermi-level control in nanostructured n¼ DðeÞf ðe  eF Þde ð10Þ
electrodes is connected to specific electrode characteristics 1

such as the small size of the building blocks, a low level of

doping, good electronic connectivity (within the nanocrystal Here f(eeF) is the Fermi–Dirac probability distribution func-
network as well as between the film and the conducting sub- tion [Eq. (11)]:
strate) and the presence of a surrounding equipotential surface
(electrolyte permeation). The semiconductor/electrolyte inter- 1
f ðe  eF Þ ¼ ee  ð11Þ
face (SEI) of nanostructured electrodes can thus be described 1 þ exp kT F
by a combination of homogeneous electron accumulation
throughout the film (but not necessarily within each particle) with k being the Boltzmann constant and T the absolute tem-
and partial band unpinning.[164] In this case, Csolid is associated perature. It can be shown that in the zero-temperature approx-
with an intrinsic film capacitance, the chemical capacitance per imation (i.e. assuming that f(eeF) is a step function) the chem-
volume unit (Cm,v). The total capacitance has been proposed to ical capacitance can be expressed by [Eq. (12)]:[166]
result from a series combination of Cm,v and CH,v [Eq. (7)]:
dn
 1 Cm;v ¼ e2 ¼ e2 DðeF Þ ð12Þ
Cv ¼ C 1
þC1
ð7Þ deF
m;v H;v

Corrections to the zero-temperature approximation have

In nanostructured electrodes, the chemical capacitance also been reported.[167] Equation (12) shows that the system re-
often dominates the total capacitance in a wide potential sponds to a perturbation deF by filling a slice of states at the
range (i.e. Cm,v ! CH,v , band edge level pinning). As a conse- Fermi level.
quence, the electronic conductivity in the films strongly de- Equation (12) opens up the possibility of extracting informa-
pends on electrode potential. In the case of nanoporous TiO2 tion on the density of electrochemically accessible electronic
electrodes in aqueous solutions, the conductivity varies over states in a semiconductor electrode. In this case, the DOS is
more than eight orders of magnitude.[165] The contribution to not exclusively determined by the intrinsic semiconductor
the capacitance coming from the substrate/electrolyte inter- properties, but contains additional contributions from persis-
face is being neglected. However, it will dominate the behavior tent or transient states resulting from the interaction of the
of the nanoporous electrode at sufficiently positive applied po- semiconductor with the electrolyte at a given electrode poten-
tentials. tial. Such extrinsic states may result from adsorption at the
The electrochemical potential of electrons is defined as semiconductor surface or intercalation of species in solution
[Eq. (8)]: into the crystal structure.

eF ¼ e þ m ¼ eE þ const: ð8Þ

3.2. Chemical Capacitance of Nanostructured Semiconduc-
where  is the local electrostatic potential (inner or Galvani po- tor Films
tential) and m the chemical potential of electrons. Note that
3.2.1. Conduction-Band States
the latter would be given per electron instead of the typical
per mole. Let us recall that the distinction between the two As shown above, the chemical capacitance is associated with
contributions to the electrochemical potential is entirely con- the DOS in the semiconductor electrode. Fundamentals of

&10& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

semiconductor physics can therefore be applied to gain by [Eq. (20)]:

a quantitative interpretation of this parameter.[159] In a semicon-
4 e  e 3=2
ductor, the thermodynamic equilibrium concentration of CB n ¼ pffiffiffi Nc F c
ð20Þ
electrons per unit volume can be calculated according to 3 p kT
[Eq. (13)]:
The CB electron concentration can be related to the chemi-
Z1 Z1
1c ðeÞde cal capacitance according to [Eq. (21)]:
n¼ 1c ðeÞf ðe  eF Þde ¼ ee  ð13Þ
1 þ exp kT F
ec ec dqv
CB
Cm;v ¼p ð21Þ
dE
where 1c ðeÞ is the DOS in the CB and ec the lower CB edge
energy. Assuming a parabolic band structure, the density of CB where dqv ¼ edn and dE ¼ deF =e with qv being the total
states is given by [Eq. (14)]: volume charge density. We are assuming that the band edges
are pinned, otherwise a change in electrode potential would
1 not translate into an equivalent change in the Fermi level. The
1c ðeÞ ¼ ð2mc Þ3=2 ðe  ec Þ1=2 ð14Þ
h3
2p2  factor p accounts for the film porosity and is defined as the
fraction of electrode (film) volume occupied by the semicon-
where mc is the effective mass of the electrons in the CB. The ductor. The capacitance per geometric unit area can be ob-
value of 1c ðeÞde gives the number of electronic states per unit tained by multiplication of Equation (21) by the film thickness
volume corresponding to the energy interval ðe; e þ deÞ. Sub- (d) [Eq. (22)]:
stituting Equation (14) into Equation (13) one obtains
[Eq. (15)]:[159] dn
CB
Cm;s ¼ pe2 d ð22Þ
deF
n ¼ Nc F1=2 ð15Þ

with Nc being the effective DOS in the CB [Eq. (16)]: The capacitance is directly proportional to the derivative of
the electron concentration with respect to the Fermi level posi-
2pmc kT 3=2 tion. For a degenerate semiconductor [fðe  eF Þ @ 0] [Eq. (23)]:
Nc ¼ 2 ð16Þ
hÞ 2
ð2p
CB e2 Nc d eF  ec 1=2
Cm;s ¼ 2p pffiffiffi ð23Þ
kT p kT
The dimensionless function [Eq. (17)]:
Z while for the nondegenerate limit [fðe  eF Þ ! 1] [Eq. (24)]:
2 1
z1=2
F1=2 ðzÞ ¼ pffiffiffi dz ð17Þ
p 0 1 þ expðz  zÞ
e2 Nc d e  e 
CB
Cm;s ¼p exp F c
ð24Þ
kT kT
is the Fermi–Dirac integral with index 1/2, which has been de-
fined using the dimensionless variables z  ðe  ec Þ=kT and
z  ðeF  ec Þ=kT. The evaluation of the Fermi–Dirac integral Equations (23) and (24) can be expressed as a function of
yields for the nondegenerate (z < 0 and j z j @ 1) and the de- electrode potential by substitution of
generate (z @ 1) limits [Eq. (18)]: ðeF  ec Þ=kT ¼ eðEc  E Þ=kT. Recently, Fabregat-Santiago
et al.[164] deduced an equivalent expression for
 the CB capaci-

( CB
expðzÞ; z < 0 and jzj  1 tance. From Equation (24) it follows that dE=d log Cm;s =
F1=2 ðzÞ ¼ 4 ð18Þ 59 mV decade1 (at 25 8C). The logarithmic form [Eq. (25)]:
pffiffi z ;
3 p
3=2
z1
 2 
e Nc d eE e
CB
ln Cm;s ¼ ln p þ c  E ð25Þ
The number of electrons in the CB may therefore be de- kT kT kT
scribed by two limiting functions. If ðeC  eF Þ @ kT, then
fðe  eF Þ ! 1 and the Fermi–Dirac function can be approximat- furthermore allows for the determination of the position of the
CB
ed in the nondegenerate case by the Boltzmann function and CB edge (Ec). By representing ln Cm;s versus E, a straight  line
e2 N d eE
the electron concentration is given by [Eq. (19)]: with a slope of e/kT and a y-axis intercept of ln p kTc þ kTc
e  e  is obtained. From the intercept, Ec can be extracted. Equa-
F c
n ¼ Nc exp ð19Þ tion (25) shows that, in the case of a nondegenerate semicon-
kT
ductor, the onset potential for charge accumulation does not
coincide with the lower edge of the CB. Ec is localized at lower
However, if ðeF  ec Þ @ kT, then fðe  eF Þ @ 0 and the semi- potentials (within the accumulation region). Figure 9 shows
conductor is degenerate. In this case the Fermi level, as for the CB capacitance for the two limiting cases as calculated
metals, lies in the CB and the electron concentration is given from Equations (23) and (24) for the degenerate and the non-

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &11&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

Comparison with Equation (24) shows that the only differ-

ence is the presence of the parameter Tc, instead of the tem-
perature. Note that Equation (27) has been obtained using the
approximation of Equation (12) [i.e. assuming a step function
for the Fermi–Dirac function instead of integrating Equa-
tion (10)]. This means that it was assumed that kT!0 (zero
temperature limit). Equation (27) completely neglects the
effect of the thermal spread of the distribution function in the
region around the Fermi level. Therefore, some features of the
true DOS can be hidden. We can consider it as a good approxi-
mation exclusively in the case that the surface-state distribu-
tion varies smoothly. Corrections to take into account the ther-
mal effects have been proposed by Bisquert and co-work-
ers.[167, 198]
Considering the contributions of both CB and surface-state
Figure 9. Conduction band capacitance as calculated from Equations (23)
electrons, the total capacitance is given by [Eq. (28)]:
and (24) for the degenerate and the nondegenerate limits. The dashed line
CB ss
represents the approximate capacitance in the region where eF is close to ec. Cm;s ¼ Cm;s þ Cm;s ð28Þ
Parameters used: T = 298.15 K, mc/m0 = 9, d = 360 nm, Nc = 6.75  1020 cm3,
p = 0.5.
Additionally to a surface-state distribution in the band gap,
some authors have proposed the existence of monoenergetic
degenerate limits, respectively. The dashed line represents the states or traps at the band gap. In the case of anatase, the
capacitance in the region where eF is close to ec and none of energy of these states coincides with the energy of the onset
the limiting cases yields a valid solution. of the surface-state distribution.[166]
Equation (24) has been verified experimentally with rutile For a monoenergetic state of electrode total volume density
NW electrodes,[83] while the experimental verification of Equa- NGB, the capacitance can be described by [Eq. (29)]:
tion (23) is still awaited. The main problem is that the latter
equation would apply under strong accumulation conditions, NGB e2
GB
Cm;s ¼d f ð1  f Þ ð29Þ
which may be accompanied by partial band edge unpinning. kT
This fact would preclude a straightforward validation of Equa-
tion (23). where f is the average occupancy. Assuming that the latter can
be described by a Fermi–Dirac distribution, the chemical ca-
pacitance associated with the band-gap states can be calculat-
3.2.2. Band-Gap States ed as [Eq. (30)]:
Up to now, only the CB capacitance has been taken into ac-  
eðEE Þ
count. In addition, band-gap states may also contribute to the NGB e2 exp  kT GB
GB
total capacitance. For nanostructured electrodes (of anatase), Cm;s ¼d    ð30Þ
kT 1 þ exp  eðEEGB Þ 2
kT
an exponential distribution of band-gap states just below the
CB edge, which was proposed to be conected to surface states
(ss), has been found to be highly relevant (see Section 3.3). where EGB is the potential corresponding to the GB monoener-
This distribution can be defined by [Eq. (26)]:[168–171] getic state level.
  Electrons trapped at monoenergetic states also contribute
Vt e  ec to the experimental capacitance. Hence [Eq. (31)]:
gss ðeÞ ¼ exp ð26Þ
kTc kTc
CB ss GB
Cm;s ¼ Cm;s þ Cm;s þ Cm;s ð31Þ
where gss ðeÞ is the density of surface states, Vt is the electrode
total volume density of traps and Tc is a characteristic tempera- It is important to reiterate that:
ture that defines the broadening of the exponential distribu-
tion. The broadening can alternatively be described by the co- * The experimental capacitance coincides with the chemical
efficient ac ¼ T=Tc . It has been shown that, as in the case of capacitance only if the bands are pinned (i.e. Cm ! CH). For
the CB capacitance, the capacitance associated with an expo- (partial) band unpinning, corrections should be introduced
nential distribution of band-gap states depends exponentially as, in this case, due to a change of the potential drop
on potential [Eq. (27)]:[164] through the Helmholtz layer, electron potential energy and
  electrode potential no longer scale linearly.
ss e2 Vt d eðEC  E Þ ss
* The expression for Cm;s has been obtained using a zero-tem-
Cm;s ¼ ð27Þ
kTc kTc CB
perature approximation. This is not the case for Cm;s GB
or Cm;s .

&12& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

3.3. Charge Accumulation in Nanostructured TiO2 Electrodes charge extracted during the reverse scan. In the presence of
a significant ohmic drop, a model fitting must be performed to
The capacitance of nanostructured electrodes has been deter- account for the distortion induced by a series resistance.[164] In
mined experimentally by means of current integration at differ- the case of partial band unpinning the capacitance is given by
ential potential steps,[172] cyclic voltammetry (CV),[164, 173] charge Equation (7). When Cm becomes larger than CH the cyclic vol-
extraction,[174–176] impedance spectroscopy,[177, 178] chronoamper- tammogram flattens due to the fact that the bands shift simul-
ometry,[179, 180] intensity modulated photovoltage spectrosco- taneously with a displacement of the Fermi level. In this case
py,[181] electro-optical techniques,[182] spectroelectrochemis- dDEsolid =dE decreases, thus indicating a retardation of electron
try[38, 183–185] and open-circuit photovoltage decay measure- accumulation in favor of charging the Helmholtz layer. This sit-
ments.[186] uation is illustrated in Figure 11, which depicts simulated vol-
tammograms for different ratios of Cm/CH.

Figure 10. a) Cyclic voltammogram of a nanoparticulate TiO2 film in aqueous

solution at pH 2. b) Capacitance of the same sample as determined by impe-
dance spectroscopy. Reprinted with permission from [166]. Copyright 2008
Elsevier.

Figure 10 depicts the typical response of a TiO2 electrode in

CV and impedance spectroscopy measurements, two methods
widely applied for studying the capacitance in nanoporous
Figure 11. a) Simulation of cyclic voltammograms for an exponential capaci-
films. Impedance spectroscopy is a small-voltage perturbation tance, Cm = Ca exp[ae(EcE)/(kT)], with a series Helmholtz capacitance, CH,
frequency modulation method, which requires a separate mea- using different ratios of Ca/CH : a) 0.001, b) 0.01, c) 0.1 and d) 10. For more de-
surement at each applied potential. The main advantage of im- tails see ref. [164]. Reprinted with permission from [164]. Copyright 2003
pedance spectroscopy is its capability of studying different American Chemical Society.
processes occurring simultaneously, such as charge accumula-
tion, charge transport and charge transfer. CV, on the other
hand, is a large-voltage perturbation method, where the cur- Different features contribute to the typical capacitive re-
rent injected in the electrode is monitored as the potential sponse of nanostructured TiO2 electrodes in aqueous electro-
varies at a constant scan rate v = dE/dt. The capacitance per lyte (Figure 10). The main contribution originates from a feature
unit area can then be obtained from the capacitive current that increases exponentially from the accumulation onset
density [Eqs. (32), (33)]: toward more negative potentials.[83, 164, 178, 185] This capacitance,
which can be fitted by Equation (27), has been associated with
dqs dqs dE dE an exponential DOS in the band gap just below the CB edge
jC ¼ ¼ ¼ Cs ð32Þ
dt dE dt dt for anatase electrodes. Typically, ac (= T/Tc) values between 0.2
jC and 0.5 have been reported. When approaching the CB, the ca-
Cs ¼ ð33Þ
v pacitance flattens due to the fact that the bands shift simulta-
neously with the displacement of the Fermi level. As a conse-
where qs is the charge per geometric (projected) electrode sur- quence the capacitance is partially controlled by the Helmholtz
face area. For Cs = Cm,s (i.e. Cm,s ! CH,s, band pinning) and ne- capacitance at the inner electrode surface.
glecting the thermal spread of the distribution function (zero- Near the onset of the exponential capacitance, an additional
temperature limit), Equation (33) provides a direct means for feature in the form of a capacitive peak is often observed (see
determining the DOS from the CV. However, as a prerequisite Section 3.3.2). This peak is characteristic of nanocrystalline elec-
the scan rate must be slow enough for the film to reach equi- trodes and it has not been observed on structurally well-de-
librium at each state of charging. Furthermore, jC can be deter- fined single-crystal electrodes.[187, 188] It should be mentioned at
mined by CV only for an ideally polarizable electrode as large this point that the characteristic capacitive features of nano-
faradaic currents mask the capacitive current. The experimental structured TiO2 electrodes described above are not restricted
criterion for this is that the charge injected in the electrode to aqueous electrolytes only, but have been observed in or-
during the negative-going scan is approximately equal to the ganic electrolytes and in room-temperature ionic liquids as

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &13&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

band gap.[170] Furthermore, nonexponential kinetics for excited-

state electron injection can be rationalized by invoking an ex-
ponential state distribution.[189] An exponential DOS has been
observed in DSCs for anatase TiO2 and ZnO electrodes by
charge extraction,[174–176] impedance spectroscopy[177, 190, 191] and
intensity modulated photovoltage spectroscopy.[181]
Recently, Ardo and Meyer[193] and Henderson[194] critically re-
viewed the present understanding of the DOS in nanocrystal-
line TiO2 thin films. The origin of the exponential DOS distribu-
tion is indeed not yet well understood, though site heteroge-
neity appears to be a probable cause. High densities of elec-
tronic band-gap states, as typically observed for nanostruc-
tured semiconductors, originate to a great extent from the
high surface area of these materials. These states result from
the truncation of the crystal lattice at the particle surface as
well as from intrinsic and extrinsic point defects. The sintering
of nanosized particles, furthermore, yields a network of crystal-
lographically misaligned crystallites with a high density of par-
ticle/particle interfaces. These interfaces are considered regions
of high concentration of defects.[188]
In contrast to solid-state studies, where the DOS is deter-
mined by photoemission spectroscopy in vacuum, contribu-
tions from the environment should be taken into account
when studying the SEI. A dynamic view of the DOS should be
envisaged for semiconductor electrodes, as a change of the
electron concentration in the film may induce the generation
of new states. Westermark et al.[195] studied the DOS in nano-
structured TiO2 electrodes by photoelectron spectroscopy
before and after electrochemical Li + intercalation (Figure 13 a).

Figure 12. a) Cyclic voltammograms of a nanocrystalline TiO2 film measured

at different scan rates (the arrows indicate increasing scan rate). The experi-
ment employed an anhydrous acetonitrile electrolyte containing 0.1 m tetra-
butylammonium perchlorate and 0.1 m lithium perchlorate. Reprinted with
permission from [179]. Copyright 2002 American Chemical Society. b) Cyclic
voltammograms for different negative potential limits of a TiO2 nanoporous
electrode in a room-temperature ionic liquid with 0.1 m lithium bis(trifluoro-
methanesulfonyl)imide solution. Reprinted with permission from [192].
Copyright 2006 Royal Society of Chemistry.
Figure 13. a) Photoelectron spectra of nanocrystalline anatase TiO2 films
before and after electrochemical Li + intercalation. The spectra are calibrated
with respect to the vacuum level of the as-prepared TiO2 film. Reprinted
with permission from [195]. Copyright 2002 Elsevier. b) Spectra before and
after Ar sputtering. The spectra are calibrated with respect to the Fermi
well (Figure 12 a, b). These electrolytes are frequently used in level. Reprinted with permission from [196]. Copyright 2002 American
functional devices, for example dye-sensitized solar cells Chemical Society.
(DSCs).

A broad distribution of band-gap states  1 eV below the CB

3.3.1. State Distribution in the Band Gap
was observed to scale with the number of inserted Li + ions.
Let us now review the evidence supporting the existence of an The chemical nature of these states was identified as Ti3 + . A
exponential state distribution below the CB edge. In the con- similar evolution of band-gap states was reported by Liu
text of DSCs, the electron transport in the nanostructured sem- et al.[196] upon Ar sputtering of a nanostructured TiO2 film (Fig-
iconductor electrodes was found to strongly depend on light ure 13 b). Importantly, in both cases no exponential DOS was
intensity and/or bias voltage. The observed transport dynamics observed either before or after thin-film modification.
as well as the nonexponential recombination kinetics have Recently, Peter suggested that Coulombic trapping of elec-
been modeled successfully assuming trap-limited transport in- trons near the interface between TiO2 particles and the adja-
volving an exponential distribution of localized states in the cent ionic solution could result in the formation of a nonideal

&14& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

two-dimensional electron gas with strong electron–electron re-

pulsions.[197] In this case “electron traps” could be a manifesta-
tion of Coulombic trapping rather than being associated with
physical defects in the semiconductor. Finally, Abayev et al. as-
cribed the DOS in 10 nm TiO2 particles to a confinement effect
due to the dielectric mismatch at the boundary between the
particle and the surrounding medium.[198] The exponential dis-
tribution in larger particles, on the other hand, was rationalized
by local Coulombic interactions and dislocations of the bulk
lattice. In conclusion, no definitive assignment of the exponen-
tial state distribution in TiO2 has been reported so far. Clearly
a better identification of the physicochemical fundamentals in-
volved is needed.
In an unprecedented approach to the study of the nano-
structured TiO2 electrochemistry, Marken et al.[199] reported the
preparation of a monolayer of anatase NPs on boron-doped di-
Figure 14. Cyclic voltammograms for anatase and rutile NW electrodes in
amond surfaces. Interestingly, these monolayers showed good the dark. Electrolyte: a) N2-purged 0.1 m HClO4 aqueous solution; b) N2-
electronic connectivity with the substrate, thereby allowing purged 0.1 m NaOH. Reprinted with permission from [83]. Copyright 2012
their application as electrodes. Cyclic voltammograms in aque- Elsevier.
ous solution at different proton concentrations were character-
ized by two couples of capacitive waves, which were attribut-
ed to two distinct binding sites for protons in the vicinity of
TiIII sites. By cathodic polarization up to  1000 electrons could
be accumulated per TiO2 particle with a diameter of 6 nm.
Jankulovska et al.[83] recently addressed the DOS of morpho-
logically well-defined, quantum-sized and pure-phase anatase
and rutile TiO2 films in aqueous electrolytes. Films consisting
of oriented NWs with a diameter of  2 nm were prepared by
CBD using different precursors to obtain the respective poly-
morph (see Section 2, Figure 4). The comparative electrochemi-
cal study of the influence of the crystal structure on the DOS
relied on the unique and comparable morphology of these
NW films. As a consequence of the small NW diameter, the
whole electrode structure can be regarded as belonging to the
surface and subsurface region, a bulk region far from the elec-
trolyte being absent. Thus, cyclic voltammograms may provide
Figure 15. Schematic illustration of the electronic band structure of anatase
for NW electrodes a direct measurement of the DOS, though
and rutile NW thin films (pH 1). Reprinted with permission from [83]. Copy-
smeared to some extent by thermal effects.[165] In spite of the right 2012 Elsevier.
fact that the CB edge was found (by photocurrent onset meas-
urements) to be located at more negative potentials for ana-
tase than for rutile, capacitance onset was observed at more at half maximum (FWHM) for the associated capacitive peak
negative potentials in the latter case (Figure 14). The capaci- measured by CV would be given by [Eq. (34)]:
tance profiles were fitted by Equation (25) yielding an ac value
of 1 for rutile and of  0.3 for anatase. An exponential state kT  pffiffiffi
FWHM ¼ ln 17 þ 6 8 ð34Þ
distribution just below the CB edge was therefore claimed to e
be exclusively present at anatase NW electrodes, while the ca-
pacitance in rutile NW electrodes was attributed to the filling and attains a value of 90 mV at 296 K. In basic media, values
of CB states (Figure 15). for FWHM as low as 64 and 54 mV have been reported for
rutile and anatase NW electrodes, respectively.[83] Other studies
have reported similar values.[82, 164, 259] These results suggest the
3.3.2. Monoenergetic Band-Gap States
presence of monoenergetic states located at the band gap in
Near the onset of the monotonic capacitance, an additional nanoporous electrodes of both rutile and anatase. Berger et al.
feature in the form of a capacitive peak is observed in nano- have ascribed these traps to grain boundaries.[188]
structured electrodes (Figure 14) and has been attributed to Wang et al.[180] and Boschloo et al.[173] studied the pH de-
the reversible filling of monoenergetic band-gap states below pendence and the spectroscopic properties of electronic states
the CB edge.[37, 164, 173, 180, 188, 200] For such states the capacitance associated with this capacitive peak in a random NP network
can be described by Equation (30). Accordingly, the full width of anatase TiO2. For particle sizes of  12 nm and at pH 6.2, the

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &15&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

Figure 18. CV of a nanostructured electrode consisting of wire-like nanocrys-

Figure 16. a) Cyclic voltammograms of a rutile NW electrode (c) and tal aggregates (NCs) and an electrode consisting of a random NP network
a nanoparticulate rutile electrode (a). b) Rutile (110) single-crystal elec- with the same active surface area a) in the dark and b) under polychromatic
trode. Electrolyte: 0.1 m HClO4, scan rate: 20 mV s1. Reprinted with permis- illumination (660 mW cm2). Electrolyte: nitrogen-purged 0.1 m HClO4 ; scan
sion from [188]. Copyright 2007 American Chemical Society. rate: 20 mV s1; film thicknesses: 5 mm (NC) and 0.7 mm (NP). Reprinted with
permission from [201]. Copyright 2010 Royal Society of Chemistry.

Q=C  cm2  3:0  106  e0:53pH ð35Þ

Based on the observed pH dependence, the capacitive peak

was associated with surface-related trap states in that study.
On the other hand, recently it was shown that the trap-state
concentration strongly depends on the morphological struc-
ture of the electrode. In particular it was found for rutile that
the number of electron traps referred to the inner surface area
of the sample is high in random NP networks, much lower in
electrodes consisting of oriented NWs, and virtually absent in
structurally well-defined single-crystal electrodes (Figure 16).[188]
Furthermore, surface modification of random NP networks by
NW deposition or adsorption of catechol had no influence on
either the energy or the concentration of the respective trap
states (Figure 17). Based on these observations the capacitive
peak was attributed to electronic states at grain boundaries. In
line with this interpretation, a higher trap concentration in the
band gap was also evidenced for anatase TiO2 electrodes con-
sisting of a random NP network when compared to electrodes
Figure 17. a) Cyclic voltammograms of a nanoparticulate rutile electrode
before and after NW deposition. Deposition times: 0, 5, 10, 15, 30, 60, and
consisting of wire-like nanocrystal aggregates (Figure 18 a).[201]
90 min. b) Cyclic voltammograms of a nanoparticulate electrode in the ab- The higher trap concentration was associated with an en-
sence (a) and presence (c) of 0.1 m catechol. Electrolyte: 0.1 m HClO4, hanced electron recombination, as reflected by a significant
scan rate: 20 mV s1, electrode area: 1.5 cm2. Reprinted with permission from shift of the photocurrent onset for water oxidation to more
[188]. Copyright 2007 American Chemical Society.
positive potential values (Figure 18 b).

number of trapped electrons associated with the capacitive

3.3.3. Other Aspects of Charge Accumulation
peak was estimated to be 12 electrons per particle. These elec-
trons were associated with an absorption band at  400 nm, It is well known that the band positions of semiconductor
where the extinction coefficient was determined to be oxides strongly depend on pH.[16] A Nernstian dependence has
1900 m1 cm1. The number of trapped electrons Q was found generally been observed and attributed to surface acid–base
to strongly depend on pH [Eq. (35)]: equilibria. In a spectroelectrochemical study, Rothenberger

&16& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

et al.[31] observed what the authors called the quasi-flatband

potential of nanostructured TiO2 films to shift by 60 mV per
pH unit in the range from pH 2 to 12. On the other hand, Lyon
and Hupp[202] observed a + 64 mV shift per log (proton activity)
of the CB edge over  31 pH units (Figure 19). As the Nernstian

Figure 19. Dependence of reflectance-derived CB edge energy on log(pro-

ton activity). (s.s.c.e.: conventional saturated (NaCl) calomel electrode) Re-
printed with permission from [202]. Copyright 1999 American Chemical Soci-
ety.

dependence persisted at pH values far above and below the

pKa values of the surface oxo and hydroxo groups, it was
argued that intercalation rather than surface protonation–de-
protonation should induce the variation of the band position
with proton activity. Based on electrochemical quartz-crystal
microbalance measurements it was concluded that the electro-
chemical generation of Ti3 + states is quantitatively accompa-
nied by proton uptake or intercalation (Figure 20) and that the
CB edge is controlled by the pH-dependent TiIII/TiIV couple ac-
cording to [Eq. (36)]:
Figure 20. A) Potential-dependent reflectance at 786 nm of a nanocrystalline
TiIV O2 þ e þ Hþ Ð TiIII ðOÞðOHÞ ð36Þ TiO2 film in 0.1 m HClO4. Scan rate: 50 mV s1. B) Cyclic voltammogram re-
corded simultaneously with the reflectance scan in (A). C) Potential-depen-
dent frequency response of a TiO2-coated crystal, recorded simultaneously
It is well known that small cations such as Li + can readily be with the data in (A) and (B) (s.s.c.e.: conventional saturated (NaCl) calomel
intercalated in the TiO2 lattice in aprotic electrolytes,[203] there- electrode). Reprinted with permission from [202]. Copyright 1999 American
by forming LixTiO2 phases, and this effect has been exploited Chemical Society.
in intercalation batteries.[204] In aqueous electrolytes, on the
other hand, Li + intercalation was claimed to occur only in
strongly basic media under strong accumulation conditions, intercalation properties were attributed to the characteristics
thereby forming Ti3 + species [Eq. (37)]:[184] of the ordered nanostructure, which, in addition to providing
a high surface area, also facilitates the diffusion of H + to/from
TiO2 þ xLiþ þ xe ! Lix TiIIIx TiIV1x O2 ð37Þ the surface. It was concluded that the intercalation process is
limited to the near-surface region of the oxide. The storage ca-
whereas at weak accumulation, electrons get compensated by pacity was thus expected to depend on the surface area rather
cation adsorption. The supercapacitor-to-pseudocapacitor tran- than on the bulk amount of TiO2. As a consequence, charge ac-
sition in anatase TiO2 nanosheets, which is associated with cumulation by Equation (36) may be used to estimate the elec-
a transition from Li + adsorption to Li + intercalation, has re- trochemically active surface area of nanostructured electrodes
cently been addressed by quantum chemical calculations.[205] in situ. In this context, a linear dependence of charge accumu-
Ghicov et al.[158] studied the electrochemical H + uptake lated in rutile TiO2 NW films on film thickness has been ob-
[Eq. (36)] in anatase TiO2 NT layers and observed a high storage served (Figure 22).[188] In line with this interpretation, Krlikow-
capacity for H + as well as a high specific optical contrast when ska et al.[206] studied hydrogen bronze formation on WO3 films
switching the layer from the H-loaded to the H-unloaded state by a spectroelectrochemical Raman approach, and successfully
(Figure 21). The favorable optoelectronic and intercalation/de- resolved on the potential scale, proton intercalation into the

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &17&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

Figure 22. a) Cyclic voltammograms for rutile TiO2 NW electrodes of different

thickness (deposition time: 1–6 h). b) Film thickness and total cathodic
charge accumulated (down to 0.6 V) as a function of deposition time. Elec-
trolyte: 0.1 m HClO4, scan rate: 20 mV s1, electrode area: 2.3 cm2. Reprinted
with permission from [188]. Copyright 2007 American Chemical Society.

sition of electronic states as well as on the transition energies

between ground and excited states.
O’Regan et al.[183] were the first to study the spectroscopic
properties of accumulated electrons in nanocrystalline anatase
Figure 21. a) Chronoamperometric profiles upon a stepwise change in elec- TiO2 electrodes. Rothenberger et al.[31] observed that upon
trode polarization from 0 to 1.5 V and to 1.5 V versus Ag/AgCl (1 m KCl) of
an amorphous TiO2 NT layer, an anatase TiO2 NT layer and a compact anodic
cathodic polarization in 0.2 m LiClO4 aqueous solution at pH 3,
oxide layer in 0.1 m HClO4. b) Reflectance at 480 nm as measured simultane- a broad absorption band between 400 and 1100 nm with
ously. Reprinted with permission from [158]. Copyright 2006 Elsevier. a maximum at 850 nm and a decadic extinction coefficient (at
800 nm) of  1200 m1 cm1 appears (Figure 23).[173, 208] At
pH 11.6 the maximum shifted to 960 nm. The absorption in-
NP “shell” and “core” regions. Consequently, the capacitance crease in the Vis/near-infrared (NIR) was attributed to electrons
associated with hydrogen insertion in the particle outermost trapped at grain boundaries. In addition to the absorption in-
layer can be used for the determination of the real (i.e. active)
surface area.

3.3.4. Spectroelectrochemical Studies of the Charge Accumu-

lation Process
Electron accumulation in TiO2 films has been extensively stud-
ied by spectroelectrochemical measurements.[38, 183–185, 207] The
challenge in this type of experiment is the correlation of the
broad spectral signatures typically observed with specific sites
in or on the oxide semiconductor.[194] Furthermore, identifica-
tion of the excited-state level of an optical transition is chal-
lenging, as excitation of a trapped electron could be via a po-
laronic excited state or into a delocalized CB state.[194] In this
regard, optical spectroscopy can yield the transition energies
between the initial and final electronic states, but it does not Figure 23. a) Difference optical absorption spectrum of a TiO2 electrode in
aqueous 0.2 m LiClO4 (pH 3.0) measured at 1.00 and + 1.0 V (SCE). b) Differ-
provide information on their energetic positions. A combina-
ence spectrum in aqueous 0.2 m LiClO4 (pH 11.6) measured at 1.30 and
tion of spectroscopic and electrochemical methods, however, + 1.0 V (SCE). Reprinted with permission from [31]. Copyright 1992 American
may yield information on the concentration and energetic po- Chemical Society.

&18& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

crease at wavelengths > 400 nm, an absorption decrease The d–d transition is symmetry forbidden; however, symmetry
below the fundamental absorption threshold of the semicon- breaking by asymmetric ligand field splitting or vibronic cou-
ductor was observed and was associated with the so-called pling makes weak absorption possible. In addition to a broad
Burstein–Moss shift, that is, a shift of the absorption onset to absorption in the Vis/NIR range, Cao et al.[38] evidenced by elec-
higher photon energies as a consequence of the filling of elec- tron paramagnetic resonance spectroscopy the presence of
tronic states near the CB edge. Adsorption of catechol or iso- Ti3 + species in TiO2 particles after negative polarization in
phthalic acid on a TiO2 film induced only a small absorbance acidic aqueous solution. The observation of a monotonic ab-
decrease in the visible of less than 10 %. On the other hand, sorbance increase toward longer wavelengths, on the other
a more recent study by de la Garza et al.[207] reports on signifi- hand, has been rationalized by the Drude absorption of free
cant changes of the visible absorption of accumulated elec- CB electrons.[184] In addition, Boschloo and Fitzmaurice[173] asso-
trons upon adsorption of enediol ligands. These changes were ciated the absorbance at  400 nm with the capacitive peak
attributed to the modification of the energetics and the distri- near the onset of the exponential capacitance and attributed it
bution of surface trapping sites. In this study, however, enediol to electrons trapped at surface states (see above). Upon small
ligands were adsorbed on TiO2 NPs prior to thin-film deposi- cation (Li + , Na + ) intercalation into TiO2 electrodes in contact
tion. with acetonitrile or strongly basic aqueous electrolytes
Introducing a band filling model and assuming band edge [Eq. (37)], a pronounced absorption at  750 nm was also ob-
pinning, spectroelectrochemical data were used to determine served.[184]
what the authors considered to be the flatband potential (ac- Recently, the spectral fingerprint of electrons accumulated in
tually the potential onset of the accumulation region) of nano- anatase TiO2 electrodes in the energy range between the fun-
crystalline anatase TiO2 films.[31] In such a case, intraband tran- damental absorption threshold and the onset of lattice absorp-
sitions were assumed to be the main contribution to the ab- tion (0.1–3.3 eV) has been reported.[185] In addition to the well-
sorption in the visible. The pH dependence of this parameter known absorption in the Vis/NIR region, a broad mid-infrared
was determined as [Eq. (38)]: (MIR) absorption, monotonically increasing toward lower wave-
numbers, has been observed (Figure 24). Importantly, signal in-
E FB =V vs: SCE ¼ 0:40  0:06 pH ð38Þ tensities in the Vis/NIR and MIR were found to be linearly cor-
related. By charge extraction experiments it was shown that
These values are in perfect agreement with those obtained the signals are associated with an exponential distribution of
for structurally well-defined anatase single crystals by Mott– band-gap states. A similar monotonic increasing MIR signal has
Schottky analysis.[187] On the other hand, significantly more been observed upon n-type doping of TiO2[211, 212] as well as
negative values for the CB edge have been estimated from the upon band-gap excitation both under high vacuum condi-
measurement of the photocurrent onset, which was found to tions[212, 213] and in the presence of hole acceptors in the aque-
occur at potentials about 0.3 to 0.4 V more negative than the ous phase.[214–216] From a technological point of view, the opti-
onset potential of charge accumulation and visible absorp- cal properties of nanostructured electrodes can be exploited in
tion.[83, 185] The difficulties associated with the definition and the electrochromic devices (see Section 5.2).[217]
determination of the CB edge in nanostructured semiconduc-
tor films have recently been highlighted by Ardo and
3.4. Electron Transfer at Nanostructured TiO2 Electrodes in
Meyer.[193] The method of choice for determining the CB edge
the Dark
position in compact TiO2 films is based on the analysis of the
potential dependence of the space charge capacitance (CSC), In this section we will briefly discuss some relevant electron-
which is associated with a depletion layer at the SEI (Mott– transfer reactions at TiO2 electrodes in the dark. Charge-trans-
Schottky analysis). However, as mentioned in Section 3.1, typi- fer reactions under above-band-gap excitation will be dis-
cal crystallite dimensions in nanocrystalline electrodes are com- cussed in detail in Section 4.
monly too small to sustain significant band bending. Therefore, In the absence of above-band-gap excitation, the semicon-
information on the band position cannot be extracted from ducting properties of TiO2 are disadvantageous for interfacial
a Mott–Schottky analysis in the case of mesoporous nanocrys- electron transfer between the electrode and a given electrolyte
talline electrodes. In this case, however, Mott–Schottky analysis species, as the electron-transfer rate is proportional to the den-
may yield some information on the triple contact of conduct- sity of electronic states near the Fermi level, which is typically
ing substrate, mesoporous film and electrolyte, such as the low for semiconductors.[44, 159] One of the most studied cathodic
TiO2 NP coverage degree on the conducting substrate.[209] reactions on TiO2 electrodes is the oxygen reduction reaction
There has been some controversy in the literature concern- (ORR). An understanding of oxygen reduction is important as it
ing the broad Vis/NIR absorption of TiO2 nanostructured elec- comprises in many processes the counter-electrode reaction
trodes, which has been attributed alternatively to electrons lo- that may even control the desired overall reaction process. In
calized in band-gap states (Ti3 + centers), to electrons in the CB, photocatalysis, for example, electron transfer to oxygen is con-
or to a superposition of both. The leveling off of the absorb- sidered the rate-limiting step.[218, 219] However, the fact that in
ance at longer wavelengths in the Vis/NIR range, which in photocatalysis oxidation and reduction reactions have to occur
some cases reaches a pronounced maximum, has been inter- in parallel at semiconductor particles complicates a separate
preted in terms of d–d transitions of localized Ti3 + states.[38, 210] analysis of cathodic and anodic processes. In this context, one

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &19&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

It is worth noting that the

electrocatalysis of TiO2 is restrict-
ed, in the absence of above-
band-gap excitation, exclusively
to reduction reactions. Oxidation
reactions, however, can be in-
duced by exposure to photons
with energies exceeding the
band gap (see next section).

4. Photoelectrochemis-
try of Nanostructured
TiO2 Electrodes
In this section, the bases of the
photoelectrochemistry of TiO2
nanostructured electrodes are
presented and compared with
those of bulk electrodes. The
equations governing the fate of
charge carriers in nanoporous
semiconductor electrodes are
discussed, first from a purely
physical point of view, and then
by introducing the kinetic infor-
mation of the elementary reac-
tions taking place at the semi-
conductor/solution interface.
Next, the interpretation of typi-
cal electrochemical measure-
ments (photocurrent and photo-
Figure 24. a) UV/Vis/NIR and d) ATR-MIR difference spectra of anatase TiO2 nanocrystalline electrodes at different potential) on a quantitative level
electrode potentials. The reference spectra were taken at EAg/AgCl = 0.4 V. b,e) Semi-logarithmic plot of absorbance
is presented, with a focus on the
and extracted charge versus electrode potential. c,f) Correlation plots of absorbance versus extracted charge. Elec-
trolyte: N2-saturated 0.1 m HClO4 aqueous solution. Reprinted with permission from [185]. Copyright 2012 Ameri- information that can be gath-
can Chemical Society. ered from them. Then, reference
is briefly made to the combined
of the main advantages of electrochemical studies is the possi- use of photoelectrochemistry and spectroscopic measurements
bility of addressing anodic and cathodic processes separately for studying the reactive SEI. Finally, the photoelectrochemistry
(Section 5.3). As a consequence, the ORR was extensively stud- of hybrid or mixed systems (inorganic–inorganic and organic–
ied in the past few decades and results have been reviewed re- inorganic) and surface- or bulk-modified electrodes is intro-
cently.[16] Nowadays, the ORR is especially interesting as being duced. Note that, at this point in the development of photo-
the main cathodic reaction in fuel cells, where platinum was electrochemistry, researchers also try to understand the
found to be one of the most active electrocatalysts. The ORR charge-transfer mechanism in the photoexcited semiconduc-
takes place at much higher overvoltage on TiO2 than on plati- tor/solution interface on a microscopic level. Although beyond
num; however, nanostructured TiO2 electrodes have recently the scope of this review, the Gerischer model keeps on consti-
received attention as an alternative catalyst support in fuel tuting nowadays an adequate framework for the rationalization
cells providing efficient dispersion of the active catalyst.[220–225] of the behavior of different systems, such as NP suspensions,
Electrocatalytic activity of nanostructured TiO2 electrodes quantum dots, nanocrystalline electrodes, and so forth.[228]
[226] [227]
has been reported for some oximino and carbonyl com-
pounds as well as for some olefins with carboxylate groups.[199]
Marken et al.[199] reported on the electrocatalytic reduction of 4.1. Photoinduced Carrier Generation, Transport and Re-
maleic acid on monolayers of TiO2 NPs deposited on boron- combination in Nanostructured Titanium Dioxide Electrodes
doped diamond surfaces. A faradaic cathodic current was at-
4.1.1. Bulk and Nanostructured Electrodes
tributed to the two-proton reduction of maleic acid to succinic
acid. These authors highlighted, furthermore, the potential Although photocurrents and photopotentials can originate
benefits of NP monolayers for controlling the reactivity and se- from sub-band-gap illumination when band-gap states partici-
lectivity in electrocatalytic applications. pate in the photoexcitation process, in the following we will

&20& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

focus on the photoeffects resulting from ultra-band-gap illumi-

nation. Such illumination leads to the generation of electron–
hole pairs whose tendency is to recombine (radiatively or not).
To avoid recombination, the photogenerated carriers need to
be separated. The mechanism by which such a separation is
achieved and, therefore, sizable photoresponses are obtained
depends very much on the semiconductor crystal size and as-
sembly in the electrode. In the case of bulk electrodes, either
monocrystalline or polycrystalline, a space charge layer is
formed at the solid side of the SEI. The electric field associated
with the SEI contributes to the separation of the carrier gener-
ated by photon absorption in this region. The carriers that can
reach the space charge layer by diffusion also contribute to
the photoeffect. In such a way, for adequate band bending,
only one type of carrier reaches the semiconductor surface,
thus minimizing surface recombination. Figure 25. Electrochemical (V vs. NHE) and potential energy (eV) diagram de-
As in the case of bulk electrodes, when a nanoparticulate picting the conduction (Ec) and valence (Ev) band positions for anatase and
semiconductor electrode is illuminated with ultra-band-gap rutile at pH 0, as well as the formal redox potentials (E8’) at pH 0 of different
relevant redox couples involved in photocatalysis and solar fuel production.
light, charge separation processes occur. However, this separa-
Band positions: Ec = 0.20 V (anatase) and 0.00 V (rutile); Ev = + 3.00 V (ana-
tion is not caused by the existence of a space charge layer in tase and rutile). Formal potentials: 0.284 V (O2/O2C), 0.199 V (CO2/
the NPs (like in bulk electrodes); on the contrary, it is primarily HCOOH), 0.106 V (CO2/CO), 0.071 V (CO2/HCHO), 0.00 V (H + /H2),
caused by the different reactivity of electrons and holes + 0.030 V (CO2/CH3OH), + 0.169 V (CO2/CH4), + 0.695 V (O2/H2O2), + 1.229 V
(O2/H2O), + 2.380 V (OHC/H2O). Data were taken from refs. [233] and [234].
toward species in solution and/or at the interface.[229] Charge
separation will be sustainable if there is in the electrode
enough driving force for the transport of carriers. Such a driv- including TiO2, the reactivity of holes is much higher than that
ing force for nanoporous electrodes results from the carrier of electrons, which leads to a rather efficient carrier separation
concentration gradients within the NP network, that is, the at the SEI.
charge transport is governed by diffusion.[230] Nonetheless, Van-
maekelbergh and de Jongh[231] pointed out that drift (migra-
tion) could also contribute to the transport of electrons within 4.1.2. Equations Governing the Fate of Carriers in Nanostruc-
the NP network, thus affecting the photogenerated carrier dif- tured Titanium Dioxide Electrodes
fusion coefficients.[232] Furthermore, in the case of ordered From a phenomenological point of view, the fate of a charged
nanostructures with particle diameters (dimensions) larger species i in a nanoporous electrode depends on three main
than the space charge layer width (e.g. NTs), an interphasial terms: transport, generation and recombination (reaction).
electric field (band-bending) can contribute to charge trans- Considering that the process is limited just to one spatial di-
port by migration.[239] mension (x, normal to the surface of the semiconductor elec-
Once the charge carriers reach the proper reactive sites at trode considered as a flat thin film), the variation of the con-
the particle surface, different electron- or hole-transfer process- centration of the species i (in the discussion below photogen-
es to (from) dissolved or adsorbed species at the illuminated erated electrons and holes) with time (t) is given by Equa-
SEI can occur.[233] Because of kinetic limitations, charge transfer tion (39):
requires a large enough difference between the potential level
of the carriers (conduction or valence band edges for photo- @ci ðx; tÞ @Ji ðx; tÞ
¼ þ Gi ðx; tÞ  Ri ðx; tÞ ð39Þ
generated free electrons and holes, respectively) and that of @t @x
the reactant species either adsorbed or in solution (redox po-
tentials).[234] In Figure 25, we summarize the potential (energy) where ci(x,t) is the concentration (cm3), Ji(x,t) is the flux
positions of some common redox couples in photocatalysis (cm2 s1), and Gi(x,t) and Ri(x,t) are, respectively, the generation
and the band edges for anatase and rutile at pH 0. As ob- and recombination terms (cm3 s1) for the charged species i.
served, photogenerated valence band (VB) holes and OH radi- For charge carriers, the transport is governed by both diffusion
cals have a very large oxidizing power (positive potential) with and migration and the flux is proportional to the electrochemi-
respect to species at the SEI,[233] but on the contrary the CB mi ) gradient according to Equation (40):[235]
cal potential (
electrons show a relatively small reducing power (moderate  
negative potential) to effectively reduce species at the SEI.[218] Di ci ðx; tÞ @
mi ðx; tÞ
Ji ðx; tÞ ¼  ð40Þ
For such a reason, illuminated bare titania has been commonly kT @x
used as photoanode for water and organics oxidations, but it
requires the use of co-catalysts for promoting proton coupled where Di is the diffusion coefficient (cm2 s1). In the case of illu-
electron-transfer processes (e.g. hydrogen generation and CO2 minated semiconductors, this electrochemical potential will be
reduction processes). In fact, for most semiconductor oxides, equal to the quasi-Fermi level of electrons (eF;n * ) or holes

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &21&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

pffiffiffiffiffiffiffiffi
(e*F;p ).[159, 236, 237] The electrochemical potential can be calculated Li ¼ Di ti ð46Þ
according to Equation (41):
On the other hand, the electric potential gradient term
i ¼ m0i þ kT lnðci =c0 Þ þ zi e
m ð41Þ could be obtained by solving the Poisson equation with the
particular boundary conditions for the semiconductor phase.
where m0i is the standard chemical potential (eV), c0 is the stan- This equation is given by [Eq. (47)]:
dard concentration (cm3), zi is the charge number of species i,
and f is the electrostatic (inner) potential. Hence, the flux will @ 2  e1
¼ ð47Þ
be given by Equation (42): @x 2 220

@ci ðx; tÞ zi e @ where 2 is the relative dielectric constant of the semiconduc-

Ji ðx; tÞ ¼ Di  Di Ci ð42Þ
@x kT @x tor, 20 is the vacuum permittivity and 1 is the volume charge
density (C cm3). However, in the case of electrolyte-shielded
where the first term is related to diffusion (first Fick’s law, con- NPs (i.e. nanoparticulate electrodes), no significant potential
centration gradient) and the second to migration or drift (po- gradients appear within the film. In nanostructured electrodes
tential gradient).[238, 239] Finally, by introducing Equation (42) consisting of large enough particles (nanorods, NTs, etc.), sig-
into Equation (39) we get Equation (43): nificant band bending could contribute to the transport (and
separation) of carriers.[239]
 
@ci @ 2 ci zi e @ @
¼ Di 2 þ ci þ Gi  R i ð43Þ
@t @x kT @x @x
4.1.3. Equations Governing the Fate of Carriers in Nanostruc-
This is the general form of the continuity equation for tured Titanium Dioxide Electrodes: Diffusion-Driven Transport
a charged species i in a homogeneous medium under non- In the following, we will focus on the typical case of nanopo-
steady-state conditions. In the particular case of electrons and rous TiO2 electrodes where no significant migration contribu-
holes in a semiconductor,[159] the generation term (Gi) is linked tion exists. In such a case, the term containing the potential
to both photon absorption by the semiconductor and any gradient is negligible in Equation (43), and the continuity equa-
other charge injection process (from an intermediate, from tion of charge carriers is governed solely by diffusion [Eq. (48)]:
a photoexcited sensitizer, etc.). Considering only semiconduc-
tor ultra-band-gap excitation (hn > eg ), then @ci ðx; tÞ @ 2 ci ðx; tÞ
¼ Di þ Gi ðx; tÞ  Ri ðx; tÞ ð48Þ
Gi ¼ Gn ¼ Gp ¼ Gðl; xÞ.[159] If we assume that light absorption @t @x 2
follows the Lambert–Beer law (transparent electrodes), Equa-
tion (44) represents the generation term:
However, the application of Equation (48) to nanostructured
TiO2 electrodes presents some particularities. As long as there
Gðl; xÞ ¼ aðlÞF0 expðaðlÞx Þ ð44Þ
is a strong coupling between electrons in the TiO2 matrix and
ions in the electrolyte, ambipolar diffusion applies. In fact, TiO2
where a(l) is the wavelength-dependent absorption coefficient nanoporous electrodes constitute the typical case in which
(cm1) and F0 is the monochromatic-light incident photon flux electrons in the oxide diffuse in a sea of highly mobile oppo-
(cm2 s1). sitely charged carriers (cations in the electrolyte in our case).
The recombination term (Ri) includes not only the electron– This means that the diffusion coefficient in Equation (48) ac-
hole recombination contribution, but also all the charge-carrier tually is an ambipolar diffusion coefficient.[240–244] It is then clear
transfer processes that occur at the SEI. If we assume that the that Di strongly depends on the electrolyte composition. In ad-
recombination processes follow a first-order kinetics [Eq. (45)]: dition, the fact that nanostructured TiO2 electrodes are highly
disordered systems justifies the existence of several plausible
transport models,[239] which explains the fact that the diffusion
ci  ci;0
Ri ¼ ð45Þ coefficient is not necessarily constant, being in general a func-
ti
tion of both charge-carrier concentration and incident light in-
tensity.[168, 241, 245–247]
where ci,0 is the dark equilibrium concentration (cm3) of As mentioned before, for photoanodes and, particularly, for
charge carriers and ti their average lifetime (s). TiO2, hole reactivity at the SEI is much higher than that of elec-
An interesting magnitude for rationalizing the behavior of trons (particularly at the TiO2/aqueous solution interface). In
nanostructured electrodes is the diffusion length (Li, cm) of addition, hole mobility seems to be much lower than electron
charge carriers. This magnitude represents the distance that mobility.[171, 248] . For these reasons, we can consider that, in gen-
the photogenerated charges can travel through the nanostruc- eral, holes do not contribute significantly to charge transport,
ture by diffusion before recombining or being trapped (in the which is dominated by electron diffusion.
bulk or at the surface of a NP). It is defined according to Equa- For stationary conditions, Equation (48) is particularized for
tion (46): electrons yielding Equation (49):

&22& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

ings. The photocatalytic behavior of titanium dioxide has been

extensively studied and modeled,[250] but in the last few years
a new kinetic scheme for photoelectrochemical measurements
in aqueous solutions has been proposed: the direct–indirect
model.[251] The model is based on the following assump-
tions:[252] 1) steady-state conditions are reached, 2) mass-trans-
port limitations are negligible, 3) the initial photooxidation of
the pollutant (organic molecules) occurs through the transfer
of two electrons, 4) the pollutant photooxidation is competi-
Figure 26. a) Cross-sectional illustration of a nanocrystalline electrode, show-
ing the EE (electrolyte–electrode) and SE (substrate–electrode) illumination tive with that of water, and 5) the main oxidation step pro-
directions, as well as the coordinate system chosen for each of them, where ceeds through photogenerated holes with two limiting cases:
d is the average electrode thickness. b) Photogenerated electron concentra-
tion (nn0) versus distance (x) profile for a nanocrystalline electrode, under
1. Indirect transfer (IT), where a hole trapped at the semicon-
unbiased (open-circuit, a) and biased (c) conditions (positive bias). n0
is the electron concentration in the dark. ductor surface is isoenergetically (without energy loss, elas-
tic process) transferred to dissolved pollutant molecules.
@nðx; tÞ d2 nðxÞ ð49Þ 2. Direct transfer (DT), where a VB free hole is adiabatically
¼ 0 ) Dn þ Gn ðxÞ  Rn ðxÞ ¼ 0
@t dx 2 transferred (with energy loss, inelastic process) to (specifi-
cally) adsorbed pollutant molecules.
where n(x) corresponds to the steady-state concentration of
electrons and the subscript n refers to electrons. To solve this Therefore, the kinetic steps for the photooxidation of a ge-
equation the proper boundary conditions are required, which neric substrate RH2 to R at TiO2 electrodes would correspond
will depend on the illumination side.[42, 249] To distinguish be- to Equations (50)–(57) (Figure 27). Note that for the sake of
tween usual illumination directions, the next nomenclature is simplicity the case of an easily oxidizable substrate is consid-
usually employed (Figure 26): ered and therefore water photooxidation can be neglected.

* EE illumination: the electrode is exposed to the light from TiO2 þ hn ! hf þ þ ef  ð50Þ

the front, that is, the electrolyte/electrode interface.
* SE illumination: the electrode is exposed to the light from
hf þ þ ¼Os ! ¼Os þ C ðhs þ Þ ð51Þ
the back contact, that is, the substrate/electrode interface.

The charge-carrier concentration profile will also depend on

the potential applied to the conducting substrate.[42] Consider-
ing the profile of photogenerated electrons in a nanocrystalline
electrode (Figure 26), under open-circuit conditions and under
illumination, the concentration in the contact will be larger
than the concentration in the dark (n0). Applying an adequate
potential to efficiently extract the electrons at the back con-
tact, a high electron gradient can be established in the vicinity
of the conducting substrate, thereby sustaining a large anodic
photocurrent.

4.2. Photoelectrochemical Measurements: Photocurrent

In this section, we will focus on the photoelectrochemistry of
nanostructured TiO2 electrodes in contact with aqueous elec-
trolytes, particularly in the context of heterogeneous photoca-
talysis. The general physical ideas presented above will be
Figure 27. Energy band scheme of the SEI under illumination, in the pres-
adapted to make explicit the series of elementary chemical ence of dissolved, (RH2)aq, and adsorbed generic pollutants, (RH2)s. An expo-
steps that may take place at the reactive SEI. In such a way, nential distribution of band-gap trap states below the CB is assumed. The
a protocol will be provided for correlating the typical electro- redox species are represented taking into account their reorganization
energy (fluctuating level model). For the sake of simplicity, the Helmholtz
chemical measurements with the chemical events occurring at
double layer is not depicted. The diagram represents the main steps of the
the interface. direct–indirect hole-transfer kinetic mechanism: 1) electron–hole photogen-
eration, 2) surface trapping of VB free holes (hf + ), 3) CB free electron (ef)–
trapped hole (hs + ) recombination, 4) electron transfer to a dissolved oxidant
4.2.1. The Direct–Indirect Kinetic Model (Ox), 5) current-doubling step from a reaction intermediate (RHaqC), 6) recom-
bination step mediated by a reaction intermediate. DT, direct free hole trans-
A proper interpretation of any kinetic measurement requires fer to an adsorbed pollutant; IT, indirect trapped hole transfer to a dissolved
a model that supports and backs up all the experimental find- pollutant.

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &23&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

ef  þ ¼Os þ C ðhs þ Þ ! ¼Os ð52Þ [Eq. (58)] and recombination [Eq. (59)] terms of the continuity
equation yields the equation governing the fate of electrons in
ðRH2 Þaq þ hs þ ! RHaq C þ Haq þ ð53Þ the nanostructured electrode. It can be solved with the boun-
þ
ðRH2 Þs þ hf ! RHs C þ Haq þ
ð54Þ dary conditions corresponding to a particular photoelectro-
chemical experiment (see below). This model has been used
RHC ! R þ Hþ þ ef  ð55Þ for titanium dioxide polycrystalline electrodes,[252, 256] NP sus-
RHC þ Hþ þ ef  ! RH2 ð56Þ pensions[251, 257] and nanostructured electrodes (NPs,[255, 258]
nanocolumns[201] and NWs[259]).
 
Ox þ ef ! OxC ð57Þ

4.2.2. Photocurrent Definition and Measurement

Equation (50) represents the charge-carrier photogeneration
step upon supra-band-gap illumination, which generates CB The photocurrent density (jph, A cm2) is operatively defined as
free electrons (ef) and VB free holes (hf + ). Equation (51) is the the difference between the current density under illumination
hole trapping step that proceeds at surface bridging oxygen (jon) and that in the dark (joff) [Eq. (60)]:
(=Os), thus generating radical oxygen species (=Os + C or surface-
trapped holes, hs + ). Equation (52) represents the trapped hole jph ¼ jon joff ð60Þ
free electron recombination process; as TiO2 is an indirect sem-
iconductor the recombination between free electrons and free This magnitude represents a direct measurement of the effi-
holes is neglected. Equation (53) is the indirect trapped hole ciency of the photoinduced process (electron or hole transfer)
transfer to a dissolved pollutant, (RH2)aq, thereby generating at a certain potential value. In addition, its sign determines
the intermediate radical species RHaqC. Equation (54) is the whether the illuminated electrode behaves as a photoanode
direct free hole transfer to an adsorbed pollutant, (RH2)s. Equa- (jph > 0, jon > joff) or photocathode (jph < 0, jon < joff). It can be
tion (55) is the current-doubling step (occurring for a number measured under two different conditions:
of reductant species),[253] where the intermediate radical spe-
cies (dissolved or adsorbed) injects an additional electron into 1. Potentiostatic conditions: the electrode potential is fixed
the TiO2 CB, which generates the R species. Equation (56) rep- (kept constant) during the measurement. The photocurrent
resents the back reaction, where the intermediate acts as a re- for interrupted illumination is represented versus time (pho-
combination center accepting electrons and generating RH2 tocurrent transient).
again. Finally, Equation (57) is the electron transfer to an exter- 2. Potentiodynamic conditions: for instance, during potential
nal oxidant species (Ox). In addition to all previous equations, scans (voltammetry). In addition, the light can be continu-
in a general case water photooxidation steps should also be ously shined on the electrode, or chopped with a defined
included.[254] on–off period. An important magnitude that can be ob-
The kinetic rates of the Equation (50) and (55) steps are asso- tained with voltammetry (or in a series of potentiostatic
ciated with the generation term, whereas the kinetic rates of transients) is the onset potential (Eonset, V). It is the potential
the Equation (52), (56) and (57) steps belong to the recombina- value at which the photocurrent acquires a non-negligible
tion term. Hence [Eqs. (58), (59)]:[255] value.

1 In Figure 28, we illustrate from a phenomenological point of

Gn ðxÞ ¼ aF0 expðaxÞ þ kinj ½RH  ð58Þ
B view, simple potentiostatic and potentiodynamic measure-
1 ments for a generic semiconductor nanostructured electrode
Rn ðxÞ ¼ ðkr Nfn þ kback n½Hþ ½RH  þ kox n½OxÞ ð59Þ
B (n-type photoelectrode; photoanode). At potentials below the
onset (E1) no photocurrent is detected; when the potential is
where B is a conversion unit factor (cm) between surface and slightly more positive than the onset (E2), a small photocurrent
volumic rates and its value is close to the volume-to-surface with an initial spike when light is turned on is observed, and in
ratio of the nanostructured thin-film electrode; kinj (cm s1) is some cases also when light is turned off. The spikes arise from
the rate constant for the current-doubling step; N is the total electron recombination with surface-trapped holes or photoox-
surface density of bridging oxygen groups (=Os) (cm2), kr idation intermediates.[260] Finally, at potentials significantly
(cm3 s1) is the rate constant for recombination of free elec- more positive than the onset value (E3 and E4), the observed
trons and trapped holes, f is the fraction of surface-trapped photocurrent increases and the spike intensities decrease due
holes at the bridging oxygen sites of TiO2, and hence Nf repre- to a reduced recombination, eventually disappearing (E4).
sents the surface concentration of surface-trapped holes (hs + ); Besides the classical measurements based on large perturba-
kback (cm7 s1) is the rate constant for the dissolved RHC inter- tions of potential, methods based on small perturbations have
mediate-mediated recombination step; and kox (cm4 s1) is the also been extensively used to probe the steady-state behavior
rate constant for electron scavenging by the external oxidant. of electrodes.[239] There are methods either in time (small tem-
The steady-state approximation is applied to obtain the equa- poral perturbation of light, triggering photocurrent transi-
tions for the concentration of the intermediates (trapped hole, ents[261]) or frequency domains. Among the latter, small period-
RHC). Introducing these concentrations into the generation ic perturbations of light intensity may be applied, giving rise

&24& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

namics for water photooxidation. Diffusion lengths were deter-

mined by two independent methods based on the IPCE versus
wavelength profiles and on the analysis of small perturbation
photovoltage transients, respectively. Similar and constant
values for the diffusion length were obtained by both ap-
proaches. The obtained diffusion length values were much
smaller than those determined for electrons photoinjected
from adsorbed catecholate into TiO2 nanoporous electrodes in
aqueous media, as in the latter case recombination is mini-
mized.[265]

4.2.3. Modeling the Photocurrent

From the expressions for the fluxes and the continuity equa-
tions of electrons and holes [Eq. (48)] one can obtain the corre-
sponding expressions for the photocurrent, by recalling that
the current density (ji ) is directly proportional to the charge-
carrier flux (Ji ) at the electric contact of the nanoporous film
(ji(x,t) = eJi(x,t)). The total photocurrent density (jph) would
result from the contribution of both photogenerated electrons
Figure 28. Generic photocurrent versus potential (voltammetry) or photocur- and holes (jph = jph,n + jph,p) as in Equation (61), valid under
rent versus time (transient) curves for an illuminated semiconductor elec-
trode behaving as a photoanode. The top panel corresponds to the voltam-
steady-state conditions:[266]
mogram (jph vs. E) and the bottom ones to the transients (jph vs. t). In the  
voltammogram, the gray lines correspond to chopped illumination and the dn dp
black line to continuous illumination. The transients correspond to four se-
jph ¼ e Dn  Dp ð61Þ
dx dx contact
lected potentials: before (E1) and after (E2, E3) the photocurrent onset, and in
the plateau (E4).
Two cases can be considered: 1) if Dn(dn/dx) > Dp(dp/dx), the
semiconductor has an n-type behavior and the photocurrent is
positive (photoanode, jph > 0); and 2) if Dn(dn/dx) < Dp(dp/dx),
to so-called intensity modulated photocurrent spectroscopy
the semiconductor has a p-type behavior and the photocur-
(IMPS), or of potential under illumination, giving electrochemi-
rent is negative (jph < 0). Under steady-state conditions, the
cal impedance spectroscopy (EIS).[262] IMPS measures the pho-
photocurrent sign will depend on the electron and hole con-
tocurrent response to modulated incident light intensity super-
centration profiles. Similarly, these concentrations will depend
imposed on a constant illumination background; on the con-
on the reactivity of both charge carriers at the SEI (i.e. the ki-
trary, EIS measures the current response to a modulated ap-
netics of the charge-transfer processes).
plied potential superimposed on a constant applied potential
In fact, cathodic photocurrents for typical n-type semicon-
background (in the dark but also under constant illumination).
ductors have been reported in the literature. Hodes et al. were
Another interesting tool for studying the photoelectrochemi-
the first to observe a photocathodic behavior for CdS and
cal properties of nanostructured films is the (photo)action
CdSe nanocrystalline electrodes.[267] Later, Tsujiko et al.[268] ob-
spectrum.[236] This is obtained when the steady-state photocur-
served negative photocurrents for TiO2 with O2-purged basic
rent is registered under monochromatic illumination and plot-
solutions. Likewise, Lana-Villarreal and Gmez[266] showed the
ted as a function of the incident wavelength (or photon
same behavior for TiO2 electrodes modified with Au NPs
energy), either as photocurrent or incident photon-to-current
(Figure 29). More recently, the group of Szaciłowski et al. have
efficiency (IPCE), defined as the number of extracted electrons
suggested that this behavior could be used for logic devices,
per incident photon. Action spectra can be used to precisely
by alternately switching from anodic to cathodic behavior
determine the photocurrent wavelength onset, which in the
(photoelectrochemical photocurrent switching effect, PEPS).[269]
absence of band-gap states corresponds to the band gap of
To obtain an explicit expression for the steady-state maxi-
the semiconductor. This approach was recently used with
mum photocurrent, we need to solve the continuity equation
rutile and anatase NW electrodes, and unveiled an increase of
under diffusion control with the corresponding boundary con-
the band-gap energy with respect to that of the corresponding
ditions. If the current is carried mainly by electrons (photoano-
bulk materials, due to the ultrasmall size of the NWs (quantum
des), Equation (49) should be solved. Under EE illumination,
confinement).[83, 259] Likewise, these spectra have been used to
Equations (62) and (63) constitute the boundary conditions:
study the charge transport in nanostructured thin-film electro-
des under steady-state operation.[263] The electron diffusion nðx ¼ dÞ ¼ n0 ð62Þ
length can be estimated by analyzing the ratio of IPCE values
 
measured under back side (SE) and front side (EE) illumination. dn
¼0 ð63Þ
Only recently, Leng et al.[264] studied the electron collection dy- dx x¼0

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &25&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

Sçdergren et al.[42] also solved Equation (49) by changing the

boundary conditions at the electric contact [Eqs. (62) and (64)]
to simulate the whole jph versus E curve. To do so they as-
sumed that the concentration of TiO2 electrons in the vicinity
of the conducting substrate was dictated by the potential ap-
plied to the substrate (see below). The more positive the ap-
plied potential, the lower is the concentration of electrons at
the back contact and the larger the electron concentration gra-
dient and the photocurrent.
Three regions can be distinguished in a typical photocur-
rent–potential curve (Figure 30). In region A, there is no driving
force to transfer electrons to the conducting substrate and re-
combination predominates. Electron transfer will occur when
the applied potential (conducting substrate Fermi level, eFTO F )
Figure 29. Positive-going voltammetric scans obtained at a rate of 2 mV s1 matches the onset potential. In region B, there is a significant
under chopped illumination for an anatase nanoporous electrode before transfer of electrons to the conducting substrate, even when
(a) and after (c) deposition of Au NPs capped with citrate. Working so-
there is a relatively high concentration of electrons in the con-
lution: O2-saturated 0.05 m NaOH; illumination: 150 W Xe arc lamp. Arrows
indicate that light is on; upward and downward arrows indicate anodic or tact. The higher the applied potential, the lower the electron
cathodic photocurrents, respectively. Reprinted with permission from [266]. concentration at the contact and the larger is the gradient
Copyright 2005 Elsevier. near it. Finally, in region C the electron concentration in the

where n0 is the dark quasi equi-

librium concentration of elec-
trons at the applied potential.
Likewise, for SE illumination,
Equations (64) and (65) give the
boundary conditions:

nðx ¼ 0Þ ¼ n0 ð64Þ
 
dn
¼0 ð65Þ
dx x¼d

Sçdergren et al.[42] were the

first to obtain generic expres-
sions for the steady-state diffu-
sion-limited photocurrent of
a photoanode. The authors
solved the electron continuity
equation [Eq. (49)] by assuming
that the generation term was
given by the Lambert–Beer law
[Eq. (44)] and that recombination
followed first-order kinetics Figure 30. Schematic photocurrent versus potential curve for a photoanode (SE illumination), where three poten-
tial regions are distinguished: one negative to the onset potential (A) and two positive to the onset potential (B
[Eq. (45)]. Under such conditions, and C). For the onset potential and for regions B and C, the electron concentration (n) versus distance (x) curves
the stationary photocurrents for and the corresponding band diagrams are represented. eFFTO represents the Fermi level of the conducting sub-
EE (jphEE) and SE (jphSE) illumina- strate (FTO), and eF,n* the quasi-Fermi level of the photogenerated electrons.
tion are given by Equations (66)
and (67):
contact is negligible and the current no longer depends on po-
 
EE La expðdaÞ La expðdaÞ tential (diffusion control, maximum gradient).
jph ¼ eF0 La þ tanhðd=LÞ  Equations (66) and (67) do not take into account the kinetic
1  L2 a 2 coshðd=LÞ
ð66Þ steps and constants involved at the SEI, as all of them are aver-
aged as a global first-order kinetic constant (lifetime inverse)
 
SE La La expðdaÞ through the diffusion length. To complete the picture consider-
jph ¼ eF0 La þ tanhðd=LÞ þ ð67Þ
1  L2 a2 coshðd=LÞ ing the chemical specificity, the direct–indirect kinetic scheme

&26& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

presented in Section 4.2.1. can be used to deduce the photo- shown to reproduce the electron transport, transfer and re-
current under steady-state conditions. Neglecting the contribu- combination processes in polycrystalline[287] and nanoporous
tion of holes to the photocurrent, Equation (49) has been electrodes,[178, 288, 289] in the dark and under illumination. In an
solved. In addition, an experimental procedure has been devel- applied vein, IMPS has been used to elucidate the effect of
oped to distinguish between the two limiting cases (indirect ethanol on the reduction of surface recombination for an illu-
and direct hole transfer). The dependence on the photon flux minated anatase electrode; on the other hand, EIS has been
of the slope of the photocurrent versus pollutant concentra- used for studying the degradation of salicylic acid[290] and sul-
tion curves in the limit of very low pollutant concentrations fosalicylic acid on anatase electrodes,[291, 292] methyl orange on
has been shown to follow Equation (68): titania NTs[293] and methylene blue on Ag fiber/TiO2 nanocom-
  posites.[294]
@jph
/ ðF0 Þn ð68Þ
@ ½RH2  ½RH2 !0
4.3. Photoelectrochemical Measurements: Photopotential
where n 1/2 for indirect hole transfer (IT) and n 1 for direct
4.3.1. Photopotential Definition and Measurement
hole transfer (DT). Equalities will hold for the particular case of
negligible recombination, that is, for mono- or polycrystalline The photopotential (Vph) can operatively be defined as the ab-
electrodes at high band bending. The model has been success- solute magnitude of the difference between the stationary
fully applied to the photooxidation of model organic com- open-circuit potential (OCP) under illumination (Eph) and that
pounds, such as methanol,[252, 255, 259] formic acid,[201, 252, 255, 258, 259] in the dark (E0) [Eq. (69)]:
and oxalate.[256] Furthermore, it has been validated not only for
TiO2[251, 252, 255, 257, 259] but also for WO3.[258] Vph ¼ Eph  E0 ð69Þ
Similarly, Bilmes et al. have used photocurrent measure-
ments to study the degradation on TiO2 of organic compounds In the literature, often the terms photopotential and photo-
such as methanol and other alcohols,[270–272] salicylate,[271, 272] ox- voltage are equivalently used, although they refer to different
alate[272, 273] and their respective mixtures. To model the meas- concepts. The former is frequently used in photoelectrochemis-
urements a simple kinetic scheme based on the reactivity of try (photoelectrocatalysis), and refers to the measured OCP of
OH radicals was proposed, without considering the continuity an illuminated semiconductor electrode that sustains a photo-
equation. More recently, Amal et al. studied the degradation of induced reaction, normally in a conventional three-electrode
oxalic acid,[274] glucose[275, 276] and succinic acid,[277] both with electrochemical cell, while the latter is employed by the solar
nanoparticulate[274, 275, 277] and nanotube–nanorod electrodes.[276] cell community and refers to the maximum bias (Voc) attained
In some cases, they analyzed on a semiquantitative level the by a two-electrode solar cell configuration working under illu-
saturation photocurrent obtained from photocurrent transi- mination.
ents[274] as a function of the concentration and the incident Photopotential measurements reflect the open-circuit be-
photon flux.[275] havior of an electrode, reproducing the conditions typical of
In contrast, Zhao’s group proposed an alternative modeling photocatalysis with either suspended or supported (immobi-
scheme for photocurrent measurements that does not require lized) particles under illumination. In this case, there is no ap-
solving the continuity equation. Their model is based on the plied bias and the conducting substrate cannot play the role
rising linear part of the voltammetric curve, which is arbitrarily of an electron sink (or source) as in photocurrent measure-
considered to follow Ohm’s law. They obtained equivalent re- ments. The measured potential reflects the overall kinetic be-
sistances for the transfer of photogenerated charges and sub- havior (anodic or cathodic) at the deposited particles. Com-
sequently analyzed how this resistance changes with organic monly, the OCP is a mixed potential (as in corrosion theory)
concentration, photon flux and other experimental parameters. both in the dark and under illumination. Similarly to photocur-
They tested their model with voltammetric measurements for rent measurements, when the OCP under illumination is more
the photooxidation of diverse organic compounds, such as negative than the OCP in the dark (Eph < E0), holes are preferen-
methanol,[278] glucose[279] and various dicarboxylic acids,[280] tially transferred to solution, the electrode behaving as a photo-
both with TiO2 nanoparticulate[278–281] and nanostructured anode; conversely, when Eph > E0 electrons are preferentially
(NTs[282]) electrodes. Additionally, they used photocurrent meas- transferred to solution (photocathode). The OCP is normally
urements for modeling the adsorption of organic compounds, measured versus time (transient) by open-circuit chronopoten-
such as phthalic acid.[281] tiometry, both in the dark and under illumination. In any case,
Regarding small perturbation measurements, laser-pulsed il- the measured potential corresponds to that of the conducting
lumination has been used to address the time-dependent substrate (normally, FTO), equilibrated with the deposited TiO2
charge diffusion in nanostructured electrodes and to deter- layer.
mine the diffusion coefficient of photogenerated elec- Analogously to the photocurrent, the photopotential (or
trons.[43, 283] Similarly, IMPS has been used for studying the elec- photovoltage) can also be measured in the frequency domain,
tronic transport properties of these electrodes proving that by means of intensity modulated photovoltage spectroscopy
charge trapping–detrapping in band-gap states plays an im- (IMVS).[181, 295] IMVS measures the photopotential response to
portant role in the transport mechanism.[284–286] EIS has been a modulated incident light intensity superimposed on a con-

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &27&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

stant-illumination background, under either open-circuit condi-

tions or constant current, although the former is the more fre-
quently used condition.

4.3.2. Open-Circuit Potential–Time Curve for a Photoanode

Let us comment on the different features of a typical OCP tran-
sient for TiO2, obtained by chronopotentiometry (Figure 31).
Depending on the illumination conditions, we can distinguish Figure 32. a) Relative position of the FTO and oxide Fermi levels in the dark
five regions. In region A, the electrode is in the dark and the (eF). b) Photogeneration of electrons and holes by oxide excitation, and
measured stationary potential (E0) is fixed by the composition transfer of holes to solution or surface traps and electrons to FTO. c) FTO
Fermi-level equilibration with the quasi-Fermi level of semiconductor elec-
trons (eF,n*).

the semiconductor CB edge; this value will be reached de-

pending on kinetic factors (generation, recombination and
transport).
It is worth mentioning that the application of an ultrasonic
field (20 kHz) to the working solution has been shown to
induce a change in the OCP of a NW rutile electrode similar to
that caused by illumination. The change in the OCP upon ultra-
sonic irradiation has been called sonopotential by analogy to
photopotential. Water sonolysis giving rise to hydrogen and
hydroxyl radicals and the different reactivity of these species at
the TiO2/solution interface would underlie this phenomen-
on.[296]
Figure 31. Illustration of a generic photoanode OCP transient. The different
stationary and nonstationary regions upon turning light on and off are de-
picted.
4.3.3. Modeling the Stationary Photopotential

(and temperature) of the electrolyte. If there is no reversible There are a number of reports on this issue,[297] but most of
redox couple in solution, the potential is kinetically determined them deal with the behavior of photoanodes in DSCs (photo-
by all the half-reactions that can take place at the electrode voltage).[175, 298, 299] Nonetheless, some researchers have related
(mixed potential). the photopotential measurements with photocatalysis. The
In region B, the electrode is illuminated. As it behaves as early works by Neumann-Spallart et al. successfully correlated
a photoanode, an electron excess is generated in the nanocrys- the photostationary OCP of TiO2[300] and WO3[301] polycrystalline
talline film, revealed by a decrease in OCP. After some time, electrodes in contact with methyl viologen and FeIII solutions,
a photostationary steady state is reached (region C) and the respectively, with their reduction rates in the corresponding
potential attains a constant value (Eph). Upon turning the illu- suspensions. Likewise, Vinodgopal et al.[33] used OCP measure-
mination off (region D), the photostationary potential relaxes ments to show the effect of oxygen as electron scavenger at
(decays), eventually reaching its initial value in the dark (re- TiO2 nanocrystalline thin films. Byrne et al.[302] studied the pho-
gion E). The time required to reach a stationary state (with and tooxidation of oxalate with TiO2 electrodes, proving a linear re-
without illumination) depends on the kinetics of the redox pro- lationship between the photostationary OCP and the oxalate
cesses at the SEI (regions B and D). concentration. Later, they showed the oxygen effect in the
When a nanostructured semiconductor thin film is photoex- presence of formic acid with similar electrodes.[303]
cited, there is a change in both electron and hole concentra- In spite of all the previous studies, no analytical model has
tions, which affects the value of the substrate Fermi level (eF, been developed directly linking the photopotential measure-
Figure 32 b). In the case of photoanodes, holes are either trans- ments with the corresponding photoinduced processes and
ferred to solution or trapped at interfaces at a faster rate than their kinetics. Gmez and Salvador[304] were the first to use
electrons, thereby generating an excess negative charge in the a model similar to that of Sçdergren et al.[42] for solving the dif-
film that induces a rise in the Fermi level of the conducting fusion equation for electrons under steady state, with the ap-
substrate, until it equilibrates with the quasi-Fermi level of the propriate boundary conditions [Eq. (70)]:
photogenerated electrons in the semiconductor film (e*F;n , Fig-    
ure 32 c). The maximum photopotential value will be the differ- dn dn
¼ ¼0 ð70Þ
ence between the initial substrate Fermi level (in the dark) and dx x¼d dx x¼0

&28& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

The photopotential is defined by Equation (71):[298, 304]

 
kT ncontact
Vph ¼ m ln ð71Þ
e n0

where ncontact is the electron concentration in the electrode

contact under illumination, n0 is the same concentration in the
dark, and m is known as the ideality factor (adimensional, m
1). Assuming a Boltzmann distribution for electrons in the dark
(nondegenerate semiconductor), n0 is given by Equation (72):
he i
n0 ¼ Nc exp ð Ec  E0 Þ ð72Þ
kT

where Nc is the CB effective DOS (  1021 cm3 for TiO2[238]), Ec is

the potential corresponding to the CB edge and E0 is the sta-
tionary rest potential of the electrode in the dark. Likewise,
ncontact has two contributions: ncontact = n0 + nph, the electron Figure 33. Simulated difference between the photopotential under SE and
concentration in the dark (n0) and the photogenerated elec- EE illumination as a function of the electrode thickness (d) and for two dif-
tron concentration (nph). Hence, the photopotential will be ferent values of the electron diffusion length (L). The parameters used for
the simulation were: F0 = 1  1017 cm2 s1, a = 5  103 cm1,
given by Equation (73): D = 2  105 cm2 s1. Reprinted with permission from [304]. Copyright 2005
  Elsevier.
kT nph
Vph ¼ m ln 1 þ ð73Þ
e n0

Similarly to the photocurrent case, an accurate description

The electron concentration depends on the illumination side of the reactive SEI with OCP measurements requires solving
(nEE SE
ph 6¼nph ). Taking into account the coordinate system for each the continuity equation with an associated kinetic scheme that
case (Figure 26), the concentrations can be equivalently de- describes the processes taking place at the SEI. In the particu-
fined as in Equations (74) and (75): lar case of large electron diffusion lengths (L > d,[264, 265] that is,
a photopotential independent of the illumination side), a uni-
nEE
ph ¼ nðx ¼ dÞ  n0 ð74Þ form concentration of photogenerated electrons is obtained
(n ¼
6 n(x)); therefore, the diffusion term in the continuity equa-
nSE
ph ¼ nðx ¼ 0Þ  n0 ð75Þ
tion is negligible [Eq. (78)]:

Solving the electron continuity equation [Eq. (49)], the fol- d2 nðxÞ ð78Þ
 0 ) GðxÞ  RðxÞ ¼ 0; GðxÞ ¼ RðxÞ
lowing expressions for the excess concentrations of photogen- dx 2
erated electrons were obtained [Eqs. (76) and (77)]:[304]
Consequently, by simply equating the generation and
EE Fat expðad Þ  expðd=LÞ recombination terms, an explicit expression for the pho-
n ph ¼ aL þ aL expðd=LÞ þ expðad Þ
1  L2 a 2 tanhðd=LÞ togenerated electron concentration, and therefore the
ð76Þ photopotential according to Equation (72), can be ob-
tained as a function of kinetic parameters (rate constants, oxi-
Fat expðad Þ  expðd=LÞ
nSE
ph ¼ aL þ aL þ 1 ð77Þ dant and reductant concentrations, incident photon flux, etc.).
1  L2 a 2 sinhðd=LÞ

where t is the average electron lifetime. By introducing these 4.3.4. Modeling the Open-Circuit Potential Decay after Light
equations into Equation (73), we finally get the photopotential is Turned Off
EE SE
for each illumination side (Vph and Vph ). One of the most important parameters related to the overall
From previous equations, the photopotential will be inde- efficiency of charge collection at a nanostructured TiO2 elec-
pendent of the illumination side when L > d and (1/a) > d, that trode is the electron lifetime (tn). Its measurement has attract-
is, for very thin and weakly absorbing electrodes. In Figure 33 ed much attention, mainly in the dye- and quantum-dot-sensi-
we depict the simulated difference between the photopoten- tized solar cell field.[245] IMVS has been the preferred technique
tial for SE and EE illumination as a function of the electrode for determining electron lifetimes;[295] however, due to its small
thickness. For typical nanoparticulate films (d  10 mm) and perturbation nature IMVS does not allow the analysis of life-
considering a rather large value for the diffusion length (L = times associated with larger OCP variations.[298] To avoid such
d = 10 mm), the photopotential difference will be small a limitation, two main experimental techniques have been de-
SE EE
( Vph  Vph = 32 mV). veloped: open-circuit photovoltage decay measurements

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &29&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

(OCVD)[186, 298] and a charge extraction method.[175] The latter

technique is particularly useful for analyzing the trapped elec-
tron decay, whereas the former is derived, in principle, for free
electrons. The OCVD analysis is based on measuring the open-
circuit voltage decay (Voc) when the light is turned off, after
reaching a photostationary state. The electron lifetime is calcu-
lated as [Eq. (79)]:
 
kT dVoc 1
tn ¼ ð79Þ
e dt

This technique has the advantage of a higher resolution

than other small perturbation techniques (e.g. IMVS), tn being
obtained in a unique and fast measurement. As presented ini-
tially by Zaban et al.,[298] this method admits that there is
a quasi-equilibrium between free and trapped electrons,[245] so
that the determination of the trapped electron lifetime re-
quires knowing their distribution function.
In spite of the wide use of the OCVD method in dye- and
quantum-dot-sensitized solar cell research, no equivalent anal-
ysis for photo(electro)catalysis existed until recently, when
Monllor-Satoca and Gmez[305] proposed a method for obtain-
ing (pseudo)first-order kinetic rate constants (inverse of the
lifetime) for both recombination and charge-transfer processes
at the reactive nanostructured SEI. The method is based on
the combined used of voltammograms in the dark and OCP
decay measurements (Figure 34). Importantly, this method cir-
cumvents the need for both distinguishing between free and
Figure 34. Illustration of the algorithm required for calculating average
trapped electrons and considering the equilibrium between pseudo-first-order rate constants on the basis of OCP decay measurements.
them. The electrode was a P25 nanocrystalline thin layer in contact with a N2-
For obtaining the rate constants, the required experimental purged 0.1 m HClO4 solution (electrode thickness, 7 mm; electrode area,
measurements are: 1) a cyclic voltammogram recorded in the 1.54 cm2 ; incident light intensity, 0.24 W from a Xe arc lamp). A) Cyclic vol-
tammogram obtained in the dark (v = 10 mV s1). B) (Photo)potential transi-
dark at a relatively low scan rate (v), which displays a quasi-re- ent obtained after interrupting illumination. C) Variation of the photogener-
versible behavior (symmetric shape with respect to the poten- ated electron concentration with time, obtained from (A) and (B). D) Varia-
tial axis, Figure 34 A), and 2) the OCP versus time relaxation tion of the average pseudo-first-order rate constant with the photogenerat-
dark
transient after turning off illumination (Figure 34 B). The elec- ed electron concentration, obtained from (C). Eoc in this figure corresponds
to E0 (open circuit potential in the dark). Reprinted with permission from
tron concentration during the decay can be calculated by inte- [305]. Copyright 2008 American Chemical Society.
gration of the voltammogram between the stationary potential
in the dark and the potential at a certain time during the
decay (Eoc ðtÞ) according to Equation (80): intermediates. In the presence of an oxidant, the constant
Z   mainly refers to the charge transfer to the electron acceptor.
Eoc ðtÞ
1 I In addition, the proposed algorithm allows one to obtain
nph ðEÞ ¼ dE ð80Þ
eAd E0 v the dependence of the rate constant on the electronic (poten-
tial) energy, by calculating the corresponding microcanonical
where A is the electrode geometric area and I the voltammet- rate constants (k(E)) from their respective average constants (k)
ric current. Once the photogenerated electron concentration through Equation (82):
decay curves are known (nph vs. t, Figure 34 C), the average
pseudo-first-order rate constant as a function of electron con-  
centration (k vs. nph, Figure 34 D) can be calculated with Equa- v dkðEÞ
kðEÞ ¼ kðEÞ þ nph ðEÞeAd ð82Þ
tion (81): IðEÞ dE

 
k nph ¼ 1 dnph ð81Þ
nph dt where I(E) is the voltammetric current for each potential value
E. This procedure was used for studying the effect of fluorina-
In the absence of an effective oxidant (electron acceptor), tion on the rate constants of charge recombination and elec-
this constant refers to the recombination process of accumu- tron transfer to oxygen.[305] Recently, Guijarro et al. used the
lated electrons with surface-trapped holes or photooxidation same method with CdSe quantum-dot-sensitized TiO2 electro-

&30& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

des to unveil the different recombination pathways in the ies).[312, 313] The temporal evolution of the photocurrent and of
composite system.[306] vibrational bands associated with adsorbate species and reac-
tion products were tracked simultaneously in the course of
oxalic acid photooxidation (see Section 5.3. for further details).
4.4. Photoelectrochemistry and Spectroscopy: A Combined
As already mentioned, it was shown that electrons accumulat-
Approach
ed in anatase TiO2 electrodes upon band-gap excitation or ex-
Whereas electrochemical approaches successfully address the ternal polarization may readily be detected by IR spectrosco-
macroscopic electrode properties, they lack molecular specifici- py.[185, 314] An exponential distribution of band-gap states, which
ty. Vibrational spectroscopy, on the other hand, has been gives rise to broad absorptions in the Vis/NIR and MIR regions,
proven useful for studying processes at the semiconductor/so- has been evidenced for anatase TiO2. Importantly, it was
lution interface on a molecular level, thus providing a micro- shown that the electron population is the same upon band-
scopic view of the reactive interface.[307, 308, 356] The high sensitiv- gap excitation at open circuit and upon external polarization
ity of IR spectroscopy is beneficial for the detection of low con- in the dark as long as the electrode potential is kept constant
centrations of reactants, intermediates or products, but the in- (Figure 35). This demonstrates that the Fermi level throughout
terference of typically strong IR absorption by the working (fre-
quently aqueous) solution presents a major challenge.
Measurements are therefore frequently performed in attenuat-
ed total reflection (ATR) mode. IR spectroscopy has been ap-
plied extensively to the study of the semiconductor/solution
interface under both reactive and nonreactive condi-
tions,[307, 308, 356] Raman spectroscopy has been used much less
due to its low sensitivity.[356] However, signal intensification has
successfully been achieved by surface enhancement (SERS)
and resonance (RRS) mechanisms. Lana-Villarreal et al.[309] stud-
ied by SERS the adsorption of phthalic acid at nanoporous ana-
tase films deposited over roughened Au substrates. The Au
substrate features an electromagnetic magnification of the
Raman signal, which extends spatially into the anatase film.
Spectral bands corresponding to vibrations of molecules ad-
sorbed on the semiconductor were thus intensified. In addi-
tion, SERS was used to study the adsorption of formic acid and
methanol on anatase films. The interaction of salicylic acid
with TiO2 films and slurries was also addressed by Raman spec-
troscopy. In this case, detection of the surface complex formed
by salicylate on TiO2 relied on the visible resonance (RRS)
Figure 35. ATR-IR spectra of anatase TiO2 nanocrystal electrodes upon UV ex-
mechanism.[310] posure at open circuit and external polarization in the dark, respectively. The
Taking advantage of the complementarity of the information reference spectra were taken prior to UV exposure at open circuit/polariza-
extractable from spectroscopic and electrochemical measure- tion in the dark. Representation of the two types of perturbation together
ments, multi-technique approaches have successfully been de- with the electronic properties deduced. Electrolyte: N2-saturated 2 m
HCOOH/0.1 m HClO4 aqueous solution. Reprinted with permission from
veloped and were applied to the study of photo- and bias-in- [314]. Copyright 2012 American Chemical Society.
duced processes at the SEI. The charge-transfer complex
formed upon adsorption of catechol on anatase NPs in contact
with aqueous acidic solutions has been studied by photoelec- the mesoporous film is homogeneous in both cases as long as
trochemical techniques, IR spectroscopy and RRS.[311] At least the concentration of electron acceptors in solution is kept low.
two adsorbate configurations, catecholate in a chelate configu- More generally, the IR–spectrophotoelectrochemical approach
ration and molecularly adsorbed catechol, were evidenced. allows for studying both the electronic properties of the semi-
Upon electron injection from the chelating catecholate, which conductor as well as the vibrational properties of solution spe-
was found to form a charge-transfer complex with the semi- cies during a photoinduced reaction. Consequently, combined
conductor, catechol polymerization was tracked by an increas- spectroscopic and photoelectrochemical approaches may find
ing fluorescence signal in the Raman spectra. Later, the adsorp- direct application in a wide range of systems where following
tion and photoreactivity of catechol was studied for slurries the reactivity of the SEI at a chemical level is critical.
and films of WO3 and TiO2, respectively, and it was concluded
that catechol adsorption is initially faster on WO3, but yields
4.5. Photoelectrochemistry of Mixed and Modified Titanium
larger final coverages with TiO2.[356]
Dioxide Nanomaterials
Finally, a combined IR spectroscopic and electrochemical ap-
proach was used to address the photoreactivity of the SEI The use of TiO2 nanostructured thin films for both photocata-
under potential control (spectrophotoelectrochemical stud- lytic and energy conversion applications is hindered by a series

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &31&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

of intrinsic limitations that can reduce the efficiency of the

photoinduced redox processes, such as:[6, 69, 194]

1. A low photonic efficiency (i.e. IPCE).

2. Recombination can be much faster than the interfacial
charge-transfer processes.
3. A reduced visible-light absorption, which prevents signifi-
cant utilization of the solar spectrum.
4. Different interfacial charge-transfer rates. As a consequence
single rate-determining steps may limit the global process
efficiency.

The use of semiconductor oxides with a wide band gap,

such as TiO2, represents an advantage from the point of view
of their stability, as they tend to (photo)corrode much less
than semiconductors with a lower band gap (e.g. Si, GaAs).[315] Figure 36. Illustration of metal–semiconductor (a, c) and semiconductor–
However, this property impedes a full exploitation of the solar semiconductor (b, d) nanocomposites, prepared by contacting one basal ma-
terial with another material (a, b) or forming a core–shell structure (c, d). In
spectrum, as only the UV region and, at best, a small part of the metallic composites, the photogenerated electrons are transferred to
the visible range is actually used. The TiO2 spectral utilization the metal center (a preferential transfer of the holes to the metal cannot be
can be mainly improved by:[316] excluded). In the all-semiconductor composites, the semiconductors are
photoexcited and photogenerated electrons or holes are eventually trans-
ferred. The lines inside the particles represent the valence and conduction
1. Bulk modification (doping). The introduction of dopants, or band edges of the semiconductors, or the Fermi level of the metal.
the creation of vacancies in the oxide, induces new energy
levels located near the band edges or as mid-gap states
that may reduce the effective band-gap width. Sometimes the composite is formed by TiO2 NPs and
2. Surface modification (sensitization). The surface adsorption a second component that is an electrocatalyst for the process
of molecules that absorb visible light allows electron injec- of interest (for instance, water photooxidation). In the context
tion from their excited state (LUMO level) to the semicon- of heterogeneous catalysis this second component is called
ductor CB when an adequate band alignment exists.[317] the co-catalyst. Some of the co-catalysts for promoting hole
transfer from TiO2 to solution are inorganic compounds, not
Concerning carrier recombination and transfer limitations, always well defined, such as RuO2, IrO2, NiOx, Co-P or Ni-P.[11, 323]
the application of an external potential (bias) can reduce the These composites cannot be included in any of the categories
electron–hole recombination and improve the rate of electron listed below.
and/or hole transfer to their respective acceptors at the SEI
(photoelectrocatalysis).[318] Likewise, the use of composite ma-
Titanium Dioxide–Semiconductor Composites
terials (nanocomposites) and quantum-confined systems
(quantum dots, NWs, etc.) could also circumvent these limita- The coupling of two semiconductors allows one to: 1) increase
tions.[319–321] In the following, we will discuss some of the meth- the charge separation (rectification) and quantum yield of the
ods for improving the photoactivity of TiO2 nanocrystalline global photocatalyzed process, and 2) extend the range of
electrodes,[322] illustrating their merits with some selected ex- photoexcitation energies (wavelengths) due to the different
amples. Additional examples will be given in Section 5.3. band-gap values.[319] Kamat and co-workers have devoted
many efforts to studying the photoelectrochemistry of semi-
conductor coupled systems.[229] Among others, TiO2 has been
4.5.1. Titanium Dioxide–Inorganic Composites
combined with WO3,[324, 325] PbS,[326] CdS,[327–329] SnO2,[330–332]]
Nanocomposites are generally hybrid or mixed semiconduc- CdSe[333, 334] and CuI.[335]
tor–semiconductor or metal–semiconductor systems, where at
least the dimension of one of the phases lies in the nanometer
Titanium Dioxide–Quantum Dot Nanocomposites
range. For their preparation, some general strategies can be
employed:[319] This is a particular case of semiconductor/semiconductor cou-
pling. Quantum dots are nanoscopic semiconductor particles
1. Disperse the particles of one component in a continuous with a size small enough to show an optical behavior inter-
matrix of the other component (support). mediate between that of a bulk semiconductor (with energy
2. Pile the components as stacked layers. bands) and that of a semiconductor cluster (with discrete
3. Connect (without covering) the particles of one component energy levels).[336] Within this size range, the particle dimen-
with those of the other (Figure 36 a and b). sions are comparable to the De Broglie wavelength of the sem-
4. Cover the particles of one component with those of the iconductor charge carriers[337] and quantum effects are pre-
other (Figure 36 c and d), in a core–shell fashion. dominant. These effects are responsible for: 1) a widening of

&32& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

the particle band gap, and hence a blue shift of the semicon- harvesters.[351] Graphite oxide was also used as co-catalyst for
ductor absorption edge; and 2) a shift of the conduction and hydrogen production, upon exfoliation and subsequent cover-
valence band edges to more negative and positive potentials, age of titanium dioxide NPs.[352] More recently, large reduced
respectively. graphene oxide[353] and nanosized reduced graphene oxide
Quantum dots are used not only as solar cell sensitizers (visi- composites with titania have shown increased photoactivi-
ble-light harvesters, Figure 37 a) but also as efficient photocata- ties.[354]
lysts. The band-edge shift leads to an increased photocatalytic

4.5.2. Titanium Dioxide–Organic Nanocomposites: Titanium

Dioxide–Polymer
The charge separation and reactivity of titanium dioxide pho-
togenerated charge carriers can be enhanced, not only by
means of all-inorganic composites, but also with organic mate-
rials. In particular, the use of polymeric composites has proven
suitable for photocatalysis.[322] Heterojunction donor–acceptor
metal oxide films constitute the basis of a common strategy
Figure 37. Comparison of TiO2 NPs sensitized to the visible region with for increasing charge separation.[355] The role of charge donor
either quantum dots (a) or adsorbed dye molecules (b). Quantum dots show or acceptor can be played by either the polymer or the semi-
a discrete distribution of states, and dye molecules are energetically charac-
conductor. Upon photoexcitation, the charge carriers can be
terized by a couple of localized states (HOMO–LUMO levels).
effectively separated at the heterojunction as it provides the
needed driving force. In general, these films can be synthe-
sized by: 1) filling the empty pores of a metal oxide film with
activity, as both the reducing and oxidizing power of photo-
the polymer, or 2) blending into a film the polymer and metal
generated electrons and holes are increased. Titanium dioxide
oxide NPs by codeposition. In both cases, either the polymer is
has been frequently coupled with CdS,[338] CdSe,[339] MoS2, or
generated in situ from starting monomers or it is deposited as
PbS[340] quantum dots.
a preformed polymer. In the former case, for photochemical
synthesis, polymerization drives away one of the photogener-
Titanium Dioxide–Metal Nanocomposites ated charge carriers by stepwise polymer growth, thus enhanc-
ing the charge separation process.[356] Titanium dioxide has
The modification of titanium dioxide by the deposition of
been cast as hybrid polymeric films mostly with polyani-
(noble) metals is another common approach to improve the
line,[357, 358] but also with polyamide[359] and polystyrene,[360] to
photoelectrochemical and photocatalytic properties of TiO2.
mention a few.
When the metal work function (F, eV) is larger than that of
On the other hand, Beranek et al.[361–363] studied TiO2 surface
TiO2, electrons are effectively transferred to the metal NPs
modification with carbon nitride generated by the pyrolysis of
forming a Schottky barrier.[6] This phenomenon brings addi-
urea. They were able to obtain visible light photoactive materi-
tional advantages, such as:[6, 69, 194] 1) an increase in the charge-
als, both as nanoparticulate[361, 362] and NT electrodes.[363] In
transfer rate to dissolved or adsorbed molecules, the metal
such a case, one of the major roles of the polymer would be
acting as an electrocatalyst (co-catalyst);[11] and 2) enhance-
to sensitize TiO2 to the visible region.
ment of the charge separation efficiency, the metal behaving
as an electron sink (electron trapping sites) or, eventually, as
a hole sink.[341] In some particular cases, the metal NPs could
4.5.3. Surface Modification
also sensitize the oxide to the visible light due to their surface
plasmons.[342] Some of the metals used for the TiO2 nanocom- The different roles of adsorption in this type of modification
posites are Au,[266, 341, 343] Pt,[344] Ag,[345] Cu[346] and Ir.[344] are determinant.[364, 365] On many occasions the adsorbed spe-
cies (molecules, ions, complexes) can either catalyze or block
the charge-transfer process. Surface modifiers can be classified
Titanium Dioxide–Carbon Nanocomposites
into two main groups:
The combined use of titanium dioxide with nanostructured
carbon materials provides additional photochemical activity 1. Spectral sensitizers. Upon adsorption, they expand the ef-
and enhances the electron transfer to species at the SEI.[347] fective spectrum absorption of the bare semiconductor.
Fullerenes can act as electron acceptors and promote both 2. Interface conditioners. Upon adsorption, they change the
charge separation and electron-transfer processes.[348] Fullerol, band edge positions, thereby altering the charge-transfer
a hydroxylated fullerene derivative, has also been used as sur- characteristics of the bare semiconductor.
face modifier for enhancing TiO2 visible photoactivity.[349]
Single-walled carbon nanotubes can act as support and an- Here the term “interface conditioners” refers to those adsor-
choring scaffold to TiO2 NPs, thereby enhancing the overall bates that do not induce any spectral sensitization. Both of
photoconversion efficiency,[350] as well as acting as visible-light them can play a catalytic role.

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &33&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

Spectral Sensitizers

Classically, sensitization has been achieved using adsorbed

dyes. The molecule of adsorbed dye is excited by the absorp-
tion of visible light, thus generating an electronic vacancy in
the molecule HOMO level and a photoexcited electron in its
LUMO level. This electron can be injected into the CB of the
semiconductor provided that the dye LUMO level and the CB
edge have the proper alignment (Figure 37 b).[316] The forma-
tion of surface charge-transfer complexes (for instance, in the
case of catechol[311]) makes possible the direct excitation of
HOMO electrons into the TiO2 CB. Visible sensitization is nor-
mally attained with transition-metal complexes with low-
energy excited states, such as complexes with polypyridine,
phthalocyanines and metalloporphyrins (organic dyes), or, al-
ternatively, with inorganic NPs (quantum dots). In the first Figure 38. Illustration of the change in the energy of the conduction and va-
case, the dye molecules bind to the semiconductor surface lence band edges upon adsorption at the semiconductor surface of a) the
negative charge density of a dipolar molecule (middle) or an anion (right),
through their functional groups (e.g. carboxylate groups).
and b) the positive charge density of a dipolar molecule (middle) or a cation
Among the most stable and frequently studied sensitizers, we (right). HP refers to the outer Helmholtz plane.
find the RuII complexes with polypyridyl ligands.[317]
Sensitization has also shown promising results in photocatal-
ysis. In this case, the excited dye acts as an oxidant of the dis- sivating agents of surface recombination sites. This type of sur-
solved electron donors, generally pollutants.[366] The conditions face modification method has been largely studied, and nowa-
for a sustainable regenerative sensitized photocatalytic oxida- days there is a wide variety of modifiers. Possibly, the most
tion were established as:[367, 368] studied has been fluoride,[371, 372] although AlIII,[373] ZnII,[374] anion-
ic polyoxometalates,[375] phosphate[376] and neutral organophos-
1. The surface properties of the semiconductor (charge, ad- phonated[377] molecules have also been used.
sorption capability) should facilitate the simultaneous pres-
ence of sensitizer and substrate (not necessarily adsorbed)
4.5.4. Electrochemical and Chemical Doping of Titanium Diox-
in the vicinity of the semiconductor surface.
ide
2. The sensitizer oxidized and reduced forms should be quite
stable and its reduction and oxidation should be rather re- The (photo)electrocatalytic performance of nanostructured
versible. electrodes can also be tailored by changing the bulk properties
3. The electronic transfer efficiency from the oxide CB to the of the particles, by the application of a negative potential (re-
oxidized sensitizer (back reaction) should be very small. ductive bias) for some time, as was already proven for single-
crystal[378–380] and polycrystalline electrodes.[381, 382] The cathodic
An important application entails the use of dye-sensitized polarization of nanocrystalline TiO2 electrodes in aqueous elec-
photoelectrosynthesis cells. The group of Meyer has extensive- trolytes was found to enhance in some cases the photocatalyt-
ly studied the use of Ru dyes for multi-electron and proton ic performance.[200, 383] The beneficial effect was attributed to
coupled transfer reactions, mainly oxygen generation.[369, 370] the electrochemical doping of the films by generation of Ti3 +
species and concomitant intercalation of H + or Li + . The
degree of such an improvement was found to strongly depend
Interface Conditioners
on the nature of the electrode.[200]
These modifiers are adsorbed species with no particular photo- Berger et al.[200] were the first to observe a photocurrent en-
chemical activity, which can be either ionic (cations or anions) hancement for water photooxidation in acidic media upon a re-
or neutral (dipolar). Upon adsorption at the semiconductor sur- ductive treatment (at 0.6 V vs. Ag/AgCl) of different TiO2
face (contact adsorption), the valence and conduction band nanoparticulate electrodes; besides, the doping process shifted
edges are shifted due to the net charge or dipolar moment of the peak associated with grain boundaries to more positive
the adsorbed modifiers, thereby altering the potential drop in potentials. The extent of the observed phenomena, both in
the Helmholtz layer (band edge level unpinning). If the modifi- the dark (grain boundary peak shift) and under illumination
er is an anion or a dipole with its negative end oriented (photocurrent enhancement) strongly depended on the partic-
toward the surface, bands shift to more negative potentials; ular TiO2 sample (Figure 39), and it was rationalized in terms of
on the contrary, for cations or dipoles with their positive end the buildup of a space charge layer (i.e. band bending) along
facing toward the solid, bands shift to more positive potentials the larger particle aggregates. The doping effect was long-last-
(Figure 38). This band shift could facilitate the charge-transfer ing but reversible, and the initial photocurrent was recovered
processes between the semiconductor and redox species in so- upon 20 h of relaxation (for electrodes prepared from PI-KEM
lution. In some cases, the surface modifiers can also act as pas- commercial nanopowders).

&34& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

conductor in photovoltaic applications[387, 388] is beyond the

scope of this review and has been intentionally omitted.

5.1. Lithium-Ion Batteries

The high specific energy and power available from Li-ion bat-
teries, together with a high reversibility of charge/discharge
cycles, are the reasons for their key importance in electronic
portable devices and electric vehicles.[389] Lithium-ion batteries
are composed of cells that employ lithium intercalation com-
pounds as the positive and negative electrodes. When the bat-
tery is cycled, lithium ions (Li + ) exchange between the positive
and negative electrodes. The positive electrode material is typi-
cally a metal oxide (LiCoO2, LiMn2O4), while the negative elec-
trode material is typically a graphitic carbon or highly graphi-
tized compound (see Figure 40). Despite its utilization, the

Figure 39. a) Cyclic voltammograms showing the photooxidation of water

on a PI-KEM electrode as a function of the electrochemical reduction time. Figure 40. Representation of a lithium-ion battery. Negative electrode
b) Enhancement factors for electrodes prepared from PI-KEM, Degussa P25, (graphite), positive electrode (LiCoO2), separated by a nonaqueous liquid
Sachtleben and Alfa Aesar commercial powders. Scan rate: 20 mV s1, film electrolyte. Reprinted with permission from [395]. Copyright 2008 John
thickness: 6–8 mm, polychromatic irradiation I = 500 mW cm2 ; working solu- Wiley and Sons.
tion: N2-purged 0.1 m HClO4. Reprinted with permission from [200]. Copy-
right 2006 Elsevier.

graphite electrode has some disadvantages, such as the initial

loss of capacity and structural deformation under working con-
Not many relevant studies dealing with the photoelectro- ditions. To avoid these drawbacks, in recent years an increasing
chemistry of TiO2 doped with either cations or anions can be research effort has been directed toward the development of
found. However, a few reports have appeared clearly evidenc- alternative negative-electrode materials with enhanced kinet-
ing the potentialities of a photoelectrochemical approach in ics.[390]
the study of these modified materials. For both N-doped and TiO2 is a promising candidate as alternative material for the
C-doped TiO2,[384–386] relevant information could be obtained re- negative electrode because of its high capacity, low voltage for
garding both the electronic structure of the modified elec- lithium intercalation, and its advantages in terms of cost,
trode, characterized by a dopant-induced midgap level, and safety and toxicity.[391] Typically, Li + insertion/extraction at TiO2
the mechanism of visible-light response of these materials. occurs in the potential range of 1.4–1.8 V versus Li/Li + . The
overall reaction can be written as Equation (83):
5. Applications of Nanostructured TiO2 Elec- TiO2 þ x Liþ þ x e Ð Lix TiO2 ð83Þ
trodes
As mentioned in the Introduction, a large number of applica- where x is the insertion coefficient. The maximum theoretical
tions employ nanoporous TiO2 electrodes (Figure 1), the final capacity is 330 mAh g1 for anatase,[392] which corresponds to
performance of the devices being determined by the TiO2 elec- x = 1 and, therefore, to the complete reduction of Ti4 + to Ti3 + .
trochemical properties. In this section, we cover very briefly All crystalline phases of TiO2 (anatase, rutile, brookite and
some of the applications of nanostructured electrodes such as TiO2(B)) can accommodate Li + ions in their structure to some
batteries, electrochromic devices, photo(electro)catalysis, and extent. Rutile and brookite were initially thought to accommo-
artificial photosynthesis. The application of TiO2 as electron date only small amounts of lithium, due to their compact

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &35&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

structure. Anatase and TiO2(B) forms were expected to show of the Li + insertion/extraction process is evidenced by the vol-
better performance.[393–395] In fact, anatase was considered to tammogram symmetry and the cathodic/anodic current values
be the most promising candidate as anode material for Li-ion plotted in the inset of Figure 41.
batteries due to its fast Li + insertion/extraction kinetics and In summary, TiO2 is a promising candidate for the negative
high insertion capacity.[396–398] The reversible insertion into ana- material in lithium-ion batteries, although there is still room for
tase TiO2 takes place up to about 0.6 mol Li + per TiO2 mol at improvement of the structure of TiO2 by optimizing the prepa-
1.78 V versus Li + /Li.[399] The presence of a Li content greater ration method and, therefore, the final nanostructure. It should
than x = 0.5 in LixTiO2 leads to strong Li–Li interaction in the be mentioned that other aspects of the Li + -based batteries are
lattice and, finally, to a loss of reversibility. In fact, the stoichi- also under investigation, such as the electrolyte composition
ometry Li1TiO2 can only be obtained from anatase at high tem- and nature. Mainly, four types of electrolytes have been used:
perature, or with a particle size inferior to 7 nm.[400, 401] liquid,[410] gel,[411] polymer[412] and ceramic electrolytes. The uti-
Lately, rutile has also been investigated as a possible candi- lization of solid electrolytes has the advantage of eliminating
date for lithium-ion batteries. Reversible Li + insertion/extrac- the need to use flammable solvents, thus increasing the safety
tion up to  0.5 mol Li per mol of TiO2 at room temperature of the Li-ion batteries.[413]
into/from nanosized rutile has been demonstrated.[123, 131] Kinet-
ic limitations are responsible for the difference between micro-
sized and nanosized TiO2 rutile.[402, 403] The latter materials are 5.2. TiO2 Electrochromics
beneficial for this particular application because they provide Electrochromism can be defined as the ability of a material to
a larger electrode/electrolyte interface, and because the inser- undergo color change upon oxidation or reduction. Electro-
tion/extraction kinetics is faster.[395, 404, 405] chromic properties can be found in almost all the transition-
Besides its crystalline phase and the particle size, a third cru- metal oxides.[132] These materials have been extensively investi-
cial characteristic of TiO2 that should be taken into account for gated because of their potential applications in practical devi-
its application in Li + batteries is the morphology of the parti- ces.[128, 131, 414, 415]
cles. Recently, one-dimensional nanostructures, such as those The mechanism for electrochromics, as in the case of the in-
constituted by NTs, nanorods, and NWs have been stud- tercalation batteries, is the double injection of electrons and
ied.[142, 406, 407] Employing TiO2(B) NWs, compositions up to ions into the oxide matrix [Eq. (84)]:[416]
Li0.91TiO2 have been obtained without structural degradation.
TiO2(B) has a lower density than rutile and anatase, thus TiO2 þ xðMþ þ e Þ Ð Mx TiO2 ð84Þ
making it an ideal host for Li + intercalation.[134, 136, 396, 408]
Figure 41 shows cyclic voltammograms at different scan rates
where M + can be either H + or Li + (x is the insertion coeffi-
for TiO2(B) in a Li + -containing electrolyte.[409] The reversibility
cient, whose value depends on the micro- and nanostructure
of the deposited thin film). During charge injection, electrons
are localized at titanium sites, thereby changing the valence of
Ti ions to 3. As mentioned in Section 3.3.3, the absorption in
the visible region was attributed to localized states near the
CB edge or to intraband transitions of free electrons in the CB.
To quantitatively compare the electrochromic properties of
different nanomaterials, spectroelectrochemical measurements
are performed, the coloration efficiency (CE) being one of the
characteristic parameters employed for such a purpose. It is
defined according to Equations (85) and (86):

DODðlÞ
CEðlÞ ¼ ð85Þ
Q
Tb ðlÞ
DODðlÞ ¼ log ð86Þ
Tc ðlÞ

where DOD is the change in the optical density of the film be-
tween its colored (Tc) and bleached (Tb) states at a certain
wavelength (l), and Q is the corresponding injected (or eject-
Figure 41. Cyclic voltammograms of TiO2(B) in 1 m LiN(CF3SO2)2 + ethylene ed) charge density per unit area. Tungsten oxide (WO3) is one
carbonate/1,2-dimethoxyethane (1:1, v/v); scan rate 0.1–1.2 mV s1 (in of the most investigated electrochromic inorganic materials
0.1 mV s1 steps for plots from bottom to top). Inset: the normalized peak because of its fast switching between colored and bleached
current, i/i01, where i01 is the peak current at the slowest scan rate
states and long-term durability, its properties serving as a refer-
(0.1 mV s1) and i is the peak current at the actual scan rate. Circles and
crosses denote two individual peaks. Reprinted with permission from [409]. ence in this field. It shows a blue color when a sufficiently neg-
Copyright 2005 American Chemical Society. ative potential is applied.[417–420] More recently, efforts have

&36& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

been made to enhance the electrochromic performance of 5.3. Photocatalysis

tungsten oxide by doping with Ti. However, depending on the
Ti content, the structure of WO3 evolves from monoclinic to Many organic compounds can be decomposed on TiO2 with
amorphous, and finally to a mixture of WO3–TiO2.[421] Up to the aid of photons with energies exceeding the semiconductor
now, a variety of results have been reported on the effect of band gap. This capability, together with the high chemical and
the Ti concentration on the final CE.[143] It seems that an opti- photochemical stability, nontoxicity and availability at low cost,
mal concentration is needed to improve the efficiency and makes TiO2 the prototype material for photocatalytic applica-
that only small amounts are favorable.[422] tions. These applications are mainly related to environmental
The values reported for CE are in general smaller for TiO2 concerns and span from water and air purification to self-
thin films than for WO3. However, CEs for TiO2 (l = 550 nm) as cleaning surfaces.[16, 218, 428]
high as 24.5 [423] and 33.7 cm2 C1 [415] have recently been report- For photocatalysis to take place, at least two types of chemi-
ed, which are comparable to those of WO3[424, 425] and much cal processes have to occur simultaneously on the catalyst—an
higher than those previously reported.[128] oxidation based on photogenerated hole transfer and a reduc-
The electrochromic properties, such as CE, cyclic durability tion involving photogenerated electrons. The overall reaction
and kinetics of the coloration and bleaching process of TiO2, relies on a precise balance of these two processes. For photo-
strongly depend on its structural, morphological and composi- catalysis based on catalyst particles or particle agglomerates,
tional characteristics, and therefore on the deposition tech- this balance is limited to discrete units where oxidation and re-
nique.[158] As ion intercalation is typically limited to a very thin duction processes occur in parallel with electron–hole recom-
metal oxide layer, to achieve a substantial contrast a high sur- bination. A single photocatalyst NP may be considered there-
face area is required.[426, 427] In this sense, Berger et al. have re- fore as a nanosized photoelectrochemical cell under short-cir-
ported remarkable electrochromic properties for electrodes cuit conditions.[429]
made of extremely thin TiO2 NWs of both anatase and The high relevance of electrochemistry for the analysis and
rutile.[82, 83] Importantly, these NW thin films have the unique manipulation of photocatalytic reactions is based on the possi-
property, among other TiO2 nanostructures, of yielding a signifi- bility of separating anodic and cathodic processes. In this con-
cant coloration in acidic aqueous solutions (0.1 m HClO4) when text, the impact of electrochemistry on photocatalysis can be
they are polarized only at 0.6 V versus Ag/AgCl. Both TiO2 ascribed to two main aspects: the possibility to gain funda-
anatase and rutile NWs behave similarly, showing fast colora- mental knowledge by using electrochemical methods as an an-
tion–decoloration kinetics (Figure 42). alytical tool, and the improvement of the performance of an
immobilized catalyst film by the action of an externally applied
electrical field.

5.3.1. Analytical Aspects

From a fundamental point of view, standard electrochemical
methods provide a straightforward means for studying inde-
pendently the anodic and cathodic parts of the overall reac-
tion.[430] Both thermodynamic and kinetic information can be
gained from such experiments. Photoelectrochemical process-
es have been studied extensively to gain fundamental knowl-
edge on the factors governing the photocatalytic behavior of
the semiconductor.
Using single-crystal electrodes, Kesselman et al.[431] presented
an approach which aims at the elucidation of important kinetic
factors of the TiO2-catalyzed photodegradation of organic com-
pounds. The overall rate of the interfacial process was associat-
ed with a steady-state condition where the flux of holes across
the solid/liquid interface is balanced by an equivalent flux of
electrons (Figure 43). This flux-matching condition was deter-
mined by measuring separately the kinetics of O2 reduction as
a function of applied potential in the dark and the photoano-
dic current characteristics in the absence of oxygen. The cur-
rent-matching condition can yield an estimate for the net flux
of electrons and holes that would be obtained in TiO2 particles
Figure 42. Chronoamperometric data corresponding to a stepwise change at a given light intensity as well as the “operating potential” of
of the electrode potential from 0.8 to 1 V and back to 0.8 V in 0.1 m HClO4
the photocatalyst in the absence of an external bias. Further-
for A) rutile (film thickness: 250 nm) and B) anatase (film thickness: 400 nm)
NW thin films , and the corresponding absorbance change measured at more, such an analysis may highlight possible strategies to im-
725 nm. Reprinted with permission from [83]. Copyright 2012 Elsevier. prove the performance of a photocatalyst.[431] However, due to

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &37&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

served for the photooxidation of small organic molecules

(methanol, formic acid) at nanocrystalline TiO2 electrodes.[433]
Mass transport from the electrolyte bulk to the semiconduc-
tor surface and vice versa is a process that gains importance
for nanocrystalline electrodes as a consequence of their meso-
porous structure. Using electrochemical[302, 436] and spectroelec-
trochemical approaches,[312] it was shown that mass-transport

Figure 43. Schematic diagram of the photoelectrochemical behavior of TiO2

in aqueous solutions. a) Dark oxygen reduction current density, qUn(E), in an
air-saturated solution as a function of potential. b) Oxidation of H2O/organic
compound under illumination, qUp(E). The quantity qG0 is the limiting photo-
generated minority carrier current density, which equals the electron–hole
pair generation flux multiplied by the electron charge, at the voltage indicat-
ed. c) Light limited case. d) Charge-transfer limited case. Reprinted with per-
mission from [431]. Copyright 1994 American Chemical Society.

the mesoporous structure of nanocrystalline thin films, addi-

tional aspects gain significance in the interactively related reac-
tion steps associated with mass transfer in solution, surface ad-
sorption and desorption, charge-carrier generation and trans-
port in the semiconductor as well as interfacial charge transfer
and recombination.[432, 433] These factors make it necessary to
study each particular system (TiO2 nanomaterial, substrate to
be oxidized, solution, etc.) to obtain relevant conclusions and/
or improvement strategies.
On the one hand, products of the hole transfer may act as
scavengers of photogenerated electrons thus causing low
IPCEs, even at strong anodic polarization. This is in contrast to
bulk TiO2 electrodes (single crystals or compact polycrystalline
films) where a depletion layer may render insignificant this in-
direct recombination (“redox cycling”) as well as direct elec-
tron–hole recombination pathways even at relatively small
anodic bias. On nanocrystalline electrodes, in contrast, signifi-
cant band bending does not occur and recombination takes
place, not only under open-circuit conditions but also under
anodic polarization. Water photooxidation, for example, is char-
acterized by relatively high IPCE values on bulk anatase and
rutile electrodes, whereas high recombination is observed at
nanocrystalline TiO2 (see Section 5.4).[36, 434] In the presence of
easily oxidizable organic molecules, higher photocurrents are
often measured. In addition, some reaction intermediates
evolving from the photooxidation may inject in a second reac- Figure 44. a) ATR-IR spectra of 0.1 and 10 mm H2C2O4 + 100 mm HClO4 aque-
tion step an electron into the CB of TiO2 (see Section 4.2.1). ous solutions in contact with a NW TiO2@Au@Si thin film. Background spec-
tra were taken using H2C2O4-free blank solutions. b, c) Difference IR spectra
This provokes a “multiplication” of the photocurrent,[435] which
(b) and photocurrent transients (c) taken simultaneously at EAg/AgCl = 0.06 V.
may be more efficient on nanoporous electrodes than on d) Spectrophotoelectrochemical setup. Reprinted with permission from
single crystals. IPCE values above 100 % have readily been ob- [312]. Copyright 2010 Royal Society of Chemistry.

&38& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

Table 2. List of aqueous organic pollutants degraded by photoelectroca- Table 2. (Continued)

talysis on nanostructured TiO2 electrodes.
Class of organic com- Examples References
pounds
Class of organic com- Examples References [526, 527]
methylene blue
pounds [528–531]
methyl orange
[496, 532]
Reactive Orange 16
[255, 259, 278, 441, 501, 502] [533]
Aliphatic alcohols methanol Reactive Brilliant Red X-3B
[279, 441, 501, 502] [523]
glucose rhodamine B
[505] [523]
ethylene glycol sulforhodamine B
[505]
diethylene glycol

[255, 259, 303, 441, 501, 503–506]

Aliphatic carboxylic formic acid
acids
[441, 505]
acetic acid
oxalic acid [280, 302, 441, 496, 501, 502, 507]
limitations may become critical, especially at low substrate
[280, 441, 501, 502]
malonic acid concentrations and for narrow pore structures. As observed in
[280, 441, 501, 502]
succinic acid
[280, 441, 501, 502, 508] Figure 44, the photocurrent dramatically decreases for oxalic
glutaric acid
acid at low concentration (0.1 mm) due to organics depletion
Aromatic alcohols phenol [260, 441, 503, 509–511]
at the oxide surface (mass-transport limitations) as confirmed
[260, 441, 512]
catechol
[512]
by means of IR spectroscopy. In this context, Wen et al.[437] de-
resorcinol
[441] termined the effective diffusion coefficients of various organic
hydroquinone
1,2,4-trihydroxybenzene [512] compounds from photocurrent measurements under diffusion-
bisphenol A [513]
controlled conditions.
[505]
2-phenoxyethanol Sun et al.[438] used chronoamperometry to study photoin-
[441 duced charge accumulation in the presence of various metal
Aromatic carboxylic benzoic acid
acids ions and co-additives and experimental potential transients
phthalic acid [441, 501, 502]
were simulated by a simple kinetic model. Mora-Ser et al.[255]
[441, 501]
salicylic acid developed a kinetic model describing the photoelectrochemi-
[512]
3,4-dihydroxybenzoic acid
[291] cal behavior of nanostructured TiO2 electrodes in contact with
sulfosalicylic acid
aqueous electrolytes containing dissolved pollutant species. As
Chloroaromatic com- 2-chlorophenol [441, 502]
discussed in Section 4, the model correlates the steady-state
pounds photocurrent, the illumination intensity and the specific photo-
[441]
3-chlorophenol
[33, 34, 441, 503] oxidation mechanism for dissolved pollutant species in compe-
4-chlorophenol
4-chlorocatechol [514] tition with water molecules.
4-chlororesorcinol [512]
Nanostructured TiO2 electrodes and photoinduced reactions
[512]
4,6-dichlororesorcinol
[496, 515, 516]
have furthermore been characterized by EIS[178, 290–294, 439] and
pentachlorophenol
IMPS.[284, 285, 440] In summary, electrochemical methods have
Aromatic nitro com- 2-nitrophenol [517] been used extensively to characterize a large number of TiO2-
pounds based nanostructured electrodes. These studies aim at eluci-
[517, 518]
4-nitrophenol dating how the photoelectrochemical and photoelectrocatalyt-
[441] ic behavior depend on thin-film characteristics including TiO2
Amino acids and de- glycine
rivatives crystal structure,[441–443] electrode morpholo-
[82, 83, 103, 201, 259, 444–453]
glutamic acid [441, 508]
gy, doping by anions,[384–386, 454, 455] doping by
[441, 508]
phenylalanine
[508]
metals,[456–459] boron doping,[460] reductive doping[200, 383, 461, 468] as
lysine
[508] well as surface modification by fluorine adsorption[305, 371, 462] or
b-alanine
8-aminooctanoic acid [508] N modification.[361–363, 463] Furthermore, compound materials
glutamine [508]
have also been studied, including TiO2/semiconductor,[464–477]
[519]
TiO2/metal,[341, 343, 478–481] TiO2/carbon[353, 482–485] and TiO2/polymer
Aromatic amines 4,4’-oxydianiline
compounds.[486–490]
Surfactants sodium dodecylbenzene- [445] Substrates that have been photoelectrocatalytically convert-
sulfonate ed at nanostructured TiO2 electrodes in aqueous solution com-
[520]
prise bacteria,[491–496] and inorganic[497–499] and organic com-
Herbicides alachlor
pounds including dyes (Table 2) as well as mixtures of organic
Dyes Acid Orange 7 [464, 521, 522] molecules[273] or dyes.[500]
[523]
alizarin red
[524]
azo dye DG 26
indigo carmine [525] 5.3.2. Technological Aspects
[331]
naphthol blue black
malachite green [523] From a technological point of view, external application of an
electric field can improve the performance of nanostructured

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &39&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

semiconductor thin-film electrodes by enhancing charge sepa-

ration.[33–35, 271, 534] Electrochemically assisted photocatalysis over-
comes the limitation of electron scavenging by separating the
anodic and the cathodic half-reactions. Whereas the semicon-
ductor anode catalyzes the oxidation of the organic com-
pound, oxygen reduction is catalyzed at the metal counter
electrode. The beneficial effect of an electric field has been ex-
ploited in photoelectrocatalytic reactors of different
design.[492, 503, 535–543] Some reactor designs are shown in Fig-
ures 45–47.

Figure 46. Arrangement consisting of a cylindrical TiO2 anode and nickel

cathode to form an electrode cassette around the annular well used to
house the UV lamps. Reprinted with permission from [496]. Copyright 2011
John Wiley and Sons.

Figure 47. Schematic cross section of: a) a one-compartment reactor, b) a

two-compartment reactor. GSL: global solar light, CDSL: concentrated direct
Figure 45. Schematic representation of a batch-mode plate reactor with
solar light. Reprinted with permission from [536]. Copyright 1999 Elsevier.
flowing film of aqueous solution. Reprinted with permission from [537].
Reflector profile and typical reactor layout for c) a parabolic trough reactor
Copyright 2002 Springer.
(PTR) and d) a compound parabolic collector (CPC). Reprinted with permis-
sion from [546]. Copyright 2009 American Chemical Society.

Recently, Egerton[496] presented a comparative study of four

representative photoelectrocatalytic reactions: nitrophenol oxi-
dation, oxalate degradation, E. coli inactivation and dye decol- deployment and activation by UV light. In this context, several
oration. Importantly, reaction rates of photoelectrocatalytic and concepts of photocatalytic reactors for environmental remedia-
photocatalytic experiments were directly compared using the tion have been presented.[544–547] The future success of photo-
same reactor and the same lamps in both cases. For this pur- catalysis technology clearly will depend on both an optimiza-
pose TiO2 layers thermally grown on Ti substrates, sol–gel elec- tion of the reactor design and the development of more effec-
trodes and NT arrays were studied. The electric field enhance- tive photocatalysts.
ment factors, defined as the ratio of reaction rates of the pho-
toelectrocatalytic and photocatalytic experiments, were found
5.4. Artificial Photosynthesis
to be probably too small to make photoelectrocatalysis (as
compared to photocatalysis) a commercially attractive process. Mankind is going to face in this century major challenges that
Photoelectrocatalytic and photocatalytic reactors must be will determine its future. As Kalyanasundaram and Grtzel have
designed to provide efficient pollutant transport, photocatalyst pointed out,[548] two of the most important will be to sustain

&40& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

a growing energy demand as part of a progressive depletion

of fossil fuels, and to limit emissions of greenhouse effect
gases (CO2). The emphasis should be on developing clean and
renewable energy sources and on providing an adequate level
of environmental protection. In addition, solar energy reaches
the Earth’s surface with a speed of 120 000 TW, an amount
which exceeds that consumed by present man-made technolo-
gy by four orders of magnitude. It is therefore attractive to try
to address these challenges from the standpoint of photo-
chemical conversion and accumulation of solar energy. It is
known that natural photosynthesis is the main method of con-
verting solar energy into chemical energy through the absorp-
tion of CO2, but its conversion efficiency is limited to roughly
1 %.
As mentioned in the Introduction,[25] the first device to per-
Figure 48. Sketch showing the potential corresponding to the conduction
form artificial photosynthesis was a photoelectrochemical cell and valence band edges of TiO2, WO3 and Fe2O3 at pH 0 together with the
consisting of a titanium dioxide (rutile) photoanode and a Pt standard potentials for the redox processes implied in water photosplitting.
counter electrode. Whereas at the Pt cathode hydrogen was
evolved, oxygen was generated
at the photoanode and, there-
fore, the device sustained the
water photosplitting reaction
[Eq. (87)]:

1
H2 OðlÞ hn! H2 ðgÞ þ O2 ðgÞ
2 ð87Þ
DG0298 K ¼ 237 kJ mol1

Unlike typical photocatalytic

processes,[218] in this case the re-
action is not spontaneous, being
a process in which light is con-
verted into chemical energy. Im- Figure 49. Scheme of the water photooxidation reaction on TiO2 electrodes as a function of pH. Reprinted with
portantly, almost 40 years later, permission from [254]. Copyright 2007 American Chemical Society.
the question is still under in-
tense investigation because
a system that works with a substantial efficiency under visible plates were successfully employed for water photosplitting.[27]
light has not been found yet. Nevertheless, significant progress In 1994, Augustynski and co-workers studied several photooxi-
has been achieved in recent years.[11, 12, 549, 550] dation reactions in aqueous media on both nanocrystalline
The fact that TiO2 has been employed within this context is and sol–gel-derived anatase electrodes. The compact sol–gel
not casual. It derives from the electronic structure of this oxide thin-film electrodes were demonstrated to be much more effi-
in which the typical high oxidative power of holes in the VB is cient toward water oxidation than the nanoparticulate ones
combined with a sufficient reductive power of the photogener- (Figure 50).[36] Recently, Hartmann et al. made similar observa-
ated CB electrons (Figure 48). Other oxides commonly em- tions.[551] The absence of a field-induced separation of the pho-
ployed in water-splitting studies need an applied bias to drive togenerated carriers facilitates substantial recombination in
the reduction process. nanoparticulate films.
The mechanism of water photooxidation on rutile single- Although the nanoparticulate electrodes do not seem to be
crystal electrodes has been studied in detail by Nakato and co- particularly effective for water photosplitting, they can be used
workers by means of photoelectrochemical and spectroscopic for analyzing in a fast and reliable way the influence of dop-
techniques (Figure 49).[254] The fact that holes need to be ants or adsorbates on the water photooxidation process at
trapped and accumulated at the surface in combination with supported or dispersed NPs. In such a way, detailed photoelec-
the evolution of several oxidized intermediates explains the ki- trochemical experiments have been used to rationalize the
netic difficulties of the process, as recombination of these spe- effect of nonmetallic dopants[384, 386] or to uncover the influence
cies with photogenerated electrons is facile. of TiO2 fluorination on the water photooxidation process.[372]
Different TiO2 electrode morphologies have been employed In contrast with the behavior of nanoparticulate samples,
apart from the single crystals typical of the early studies. In TiO2 NTs grown by anodization of Ti plates have been shown
1975, rutile thin films prepared by thermal annealing of Ti to be quite effective for water photooxidation, as demonstrat-

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &41&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

In addition to the photoelectrochemical methods them-

selves, the systems formed by dispersed or supported NPs are
based on the same ideas, although each of the particles in this
case constitutes a microreactor (as described in Section 5.3).
Most research in this field focuses on suspensions because
they are cheaper, but have the disadvantage that the photo-
dissociation products must be separated.[12]
On the other hand, there is a growing concern about the
CO2 emission resulting from fossil fuel combustion, which is
leading to an increase in CO2 concentration in the atmos-
phere.[553–555] This gas is the principal contributor to the green-
house effect, which is leading to global warming. In this con-
text, several technologies for CO2 capture have been devel-
oped, although nowadays there does not exist an effective
and practical method of post-treatment.[556] Among them, the
Figure 50. Photocurrent–voltage plots for a) a nanoparticulate P25 TiO2 film photoelectrocatalytic reduction of CO2 with TiO2 and other
and b) a compact, sol–gel-derived anatase TiO2 electrode recorded in a 0.1 m oxides and semiconductors was demonstrated for the first
NaOH aqueous solution and after addition of 0.1 m methanol. The electrodes
time in 1979.[557] Many other photocatalytic studies, as re-
were irradiated with the full output of a 150 W Xe lamp. Reprinted from
[36]. Copyright 1994 Royal Society of Chemistry. viewed recently,[554] with TiO2 as photoactive material have
been published since then.
We finally remark on the use of TiO2 nanotubular structures
ed for the first time by Grimes and co-workers.[119] This seminal as photoanodes in photoelectrochemical cells based on the
paper has triggered frenetic research in this field, partly be- concept of the proton exchange membrane (PEM) fuel cells
cause the NT structure can be controlled in a convenient way that can sustain both hydrogen generation and CO2 reduction
from both a morphological and compositional point of view. (Figure 52).[558]
Recent reviews cover this topic in a comprehensive way.[549, 552]
Other quasi-one-dimensional nanostructures such as those
constituted by rutile NWs have given impressive results once
thermally treated in hydrogen (see Figure 51).[461]

Figure 52. a) View of the lab-scale photoelectrochemical cell (PEC) device.

b) Image of the photo/electrocatalytic disc. c) Scheme of the PEC device for
Figure 51. a) IPCE spectra of pristine TiO2 and H:TiO2 NWs prepared at differ- CO2 reduction to fuels and H2 production. Reprinted from [558]. Copyright
ent temperatures at a potential of 0.6 V versus Ag/AgCl in 1 m NaOH. 2010 Royal Society of Chemistry.
Inset: magnified IPCE spectra are highlighted in the dashed box, at the inci-
dent wavelength range from 440 to 650 nm. b) Simulated efficiencies for the
pristine TiO2 and H:TiO2 NWs as a function of wavelength, by integrating
their IPCE spectra collected at 0.6 V versus Ag/AgCl in 1 m NaOH with
a standard AM 1.5G solar spectrum. Reprinted from [461]. Copyright 2011
American Chemical Society.

&42& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

6. Conclusions and Outlook taking into account the fact that many reactions in heteroge-
neous catalysis actually result from the sum of two redox half-
Semiconductor oxides, particularly titanium dioxide, are materi- reactions, which imply an interfacial charge transfer that
als receiving a growing interest due to their multiple industrial should be affected by changes in the solid/electrolyte interface
and technological applications (solar cells, catalysts, photocata- conditions. From these ideas, it can be concluded that the
lysts, sensors, etc.). Most of these applications require nano- electrochemical measurements may play a fundamental role in
scopic materials (NPs), which show peculiar properties depend- studies that aim at enhancing the TiO2 photoactivity. In fact,
ing frequently on their shape and size. The study of such mate- electrochemical experiments in the dark may provide informa-
rials has also triggered important progress in science and tech- tion on oxygen, water or CO2 reduction processes, while those
nology. It is worthwhile to emphasize the relation existing be- under illumination may allow us to study reactions of hole
tween the nanoscopic oxide electrochemical properties and transfer to water or organic species in solution.
others much more studied, such as the structural, electrical, It is important to mention that electrochemistry not only
electronic, optical and catalytic properties. This relation is often provides techniques for analyzing the interfacial processes. It
neglected, misusing its potential advantages, probably be- can also be used to prepare nanostructures (for instance, syn-
cause many groups working in the science and technology of thesis of NTs by anodization) or to modify their properties, for
titanium dioxide nanomaterials lack a good background in example, by electrochemical doping. This possibility can be
electrochemical science. Being aware of this fact, in this review, employed to optimize the nanostructures for some particular
we have tried to summarize the main conceptual tools that applications.
can be employed to rationalize the electrochemical behavior Although our understanding of the electrochemistry of
of TiO2 nanostructured materials, and to briefly present some nanostructured TiO2 electrodes has improved dramatically in
applications of these materials based on their electrochemical the last two decades, there are still significant scientific chal-
properties. lenges that will require the attention of the scientific commun-
It is noteworthy that charge generation, injection, separa- ity in the future. For instance, the factors that govern band
tion, transport and/or accumulation in the nanostructure un- edge pinning/unpinning are not fully understood. This is im-
derlie many of the TiO2 nanomaterial applications, which clear- portant not only from a fundamental point of view, as the DOS
ly shows that an exhaustive study of the electrochemistry of can easily be derived from the electrochemical response in the
these materials is a key point to understand and enhance or case of band edge pinning, but also from an applied point of
optimize some of the effects. Some of these applications are view, as electrochromism or charge accumulation would be en-
related to environmental issues (heterogeneous photocatalysis) hanced. On the other hand, the reactivity under illumination
and others to energy issues, including both saving (electro- should strongly depend on the NP morphology and assembly
chromic devices) and generation or accumulation (third-gener- within the film. However, to date, the dependency of the elec-
ation solar cells, lithium batteries). We believe that once the in- trochemical behavior on NP size and shape is not systematical-
fluence of the TiO2 structure on the electrochemical properties ly known. Other questions, such as the optimization of elec-
is understood, the structure of the nanomaterial will be opti- trode properties via the irreversible adsorption of simple ions
mized using the electrochemical response as a guideline and (fluoride, for instance), are still in their infancy, as is the possi-
taking into account the requirements of each practical applica- bility of doping/modifying the TiO2 nanomaterial via (photo)-
tion. electrochemical processes. Undoubtedly, many open questions
Already by the 1990s, it was clearly shown that the behavior remain and we just hope that this work may contribute to im-
of nanostructured TiO2 electrodes was different from that of proving the electrochemical background of scientists working
single crystals and polycrystalline samples. Concretely, charge or planning to work in the fascinating world of nanostructured
separation (and therefore photocurrent generation) in nanopo- TiO2 materials.
rous electrodes was shown to originate from the difference in
the charge-transfer kinetics for both carriers (electrons and Abbreviations and Symbols
holes) at the solid/solution interface. On the other hand, it was
also evidenced that electron transport inside the electrode A Electrode geometric area [cm2]
takes place mainly by diffusion. ATR Attenuated total reflection mode of IR spectroscopy
Although not as popular as their photoelectrochemistry, the B Conversion unit factor between surface and volumic
dark electrochemistry of these materials provides a wealth of rates [cm]
information that spans from the electronic structure of the Cm Chemical capacitance [F]
oxide to the determination of the electrochemically active in- Cm,v Chemical capacitance by volume unit [F cm3]
CB
terfacial area, or the presence of grain boundaries. These dark Cm;v Chemical capacitance by volume unit associated to
electrochemical properties of TiO2 underlie applications as im- the CB [F cm3]
GB
portant as electrochromism and intercalation electrodes for Cm;s Chemical capacitance by area unit associated to
batteries. monoenergetic states [F cm2]
That the electrochemistry of TiO2 nanoparticulate samples is ci(x,t) Concentration of species i [cm3]
ss
central in the understanding of heterogeneous photocatalytic Cm;s Chemical capacitance by area unit associated to
or photosynthetic processes is not particularly surprising, a distribution of surface states [F cm2]

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &43&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

CH Double-layer capacitance (Helmholtz layer) [F] Ji(x,t) Flux of charge species i [cm2 s1]
CH,v Double-layer capacitance by volume unit [F cm3] joff Current density in the dark [A cm2]
c0 Standard concentration [cm3] jon Current density under illumination [A cm2]
ci,0 Dark equilibrium concentration of charge car- jph Photocurrent density [A cm2]
riers [cm3] jphEE Stationary photocurrent for EE illumination [A cm2]
Cs Experimental capacitance by area unit [F cm2] jphSE Stationary photocurrent for SE illumination [A cm2]
CSC Space charge capacitance [F] jph,n Photocurrent density for electrons [A cm2]
Csolid Capacitance associated with the solid [F] jph,p Photocurrent density for holes [A cm2]
CB Conduction band k Boltzmann constant
CBD Chemical bath deposition k Average pseudo-first-order rate constant [s1]
CE Coloration efficiency k(E) Microcanonical rate constants from their respective
CV Cyclic voltammetry averages [k] [s1]
CVD Chemical vapor deposition kback Rate constant for the RHC intermediate mediated re-
d Film thickness [cm] combination step [cm7 s1]
Di Diffusion coefficient [cm2 s1] kinj Rate constant for the current-doubling step [cm s1]
D(e) Density of states function [cm-3 eV1] kox Rate constant for the electron scavenging by the ex-
DOS Density of states [cm3] ternal oxidant [cm4 s1]
DSCs Dye-sensitized solar cells kr Rate constant for recombination of free electrons
DT Direct transfer, if a valence band free hole is adiabat- and trapped holes [cm3 s1]
ically transferred to pollutant molecules L Diffusion length of charge carriers [cm]
e Elementary charge [C] mc Effective mass of electrons in the conduction band
ef  Conduction band free electron n Concentration of conduction band electrons per
E Applied potential with respect to a reference elec- unit volume [cm3]
trode [V] N Total surface density of bridging oxygen
E0 Open-circuit potential in the dark [V] groups [cm2]
EC Potential corresponding to the bottom CB edge [V] Nc Effective density of states in the conduction
EF Potential corresponding to the Fermi level of the band [cm3]
electrode [V] ncontact Electron concentration in the electrode contact
EFB Flatband potential [V] under illumination [cm3]
EGB Potential corresponding to the GB level [V] nph Photogenerated electron concentration under illu-
dark
Eoc Stationary potential in the dark [V] mination [cm3]
Eoc ðtÞ Potential at any time after the illumination [V] n0 Electron concentration in the electrode contact in
Eph Open-circuit potential under illumination [V] the dark [cm3]
ERed/Ox Redox potential of the electrolyte [V] ND Donor density [cm3]
EV Potential corresponding to the top VB edge [V] Nf Surface concentration of surface-trapped
EE Electrolyte–electrode illumination direction holes [cm2]
EIS Electrochemical impedance spectroscopy NGB Total electrode volume density of monoenergetic
f Fraction of surface-trapped holes in the bridging states [cm3]
oxygen sites of TiO2 NP Nanoparticle
f(e-eF) Average occupancy or the Fermi–Dirac function NT Nanotube
FTO Fluorine tin oxide conducting glass NW Nanowire
FWHM Full width at half maximum OCP Open-circuit potential
Gi(x,t) Generation of charged species i [cm3 s1] OCVD Open-circuit photovoltage decay
GB Grain boundary ORR Oxygen reduction reaction
gss(e) Density of surface states function [cm3 eV1] Ox Dissolved oxidant
hf + Valence band free hole p Fraction of electrode (film) volume not occupied by
hs + Surface-trapped hole the SC
I Current [A] PEPS Photoelectrochemical photocurrent switching effect
IMPS Intensity modulated photocurrent spectroscopy Q Charge [C]
IMVS Intensity modulated photovoltage spectroscopy qv Charge per unit volume [C cm3]
IPCE Incident photon-to-current efficiency qs Charge per geometric (projected) electrode surface
IR Infrared area [C cm2]
IT Indirect transfer if a hole trapped at the semicon- Ri(x,t) Recombination of the charged species i (cm3 s1)
ductor surface is isoenergetically transferred to dis- RH2 Oxidizable species in solution
solved pollutant molecules RHaqC Reaction intermediate
jC Capacitive current density [A cm2] RRS Resonance Raman spectroscopy

&44& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

SC Semiconductor ackowledged. T.B. and M.J. also thank the Spanish MICINN for
SE Substrate–electrode illumination direction the award of a “Ramn y Cajal” contract and an FPI grant, re-
SEI Semiconductor/electrolyte interface spectively.
SERS Surface-enhanced Raman spectroscopy
t Time [s] Keywords: electrochemistry · heterogeneous catalysis ·
T Temperature [K] nanostructures · photochemistry · titanium dioxide
Tc Characteristic temperature that defines the broaden-
ing of the surface-state exponential distribution [K] [1] X. Chen, S. S. Mao, Chem. Rev. 2007, 107, 2891 – 2959.
v Potential scan rate [V s1] [2] A. A. Ismail, D. W. Bahnemann, J. Mater. Chem. 2011, 21, 11686 – 11707.
[3] F. Han, V. S. R. Kambala, M. Srinivasan, D. Rajarathnam, R. Naidu, Appl.
Voc Open-circuit voltage [V] Catal. A 2009, 359, 25 – 40.
Vph Photopotential [V] [4] D. P. Macwan, P. N. Dave, S. Chaturvedi, J. Mater. Sci. 2011, 46, 3669 –
VphEE Photopotential under EE illumination [V] 3686.
VphSE Photopotential under SE illumination [V] [5] S. Girish Kumar, L. Gomathi Devi, J. Phys. Chem. A 2011, 115, 13211 –
13241.
VB Valence band [6] O. Carp, C. L. Huisman, A. Reller, Prog. Solid State Chem. 2004, 32, 33 –
W Width of space charge region [cm] 177.
x Distance [cm] [7] M. Grtzel, Inorg. Chem. 2005, 44, 6841 – 6851.
zi Charge number of species i [8] S. Gnes, N. S. Sariciftci, Inorg. Chim. Acta 2008, 361, 581 – 588.
[9] J. H. Yum, P. Chen, M. Grtzel, M. K. Nazeeruddin, ChemSusChem 2008,
a(l) Wavelength-dependent absorption coeffi- 1, 699 – 707.
cient [cm1] [10] I. Mora-Ser, S. Gimnez, F. Fabregat-Santiago, R. Gmez, Q. Shen, T.
ac Coefficient describing the broadening of the elec- Toyoda, J. Bisquert, Acc. Chem. Res. 2009, 42, 1848 – 1857.
tron state distribution in the band gap [11] R. Abe, J. Photochem. Photobiol. C 2010, 11, 179 – 209.
[12] K. Maeda, K. Domen, J. Phys. Chem. Lett. 2010, 1, 2655 – 2661.
DH Electrostatic potential drop across the double [13] R. Beranek, Adv. Phys. Chem. 2011, 786759.
layer [V] [14] T. Djenizian, I. Hanzu, P. Knauth, J. Mater. Chem. 2011, 21, 9925 – 9937.
DSC Potential difference associated to the band energy [15] N. Nang Dinh, N. Minh Quyen, D. N. Chung, M. Zikova, V.-V. Truong,
bending [V] Sol. Energy Mater. Sol. Cells 2011, 95, 618 – 623.
[16] A. Fujishima, X. Zhang, D. A. Tryk, Surf. Sci. Rep. 2008, 63, 515 – 582.
Dsolid Potential drop over the interfacial part of the [17] K. Honda, J. Photochem. Photobiol. A 2004, 166, 63 – 68.
solid [V] [18] E. Becquerel, C. R. Acad. Sci.1839, 9, 561-567.
DOD Change in the optical density of the film between [19] R. Audubert, Les Piles Sensibles  l’Action de la Lumire, Hermann et
its colored and bleached states at a certain wave- Cie, Paris, 1931, p. 16.
[20] W. H. Brattain, C. G. B. Garrett, Bell Syst. Tech. J. 1955, 34, 129.
length [21] A. V. Byalobzheskii, V. D. Val’kov, Dokl. Akad. Nauk SSSR 1960, 134,
eF Electrochemical potential, or equivalently the Fermi 121 – 124.
level, of electrons in the electrode [eV] [22] A. Fujishima, S. Kikuchi, K. Honda, J. Chem. Soc. Jpn. Ind. Chem. Sect.
*
eF;n Quasi-Fermi level of electrons [eV] 1969, 72, 108 – 113.
[23] A. Fujishima, A. Sakamoto, K. Honda, Seisan Kenkyu 1969, 21, 450 –
*
eF;p Quasi-Fermi level of holes [eV] 452.
ec Lower CB edge energy [eV] [24] A. Fujishima, K. Honda, Bull. Chem. Soc. Jpn. 1971, 44, 1148 – 1150.
eFTO
F Applied potential (conducting substrate Fermi leve- [25] A. Fujishima, K. Honda, Nature 1972, 238, 37 – 38.
l) [eV] [26] A. Fujishima, E. Hayashitani, K. Honda, Seisan Kenkyu 1971, 23, 363 –
365.
220 Permittivity of the semiconductor [27] A. Fujishima, K. Kohayakawa, K. Honda, J. Electrochem. Soc. 1975, 122,
Vt Total volume density of traps [cm3] 1487 – 1489.
l Wavelength [cm] [28] B. O’Regan, J. Moser, M. Anderson, M. Grtzel, J. Phys. Chem. 1990, 94,
m Chemical potential of electrons [eV] 8720 – 8726.
[29] B. O’Regan, M. Grtzel, Nature 1991, 353, 737 – 740.
m0i Standard chemical potential of i [eV] [30] B. O’Regan, M. Grtzel, D. Fitzmaurice, J. Phys. Chem. 1991, 95, 10525 –
1c ðeÞ Density of states function in the conduction 10528.
band [cm3 eV1] [31] G. Rothenberger, D. Fitzmaurice, M. Grtzel, J. Phys. Chem. 1992, 96,
ti Average lifetime of species i [s] 5983 – 5986.
[32] G. Redmond, D. Fitzmaurice, M. Grtzel, J. Phys. Chem. 1993, 97, 6951 –
 Local electrostatic potential (inner or Galvani poten- 6954.
tial) [V] [33] K. Vinodgopal, S. Hotchandani, P. V. Kamat, J. Phys. Chem. 1993, 97,
F0 Monochromatic-light incident photon flux [cm2 s1] 9040 – 9044.
F1=2 ðzÞ Fermi–Dirac integral with index 1/2 [34] K. Vinodgopal, U. Stafford, K. A. Gray, P. V. Kamat, J. Phys. Chem. 1994,
98, 6797 – 6803.
[35] D. H. Kim, M. A. Anderson, Environ. Sci. Technol. 1994, 28, 479 – 483.
[36] A. Wahl, M. Ulmann, A. Carroy, J. Augustynski, J. Chem. Soc. Chem.
Commun. 1994, 2277 – 2278.
Acknowledgements [37] L. Kavan, K. Kratochvilov, M. Grtzel, J. Electroanal. Chem. 1995, 394,
93 – 102.
[38] F. Cao, G. Oskam, P. C. Searson, J. M. Stipkala, T. A. Heimer, F. Farzad,
Financial support from the Spanish Ministry of Science and Inno-
G. J. Meyer, J. Phys. Chem. 1995, 99, 11974 – 11980.
vation (MCINN) through projects MAT2009-14004 (Fondos FEDER) [39] G. K. Boschloo, A. Goossens, J. Schoonman, J. Electroanal. Chem. 1997,
and HOPE CSD2007-00007 (Consolider-Ingenio 2010) is gratefully 428, 25 – 32.

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &45&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

[40] S. K. Poznyak, A. I. Kokorin, A. I. Kulak, J. Electroanal. Chem. 1998, 442, [78] C. D. Lokhande, E.-H. Lee, K.-D. Jung, O.-S. Joo, J. Mater. Sci. 2004, 39,
99 – 105. 2915 – 2918.
[41] Y. Wang, Y. Hao, H. Cheng, J. Ma, B. Xu, W. Li, S. Cai, J. Mater. Sci. 1999, [79] B. Liu, E. S. Aydil, J. Am. Chem. Soc. 2009, 131, 3985 – 3990.
34, 2773 – 2779. [80] Q. Zeng, L. Wu, Y. Zhang, B. Qi, J. Zhi, Scr. Mater. 2010, 62, 810 – 813.
[42] S. Soedergren, A. Hagfeldt, J. Olsson, S.-E. Lindquist, J. Phys. Chem. [81] S. Yamabi, H. Imai, Chem. Mater. 2002, 14, 609 – 614.
1994, 98, 5552 – 5556. [82] T. Berger, T. Lana-Villarreal, D. Monllor-Satoca, R. Gmez, Chem. Phys.
[43] A. Solbrand, H. Lindstrçm, H. Rensmo, A. Hagfeldt, S.-E. Lindquist, J. Lett. 2007, 447, 91 – 95.
Phys. Chem. B 1997, 101, 2514 – 2518. [83] M. Jankulovska, T. Berger, T. Lana-Villarreal, R. Gmez, Electrochim. Acta
[44] H. Gerischer, Electrochim. Acta 1990, 35, 1677 – 1699. 2012, 62, 172 – 180.
[45] W. W. Dunn, Y. Aikawa, A. J. Bard, J. Am. Chem. Soc. 1981, 103, 3456 – [84] E. Hosono, S. Fujihara, K. Kakiuchi, H. Imai, J. Am. Chem. Soc. 2004,
3459. 126, 7790 – 7791.
[46] K. Shankar, J. I. Basham, N. K. Allam, O. K. Varghesem, G. K. Mor, X. [85] X. Feng, K. Shankar, O. K. Varghese, M. Paulose, T. J. Latempa, C. A.
Feng, M. Paulose, J. A. Seabold, K.-S. Choi, C. A. Grimes, J. Phys. Chem. Grimes, Nano Lett. 2008, 8, 3781 – 3786.
C 2009, 113, 6327 – 6359. [86] X. Yang, J. Zhuang, X. Li, D. Chen, G. Ouyang, Z. Mao, Y. Han, Z. He, C.
[47] J.G. Lu, P. Chang, Z. Fan, Mater. Sci. Eng. 2006, 52, 49 – 91. Liang, M. Wu, J. C. Yu, ACS Nano 2009, 3, 1212 – 1218.
[48] M. Hernndez-Vlez, Thin Solid Films 2006, 495, 51 – 63. [87] M. O. Abou-Helal, W. T. Seeber, Appl. Surf. Sci. 2002, 195, 53 – 62.
[49] S. V. N. T. Kuchibhatla, A. S. Karakoti, D. Bera, S. Seal, Prog. Mater. Sci. [88] M. Okuya, N. A. Prokudina, K. Mushika, S. Kaneko, J. Eur. Ceram. Soc.
2007, 52, 699 – 913. 1999, 19, 903 – 906.
[50] H. Tada, A. Hattori, Y. Tokihisa, K. Imai, N. Tohge, S. Ito, J. Phys. Chem. B [89] P. J. Cameron, L. M. Peter, J. Phys. Chem. B 2003, 107, 14394 – 14400.
2000, 104, 4585 – 4587. [90] H. Sun, C. Wang, S. Pang, X. Li, Y. Tao, H. Tang, M. Liu J. Non-Cryst.
[51] L. Andronic, A. Duta, Adv. Mater. Res. 2007, 23, 325 – 328. Solids 2008, 354, 1440 – 1443.
[52] I. M. Arabatzis, S. Antonarakj, T. Stergiopoulous, A. Hiskia, E. Papacon- [91] S. A. O’Neill, R. J. H. Clark, I. P. Parkin, N. Elliot, A. Mills, Chem. Mater.
stantinou, M. C. Bernard, P. Falaras, J. Photochem. Photobiol. A 2002, 2003, 15, 46 – 50.
149, 237 – 245. [92] K. Nakaso, K. Okuyama, M. Shimada, S. E. Pratsinis, Chem. Eng. Sci.
[53] L. Andronic, D. Andrasi, A. Enesca, M. Visa, A. Duta, J. Sol-Gel Sci. Tech- 2003, 58, 3327 – 3335.
nol. 2011, 58, 201 – 208. [93] H. Lee, M. Y. Song, J. Jurng, Y.-K. Park, Powder Technol. 2011, 214, 64 –
[54] G. Syrrokostas, M. Giannouli, P. Yianoulis, Renewable Energy 2009, 34, 68.
1759 – 1764. [94] A. Brevet, M. C. Marco de Lucas, V. Potin, R. Chassagnon, L. Imhoff, B.
[55] C. J. Barb, F. Arendse, P. Comte, M. Jirousek, F. Lenzmann, V. Shklover, Domenichini, S. Bourgeois, J. Phys. D 2009, 42, 175302.
M. Grtzel, J. Am. Ceram. Soc. 1997, 80, 3157 – 3171. [95] J.-H. Chang, A. V. Ellis, Y.-H. Hsieh, C.-H. Tung, S.-Y. Shen, Sci. Total Envi-
[56] M. K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker, E. Mller, P. ron. 2009, 407, 5914 – 5920.
Liska, N. Vlachopoulos, M. Grtzel, J. Am. Chem. Soc. 1993, 115, 6382 – [96] A.R. Zainun, S. Tomoya, U. M. Noor, M. Rusop, I. Masaya, Mater. Lett.
6390. 2012, 66, 254 – 256.
[57] B. van der Zanden, A. Goossens, J. Phys. Chem. B 2000, 104, 7171 – [97] T.-Y. Tsai, S.-Y. Lu, Electrochem. Commun. 2009, 11, 2180 – 2183.
7178. [98] S. Karuppuchamy, K. Nonomura, T. Yoshida, T. Sugiura, H. Minoura,
[58] D. S. Tsoukleris, I. M. Arabatzis, E. Chatzivasiloglou, A. I. Kontos, V. Be- Solid State Ionics 2002, 151, 19 – 27.
lessi, M. C. Bernard, P. Falaras, Sol. Energy 2005, 79, 422 – 430. [99] I. J. Zhitomirsky, L. Gal-Or, A. Kohn, H.W. Hennicke J.Mater. Sci. 1995,
[59] K. Fan, M. Liu, T. Peng, L. Ma, K. Dai, Renewable Energy 2010, 35, 555 – 30, 5307 – 5312.
561. [100] M.-S. Wu, M.-J. Wang, J.-J. Jow, W.-D. Yang, C.-Y. Hsieh, H.-M. Tsai, J.
[60] T. Ma, T. Kida, M. Akiyama, K. Inoue, S. Tsunematsu, K. Yao, H. Noma, E. Power Sources 2008, 185, 1420 – 1424.
Abe, Electrochem. Commun. 2003, 5, 369 – 372. [101] C.-C. Hu, C.-C. Huang, K.-H. Chang, Electrochem. Commun. 2009, 11,
[61] E. M. El-Maghraby, Y. Nakamura, S. Rengakuji, Catal. Commun. 2008, 9, 434 – 437.
2357 – 2360. [102] S. Sawatani, T. Yoshida, T. Ohya, T. Ban, Y. Takahashi, H. Minoura, Elec-
[62] H. Cui, H. S. Shen, Y. M. Gao, K. Dwight, A. Wold, Mater. Res. Bull. 1993, trochem. Solid-State Lett. 2005, 8, C69 – C71.
28, 195 – 201. [103] K. Wessels, M. Minnermann, J. Rathousky, M. Wark, T. Oekermann, J.
[63] R. Mechiakh, N. Ben Sedrine, R. Chtourou, R. Bensaha, Appl. Surf. Sci. Phys. Chem. C 2008, 112, 15122 – 15128.
2010, 257, 670 – 676. [104] M. S. Wu, C.-H. Tsai, J.-J. Jow, T.-C. Wei, Electrochim. Acta 2011, 56,
[64] K. K. Saini, S. D. Sharma, Chanderkant, M. Kar, D. Singh, C. P. Sharma, J. 8906 – 8911.
Non-Cryst. Solids 2007, 353, 2469 – 2473. [105] K. Wessels, A. Feldhoff, M. Wark, J. Rathousky, T. Oekermann, Electro-
[65] S. Sokolov, E. Ortel, J. Radnik, R. Kraehnert, Thin Solid Films 2009, 518, chem. Solid-State Lett. 2006, 9, C93 – C96.
27 – 35. [106] V. Zwilling, M. Aucouturier, E. Darque-Ceretti, Electrochim. Acta 1999,
[66] A. G. Agrios, I. Cesar, P. Comte, M. K. Nazeeruddin, M. Grtzel, Chem. 45, 921 – 929.
Mater. 2006, 18, 5395 – 5397. [107] V. Zwilling, E. Darque-Ceretti, A. Boutry-Forveille, D. David, M. Y. Perrin,
[67] J. Yu, X. Zhao, Q. Zhao, Thin Solid Films 2000, 379, 7 – 14. M. Aucouturier, Surf. Interface Anal. 1999, 27, 629 – 637.
[68] C. Su, B.-Y. Hong, C.-M. Tseng, Catal. Today 2004, 96, 119 – 126. [108] D. Gong, C. A. Grimes, O. K. Varghese, W. Hu, R. S. Singh, Z. Chen, E. C.
[69] R. Zhang, L. Gao, Q. Zhang, Chemosphere 2004, 54, 405 – 411. Dickey, J. Mater. Res. 2001, 16, 3331 – 3334.
[70] L. Spanhel, M. A. Anderson, J. Am. Chem. Soc. 1991, 113, 2826 – 2833. [109] G. K. Mor, O. K. Varghese, M. Paulose, K. Shankar, C. A. Grimes, Sol.
[71] B. B. Lakshmi, C. J. Patrissi, C. R. Martin, Chem. Mater. 1997, 9, 2544 – Energy Mater. Sol. Cells 2006, 90, 2011 – 2075.
2550. [110] P. Roy, S. Berger, P. Schmuki, Angew. Chem. 2011, 123, 2956 – 2995;
[72] B. B. Lakshmi, P. K. Dorhout, C. R. Martin, Chem. Mater. 1997, 9, 857 – Angew. Chem. Int. Ed. 2011, 50, 2904 – 2939.
862. [111] J. Bai, B. Zhou, L. Li, Y. Liu, Q. Zheng, J. Shao, X. Zhu, W. Cai, J. Liao, L.
[73] Z. Miao, D. Xu, J. Ouyang, G. Guo, X. Zhao, Y. Tang, Nano Lett. 2002, 2, Zou, J. Mater. Sci. 2008, 43, 1880 – 1884.
717 – 720. [112] O. K. Varghese, D. Gong, M. Paulose, C. A. Grimes, E. C. Dickey, J. Mater.
[74] L. Zhou, R. C. Hoffmann, Z. Zhao, J. Bill, F. Aldinger, Thin Solid Films Res. 2003, 18, 156 – 165.
2008, 516, 7661 – 7666. [113] S. P. Albu, A. Ghicov, J. M. Macak, P. Schmuki, Phys. Status Solidi RRL
[75] S. M. Pawar, B. S. Pawar, J. H. Kim, O.-S. Joo, C. D. Lokhande, Curr. Appl. 2007, 1, R65 – R67.
Phys. 2011, 11, 117 – 161. [114] M. Paulose, H. E. Prakasam, O. K. Varghese, L. Peng, K. C. Popat, G. K.
[76] A. M. More, T. P. Gujar, J. L. Gunjakar, C. D. Lokhande, O.-S. Joo, Appl. Mor, T. A. Desai, C. A. Grimes, J. Phys. Chem. C 2007, 111, 14992 – 14997.
Surf. Sci. 2008, 255, 2682 – 2687. [115] N. K. Allam, C. A. Grimes, J. Phys. Chem. C 2007, 111, 13028 – 13032.
[77] R. S. Mane, Y. H. Hwang, C. D. Lokhande, S. D. Sartale, S.-H. Han, Appl. [116] N. K. Allam, K. Shankar, C. A. Grimes, J. Mater. Chem. 2008, 18, 2341 –
Surf. Sci. 2005, 246, 271 – 278. 2348.

&46& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

[117] G. K. Mor, O. K. Varghese, M. Paulose, N. Mukherjee, C. A. Grimes, J. [154] D. Guan, C. Cai, Y. Wang, J. Nanosci. Nanotechnol. 2011, 11, 3641 –
Mater. Res. 2003, 18, 2588 – 2593. 3650.
[118] Q. Cai, M. Paulose, O. K. Varghese, C. A. Grimes, J. Mater. Res. 2005, 20, [155] Y. Sun, G. Wang, K. Yan, Int. J. Hydrogen Energy 2011, 36, 15502 –
230 – 236. 15508.
[119] G. K. Mor, K. Shankar, M. Paulose, O. K. Varghese, C. A. Grimes, Nano [156] Q.-Q. Meng, J.-G. Wang, Q. Xie, H.-Q. Dong, X.-N. Li, Catal. Today 2011,
Lett. 2005, 5, 191 – 195. 165, 145 – 149.
[120] G.-J. Wang, S.-W. Chou, Nanotechnology 2010, 21, 115206. [157] O. K. Varghese, M. Paulose, K. Shankar, G. K. Mor, C. A. Grimes, J. Nano-
[121] J.-H. Yum, S.-S. Kim, D.-Y. Kim, Y.-E. Sung, J. Photochem. Photobiol. A sci. Nanotechnol. 2005, 5, 1158 – 1165.
2005, 173, 1 – 6. [158] A. Ghicov, H. Tsuchiya, R. Hahn, J. M. Macak, A. G. MuÇoz, P. Schmuki,
[122] W. Tan, X. Yin, X. Zhou, J. Zhang, X. Xiao, Y. Lin, Electrochim. Acta 2009, Electrochem. Commun. 2006, 8, 528 – 532.
54, 4467 – 4472. [159] Y. V. Pleskov, Y. Y. Gurevich, Semiconductor Photoelectrochemistry, Con-
[123] Y.-S. Hu, L. Kienle, Y.-G. Guo, J. Maier, Adv. Mater. 2006, 18, 1421 – 1426. sultants Bureau, New York, 1986.
[124] D. S. Tsoukleris, A. I. Kontos, P. Aloupogiannis, P. Falaras, Catal. Today [160] D. Vanmaekelbergh, A. J. Houtepen, J. J. Kelly, Electrochim. Acta 2007,
2007, 124, 110 – 117. 53, 1140 – 1149.
[125] C. Euvananont, C. Junin, K. Inpor, P. Limthongkul, C. Thanachayanont, [161] A. Natarajan, G. Oskam, P. C. Searson, J. Phys. Chem. B 1998, 102,
Ceram. Int. 2008, 34, 1067 – 1071. 7793 – 7799.
[126] M. C. Carotta, M. Ferroni, V. Guidi, G. Martinelli, Adv. Mater. 1999, 11, [162] J. J. Kelly, D. Vanmaekelbergh, Electrochim. Acta 1998, 43, 2773 – 2780.
943 – 946. [163] A. Wahl, J. Augustynski, J. Phys. Chem. B 1998, 102, 7820 – 7828.
[127] M. Sasani Ghamsari, A. R. Bahramian, Mater. Lett. 2008, 62, 361 – 364. [164] F. Fabregat-Santiago, I. Mora-Ser, G. Garcia-Belmonte, J. Bisquert, J.
[128] N. N. Dinh, N. T. T. Oanh, P. D. Long, M. C. Bernard, A. H. L. Goff, Thin Phys. Chem. B 2003, 107, 758 – 768.
Solid Films 2003, 423, 70 – 76. [165] I. Abayev, A. Zaban, F. Fabregat-Santiago, J. Bisquert, Phys. Status Solidi
[129] R. Jourdani, D. Aitelhabti, E. L. Ameziane, S. Barazzouk, S. Hotchandani, A 2003, 196, R4 – R6.
Act. Passive Electron. Compon. 2004, 27, 125 – 131. [166] J. Bisquert, F. Fabregat-Santiago, I. Mora-Ser, G. Garcia-Belmonte,
[130] J. Zhu, J. Yang, Z.-F. Bian, J. Ren, Y.-M. Liu, Y. Cao, H.-X. Li, H.-Y. He, K.-N. E. M. Barea, E. Palomares, Inorg. Chim. Acta 2008, 361, 684 – 698.
Fan, Appl. Catal. B 2007, 76, 82 – 91. [167] V. G. Kytin, J. Bisquert, I. Abayev, A. Zaban, Phys. Rev. B 2004, 70,
[131] M. A. Reddy, M. S. Kishore, V. Pralong, V. Caignaert, U. V. Varadaraju, B. 1933041.
Raveau, Electrochem. Commun. 2006, 8, 1299 – 1303. [168] A. C. Fisher, L. M. Peter, E. A. Ponomarev, A. B. Walter, K. G. U. Wijayan-
[132] C. G. Granqvist, P. C. Lans ker, N. R. Mlyuka, G. A. Niklasson, E. Avenda- tha, J. Phys. Chem. B 2000, 104, 949 – 958.
[169] J. van de Lagemaat, A. J. Frank, J. Phys. Chem. B 2000, 104, 4292 – 4294.
Ço, Sol. Energy Mater. Sol. Cells 2009, 93, 2032 – 2039.
[170] J. Nelson, Phys. Rev. B 1999, 59, 15374 – 15380.
[133] G. K. Mor, K. Shankar, M. Paulose, O. K. Varghese, C. A. Grimes, Nano
[171] R. Kçnenkamp, Phys. Rev. B 2000, 61, 11057 – 11064.
Lett. 2006, 6, 215 – 218.
[172] A. L. Roest, J. J. Kelly, D. Vanmaekelbergh, E. A. Meulenkamp, Phys. Rev.
[134] J. Li, W. Wan, H. Zhou, J. Li, D. Xu, Chem. Commun. 2011, 47, 3439 –
Lett. 2002, 89, 0368011.
3441.
[173] G. Boschloo, D. Fitzmaurice, J. Phys. Chem. B 1999, 103, 2228 – 2231.
[135] A. R. Armstrong, G. Armstrong, J. Canales, P. G. Bruce, J. Power Sources
[174] L. M. Peter, N. W. Duffy, R. L. Wang, K. G. U. Wijayantha, J. Electroanal.
2005, 146, 501 – 506.
Chem. 2002, 524 – 525, 127 – 136.
[136] Z. Yang, G. Du, Z. Guo, X. Yu, Z. Chen, T. Guo, N. Sharma, H. Liu, Electro-
[175] M. Bailes, P. J. Cameron, K. Lobato, L. M. Peter, J. Phys. Chem. B 2005,
chem. Commun. 2011, 13, 46 – 49.
109, 15429 – 15435.
[137] Q. Mu, Y. Li, Q. Zhang, H. Wang, J. Hazard. Mater. 2011, 188, 363 – 368.
[176] G. Boschloo, A. Hagfeldt, J. Phys. Chem. B 2005, 109, 12093 – 12098.
[138] S. Pavasupree, S. Ngamsinlapasathian, M. Nakajima, Y. Suzuki, S. Yoshi-
[177] J. van de Lagemaat, N. G. Park, A. J. Frank, J. Phys. Chem. B 2000, 104,
kawa, J. Photochem. Photobiol. A 2006, 184, 163 – 169.
2044 – 2052.
[139] X. Gao, H. Zhu, G. Pan, S. Ye, Y. Lan, F. Wu, D. Song, J. Phys. Chem. B
[178] F. Fabregat-Santiago, G. Garcia-Belmonte, J. Bisquert, A. Zaban, P. Sal-
2004, 108, 2868 – 2872.
vador, J. Phys. Chem. B 2002, 106, 334 – 339.
[140] Z. Liu, B. Pesic, K. S. Raja, R. R. Rangaraju, M. Misra, Int. J. Hydrogen
[179] R. L. Willis, C. Olson, B. O’Regan, T. Lutz, J. Nelson, J. R. Durrant, J. Phys.
Energy 2009, 34, 3250 – 3257. Chem. B 2002, 106, 7605 – 7613.
[141] R. Hahn, A. Ghicov, H. Tsuchiya, J. M. Macak, A. G. MuÇoz, P. Schmuki, [180] H. Wang, J. He, G. Boschloo, H. Lindstrçm, A. Hagfeldt, S. E. Lindquist,
Phys. Status Solidi A 2007, 204, 1281 – 1285. J. Phys. Chem. B 2001, 105, 2529 – 2533.
[142] J. R. Li, Z. L. Tang, Z. T. Zhang, Electrochem. Solid-State Lett. 2005, 8, [181] G. Schlichthçrl, S. Y. Huang, J. Sprague, A. J. Frank, J. Phys. Chem. B
A316 – A319. 1997, 101, 8141 – 8155.
[143] P. S. Patil, S. H. Mujawar, A. I. Inamdar, S. B. Sadale, Appl. Surf. Sci. 2005, [182] G. Franco, J. Gehring, L. M. Peter, E. A. Ponomarev, I. Uhlendorf, J. Phys.
250, 117 – 123. Chem. B 1999, 103, 692 – 698.
[144] A. Sobczyk-Guzenda, M. Gazicki-Lipman, H. Szymanowski, J. Kowalski, [183] B. O’Regan, M. Grtzel, D. Fitzmaurice, Chem. Phys. Lett. 1991, 183, 89 –
P. Wojciechowski, T. Halamus, A. Tracz, Thin Solid Films 2009, 517, 93.
5409 – 5414. [184] G. Boschloo, D. Fitzmaurice, J. Phys. Chem. B 1999, 103, 7860 – 7868.
[145] J. Aarik, A. Aidla, A.-A. Kiisler, T. Uustare, J. Aarik, V. Sammelselg, Thin [185] T. Berger, J. A. Anta, V. Morales-Flrez, J. Phys. Chem. C 2012, 116,
Solid Films 1997, 305, 270 – 273. 11444 – 11455.
[146] M. Wu, Z. H. Yang, Y. H. Jiang, J. J. Zhang, S. Q. Liu, Y. M. Sun, J. Solid [186] J. Bisquert, A. Zaban, M. Greenshtein, I. Mora-Ser, J. Am. Chem. Soc.
State Electrochem. 2010, 14, 857 – 863. 2004, 126, 13550 – 13559.
[147] W.-J. Lee, M. Alhoshan, W. H. Smyrl, J. Electrochem. Soc. 2006, 153, [187] L. Kavan, M. Grtzel, S. E. Gilbert, C. Klemenz, H. J. Scheel, J. Am. Chem.
B499 – B505. Soc. 1996, 118, 6716 – 6723.
[148] J. M. Macak, H. Tsuchiya, A. Ghicov, K. Yasuda, R. Hahn, S. Bauer, P. [188] T. Berger, T. Lana-Villarreal, D. Monllor-Satoca, R. Gmez, J. Phys. Chem.
Schmuki, Curr. Opin. Solid State Mater. Sci. 2007, 11, 3 – 18. C 2007, 111, 9936 – 9942.
[149] M. Adachi, Y. Murata, M. Harada, S. Yoshikawa, Chem. Lett. 2000, 29, [189] J. Durrant, J. Photochem. Photobiol. A 2002, 148, 5 – 10.
942 – 943. [190] F. Fabregat-Santiago, J. Bisquert, G. Garcia-Belmonte, G. Boschloo, A.
[150] S. Z. Chu, S. Inoue, K. Wada, D. Li, H. Haneda, S. Awatsu, J. Phys. Chem. Hagfeldt, Sol. Energy Mater. Sol. Cells 2005, 87, 117 – 131.
B 2003, 107, 6586 – 6589. [191] E. Guilln, L. M. Peter, J. A. Anta, J. Phys. Chem. C 2011, 115, 22622 –
[151] M. Paulose, O. K. Varghese, G. K. Mor, C. A. Grimes, K. G. Ong, Nano- 22632.
technology 2006, 17, 398 – 402. [192] F. Fabregat-Santiago, H. Randriamahazaka, A. Zaban, J. Garc a-CaÇadas,
[152] K. Shankar, G. K. Mor, H. E. Prakasam, S. Yoriya, M. Paulose, O. K. Var- G. Garcia-Belmonte, J. Bisquert, Phys. Chem. Chem. Phys. 2006, 8,
ghese, C. A. Grimes, Nanotechnology 2007, 18, 065707. 1827 – 1833.
[153] J. Yan, F. Zhou, J. Mater. Chem. 2011, 21, 9406 – 9418. [193] S. Ardo, G. J. Meyer, Chem. Soc. Rev. 2009, 38, 115 – 164.

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &47&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

[194] M. A. Henderson, Surf. Sci. Rep. 2011, 66, 185 – 297. [232] J.-J. Lee, G. M. Coia, N. S. Lewis, J. Phys. Chem. B 2004, 108, 5269 –
[195] K. Westermark, A. Henningsson, H. Rensmo, S. Sodergren, H. Siegbahn, 5281.
A. Hagfeldt, Chem. Phys. 2002, 285, 157 – 165. [233] E. Pelizzetti, C. Minero, Electrochim. Acta 1993, 38, 47 – 55.
[196] G. Liu, W. Jgermann, J. He, V. Sundstrçm, L. Sun, J. Phys. Chem. B [234] A. J. Bard, R. Parsons, J. Jordan, Standard Potentials in Aqueous Solu-
2002, 106, 5814 – 5819. tions, Marcel Dekker, New York, 1985.
[197] L. M. Peter, Acc. Chem. Res. 2009, 42, 1839 – 1847. [235] A. J. Bard, L. R. Faulkner, Electrochemical Methods, 2nd ed., Wiley, New
[198] I. Abayev, A. Zaban, V. G. Kytin, A. A. Danilin, G. Garcia-Belmonte, J. Bis- York, 2001.
quert, J. Solid State Electrochem. 2007, 11, 647 – 653. [236] J. Nelson in Encyclopedia of Electrochemistry, Vol. 6 (Eds.: A. J. Bard, M.
[199] F. Marken, A. S. Bhambra, D. H. Kim, R. J. Mortimer, S. J. Stott, Electro- Stratmann), Wiley-VCH, Weinheim, 2002.
chem. Commun. 2004, 6, 1153 – 1158. [237] N. Sato, Electrochemistry at Metal and Semiconductor Electrodes, Elsevi-
[200] T. Berger, T. Lana-Villarreal, D. Monllor-Satoca, R. Gmez, Electrochem. er, Amsterdam, 1998.
Commun. 2006, 8, 1713 – 1718. [238] L. M. Peter, Phys. Chem. Chem. Phys. 2007, 9, 2630 – 2642.
[201] T. Lana-Villarreal, Y. Mao, S. S. Wong, R. Gmez, Nanoscale 2010, 2, [239] J. Bisquert, Phys. Chem. Chem. Phys. 2008, 10, 49 – 72.
1690 – 1698. [240] L. M. Peter, J. Phys. Chem. C 2007, 111, 6601 – 6612.
[202] L. A. Lyon, J. T. Hupp, J. Phys. Chem. B 1999, 103, 4623 – 4628. [241] N. Kopidakis, E. A. Schiff, N.-G. Park, J. van de Lagemaat, A. J. Frank, J.
[203] L. A. Lyon, J. T. Hupp, J. Phys. Chem. 1995, 99, 15718 – 15720. Phys. Chem. B 2000, 104, 3930 – 3936.
[204] S. Y. Huang, L. Kavan, I. Exnar, M. Grtzel, J. Electrochem. Soc. 1995, [242] S. Nakade, S. Lambe, T. Kitamura, Y. Wada, S. Yanagida, J. Phys. Chem. B
142, L142 – L144. 2001, 105, 9150 – 9152.
[205] J. Kang, S. H. Wei, K. Zhu, Y. H. Kim, J. Phys. Chem. C 2011, 115, 4909 – [243] S. Nakade, S. Lambe, M. Matsuda, Y. Saito, T. Kitamura, Y. Wada, S. Ya-
4915. nagida, Phys. E 2002, 14, 210 – 214.
[206] A. Krlikowska, P. Barczuk, R. Jurczakowski, J. Augustynski, J. Electroa- [244] S. Lambe, S. Nakade, T. Kitamura, Y. Wada, S. Yanagida, J. Phys. Chem. B
nal. Chem. 2011, 662, 229 – 239. 2002, 106, 2967 – 2972.
[207] L. de La Garza, Z. V. Saponjic, N. M. Dimitrijevic, M. Thurnauer, T. Rajh, J. [245] J. Bisquert, V. S. Vilhrenko, J. Phys. Chem. B 2004, 108, 2313 – 2322.
Phys. Chem. B 2006, 110, 680 – 686. [246] N. W. Duffy, L. M. Peter, K. G. U. Wijayantha, Electrochem. Commun.
[208] A. Kay, R. Humphry-Baker, M. Grtzel, J. Phys. Chem. 1994, 98, 952 – 2000, 2, 262 – 266.
959. [247] L. Dloczik, O. Ileperuma, I. Lauermann, L. M. Peter, E. A. Ponomarev, G.
[209] F. Fabregat-Santiago, G. Garcia-Belmonte, J. Bisquert, P. Bogdanoff, A. Redmond, N. J. Shaw, I. Uhlendorf, J. Phys. Chem. B 1997, 101, 10281 –
Zaban, J. Electrochem. Soc. 2003, 150, E293 – E298. 10289.
[248] V. Duzhko, V. Y. Timoshenko, F. Koch, T. Dittrich Phys. Rev. B 2001, 64,
[210] V. M. Khomenko, K. Langer, H. Rager, A. Fett, Phys. Chem. Miner. 1998,
075204.
25, 338 – 346.
[249] S.-E. Lindquist, B. Finnstrçm, L. Tegnr, J. Electrochem. Soc. 1983, 130,
[211] D. A. Panayotov, S. P. Burrows, J. T. Yates, J. R. Morris, J. Phys. Chem. C
351 – 358.
2011, 115, 22400 – 22408.
[250] C. Minero, Catal. Today 1999, 54, 205 – 216.
[212] T. Berger, M. Sterrer, O. Diwald, E. Knçzinger, D. Panayotov, T. L.
[251] D. Monllor-Satoca, R. Gmez, M. Gonzlez-Hidalgo, P. Salvador, Catal.
Thompson, J. T. Yates, J. Phys. Chem. B 2005, 109, 6061 – 6068.
Today 2007, 129, 247 – 255.
[213] S. Szczepankiewicz, J. A. Moss, M. R. Hoffmann, J. Phys. Chem. B 2002,
[252] T. Lana-Villarreal, R. Gmez, M. Neumann-Spallart, N. Alonso-Vante, P.
106, 2922 – 2927.
Salvador, J. Phys. Chem. B 2004, 108, 15172 – 15181.
[214] T. Yoshihara, R. Katoh, A. Furube, Y. Tamaki, M. Murai, K. Hara, S.
[253] N. Hykaway, W. M. Sears, H. Morisaki, S. R. Morrison, J. Phys. Chem.
Murata, H. Arakawa, M. Tachiya, J. Phys. Chem. B 2004, 108, 3817 –
1986, 90, 6663 – 6667.
3823.
[254] A. Imanishi, T. Okamura, N. Ohashi, R. Nakamura, Y. Nakato, J. Am.
[215] D. S. Warren, A. J. McQuillan, J. Phys. Chem. B 2004, 108, 19373 – 19379.
Chem. Soc. 2007, 129, 11569 – 11578.
[216] D. M. Savory, D. S. Warren, A. J. McQuillan, J. Phys. Chem. C 2011, 115,
[255] I. Mora-Ser, T. Lana-Villarreal, J. Bisquert, . Pitarch, R. Gmez, P. Salva-
902 – 907.
dor, J. Phys. Chem. B 2005, 109, 3371 – 3380.
[217] A. Hagfeldt, N. Vlachopoulos, M. Grtzel, J. Electrochem. Soc. 1994, 141,
[256] G. Waldner, R. Gmez, M. Neumann-Spallart, Electrochim. Acta 2007,
L82 – L84. 52, 2634 – 2639.
[218] M. R. Hoffmann, S. T. Martin, W. Choi, D. W. Bahnemann, Chem. Rev. [257] T. Lana-Villarreal, R. Gmez, M. Gonzlez-Hidalgo, P. Salvador, J. Phys.
1995, 95, 69 – 96. Chem. B 2004, 108, 20278 – 20290.
[219] M. A. Fox, M. T. Dulay, Chem. Rev. 1993, 93, 341 – 357. [258] D. Monllor-Satoca, L. Borja, A. Rodes, R. Gmez, P. Salvador, ChemPhys-
[220] J. M. Macak, P. J. Barczuk, H. Tsuchiya, M. Z. Nowakowska, A. Ghicov, M. Chem 2006, 7, 2540 – 2551.
Chojak, S. Bauer, S. Virtanen, P. J. Kulesza, P. Schmuki, Electrochem. [259] T. Berger, T. Lana-Villarreal, D. Monllor-Satoca, R. Gmez, J. Phys. Chem.
Commun. 2005, 7, 1417 – 1422. C 2008, 112, 15920 – 15928.
[221] K. W. Park, K. S. Seol, Electrochem. Commun. 2007, 9, 2256 – 2260. [260] D. Monllor-Satoca, R. Gmez, Electrochim. Acta 2010, 55, 4661 – 4668.
[222] H. Chhina, S. Campbell, O. Kesler, J. Electrochem. Soc. 2009, 156, [261] A. Solbrand, A. Henningsson, S. Sçdergren, H. Lindstrçm, A. Hagfeldt,
B1232 – B1237. S.-E. Lindquist, J. Phys. Chem. B 1999, 103, 1078 – 1083.
[223] S.-Y. Huang, P. Ganesan, S. Park, B. N. Popov, J. Am. Chem. Soc. 2009, [262] A. J. Frank, N. Kopidakis, J. van de Lagemaat, Coord. Chem. Rev. 2004,
131, 13898 – 13899. 248, 1165 – 1179.
[224] S. Shanmugam, A. Gedanken, J. Phys. Chem. C 2009, 113, 18707 – [263] S. E. Lindquist, A. Hagfeldt, S. Sçdergren, H. Lindstrçm, Charge trans-
18712. port in nanostructured thin-film electrodes in Electrochemistry of Nano-
[225] A. Bauer, K. Lee, C. Song, Y. Xie, J. Zhang, R. Hui, J. Power Sources materials (Ed.: G. Hodes), Wiley-VCH, Weinheim, 2001.
2010, 195, 3105 – 3110. [264] W. H. Leng, P. R. F. Barnes, M. Juozapavicius, B. C. O’Regan, J. R. Dur-
[226] D. Chu, M. Xu, J. Lu, P. Zheng, G. Qin, X. Yuan, Electrochem. Commun. rant, J. Phys. Chem. Lett. 2010, 1, 967 – 972.
2008, 10, 350 – 353. [265] T. Lana-Villarreal, D. Monllor-Satoca, R. Gmez, P. Salvador, Electrochem.
[227] D. Chu, Y. Hou, J. He, M. Xu, Y. Wang, S. Wang, J. Wang, L. Zha, J. Nano- Commun. 2006, 8, 1784 – 1790.
part. Res. 2009, 11, 1805 – 1809. [266] T. Lana-Villarreal, R. Gmez, Chem. Phys. Lett. 2005, 414, 489 – 494.
[228] C. A. Koval, J. N. Howard, Chem. Rev. 1992, 92, 411 – 433. [267] G. Hodes, I. D. J. Howell, L. M. Peter, J. Electrochem. Soc. 1992, 139,
[229] P. V. Kamat, K. Tvrdy, D. R. Baker, J. G. Radich, Chem. Rev. 2010, 110, 3136 – 3140.
6664 – 6688. [268] A. Tsujiko, H. Itoh, T. Kisumi, A. Shiga, K. Murakoshi, Y. Nakato, J. Phys.
[230] A. Hagfeldt, H. Lindstrçm, S. Sçdergren, S.-E. Lindquist, J. Electroanal. Chem. B 2002, 106, 5878 – 5885.
Chem. 1995, 381, 39 – 46. [269] K. Szaciłowski, W. Macyk, Solid-State Electron. 2006, 50, 1649 – 1655.
[231] D. Vanmaekelbergh, P. E. de Jongh, J. Phys. Chem. B 1999, 103, 747 – [270] P. A. Mandelbaum, A. E. Regazzoni, M. A. Blesa, J. Phys. Chem. B 1999,
750. 103, 5505 – 5511.

&48& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

[271] P. Mandelbaum, S. A. Bilmes, A. E. Regazzoni, M. A. Blesa, Solar Energy [315] M. X. Tan, R. E. Laibinis, S. T. Nguyen, J. M. Kesselman, C. E. Stanton,
1999, 65, 75 – 80. N. S. Lewis, Prog. Inorg. Chem. 1994, 41, 21 – 144.
[272] M. E. Calvo, R. J. Candal, S. A. Bilmes, Catal. Today 2002, 76, 133 – 139. [316] A. Islam, H. Sugihara, K. Hara, L. P. Singh, R. Katoh, M. Yanagida, Y. Taka-
[273] M. E. Calvo, R. J. Candal, S. A. Bilmes, Environ. Sci. Technol. 2001, 35, hashi, S. Murata, H. Arakawa, G. Fujihashi, Inorg. Chem. 2001, 40,
4132 – 4138. 5371 – 5380.
[274] W. Y. Gan, M. W. Lee, R. Amal, H. Zhao, K. Chiang, J. Appl. Electrochem. [317] A. Hagfeldt, M. Grtzel, Acc. Chem. Res. 2000, 33, 269 – 277.
2008, 38, 703 – 712. [318] M. Neumann-Spallart, Chimia 2007, 61, 806 – 809.
[275] W. Y. Gan, D. Friedmann, R. Amal, S. Zhang, K. Chiang, H. Zhao, Chem. [319] K. Rajeshwar, N. R. De Tacconi, C. R. Chenthamarakshan, Chem. Mater.
Eng. J. 2010, 158, 482 – 488. 2001, 13, 2765 – 2782.
[276] H. Zhang, P. Liu, X. Liu, S. Zhang, X. Yao, T. An, R. Amal, H. Zhao, Lang- [320] P. V. Kamat, J. Phys. Chem. C 2007, 111, 2834 – 2860.
muir 2010, 26, 11226 – 11232. [321] H. Zhang, G. Chen, D. W. Bahnemann, J. Mater. Chem. 2009, 19, 5089 –
[277] W. Y. Gan, H. Zhao, R. Amal, Appl. Catal. A 2009, 354, 8 – 16. 5121.
[278] D. Jiang, H. Zhao, Z. Jia, J. Cao, R. John, J. Photochem. Photobiol. A [322] A. N. Banerjee, Nanotechnol. Sci. Appl. 2011, 4, 35 – 65.
2001, 144, 197 – 204. [323] A. A. Aal, H. B. Hassan, M. A. Abdel Rahim, J. Electroanal. Chem. 2008,
[279] D. Jiang, H. Zhao, S. Zhang, R. John, J. Phys. Chem. B 2003, 107, 619 – 620, 17 – 25.
12774 – 12780. [324] J. Georgieva, S. Armyanov, E. Valova, I. Poulios, S. Sotiropoulos, Electro-
[280] D. Jiang, H. Zhao, S. Zhang, R. John, J. Catal. 2004, 223, 212 – 220. chem. Commun. 2007, 9, 365 – 370.
[281] D. Jiang, H. Zhao, S. Zhang, R. John, G. D. Will, J. Photochem. Photobiol. [325] J. Georgieva, S. Sotiropoulos, S. Armyanov, N. Philippidis, I. Poulios, J.
A 2003, 156, 201 – 206. Appl. Electrochem. 2011, 41, 173 – 181.
[282] H. Zhang, H. Zhao, S. Zhang, X. Quan, ChemPhysChem 2008, 9, 117 – [326] Y. Sun, E. Hao, X. Zhang, B. Yang, J. Shen, L. Chi, H. Fuchs, Langmuir
123. 1997, 13, 5168 – 5174.
[283] A. Solbrand, K. Keis, S. Sçdergren, H. Lindstrçm, S.-E. Lindquist, A. Hag- [327] H. Gerischer, M. Lbke, J. Electroanal. Chem. 1986, 204, 225 – 227.
feldt, Sol. Energy Mater. Sol. Cells 2000, 60, 181 – 193. [328] S. Chen, M. Paulose, C. Ruan, G. K. Mor, O. K. Varghese, D. Kouzoudis,
[284] P. E. de Jongh, D. Vanmaekelbergh, Phys. Rev. Lett. 1996, 77, 3427 – C. A. Grimes, J. Photochem. Photobiol. A 2006, 177, 177 – 184.
3430. [329] D. R. Baker, P. V. Kamat, Adv. Funct. Mater. 2009, 19, 805 – 811.
[285] P. E. de Jongh, D. Vanmaekelbergh, J. Phys. Chem. B 1997, 101, 2716 – [330] I. Bedja, P. V. Kamat, J. Phys. Chem. 1995, 99, 9182 – 9188.
2722. [331] K. Vinodgopal, I. Bedja, P. V. Kamat, Chem. Mater. 1996, 8, 2180 – 2187.
[286] D. Vanmaekelbergh, P. E. de Jongh, Phys. Rev. B 2000, 61, 4699 – 4704. [332] Y. Cao, X. Zhang, W. Yang, H. Du, Y. Bai, T. Li, J. Yao, Chem. Mater. 2000,
[287] W. P. Gomes, D. Vanmaekelbergh, Electrochim. Acta 1996, 41, 967 – 973.
12, 3445 – 3448.
[288] J. Bisquert, G. Garcia-Belmonte, F. Fabregat-Santiago, N. S. Ferriols, P.
[333] D. Liu, P. V. Kamat, J. Electroanal. Chem. 1993, 347, 451 – 456.
Bogdanoff, E. C. Pereira, J. Phys. Chem. B 2000, 104, 2287 – 2298.
[334] D. Liu, P. V. Kamat, J. Phys. Chem. 1993, 97, 10769 – 10773.
[289] J. Bisquert, J. Phys. Chem. B 2002, 106, 325 – 333.
[335] R. Beranek, H. Kisch, Angew. Chem. 2008, 120, 1340 – 1342; Angew.
[290] W. H. Leng, Z. Zhang, J. Q. Zhang, C. N. Cao, J. Phys. Chem. B 2005,
Chem. Int. Ed. 2008, 47, 1320 – 1322.
109, 15008 – 15023.
[336] A. Henglein, Chem. Rev. 1989, 89, 1861 – 1873.
[291] H. Liu, S. Cheng, M. Wu, H. Wu, J. Zhang, W. Li, C. Cao, J. Phys. Chem. A
[337] A. L. Linsebigler, G. Lu, J. T. Yates, Chem. Rev. 1995, 95, 735 – 758.
2000, 104, 7016 – 7020.
[338] R. S. Dibbell, G. R. Soja, R. M. Hoth, D. F. Watson, Langmuir 2007, 23,
[292] H. Liu, X. Z. Li, Y. J. Leng, W. Z. Li, J. Phys. Chem. B 2003, 107, 8988 –
3432 – 3439.
8996.
[339] Q. Shen, D. Arae, T. Toyoda, J. Photochem. Photobiol. A 2004, 164, 75 –
[293] N. Baram, Y. Ein-Eli, J. Phys. Chem. C 2010, 114, 9781 – 9790.
80.
[294] Z. Hosseini, N. Taghavinia, N. Sharifi, M. Chavoshi, M. Rahnman, J. Phys.
[340] H. Tada, M. Fujishima, H. Kobayashi, Chem. Soc. Rev. 2011, 40, 4232 –
Chem. C 2008, 112, 18686 – 18689.
4243.
[295] L. Peter, K. G. U. Wijayantha, Electrochem. Commun. 1999, 1, 576 – 580.
[341] T. Lana-Villarreal, R. Gmez, Electrochem. Commun. 2005, 7, 1218 –
[296] E. Marchante, T. Lana-Villarreal, V. Sez, J. Gonzlez-Garc a, R. Gmez,
1224.
Chem. Commun. 2009, 4127 – 4129.
[297] N. S. Lewis, J. Electrochem. Soc. 1984, 131, 2496 – 2503. [342] M. Anpo, M. Takeuchi, J. Catal. 2003, 216, 505 – 516.
[298] A. Zaban, M. Greenshtein, J. Bisquert, ChemPhysChem 2003, 4, 859 – [343] N. Chandrasekharan, P. V. Kamat, J. Phys. Chem. B 2000, 104, 10851 –
864. 10857.
[299] I. Mora-Ser, T. Dittrich, A. Belaidi, G. Garc a-Belmonte, J. Bisquert, J. [344] V. Subramanian, E. Wolf, P. V. Kamat, J. Phys. Chem. B 2001, 105,
Phys. Chem. B 2005, 109, 14932 – 14938. 11439 – 11446.
[300] M. Neumann-Spallart, O. Eneo, J. Electrochem. Soc. 1984, 131, 2767 – [345] C. He, Y. Xiong, X. H. Zhu, Thin Solid Films 2002, 422, 235 – 238.
2772. [346] C. He, X. Li, Y. Xiong, X. Zhu, S. Liu, Chemosphere 2005, 58, 381 – 389.
[301] J. Desilvestro, M. Neumann-Spallart, J. Phys. Chem. 1985, 89, 3684 – [347] P. V. Kamat, J. Phys. Chem. C 2008, 112, 18737 – 18753.
3689. [348] D. Yu, E. Nogelli, F. Du, L. Dai, J. Phys. Chem. Lett. 2010, 1, 2165 – 2173.
[302] J. A. Byrne, B. R. Eggins, J. Electroanal. Chem. 1998, 457, 61 – 72. [349] Y. Park, N. J. Singh, K. S. Kim, T. Tachikawa, T. Majima, W. Choi, Chem.
[303] J. A. Byrne, A. Davidson, S. P. M. Dunlop, B. R. Eggins, J. Photochem. Eur. J. 2009, 15, 10843 – 10850.
Photobiol. A 2002, 148, 365 – 374. [350] A. Kongkanand, R. M. Dom nguez, P. V. Kamat, Nano Lett. 2007, 7,
[304] R. Gmez, P. Salvador, Sol. Energy Mater. Sol. Cells 2005, 88, 377 – 388. 676 – 680.
[305] D. Monllor-Satoca, R. Gmez, J. Phys. Chem. C 2008, 112, 139 – 147. [351] S. Barazzouk, S. Hotchandani, K. Vinodgopal, P. V. Kamat, J. Phys. Chem.
[306] N. Guijarro, Q. Shen, S. Gimnez, I. Mora-Ser, J. Bisquert, T. Lana-Villar- B 2004, 108, 17015 – 17018.
real, T. Toyoda, R. Gmez, J. Phys. Chem. C 2010, 114, 22352 – 22360. [352] Y. Park, S.-H. Kang, W. Choi, Phys. Chem. Chem. Phys. 2011, 13, 9425 –
[307] T. Brgi, A. Baiker, Adv. Catal. 2006, 50, 227 – 283. 9431.
[308] P. Araujo, C. B. Mendive, L. A. Garc a Rodenas, P. J. Morando, A. E. Re- [353] Y. H. Ng, I. V. Lightcap, K. Goodwin, M. Matsumura, P. V. Kamat, J. Phys.
gazzoni, M. A. Blesa, D. Bahnemann, Colloids Surf. A 2005, 265, 73 – 80. Chem. Lett. 2010, 1, 2222 – 2227.
[309] T. Lana-Villarreal, J. M. Prez, R. Gmez, Surf. Sci. 2004, 572, 329 – 336. [354] H.-i. Kim, G.-h. Moon, D. Monllor-Satoca, Y. Park, W. Choi, J. Phys. Chem.
[310] T. Lana-Villarreal, J. M. Prez, R. Gmez, C. R. Chim. 2006, 9, 806 – 816. C 2012, 116, 1535 – 1543.
[311] T. Lana-Villarreal, A. Rodes, J. M. Prez, R. Gmez, J. Am. Chem. Soc. [355] J. Boucl, P. Ravirajan, J. Nelson, J. Mater. Chem. 2007, 17, 3141 – 3153.
2005, 127, 12601 – 12611. [356] T. Lana-Villarreal, D. Monllor-Satoca, A. Rodes, R. Gmez, Catal. Today
[312] T. Berger, A. Rodes, R. Gmez, Chem. Commun. 2010, 46, 2992 – 2994. 2007, 129, 86 – 95.
[313] T. Berger, A. Rodes, R. Gmez, Phys. Chem. Chem. Phys. 2010, 12, [357] H. Xia, Q. Wang, Chem. Mater. 2002, 14, 2158 – 2165.
10503 – 10511. [358] M. R. Nabid, M. Golbabaee, A. B. Moghaddam, R. Dinarvand, R. Sedghi,
[314] T. Berger, J. A. Anta, Anal. Chem. 2012, 84, 3053 – 3057. Int. J. Electrochem. Sci. 2008, 3, 1117 – 1126.

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &49&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

[359] S.-Y. Kwak, S. H. Kim, S. S. Kim, Environ. Sci. Technol. 2001, 35, 2388 – [398] L. Kavan, J. Rathousky, M. Grtzel, V. Shklover, A. Zukal, J. Phys. Chem. B
2394. 2000, 104, 12012 – 12020.
[360] Y. Liu, A. Wang, R. Claus, J. Phys. Chem. B 1997, 101, 1385 – 1388. [399] S. Huang, L. Kavan, I. Exnar, M. Grtzel, J. Electrochem. Soc. 1995, 142,
[361] R. Beranek, H. Kisch, Electrochem. Commun. 2007, 9, 761 – 766. L142 – L144.
[362] R. Beranek, H. Kisch, Photochem. Photobiol. Sci. 2008, 7, 40 – 48. [400] W. J. Macklin, R. J. Neat, Solid State Ionics 1992, 53 – 56, 694 – 700.
[363] R. Beranek, J. M. Macak, M. Grtner, K. Meyer, P. Schmuki, Electrochim. [401] U. Lafont, D. Carta, G. Mountjoy, A. V. Chadwick, E. M. Kelder, J. Phys.
Acta 2009, 54, 2640 – 2646. Chem. C 2010, 114, 1372 – 1378.
[364] P. Calza, E. Pelizzetti, Pure Appl. Chem. 2001, 73, 1839 – 1848. [402] E. Baudrin, S. Cassaignon, M. Koelsch, J.-P. Jolivet, L. Dupont, J.-M. Tar-
[365] M. Abdullah, G. K.-C. Low, R. W. Matthews, J. Phys. Chem. 1990, 94, ascon, Electrochem. Commun. 2007, 9, 337 – 342.
6820 – 6825. [403] S. Bach, J. P. Pereira-Ramos, P. Willman, Electrochim. Acta 2010, 55,
[366] H. Ross, J. Bendig, S. Hecht, Sol. Energy Mater. Sol. Cells 1994, 33, 475 – 4952 – 4959.
481. [404] S. Panero, B. Scrosati, M. Wachtler, F. Croce, J. Power Sources 2004, 129,
[367] J. Lobedank, E. Bellmann, J. Bendig, J. Photochem. Photobiol. A 1997, 90 – 95.
108, 89 – 93. [405] M. Wagemaker, W. J. H. Borghols, F. M. Mulder, J. Am. Chem. Soc. 2007,
[368] M. Stylidi, D. I. Kondarides, X. E. Verykios, Appl. Catal. B 2004, 47, 189 – 129, 4323 – 4327.
201. [406] Y. K. Zhou, L. Cao, F. B. Zhang, B. L. He, H. L. Li, J. Electrochem. Soc.
[369] J. J. Concepcion, J. W. Jurss, M. K. Brennaman, P. G. Hoertz, A. O. T. Pa- 2003, 150, A1246 – A1249.
trocinio, N. Y. M. Iha, J. L. Templeton, T. J. Meyer, Acc. Chem. Res. 2009, [407] A. R. Armstrong, G. Armstrong, J. Canales, R. Garc a, P. G. Bruce, Adv.
42, 1954 – 1965. Mater. 2005, 17, 862 – 865.
[370] W. Song, M. K. Brennaman, J. J. Concepcion, J. W. Jurss, P. G. Hoertz, H. [408] G. Armstrong, A. R. Armstrong, P. G. Bruce, P. Reale, B. Scrosati, Adv.
Luo, C. Chen, K. Hanson, T. J. Meyer, J. Phys. Chem. C 2011, 115, 7081 – Mater. 2006, 18, 2597 – 2600.
7091. [409] M. Zukalov, M. Kalbc, L. Kavan, I. Exnar, M. Graetzel, Chem. Mater.
[371] H. Park, W. Choi, J. Phys. Chem. B 2004, 108, 4086 – 4093. 2005, 17, 1248 – 1255.
[372] D. Monllor-Satoca, T. Lana-Villarreal, R. Gmez, Langmuir 2011, 27, [410] K. Xu, Chem. Rev. 2004, 104, 4303 – 4418.
15312 – 15321. [411] N. H. Idris, Md. M. Rahman, J.-Z. Wang, H.-K. Liu, J. Power Sources 2012,
[373] M. I. Franch, J. Peral, X. Dom nech, J. A. Aylln, Chem. Commun. 2005, 201, 294 – 300.
1851 – 1853. [412] A. M. Stephan, K. S. Nahm, Polymer 2006, 47, 5952 – 5964.
[374] V. Maurino, C. Minero, E. Pelizzetti, G. Mariella, A. Arbezzano, F. Ruber- [413] S. Kalnaus, A. S. Sabau, W. E. Tenhaeff, N. J. Dudney, C. Daniel, J. Power
Sources 2012, 201, 280 – 287.
telli, Res. Chem. Intermed. 2007, 33, 319 – 332.
[414] C. G. Granqvist, A. Azens, P. Heszler, L. B. Kish, L. sterlund, Sol. Energy
[375] H. Park, W. Choi, J. Phys. Chem. B 2003, 107, 3885 – 3890.
Mater. Sol. Cells 2007, 91, 355 – 365.
[376] J. D. Zhao, C. Chen, Y. Wang, H. Ji, W. Ma, L. Zang, J. Zhao, J. Phys.
[415] N. N. Dinh, N. M. Quyen, D. N. Chung, M. Zikova, V.-V. Truong, Sol.
Chem. C 2008, 112, 5993 – 6001.
Energy Mater. Sol. Cells 2011, 95, 618 – 623.
[377] I. C. Cheon, M. S. Lee, H. K. Shim, Y. I. Kim, Bull. Korean Chem. Soc.
[416] C. G. Granqvist, Appl. Phys. A 1991, 52, 83 – 93.
2003, 24, 1535 – 1537.
[417] J. Zhang, X. L. Wang, X. H. Xia, C. D. Gu, Z. J. Zhao, J. P. Tu, Electrochim.
[378] H. Pelouchova, P. Janda, J. Weber, L. Kavan, J. Electroanal. Chem. 2004,
Acta 2010, 55, 6953 – 6958.
566, 73 – 83.
[418] V. V. Abramova, A. S. Sinitskii, E. A. Goodilin, Y. D. Tretyakov, Mendeleev
[379] L. A. Harris, M. E. Gerstner, R. H. Wilson, J. Electrochem. Soc. 1979, 126,
Commun. 2005, 15, 178 – 180.
850 – 855.
[419] A. A. Joraid, Curr. Appl. Phys. 2009, 9, 73 – 79.
[380] D. S. Ginley, M. L. Knotek, J. Electrochem. Soc. 1979, 126, 2163 – 2166.
[420] D. Isik, M. Ak, C. Durucan, Thin Solid Films 2009, 518, 104 – 111.
[381] P. Clechet, C. Martelet, R. Olier, J. P. Thomas, M. Fallavier, J. Electrochem.
[421] S. R. Bathe, P. S. Patil, Solid State Ionics 2008, 179, 314 – 323.
Soc. 1983, 130, 1795 – 1796.
[422] N. R. de Tacconi, C. R. Chenthamarakshan, K. L. Wouters, F. M. MacDon-
[382] M. F. Weber, L. C. Schumacher, M. J. Dignam, J. Electrochem. Soc. 1982,
nell, K. Rajeshwar, J. Electroanal. Chem. 2004, 566, 249 – 256.
129, 2022 – 2028.
[423] C.-M. Wang, S.-Y. Lin, Y.-C. Chen, J. Phys. Chem. Solids 2008, 69, 451 –
[383] B. H. Meekins, P. V. Kamat, ACS Nano 2009, 3, 3437 – 3446. 455.
[384] R. Nakamura, T. Tanaka, Y. Nakato, J. Phys. Chem. B 2004, 108, 10617 – [424] P. Delich re, P. Falaras, A. Hugot-Le Goff, Sol. Energy Mater., 1989, 19,
10620. 323 – 333.
[385] B. Neumann, P. Bogdanoff, H. Tributsch, S. Sakthivel, H. Kisch, J. Phys. [425] S. Sallard, T. Brezesinski, B. M. Smarsly, J. Phys. Chem. C 2007, 111,
Chem. B 2005, 109, 16579 – 16586. 7200 – 7206.
[386] H. Liu, A. Imanishi, Y. Nakato, J. Phys. Chem. C 2007, 111, 8603 – 8610. [426] J. Wang, X. W. Sun, Z. Jiao, Materials 2010, 3, 5029 – 5053.
[387] A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo, H. Pettersson, Chem. Rev. [427] T. Gaoa, A. Gustavsen, B. P. Jelleb, Zero emission buildings—proceedings
2010, 110, 6595 – 6663. of Renewable Energy Conference 2010, Trondheim, Norway.
[388] S. Rhle, M. Shalom, A. Zaban, ChemPhysChem 2010, 11, 2290 – 2304. [428] A. Mills, S. Le Hunte, J. Photochem. Photobiol. A 1997, 108, 1 – 35.
[389] L. F. Nazar, G. Goward, F. Leroux, M. Duncan, H. Huang, T. Kerr, J. Gau- [429] A. J. Bard, J. Photochem. 1979, 10, 59 – 75.
bicher, Int. J. Inorg. Mater. 2001, 3, 191 – 200. [430] K. Rajeshwar, J. Appl. Electrochem. 1995, 25, 1067 – 1082.
[390] P. Poizot, S. Laruelle, S. Grugeon, L. Dupont, J. M. Tarascon, J. Power [431] J. M. Kesselman, G. A. Shreve, M. R. Hoffmann, N. S. Lewis, J. Phys.
Sources 2001, 97 – 98, 235 – 239. Chem. 1994, 98, 13385 – 13395.
[391] C. Natarajan, K. Setoguchi, G. Nogami, Electrochim. Acta 1998, 43, [432] H. Rensmo, H. Lindstrçm, S. Sçdergren, A. K. Willstedt, A. Solbrand, A.
3371 – 3374. Hagfeldt, S. E. Lindquist, J. Electrochem. Soc. 1996, 143, 3173 – 3178.
[392] V. Subramanian, A. Karki, K. I. Gnanasekar, F. Posey Eddy, B. Rambabu, [433] R. Solarska, I. Rutkowska, R. Morand, J. Augustynski, Electrochim. Acta
J. Power Sources 2006, 159, 186 – 192. 2006, 51, 2230 – 2236.
[393] B. Zachau-Christiansen, K. West, T. Jacobsen, S. Atlung, Solid State [434] R. Solarska, I. Rutkowska, J. Augustynski, Inorg. Chim. Acta 2008, 361,
Ionics 1988, 28 – 30, 1176 – 1182. 792 – 797.
[394] G. Armstrong, A. R. Armstrong, J. Canales, P. G. Bruce, Electrochem. [435] S. R. Morrison, T. Freund, J. Chem. Phys. 1967, 47, 1543 – 1551.
Solid-State Lett. 2006, 9, A139 – A143. [436] R. Solarska, I. Rutkowska, R. Morand, J. Augustynski, Electrochim. Acta
[395] P. G. Bruce, B. Scrosati, J.-M. Tarascon, Angew. Chem. 2008, 120, 2972 – 2006, 52, 694 – 703.
2989; Angew. Chem. Int. Ed. 2008, 47, 2930 – 2946. [437] W. Wen, H. Zhao, S. Zhang, V. Pires, J. Phys. Chem. C 2008, 112, 3875 –
[396] C. Jiang, M. Wei, Z. Qi, T. Kudo, I. Honma, H. Zhou, J. Power Sources 3880.
2007, 166, 239 – 243. [438] W. Sun, C. R. Chenthamarakshan, K. Rajeshwar, J. Phys. Chem. B 2002,
[397] G. Sudant, E. Baudrin, D. Larcher, J. M. Tarascon, J. Mater. Chem. 2005, 106, 11531 – 11538.
15, 1263 – 1269. [439] A. G. MuÇoz, Electrochim. Acta 2007, 52, 4167 – 4176.

&50& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
The Electrochemistry of Nanostructured TiO2 Electrodes

[440] O. A. Semenikhin, V. E. Kazarinov, L. Jiang, K. Hashimoto, A. Fujishima, [478] C. He, Y. Xiong, J. Chen, C. Zha, X. Zhu, J. Photochem. Photobiol. A
Langmuir 1999, 15, 3731 – 3737. 2003, 157, 71 – 79.
[441] D. Jiang, S. Zhang, H. Zhao, Environ. Sci. Technol. 2007, 41, 303 – 308. [479] H. Chen, S. Chen, X. Quan, H. Yu, H. Zhao, Y. Zhang, J. Phys. Chem. C
[442] P. C. Angelom, L. Andrini, M. E. Calvo, F. G. Requejo, S. A. Bilmes, 2008, 112, 9285 – 9290.
G. J. A. A. Soler-Illia, J. Phys. Chem. C 2007, 111, 10886 – 10893. [480] K. Xie, L. Sun, C. Wang, Y. Lai, M. Wanga, H. Chen, C. Lin, Electrochim.
[443] I. Bilecka, P. J. Barczuk, J. Augustynski, Electrochim. Acta 2010, 55, 979 – Acta 2010, 55, 7211 – 7218.
984. [481] A. A. Ismail, D. W. Bahnemann, J. Rathousky, V. Yarovyi, M. Wark, J.
[444] S. K. Poznyak, A. I. Kokorin, A. I. Kulak, J. Electroanal. Chem. 1998, 442, Mater. Chem. 2011, 21, 7802 – 7810.
99 – 105. [482] H. Yu, X. Quan, S. Chen, H. Zhao, Y. Zhang, J. Photochem. Photobiol. A
[445] H. Hidaka, K. Ajisaka, S. Horikoshi, T. Oyama, K. Takeuchi, J. Zhao, N. 2008, 200, 301 – 306.
Serpone, J. Photochem. Photobiol. A 2001, 138, 185 – 192. [483] B. Gao, C. Peng, G. Z. Chen, G. L. Puma, Appl. Catal. B 2008, 85, 17 – 23.
[446] J. Krýsa, M. Keppert, G. Waldner, J. Jirkovský, Electrochim. Acta 2005, [484] J. Lin, R. Zong, M. Zhou, Y. Zhu, Appl. Catal. B 2009, 89, 425 – 431.
50, 5255 – 5260. [485] L. C. Chen, Y. C. Ho, W. S. Guo, C. M. Huang, T. C. Pan, Electrochim. Acta
[447] H. Kikuchi, M. Kitano, M. Takeuchi, M. Matsuoka, M. Anpo, P. V. Kamat, 2009, 54, 3884 – 3891.
J. Phys. Chem. B 2006, 110, 5537 – 5541. [486] Y. Xie, Adv. Funct. Mater. 2006, 16, 1823 – 1831.
[448] D. Chen, Y. Gao, G. Wang, H. Zhang, W. Lu, J. Li, J. Phys. Chem. C 2007, [487] D. Mitoraj, R. Bernek, H. Kisch, Photochem. Photobiol. Sci. 2010, 9, 31 –
111, 13163 – 13169. 38.
[449] J. Y. Kim, S. B. Choi, D. W. Kim, S. Lee, H. S. Jung, J. K. Lee, K. S. Hong, [488] X. Li, W. Teng, Q. Zhao, L. Wang, J. Nanopart. Res. 2011, 13, 6813 –
Langmuir 2008, 24, 4316 – 4319. 6820.
[450] H. C. Liang, X. Z. Li, J. Hazard. Mater. 2009, 162, 1415 – 1422. [489] X. Zhou, F. Peng, H. Wang, H. Yua, Y. Fang, Chem. Commun. 2011, 47,
[451] H. J. Yun, H. Lee, J. B. Joo, W. Kim, J. Yi, J. Phys. Chem. C 2009, 113, 10323 – 10325.
3050 – 3055. [490] M. Bledowski, L. Wang, A. Ramakrishnan, O. V. Khavryuchenko, V. D.
[452] A. Atyaoui, L. Bousselmi, H. Cachet, P. Pu, E. M. M. Sutter, J. Photochem. Khavryuchenko, P. C. Ricci, J. Strunk, T. Cremer, C. Kolbeck, R. Beranek,
Photobiol. A 2011, 224, 71 – 79. Phys. Chem. Chem. Phys. 2011, 13, 21511 – 21519.
[453] F. Sordello, V. Maurino, C. Minero, J. Mater. Chem. 2011, 21, 19144 – [491] P. S. M. Dunlop, J. A. Byrne, N. Manga, B. R. Eggins, J. Photochem. Pho-
19152. tobiol. A 2002, 148, 355 – 363.
[454] Y. Su, X. Zhang, M. Zhou, S. Han, L. Lei, J. Photochem. Photobiol. A [492] P. A. Christensen, T. P. Curtis, T. A. Egerton, S. A. M. Kosa, J. R. Tinlin,
2008, 194, 152 – 160.
Appl. Catal. B 2003, 41, 371 – 386.
[455] X. Zhang, K. Udagawa, Z. Liu, S. Nishimoto, C. Xu, Y. Liu, H. Sakai, M.
[493] T. A. Egerton, S. A. M. Kosa, P. A. Christensen, Phys. Chem. Chem. Phys.
Abe, T. Murakami, A. Fujishima, J. Photochem. Photobiol. A 2009, 202,
2006, 8, 398 – 406.
39 – 47.
[494] N. Baram, D. Starosvetsky, J. Starosvetsky, M. Epshtein, R. Armon, Y.
[456] Y. Wang, H. Cheng, L. Zhang, Y. Hao, J. Ma, B. Xu, W. Li, J. Mol. Catal. A
Ein-Eli, Electrochim. Acta 2009, 54, 3381 – 3386.
2000, 151, 205 – 216.
[495] N. Philippidis, E. Nikolakaki, S. Sotiropoulos, I. Poulios, J. Chem. Technol.
[457] A. V. Emeline, Y. Furubayashi, X. Zhang, M. Jin, T. Murakami, A. Fujishi-
Biotechnol. 2010, 85, 1054 – 1060.
ma, J. Phys. Chem. B 2005, 109, 24441 – 24444.
[496] T. A. Egerton, J. Chem. Technol. Biotechnol. 2011, 86, 1024 – 1031.
[458] B. Xin, Z. Ren, P. Wang, J. Liu, L. Jing, H. Fu, Appl. Surf. Sci. 2007, 253,
[497] M. V. Boldrin Zanoni, J. J. Sene, H. Selcuk, M. A. Anderson, Environ. Sci.
4390 – 4395.
Technol. 2004, 38, 3203 – 3208.
[459] K. Esquivel, M. G. Garc a, J. F. J. Rodr guez, M. Vega Gonzlez, L. Esco-
[498] J. Nemoto, N. Gokan, H. Ueno, M. Kaneko, J. Photochem. Photobiol. A
bar-Alarcn, L. Ortiz-Frade, L. A. God nez, J. Nanopart. Res. 2011, 13,
2007, 185, 295 – 300.
3313 – 3325.
[499] F. M. Monteiro Paschoal, G. Pepping, M. V. Boldrin Zanoni, M. A. Ander-
[460] N. Lu, X. Quan, J. Y. Li, S. Chen, H. T. Yu, G. H. Chen, J. Phys. Chem. C
son, Environ. Sci. Technol. 2009, 43, 7496 – 7502.
2007, 111, 11836 – 11842.
[500] F. M. Monteiro Paschoal, M. A. Anderson, M. V. Boldrin Zanoni, Desali-
[461] G. Wang, H. Wang, Y. Ling, Y. Tang, X. Yang, R. C. Fitzmorris, C. Wang,
nation 2009, 249, 1350 – 1355.
J. Z. Zhang, Y. Li, Nano Lett. 2011, 11, 3026 – 3033.
[462] X. F. Cheng, W. H. Leng, D. P. Liu, Y. M. Xu, J. Q. Zhang, C. N. Cao, J. [501] D. Jiang, H. Zhao, S. Zhang, R. John, J. Photochem. Photobiol. A 2006,
Phys. Chem. C 2008, 112, 8725 – 8734. 177, 253 – 260.
[463] R. Bernek, B. Neumann, S. Sakthivel, M. Janczarek, T. Dittrich, H. Trib- [502] H. Zhao, D. Jiang, S. Zhang, W. Wen, J. Catal. 2007, 250, 102 – 109.
utsch, H. Kisch, Chem. Phys. 2007, 339, 11 – 19. [503] I. M. Butterfield, P. A. Christensen, A. Hamnett, K. E. Shaw, G. M. Walker,
[464] K. Vinodgopal, P. V. Kamat, Environ. Sci. Technol. 1995, 29, 841 – 845. S. A. Walker, C. R. Howarth, J. Appl. Electrochem. 1997, 27, 385 – 395.
[465] M. G. Kang, H. E. Han, K. J. Kim, J. Photochem. Photobiol. A 1999, 125, [504] R. J. Candal, W. A. Zeltner, M. A. Anderson, Environ. Sci. Technol. 2000,
119 – 125. 34, 3443 – 3451.
[466] Y. Shaogui, Q. Xie, L. Xinyong, L. Yazi, C. Shuoa, C. Guohua, Phys. Chem. [505] H. Hidaka, H. Nagaoka, K. Nohara, T. Shimura, S. Horikoshi, J. Zhao, N.
Chem. Phys. 2004, 6, 659 – 664. Serpone, J. Photochem. Photobiol. A 1996, 98, 73 – 78.
[467] S. Kuang, L. Yang, S. Luo, Q. Cai, Appl. Surf. Sci. 2009, 255, 7385 – 7388. [506] D. H. Kim, M. A. Anderson, J. Photochem. Photobiol. A 1996, 94, 221 –
[468] F. Fabregat-Santiago, E.M. Barea, J. Bisquert, G.K. Mor, K. Shankar, C.A. 229.
Grimes, J. Am. Chem. Soc. 2008, 130, 11312 – 11316. [507] J. A. Byrne, B. R. Eggins, W. Byers, N. M. D. Brown, Appl. Catal. B 1999,
[469] Y. Hou, X. Li, X. Zou, X. Quan, G. Chen, Environ. Sci. Technol. 2009, 43, 20, L85 – L89.
858 – 863. [508] H. Hidaka, T. Shimura, K. Ajisaka, S. Horikoshi, J. Zhao, N. Serpone, J.
[470] Y. Hou, X. Li, Q. Zhao, X. Quan, G. Chen, Environ. Sci. Technol. 2010, 44, Photochem. Photobiol. A 1997, 109, 165 – 170.
5098 – 5103. [509] Z. Liu, X. Zhang, S. Nishimoto, M. Jin, D. A. Tryk, T. Murakami, A. Fujishi-
[471] W. Zhu, X. Liu, H. Liu, D. Tong, J. Yang, J. Peng, J. Am. Chem. Soc. 2010, ma, J. Phys. Chem. C 2008, 112, 253 – 259.
132, 12619 – 12626. [510] H. G. Oliveira, D. C. Nery, C. Longo, Appl. Catal. B 2010, 93, 205 – 211.
[472] J. Zhang, J. H. Bang, C. Tang, P. V. Kamat, ACS Nano 2010, 4, 387 – 395. [511] M. Zhou, X. Ma, Electrochem. Commun. 2009, 11, 921 – 924.
[473] H. I. Kim, J. Kim, W. Kim, W. Choi, J. Phys. Chem. C 2011, 115, 9797 – [512] A. Taghizadeh, M. F. Lawrence, L. Miller, M. A. Anderson, N. Serpone, J.
9805. Photochem. Photobiol. A 2000, 130, 145 – 156.
[474] H. Zhao, W. Fu, H. Yang, Y. Xu, W. Zhao, Y. Zhang, H. Chen, Q. Jing, X. [513] Y. Xie, Electrochim. Acta 2006, 51, 3399 – 3406.
Qi, J. Cao, X. Zhou, Y. Li, Appl. Surf. Sci. 2011, 257, 8778 – 8783. [514] J. M. Kesselman, N. S. Lewis, M. R. Hoffmann, Environ. Sci. Technol.
[475] X. Zhao, H. Liu, J. Qu, Appl. Surf. Sci. 2011, 257, 4621 – 4624. 1997, 31, 2298 – 2302.
[476] S. Chai, G. Zhao, P. Li, Y. Lei, Y. N. Zhang, D. Li, J. Phys. Chem. C 2011, [515] X. Quan, S. Yang, X. Ruan, H. Zhao, Environ. Sci. Technol. 2005, 39,
115, 18261 – 18269. 3770 – 3775.
[477] T. H. Jeon, W. Choi, H. Park, J. Phys. Chem. C 2011, 115, 7134 – 7142. [516] S. Yang, Y. Liu, C. Sun, Appl. Catal. A 2006, 301, 284 – 291.

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &51&
These are not the final page numbers! ÞÞ
 R. Gmez et al.

[517] M. Tian, G. Wu, B. Adams, J. Wen, A. Chen, J. Phys. Chem. C 2008, 112, [539] P. A. Christensen, T. A. Egerton, S. A. M. Kosa, J. R. Tinlin, K. Scott, J.
825 – 831. Appl. Electrochem. 2005, 35, 683 – 692.
[518] B. Su, Y. Ma, Y. Du, C. Wang, Electrochem. Commun. 2009, 11, 1154 – [540] Y. Xu, Y. He, X. Cao, D. Zhong, A. Jia, Environ. Sci. Technol. 2008, 42,
1157. 2612 – 2617.
[519] J. Carvalho Cardoso, T. Mescoloto Lizier, M. V. Boldrin Zanoni, Appl. [541] Y. Xu, J. Jia, D. Zhong, Y. Wang, Chem. Eng. J. 2009, 150, 302 – 307.
Catal. B 2010, 99, 96 – 102. [542] J. Bai, Y. Liu, J. Li, B. Zhou, Q. Zheng, W. Cai, Appl. Catal. B 2010, 98,
[520] Y. Xin, H. Liu, L. Han, Y. Zhou, J. Hazard. Mater. 2011, 192, 1812 – 1818. 154 – 160.
[521] K. Vinodgopal, P. V. Kamat, Sol. Energy Mater. Sol. Cells 1995, 38, 401 – [543] A. Zhang, M. Zhou, L. Liu, W. Wang, Y. Jiao, Q. Zhou, Electrochim. Acta
410. 2010, 55, 5091 – 5099.
[522] M. Zlamal, J. M. Macak, P. Schmuki, J. Krýsa, Electrochem. Commun. [544] D. Bahnemann, Sol. Energy 2004, 77, 445 – 459.
2007, 9, 2822 – 2826. [545] S. Malato, J. Blanco, D. C. Alarcn, M. I. Maldonado, P. Fernndez-
[523] J. Yang, C. Chen, H. Ji, W. Ma, J. Zhao, J. Phys. Chem. B 2005, 109, IbÇez, W. Gernjak, Catal. Today 2007, 122, 137 – 149.
21900 – 21907. [546] R. J. Braham, A. T. Harris, Ind. Eng. Chem. Res. 2009, 48, 8890 – 8905.
[524] A. Turolla, M. Fumagalli, M. Bestetti, M. Antonelli, Desalination 2012, [547] C. McCullagh, N. Skillen, M. Adams, P. K. J. Robertson, J. Chem. Technol.
285, 377 – 382. Biotechnol. 2011, 86, 1002 – 1017.
[525] T. T. Guaraldo, S. H. Pulcinelli, M. V. Boldrin Zanoni, J. Photochem. Photo- [548] K. Kalyanasundaram, M. Grtzel, Curr. Opin. Biotechnol. 2010, 21, 298 –
biol. A 2011, 217, 259 – 266. 310.
[526] T. C. An, X. H. Zhu, Y. Xiong, Chemosphere 2002, 46, 897 – 903. [549] J. Augustyr ski, B. D. Alexander, R. Solarska, Top. Curr. Chem. 2011, 303,
[527] X. Wu, Y. Ling, L. Liu, Z. Huang, J. Electrochem. Soc. 2009, 156, K65 – 1 – 38.
K71. [550] S. Caramori, V. Cristino, L. Meda, R. Argazzi, C. A. Bignozzi, Top. Curr.
[528] Z. Zhang, Y. Yuan, G. Shi, Y. Fang, L. Liang, H. Ding, L. Jin, Environ. Sci. Chem. 2011, 303, 39 – 94.
Technol. 2007, 41, 6259 – 6263. [551] P. Hartmann, D.-K. Lee, B. M. Smarsly, J. Janek, ACS Nano 2010, 4,
[529] A. Kar, Y. R. Smith, V. Subramanian, Environ. Sci. Technol. 2009, 43, 3147 – 3154.
3260 – 3265. [552] Z. Su, W. Zhou, J. Mater. Chem. 2011, 21, 8955 – 8970.
[530] Y. R. Smith, A. Kar, V. Subramanian, Ind. Eng. Chem. Res. 2009, 48, [553] G. Centi, S. Perathoner, Catal. Today 2009, 148, 191 – 205.
10268 – 10276. [554] T. Yui, Y. Tamaki, K. Sekizawa, O. Ishitani, Top. Curr. Chem. 2011, 303,
[531] Y. S. Sohn, Y. R. Smith, M. Misra, V. Subramanian, Appl. Catal. B 2008, 151 – 184.
84, 372 – 378. [555] S. C. Roy, O. K. Varghese, M. Paulose, C. A. Grimes, ACS Nano 2010, 4,
[532] P. A. Carneiro, M. E. Osugi, J. J. Sene, M. A. Anderson, M. V. B. Zanoni, 1259 – 1278.
Electrochim. Acta 2004, 49, 3807 – 3820. [556] P. Usubharatana, D. McMartin, A. Veawab, P. Tontiwachwuthikul, Ind.
[533] Z. Jiang, H. Wang, H. Huang, C. Cao, Chemosphere 2004, 56, 503 – 508. Eng. Chem. Res. 2006, 45, 2558 – 2568.
[534] G. Waldner, J. Krýsa, Electrochim. Acta 2005, 50, 4498 – 4504. [557] T. Inoue, A. Fujishima, S. Konichi, K. Honda, Nature 1979, 277, 637 –
[535] P. Christensen, G. Hutson, Photoelectrochemical Reactor, European 638.
Patent 9592 1936.1, 1996. [558] C. Ampelli, G. Centi, R. Passalacqua, S. Perathoner, Energy Environ. Sci.
[536] P. Fernandez-IbaÇez, S. Malato, O. Enea, Catal. Today 1999, 54, 329 – 2010, 3, 292 – 301.
339.
[537] J. Krýsa, J. Jirkovský, J. Appl. Electrochem. 2002, 32, 591 – 596.
[538] T. An, Y. Xiong, G. Li, C. Zha, X. Zhu, J. Photochem. Photobiol. A 2002, Received: January 29, 2012
152, 155 – 165. Published online on && &&, 2012

&52& www.chemphyschem.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 0000, 00, 1 – 53
ÝÝ These are not the final page numbers!
REVIEWS
Ubiquitous titania electrodes: The T. Berger, D. Monllor-Satoca,
electrochemistry of nanostructured tita- M. Jankulovska, T. Lana-Villarreal,
nium dioxide electrodes is reviewed R. Gmez*
with a focus on the fundamental as-
&& – &&
pects that determine their behavior
both in the dark and under illumination The Electrochemistry of
(see picture). Some applications in the Nanostructured Titanium Dioxide
fields of environmental remediation Electrodes
(heterogeneous photocatalysis), and
energy saving (electrochromism), gener-
ation (artificial photosynthesis) and ac-
cumulation (Li-ion batteries) are also
dealt with.

ChemPhysChem 0000, 00, 1 – 53  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemphyschem.org &53&
These are not the final page numbers! ÞÞ